CHARACTERISATION AND IN VITRO DRUG SENSITIVITIES OF

HUMAN TRANSITIONAL CELL CARCINOMA CELL LINES:

EVALUATION OF THEIR USE AS A MODEL SYSTEM FOR BLADDER CANCER

by

PETER JOHN HEPBURN

A thesis submitted to the Faculty of Medicine for

the Degree of Doctor of Philosophy

Institute of Urology

University College London

April 1990

ProQuest Number: U051114

All rights reserved

INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a com p le te manuscript and there are missing pages, these will be noted. Also, if material had to be removed,

a note will indicate the deletion.

uestProQuest U051114

Published by ProQuest LLC(2017). Copyright of the Dissertation is held by the Author.

All rights reserved.This work is protected against unauthorized copying under Title 17, United States C ode

Microform Edition © ProQuest LLC.

ProQuest LLC.789 East Eisenhower Parkway

P.O. Box 1346 Ann Arbor, Ml 48106- 1346

ABSTRACT

The concept that continuous cell lines derived from one

histological type of tumour provide an in vitro model

system for that disease was examined. Biological

characteristics and in vitro drug sensitivities of

lines derived from transitional cell carcinomas (TCC)

of the human bladder are described.

The isozyme phenotype, cytological appearance, cell

kinetics, colony-forming efficiency, tumorigenicity and

monoclonal antibody reactivities of twenty-two

urothelial lines were assessed. Three types were

identified; 1) fourteen distinct lines derived from

different patients, 2) five cross-contaminated with

one of the distinct lines (T24) and 3) three

non-tumorigenic lines. Xenograft morphology and

isozyme pattern were used to select representative TCC

cell lines.

Ten clonogenic assay procedures for measuring

cytotoxicity were compared using one cell line (RT112)

exposed to adriamycin and methotrexate. Significant

differences between dose-response curves were obtained

for the same drug indicating methods should be

standardised.

The relative cytotoxicities of a 24h exposure to twelve

chemotherapeutic drugs against RT112 cells were

compared using a clonogenic assay. Drug concentrations

reducing clonogenic cell survival by 10% ranged from

2.9 ng/ml for mitoxantrone to 27.0 pg/ml for

hydroxyanisole.

The range and reproducibility of in vitro sensitivities

to adriamycin and methotrexate were compared using

eight distinct lines and four lines cross-contaminated

with T24. The cross-contaminated lines had similar

sensitivities, demonstrating stability during long-term

culture. However there was a broad spectrum of

sensitivities (>50 fold) amongst the distinct lines to

methotrexate and a narrow range (<3 fold) to

adriamycin.

In summary, these data show that cell lines derived

from human bladder tumours can retain certain

biological characteristics of the tumours of origin.

Taken in conjunction with published data, these results

indicate that cell lines also retain the pattern of

drug sensitivities of the tissue of origin, and if the

differences in pharmacokinetics in vivo can be taken

into account, might be used for screening new agents.

DEDICATION

This thesis is dedicated to my mother, my daughters Lucy and Sophie and to a very special friend, Michele.

ACKNOWLEDGEMENTS

This study was carried out at St Paul's Hospital (Institute of Urology) and the Cellular Chemotherapy Laboratory, Imperial Cancer Research Fund (where I was appointed as an honorary visiting worker).I am grateful for the use of facilities, equipment and collaborationwith many members of both establishments enabling me to perform thiswork.

My major acknowledgement must go to my supervisor, Dr John Masters, for his guidance, advice and willingness to discuss my project at alltimes. My thanks also to Dr Bridget Hill, Head of the Department ofCellular Chemotherapy, for her interest in my work.

I am grateful to many colleagues at ICI Pharmaceuticals for their constructive comments on this manuscript and members of the Department of Pathology are thanked for making St Paul's such an enjoyable place in which to work.

I am extremely grateful to my typist for all her time, effort and patience spent on deciphering my handwriting and so professionally typing this manuscript.

Finally, I must thank Michele who deserves special acknowledgement for her kindness and understanding while writing this manuscript and without whose support completion would not have been possible.

TABLE OF CONTENTS

PAGE

TITLE PAGE 1

ABSTRACT 2*

DEDICATION 4

ACKNOWLEDGEMENTS 5

TABLE OF CONTENTS 6

ABBREVIATIONS 8

LIST OF FIGURES 10

LIST OF TABLES 14

CHAPTER 1 18General Introduction

CHAPTER 2 107Materials and Methods

CHAPTER 3 164Characterisation of Human Urothelial CellLines

CHAPTER 4 194Comparison of Clonogenic Assay Procedures for Measuring the Drug Sensitivities in vitro of Human Tumour Cell Lines

CHAPTER 5 217Cytotoxic Activities of Chemotherapeutic Drugs and the Evaluation of Drug Combinations with Spirogermanium Using the Human Bladder Cancer Cell Line, RTU2

PAGE

CHAPTER 6In Vitro Drug Sensitivities of a Panel of Continuous Cell Lines Derived from Human Bladder Cancer to Adriamycin and Methotrexate: Range and Stability in Long-Term Culture

CHAPTER 7General Discussion and Summary

253

279

REFERENCES 299

ABBREVIATIONS

CEA - Carcinoembryonic antigen

AFP - Alphafetoprotein

ATCH - Adrenocorticotropic hormone

ADH - Antidiuretic hormone

PTH - Parathormone

HCG - Human chorionic gonadotropin

HPL - Human placental lactogen

PAP - Prostatic acid phosphatase

LDH - Lactic dehydrogenase

SRC - Subrenal capsule assay

ADC - Agar diffusion chamber assay

CMI - Cytotoxic metabolic inhibition test

SDI - Succinic dehydrogenase inhibition test

FMF - Flow microfluoriraetry

HTSCA - Human tumour stem cell assay

TCC - Transitional cell carcinoma

TUR - Trans urethral resection

GR - Growth fraction

CFE - Colony forming efficiency

PDT - Population doubling time

LI - Labelling index

MR - Mitotic rate

ADR - Adriamycin

BLM - Bleomycin

Cisplatin - Cis-dichlorodiammine platinum II

CDBDCA - Cis-diammine-1,1-cyclobutane

decarboxylate platinum II

DBD Dibromodulci tol

5-FU 5-Fluorouracil

4-OHA 4-Hydroxyanisole

MTX Methotrexate

Mit C Mitomycin C

MX Mitoxantrone

Sp Spirogermanium

VP16-213 - 4-demethyl epipodophyllotoxin-£-D-(Etoposide) ethylidene glucoside

LIST OF FIGURES

Figure 1.1Diagramatic representation of the UICC 1978 histopathological classification of human bladder cancer.

Figure 1.2 The cell cycle.

Figure 1.3Survival of mammalian cell reproductive capacity after 24 hour exposure to antitumour agents.

Figure 1.4Five patterns of cell survival response following treatment of a human lymphoma cell line (T ) with anti-tumour drugs for 1 hour in vitro.

Figure 2.1Pathways for metabolism of 5-FU.

Figure 2.2Mechanism of action of methotrexate.

Figure 2.3Diagramatic representation of starch gel electrophoresis.

Figure 2.4Comparison of growth rate curves for the cell line RT112 seeded in 35mm petri dishes and 96 well flat bottom micro test plates.

Figure 2.5Growth curve of the cell line RT112.

Figure 3.1Comparison of original patient biopsy and xenograft of resultant cell lines, RT4 and RT112.

Figure 3.2Histological section (H & E) of tumours formed in nude mice as a result of subcutaneous injection of each cell line.

Figure 3.3Immunofluorescence micrographs shoving characteristic cell-surface and intracellular immunolocalisation patterns seen with the LBS- series monoclonal antibodies on human bladder cell lines.

Figure 4.1Clonogenic assay procedures used to compare in vitro drug sensitivities of human tumour cell lines.

Figure 4.2Comparison of dose response curves obtained using various procedures after a 24 hour exposure to methotrexate or adriamycin.

Figure 4.3Comparison of the fraction of labelled cells plotted against time of exposure

3to [methyl H] thymidine, with the cell densities equivalent to high density transfer, low density transfer, in situ assay.

PAGE

Figure 5.1 (i-xii) 233Clonogenic cell survival of RT112 against a range of chemotherapeutic drugs.

Figure 5.2 246Effective chemotherapeutic drug concentration required to reduce clonogenic cell survival from 50 - 90%.

Figure 5.3 251Lethal effects of a 24 hour exposure to combinations of spirogermanium (1.28yg/ml) with cisplatin, 5-fluorouracil, adriamycin and methotrexate on RT112 cells.

Figure 6.1 271Effective adriamycin concentrations required to reduce clonogenic cell survival from 50 - 90%.

Figure 6.2 272Effective methotrexate concentrations required to reduce clonogenic cell survival from 50 - 90%.

Figure 6.3 273Clonogenic cell survival of a panel ofeight distinct human bladder cancercontinuous cell lines shown on a logarithmicscale against a range of concentrations ofadriamycin plotted in ngml” on a linear scale.

PAGE

Figure 6.4 274Clonogenic cell survival of a panel of eightdistinct human bladder cancer continuous celllines shown on a logarithmic scale against arange of concentrations of methotrexate plottedin ngml~ on a linear scale.

Figure 6.5 275Clonogenic cell survival of T24 and the fourcross-contaminated sublines shown on alogarithmic scale against a range ofconcentrations of adriamycin plotted in ngml”on a linear scale.

Figure 6.6 276Clonogenic cell survival of T24 and thefour cross-contaminated sublines shown on alogarithmic scale against a range ofconcentrations of methotrexate plotted inngml” on a linear scale.

Figure 6.7 277Scatter plots for the eight distinct humanbladder cancer continuous cell lines showing thelack of correlation between LI and PDT againstthe ID70 values for adriamycin.

Figure 6.8 278Scatter plots for the eight distinct humanbladder cancer continuous cell lines showing thelack of correlation between LI and PDT againstthe ID70 values for methotrexate.

LIST OF TABLES

Table 1.1Comparison of the 1978 UICC, TNM, St Peter's group proposed and Jewett-Strong-Marshall, histopathological classification of human bladder cancer.

Table 1.2Classification of tumour markers.

Table 1.3Advantages and disadvantages of in vivo model systems.

Table 1.4Major advantages and limitations associated with in vitro model systems.

Table 1.5Chemosensitivity studies of carcinoma of the human bladder using the HTSCA.

Table 1.6Kinetic classification of anti tumour drugs used in these studies.

Table 1.7Phases of the cell cycle when certain anti-tumour agents exert their effects.

Table 2.1RPMI 1640 medium.

Table 2.2Trypsin 0.25% in Tris saline.

PAGE

26

33

44

61

70

101

102

112

116

Table 2.3Versene (EDTA) 0.02% in PBS (solution A).

Table 2.4Baker's formol-calcium.

Table 2.5Chrom alum/gelatin.

Table 2.6Bridge buffers used for starch gell electro­phoresis.

Table 2.7Acrylamide gel for isoelectric focusing.

Table 2.8Chemotherapeutic drugs.

Table 2.9Origin of cell lines.

Table 2.10Constituents to make PPLO agar.

Table 2.11Effect of various concentrations and volumes of a 1:5 trypsin/versene mixture on the secondary colony forming efficiency of RT112 cells.

Table 3.1Isozyme profiles of human bladder cell lines.

Table 3.2Cell line morphology and tumorigenicity.

PAGE

Table 3.3 185Reactivity of the human bladder cell line panel with the LBS-series monoclonal antibodies.

Table 3.4 186Growth characteristics of the human bladder cell line panel.

Table 3.5 187Estimates of cell loss rate and intermitotic times of HT1197 and T24 followed for two cell generations by time-lapse cinemicroscopy.

Table 4.1 214A comparison of in vitro drug concentrations requested to reduce survival by 70% (ID70), depending on the assay procedure used.

Table 4.2 215The labelling indices of cells plated at

3 4densities of 1 x 10 , 6.1 x 10 and5 27.8 x 10 /5cm dish, equivalent to the cell

concentrations used for the in situ, lowand high cell density transfer assays respectively

3and exposed to H-thymidine continuously for periods ranging from 15 minutes to 96 hours.

Table 5.1 247The concentration of each drug, expressed as(a) yg/ml (b) molar concentrations, required to reduce clonogenic cell survival by 50%(ID50), 70% (ID70) and 90% (ID90).

PAGE

Table 5.2 249A summary of some of the published datadescribing the pharmacokinetics in patients ofdrugs used in this study including the dosesand routes of administration, measurementsof the ranges of peak plasma concentrations,concentration x time characteristics andterminal half-lives.

Table 5.3 250Comparison of in vitro concentrations required to reduce clonogenic cell survival by 70% (ID70) with C x T values obtained from clinical studies.

Table 6.1 268Comparison of the dose of adriamycin required to reduce the clonogenic cell survival by 70% (ID70) of a panel of human bladder cancer cell lines after a 24 hour exposure.

Table 6.2 269Comparison of the dose of methotrexate required to reduce the clonogenic cell survival by 70%(ID70) of a panel of human bladder cancer cell lines after a 24 hour exposure.

Table 6.3 270Comparison of the dose of adriamycin/methotrexate required to reduce the clonogenic cell survival by 70% (ID70) with the population doubling time (PDT) and labelling index (LI) of eight distinct human bladder cancer cell lines.

CHAPTER I. GENERAL INTRODUCTION

PAGE22

2328

29

31

3132343637384040

4041

41

42424345464848

CONTENTS

RATIONALE OF THESISHUMAN BLADDER CANCER? CLINICAL BACKGROUND CYTOTOXIC DRUG TESTING IN VITRO? HISTORICAL BACKGROUNDHUMAN CONTINUOUS'CELL LINES AS MODELS OF DISEASE; CRITERIAEXTENT TO WHICH HUMAN CONTINUOUS CELL LINES FULFIL THESE CRITERIA; RETENTION OF CHARACTERISTICS OF THE PARENT TUMOUR(A) Isozymes(B) Biochemical characteristics(C) Chromosomal studies (karyotyping)(D) Tumorigenicity(E) Growth characteristics(F) Morphology(G) Sensitivity to chemotherapeutic drugs

(i) Use of patients' individual tumourcells

(ii) Use of continuous cell lines (iii) Drug sensitivity assays using

continuous cell lines (iv) In vitro/in vivo comparisons

DRUG SCREENING(A) Drug development(B) Experimental systems in vivo

(i) Animal models(ii) Human tumour xenografts(iii) Subrenal capsule (SRC) assay(iv) Agar diffusion chamber (ADC) assay

PAGE(C) Experimental systems in vitro 49

(i) Types of in vitro assay; brief 51description(a) . Morphological assessment 51(b) Cellular attachment 52(c) Cell volume analysis 52(d) Cytotoxic metabolic inhibition 52

(CMI) test(e) Measurement of respiration and 53

glycolysis(f) Succinic dehydrogenase 53

inhibition (SDI) test(g) Growth kinetic determinations 54(h) Flow microfluorimetry (FMF) 54(i) Dye exclusion assays 55(j) Uptake of radioactive precursors 56(k) Organ culture 56(1) Multicellular tumour spheroids 57(m) Clonogenic assay 57

(ii) The uses and limitations of in vitro 59models

(D) Factors affecting measurements of in vitro drug 72sensitivities using cell lines(i) Cell cycle 73(ii) Culture conditions 76

(a) Techniques 76(b) Media 78

(iii) Serum and hormones 80(a) Serum 80(b) Hormones 82

(iv) Scheduling and period of exposure 83(v) Growth status (log/plateau) and density 85(vi) Feeder cells 86(vii) Proteolytic enzymes 87

PAGE

(viii) Hydrogen ion concentration (pH) 89(ix) Oxygen 91(x) Temperature 93(xi) Concentration of agar 94

(E) Relationship between in vivo and in vitro 95cytotoxicity(i) Drug storage and stability 95(ii) Drug metabolism 97(iii) Classification of antitumour agents 97(iv) Pharmacokinetics 104

1.7 APPLICATIONS AND FUTURE PROSPECTS 105

1.1 RATIONALE OF THESIS

The aims of this thesis were (i) to evaluate the use of continuous cell lines derived from human transitional cell carcinoma as a representative model of this disease and(ii) assess the extent to which these cell lines are a valid in vitro model for pre—clinical screening of chemotherapeutic drugs. While the clinical trial is the only acceptable means of evaluating new drugs or treatment schedules in patients, the numbers of new agents and possible therapeutic permutations has far exceeded the rate at which such trials can be conducted.Thus, the assessment of new drugs is both a slow and empirical process which so far has been hampered by the lack of a proven model system.

1.2 HUMAN BLADDER CANCER; CLINICAL BACKGROUND

The majority of urinary tract tumours seen in the UK are of the transitional cell type with pure squamous cell carcinoma of the bladder only accounting for between 1 — 2% of urothelial neoplasms (Costello et al, 1984). However, this incidence rate varies throughout the world. For example, squamous cell carcinoma is the predominant type of bladder cancer in areas where Schistosomiasis is endemic, such as Egypt (Webb, 1984).

The highest incidence of transitional cell carcinoma (TCC) of the bladder occurs between the fifth and seventh decades, with that in males approximately threefold that in females. In the United States of America, carcinoma of the urinary bladder ranks seventh in males and thirteenth in females among the causes of death due to cancer, with approximately 45,000 new cases and over 11,000 deaths per annum (Silverberg and Lubera, 1987). In England and Wales there is an incidence of approximately 8,000 new patients (Tate et al, 1979), with an annual death rate of over 4,000 (Office of population censuses and surveys; Mortality statistics, 1982). There has been little reduction in this mortality rate despite the fact that the majority (approximately 80%) of patients present with superficial tumours (Lum, 1983; Prout, 1984). These tumours fall into the Jewett—Strong-Marshall classification of Stage 0 or A tumours and would be classified at the clinico-histopathological stage pTis, pTa or pTl, in the TNM system (Jewett and Strong, 1946; Marshall, 1952; UICC, 1978,Pugh, 1981). However, approximately two—thirds of these tumours after treatment developed recurrences (Lum, 1983), of which, approximately 20% recurred at a higher grade or stage (Gilbert et al, 1978). Different histopathological classifications of human bladder tumours are listed in Table 1.1 with a diagramatic representation of the UICC 1978 classification shown in Figure 1.1.

Superficial disease can be controlled by transurethral resection (TUR). However, until recently, the choice of treatment for the patient with advanced bladder cancer has been limited to radiotherapy or surgery, or a combination of both (recently reviewed, Grossman, 1979; Smith, 1984). A number of clinical reports indicate that preoperative radiotherapy and cystectomy offers a 5 year survival of 40 — 50%, compared to 20 — 30% for radical radiotherapy alone (Smith, 1981). For Stage T4 patients however, these survival figures are greatly reduced, with fewer than 10% of patients surviving 5 years (Edsmyr,1981).

Intravesical therapy of superficial bladder tumours began as long ago as 1903, with Herring's report on the use of silver nitrate (Herring, 1903). Other early studies included the use of trichloroacetic acid (Joseph, 1919) and podophyllin (Semple,1948; Duckworth, 1950). In the early 1960s a number of new cytotoxic drugs became available. These were instilled directly into the bladder in an attempt to treat superficial bladder cancer (reviewed by Lum, 1983). One of the first drugs to be used in this manner was thiotepa (Jones and Swinney, 1961;Veenema et al, 1962; Abbassian and Wallace, 1966). A number of other cytotoxic drugs, in addition to thiotepa, are now used topically in the intravesical treatment of superficial papillary tumours (Grossman, 1979; Smith, 1981; Prout and Hagen, 1982; Lum, 1983; Prout, 1984; Pavone-Maculuso et al, 1984). Those agents that appear to be most effective are adriamycin, mitomycin C, thiotepa and epodyl. A number of publications have provided evidence to suggest that dimethyl sulphoxide (DMS0) can increase the anticancer activity of chemotherapeutic drugs (Warren et al, 1975; Thuning et al, 1983). However, an in vitro investigation to test this hypothesis for human bladder cancer demonstrated that the addition of DMS0 to the four drugs most frequently used for intravesical chemotherapy did not increase tumour cell kill (Walker et al, 1988). In addition to chemotherapy for the

treatment of superficial bladder cancer a number of publications have reported success using an immunological approach. Such an approach using Calmette-Guerin bacillus (BCG) was initiated in 1976 (Morales et al, 1976). More recently intravesical administration of BCG, after conventional chemotherapeutic drugs have failed, has resulted in a reduction of the recurrence index and increased the disease—free interval (Soloway and Perry, 1987; Llopis et al, 1989; Steg et al, 1989).

The most common cytotoxic drugs used systemically as single agents for advanced disease in the early studies of the 1960s were 5—fluorouracil (Wilson, 1960; Glenn et al, 1963), cyclophosphamide (Fox, 1965) and methotrexate (Elliot, 1962;Burn et al, 1966). The role of systemic chemotherapy in the management of bladder cancer, particularly metastatic disease, has substantially Increased within the last decade.

Currently, the cytotoxic drugs which appear to be the most effective as single agents for advanced disease are cisplatln, adriamycin, methotrexate and mitomycin C (deKernlon, 1977;Yagoda, 1980; Smith, 1981,1984; Harker and Torti, 1983; Corder et al, 1984; Edsmyr et al, 1984; Oliver et al, 1984). Other agents, cited in these publications, which may also show possible activity as single agents are cyclophosphamide, 5—fluorouracil, hexamethylmelamine, vincristine and neocarzinostatin. However, studies from these reports, collectively, also show that combination chemotherapy has generally been found to be more effective than single agents in inducing remissions and increasing the 5 year survival figures. The use of neoadjuvant M—VAC (methotrexate, vinblastine, adriamycin and cisplatin) therapy has been widely reported (Soloway, 1988; Nakagawa et al, 1988; Sternberg et al, 1988; Scher et al, 1988; Valente et al, 1989). While success has been shown for selected patients with advanced urothelial transitional cell carcinoma, this therapy seems to have limited antitumour action against non transitional

Table 1.1 Comparison of the 1978 UICC* TNM, St Peter's group proposed, and Jewett—Strong—Marshall, histopathological classifications of human bladder cancer

1978 UICC* TNM St Peter's Group

Jewett—Strong- Marshall

pTis Flat pre—invasive pTa Papillary non-invasive pTl Lamina propria )

)pT2 Superficial muscle pT3a Deep muscle

b Through muscle

Pis in situ PaPla stromal corePlbP2P3aP3b perivesical

*1B2c

Fixed to or invading: P4aapT4 a Prostate, uterus, vagina **P4a P4ab

b Pelvic/abdominal wall P4b D1

j)2 Distant metastases

* Union Internationale Contre le Cancrum (UICC)

4aa ducts and acini only** P4a tumours involving prostate

interstitial tissue

Figure 1.1 Diagramatic representation of the UICC8 1978 histopathological classification of human bladder cancer

M uscle coat

* Union Internationale Contre le Cancrum (UICC)

cell cancers (Sternberg et al, 1988). In addition, toxicity of the regimen has been shown to be severe by Tannock et al (1989) and they suggest caution in the widespread application of this protocol. Successful therapy of unresectable urothelial tumours has also been reported with another cisplatin—based drug combination, CISCA, (clsplatin, cyclophosphamide and adriamycin) (Logothetis et al, 1989). However, as shown for M—VAC, the greatest success was shown for patients with pure transitional cell cancer.

1.3 CYTOTOXIC DRUG TESTING IN VITRO; HISTORICAL BACKGROUND

Tissue culture techniques have advanced rapidly since the report by Harrison (1907), who demonstrated the outgrowth of nerves from tadpole spinal cord tissues in hanging-drop culture. Harrison*s work prompted many other investigators to improve on the techniques of tissue culture. Successful in vitro growth of human tumours dates back to 1911 with the reported growth of human sarcomata and carcinomata (Carrel and Burrows, 1911). The first human urinary bladder tumour cultured in vitro was reported by Burrows et al, (1971), who used a substrate of clotted human plasma and agar.

Techniques of cell culture continued to advance rapidly and by 1955 Eagle reported the minimal nutritional requirements essential for optimal growth of mammalian cells in vitro. This was followed by reports on a quantitative method for measuring the cytotoxic action of drugs in tissue culture using standardised experimental conditions (Eagle and Foley, 1956, 1958). There are now many reports in the literature citing the use of cell lines for screening potential anticancer agents (see, Cobb et al, 1961; Ambrose and Easty, 1963; Foley and Epstein, 1964; Hakala and Rustum, 1979; Hill, 1983a).

Most of the early studies assessing the potential value of cytotoxic drugs were relatively simple, for example, measuring the colour Intensity of lysed cells stained with crystal violet (Grady et al, 1960). An agar plate assay was devised by Schuurmans et al, (1961), which was very similar to that used for microbiological tests evaluating the effects of antibiotics. Solutions of cytotoxic drugs were dispensed into stainless steel cylinders in the agar plate. The cytostatic effect of a drug was assessed according to the diameter of the zone produced as a result of cytotoxicity. A similar technique was used in other investigations (Knock, 1965, 1971; Kondo et al, 1966; Kondo, 1971), except in these studies the cytostatic effect was measured by determination of the inhibition of succinic dehydrogenase activity evaluated by colour changes of various dyes.

1.4 HUMAN CONTINUOUS CELL LINES AS MODELS OF DISEASE; CRITERIA

There have been many studies described in the literature using both animal tumours in vitro and early studies using human tumour cells. A number of these reports suggest the experimental in vitro results reflect clinical therapeutic data (see, Johnson and Goldin, 1975; Mattern and Volm, 1982; Ekwall, 1983; Carney and Winkler, 1985). However, for the in vitro screening of anticancer agents most likely to prove cytotoxic for human neoplasms, it is reasonable to suggest that only human tumour cells should be used. Ideally these should be human continuous cell lines derived from the same histological tumour type for which the efficacy of the cytotoxic drugs are being assessed.

It is possible to grow neoplastic cells derived from a biopsy of a human tumour in vitro. If a pure population of these tumour cells is established with an indefinite life-span, it is defined as a continuous cell line (Schaeffer, 1984). Such lines provide an unlimited supply of neoplastic cells derived from a single human tumour, and thus are useful for in vitro

models for human cancer. The first human continuous cell line, HeLa, was established in 1952 (Gey et al, 1952). It was derived from a cervical carcinoma, the cells of which were initially maintained, using roller tube cultures, in a composite medium of chicken plasma, bovine embryo extract and human placental cord serum. There are now many hundreds of continuous cell lines reported in the literature (see reviews, Fogh and Trempe, 1975; Fogh, 1978; Sinkovics et al, 1978; Smith and Dollbaum, 1981; Fogh et al, 1982). Consequently, there are now available panels of cell lines, each derived from human tumours of one histological type, including, for example, bladder (Hepburn and Masters,1983), breast (Engel and Young, 1978), colorectal (Leibovitz et al, 1976), head and neck (Easty et al, 1981), melanoma (Tveit e£ al, 1981b), neuroblastoma (Brodeur et al, 1981), small—cell lung cancer (Pettengill et al, 1980; Gazdar et al, 1980, Carney et al, 1983), sarcomas (Bruland et al, 1985) and testicular germ cell tumours (Andrews et al, 1980). Thus, it is now possible not only to use human cancer cells for in vitro studies of experimental chemotherapy, but also to specify which histological type of tumour is employed. This is a major advance because different drugs and schedules are used for each histological type of tumour.

Continuous cell lines grow in tissue culture either as monolayers or in suspension in an environment uninfluenced by humoral factors. Their interrelationships with blood supply and normal cells are lost. Only a relatively small proportion of human tumour cells can be cultured indefinitely in vitro, and thus those cells successfully propagated probably represent a selected subpopulation from a heterogeneous tumour. To validate the use of continuous cell lines as in vitro models of disease, and for the purposes of selecting effective anticancer drugs relevant to patient treatment, it is essential to determine the extent to which a number of criteria are fulfilled. Firstly, the biological characteristics shown by the cell lines should be

examined In order to determine If they retain the characteristics of the parent tumour. Secondly, the extent to which a panel of cell lines from one histological type of tumour reflect the spectrum of characteristics encountered clinically should be determined. Thirdly, it is important to know if these features are stable over many generations in culture. Finally, with regard to chemotherapy treatment, the extent to which continuous cell lines can select the most effective anticancer agents needs to be assessed. Experiments designed to answer these points are described in the following chapters of this thesis.

1.5 EXTENT TO WHICH HUMAN CONTINUOUS CELL LINES FULFIL THESE CRITERIA; RETENTION OF CHARACTERISTICS OF THE PARENT TUMOUR

A. Isozymes

Within cells there are certain enzymes which occur in different forms known as isozymes. All tissues, within an individual and with demonstrable enzyme activity, have the same phenotype (Wright et al, 1981). This genetic polymorphism can be used to demonstrate the uniqueness of a particular cell line. A common problem with continuous cell lines is cross-contamination. This has been shown to occur between cells derived from tumours of the same histological type (O'Toole et al, 1983), or with other human tumour cells, often the ubiquitous HeLa cells (Nelson-Rees et al, 1981) or with cells of other species (Harris et al, 1981). By applying isozyme analysis it is possible to investigate whether a cell line is of human origin, derived from a particular patient or has been cross-contaminated.Furthermore, isozyme analysis can be used to ascertain whether certain biochemical characteristics have been retained and the stability of the phenotype expressed.

Ideally, when attempting to establish a new cell line from tumour material, peripheral blood cells should also be obtained so that the isozyme patterns of the normal cells and the established cell line can subsequently be shown to be identical. Polymorphic enzyme analyses of large numbers of cultured human and animal cell lines have been described in the literature (Gartler, 1967? Fogh et al, 1977? Wright et al, 1979, 1981; O'Brien et al, 1980; Povey et al, 1980; Nelson-Rees et al, 1981; Dracopoli and Fogh, 1983a,b). As more lines are analysed the statistical probability of a small proportion demonstrating identical phenotypes will increase, and yet they may not be cross-contaminated. In such instances, in order to confirm their individuality, additional polymorphic loci will need to be analysed. A number of loci (approximately 15%) may not be expressed in cell lines (Povey et al, 1980; Dracopoli and Fogh, 1983a). This loss of expression may have occurred during tumour cell growth in vivo or early during the establishment of the cell lines in vitro. Nevertheless, the loci that are expressed appear to remain stable in culture. For example, Wright and co-workers (1979), showed that pairs of lines, derived either at the same or different times from one tumour, demonstrated phenotypic stability. In addition, phenotypic stability can also be shown in lines such as the well known HeLa cells which have been cultured over many hundreds of passages (Dracopoli and Fogh, 1983a).

B. Biochemical characteristics

In addition to the different forms of enzymes (isozymes), cells can produce a variety of substances which are sometimes more broadly defined as tumour markers. These include hormones (Amatruda et al, 1963; Yalow et al, 1979; Carney et al, 1983), oncofetal antigens (Gold et al, 1978; Neuwald et al, 1980), hormone receptors (Engel and Young, 1978) and tissue specific products such as melanin (Sinkovics et al, 1978) and ferritin (Grail et al, 1982). All these different substances have been classified by Bates and Longo (1985), although such a classification is arbitrary (see Table 1.2).

Table 1.2 Classification of Tumour Markers

1. ONCOFETAL ANTIGENSCarcinoembryonic antigen (CEA) Alphafetoprotein (AFP)

2. ECTOPIC HORMONESAdrenocorticotropic hormone (ACTH) Antidiuretic hormone (ADH)Parathormone (PTH)Calcitonin

3. PLACENTAL PROTEINSHuman chorionic gonadotropin (HCG)Human placental lactogen (HPL)Regan isoenzyme (of alkaline phosphatase)

4. ENZYMESProstatic acid phosphatase (PAP)Lactic dehydrogenase (LDH)

5. SERUM PROTEINSImmunoglobulins

6. MISCELLANEOUSPolyaminesFerritinMelaninHydroxyproline Catecholamine metabolites

Although it has been demonstrated that cell lines can express many biochemical features, it is less clear to what extent individual lines reflect the biochemical properties of their parent tumour. For example, Easty and co-workers (1981) described ten human squamous carcinoma cell lines, derived from head and neck tumours, producing human chorionic gonadotropin (HCG), a feature not observed in the original tumours. In contrast, relatively few non—seminomatOus testicular tumour lines produce alphafetoprotein (AFP) or HCG, in contrast to most tumours of this histological type (Andrews et al, 1980).

Cell lines usually retain tissue—specific and tumour- specific properties, and therefore provide a tool for molecular biologists, such as monoclonal antibody production. However, although monoclonal antibodies specific for a variety of human malignancies have been described (see references cited by Bates and Longo, 1985), there is little information available on the clinical relevance of the use of these antibodies as markers. In contrast to the stability of many tissue— and tumour—specific properties, certain biochemical properties can be lost from cells maintained in vitro. For example, cytochrome P450, a drug- metabolising enzyme, the activity of which is particularly relevant to cancer chemotherapy. Furthermore, the activity of these drug-metabolising enzymes can also be modulated by culture conditions (Bridges, 1980).

C. Chromosomal Studies (Karyotyping)

Chromosome abnormalities are characteristic of tumour cells (German, 1974; Atkin, 1976) and both chromosome mode number and markers are two features of a continuous cell line which, if the facilities and expertise are available, should be determined. In addition, karyotyping, now possible with banding techniques (Kakati et al, 1976; Harnden and Taylor, 1979), can be used to identify interspecies and intraspecies cross—contamination of

cell lines (Nelson-Rees et al, 1981), and to study tumour histiogenesis (Wang et al, 1981). Numerous investigations of the DNA content, chromosome number and karyotype of human cell lines have been described (for review see Gilbert, 1983). In addition, the technique of somatic cell hybridisation and DNA transfer between cells have been used to map the human genome and identify sequence coding for oncogenes (Hamlyn and Sikora, 1983).

Some studies on bladder carcinomas have shown that the differentiated tumours have near—diploid karyotypes, whereas poorly differentiated tumours have chromosome populations ranging from triploid to hypotetraploid (Spooner and Cooper, 1972; Hastings and Franks, 1981). Alterations in chromosomal constitution have been reported between cells of the same line at different passage levels and between clones (Fogh et al, 1982). Thus, cells of early, rather than late, passages probably more closely reflect the chromosomal constitution of the parent tumour. A finding often reported is that the modal number of chromosomes decreases with long-term culture. For instance, the modal number of the bladder carcinoma cell line T24 was shown to be 86 at passage 26, compared to 78 at passage 66 (Malkovsky and Bubenik, 1977). Conversely, Tveit and Pihl (1981), reported that consecutive subcultures of most cell lines derived from melanoma xenografts remained stable. In addition, three lines derived from three separate metastases in a patient with a testicular germ cell tumour, established from biopsies taken over a three year period, all had similar chromosome numbers and identical modes (Wang et al, 1980). Cells within each of the three lines contain structural changes to chromosome 1 and have six identical markers. Marker chromosomes are also a common feature of continuous cell lines derived from human bladder tumours (Hastings and Franks, 1981).

From numerous chromosome studies it is clear that no generalisation can be applied to changes in the chromosomal pattern of cells in long-term culture. Furthermore, some data (Wang et al, 1980) provide evidence that chromosomal alterations are not artefacts of in vitro culture, since such changes are unlikely to have occurred independently in cell lines derived from a primary tumour and its metastasis.

D. Tumorigenicity

Tumour production in immunosuppressed animals is an essential part of the characterisation of a continuous cell line. It is said to be the best available system for testing the malignancy of human cells (Stiles et al, 1976), demonstrating the retention of both the neoplastic phenotype and the morphology of the parent tumour. Most cell lines have the capacity to develop tumours on transplantation into an immunosuppressed animal, usually the congenitally athymic nude mouse (Rygaard and Povlsen,1969). The largest published series of xenotransplantation was made by Fogh et al, (1977b) in which 127 out of 162 (78%) human lines formed tumours in nude mice. Occasionally, loss of tumorigenicity can occur between different passage levels and clones of the same line. Work presented in this thesis (see Chapter 3) show that tumorigenicity was highly variable among the five T24 sublines tested. Clones of one of these sublines, MGH— UI, designated EJ in some early studies, also resulted in lines differing widely in tumorigenicity (Hastings and Franks, 1981). Loss of tumorigenicity has also been observed in melanoma (Tveit and Pihl, 1981) and other cell lines (Fogh et al, 1977b). These findings have not been fully explained, although the differences in tumorigenicity may reflect changes in the immunogenicity of the cells. Furthermore, tumorigenicity is influenced by a variety of factors extrinsic to the cell lines such as implantation site and the immunological state, age and health of the recipient animal.

All the xenografts derived from 127 lines described by Fogh and colleagues (1977b), had a histology compatible with that of the tumour of origin. In addition, replicate xenografts of a number of human bladder cell lines (see Chapter 3) match the histopathology of the original biopsy taken more than 18 years earlier. Such reports demonstrate that xenografts provide the most convincing evidence that the majority of cell lines retain the biological properties of the parent tumour and that these are stable in long-term culture. However, not all cell lines will produce xenografts histologically identical with the tumour of origin. Nevertheless, using this technique, cells which retain this capacity can be selected. A number of such lines can then be used as an in vitro model for that particular histological type of human tumour.

E. Growth characteristics

Many aspects of the growth characteristics of cell lines have been described, including population doubling and generation times, growth fractions, labelling and mitotic indices, DNA distributions and colony forming efficiencies on plastic and in agar. It is difficult to draw general conclusions concerning growth rates of cell lines as these are influenced by the media, serum, substrate and other media supplements and also by cell density and the frequency of subculture. Variations in growth rate have been observed between different passage levels and clones derived from the same line. Nevertheless, under standardised conditions, growth rates usually remain stable over many generations in culture and it has been suggested that they may be considered as an inherent and distinctive characteristic (Pettengill et al, 1980).

Population doubling times of most human cell lines are in the order of 20 — 50 hours, bearing in mind the variations that may occur as a result of growth conditions. In contrast, of

sixteen lines derived from small—cell lung cancers, seven had doubling times in excess of 4 weeks (Pettengill et al, 1980).Slow growing lines such as these are less easily managed and difficult to use experimentally, and thus are less frequently described.

Growth of cell lines in agar is used as a marker for cancer cells and is often correlated with their ability to form tumours in vivo (Freedman and Shin, 1974, 1978). However, not all cancer—derived cells produce colonies in agar and a number of reports show no correlation between growth in agar and tumorigenicity (Marshall et al, 1977; Hastings and Franks, 1981). Furthermore, colony formation in agar can depend both on the type and concentration of agar and the density at which the cells are seeded. The ability of transformed fibroblasts to grow in agar appears to result not only from a loss of anchorage dependence but also from an increased resistance to the polyanions contained in agar (Montagnier, 1971). For this reason many tumour cell lines are grown in agarose which lacks polyanions. Clonal growth of such cells in agarose is not only often superior to that in agar but is also more sensitive, particularly for cells that have lost anchorage dependence but are still inhibited by polyanions.

F. Morphology

Continuous cell lines show a range of morphological features at both light and electron microscope levels. The cells may exhibit differences in the amount of cytoplasm, organelles, size of the nucleus and cellular inclusions such as glycogen (glycogen is found in many human bladder carcinoma cells, Rigby and Franks,1970). Most published characteristics of cell lines generally described their cellular morphology. The similarity in the appearance of some lines, derived from different histological types of tumour and the contrasting disparities in the

morphology of those derived from the same histological type of tumour, suggests the value of such morphological descriptions are limited (Springer, 1980). However, the observations of characteristic features, such as keratin in squamous cell carcinoma lines (Easty et al, 1981; Tilgen et al, 1983) melanin in melanoma lines (Sinkovics et al, 1978) or specific tumour cell markers (Bates and Longo, 1985) provide some reassurance.

A number of continuous cell lines show cell populations consisting of more than one morphological cell type. This feature is well documented for a number of human bladder cancer cell lines and has been confirmed by electron microscopy. For example, a population of T24 cells has been described as showing one cell type consisting of large, round, light nuclei, frequently with seven to eight nucleoli and abundant, slightly pyronin—positive cytoplasm. The other cell type, within the same population, are described as elongated cells with oval dark nuclei, also with several nucleoli and a more strongly pyroninophilic cytoplasm (Bubenik et al, 1973). Another human bladder cancer cell line demonstrating an even wider range of morphological features is that exhibited by HT—1197, with three morphologically distinct cell types (Rasheed et al, 1977).

In contrast to the observations of living cells under phase microscopy, the use of routine cytological examination of specially fixed and stained cells can permit the histological identification of a cell line. It has been shown that when cell lines are processed using cytological techniques to produce a stained smear the morphology of the cells is similar or identical to that of exfoliated cells normally examined by cytopathologists (Hajdu et al, 1974), see also Chapter 3.

G. Sensitivity to chemotherapeutic drugs

i. Use of patients' individual tumour cells

Many attempts have been made over the past decades to develop in vitro assays to predict sensitivity or resistance of an individual patient's tumour. These in vitro assays might also help define the activity of new experimental chemotherapeutic drugs. It has been said that in vitro screens, employing cells derived from freshly resected human tumour tissue in a clonogenic assay, provide information of the greatest relevance on antineoplastic compounds used in the clinical treatment of this disease (Salmon, 1980). Furthermore, it is probably the best characterised in vitro assay (for review see Hanauske et al, 1985). Although a number of recent improvements have been made (Hanauske et al, 1987), there are many technical difficulties and problems inherent in the use of such specimens (Lieber and Kovach, 1982; Mattern and Volm, 1982; Selby et al, 1983; Carney and Winkler, 1985). Furthermore, many studies have demonstrated that individual patients with the same histological type of cancer do not respond uniformly to chemotherapeutic drugs, demonstrating the considerable heterogeneity of responses which may be observed (Kern et al, 1984; Von Hoff and Clark, 1984).

ii. Use of continuous cell lines

For the reasons described above, a different approach to determine the sensitivity of human tumour cells to chemotherapeutic drugs was investigated. This approach used continuous cell lines derived from human tumours, of which there are now many well characterised examples (for references see section 1.4). The use of such cells provides the multiple and reproducible comparisons of drug potency over protracted periods of time. In addition, the use of a panel of lines derived from one histological tumour type would indicate the variation in drug sensitivity that may be seen clinically in that tumour type.

Cell lines are now probably the most widely used in vitro model system for investigating the mechanisms of action of, and resistance to, chemotherapeutic drugs (Hill, 1983a). The basicproblem of all in vitro tests is that of being able to mimick invivo drug exposure conditions. Therefore, doubt has been expressed as to whether cell lines reflect the drug sensitivities of their tumours of origin and if the action of a drug in vitrois comparable with that in vivo.

iii. Drug sensitivity assays using continuous cell lines

A number of in vitro methods employing cell lines have been used to measure in vitro drug sensitivities, see Section 1.6 C (for reviews see Mattern and Volm, 1982; Carney and Winkler, 1985). Those most commonly used include (a) clonogenic assays, which measure the tumour cell reproductive capacity, (b) medium term assays, relying on some measure of cell proliferation and/or viability, using a radiolabel or vital dye and (c) short term biochemical assays using radioactive precursors of either DNA or RNA. Although there are exceptions, in general there is much evidence which shows agreement between in vitro cytotoxicity test results and clinical data (Foley and Epstein, 1964; Hill, 1983b; Finley and Baguley, 1984; Carney and Winkler, 1985).

iv. In vitro/in vivo comparisons

The ideal comparison would be to compare the effect of single agents in patients with the response in vitro of cell lines derived from the same tumour. Although logistically this is virtually impossible, a study by Minna's group, using human small cell lung cancer cell lines, most closely approached this criterion (Carney et al, 1983). From their findings they concluded that there was an excellent correlation between the prior treatment status of the patient and the in vitro sensitivities of derived cell lines. A slightly different

approach comparing in vivo and in vitro drug sensitivities was reported by Baguley and Wilson (1987). They compared the sensitivity of a Lewis lung carcinoma cell line and the P388 murine leukaemia cell line in vitro to the in vivo response of these lines after injecting into mice. Their findings emphasise the importance of tumour site and drug administration in influencing apparent tumour selectivity in experimental cancer chemotherapy. Furthermore, there is also convincing evidence that human cell lines do retain the characteristic patterns of drug sensitivity of tumours of that histological type (Arlett and Harcourt, 1980; Oosterhuis et al, 1984). In addition, in vitro drug sensitivity of cell lines are generally stable in long-term culture (Foley and Epstein, 1964; Tveit et al, 1981b; Shoemaker et al, 1983).

A reservation concerning the use of cell lines for assaying chemotherapeutic drugs in vitro is that some, for example cyclophosphamide, require metabolic activation. However, this problem can be circumvented, for example with the addition of metabolising systems to the culture (Metelmann and Von Hoff, 1983). Furthermore, current data indicate that most drugs have similar cytotoxicity in vivo and in vitro (Ekwall, 1983).

1.6 DRUG SCREENING

A. Drug development

For chemotherapy to be successfully employed In the treatment of human bladder cancer the most effective agents, and combinations, must be utilised and new cytotoxic drugs evaluated as they become available. Today there are a large number of potential cytotoxic drugs available for the treatment of human neoplastic disease. There are successive steps in the clinical development of new anticancer agents. Those that show activity in a number of transplantable animal tumours are entered into phase I clinical trials to measure the toxicities and the

maximally tolerated dose in man. Because of the importance of patient and disease—related parameters, phase II clinical trials use the maximally tolerated dose of a drug in man against a panel of human tumours (Marsoni and Wittes, 1984). The National Cancer Institute (NCI) human tumour panel was created in 1975 and includes the use of a P388 mouse leukaemia prescreen (Von Hoff et al, 1979). Agents which appear the most effective in this prescreen are selected for further screening against lung, colon and breast carcinomas, lymphoma, leukaemia and melanoma. Those that show activity in the phase II trials are then selected for comparison with drugs currently used for the treatment of a particular disease (phase III). In addition, the therapeutic potential of these agents is further explored by comparing with other agents and/or used by alternative routes of administration. In the United States of America, using the screening system described above, approximately 15,000 potential anticancer agents are screened by the NCI each year (Goldin et al, 1981).

At the present time there is little clinical data concerning the effectiveness of these agents against bladder cancer. This is primarily due to the difficulty and expense of clinically evaluating all the drugs exhibiting some antineoplastic activity in broad screening programmes. It is clearly evident that experimental model systems would be of great value as a primary screen for anticancer agents and only those which appear potentially effective for a particular histological type of tumour would have to be clinically assessed.

B. Experimental systems in vivo

A method of obtained preliminary data to assist the clinician in the selection of potentially active cytotoxic drugs is the use of an in vivo experimental system. A number of such systems have been developed (see Table 1.3) and one of these animal model systems has been particularly useful for the investigation of bladder cancer in evaluating both potential

Table 1.3 Advantages and disadvantages of In vivo model systems

In vivo model I Disadvantages Advantages

I * More resource demanding * Large supply of animalsI than in vitro models. with virtually identicalI * Dose levels used tumours can be attained.

Animal 1 frequently exceed those * Treatment of tumourModels t tolerable in humans. bearing animals akin to

I * Tumour burden usually those in patients.I small compared to those • Cytotoxic drugs thatI in patients at the require metabolic1 commencement of activation easilyI treatment. assessed.

I * Not all tumours grow, ■ Drugs can be testedI take— rates sometimes without disruption ofI low. tumour/host tissue.I * Results may take up to * Effect of drugs measured

Xenografts I 2— 3 months. over several cellI • Growth kinetics of cycles.1 tumour in animal may be ■ Test drug combinations.I different from that in * Cytotoxic drugs thatI humans. require metabolic1 * High cost of animals. activation easily

assessed.

I * Evaluation of results * More rapid technique1 may be impaired due to than animal models and

SRC assay I lymphocytic xenograft techniques.I infiltration. * Small tumour burden more1 * Expensive in terms of closely resembles aI animals required. metastasis.

* Retains the tumourmicro—architecture.

1 * Expensive in terms of * Human tumour xenograftsI animals required. can be used as good

ADC assay 1 * Agar may affect access supply of tumour cells.I and clearance of some • Cytotoxic drugs thatl cytotoxic drugs. require metabolic

activation easilyassessed.

Abbreviations.

SRC assay a Subrenal capsule assay

ADC assay = Agar diffusion chamber assay

intravesical and systemic agents (Soloway et al, 1973? Soloway and Murphy 1979, 1981? UICC Technical Report, 1981; Corder et al, 1984; Drago, 1984), It is said that in in vivo models the anticancer activity of drugs is measured in tumour—bearing animals under conditions more akin to those prevailing in patients. Conversely, treatment in animal experiments, usually starts at an early stage and the tumour burden is small, whereas patients are usually treated in an advanced stage with a large tumour burden. Moreover, in animal experiments, dose levels are frequently used which strongly exceed those that could be used in humans. A further argument against in vivo methods is that they are more resource-demanding than in vitro models.

i. Animal models

There are four main categories of animal model tumour,(i) spontaneous, as a result of inbreeding animals which have a high incidence of a particular neoplasm, (ii) carcinogen—induced, as a result of feeding animals with carcinogenically active chemicals, a method particularly used in the area of bladder cancer research, (iii) virally—induced,.the tumours being categorised by the genetic make-up of the virus, DNA or RNA, and (iv) transplantable tumours. The advantage of such animal models is that tumours which initially arise, due to one of the above mentioned routes, in strains of highly inbred mice and rats may be transplanted without rejection. Thus, a large supply of animals with virtually identical tumours in the same stage of growth, and which retain their initial histological appearance, can be easily attained (Soloway and Murphy, 1979; Drago, 1984). Furthermore, from studies so far there has been a relatively close correlation between agents found to be effective in animal models, particularly the murine model, and in clinical studies (Soloway and Murphy, 1981).

It is now generally accepted that anticancer agents which show a cytotoxic effect for one type of animal tumour will not necessarily be active against tumours of a different histological type (Connors and Phillips, 1975). These differences in neoplastic activity may be due to the individual metabolic characteristics and general cell kinetics of each type of cancer, or its vascularity, antigenicity and differences in the aquisition of drug resistance (Johnson and Goldin, 1975). Thus, if animal models are to be used to select effective anticancer agents, a specific model system for each type of neoplastic disease would have to be employed. A logical development of this concept, in order to simulate more precisely the clinical situation, would be the use of human tumours that were transplanted into Immunosuppressed or athymic animals with the resultant formation of a xenograft (Houghton and Houghton,1983).

ii. Human tumour xenografts

The transplantation of human tumours to experimental animals has been investigated for many years. Initially, successful transplants were made in immunologically incompetent animals or to immunologically privileged sites, such as the anterior eye chamber of rodents (Greene, 1952) or the cheek pouch of hamsters (Goldenburg and Witte, 1967). It was then shown that transplants could be achieved in other sites if the animal's immune system was suppressed or newborn animals were used in which the Immune system is not completely developed. With the availability of a thymus aplastic mouse mutant (Flanagan, 1966; Pantelouris, 1968), known as the 'nude' mouse, techniques consequently became simpler because additional immunosuppression is not necessary.

The first report of human tumours transplanted to thymus aplastic nude mice was made by Rygaard and Povlsen, (1969). Since then there have been many publications which have reported

successful transplantations of human tumours to nude mice (see review by Mattern and Volm, 1982). A number of animals have since been used as recipients of human tumours and these, together with the take—rates of different human malignancies, have been reviewed (Steel, 1978). Human tumours are either transplanted as small fragments or cell suspensions inoculated usually into the hind limb, peritoneal cavity or subcutaneously in one or both flanks of the animal. The effect of a cytotoxic drug on the xenograft is either assessed morphologically or by measurement of survival time or, using several parameters, by tumour size. One of the major advantages of the xenograft system, in contrast to in vitro methods, Is that the technique permits in vivo sensitivity to be measured without disruption of the tumour tissue, which becomes integrated with the vascular system of the host. In addition, the drugs can act throughout several cell cycles and, moreover, it is possible to test agents such as cyclophosphamide, which require metabolic activation. Furthermore, drugs or combinations of drugs can be administered in a manner similar to their clinical application, a feature not possible with in vitro methods. There are nevertheless a number of disadvantages with the xenograft model. Not all human tumours grow in animals and the take—rates for some are very low, for example breast tumours (for review see Mattern and Volm, 1982). Thus, the use of this model for predictive testing is limited. From those tumours which do successfully take, the results of drug treatment may take several months. Furthermore, investigations with human bronchogenic carcinoma have shown that the cell kinetics of the transplanted tumour can be significantly altered, compared to the original tumour (Mattern et al, 1981).An additional drawback is that the measured response may reflect the combined result of a direct effect of treatment upon the tumour, and an indirect effect of treatment in modifying the level of host response.

iii. Subrenal capsule (SRC) assay

Tumour tissue from xenografts can also be used as a source of human tumour material for the SRC assay, first described by Bogden and co-workers, (1978). Small pieces of tumour tissue obtained from the xenograft are implanted under the renal capsule of conventional mice and the size of the graft is measured using a stereoscopic microscope fitted with an ocular micrometer. The animals are treated with the test drug, and after 6 days the size of the graft is again measured. The assay time gives this technique its more commonly used name of 6—day SRC assay. An advantage of this assay, apart from speed of results, is that the cell membrane integrity, cell to cell contact, and the spatial relationship of the heterogeneous cell population within the tumour fragment are maintained. In addition, the low tumour burden more closely resembles small metastases than does a subcutaneous palpable mass. The disadvantages on the other hand are the large number of animals required and the volume of drug used. Furthermore, tumour size as a parameter for measuring tumour response may be invalidated due to host cell infiltration. However, it has been suggested that a four—day SRC assay would yield more accurate results due to increased growth and better preservation of tumour cells and minimal lymphocytic infiltration (Levi et al, 1984).

iv. Agar diffusion chamber (ADC) assay

This is a further in vivo technique which can use cells derived from human tumour xenografts as a source of material.The assay was initially developed for the colony formation of haemopoietic precursor cells described by Gordon et al, (1975). Using a modification of this technique Smith and co-workers (1976) successfully grew cells from human tumour xenografts. Briefly, the ADC assay technique consists of implanting agar chambers, containing tumour cells, into the peritoneal cavity of

mice, pre—irradiated with 900 rads whole—body irradiation, usually from a ^^Co source (Smith et al, 1976). The lethal effects of irradiation at this dose can be overcome if mice are treated, two days before irradiation, with cytosine arabinoside (Smith et al, 1976).

An advantage of the technique, which also applies to the SRC assay, is that cells derived from human tumour xenografts can more reliably provide a continuous supply of tumour cells, rather than directly from surgical specimens. In addition, it has been suggested that the ADC technique may be the assay of choice for drugs that require metabolic activation (Sobrero and Marsh,1984). One of the major disadvantages of the ADC assay as a screening system for anticancer agents is based on the grounds of cost. Ideally, xenografts of human neoplasms are used as the source of tumour cells. Thus, in addition to the immunosuppressed animals required for xenograft development, further animals are necessary for the implantation of the agar diffusion chamber. Furthermore, Selby and Steel (1982) suggested that the agar used in the chambers may affect the access and clearance of some drugs, for example, they showed that the activity of some drugs, such as melphalan, against cells treated using the ADC assay and other in vivo techniques differed significantly.

C. Experimental Systems in vitro

Currently the number of effective anticancer agents being discovered has diminished, and it could be argued that further animal screening systems will select only more drugs with a similar activity to those already available. In addition, many agents which are selected have no clinical value and are known as 'false—positive' agents. There is also an unknown number of other agents which show no activity in the experimental system yet would be active clinically, known as 'false—negative' agents,

such as cyclophosphamide, prednisolone and busulphan (Connors and Phillips, 1975? Johnson and Goldin, 1975), Cyclophosphamide is the best known example which becomes cytotoxically active after enzymatic activation in the liver (Sladek, 1971). The difficulty of using cyclophosphamide for in vitro studies can be overcome by either incubating the agent with liver microsomes before the treatment of tumour cells (Phillips, 1974), or by using the derivative 4—hydroperoxycyclophosphamide, which is stable at low temperatures but converts in vitro to the active 4— hydrocyclophosphamide at normal physiological temperature (Kaufmann et al, 1975).

Alternative systems to animal testing are in vitro techniques using cells derived from human tumours which have several advantages over animal tumour models (Penso and Balducci, 1963? Hakala and Rustum, 1979). They provide a rapid, reproducible and inexpensive method of assessing the efficacy of anticancer agents and might identify clinically active agents missed using in vivo animal tumour systems. Furthermore, the environment in which cultures are maintained is more easily controlled, a variety of precise endpoints are available, samples can be divided for multiple analysis and the effect of many body systems, including the nervous and hormonal systems, which complicate in vivo studies, are absent. However, it must be recognised that tissue cultures also have disadvantages. The techniques used are not straightforward. They require training, expertise, experience and sometimes expensive equipment. Furthermore, the absence of the influences and mechanisms which so drastically affect the behaviour of cells in the whole body and the uptake, distribution and modification of chemicals taken into it, may mean that tissue culture techniques will give misleading information. However, it is not envisaged that in vitro models could completely replace animal studies, because, for example, it is not currently possible to measure, in vitro, the degree of toxicity to non—tumour cells which cannot be

maintained in culture as continuous cell lines. Neither is it possible to measure certain pharmacological parameters which must be known before a drug can be used clinically (Connors and Phillips, 1975). Nevertheless, the use of in vitro models would significantly reduce the number of animals killed in an attempt to identify more effective agents for cancer chemotherapy. A summary of the most important predictive tests using a number of model systems has been reviewed, (see Tanneberger and Nissen, 1983? Carney and Winkler, 1985).

Over the past 20 — 30 years many in vitro predictive test systems measuring drug induced cytotoxicity have been developed. These in vitro systems can be categorised into those that measure! (i) gross cellular morphology changes, (ii) cell viability, (iii) cellular metabolism, (iv) the inhibition of radioactive precursor incorporation and (v) cellular reproductive capacity. These assays are either short-term, giving results in a matter of hours? medium-term, requiring 5 — 6 days? clonogenic, usually requiring between 10 — 21 days. The tumour cell material can be either unselected (taken directly from the patient), partly selected, involving days of selection (primary cultures), or highly selected, as in the case of established cell lines.

A variety of in vitro methods used to measure the efficacy of cytotoxic drugs are outlined briefly in the following section. As the work in this thesis is based wholly on a clonogenic assay, it will be described in more detail. The uses and limitations of the most important methods will then be compared.

(i) Types of in vitro assay? brief description

a. Morphological assessment

One of the simplest means of evaluating the cytotoxic effects of anticancer agents in vitro is by morphological assessment. Usually malignant tissue from the patient is used

for the establishment of short-term cell cultures. The morphological characteristics of each cell culture are determined microscopically, then reobserved usually 24 — 48 hours after exposure to cytotoxic drugs. The type of morphological characteristics usually assessed ares the general form and size of the cells, number of mitoses, granulation, nuclear changes, cytolysis, amount of cellular debris and the extent of cell separation from the substrate to which the cells are attached (Wright et al, 1962; Hurley and Yount, 1965? Formina et al, 1970; Dendy et al, 1976; Morasca et al, 1980).

b. Cellular attachment

The attachment of cells to a solid substrate, such as glass or plastic, has been used as a simple and rapid assessment of cell survival after treatment with various agents (Harris and Griffiths, 1974; Tennant, 1974; Bhuyan et al, 1976). This technique, also known as a recovery index, involves the exposure of a monolayer of tumour cells to a drug for a set time. Cells are then washed and detached, usually by an enzymatic process. Known numbers of cells are then resuspended in medium and plated into petri dishes. After a short incubation period, usually up to 24 hours, cell counts are made of the attached and unattached fractions.

c. Cell volume analysis

The volume of cells measured by means of electronic volume analysis, using a Coulter Channelyser, has been used to study the effects of single—agent chemotherapy. Zucker and Helfman (1976) used L1210 ascites tumour cells grown in the peritoneal cavity of mice. Drugs were given on the fourth day and cell volumes measured over a three day period from day five to day seven.

d. Cytotoxic metabolic inhibition (CMI test)

It has been suggested that the cytotoxicity of anticancer agents on tumour cells in culture could be measured by recording

the inhibition of acid production (McAllister et al, 1959). They mixed different numbers of tumour cells, in this case HeLa, with varying dilutions of cytotoxic drugs and compared pH readings with stable phenol—red indicator standards. Controls changed in colour from red (pH 7.6) to yellow (pH 7.0). Effectiveness of cytotoxic drugs was assessed by the prevention of cell metabolism, thus changing the colour from red (pH 7.6) to purple (pH 8.0 plus).

e. Measurement of respiration and glycolysis

The action of cytotoxic drugs in tumour cells affects bothrespiration and glycolysis. Measurements of glycolysis have beenperformed by assessing the effect of drugs on glucose oxidation which can provide information about how the drugs affect cellular respiration. It has been known for some time that alkylating agents which show activity against tumour growth also inhibit their glycolysis (Holzer, 1964). The degree of inhibitioncorresponds to the sensitivity of the tumour to the drug(Larionov, 1965). Assays measuring glycolysis normally use tumour samples which are cut into slices and the effects of cytotoxic drugs measured using classical Warburg manometry (for details see Bickis et al, 1966; Dickson and Suzanger, 1976).

f. Succinic dehydrogenase inhibition (SDI) test

Tests that measure the in vitro effect on cellular dehydrogenase activity usually depend on the reduction of a tetrazolium dye. Triphenyl tetrazolium chloride is added to an incubating mixture of tumour cells and a cytotoxic drug, and the colour changes measured spectrophotometrically. This method is known as the "Kondo" test (Kondo et al, 1966). The effect of a number of anticancer agents on tumour cells have been measured by this technique (Black and Speer, 1954; Kondo, 1971; Knock et al, 1974). A modification of this technique used tumour cells suspended in agar (Miyamura, 1956; DiPaolo and Moore, 1958). The

cell suspension was overlaid with disks containing a cytotoxic drug. Cellular dehydrogenase activity was measured by reduction of methylene blue dye. The area of unreduced dye indicated the inhibition of dehydrogenase activity due to cell cytotoxicity.

g. Growth kinetic determinations

These assays usually involve the counting of cells that have remained attached to a substrate for several days after exposure to a cytotoxic drug. From these results growth curves can be plotted as cell number against time. The increased efficacy of different drugs will manifest itself as an increase in the steepness of the slope of the growth curve compared to that of untreated control cells. Early studies which measured the sensitivities of human tumour cells employing this technique (Ambrose et al, 1962; Easty and Wylie, 1963) used primary cultures. A number of variations to the original technique have since been reported using primary culture cells derived from a variety of human tumours (Phillips, 1974; Holmes and Little,1974; Berry et al, 1975).

h. Flow microfluorimetry (FMF)

Initial determinations of cellular DNA content required large sample sizes and the procedures used were time-consuming (Krishan, 1975). These difficulties were overcome with the introduction of laser—based flow microfluorimetry (Van Dilla et al, 1969). This technique has since been used in a number of studies which require rapid analysis of cell cycle kinetic parameters (Zietz and Nicolini, 1978). Within the last decade it has been used as a rapid means of assessing the kinetic effects of drugs on cells. The assay can be applied to most antitumour agents, including antimetabolites, such as methotrexate and 5— fluorouraci.l; agents which modify DNA covalently, such as cis— platin; agents which intercalate DNA, such as adriamycin (Finlay and Baguley, 1984).

The technique involves the exposure of cells to a cytotoxic drug, the time of exposure ranging from 1 hour (Engelholm et al, 1983) to several days (Finlay and Baguley, 1984). After drug exposure the cells are fixed and can be stained with one of a number of fluorescent DNA selective stains (Crissman and Tobey, 1974; Krishan, 1975; Taylor, 1980). The stained cells are passed through a focussed beam of a high power, usually an argon ion laser using instruments such as a fluorescence activated cell sorter. These instruments measure both size and fluoresence of the cells, and the evaluation of the DNA histograms produced can be performed using the procedure recommended by Barfod (1977).

i. Dye exclusion assays

The ability of cells to exclude vital dye has been used for a great many years as an indicator of membrane damage and cell death (Schrek, 1936; Hoskins et al, 1956; Phillips and Terryberry, 1957; Liberman et al, 1973; Harris and Griffiths, 1974; Tennant, 1974; Warters and Hofer, 1974). The assays are normally performed by counting cells in the presence of the dye as wet preparations. However, the results are based upon the assumption that a particular dye will stain only dead and non— viable cells, and that live cells will remain unstained. After exposure to cytotoxic drugs, dyes such as eosin (Yuhas et al, 1974), nigrosin (Tsukeda et al, 1978) and trypan blue (Durkin et_ al, 1979) have all been used as an index of cell viability.

More recently, a novel dye exclusion assay for testing in vitro chemosensitivity of human tumour cells has been reported (Weisenthal et al, 1983a). Briefly, cells were either exposed to anticancer agents for 1 hour or continuously and then cultured for 4 to 6 days in liquid media. Cells are stained with fast green dye, sedimented onto slides with a cytocentrifuge and counterstained with haematoxylin and eosin. Dead cells stain with fast green dye and living cells stain with haematoxylin and

eosin. To reduce the underestimate in cell kill, often associated with these dye exclusion assays, an internal standard was developed using fixed duck erythrocytes (Weisenthal et al, 1983b).

j. Uptake of radioactive precursors

The uptake of radioactive nucleic acid and protein precursors have been widely used to measure the effectiveness of anticancer agents (Mattern and Volm, 1982). The amount of radio—isotope incorporated into cells can then be measured by autoradiography or by liquid scintillation spectrometry. The isotopes most widely used are [ H] thymidine, [^^1] —Urdr,[%] uridine, [%] leucine, [^S] methionine and [32p] orthophosphate. Many of the commonly used cytotoxic drugs block cell growth in a particular phase or phases of the cell cycle and specific pathway(s) of nucleic acid synthesis. Therefore, choice of precursor used depends on the drug whose cytotoxic action is being investigated (Volm et al, 1979; Mattern and Volm, 1982). Chemosensitivity studies which measure the inhibition of radioactive precursor incorporation have used tumour samples ranging from single cell suspensions (Volm et al, 1979), monolayer cultures (Morgan et al, 1983), cell culture in agar (Tanigawa et al, 1982; Friedman and Glaubiger, 1982) to tumour slices grown as organ cultures (Tchao et al, 1967).

k. Organ culture

The concept of organ culture was originally described in the late 1920s (Fell and Robison, 1929). A number of changes and improvements to the original technique have been made, the methodologies of which have been reviewed by Easty (1970) and Hodges (1981). The technique most commonly used is that described by Trowell (1959) which, briefly, consists of maintaining thin slices of tissue at a liquid—gas interface on some form of support such as a stainless steel grid. Organ

culture techniques have been applied to a wide range of human solid tumours and the effects of anticancer agents on some of these maligant tissues have been reported (see Masters, 1983).

1. Multicellular tumour spheroids

Spheroids are spherical aggregates of tumour cells with an intermediate level of complexity. The 3—dimensional growth of cells provides a diffusion gradient for oxygen, nutrients and chemotherapeutic drugs. If they become large enough (0.4 —0.8 mm^) necrotic regions will develop in the central parts.This configuration is said to emulate solid tumours in vivo (Sutherland and Durand, 1976). These aggregates can be cultivated in two ways: the spinner flask technique and the liquid overlay technique. For detailed methodology see references cited by Nederman (1983). The spinner flask technique has been extensively used for cultivating rodent cell spheroids. Briefly, small aggregates of cells develop from a single—cell suspension added to a rotating spinner flask containing medium. The other method, known as the liquid overlay technique, is more commonly used for cultivating human tumour spheroids. In this technique cell suspensions are placed into dishes with non- adhesive surfaces (i.e. non tissue culture standard) and within a few days spherical aggregates begin to form.

m. Clonogenic assay

There are two types of clonogenic assay routinely used today: (i) the cells, after exposure to anticancer agents, are left to form colonies in a semi—solid medium, usually some form of agar, and (ii) liquid—culture techniques where the cells form colonies on a solid surface such as a plastic petri dish. The first technique is mainly used for the clonal growth of tumour stem cells derived from fresh surgical specimens, whereas, in vitro experiments employing continuous cell lines established from human tumours primarily use the liquid-culture technique for clonal growth.

Clonogenic assay for colony—forming tumour cells was first developed by Puck and Marcus (1955). Other early studies using colony—forming assays determined the cellular sensitivity of experimental animal tumours (Park et al, 1971; Thomson and Rauth, 1974; Courtenay, 1976) and of normal and malignant haemopoietic cells (Senn et al, 1967). These techniques were further developed using semi—solid agar supplemented with enriched media which supported colony growth from stem cells derived from a variety of human tumours (Hamburger and Salmon, 1977; Salmon et al, 1978; Courtenay and Mills, 1978; Courtenay et al, 1978).

The assay system used by Salmon and co-workers, commonly known as the human tumour stem cell assay (HTSCA), uses a double layer of agar. The base layer contains essential nutrients for cell growth and acts as a feederlayer, while the drug treated cells were plated in the upper agar layer for colony formation. Details of the methodology currently used were reported by Hamburger (1983). Colonies in agar were initially defined as collections of more than 40 cells (Hamburger and Salmon, 1977), however, the criterion chosen to define a colony is arbitrary (Sikic and Taber, 1981) and is usually set at 30 to 50 cells (Selby et al, 1983). The other soft agar technique, commonly known as the "Courtenay" assay, involves the growth of drug treated tumour cells in soft agar in a tube with a replenishable liquid phase. This assay differs from other soft agar assays by the addition of August rat red blood cells and a low 0£ concentration (see, Courtenay, 1983).

The majority of clonogenic assays using continuous cell lines, use the liquid—culture assay system where cells are allowed to form colonies on a solid substrate. There are basically two methods; (a) exposure of a monolayer of cells (usually in exponential growth) to drugs for a specified length of time, followed by washing, enzyme detachment and transferral

to petri dishes for colony formation, (b) plating of a known number of cells into petri dishes for drug exposure in situ and allowing the surviving cells to form colonies. Irrespective of which method is employed, counts are made of colonies formed which represent the surviving drug treated cells. These are compared to the untreated control cells that develop macroscopic colonies, usually containing at least 50 cells (Elkind and Whitmore, 1967). There are now numerous reports in the literature where continuous cell lines from a wide variety of malignancies have been established and characterised (Hill,1983a) many of which have been used for in vitro chemosensitivity studies using the clonogenic assay.

ii. The uses and limitations of in vitro models

It has been said (see review by Pihl, 1986) that, in vitro, the chemosensitivity of cells is measured under highly artificial conditions, far removed from those prevailing in a tumour-bearing host. Factors such as tissue distribution of the drugs, diffusion barriers, activation and deactivation of drugs and theinfluence of the geometry of solid tumours on the cellularresponse are not revealed. Cells freshly isolated from surgical specimens of human tumour tissue and continuous cell lines,derived from a variety of human malignancies, have been used in anumber of in vitro predictive tests which measure cell injury,observing one or more parameters (Mattern and Volm, 1982). Someof these assay techniques are still in use today, although many have been modified (see brief descriptions above, Section 1.6Ci).

The most commonly used in vitro techniques are clonogenic assays, inhibition of DNA synthesis (measured by uptake of radio isotopes), organ culture, dye exclusion techniques and more recently the use of multicellular tumour spheroids. There are a number of advantages and disadvantages associated with these and other in vitro systems (see Table 1.4). In general the tests that measure immediate metabolic derangement of the treated cells, such as inhibition of DNA synthesis and dye exclusion, will depend critically on the time of the measurement and may severely over or underestimate the true lethal effect of an anticancer drug (Pihl, 1986).

The principle advantage of assays using radioisotopes is the limited time the test requires. Also original tumour biopsy cells can be used, thereby avoiding the possible difficulty in culturing certain tumour types. However, since the cells continue the DNA—synthesis, initiated in vivo, but do not necessarily enter a new cell cycle (Rajewski, 1966), mitosis- blocking agents such as vincristine or vinblastine cannot be tested. In an assessment of the dye exclusion assay and clonogenic assay a comparison of labelling indices revealed the former to produce only qualitative changes (Rupniak et al,1983a). Dye exlusion assays were re—evaluated by Weisenthal and co-workers (1983a) in an attempt to overcome the many problems associated with them (see Table 1.4). However, for fresh surgical specimens, best results were obtained with solid tumours because of the difficulty in distinguishing normal from malignant cells in serous effusions. Furthermore, assessment of cell kill is still subject to individual interpretation.

A more recent in vitro technique based on a colour reaction is a colorimetric assay based on a tetrazolium salt (Carmichael et al, 1987). The assay is dependent on the cellular reduction of 3—(4,5-dimethylthiazol—2—yl)—2,5-diphenyl tetrazolium bromide, known as MTT, by the mitochondrial dehydrogenase of viable cells

TABLE 1 b Major Advantages and Limitations associated with in vitro Model Systems

In Vitro System I Advantages I Limitations 1

Morphologicalassessment

I Simple to perform. I Rapid assessment.

I1 Not all alterations In cell 1 viability manifest themselves I as cell structural changes.I Assessment Is subjective.I

Cellularattachment

I Simple and rapid 1 assessment of cell I viability.

iI Makes assumption that only I viable cells re-adhere to l substrate after exposure to I lethal doses of cytotoxic drug.i

Cell Volume analysis

1 Relatively rapid.ii More effective In measuring i tumour cell regrovth rather than 1 destruction.I Unlikely this type of assay can l be of clinical value.I

CMI test 1 Easy to set up and I read.

II Poor qualitative results.I Injured cells not differentiated I from reproductlvely dead cells.

SDI test 1 Simple technique, i Rapid result.

I1 Poor qualitative results.I Injured cells not differentiated I from reproductlvely dead cells, i Prediction heavily weighted I towards non—response.I

TABLE 1.4 cont.

In Vitro System I Advantages Limitations

Growth kinetic determinations

I Primary cultures I easily used.1 Results available 1 after few days.

Significantly underestimates tumour cell kill.

FMF 1 Cytotoxic effect of I most types of drugs I can be measured.

Short drug exposure times require drug concentrations In excess of those clinically achievable.Results highly dose— and time- dependent.

Dye exclusion I Short duration of 1 assay.1 >701 tumour specimens 1 can be evaluated to I one or more drugs.

Early studies:Poor correlation between the ability of a cell to exclude vital dye and its long-term reproductive ability.Re—evaluated assays: Difficulty In distinguishing between tumour and non—tumour cells in effusions.Effects due to cell cycle- specific agents may not be detected.

Radioactiveprecursors

1 Rapid short-term t assay.I Original tumour l tissue cells can be l used therefore l greater variety of l tumour types can be 1 assessed.

Mitosis—blocking agents such as vincristine, vinblastine cannot be used.Some drugs affect permutation of certain isotopes.

TABLE 1.4 cont.

In Vitro System 1 Advantages Limitations

Organ culture l Retains integrity of l three-dimensional 1 cellular lnter- 1 relationship.

Qualitative measurements of chemosensitivity not reliable.

Multlcellular tumour spheroids

1 Emulate small tumours 1 In vivo with.regard 1 to drug penetration 1 studies.1 Several ways of 1 analysing the effects 1 of cytotoxic drugs.

Spheroids originating from some cell lines can show heterogeneity.Not all tumour types can be successfully grown as spheroids.

HTSCA I Measure drug l sensitivity of 1 patients' tumour l cells directly.I Growth of stromal l cells inhibited.

Generally poor cloning efficiency (0.001 — 0.1X). Difficulties In obtaining good viable single cell suspensions. Results take 2 - 3 weeks.

Clonogenic assay using cell lines

1 Simple, sensitive,1 cheap model system.I Unlimited supply of 1 material, t High cloning 1 efficiencies.I Many characteristics I of original tumour 1 retained.I Small quantities of 1 drug required.

Selected cell population. Loss or absence of some biochemical properties. Studies of selective cytotoxicity not possible.

to a blue formazan product which can be measured spectrophotometrically. However, it cannot distinguish between cytostatic and cytocidal effects on cells (Carmichael et al,1987) and the cytocidal effects are limited to 1 log of cell kill. Furthermore, certain drugs and ionophores affect permutation at the mitochondrial membrane which may give erroneous spectrophotometric readings.

The technique of organ culture has been used to assess the effect of cytotoxic drugs on a wide range of human tumours (Masters, 1983). This technique has been described as the maintenance of tissues in vitro to allow differentiation and preservation of architecture and/or function (Schaeffer, 1984). Thus, the major concept of the technique is to retain the integrity of three-dimensional cellular interrelationships. However, in vitro studies using this technique have so far been limited by the lack of reliable quantitative measurements which not only take into account tumour heterogeneity but also the presence of normal tissue, the quantity of which may vary from sample to sample.

The technique of growing tumour cells in a spheroid configuration has been used in a variety of research applications (Sutherland et al, 1981). If the spheroids are allowed to grow large enough, necrotic regions will develop in the central part. They are therefore useful for simulating the chemotherapeutic treatment of solid tumours, which is dependent upon drug penetration, proliferation gradients and factors affecting the microenvironment, such as pH and pC^. Some of these studies have important clinical implications, particularly in relation to the problem of drug penetration. It has been shown, for example, that cell subpopulations derived from different zones of the spheroid survive bleomycin treatment to different degrees (Toburen, 1981), such that, the peripheral cells, treated as a monolayer, were less sensitive to bleomycin than the inner

most cells. This confirms previous findings that non- proliferating cells are more sensitive to bleomycin than proliferating ones (Hill, 1978a). However, differences in sensitivity to bleomycin between early and late plateau phase (non—proliferating) cells have been reported (Twentyman and Bleehan, 1975). Tissue penetration differences have also been shown for other drugs (Nederman et al, 1981; Nederman, 1984).

There are three different methods of analysing the efficacy of cytotoxic drugs: (i) measuring intact spheroid growth, (ii) assessing the capacity of cellular attachment and resultant monolayer outgrowth and (iii) clonogenicity of individual cells from disaggregated spheroids. For volume growth the spheroids are examined by inverted microscopy and after repeated measurements of the diameter the volume is mathematically calculated (Nederman, 1983). For cellular attachment and monolayer outgrowth the spheroids, after drug exposure, are transferred to tissue culture dishes. Attachment occurs within a few days and the rate of proliferation of the monolayer outgrowth is assessed. For clonogenic analysis the spheroid is disaggregated, usually enzymatically, into a single—cell suspension and the numbers of surviving cells able to form colonies are measured. Using this technique the differences in drug effectiveness, as a function of depth in the spheroid, can be analysed by sequential trypsinisation (Sutherland, 1980; Wibe,1980). The general conclusions of using multicellular tumour spheroids for drug sensitivity studies are that they offer many of the characteristics exhibited by tumours in.vivo which cannot be simulated by other in vitro culture systems. However, there are limitations, such as, spheroids originating from the same cell line showing heterogeneity (Durand, 1980) and, not all tumour types can be cultured as spheroids.

It is clear from the numerous reports in the literature that many methods have been used to assess the effectiveness of anticancer agents in vitro by measuring cell damage or death. However, it has been argued that the best parameter of cellular death in proliferating populations is the permanent loss of reproductive integrity, i.e. inability to proliferate and to form colonies under appropriate experimental conditions (Roper and Drewinko, 1976; Hill et al, 1980, 1981; Rupniak et al, 1983a). Furthermore, experimental evidence appears to support the theoretical consideration that the inhibition of tumour stem cell colony formation is the most significant end-point to measure the efficacy of anticancer agents (Roper and Drewinko, 1976; Steel, 1977; Agre and Williams, 1983).

Clonogenic assay methods using an agar system are particularly useful for assessing the drug sensitivities of human tumour cells derived from fresh biopsy specimens, primarily because they allow the growth of tumour stem cells whilst preventing the growth of stromal cells (Courtenay et al, 1978; Rupniak and Hill, 1980). However, from studies concerned with optimising the growth of a particular histopathological type of human cell, it has been reported that both normal and malignant cells grow equally as well (Smith et al, 1981), a disadvantage if one is only concerned with the drug sensitivities of neoplastic cells in a mixed cell population from surgical specimens. There have now been a number of publications which report colony growth has been achieved using present methodologies for a variety of tumour types (see Daniels et al, 1981; Hill, 1983c). There are however, a number of problems associated with the use of fresh surgical specimens as a source of human tumour cells for use in clonogenic assays and with the use of the agar system itself. One of the problems associated with tissue samples is the variability in quality and quantity of the biopsy material, in particular, the tumour cell content of the biopsy sample (Von Hoff et al, 1981; Bertoncello et al,

(1982). In addition, although a number of enzymatic and mechanical procedures for tumour disaggregation have been compared (Carney and Malmgren, 1967; Rosenblum et al, 1980; Mackintosh et al, 1981; Slocum et al, 1981a; also see references cited by Hill, 1983c), the difficulty of preparing a single—cell suspension is still frequently reported (Rupniak and Hill, 1980; Laboisse et al, 1981; Rosenblum et al, 1981; Mattern and Volm, 1982; Selby et al, 1983). Single-cell suspensions are an essential requirement for quantitative drug sensitivity assays, however, it has been reported that colonies form more readily from small cell aggregates of up to ten cells (Rupniak and Hill, 1980; Hill, 1983c). The most successful colony growth from human neoplasms is achieved when the minimum number of manipulations, aimed at obtaining a single—cell suspension, are used. This fact perhaps reflects the success in colony growth from human tumours such as malignant ascites and tumours of the haematopoetic system which require minimal disaggregation compared with the majority of solid tumours (Rupniak and Hill, 1980; Hill, 1983c).

The only cells of therapeutic importance within a tumour are the stem cells, because these are the cells that are responsible for tumour repopulation after treatment and also for metastatic growth. They frequently comprise 1% or less of the entire tumour mass (Steel, 1977), a finding which may have a bearing on the other major problem associated with the clonogenicity of tumour cells from primary tissue, that is, the very low colony—forming efficiency, which is reported to be in the order of 0.01 to 2% (Rupniak and Hill, 1980, Sikic and Taber, 1981; Rosenblum et al,1981). Overall, clonogenic cell growth is reported in at least 50 — 70% of specimens tested (Hill 1983c) and it is suggested that at least 30 colonies per control sample are required for drug sensitivity determinations (Von Hoff et al, 1981; Uitendaal et ai, 1983; Schlag and Flentje, 1984). However, the proportion of human tumours that grow with a plating efficiency sufficient for the assessment of drug activity is frequently low (Von Hoff

et al, 1981; Selby et al, 1983, Schlag and Flentje, 1984). As a consequence of the low plating efficiency it has been suggested that drug screening assays using cells from primary human tumours have limited applicability (Rupniak and Hill, 1980).Nevertheless, the potential applications for in vitro biological studies should not be overlooked and attempts to improve tissue disaggregation procedures and optimise the conditions for colony growth should be investigated.

A recent comparison of the two soft-agar methods, using malignant melanomas (Tveit et al, 1981), showed that the colony- forming efficiency was considerably higher in the Courtenay and Mills system compared to the method proposed by Hamburger and Salmon. However, tumour cells appeared to be more sensitive to some drugs using the latter technique. The Courtenay assay also provided the highest colony—forming efficiencies in a range of human tumour cell lines (Hill and Whelan, 1983). Because of the potential usefulness of the clonogenic assay systems using cells from surgical specimens a number of studies have been performed to maximise the colony—forming efficiency of malignant cells.

Many alterations and substitutions have been made to the basic assay systems which have shown an increase in the clonogenicity of cells from neoplastic tissue. Examples of such variations include (a) the replacement of enriched CMRL 1066 medium with cell free ascites, which was shown to improve the growth of ovarian cancer cells (Uitendaal et al, 1983) and (b) the addition of insulin and the removal of red blood cells, which increased the plating efficiency of a number of human tumours (MacKintosh et al, 1981). However, further elimination of unwanted cells should be approached with caution. Buick et al (1980) showed that adherent cells enhanced the colony—forming efficiency of a number of malignant effusions and Hamburger and

White (1981) showed that autologous macrophages enhanced the growth of human cells in soft agar. Other ways of improving the culture conditions were thought to be the use of agarose instead of agar (Laboisse et al, 1981; Carney et al, 1981). However, agarose contains sulphated polysaccharides which strongly inhibit cell growth (Montagnier, 1971). A number of other studies have reported the substitution of methyl cellulose for agar (Buick et al, 1979; Buick and Fry, 1980; Pavelic et al, 1980). None of these substances have been shown to be superior to agar, although improvements in colony growth have been shown with lower concentrations of agar and agarose (Whelan and Hill, 1981; Hill and Whelan, 1983). Nevertheless, the recovery of colonies after growth in methyl cellulose rather than agar can be achieved much more readily. Thus, subsequent cellular manipulations, such as the transplantation of colonies into nude mice or their histological examination, are much more easily achieved (Buick et al, 1979; Pavelic et al, 1980).

Despite relatively easy access to bladder tumours using minimal invasive procedures, chemosensitivity studies using in vitro clonogenic assays have only recently been evaluated.Tumour cell samples have been obtained mainly from bladder washings. Solid tumour tissue in some studies was also obtained by transurethral resection (TUR). Some examples of these studies are summarised in Table 1.5.

The original concept of in vitro clonogenic assays, using semisolid medium, was to assess the sensitivity of a tumour from an individual patient. Most early assays, such as those by Wright et al, (1957) and Cobb et al, (1961) who studied the effect of triethylenemelamine (TEM) against tumours in plasma clot cultures, did not demonstrate a consistent correlation between in vitro activity and clinical course of the individual patient. A failure to identify the subpopulation of stem cells

TABLE 1.5 Chemosensitivity studies of carcinoma of the human bladder using the HTSCA

REFERENCESOLID SUPPORT MEDIUM

(Top layer/Bottom layer)SOURCE OF MATERIAL

Buick et al, 1979 0.8% methyl cellulose 0.8% methyl cellulose/0.5% agar0.3% agar/0.5% agar

Bladderwashings/TUR

Stanisic and Buick 1980

0.3% agar/0.5% agar Bladderwashings

Daniels et al, 1981 0.3% agar/0.5% agar Bladderwashings

Von Hoff et al, 1981

0.3% agar/0.5% agar TUR/lymphnodedlsectlon

Stanisic et al, 1981

0.3% agar/0.5% agar Bladderwashings/TUR

Sarosdy et al. 1982 0.3% agar/0.5% agar Bladderwashings/TUR/Totalcystectomy

Niell and Soloway, 1983

0.3% agar/0.5% agar Bladderwashings

Kirkels et al, 1983 0.3% agar/0.5% agar Urine/bladderwashings/TUR

Kovnat et al, 1984 0.3% agar/0.5% agar 0.8% methyl cellulose/ 0.5% agar

TUR

within the tumour was thought to be the reason (Steel, 1977).More recent clonogenic assay systems (Von Hoff et al, 1981; Alberts et al, 1981b; Salmon, 1982) accurately predict clinical resistance to drugs in approximately 95% of cases and sensitivity in 60 — 65% However, the data for resistance are derived largely from retrospective correlation of the in vitro results in patients who have been exposed to the drug and failed treatment, thus, prospective studies may not yield the same results.

Repeat experiments, over a prolonged period, using cells from human tumour biopsies present a major obstacle to their use, primarily that of shortage of replicate material. However, this is not a problem associated with the use of continuous cell lines. The success rates for establishing a continuous cell line from human tumours are said to be approximately 1 — 10% (Smith and Dollbaum, 1981) and their colony forming efficiency is much higher than that reported for cells from fresh surgical tumour samples.

The ability of cells to grow in agar has been said to be a measure of the cells* malignancy. However, there are a number of examples of continuous cell lines derived from different tumour types, such as bladder (Marshall et al, 1977), breast (Engel and Young, 1978) and head and neck (Easty et al, 1981), which have been shown to lack this ability. Nevertheless, those cell lines that failed to grow in semisolid media were able to form colonies on plastic and were shown to be tumorigenic in nude mice. The majority of clonogenic assays using continous cell lines, however, use the liquid—culture assay system where cells are allowed to form colonies on a solid substrate such as a plastic petri dish.

The main advantages of using established human tumour cell lines in vitro is that they provide an unlimited supply of material which is relatively stable and retains many

characteristics of the tumours from which they were derived. The clonogenic assay using such lines provides a simple, sensitive and cheap model system for measuring the effects of anticancer agents. There are nevertheless some limitations. These include the fact that an established cell line consists of a selected cell population; drug pharmacokinetics are different from those measured in vivo; selected cytotoxicity studies are not possible (see Hill, 1983b). Nevertheless, many studies have reported in vitro chemosensitivity patterns of established cell lines of human tumours (see Hill, 1983c; Carney and Winkler, 1985). In several of these studies the in vitro cytotoxicity has been compared to results observed in patients (Carney and Winkler, 1985). From such data and because of the advantages of using cell lines over fresh biopsy specimens (including availability of cells and reproducibility of assays) the use of panels of cell lines derived from different histological types of human tumour offers an alternative approach to screening new agents in the preclinical setting. The hypothesis of using a panel of cell lines for drug screening is tested in this thesis using clonogenic assays. In addition, the use of such a panel of lines derived from human transitional cell carcinoma as a model for this disease is also investigated.

D. Factors affecting measurements of in vitro drug sensitivties using cell lines

From comparisons made between many of the above mentioned in vitro techniques it was generally concluded that the method of choice for measuring drug sensitivites in vitro is the clonogenic assay (Roper and Drewinko, 1976; Hill et al, 1980, 1981; Rupniak et al, 1983a). However, both cell cycle kinetic and assay conditions may influence the in vitro sensitivity measurements

determined using the clonogenic assay, hence, both factors should be carefully considered in the evaluation of these results. In addition, it should be remembered that in vitro studies with chemotherapeutic agents make the assumption that the results reflect those observed clinically once the drug reaches the malignant cells. Thus, pharmacokinetic determinations of tumour response are circumvented (i.e. absorption, transportation, combination, transformation and degradation).

i. Cell cycle

The concept of the cell cycle was introduced by Howard and Pelc (1953) see Fig 1.2. An understanding of cell cycle kinetics and perturbation effects, following treatment with antitumour agents, are important determinants in the final outcome of drug—cell interactions (Baserga, 1968; Lamerton, 1974; Hill and Baserga, 1975; Gray, 1983; Drewinko and Barlogie, 1984). These effects can be exploited for the design of cancer treatment protocols. The most important cells within a tumour mass are the malignant stem cells. They are thought to arise at low frequency, from the transformation of one or a few normal cells, and are capable of self—renewal and migration. Thus, they are responsible for continued primary tumour growth and initiation of distant metastases. The aim of cancer therapy is the eradication of these malignant stem cells, therefore a knowledge of their kinetic properties is essential. When designing cancer treatment protocols it is important to relate malignant stem cell kill with minimal normal stem cell kill to prevent excessive normal tissue damage (Hill, 1978a).

Our knowledge of cell cycle kinetics of animal tumour systems has continually advanced, however, few of these findings

Figure 1.2 The cell cycle

phase, most variable — up to 30 hours. For majority of cells 5 — 10 hours.S phase, 6 — 8 hours.G2 phase, 2 — 4 hours.M phase, usually completed within 1 hour.

have been extended to human tumours which show considerable cellular heterogeneity, particularly human solid tumours which represent at least 80% of human cancers (Hill, 1978b). The idea that only a proportion of cells in a given tumour, the growth fraction (GF), were actively proliferating was proposed by Mendelsohn (1962). Using autoradiographic techniques it is possible to estimate the growth fraction of viable tumour cells, defined as the ratio of actively proliferating cells to the total number of cells, including the quiescent cells which are out of the mitotic cycle. The growth fraction for a variety of solid tumours in experimental animals frequently ranges between 30 and 80%. Human tumours show a wider range of growth fractions from less than 6% in adenocarcinomas to 90% in malignant lymphomas and embryonic tumours (Hall, 1978). In general, tumours grow more slowly than would be predicted from a knowledge of the cell cycle time and growth fraction. These differences are attributed to cell loss, mainly brought about by cell death, and the realisation that, in addition to actively—proliferating cells, both slowly—proliferating and non—proliferating (G0) cells exist which can revert back to the proliferating state (Steel, 1977). Clonogenic assay procedures may encourage all of these stem cells to proliferate, whether or not they were proliferating at the time the tumour was removed from the patient. In addition, as present methods of investigation are generally unable to detect cell numbers less than 10 in man, estimations of solid tumour growth rates are usually restricted to a short period near the end of their lifespan (Hill, 1978b). Tubiana and Malaise (1976) reviewing data for the duration of the cell cycle in 41 human solid tumours of various histological types, quote values ranging from 15 to 125 hours. Clinical data suggests that rapidly growing tumours are often responsive to chemotherapy, whereas slowly growing tumours respond less favourably to therapy (Shackney et al, 1978). However, there is no simple relationship between the growth rate of a tumour and the patient's survival time (see references cited by Tubiana and Malaise, 1976).

ii. Culture conditions

a. Techniques

Commonly used in vitro systems consist of culturing cells either as monolayers on a solid substrate, single cells in a semi—solid medium, single cells in suspension or as multicellular spheroids, A wide variety of solid substrates, including glass, plastic, Sephadex beads, polyester film and cellulose acetate membranes, have been used in tissue culture systems for the maintenance and growth of anchorage dependent cells (Hodges, 1976), A number of studies have demonstrated that cells show varying affinities for the different types of substrate (see references cited by Hodges, 1976). This property of differential affinity to substrate has been utilised to separate mixed populations of epithelial and fibroblast cells (Hodges, 1976). Periodic media changes of monolayer cultures are required due to the exhaustion of essential constituents which have been metabolised by the cells or spontaneously degraded. The intervals between changes of medium will vary from one cell line to another. The usual ratio of medium volume to surface area is0.2 — 0.5 ml/cm^ (Freshney, 1983a). Cells with a high oxygen requirement grow better in a shallow medium depth. However, if the depth is greater than 5 mm, then gaseous diffusion may become a limiting factor. This may have an effect on chemosensitivity studies by promoting the efficacy of drugs with improved activity in hypoxic conditions. The problem of gaseous diffusion for cells grown as a monolayer culture can be overcome by the use of roller tube techniques. These consist of cells attached to the inside surface of a bottle or tube which is rolled around its axis (Freshney, 1983b). This system has three advantages over static monolayer culture* (i) an increased surface area, (ii) constant, gentle agitation of medium, thus allowing even gaseous exchange to take place and (iii) an increased ratio of medium surface area to volume. When all the available substrate surface is occupied, i.e. confluent monolayer, the culture must be

divided. This is usually performed by removal of the medium, dissociation of cell monolayer with enzyme treatment (e.g. trypsin) and subsequent dilution in fresh medium into new flasks, dishes or tubes.

The use of semisolid medium is reserved specifically for the growth of tumour stem cells, which are often anchorage independent. This culture system is therefore not normally used for the maintenance of cell lines, but is more specifically used for the formation of cell colonies from primary tumour cells following drug exposure. Agar, agarose and methylcellulose have all been used in systems employing a semisolid support, however, the concentration and type of matrix appear to influence the clonogenicity of the cells and therefore influence cell, survival data (Buick and Fry, 1980; Whelan and Hill, 1981; Hill and Whelan, 1983).

Cells can also be cultured and exposed to cytotoxic drugs as single cells in suspension, either because they are non—adhesive or because they are kept in suspension mechanically. A number of tumour tissues such as ascites fluids, leukaemia and lymphoma cells exist naturally as single cell suspensions. Single cell suspension cultures are usually maintained by gentle stirring at a sufficient speed to prevent cell sedimentation but not so fast as to grind or shear the cells (Freshney, 1983b). Enzyme treatment which, depending on its concentration has an effect on the drug treated cells, is not required as subculture is performed either by dilution or increase of the volume of medium. Suspension cultures have a number of advantages compared to monolayer cultures (Freshney, 1983a). The most important being the production and harvesting of large quantities of cells, achieved without increasing the surface area of the substrate, and, if dilution of the culture is continuous and the cell concentration kept constant, a steady state can be achieved.Thus, exposures to drugs at specific phases of the cell cycle can be performed.

Another cell culture technique which has been used for drug sensitivity studies is the multicellular spheroid. Two major techniques have been employed to cultivate multicellular spheroids, namely the spinner flask technique and the liquid overlay technique. Spheroids, which may consist of several hundred cells (Dewey, 1979), are important because they are said to simulate solid tumours in vivo and are therefore useful as model systems for studying radiation sensitisers (Sutherland, 1980). Spheroids are also useful for drug sensitivity studies because it is possible to assess the effects on the different zones of cells within a spheroid and thus measure drug penetration in solid tumours, which has been proposed as one of the mechanisms of drug resistance (Erlichman and Vidgen, 1984).

b. Media

A large variety of cell-culture media is commercially available, the composition of which may affect the results derived from cytotoxicity tests (Hakala and Rustum, 1979).These media differ with regard to the number of simple compounds, such as amino acids, to various bizarre ingredients in some of the more complex media. Many of the medium additives appear to be non-essential, for example, it is excessive to provide cells with both glutamine and glutamic acid. Glutamine is taken up by cells approximately ten times more readily than glutamic acid and is then rapidly deaminated in the cell to glutamic acid (Hakala et al, 1974). Also, some media contain phosphorylated compounds such as nucleotides or sugar phosphates which, unless the phosphate has been removed, enter cells poorly (Hakala and Rustum, 1979) and therefore appear wasteful.

Whilst considering additives such as those mentioned above, the choice of medium used for in vitro drug sensitivity assays has to be made carefully because other components in commercially available media may protect cells from the cytotoxic effects of

certain anti-cancer drugs. Some media, such as Ham's F—10 (Ham, 1965), CMRL (Parker et al, 1957) and Waymouth's (Gorham and Waymouth, 1965), contain thymidine and/or hypoxanthine (deoxyadenosine) at concentrations which may influence the expression of cytotoxicity of antimetabolite drugs such as methotrexate, 6—mercaptopurine and 5—fluorouracil. The effect on the cytotoxicity of these and many other antimetabolites have been reviewed by Hakala and Rustum (1979).

The source of some drugs, for instance asparaginase, may affect essential constituents in the cell—culture medium and thus influence the drug sensitivity data. In a discussion point raised at the 1970 symposium on Prediction of Response in Cancer Therapy, to the paper presented by Pragner (1971), it was said that bacterial asparaginase, such as that from Escherichia coli, destroys glutamine in the tissue culture medium. Thus, what appears to be a positive test result may in fact be due to glutamine depletion.

Some investigators routinely add antibiotics and antifungal agents to tissue culture medium. Although this may be necessary when establishing a new cell line or growing tumour cells from primary cultures, I do not believe it is necessary or advisable for drug sensitivity assays employing continuous cell lines. The use of such drugs may not only conceal a low density microbial contamination, which itself may influence the drug sensitivity results, but they may also directly or indirectly affect the cytotoxic potential of the anticancer drugs. It has been shown, for example, that amphotericin B, commonly employed to suppress fungal contamination in tissue culture medium, interacts with cell membranes and this may modify the effects of some drugs by increasing membrane permeability (Medoff et al, 1975). The effects of antibacterial and antifungal agents on the cytotoxicity of anticancer drugs can also be extended to the clinical situation, for example, penicillin, which is chemically similar to methotrexate, blocks the proximal renal tubular secretion of methotrexate, therefore delaying elimination of the

drug. The net effect would be to induce altered pharmacokinetics of methotrexate, however, significant blockage will only occur if high doses of penicillin are administered (Williams et al, 1984). The cytotoxicity of a number of antibiotics has also been investigated by Abaza et al (1978) who compared the effects of a range of concentrations and combinations of antibiotics against three human bladder carcinoma cell lines.

iii. Serum and hormones

a. Serum

It has been shown that serum, which is widely used as a constituent in the medium used for in vitro cell culture, can affect the data derived from clonogenic assays. It can influence drug sensitivity data in particular in two ways: firstly, the concentration of the serum can have a direct influence on the proliferative rate of cells and secondly, serum can interact with anticancer drugs, thus affecting their apparent biological efficacy.

The addition of serum to a monolayer of 3T3 cells was shown to initiate DNA synthesis in some of the cells (Todaro et al, 1965) and it was also demonstrated that the maximum size of a cell population was determined by the concentration of the serum (Holley and Kiernan, 1968; Castor, 1971). This was true for both normal and transformed 3T3 cultures, although the dependence on serum factors was greatly reduced in the latter. The fractionation of serum has shown that its constituents comprise a wide variety of proteins and hormones (see Riley, 1981). They are essential for the proliferation of mammalian cells, although reliance on particular serum factors appears to vary from one cell type to another. However, serum has been shown to influence the cytotoxic activity of anticancer drugs. Using human mesothelioma cells Nissen and Tanneberger (1981) demonstrated the effects of calf and foetal calf serum on vinblastine activity by

measuring the inhibition of % —thymidine incorporation. In a more detailed study, Takahashi et al, (1980) reported the effects of human serum albumin on the biological activity of 13 widely used chemotherapeutic drugs. These drugs could be categorised into three groups according to changes in their biological activity influenced by human serum albumin: Type I, drugs whose biological activity was reduced in the presence of albumin, such as cisplatin, adriamycin and mitomycin C; Type II, drugs that were not influenced by albumin, such as 5—fluorouracil; Type III, drugs that showed an increase in biological activity in the presence of albumin, such as bleomycin and vinblastine.

Because the biological activities of anticancer drugs are influenced, it is reasonable to infer that their clinical effectiveness is probably also affected. However, no correlation between the molecular weight of a drug and the influence of human serum albumin on its biological activity was found. One problem which may occur in vivo if other unrelated drugs with a higher binding capacity are administered, is the displacement of the initial anticancer drug — albumin complex, resulting in an increased clinical toxicity. The effects of serum are not only restricted to anticancer agents. Haspel et al (1984) reported a decreased cytocidal effect of the male oral contraceptive, Gossypol, on murine erythroleukaemia cells. Many of the problems encountered with serum in drug sensitivity studies can be circumvented in vitro with the use of a serum—free, chemically defined medium. For example, HITES medium has been used to selectively support the growth of small cell lung carcinoma cells (Carney et al, 1981). Another example, DH—SI was used by Messing et al (1982) to establish a human transitional cell carcinoma cell line without the contamination of fibroblasts. Inhibited fibroblast growth in primary cultures has also been reported by other authors (Taub et al, 1979; Bottenstein et al,1979). There have also been a number of publications which have

reported the use of various serum—free media formulations for growing normal human cells, such as epidermal keratinocytes (Maciag et al, 1981} Tsao et al, 1982) and bronchial epithelial cells (Lechner et al, 1982).

b. Hormones

Serum usually employed as a supplement in tissue culture medium contains various hormones and growth factors at concentrations usually adequate for cell growth. However, as mentioned in the previous section, there may be circumstances when it would be preferable to use serum—free medium. Therefore, in order that cellular growth can be maintained, culture medium will have to be supplemented with a number of hormones and growth factors normally found in serum (see Riley, 1981). A major characteristic of transformed cells is their reduced requirement for serum, nevertheless, in cell cultures derived from tissues, which in vivo depend on target—specific hormones, such as mammary tissue, it is often necessary to add the appropriate hormone(s) to the culture medium. In addition to the general protein hormones, such as insulin, and the target—specific hormones, such as ACTH and steroids, e.g. oestrogen, progesterone, there are a number of growth factors. However, it is not clear to what extent these growth factors act independently of each other.It is known that many cells exhibit epidermal growth factor (EGF) receptors (Hollenberg and Cuatrecasas, 1975), and the biological effect of EGF is proportional to the degree of binding to the plasma membrane receptors. It would appear therefore, that EGF is an essential requirement for the growth of many cell types in serum—free medium.

Evidence that cultures were able to grow in the absence of serum, if the medium was supplemented with appropriate growth factors, was provided by the work of Barnes and Sato (1979).Using a mammary cell line, MCF—7, they demonstrated that the proliferation of cultures grown in serum—free medium containing

physiological levels of insulin, transferrin, fibronectin, EGF and prostaglandin F2 was identical to that of cultures in medium supplemented with foetal bovine serum. Another cell line, ZR—75— 1, derived from a human mammary cancer (Engel et al, 1978), could be grown indefinitely in serum—free medium, but the growth required the presence of 170-oestradiol, transferrin, L— triiodothyronine (T3) and insulin (Lippman et al, 1979).

Cell lines established from human neoplasms have been shown to retain certain characteristics of the parent tumour, such as drug sensitivity profiles (Carney et al, 1982, 1983). However, hormones, which would be required in serum—free medium, are known to influence the growth of certain tumour cells in vitro, e.g. breast carcinoma, and may alter their chemosensitivity patterns (Klevjer—Anderson and Buehring, 1980). Calvo et al (1983) evaluated the influence of hormone addition on the chemosensitivity of the human breast cancer cell line MCF—7.They found that despite an increase in the colony—forming efficiency, in vitro chemosensitivity of the cells to a number of cytotoxic drugs was similar in hormone supplemented and non— supplemented medium. Insulin, which is used as a supplement in the clonogenic assay procedure of Hamburger and Salmon (1977), has been shown to increase the clonogenicity of tumour cells (MacKintosh et al, 1981). However, not all hormonal factors are growth stimulators. It has been shown that adrenocorticotropin (ACTH) has an inhibitory action on the target cells of the adrenal cortex. Ramachandran and Suyama (1975) and Liao (1975) demonstrated that glucocorticoids are inhibitory to fibroblast growth. Nevertheless, selected hormonal factors are necessary supplements for in vitro studies using cell lines in serum—free medium.

iv. Scheduling and period of exposure

The clonogenic assay procedures most often adopted are either (a) exposure of exponentially—growing cells to an

anticancer drug followed by transfer as a single cell suspension to fresh dishes to form colonies or (b) allow cells to form colonies in situ after drug exposure (Twentyman, 1979).Depending on the type of anticancer drug used the clonogenic assay procedure adopted can influence drug sensitivity measurements. Thus, careful evaluation of such data are important if clonogenic assays are to be used for comparing and evaluating new drugs (Venditti, 1983). Using the in situ type of assay it is desirable to allow a period of time before drug exposure, sufficiently long enough so that cell cycle perturbations, caused by enzyme detachment and the re-attachment of cells, have disappeared. It has been shown (Barranco et al,1980) that clonogenic cell survival values varied depending on the time which elapsed between exposure to enzymes and drug treatment. However, during this period some cell division may occur, thus, a correction factor to take this into account will have to be applied (Szechter et al, 1978). Another important time variable, which applies to the transfer type assay, is the period between the end of drug exposure and the transfer of cells for colony formation. The delay between drug treatment and cell transfer may permit some clonogenic cells to recover from potentially lethal damage. Hahn and Little (1972) were the first to demonstrate an apparent increase in cell survival after a transfer delay period of 4 — 6 hours, however, no further increase in survival occurred after a six hour delay (Hahn,1976). The fact that the fraction of surviving clonogenic cells can be influenced by transfer delay periods has since been shown by other investigators (Barranco et al, 1975; Twentyman, 1979). Nevertheless, cell survival following the exposure of some anticancer drugs appears not to be affected by transfer delay times, e.g. adriamycin (Hahn, 1976).

The length of time cells are exposed to anticancer drugs is also an important factor that can influence clonogenic cell survival (Hahn et al, 1973; Hill et al, 1981; Calabro—Jones et al, 1982). Using murine neuroblastoma cells Hill et al (1981) showed that for certain drugs, e.g. methotrexate, vincristine and VM—26, negligible cell kill occurred within a 1 hour exposure

time and thus, was unsuitable for evaluating sensitivity compared to 24 hour drug exposure times. This observation was also made for cytosine arabinoside, 5—fluorouracil, methotrexate and hydroxyurea in human lymphoma cells (Drewinko et al, 1979), and hydroxyurea in ovarian tumour cells (obtained from ascitic fluid) and a colon carcinoma cell line (Rupniak et al, 1983b). However, drugs known to cause plateau—type dose response curves with short exposure periods failed to do so when given by continuous exposure (Ludwig et al, 1984b), therefore, any distinction between cytotoxic and cytostatic drug effects is not possible.In addition, as yet, in vitro continuous exposure schedules have not been shown to be predictive of clinical response for any agent or tumour type (Ludwig et al, 1984b). Therefore, as a general rule, an assay of cell survival carried out 24 — 48 hours after drug administration is likely to give the best answer for all classes of antitumour drugs (Twentyman, 1980).

v. Growth status (log/plateau) and density

Increasing tumour cell growth in cell culture systems, where medium replacement is not continuous, will eventually result in a decrease in the proliferative rate of the cells. This is primarily due to the depletion of essential nutrients (Riley,1981) and not, as in the case of the normal cells, due to density dependent (Canagaratna and Riley, 1975) or contact inhibition of growth (Schutz and Mora, 1968; Castor, 1971). The proportion of cells proliferating at the time of drug treatment is significant. In general, cells in exponential growth are much more sensitive to those in stationary or plateau phase growth (see references cited by Drewinko et al, 1981; also section l.E iii). Nevertheless, the efficacy of some anticancer drugs is also affected by the density of exponentially—growing cells.Using the Burkitts lymphoma cell line, DND—39A, Arkin and co- workers (1984) exposed different cell densities to a number of cytotoxic drugs and measured the clonogenicity of surviving cells in soft agar. They found that cell kill by cisplatin,

carboplatin and melphalan was independent of cell density. However, vincristine, adriamycin, daunorubicin, mitoxantrone and bleomycin all became progressively less efficacious with increasing cell densities. Similar findings using three human leukaemia—lymphoma cell lines were also reported by Ohnuma and co—workers (1986).

vi. Feeder cells

It has been suggested that feeder cells provide diffusible factors that are necessary or improve cell growth (Puck and Marcus, 1955). In a later publication they reported on the clonal growth of mammalian cells with and without feeder cells (Puck et al, 1956). The feeder cells used were x—irradiated HeLa cells. These non—multiplying cells were said to "condition" the medium so as to permit the single cells to reproduce to the point at which they eventually became self-sustaining. However, although considerable attention has been paid to the identification of the diffusible factors produced by lethally irradiated cells, Puck et al (1956), concluded that although their use was an advantage for improving colony growth they were not essential. The reason for this was because the necessity to "condition" the medium reflects only its lack of nutritional adequacy. Therefore, one should investigate what changes in growth conditions will permit single cells to multiply in the absence of feeders.

Another cell type which has been extensively used as a feeder cell Is that derived by Todaro and Green (1963) from mouse embyro tissue and designated 3T3. These cells can be used In clonogenic assays to overcome the problem of density dependent growth, i.e. when the colony forming efficiency of tumour cells is dependant on the number plated. However, the feeder cells may influence clonogenic cell survival following drug exposure, for instance by metabolic co-operation. The existance of a diffusible factor from 3T3 cells has been demonstrated to play a

role in the control of macromolecular synthesis, and hence, in cell replication (Yeh and Fisher, 1969). In a study comparing clonogenic assay procedures the addition of 3T3 cells had little effect on the dose—response curves of cells exposed to adriamycin and methotrexate (Hepburn and Masters, 1986). Nevertheless, there have been reports in the literature that after treatment with radiation (Deen et al, 1979), alkylating agents (Wheeler et al, 1975) and antimetabolites (Weizsaecker and Deen, 1980), the use of feeder cells has affected clonogenic cell survival. Thechanges in the dose—response curves observed by Weizsaecker and Deen (1980) were most noticable when dialysed serum was used,I.e. under sub—optimal growth conditions when the maximum colony- forming efficiency was not attained.

In other clonogenic assay systems it has been reported that adherent and phagocytic cells are required for optimal tumour colony growth (Hamburger et al, 1978). In a subsequent publication Hamburger et al (1983) showed that irradiation enhanced the ability of adherent cells to support tumour colony growth. They suggested the most plausible explanation is that the cells which suppress colony growth, either directly or indirectly, may have been inactivated by the irradiation. Also using the clonogenic assay described by Hamburger and Salmon (1977), Uitendaal et al (1983) reported an increase in the plating efficiency of ovarian cancer cells by the addition of cell—free ascites in the top agar layer.

vii. Proteolytic enzymes

Tumour disaggregation problems are usually associated with the production of cell suspensions from primary tumour samples and xenografts. However, it is sometimes necessary to disaggregate the cells that compose multicellular spheroids for assessment of drug effects by clonogenic assay. Briefly, there are two basic approaches; mechanical separation of cells or enzymatic treatment. Mechanical methods, which are quick and simple, usually consist of slicing the tissue and then mincing

with crossed scalpel blades. This normally releases cells loosely held in the tissue matrix which can easily be washed through a mesh screen. Further mechanical treatments may improve the total cell yield, but often result in an increase in the number of non—viable cells (see references cited by Slocum et al, 1981b). Enzymatic procedures employ either one of a number of different enzymes or enzyme cocktails. Those enzymes most commonly used are trypsin, pronase, collagenase and DNAse. The choice of enzyme or enzymes, all of which may have deleterious effects on cells, appears to be dependent on the particular type and/or source of tissue (see references cited by Slocum et al, 1981b). MacKintosh and co-workers (1981) working with Lewis lung carcinoma produced cell suspensions yielding 95% viability with an enzyme mixture of pronase, collagenase and DNAse. This compares with 43% viability using 0.05% trypsin and 10% viability using the purely mechanical methods of Salmon et al (1978). A two-step method for obtaining large numbers of viable cells from human melanoma, sarcoma and lung tumours was developed by Slocum et al (1980, 1981b). This group also compared cells released mechanically and enzymatically from human melanoma, sarcoma, lung, colon and breast (Slocum et al, 1981a). Generally, a much higher yield of cells was obtained enzymatically than by mechanical disaggregation, the latter of which resulted in variable cell yields from individual tumours, even within the same pathological type. In addition, higher ribonucleotide triphosphate pools were exhibited in enzymatically released cells which also displayed less DNA damage. Nevertheless, clonogenicity in semi—solid media was shown to be similar for all tumour types they studied.

Using the clonogenic assay procedure, exposure to enzymes (usually trypsin) is likely at some time prior to or after drug treatment. These enzymes have been shown to cause a number of effects which may influence drug sensitivity survival data. Differences in cell membranes or surface components of dividing and non—dividing cells may affect the binding and transport of

some drugs into the cells and thus, ultimately cause rate- limiting differential survival responses (Barranco et al, 1980). Working with the bacterium E coli, Leive (1965) demonstrated that exposure to EDTA removed a cell surface component with a resultant increase in permeability to drugs such as actinomycin D. It has also been shown that mild trypsin treatment of ^SO^— labelled SV 3T3 cells removed 68 —84% of the cellular sulphated glycosaminoglycan from the cell membranes (Roblin et al, 1975).In addition, trypsin has been shown to cause the release of glycopeptides from the surface of plasma membranes (Kraemer, 1971). These findings may have important implications for in vitro drug sensitivity assays because glycoproteins, which are synthesised intracellularly, may have a function in the control of cell proliferation (see references cited by Barranco et al,1980). Using the mouse lymphoma cell line, L5178Y, and its actinomycin D—resistant subline, L5178Y/D, it has also been demonstrated that glycoproteins are Involved in drug resistance mechanisms (Kessel and Bosmann 1970; Bosmann, 1971). The effect of trypsin on the drug survival responses of CHO cells has also been investigated (Barranco et al, 1980). Using a number of cytotoxic agents, cells were drug treated either before exposure to trypsin or at various time intervals after. Their results showed a deviation from the normally expected survival values ranging from 4 to 50 fold. Nevertheless, greater than 12 hours post—trypsinisation the drug survival responses had returned to the expected values.

viii. Hydrogen ion concentration (pH)

Maintaining a hydrogen ion concentration near neutrality is essential for many of the metabolic processes performed by mammalian cells and it would appear that the extracellular pH environment is important in influencing the behaviour of those cells, a feature which has been investigated by Eagle (1973“).The growth and plating efficiency of normal cells exhibit an external pH optimum of approximately 7.7. Small fluctuations

from this optimum can reduce the plating efficiency by 20%. Evidence that the action of H1- is external has been shown by Manfredi (1963) whose data suggest that the intracellular pH of human cells is near 6.9. Virus—transformed cells and spontaneous tumours exhibit a lower pH optimum of 7.2 (Ceccarini and Eagle, 1971; Rubin, 1971; Eagle, 1973). The effect of pH on cellular growth has also been demonstrated (Siskin and Kinosita, 1961). They showed that the phase of the cell cycle was the phase most influenced by external pH values, thus, in vitro drug studies, using drugs whose greatest efficacy is against cells in the G^ phase, may be influenced. It has also been demonstrated that both initiation and rate of DNA synthesis is increased in cells exposed to pH values higher than the cell's optimum (Rubin, 1971).

The experimental data cited above provides evidence that results obtained from in vitro clonogenic assays may be influenced because changes in environmental pH may alter cellular growth rates. In addition, there has also been experimental evidence that demonstrates the effect of pH on the cytotoxicity of anticancer drugs. Nissen and Tanneberger (1981) studied the influence of pH changes on the cytotoxicity of 5—fluorouracil by measuring cellular incorporation of H—uridine. Using Chinese hamster cells (HA—1), Hahn and Shin (1983) measured the cell- killing ability of a number of cytotoxic drugs in a range of pH conditions. They showed that the cytotoxicity of methotrexate was not influenced by pH, whereas, bleomycin was most effective at acidic pH (less than 7.5). On the other hand, the cytotoxicity of amphotericin B was greatest at higher and lower pH values than that of normal tissue (7.4). Thiotepa, which represents one of the oldest chemotherapeutic drugs still used in clinical oncology, was shown to be relatively stable at pH values between 6 and 7, but drug concentration decreased by 40% at pH 5 (Cohen et al, 1984). However, the alkylating activity, as assessed by p—nitrobenzyl pyridine activity, remained stable at both the pH values and incubation temperatures tested. These

results therefore may have some relevance in the design of in vitro drug sensitivity assays and in vitro — in vivo correlations.

ix. Oxygen

The oxygen transport for cells cultured in vitro depends on the kinetics of gas diffusion through the overlying medium. It has been calculated, using liver parenchymal cells, that the oxygen requirement would not be met by the atmospheric partial pressure of oxygen if the medium depth was greater than 0.34 mm (Stevens, 1965). Under a standard medium overlay depth of 1.525 mm, McLimans et al (1968) calculated that a monolayer of respiring cells would exhaust the atmospheric oxygen supply in 35 minutes, suggesting growth would be dependent on the supply rate of oxygen. The micro-environment of oxygen for cultured cells incubated at 35°C in a humidified atmosphere of 5% CO2 in air has been investigated by Werrlein and Glinos (1974). Using oxygen- sensitive electrodes they measured changes in the partial pressure of oxygen in the medium and detected oscillating concentration gradients of oxygen in the region just above the attached cells. It would appear that the oscillations are due to the reduction in micro-environmental oxygen supply, an effect that was more pronounced in dense cultures than for attached subconfluent cells. The oscillations suggest that large parts of the culture respire in synchrony and they proposed that the cellular respiration in monolayer cultures responds to environmental feedback and is modified by local exhaustion of oxygen.

Using the soft—agar colony assays human tumours in primary culture have consistently shown improved colony formation when gassed with mixtures containing 3 to 5% oxygen (Richter et al, 1972; Tveit et al, 1981a; Courtenay, 1983, 1984; also see references cited by Gupta and Eberle, 1984). This is perhaps not surprising when the majority of cells in a solid tumour are in a

microenvironment with oxygen concentration ranging from 0.1 to 5% (Gupta and Eberle, 1984) and may reach levels below 0.009%(Millar et al, 1982). Oxygen can be toxic at higher than physiological concentrations due to the effects of superoxide anion, however, superoxide dismutase can, with varying degrees and depending on cell type, protect cells from oxygen toxicity (Gupta and Eberle, 1984).

The cellular effects of oxygen also have important implications for drug sensitivity data measured by clonogenic assays. It has long been considered that cells deep in a solid tumour are a major source of tumour resistance due to hypoxia as a result of poor vascular supply. Therefore, in vitro studies measuring anticancer drug action under hypoxic conditions is of great importance. With this in mind, a number of investigations have studied radio and chemo—sensitisers, such as misonidazole, which can potentiate the effectiveness of many conventional chemotherapeutic drugs (Sealy et al, 1984; Mulcahy et al, 1984; Mulcahy, 1984). In addition to these radiation and chemosensitisers, it has been shown that the conventional chemotherapeutic drugs may have selective cytotoxicity for hypoxic or aerobic cells. Teicher et al (1981) classified drugs into groups depending on the preferential toxicity to oxygenated or hypoxic cells. Using Chinese hamster ovary cells, similar studies on the effectiveness of anticancer drugs under aerobic or hypoxic conditions have also been investigated (Tannock and Guttman, 1981). It was generally concluded that some drugs under hypoxic or aerobic conditions exhibit no difference in sensitivity, e.g. 1,3—bis(2—chlorethyl 1)—1-nitrosourea (BCNU) and cisplatin. However, some drugs, such as bleomycin, showed an increase in toxicity under aerobic conditions, and others, such as the radiosensitisers, e.g. misonidazole, were more effective under conditions of hypoxia. Nevertheless, for some anticancer drugs the evidence was not consistent. Adriamycin was said by Teicher et al (1981) to be more toxic under hypoxic conditions, whereas Tannock and Guttman (1981), in agreement with other

reports, suggested that adriamycin exerts its greatest efficacy in aerobic conditions. Differences in preferential aerobic conditions for other drugs have also been reported, for example, cisplatin (Stratford et al, 1980) and mitomycin C (Ludwig et al, 1984a). However, the results for some of the anticancer drugs may be cell line dependent.

x. Temperature

As the ambient temperature is increased above 37°C, many, although certainly not all, anticancer drugs become more cytotoxic (Hahn and Li, 1982; Hahn and Shin, 1983; Hahn, 1984; Cohen et al, 1984). For some drugs 43°C represents a threshold below which there is no increase in cytotoxicity, whilst above 43°C there is. In addition, some cells may be thermotolerant (Gerner and Schneider, 1975) and thus may respond differently to drugs compared to cells that have not been pretreated. This phenomena has been discussed by Bertino et al (1984). Due to the reasons mentioned above it is important that clonogenic assays are performed at a physiological temperature of 37°C and that this temperature remains constant.

At pH values lower than that found in normal tissues Gerweck (1977) demonstrated that the sensitivity of cells to elevated temperatures was enhanced. In addition, as cells become nutritionally deprived they are also less able to withstand hyperthermic conditions (Hahn, 1984). Both lower pH values and nutritional deprivation characterise the state of many cells found in the interiors of solid tumours, thus these results suggest that excess heat should be effective clinically in killing some cells in such tumours (see Hahn, 1974; Oleson,1984). Survival curves of mammalian cells exposed to temperatures above 43°C resemble those obtained using x— irradiation, with the width of the shoulder and the steepness of the slope dependent on the temperature, whilst below 43°C the curve expresses a resistant "tail" (Sapareto et al, 1978).

xi. Concentration of agar

There are two main types of soft—agar assays which have been used to quantitate the differential sensitivities of a range of human tumour cells to anticancer drugs (Hamburger and Salmon, 1977; Courtenay et al, 1978). However, one of the major problems, which is generally accepted, is that the proportion of clonogenic cells within a heterogeneous tumour cell population is probably less than 1% and the plating efficiencies of assays using material from human tumours is often less than 0.01% (Hamburger et al, 1978; Buick et al, 1979).

In a comparison of the two major techniques (Tveit et al, 1981a), the method of Courtenay and Mills (1978) gave considerably higher plating efficiencies than that of Hamburger and Salmon (1977). However, this increase in plating efficiency was attributed to the low 0£ concentration and presence of rat red blood cells in the Courtenay assay. Nevertheless, attempts were made to increase the number and quality of colonies formed in soft—agar assays by varying the concentration of agar. Using a series of five continuous human cell lines Whelan and Hill (1981) demonstrated that lowering the concentration of the more purified agarose from 0.3 to 0.16% the colony—forming efficiency increased 2 — 10 fold. These findings were confirmed and extended in a subsequent publication (Hill and Whelan, 1983) where they also showed that reducing agar concentration from 0.3 to 0.17% also resulted in a significant increase in colony- forming efficiency. The standard Courtenay assay proved to be the best system with the highest plating efficiencies recorded, however, further attempts to increase the cloning efficiency of the cell lines by reducing the agar concentration only showed a marginal increase. Methylcellulose has also been used at 0.18% (Buick and Fry, 1980), the use of which was said to facilitate the plucking out of colonies for further studies. Whilst this may be true, the use of low concentrations of agarose have also been shown to facilitate the removal of individual colonies (Whelan and Hill, 1981).

E. Relationship between in vivo and in vitro cytotoxicity

i. Drug storage and stability

Clonogenic assays have proven useful in the evaluation of the chemosensitivities of primary cells from human cancers and continuous cell lines to both standard and experimental anticancer drugs, and also, in the prediction of clinical tumour response to these drugs (see references cited by Ludwig and Alberts, 1984). All tumour chemosensitivity assays performed in the laboratory require the frequent use of anticancer drugs, often at short notice, many of which are very expensive. Therefore, for reasons of both economy and convenience many of these drugs are stored in aliquots for use as required. No standard method of drug preparation or storage appears to be followed. Some laboratories use samples of clinical drugpreparations made up fresh prior to each assay, as performed instudies for this thesis, while others store frozen aliquots of diluted drug (Yang and Drewinko, 1983? 1985).

It is clear from a number of studies that storage containers, method and time may affect the stability of the lethal activity of some drugs (Benvenuto et al, 1981; Bosanquet, 1986). Benvenuto and co—workers (1981) determined the stability, over 24 hours, of 13 anticancer drugs (48 hours for adriamycin) in both glass and plastic containers. All drug mixtures were kept at room temperature and not protected from light. They measured drug stability by high-pressure liquid chromatography, except cyclophosphamide, which was analysed bymass spectrometry, and carmustine, which was analysed byspectrophotometry. For the majority of the drugs tested there was no significant difference between stability after storage in glass or plastic containers. However, there were exceptions, for example, carmustine and bleomycin were shown to be more stable in glass, and adriamycin and 5—fluorouracil in plastic containers. In addition, mitomycin C, when dissolved in 5%

dextrose was not stable In glass or plastic containers, whereas, when dissolved in 0.9% sodium chloride, mitomycin C was shown to be more stable in plastic containers. However, Bosanquet (1986), in a report on the stability and storage of antineoplastic agents, did not agree with some of these findings, in particular adriamycin, which was said to bind strongly to plastic and cellulose filters.

Using cells from an induced rat sarcoma in a soft agar clonogenic assay system Franco and co-vorkers (1984) determined the stability of eight common anticancer drugs after storage at —60°C. Survival of clonogenic cells was compared using freshly prepared drug, freshly frozen drug and drug stored frozen for varying lengths of time from 10 — 61 weeks. For the drugs tested they found no loss of potency following storage at —60°C for a minimum of 10 weeks. However, both melphalan and 5—fluorouracil have been shown to lose 50% of their biological activity within 2 weeks regardless of whether they were stored frozen or refrigerated (Yang and Drewinko, 1985).

Evidence of both the biological and chemical activity of a number of anticancer drugs has been presented by Ludwig and Alberts (1984). Samples of culture media were obtained immediately, 6 and 24 hours, 2, 4 and 10 days after incubation and then frozen in liquid nitrogen. Biological activity was determined by the human tumour stem cell assay using an endometrial carcinoma and melanoma cell line. Chemical activity was analysed by high-pressure liquid chromatography or radioimmunoassay techniques. Actinomycin D, bisantrene, bleomycin and vinblastine retained their biological and chemical activity even after 10 days incubation in culture media, whereas, etoposide lost 60% of its biological activity over this period and its chemical activity decreased to zero over 10 days. Due to technical difficulties an accurate quantitation of the chemical activity of adriamycin was not possible, nevertheless, it retained greater than 80% of its biological activity over the 10 day incubation period.

ii. Drug metabolism

The majority of anticancer drugs do not have to be metabolically activated before they become cytotoxic to cells in culture. However, there are exceptions. The best known, perhaps, is cyclophosphamide which is converted in vivo by the hepatic mixed—function oxidase system to 4— hydroxycyclophosphamide (Sladek, 1971; Colvin et al, 1973;Connors et al, 1974). Further metabolites are then produced, some of which are non-cytotoxic (Pratt and Rudden, 1979a). Nevertheless, cyclophosphamide can be used in in vitro chemosensitivity studies by pre—incubating with liver microsomes (Philips, 1974) or by using a derivative such as 4— hydroperoxycyclophosphamide which does not require metabolic activation (Kaufmann et al, 1975). For some other cytotoxic drugs there appears to be some confusion as to whether the drug has to be biologically activated or not. Mitomycin C is one such drug. In vitro it has no effect on purified DNA unless a cell extract is added (Pratt and Rudden, 1979b). Kennedy et al (1980a,b), using murine cell lines, EMT-6 and S—180, proposed that mitomycin C is preferentially activated and metabolised by hypoxic tumour cells. The essential step for the biological activity of the drug was said to be the reduction of thebenzoquinone ring of the mitomycin molecule todihydrobenzoquinones which preferentially occurs in the decreased oxygen tension found in hypoxic tumour cells (Lin et al, 1976).However, Ludwig et al (1984a), using the murine cell line EMT—6and the human colon carcinoma cell line WiDR, in a human tumour stem cell assay, concluded that the potentiating effect of hypoxia on mitomycin metabolism and biological activity may be peculiar to the murine cell lines and that differential efficacy against hypoxic human carcinoma cells was unlikely.

iii. Classification of antitumour agents

Bruce and his co- workers (1966) were one of the first groups to demonstrate that the cytotoxic effects of antitumour

agents may be dependent on cell proliferation. Using a quantitative spleen—colony forming assay in the mouse they assessed the sensitivities of both normal haematopoietic stem cells and lymphoma colony—forming cells to a 24 hour exposure of a variety of cytotoxic agents. Based on their results they classified anticancer agents into three groups, depending on the survival curves obtained, as shown in Figure 1.3. Class I — No difference in toxicity exerted against normal haematopoietic and lymphoma colony—forming cells, i.e. equally toxic for proliferating and resting (Go) cells. These agents were said to be non-specific, (e.g. x—irradiation); Class II — Agents exerted greater toxicity to lymphoma colony—forming cells than normal haematopoietic cells, with increasing concentrations killing a greater number of cells until a plateau was reached where there was no further increase in cell kill. The fraction of cells surviving doses in the plateau range is said to be equal to the fraction of insensitive cells in the population (Gray, 1983). Resting (Go) cells do not appear to be affected provided exposure time is short (24 hours), whereas proliferating cells are killed during a specific phase or phases of the cell cycle; Class III — Survival curves were exponential but there was a significant difference in the slope for the normal haematopoietic and lymphoma colony—forming cells, the latter being more sensitive. These agents were termed cycle specific since although they damaged both proliferating and resting cells, the former were much more sensitive and were killed throughout the cell cycle. This specificity appeared to be due to the fact that in untreated mice most haematopoietic stem cells were resting whilst all detectable lymphoma stem cells appeared to be proliferating (Bruce and Valeriote, 1968). However, this kinetically exploitable difference applied only after 24 — 36 hours, as after that time haematopoietic stem cells were recruited to restore the damaged proliferating cell population. Using cell lines this kinetic classification of antitumour agents has been confirmed and extended (Valeriote and Tolen, 1972; Hill, 1978a; Hill et al,

Figure 1.3

Survival of

mammalian

cell reproductive capacity after

24 hou

r exposure to

antitumour agents

(after Bruce

et al

, 1966).

H A

H

o —

iO

u o u o e j j j B A i A j n g | | a 3 B o n

io IO • o

k

1981; Hill and Whelan, 1981; Hill and Price, 1982; Metcalfe et al, 1983; Hepburn et al, 1985). The kinetic classification, as originally proposed by Bruce and his co-workers (1966), of drugs used in the studies presented in this thesis are listed in Table 1.6, and the cycle stages in which the lethal effect and progression delay or arrest are preferentially exerted are given in Table 1.7.

A more detailed analysis was recently suggested (Mauro et al, 1983) based on a more specific definition of drug effects in phases of the cell cycle:(a) — Agents showing a lethal effect in only one stage of thecell cycle and progression delay at the same stage;(b) — Agents lethally effective in more than one stage of thecell cycle and progression delay of one or more points along thecell cycle;(c) — Agents lethally active to the same extent in all stages of the cell cycle;(d) — Agents active on non—proliferating cells.

Using an asynchronous, exponentially—growing population of human T^ lymphoma cells (Trujillo et al, 1972) and a 1 hour exposure time, Drewinko et al (1979) classified antitumour agents according to five patterns of cell survival, see Figure 1.4. The survival curves they obtained were described as:Type A — Simple exponential, e.g. adriamycin;Type B — Biphasic exponential, e.g. bleomycin;Type C — Threshold exponential, e.g. melphalan;Type D — Exponential plateau, e.g. methotrexate;Type E — Ineffectual, e.g. arabinofuranosyl cytosine (ara—C).They suggested that the shape of the survival curve is a function of the biological properties of the treated cells and the physiochemical characteristics of the antitumour drug. Although guidelines were given for the design of chemotherapeutic protocols based on the five survival patterns they, however, regard its clinical application to be premature without further appropriate investigations.

Table 1.6 Kinetic classifiation of antitumour drugs used in these studies*5

Class II (cell cyle phase—specific)

MethotrexateVP16-21354—Hydroxyanisole

Class III (cell cycle non-specific)

AdriamycinCisplatinCBDCA5Dibromodulci tol5—Fluorouracil Mitomycin C Mitoxantrone Bleomycin+ Spirogermanium+

5 Abbreviations used: CBDCA = cis—diammine—1,1—cyclobutanedecarboxylate platinum II.

VP16—213 = 4—demethyl epipodophyllotoxin— 0—D—ethylidene glucoside.

55 Results obtained from using either in vivo spleen colony- forming assays and/or in vitro assays for colony—forming ability in a range of different cell types (adapted from Hill, 1978a).

+ Not myelosuppresive but causes exponential kill of tumour cells.

Table 1.7 Phases of the cell cycle where certain anti-tumour agents exert their effects (Adapted from Hill, 1978a)

| DRUGSPredominant phase(s) of cell cycle where effect occurs |

Lethality*1 1 | Progression delay** |1 1

| Adrlamycin Late S. M1 1 1 s, g 2 1I I

| Bleomycin ^2 * h1 1 1 S/G2, C2 1| I

| Clsplatln G l» g21 1 I S. G2 . Gi/S | 1 |

| CBDCA* -1 1

| Dlbromodulcltol G l. s G l/S 1 { |

1 5-Pluorouracil All phases1 1 i Gi / s 11 1

| 4-Hydroxyanisol S1 1

| Methotrexate Early G^t Gj/S, S i G ^ s | 1 1

| Mitomycin C Gl t ^2• M 1 s /G2 1 1 1

| Mltoxantrone -1 1

| Spirogermanium -1 1

1 1| VP16-213* s . g2

1 1

1 S, g 2 1 1 1

* Abbreviations used: CBDCA ■ Cis-Diamine-l,l-cyclobutane decarboxylateVP16-213 ■ 4-demethyl-epipodophyllotoxin-0-D-

ethylidene glucoside 11 Lethality measured on synchronous cell populations

** Progression delay measured on asynchronous cell populations - Not recorded.

Perc

enta

ge

Surv

ival

Figure 1.4 Five patterns of cell survival response following treatment of a human lymphoma cell line (T ) with anti-tumour drugs for 1 hour in vitro (after Drewinko et al, 1979).

Drug Concentration

Proliferation — dependent toxicity of antitumour agents is now well recognised and has been demonstrated in a wide range of experimental systems (Barranco et al, 1975; Mauro et al, 1974a; Hahn et al, 1974; Twentyman and Bleehan, 1975; Valeriote and van Put ten, 1975; Drewinko et al, 1981). Generally, proliferating cells are more sensitive to antitumour drugs than non—proliferating cells. However, some drugs, such as mitomycin C (Lahiri, 1973) and mechlorethamine (Ray et al, 1973), reportedly kill cells in both proliferating and non—proliferating states to the same extent. In addition, there are a few examples, bleomycin and BCNU (Barranco et al, 1973; Hill, 1978a), hydroxyurea (Mauro et al, 1974a) cisplatin and vindesine (Drewinko and Barlogie, 1984), where non—proliferating cells are more sensitive than proliferating ones. .Many of the studies measuring the kinetic effects of antitumour agents are performed using plateau—phase cultures (Hahn and Little, 1972; Barranco et_ al, 1973; Twentyman and Bleehan, 1973; Mauro et al, 1974b). The principle reason for this is because the overall patterns of the plateau—phase cell population resembles that of cell—renewal systems in vivo (Hahn and Little, 1972).

iv. Pharmacokinetics

Anticancer agents, unlike many other clinically useful agents, vary enormously in physical size, lipophilicity, reactivity, stability, cellular site(s) of action and target cell sensitivity (Levin, 1986). Each drug class presents its own unique problems, thus, useful generalisations among anticancer agents are restricted. One major problem common to all in vitro test systems lies in selection of the drug concentration to which the in vitro tumour cells are to be exposed.

Predictive studies involve many pharmacological considerations (Alberts et al, 1980), not least the possible alterations in the chemosensitivity of cells in vitro from that

in vivo. The three most important pharmacokinetic parameters ares (i) the plasma concentration.time product (CxT), (ii) the peak plasma drug concentration and (iii) the plasma terminal— phase drug half-life (ti). The methods for their calculation have been previously described (Gibaldi and Perrier, 1982;Levin, 1986). The CxT, which is probably the most important pharmacokinetic parameter, can be calculated by measuring the area under the plasma disappearance curve after drug administration. The peak concentration of a drug in plasma is usually determined by direct measurement, while the ti (terminal) is the time required to halve the concentration of the drug during its terminal elimination phase. It is important to determine the plasma half-life of a drug in order that chemical degradation or conjugation and elimination from the body, specifically by the hepatic and renal routes, can be estimated.On a pharmacokinetic basis in vitro sensitivities for any given drug are only 5 — 10% of the clinical concentration.time product (CxT) achievable. This suggests that intratumoral drug concentrations in vivo may be lower than those measured in the plasma (Salmon et al, 1980a). It has been suggested from pharmacokinetic data (Alberts et al, 1979) that the greatest cellular exposure to most drugs occurs during the first hour after administration. Further investigations (Alberts et al, 1980) showed that the result of a 1 hour incubation of a drug, at a concentration which represents l/10th of the peak plasma level in vivo, correlates best with clinical response. However, for certain schedule—dependent drugs, as well as those with a prolonged plasma half-life and those used according to a repeated daily schedule, prolonged in vitro exposure greater than one hour may be needed (Alberts et al, 1981a).

1.7 APPLICATIONS AND FUTURE PROSPECTS

Clonogenic assays have been utilised in a number of biological studies including, tumour cell heterogeneity with

respect to differentiation, biological control of tumour cell proliferation and identification of new potential cytotoxic drugs (Buick and Poliak, 19&4). The use of primary cultures to predict the response of a patient's individual tumour to chemotherapy is an over—subscribed idea. However, using the clonogenic assay, human tumour continuous cell lines have perhaps had their greatest impact in the identification of new anticancer agents and their potential clinical applicability for specific human neoplasms. They may not provide the ideal model system for evaluating drug sensitivities and it is not suggested that they can totally replace clinical trials. Nevertheless, there are a number of advantages for their use which can provide a basis for future studies provided limitations associated with their use are acknowledged (Mattern and Volm, 1982; Hill, 1983a).

The prospects for the expansion of in vitro drug sensitivity studies depend on the establishment of more human tumour cell lines, particularly from neoplasms which at present are poorly represented. This may require improvements in the present cell culture techniques now available. Nevertheless, there are a number of studies related to antitumour drug evaluations for which continuous cell lines may already be used. These studies include (i) an understanding of the mechanism(s) of drug action,(ii) elucidation of mechanisms of drug resistance, both inherent and experimentally induced, (iii) identifying mechanism(s) of drug metabolism, (iv) evaluating the potential of improved analogues of existing compounds, (v) identifying synergy or antagonism in new drug combinations and (vi) establishing dose and time dependant effects on cell survival. The first three studies are particularly important with respect to the observed heterogeneity of response to drugs amongst tumour cell populations not only within different tumours of the same histological type but also within the same individual tumour.

CHAPTER 2. MATERIALS AND METHODS

COOTENTSPAGE

2.1 MATERIALS 110(A) Tissue culture materials 110(i) Culture vessels 110(ii) Culture medium 110(iii) Serum 111(iv) L-glutamine 111(B) Reagents 113(i) Trypsin 113(ii) Versene (EDTA) 113(iii) Dulbecco's PBS solution A 113(iv) Dimethyl sulphoxide 114(v) Radioactive thymidine ( % —TdR) 114(vi) Antibiotics solution 114(vii) Histological fixative 114(viii) Cytological fixative 115(ix) Chrom alum/gelatin 115(x) Gels for polymorphic enzyme analysis 115(C) Drugs 123(i) Adriamycin (ADR) 123(ii) Bleomycin (BLM) 125(iii) Cis—Diamminedichloroplatinum II (Cisplatin) 125(iv) Cis—Diammine—1,1—Cyclobutane Dicarboxylate 126

Platinum II (CBDCA)(v) Dibromodulcitol (DBD) 127(vi) 5—Fluorouracil (5—FU) 128(vii) 4—Hydroxyanisole (4-OHA) 130(viii) Methotrexate (MTX) 131(ix) Mitomycin C (Mit C) 133(x) Mitoxantrone (MX) 134(xi) Spirogermanium (Sp) 135(xii) Etoposide (VP16-213) 136

PAGE

2.2 METHODS 137(A) Cell culture techniques 137(B) Subculturing 139(C) Cryopreservation of tumour cell lines 139(i) Freezing cells 139(ii) Thawing 140(D) Determination of mycoplasma contamination 140(i) Staining of cell cultures 140(ii) Growth in agar 141(E) Histological processing 143

2.3 EQUIPMENT 143

2.4 EXPERIMENTAL PROTOCOLS 145(A) Colony forming efficiency (CFE) 145(B) Population doubling time (PDT) 146(C) Cell labelling 146(i) Labelling index (LI) 146(ii) Continuous labelling 147(iii) Autoradiography 147(D) Time-lapse photography 148(E) Tumorigenicity in nude mice 149(F) Cytology 149(G) Enzyme polymorphism analysis 150(H) Monoclonal antibodies 151(I) . Clonogenic drug sensitivity assays 153(J) Comparison of growth rate curves of RT112 cells 155

cultured in microtest plate wells and 35 mm dishes (K) Experiment to determine at what stage RT112 cells, 157

at an equivalent cell density used for clonogenic assay, enter plateau—phase growth

(L) Experiment to determine the effect of trypsin/ 159versene (EDTA) concentration and volume on cell viability

2.1 MATERIALS

(A) Tissue culture materials

(i) Culture vessels

Nunclon Delta disposable plastic products (Gibco Bio—Cult Limited, Paisley, Scotland) were used throughout these studies. Tissue culture flasks, either 25 cm2/40 ml (code 1—63371) to which 5 ml medium was added, or 175 cm^/SOO ml (code 1—52051) to which 40 ml medium was added were used for the maintenance of stock cells. Colony forming efficiency (CFE) tests, clonogenicity of surviving cells after drug exposure and the drug exposure of cells in situ were all performed using 50 mm tissue culture petri dishes (code 1—50288) to which 5 ml of medium was added. Growth rate determinations (PDT) and labelling indices (LI) of the cell lines were both performed in 35 mm tissue culture petri dishes (code 1—53066) containing 5 ml of medium. Continuous labelling experiments were performed using 50 mm tissue culture petri dishes containing 5 ml medium. Tissue culture microtest plates (96 flat bottomed wells, code 1—67008) were used for the drug treatment of cells, 0.2 ml medium per well.

(ii) Culture medium

Cell lines were cultured in Roswell Park Memorial Institute (RPMI) 1640 medium. This was obtained from Gibco Bio—Cult Limited (code 074—1800) and prepared by the Central Research Services of the Imperial Cancer Research Fund, London. The medium in powder form was added to fresh double distilled water and NaHC03 was added at 2g per litre. The pH was adjusted using IN HC1 or IN NaOH to within 0.2— 03 pH units below the final desired pH of 7.0 — 7.2. (The pH tends to rise during filtration). The medium was then sterilised by filtration

through a 0.22 p Millipore membrane, dispensed into glass bottles in 500 ml and 100 ml aliquots and stored at +-4°C for a period not usually exceeding one month before use. Duplicate samples of the medium (1 — 2 ml) were added to Brain heart infusion, Thioglycollate medium and Sabouraud broth for sterility checks. (The Brain heart infusion for aerobic contaminants,Thioglycollate medium for anaerobic contaminants and the Sabouraud broth for the isolation of fungal contamination). Half the samples, for each test medium, were incubated at 37°C, the remaining samples were incubated at room temperature. They were checked daily for contamination over a period of 21 days. The constituents of this medium, which show a few modifications from the original (Moore et al, 1967), are listed in Table 2.1.

(iii) Serum

Maintenance of the cell lines and all experimental studies were performed using medium containing 5% of a single batch (Batch No 29121125) of foetal bovine serum (FBS), obtained as frozen 500 ml aliquots from Flow Laboratories Limited, Irvine, Scotland. Before use the serum was thawed, heat inactivated at 56°C for 30 minutes (in order to destroy the heat labile complement components, thus preventing any antigen—antibody reaction which may cause the lysis of cells), dispensed into 1,5, 10 and 20 ml aliquots and stored at —20°C.

(iv) L-glutamine

200 mM L-glutamine was obtained from Flow Laboratories Limited. It was dispensed into 1 ml and 5 ml aliquots, stored at —20°C and thawed at room temperature as required. It was added to RPMI 1640 medium to give a final concentration of 2 mM.

Table 2.1 RPMI 1640 medium (Moore et al, 1967)

Constituent mg/litre Consituent mg/litri

L—Arginine 200.00 Biotin 0.20L—Asparagine H2O 56.82 D-Calcium pentothenate 0.25L—Aspartic acid 20.00 Choline chloride 3.00L-Cystine disodium salt 59.15 Folic acid 1.00L—Glutamic acid 20.00 i—Inositol 35.00Glutathione 1.00 Nicotinamide 1.00Glycine 10.00 p—Aminobenzoic acid 1.00L—Histidine 15.00 Pyridoxine HC1 1.00L—Hydroxyproline 20.00 Riboflavin 0.20L—Isoleucine 50.00 Thiamin HC1 1.00L—Leucine 50.00 Vitamin B12 0.005L-Lysine HC1 40.00 Ca (N03)2 69.49L-Methionine 15.00 KC1 400.00L—Phenylalanine 15.00 MgS04 7H20 100.00L—Proline 20.00 NaCl 6000.00L—Serine 30.00 NaHC03 2000.00L—Threonine 20.00 Na2HP04 800.70L—Tryptophan 5.00 Glucose 2000.00L—Tyrosine 20.00 Sodium phenol red 5.00L—Valine 20.00

(B) Reagents

(i) Trypsin

Trypsin (1:250), was obtained from Difco Laboratories Limited, London, England. A stock 0.25% solution in Tris saline was prepared by the Central Research Services of the Imperial Cancer Research Fund (constituents are listed in Table 2.2).After the pH was checked (pH 7.7) the solution was prefiltered through 1.2, 0.8 and 0.4 ^ Millipore membranes, with final sterilisation by filtration through a 0.22 |i Millipore membrane, and stored in 20 ml aliquots at —20°C. A duplicate sample of the trypsin (1—2 ml) was added to Brain heart infusion,Thioglycollate medium and Sabouraud broth for sterility checks (for details see section 2.1(A)ii). The tryptic efficiency was checked by testing the digestive activity of a two fold dilution series from 1/64 — 1/1024 on charcoal gelatin discs. Each dilution and charcoal gelatin disk was incubated for 1 hour at 37°C. The highest dilution showing charcoal release was taken as the tryptic efficiency, which for 0.25% trypsin was usually 1/512.

(ii) Versene (EDTA)

Versene was obtained from BDH Chemicals Limited, Poole, England, and prepared as a 0.02% solution in phosphate buffered saline (PBS) solution A, by the Central Research Services of the Imperial Cancer Research Fund (Constituents are listed in Table 2.3). The pH was checked (7.2), dispensed into 16 ml aliquots, sterilised by autoclaving at 121°C (15 lbs p.s.i.) for 15 minutes and stored at +4°C.

(iii) Dulbecco's PBS Solution A

PBSA (Dulbecco and Vogt, 1954) was prepared by the Central Research Services of the Imperial Cancer Research Fund. The

salts (see Table 2.3) were dissolved in double distilled water, the pH checked (7.2), dispensed into 20 ml and 400 ml aliquots, sterilised by autoclaving at 121°C (15 lbs p.s.i.) for 15 minutes and stored at +4°C.

(iv) Dimethyl sulphoxide

Dimethyl sulphoxide (DMSO) was obtained from BDH Chemicals Limited. It was dispensed into 1 and 5 ml aliquots, sterilised by autoclaving at 121°C (15 lbs p.s.i.) for 15 minutes and stored at room temperature.

(v) Radioactive thymidine (%—'TdR)

Methyl—% thymidine (code TRK120) of high specific activity (25 Ci per mMol) in aqueous solution was obtained from Amersham International PLC, Amersham, England. A stock solution was prepared by adding 9 ml RPMI 1640 medium containing antibiotic snlution to 1 ml 'TdR at 1 mCi/ml, giving a final concentration of 100 pCi/ml which was stored at i-4°C for a maximum of two months.

(/i) Antibiotic Solution

Penicillin and Streptomycin were obtained as crystamycin from Glaxo Laboratories Limited, Greenford, England. Each vial cnntained 300 mg Benzyl penicillin (sodium) BP and 500 mg Streptomycin (as streptomycin sulphate BP) and was stored at room temperature until required. For use the contents of each vial were dissolved in 50 ml sterile injection water and dispensed into 1 and 5 ml aliquots and stored at —20°C.

(/ii) Histological fixative

Baker's Formol—calcium solution (Baker, 1944), see Table 2 4, was used for the fixation of xenografts excised from nude nice for histological examination.

(viii) Cytological fixative

Duplicate smears of exponentially—growing cells were prepared on glass slides, one of which was air dried before staining, the other was fixed in 95% ethyl alcohol (Absolute alcohol 100) James Burrough, London, England.

(ix) Chrom alum/gelatin

A chrom alum/gelatin solution was prepared (see Table 2.5) in distilled water and stored at +>4°C. It was used to coat fixed radio—labelled cells for a more even coating of the photographic emulsion (Brooks, 1975).

(x) Gels for polymorphic enzyme analysis

The constituents for the buffers used in the gels are listed in Tables 2.6 and 2.7. All reagents were analar grade, BDH Chemical Limited, Poole, England.

Table 2.2 Trypsin 0.25% in Tris Saline

Constituents to make 1000.0 ml

NaCl 8.0 g

KC1 (19%) 2.0 ml

Na2HP04 0.1 g

Dextrose 1.0 g

Trizma Base(Tris hydroxy methyl) amino methane) 3.0 g

Trypsin (Difco 1:250) 2.5 g

Phenol Red (1%) 1.5 ml

Table 2.3 Versene (EDTA) 0.02% in PBS (solution A)

Constituents to make 1000.0 ml

NaCl 8.0 g

KC1 0.2 g

Na2HP04 1.15 g

KH2HP04 0.2 g

Diamino ethane tetra—acetic aciddisodium salt (versene) 0.2 g

Phenol Red (1%) 1.5 ml

P>PBSA

Table 2.4 Baker's formol-calcium

CaCl2 anhydrous 1 g40% formaldehyde 10 mlDistilled water 90 ml

(Baker, 1944)

Table 2.5 Chrom alum/gelatin

Solution A

Solution B

5 g Gelatin (Griffin and George Limited, London, England)

400 ml Distilled water — heat

0.5 g Chromium Potassium Sulphate (BDH Chemicals Limited, Poole, England)

400 ml Distilled water

Mix Solution A and B — make up to 1000 ml with distilled water. Add few drops formalin to maintain sterility, store +4°C.

Table 2.6 Bridge buffers used for starch gel electrophoresis

i TEMM buffer pH 7.4

Tris 12.1 gMaloic anhydrous 9.8 gEDTA Na2 salt 3.7 gMgCl2 2.0 gDistilled water 1000 ml

ii Citrate phosphate buffer pH 6.3 — 6.4

Trisodium citrate 32.1 gNaH2P0 anhydrous 29.4 gDistilled water 1000 ml

iii Phosphate buffer pH 7.0

NaH2P04 anhydrous 24.0 gNa2HP04 28.39 gDistilled water 1000 ml

Hydrolysed starch for the gels is used at a concentration of 11% in one of the above bridge buffers.

Table 2.7 Acrylamide gel for Isoelectric focusing

Glycine 0.28 gAcrylamide 4.5 ml of a 29.1% solutionBis—acrylamide 4.5 ml of a 0.9% solutionDistilled water 17 ml

Ampholines 0.9 ml to give pH range 3.5 — 10.0(Chemically synthesised amino acids)

40 mg % riboflavin (to polymerise gel)

0.1 ml

Table 2.8 Chemotherapeutic drugs

Drui Abbreviation Source

Adriamycin

Bleomycin

Cis-Dichloro- diammine platinum (II)

Cis-Diammine-1,1* cyclobutane dicarboxylate platinum (II)

Dibromodulcitol

5-Fluorouracil

4-Hydroxyanisole

ADR

BLM

Cisplatin

CBDCA

DBD

5-FU

4-OHA

Farmitalia Carlo Erba, Barnet, England

Lundbeck Ltd, Luton, England

Mead Johnson Laboratories, Bristol Myers Co, Ltd, Langley, England

Gift from Dr A H Calvert Institute of Cancer Research, Sutton, England

Gift from Chinoln Pharmaceutical Works, Budapest, Hungary

Roche Products Ltd, Welwyn Garden City, England

Koch-Light Ltd, Colnbrook, England

Methotrexate

Mitomycin C

Mitoxantrone

Spirogermanium

MTX

Mit C

MX

SP

Lederle Laboratories, Cyanamid, Gosport,England

Kyowa Hakko Kogyo Co Ltd, Tokyo, Japan

Gift from Lederle Laboratories, Cyanamid, Gosport, England

Gift from Unimed Inc, Somerville, New Jersey, U.S.A.

4-demethyl-epipodophyllotoxin-3-D-ethylideneglucoside

Mead Johnson Laboratories, VP16-213 Bristol Myers Co Ltd,

Langley, England

Solvent

PBSA

PBSA

PBSA

PBSA

DMSO

PBSA

Medium

PBSA

PBSA

PBSA

PBSA

PBSA

(C) Drugs

Stock solutions of each drug were prepared just prior to use in PBSA and diluted in medium to yield a final concentration of PBSA not exceeding 0.1%. The only exceptions were 4— hydroxyanisole (4—OHA), which was dissolved directly in medium, and dibromodulcitol (DBD), which was solubilised in dimethyl sulphoxide and then diluted at least 1000—fold in medium before use. The drugs used and their sources are listed in Table 2.6.A brief description of each drug and their mechanism of action are described below:

(i) Adriamycin (ADR)

‘OH

OCH OH

OH

Mol. wt. 579.98

Adriamycin (ADR) is an anthracycline antibiotic isolated from Streptomyces peucetius v caesius. It is a four—ring structure linked to an amino sugar by a glycosidic bond and has activity against a broad spectrum of human tumours (Blum and Carter, 1974). Its mechanism of action has been extensively reviewed (DiMarco, 1975; Crooke and Reich, 1980). Its antitumour activity is thought to be exerted by binding tightly to DNA and intercalating between the base pairs, leading to distortion of the helix and reduction in template activity. More

recently single strand breaks in DNA have been found after treatment of mammalian cells with adriamycin (Lee and Byfield,1976), but their role in drug induced cytotoxicity is unclear (Ross et al, 1979). Adriamycin has also been shown to bind to cell membranes (Mikkelsen et al, 1977), causing altered cell membrane integrity and function at concentrations below those necessary to impair DNA function (Dorr and Alberts, 1982). The cytotoxic effect of adriamycin can be exerted without entering cells, shown in experiments with the drug linked to agarose beads (Tritton and Lee, 1982).

(ii) Bleomycin (BLM)

CH

N HN HCH

CH

HO-O H Mol. wt. approximately

1400O H

O H

O H

N H

R+ ru

B l e o m y c i n A 2 N H C H 2C H 1C H 2 S ^ 3^CH3B l e o m y c i n B2 N H C H 2CH->CH2C H2N H C ^ NH^NH2

The bleomycins, produced by a strain of Streptomyces verticillus, are water-soluble glycopeptide antibiotics, consisting of a group of basic sulphur-containing polypeptides. "Bleomycin" refers to the commercial preparation, which is a mixture of bleomycin A£ and B£ which differ only in their terminal amine moieties, the former being the predominant component in clinical preparations. It is a unique anticancer agent in that it shows an affinity for squamous cell tissue and shows very little bone marrow toxicity. The biological effects and mechanism of action of bleomycin have been reviewed (Umezawa,1974). Briefly, it markedly inhibits DNA synthesis, but RNA and protein synthesis are much less affected (Muller et al, 1975).It also causes DNA strand breaks (Terasima et al, 1970). Both single—stranded and native DNA scission in vitro can be potentiated by a number of substances, including sulphydryl compounds and magnesium ions, however, because bleomycin is a potent copper chelator its action is inhibited by copper and zinc ions (Pratt and Ruddon, 1979b). Studying the lethal effects of bleomycin on mammalian S3 cells in culture, Terasima and Umezama (1970), observed that the dose—response curve exhibited an upward concavity, suggesting the involvement of sensitive and less sensitive fractions in a randomly growing population.

(iii) Cis—Diamminedichloroplatinum (Cisplatin)

The platinum co-ordination complexes were first suggested as antitumour agents by Rosenberg et al (1965, 1969) with the parent compound, cis-dichlorodiammine platinum (II), showing activity

Mol. wt. 300

against several types of cancer, both alone and in combination with other drugs and radiation (Muggia and Rozencweig, 1978). These platinum compounds are of great interest because they represent an entirely new class of anticancer drugs. However, enthusiasm for the clinical application of cisplatin is markedly tempered by the associated nephrotoxicity, the dose—limiting factor in the use of this agent in most drug schedules, which can be life-threatening and is irreversible (Hayes et al, 1977).This drug produces a selective and persistent inhibition of DNA synthesis in a variety of cell types, with several possible modes of interaction having been suggested (Douple and Richmond, 1979). The cis—isomer has been shown to bind to DNA and can form interstrand cross-links (Roberts and Pascoe, 1972). Published data from several studies show the drug preferentially binds to guanine, although it also binds to adenine and cytosine (Munchausen and Rahn, 1975; Stone et al, 1976; Kelman et al,1977).

(iv) Cis—Diammine—1,1—Cyclobutane Dicarboxylate Platinum II (CBDCA)

00C CH2

ch2Mol. wt. 370

CH2

CBDCA, also known as JM8, is one of a group of new second- generation platinum compounds which has been shown to retain its antitumour activity but with apparently less host toxicity compared to the parent compound cisplatin (see references cited by Calvert et al, 1982; Curt et al, 1983). The only aspect of dose—limiting host toxicity associated with CBDCA was

myelosuppression, in particular thrombocytopenia. The pharmacokinetics of CBDCA differs from that of cisplatin, furthermore CBDCA does not react as extensively with plasma proteins (Curt et al, 1983).

(v) Dibromodulcitol (DBD)

OH OHH I I HC ------C ------ C C ------CH2I H H I IOH OH Br

Mol. wt. 308

The dibromo derivatives of a,w—substituted hexitols are cytotoxic to a number of animal tumours (Sellei et al, 1969;Belej et al, 1972). The diastereoisomer of dibromommanitol, known as dibromodulcitol, has been shown in addition to inhibit a number of transplantable tumours. Its chemical name is 1,6 dibromo 1,6 dideoxy dulcitol and is said to act as a bifunctional alkylating agent, able to produce either cross- linking of DNA bases on the same strand or interstrand cross­links (Sellei et al, 1969; Falkson et al, 1982). At slightly alkaline pH it is converted, in the plasma, into epoxides which are more reactive alkylating agents than the parent compound (Horvath et al, 1982), with excretion occurring exclusively via the kidneys. However, the action of dibromodulcitol cannot be entirely explained by this mechanism because, compared to typical alkylating agents, differences have been shown in its interaction with DNA, RNA and protein (Belej et al, 1972; Falkson et al, 1982).

H2CIBr

(vi) 5—Fluorouracil (5—FU)

0

HN

0NH

The development of fluorouracil began with the observation by Rutman et al (1954) who found that uracil was preferentially utilized and incorporated into nucleic acids of an experimental tumour. This finding stimulated Heidelberger and colleages (1958) to test potential uracil antagonists as cancer chemotherapeutic agents. 5—fluorouracil (5—FU) is a pyrimidine antagonist which belongs to the group of tumour inhibitory compounds known as the antimetabolites. The structural formula resembles that of uracil, the hydrogen atom in position 5 of the latter compound being replaced by a fluorine atom. This atom is very small and the carbon—fluorine bond very stable. This not only permits the antimetabolite to occupy active sites on enzymes, often in preference to normal substrates, but also prevents the attachment of groups to the substituted carbon atom of the metabolite (Heidelberger et al, 1958).

As shown in Figure 2.1, 5—FU is activated in vivo by one or more alternative routes (Piper and Fox, 1982). It may be converted to 5—fluorouridine 5'-monophosphate (FUMP) either directly by orotate phosphoribosyltransferase (Pathway I) or by the sequential action of uridine phosphorylase and uridine kinase (Pathway II). FUMP may then be incorporated into RNA or it may be converted to 5—fluorodeoxy—uridine 5'-monophosphate (FdUMP), an inhibitor of thymidylate synthetase (Heidelberger, 1975) and

Figure 2.1 Pathways for metabolism of 5-FU (After Piper and Fox, 1982).

Pathway I

Pathway II _ __ ^

PRPP

5-FU Fl'DPFUMP ■►RN'A■FUTPorotate phospho- rlbosyl transferase

Rlbonuclotidereductase

Uridinekinase

RIP ATP

FdUDPUridinephosphorylase

dRIP ATP

Pathway III 5-PUThymidine

phosphorylase

FdUrd FdUMP

Abbreviations used:

5—PU, 5-fluorouracil;PUMP, 5-fluorourldine 5’-monophosphate;PUTP, 5-fluorourldlne 5'-triphosphate;PUDP, 5—fluorourldlne 5'-diphosphate;PdUMP, 5—fluorodeoxyuridine 5’-monophosphate; PUrd, 5—f luorourldlne;PdUrd, 5-fluorodeoxyuridine;PRPP, phosphorlbosyl pyrophosphate;RIP, rlbose 1-phosphate; dRIP, deoxyrlbose l-phosphate;

Thymidinekinase

dUMP- dTMP

Thymidylatesynthetase

hence DNA synthesis. The third route (Pathway III) of 5—FU activation, catalysed by thymimidine phosphorylase and thymimidine kinase, results in synthesis of FdUMP only. Hence, only DNA synthesis can be inhibited.

5—fluorouracil has been used clinically for the treatment of a wide range of solid tumours (Carter and Soper, 1974) and has shown some success in the treatment of advanced bladder cancer either alone or in combination with other drugs, notably adriamycin (Cross et al, 1976; Smalley et al, 1981; Fossa and Gudmundsen, 1981).

(vii) 4—Hydroxyanisole (4—OHA)

HO H

In animal studies a range of substituted phenols, including several mono and dihydric phenols with non—polar side chains, have been shown to be effective depigmenting agents when applied locally to the skin (see references cited by Morgan et al, 1981). The depigmenting action of these agents is related to their structural similarity to the amino acid tyrosine. One of these agents is the tyrosine analogue, 4—hydroxyanisole (4—OHA), which is oxidised by tyrosinase (Riley, 1969b) and gives rise to cytotoxic oxidation products (Riley, 1969a, 1970; Riley et al,1975). Animal studies in mice bearing transplantable Harding— Passay melanomas showed that regression of tumours occurred in animals receiving 4-OHA in doses well below those producing generalised toxicity (Dewey et al, 1977). However, in a study where 4—OHA was used to treat patients with proven primary or secondary malignant melanomas, only when it was administered by the intra—arterial route did it prove to have a localised and

therapeutically useful cytotoxic action (Morgan et al, 1981). It would appear that the metabolism of 4-OHA and its clearance from the blood is very rapid, and thus intravenous administration is without effect.

(viii) Methotrexate (MTX)

NH 2

0IIC~NH — C H — COOH Icm

c h 2- c o o h

N CH^— NCH

Mol . wt. 1+5^

Methotrexate (MTX) was developed in 1949 and quickly superseded its parent compound, aminopterin, in the treatment of acute leukaemia, due to increased therapeutic index in L1210 mouse leukaemia (Law et al1, 1949). MTX is a folic acid antagonist which binds tightly but reversibly to dihydrofolate reductase, thereby inhibiting this enzyme from converting folic acid to tetrahydrofolate (Bleyer, 1977), see Figure 2.2. This enzyme block results in the depletion of reduced folates which are necessary for the metabolic transfer of 1—carbon units in a variety of biochemical reactions. Those reactions which are of special importance in cellular proliferation are the biosynthesis i.of thymidylic acid, the nucleotide specific to DNA and the biosynthesis of inosinic acid, the precursor of purines necessary for both DNA and RNA synthesis. In man, MTX appears to inhibit DNA synthesis to a greater extent than RNA synthesis, suggesting that inhibition of thymidylate synthesis is the most important mechanism of MTX cytotoxicity (Hoffbrand and Tripp, 1972). Hence the drug is highly cell—cycle dependent, acting primarily during DNA synthesis (S-phase).

Figure 2.2 Mechanism of action of methotrexate (After Bleyer, 1977)

DNA de novopyrimidinesynthesis

Thymidine monophosphate (dTMP) vV

Thymidylate synthetase

Deoxyuridylic acid (dUMP)

Dihydrofolic acid (FH2)

5,10-methylene-tetrahydrofolate

NADPH

MTX

NAO

Dihydrofolate reductase

Tetrahydro-folace(FH4)

serine

’methionine

homocysteine

5-eethyl-cetrahydrofolateA

5,10-methenyl-tetrahydrofolate

a

5-formyl— tetrahydrofolate (Citrovorum factor)

Reduced folates are also required as co—factors in the conversion of glycine to serine and homocysteine to methionine.By preventing these amino acid interconversions, MTX may also interfere with protein synthesis. It has been demonstrated that intracellular MTX levels in excess of those required to saturate the tight binding sites are necessary for maximal suppression of DNA synthesis. Even higher free drug levels are necessary for inhibition of RNA synthesis and higher levels yet for the cessation of protein synthesis (see references cited by Bleyer,1978). These observations confirm the need to maintain the extracellular MTX concentration above a critical level for maximum effect.

For high dose MTX regimes, where normal stem cells In the bone marrow are affected, the duration of exposure to the drug effect has to be limited. This is achieved by the addition of leucovorin (Citrovorum factor) rescue. Citrovorum factor (5— formyl—tetrahydrofolate) is converted to 5—methyl— tetrahydrofolate which is the predominant reduced folate in the body. This metabolite is converted to tetrahydrofolate or to5,10-methylene—tetrahydrolate, which enters the reduced folate cycle, bypasses the MTX—induced enzymatic block, and repletes the cell of reduced folates (Bleyer, 1978). It has been shown that this resumption of DNA synthesis in normal bone marrow can occur without concomitant recovery of DNA synthesis in neoplastic cells (Frei et al, 1975).

(ix) Mitomycin C (Mit C)

0

Mol. wt. 334

Mitomycin C (Mit C) is an antibiotic isolated from Streptomyces caespitosus (Wakaki et al, 1958). It is an alkylating agent whose anticancer activity was first reported by a number of workers in Japan (Frank and Osterberg, 1960). Mit C is known to inhibit selectively the synthesis of DNA in bacteria and mammalian cells in culture (see references cited by Schwartz et al, 1963) and has been shown to be successful in the treatment of a variety of human cancers (see references cited by Crooke and Bradner, 1976; vanHazel and Kovach, 1982). Although the drug is usually given intravenously, recent reports have demonstrated clinical effectiveness against urothelial carcinoma after topical instillation (Mishina et al, 1975; Bracken and Johnson, 1979).

As shown with the mustard-type alkylating agents, Mit C preferentially binds to DNA guanine, forming interstrand cross­links. These cross-links increase with increasing DNA guanine content (Iyer and Szybalski, 1964). The production of interstrand cross-links was first demonstrated using DNA melting experiments, in which the denatured DNA was shown to renature spontaneously on rapid chilling (Iyer and Szybelski, 1963).Rapid and complete renaturation is characteristic of DNA strands that have been kept in association by cross-linkage throughout denaturation.

(x) Mitoxantrone (MX)

NCH-CHLNCH.CH-OH

2 HC

HO NCH-CH.NCH-CH^OH

Mol. wt. 518

Mitoxantrone (MX) is an anthracenedione showing structural similarities to adriamycin. It contains an electron—rich planar ring structure which is an essential feature for intercalation (Double and Brown, 1975). Adamson (1974), postulated that the amino—sugar moiety on adriamycin was responsible for its cardiotoxicity but not its antitumour activity. Although it is often stated that mitoxantrone was developed on the basis of a rational search for adriamycin—like compounds without this sugar moiety (see references cited by Smith, 1983), its discovery really owed something to serendipity as well as to rational drug development. MX was among a large series of compounds synthesised in the American Cyanamid Laboratories by Murdock et_ al (1979). In several cardiotoxicity models MX showed either none or less toxicity than adriamycin (Henderson et al, 1982; Sparano et al, 1982).

In vivo antitumour activity does not correlate well with in vitro measurements of DNA binding capability (Johnson et al,1979), thus, an effect other than, or in addition to, DNA intercalation may account for the antitumour activity of MX.There are now a number of studies which have established that MXinhibits both RNA and DNA synthesis as well as intercalating withDNA, although its precise mechanism of action has not beendetermined (Durr et al, 1983).

(xi) Spirogermanium (Sp)

Mol. wt. 341

Spirogermanium (Sp), represents a new class of antitumour agent. It Is a synthetic azaspirane compound with the metal germanium substituted for a one—carbon moiety in the heterocyclic ring structure (Rice et al, 1974). It inhibits protein synthesis in vitro, while the synthesis of DNA and RNA is initially unchanged, however, its precise mode of action remains unclear. The cytotoxic activity of Sp against the in vivo ascitic Walker 256 carcinoma in rats (Rice et al, 1974), provided the basis for its selection for clinical trials. Its activity against a variety of human tumours has been assessed in phase I/II clinical trials with responses reported in ovarian (Schein et al, 1980) and breast cancers (Falkson and Falkson, 1983), and in malignant lymphoma (Trope et al, 1981). Laboratory studies have shown cytotoxic activity against certain tumour cell lines both in vitro and in vivo (Rice et al, 1974; Woolley et al, 1979; Mulinos and Amin, 1980; Hill et al, 1982, 1983). From the in vitro studies using human cell lines resistance to the cytotoxic effects of Sp are most noticeable with lines derived from urological malignancies, such as bladder and prostate (Hill et_ al, 1983).

(xii) Etoposide (VP16—213)

H

0

Mol. wt. 588

The root of Podophyllum peltatum, a perennial herbaceous plant from America and Canada, commonly known as May Apple or American Mandrake, was used by American Indians as a remedy for numerous ailments and its cytotoxic activity was first reported by Bentley (1861). Interest in the drug was revived by Kaplan (1942) who successfully used an oily solution of podophyllin as a topical remedy in cases of Condyloma acuminata and other benign skin diseases.

Podophyllotoxin, as such, is excessively toxic in clinical applications, however, Stahelin (1973), working in the Sandoz Laboratories, synthesised a new glucoside derivative known as etoposide (VP16—213). The mechanism of action of VP16—213 and its parent compound podophyllotoxin are very different. Although they both inhibit the uptake of nucleosides into HeLa cells,VP16—213 does not produce metaphase arrest by inhibiting microtuble assembly, but does induce single—strand breaks in DNA (Loike and Horwitz, 1976). Studies on the metabolism, mechanism of action and clinical activities of VP16—213 have been extensively reviewed (Arnold and Whitehouse, 1981; Achterrath e_t al, 1982; Estape and Milla, 1982; vanMaanen et al, 1982).

2.1 METHODS

(A) Cell culture techniques

All the cell lines were grown as monolayer cultures in 25 cm^ flasks (Nunc) and used over a maximum of ten passages (see Table 2.9). The cells were maintained in RPMI 1640 medium (Gibco) supplemented with 5% heat—inactivated foetal bovine serum (Batch No. 29121125) (Flow Laboratories Limited) and 2 mM L— glutamine (Gibco) at 36.5°C in a humidified atmosphere of 5% CO2 in air.

Table

2.9

Origin of

cell lines

a. >.a 4 4 • ub 9) C JS 3 b

iw eO 41X b41 4 co a

34) 4)

3 j3U (0

4Ob00 4) O -3 b 4 O t44

4)004 V b a 4 4Vb 4 4 ba 3e <m b o

3 O ■3 b >J 4 tw Q.O Obe jo

4)s0 4 4z z z

I ba >. o3 CMba. >% 4 b e g 41 *i h3 > 4 34 b U b U U (CH4 b 4) 4) O b 3 4C C O 3 b 3O 0 CM OZ Z M o

>» 4 —l .* 4) 3 >bao b 4

O ts i/i in cmp*. r*n r«» r«* r *O' O' O' on 4 O'

on 3 cC h n m © © Z © O ©

00 CM 0< N H nZ H 2 H H H

4>*3 •3 4bJO b b b b bb 2 2 2 2 2 ba a. a a. a.41 b b b b bb 4) 4) 4) 4) 4)b *3 -3 *3 -3 -33 3 3 3 3 33 4 4 4 4 44) b b b b b4 4 4 4 4 4

3oo44 O 4 e 4a.e« o e <hb bb 4 S3 QCb3 © V 3

CM COO H CO PI 4 COf T T T T TH cn o nc nc in © cm in « 3

3 3 3 3c c e e o o o o z z z z4) O e o o o

C E h k i u r r c z

M<nrii»i/iininio r * r * » o * r ^ r ,* r » . o » r » . ^ o

u m <n cm «»06 4 n pi ©©UOZZ©«

a ■CM CM |H H ! '*06 06 •* CM H Z Z H H 06 on z H

b J= b41 a. 33 a 3 a a b3 >•, 3 o o a 34 lb 4 b b b o yb H <M b <b 3 b eJ3 b 4 J0 >s >* >. >» >s >» >1 >> >> b 3 y «M 44 b b b b b b b b b b b b J= o b J= e z ya 3 4 e 4 4 4 4 4 4 4 4 4 3 3 b e 3 b 4 b bb a 4 b ■ ■ a a a ■ a a a U 3 b 4 3 b y 3 b b

o b b b b b b b b b b 4 3 » y 3 3 b 3 bV b 4 3 b b b b b b b b b 4 a 4 4 3 3 33 b 4 3 a. & a. a a a a a a 3 0 pH b b b b b b b ye b b e O e J3 e 3 JO S 4 3 S4> 41 4) 3 b b b b b b b b b a b 3 3 3 b 3 bb a. ■ b 3 3 3 3 3 3 3 3 3 4 y b b 3 b b 4 b b 4b o b 3 3 3 3 3 3 3 3 3 3 b 4 b 4 4 b O 4 b e3 b 4) 3 3 3 3 3 3 3 3 3 3 3* 4 a 4 b a 4 b a 4 33 b ■? 3 4 4 4 4 4 4 4 4 4 4 O b O.Z b a a. b a b4) 3 0 3 b b b b b b b b O O o06 Ctf s 06a ao00aooaCO00 a CO Z Z z

in «o -o ooint

on m oo o •» r-» <N oCM CN Ifl Mi i i i T T cni CNli i i I sf O O' H <N 00 1 i00b b *» on co r* <N H b

»-H <N??•ossCM © © H Z Z

b b CM1 1 3HH O'. © a. © CO CM 06 O'© b O O' On CM s s s u CM © O'in © © 3 b 3 b CO u u O m © On o-» O' CM CM «* 3 b b b © T T b © •* 4 > O ©9 s CO 00 H m .. H{-Huz z CO o U U U CO sSs 3 3 0£ CM Ss06H > > CL.u U6 M s s s

(B) Subculturing

The cells were passaged usually 1 — 2 times per week, depending on the growth rate of the cell line. For cell detachment the medium was removed, cells washed with a 1:5 mixture of 0.25% trypsin/0.02% versene (equivalent to a trypsin concentration of 0.05% and versene concentration of 0.016%), and incubated at 36.5°C for 2 — 5 minutes. [For the transfer of cells after exposure to anticancer agents the above enzyme mixture is further diluted in PBSA Is5, giving a final trypsin concentration of 0.01% and versene 0.003%, see section 2.4 (L)]. The detached cells were washed with 10 ml complete growth medium using a 10 ml sterile syringe and a 19G 1$ inch needle. The cell suspension from one flask was then evenly distributed into 2 — 6 flasks, depending on the growth rate of the cell line, to which 5 ml fresh medium was added and then incubated at 36.5°C in a humidified atmosphere of 5% CO2 in air. After 24 hours the medium from each flask was removed and 5 ml fresh pre—warmed medium added and the cells reincubated. This procedure removed dead cells and cell debris as a result of the passaging.Exposure of cells to trypsin and versene results in a degree of cell damage (Rinaldini, 1958; Waymouth, 1974).

(C) Cryo preservation of tumour cell lines

(i) Freezing cells

Monolayer cultures were harvested from tissue culture flasks using the 1:5 trypsin/versene enzyme mixture described previously, 2.2 (B). The cell suspension was added to 10 ml of medium plus serum and centrifuged (250 g, 10 minutes). The supernatant was discarded and the cell pellet resuspended in 8 ml medium containing FCS (10%) plus DMS0 (10%) to minimise any membrane damage done to osmotic change (Paul, 1975). 1 mlaliquots of the cell suspension, containing approximately 1 xlO^— 10? cells, were dispensed immediately into plastic Nunc cryo

tubes (code 3—66656), sealed and placed inside a polystyrene freezing canister (Union Carbide, Darlington, England). This was positioned in the neck of a liquid nitrogen fridge at a pre­determined depth so that cells were frozen (in the gas phase) at a rate of 1°C per minute until —70°C. The cryo tubes were then stored in the liquid phase at —196°C.

(ii) Thawing

The cryo tubes were retrieved from liquid nitrogen and placed in a water bath at 37°C. When thawed the cell suspension was transferred to a tissue culture flask containing 5 ml medium with serum and placed in an incubator at 36.5°C with 5% CO2 in air. (10% DMSO is near to toxic level for cells and therefore material in DMSO—medium in the liquid state was manipulated as rapidly as possible). After 24 hours incubation the medium was changed to remove dead cells and cell debris. The yield of viable cells from frozen stock was enhanced if cultures originally cryopreserved were in exponential growth.

(D) Determination of mycoplasma contamination

All the cell lines were checked for contamination by Pleuropneumonia—like organisms (PPLO) by Central Cell Services of the Imperial Cancer Research Fund. A cell suspension, approximately 1 xlO^ cells, was added to a 50 mm dish (Nunc) containing 5 ml medium plus serum and three 14 mm round glass coverslips. The dishes were incubated in the normal manner until the cells were near to confluence on the coverslips. Cell cultures on the coverslips were then checked for PPLO by a staining technique and the medium used to inoculate agar plates.

(i) Staining of cell cultures

The staining technique was the same as that described by Fogh and Fogh (1964). The coverslips were aseptically placed

into a clear petri dish to which 3.75 ml sodium citrate solution (Tri—sodium citrate (BDH), 6.84 g, distilled H2O 1000 ml) was added and left for 10 minutes. An equal volume of Carnoy's fixative (Glacial acetic acid/absolute ethyl alcohol, 1:4), was then added and left for a further 10 minutes. The coverslips were then placed on a staining rack and left to air dry. When dry, filtered Orcein stain was added and left for 7 minutes. The coverslips were then washed x3 in absolute alcohol, mounted in Euparol, allowed to set and examined by phase contrast microscopy.

Orcein stain:

Natural orcein 2 g was stirred into a small volume of boiling acetic acid, the mixture then cooled to 55°C and the volume made up to 60 ml with acetic acid. Distilled H2O was then added to a final volume of 100 ml. When the stain solution was cooled to room temperature it was filtered twice through Whatman filter paper No. 1.

(ii) Growth in agar

Medium in which the cells were grown was used to inoculate agar plates for the detection of low levels of mycoplasma contamination which may not be observed in stained cells (Dickson et al, 1980). Tissue culture dishes were seeded with BHK (baby hamster kidney) C13 cells grown to form a confluent monolayer in Eagles (E4) medium plus 10% calf serum. The medium was removed and a mixture of equal quantities of double strength E4 plus 10% calf serum and 2.5% agar was added to the dishes. When set the PPLO agar, Table 2.10 (Lemcke, 1965), was poured into the dishes to form a second layer. The medium to be tested was then incubated at 36.5°C for 7 days in an anaerobic jar with 5% CO2. [The BHK C13 cells secrete substances into the agar which enhance the growth of PPLO],

Table 2,10 Constituents to make P.P.L.O. agar

P.P.L.O. Agar (Difco) 60.0 ml

Wellcome No. 3. Horse serum 20.0 ml

Yeast extract (Gibco) 10.0 ml

Tryptose phosphate broth 20.0 ml

1/80 Thallous acetate 1.0 ml

Penicillin (10,000 units) 1.0 ml

(E) Histological processing

Formol calcium fixed xenografts were embedded in paraffin wax for sectioning using a routine histological processing schedule. This was performed automatically using a 2L processor MKII (Shandon Southern Products Limited, Runcorn, England). A nine step procedure was used to remove the water content from the tumour samples, using a series of alcohols, before embedding in paraffin wax. The procedure was as follows:—

70% absolute alcohol 1 hour70% absolute alcohol 1 hour95% absolute alcohol 1.5 hours95% absolute alcohol 1.5 hours100% absolute alcohol 2 hours100% absolute alcohol 2 hours100% absolute alcohol 2 hoursToluene 2 hoursToluene 2 hoursMolten wax 2 hoursMolten wax 2 hours under negative

pressure (—0.6 Bar)

Using a microtome (Leitz), the embedded samples were cut into sections, 5pm thick, and placed onto clean glass slides, dewaxed in xylene and brought to water by conventional procedures before applying haematoxylin and eosin (H and E) stain. Details of histological stains can be found in Carlton's Histological Technique (Drury and Wallington, 1980).

2.3 EQUIPMENT

(i) All tissue culture experiments and cell manipulationswere performed aseptically in a laminar air flow cabinet, Microflow Dent and Hellyer Limited, Andover, England.

(ii)

(iii)

(iv)

(v)

(vi)

(vii)

(viii)

Cell cultures were Incubated in GA2/3 automatic CO2 incubators, Leec, Nottingham, England.

Centrifugation of cell suspensions were performed using an MSE Chilspin bench centrifuge, MSE, Crawley, England.

Stock cells were stored in a liquid nitrogen fridge, Union Carbide, Jencons Scientific Limited, Leighton Buzzard, England.

Fixed cell colonies were counted using a binocular dissecting microscope, Nachet, France.

Routine microscopical examination of cell cultures and phase photography was performed using Leitz Diavert inverted microscope, Luton, England.

Viable cell counts using a haemocytometer (improved Neubauer) were measured with a Carl Zeiss photomicroscope, Germany, using a xlO objective. This microscope was also used for photography of fixed cell/tissue preparations and counting % —1TdR labelled cells.

Cell counts for the determination of population doubling times were performed with a Coulter counter, Model ZB, using a 100 pm orifice, Coulter Electronics Limited, Luton, England, the following settings were used:—

Lower threshold = 5Upper threshold = >1001/Amplitude = 11/Aperture current = 8

(ix) All nude mice (nu/nu), used for xenotransplantationexperiments, were kept in a negative pressure isolator, Vickers Medical, Basingstoke, England. Inoculation of

human bladder tumour cells was performed and the general health and condition of the mice monitored by the staff of the Animal House Unit of the Imperial Cancer Research Fund.

(x) Histological processing of tumour samples for embedding into paraffin wax was performed automatically using a 2L processor MKII, Shandon Southern Products Limited, Runcorn, England. Embedded wax samples were cut into 5 pm sections using a microtome (Leitz, Luton, England).

(xi) Time-lapse photography was performed using an Olympus (IMT) inverted microscope (Japan) fitted with a Bolex type H16J, 16mm cine camera (Switzerland). The camera is controlled by an Olympus time lapse control unit using automatic exposure times of between 0.5 and 1 second, the 16 mm film used was Kodak Infocapture AHU.

2.4 EXPERIMENTAL PROTOCOLS

(A) Colony forming efficiency (CFE)

Exponentially-growing cells were enzymatically detached from tissue culture flasks and 0.2 ml cell suspension was added to 0.2 ml 0.1% trypan blue in PBSA, mixed thoroughly and a drop placed on a haemocytometer (improved Neubauer). Counts were made using a photomicroscope (Carl Zeiss, Germany) using a.xlO objective. Only viable cells (those excluding trypan blue) were counted and the number per ml calculated. Serial 50% dilutions of the cell suspension were prepared in medium plus serum yielding from 60 to 6400 viable cells per ml. 1 ml aliquots of each cell dilution were plated in triplicate into 50 mm dishes (Nunc) each containing 4 ml pre—warmed/gassed medium. Using a 16 G, 4 inch needle and a 10 ml sterile syringe the medium in each dish was pumped once to mix the cells with the medium. The dishes were incubated for between 8 and 21 days depending on the growth rate of the cell line. For those cell lines that require an

incubation period of 14 days or more to form colonies of >50 cells the medium was replenished at intervals of 7 days.Following the required incubation period the medium was poured off, cells washed x2 in tap water, xl in methanol AR (BDH). The cells were fixed in methanol AR for 10 minutes, washed x2 in tap water and stained 1 — 5 minutes with 10% Giemsa (Gurr's improved R66) (BDH). After dishes were air dried, colonies consisting of 50 or more cells were counted using a binocular dissecting microscope (Nachet, France) and the number of colonies counted were expressed as a percentage of the total cells originally plated; CFE = Number of colonies formed/number of viable cells seeded xlOO.

(B) Population doubling time (PDT)

Duplicate cultures of exponentially—growing cells were plated in 35 mm dishes (Nunc) containing 5 ml medium at cell densitites ranging from 24,000 to 768,000 viable cells per dish and incubated in conditions previously described. At 24, 48, 72 and 96 hours enzymatically detached cells from duplicate dishes of each cell density were suspended in a known volume (10 ml) of filtered PBSA and two 0.5 ml sample aliquots of each duplicate were counted using a Coulter counter (Coulter Electronics Limited). The mean results from a minimum of two experiments were plotted on semi—logarithmic/linear graph paper and the PDT calculated for the period of cell growth between 48 and 96 hours after plating using the formula adapted from John Paul (1965), i.e. Tj) = tln2/ln (Nt/No), where Tp = PDT, t * time interval (48 hours), Nt = population count at the end of the time interval (96 hours), No = population count at the start of the time interval (48 hours).

(C) Cell labelling

(i) Labelling index (LI)

Exponentially—growing cells were plated in duplicate in 35 mm dishes (Nunc) containing 5 ml medium. The cell density used

was equivalent (per mm^ surface area) to that used to seed microtest plate wells in drug sensitivity assays. Following a 48 hour incubation period the medium was discarded and 2 ml fresh medium containing 20 pCi per ml methyl—^H thymidine (Amersham International PLC, 25 Ci mmol) was added to each dish. After a 15 minute pulse exposure the dishes were rinsed three times with serum free medium at 4°C, washed twice with 5% trichloroacetic acid, TCA (BDH), at 4°C for 5 minutes each (this both fixes the cells and extracts the soluble radioactivity) and finally fixed by three 5 minute washes in 95% ethyl alcohol (James Burrough) at 4°C and left inverted to air dry.

(ii) Continuous labelling

Exponentially-growing cells were plated in duplicate into 50 mm dishes (Nunc) containing 5 ml medium at cell densities equivalent to those used for the comparison of clonogenic assays, i.e. 1,000, 61,000 and 780,000 cells per 50 mm dish. Following a 48 hour incubation period the medium was aseptically removed and 5 ml fresh medium containing 1 pCi per ml methyl—% thymidine (Amersham International PLC) was added to each dish. At time intervals of 15 minutes, 2, 4, 8, 24, 48, 72 and 96 hours duplicate dishes were fixed in the same manner as that described above (2.4 C(i)) ready for autoradiography.

(iii) Autoradiography

The fixed cells from both the pulse LI and continuous labelling experiments were coated with a chrom alum/gelatin solution and left inverted to air dry (Brooks, 1975). Coating the cells with chrom alum/gelatin ensures an even coating of the photographic emulsion and in addition prevents the emulsion detaching from the surface of the dish during the developing and fixing process. The autoradiography was performed in a similar manner to that described by Brooks et al (1983). Briefly, using

standard darkroom safelight conditions, a small volume (approximately 2—3 ml) Ilford K5 emulsion (Ilford Nuclear Research Limited, Knutsford, England) diluted Is3 with distilled water and warmed to 43°C was added to each dish. After gentle rotation, to ensure even coating, the dishes were inverted and tilted at an angle to allow excess emulsion to be removed. When almost dry the dishes were placed inverted in a box and double wrapped in aluminium foil to prevent penetration of light exposing the emulsion. After 7 days the dishes were developed by immersion for 2 minutes in Kodak D19 developer (Kodak—Pathe, France) diluted 1:1 with distilled water, rinsed once in distilled water, acidified with a few drops of 1% acetic acid, and fixed by immersion for 4 minutes in Ilford Hypan fixer (Ilford Limited) diluted 1:4 with distilled water. The dishes were then left to rinse in slow running cold tap water for 30 minutes, lightly stained with 10% Giemsa (Gurr's improved R66) (BDH) and left inverted to air dry. Counting of labelled nuclei was performed microscopically viewing through a x25 objective (Carl Zeiss photomicroscope, Germany). A minimum of 500 nuclei were counted (or the maximum number of cells for dishes plated at low density) and fifteen individual silver grains per nucleus was scored as a positive label.

(D) Time-lapse photography

Between 2 — 4 xlO^ viable exponentially—growing cells were plated into a 50 mm plastic petri dish (Nunc) containing 5 ml medium, to yield approximately 10 to 20 cells per mm^. After 48 hours incubation, at 36.5°C in a humidified atmosphere of 5% C0£ in air, the medium was aspirated, the cells washed with medium and 5 ml fresh medium plus normal supplements were added. The dish was placed into a stainless steel chamber with a glass top, and warmed to 36.5°C to remove condensation from lid of petri dish. The chamber was then placed into a perspex incubator fitted to the Olympus IMT microscope to maintain a humidifed 5% CO2 in air atmosphere. (The equipment is similar to that described by Riddle, 1979). The Olympus time-lapse control unit (Japan) was set to activate the 16 mm Bolex type J cine camera

(Switzerland) for automatic exposures of between 0.5 and 1 second. Green and heat filters were fitted to the light source. The number of frames per hour set on the control unit ranged from 9.8 to 19.4. After the cells within the field of view had divided four to six times the film was stopped. Analyses of the films were performed using an analysing projector at the Cinemicroscopy Laboratory of the Imperial Cancer Research Fund.

(E) Tumorigenicity in nude mice

Exponentially growing cells were detached from 25 cm^ flasks (Nunc) using trypsin/versene 1:5 and centrifuged at 250 g for 10 minutes in RPMI 1640 medium plus serum at 4°C. The cells were washed twice in serum and glutamine free medium and finally the cell pellet was resuspended in unsupplemented medium to give a cell concentration of approximately 1 x 10® cells per ml. A 0.1 ml sample of this cell suspension was injected subcutaneously into one flank of a nude mouse (nu/nu), with four mice being used for.each cell line. The mice were examined weekly for gross evidence of tumour formation. When these tumours were approximately 0.5 to 1 cm in diameter the mice were sacrificed by cervical dislocation and the tumour excised and fixed in Baker's formol calcium. Following routine histological processing (for details see this Chapter Section 2.2 (E)) performed by the Histology Department of St Paul's Hospital, paraffin sections of each tumour were stained with haematoxylin and eosin and examined using light microscopy.

(F) Cytology

Cytological examination of the cell lines was performed by microscopical examination of stained smears made from exponentially-growing cells. The medium from a 50 mm dish (Nunc) was discarded and the cell monolayer washed twice with PBSA.Using a glass slide cells were scraped off the dish and with the glass slide held at an angle of approximately 45°, smears of the

detached cells were made on two separate glass slides, one of which was air dried, the other immersed in 95% ethyl alcohol (Absolute alcohol 100, James Burrough). The cell smear on the air dried slide was stained with May—Grunwald (R.A. Lamb, London, England) and Giemsa stain (BDH), whilst the cell smear fixed in alcohol was stained with Papanicalaou stain (Ortho Diagnostic Systems Limited, High Wycombe, England). The staining of the smears and critical examination was performed by the Cytology Department of St Paul's Hospital. Details of the stains and their uses are described by Koss (1979).

(G) Enzyme polymorphism analysis

The enzyme phenotypes of the polymorphic loci expressed by the cell lines were identified using electrophoretic techniques performed at the MRC Human Biochemical Genetics Unit of University College, London. Exponentially-growing cells from four 175 cm flasks (Nunc) were harvested using 0.25% trypsin and 0.02% versene It5, and centrifuged in RPMI 1640 medium plus 5%FBS at 1,000 g for 10 minutes. The cells were then washed twice with PBSA in cryotubes (Nunc). After the final wash the supernatant was aspirated and the tubes inverted for 15 minutes to drain. The cell pellet in the cryo tube was then stored in liquid N2 until anlaysis was performed. Starch gels (Starch- hydrolysed, Connaught Laboratories Limited, Ontario, Canada) were prepared as described by Harris and Hopkinson (1976). The constituents for the various bridge buffers are listed in Table 2.6. Gels made with TEMM buffer diluted 1/10 were used for the enzyme loci PGMI and 3, G0TM, G0TS, ESD, GL0 and PGP, while citrate phosphate buffer, diluted 1/40 was used for gels running the enzyme loci AKI, ADA and ACPI, and phosphate buffer for PGD and G6PD. The gels for GL0 and PGP also contained 0.1 ml per 100 ml gel, mercaptoethanol to act as a reducing agent for these enzymes. The enzyme loci a—FUC requires the technique of isoelectric focussing using an acrylamide gel. See Table 2.7.All gels were covered with a sheet of Melinex (ICI, Manchester, England) when setting to avoid skin formation on the surface.

Preparation of the cell samples consisted of adding 100 —200 pi distilled water to the frozen cell pellet and sonicating the suspension with two 2 second bursts using a Dawe soniprobe type II 30A (Dawe Instruments Limited, London, England). The samples were then centrifuged at 5,000 rpm (MSE bench centrifuge). Pieces of 17 mm thick Whatman chromatography paper, approximately 0.5 cm^ were dipped into the supernatant and placed end on in small slots cut into the gel see Figure 2.3. The gel plates were continually cooled while a current of 110V, 60 —70 mA was applied overnight, except for AKI which was run for 4 to 5 hours. All enzyme loci ran from cathode to anode, except GQTM which ran from anode to cathode. All starch gels were carefully sliced length-wise and various stains (see handbook, Harris and Hopkinson, 1976) applied to the freshly cut surface.

(H) Monoclonal antibodies

The characteristic binding patterns of a series of eleven monoclonal antibodies (the LBS series) to seventeen continuous cell lines derived from transitional cell carcinoma of the human bladder was assessed. The LBS series of antibodies was derived from Balb/c mice immunised with intra—peritoneal injections of one of these cell lines, RT112 (Trejdosiewicz et al, 1985).

Exponentially—growing cells were plated, in duplicate, onto two Teflon—coated "Multitest" slides (Flow Laboratories Limited), to achieve semi-confluence after incubating at 36.5°C in a humidified atmosphere of 5% CO2 in air for 48 hours. After incubation the slides were rinsed twice with normal saline (5 minutes each). The cells on one slide were then fixed for 15 minutes in 10% formalin in PBSC (phosphate buffered saline containing 1 mM CaCl2 and 1 mM MgCl2) and washed in serum free RPMI 1640 medium (never allowing the slides to become dry). The cells on the other duplicate slide were fixed with two changes (5 minutes each) of a 1:1 mixture of methanol and acetone and air dried. Formalin—fixed cells were used to visualise cell membrane

Figure 2.3

Diagrammatic representation of starch gel electrophoresis

Cathode —ve

1 1 1 i 91 1 1 1i ft1 1 1 « 11 1 1 1 1 »1 1 9 ■ ft1 1 1 i ft

Anode■tve

binding, while the solvent—fixed cells, in which the membrane lipids had been solubilised, thus allowing the penetration of the antibodies into the cells, were used to visualise intra cellular binding.

After the cells were fixed, 25 ^1 of each primary antibody, LBS series, containing 0.01% azide to prevent microbial contamination, was added to duplicate wells on each slide and incubated in humidified conditions at 36.5°C for 1 hour. The slides were then rinsed with three 5 minute washes in normal saline and 5 pi of fluorochrome-conjugated goat anti-mouse IgG (H+L) diluted 1/16 in tris buffered saline, added to each well and incubated for a further 30 minutes. The fluorochrome used was TRITC (tetraethyl rhodamine isothyocyanate) also known as Rhodamine (Miles Scientific, Stoke Poges, England). The slides were then washed three times (5 minutes each) in normal saline and mounted in Gurr's aqueous Uvinert (BDH) and left for 24 hours in the dark to reduce scatter glare. Finally, the cells were examined at the Tissue Interaction Laboratory at the Imperial Cancer Research Fund, using a Zeiss photomicroscope equipped with epifluorescent illumination and selective filters. Control cells were set up in the "Multitest" slides containing medium only and treated throughout in the same manner.

(I) Clonogenic drug sensitivity assays

All drug sensitivities in vitro were performed using a clonogenic assay on plastic. Viable exponentially—growing cells (as shown by trypan blue exclusion) were plated in microtest plate wells (Nunc, 96 wells flat—bottom) in 0.2 ml medium at cell densities, depending on the population doubling time, between 3,200 and 12,800 cells per well. The cell lines T24, MGH-U1, MGH-U2, HU961T and 253J were plated at 3,200 per well, HU456,VM—CUB—3 and TCCSUP at 6,400 per well, while RT112, RT4, HT1197 and HT1376 were all plated at 12,800 per well. The outer rows of wells of the microtest plate were not used to avoid edge effects (Belfield and Simmonds, 1982).

After 48 hours incubation at 36.5°C in a humidified atmosphere of 5% C0£ in air, to ensure logarithmic growth, the medium was aspirated and 0.2 ml aliquots of fresh medium containing a range of seven to twelve drug concentrations (three replicates per drug concentration) was added to the wells. Five control wells received drug—free medium. Following a 24 hour exposure at 36.5°C, the cells were washed three times each with 0.2 ml serum free medium and once with calcium and magnesium free phosphate buffered saline (PBSA). The cells were then detached using 0.2 ml of a 0.01% trypsin in 0.003% versene (Ethylenediaminetetra—acetic acid disodium salt, EDTA) mixture. This concentration and volume of enzyme mixture had previously been determined as being able to detach cells while causing the least toxic effect (see this Chapter Section 2.4 (L)).

The detached cells were removed from the wells using a Finnpipette (Jencons Scientific Limited) fitted with a 1 ml tip and diluted as necessary in medium to yield between 100 — 200 colonies after a 0.2 ml sample was plated in 50 mm dishes (Nunc) containing 5 ml pre—warmed and gassed medium. The contents of each dish were thoroughly mixed to ensure even cell distribution using a 10 ml sterile syringe fitted with a 16G 4 inch needle. Assays using the cell lines T24, MGH—U1 and MGH—U2 were incubated for 8 days with no media change. Cell lines HU961T and 253J for 10 days and HU456 for 11 days again all with no medium change.The lines RT112, VM—CUB—3 and TCCSUP were incubated for 14 days and the medium replenished on day 7, while RT4, HT1197 and HT1376 were incubated for 21 days and the medium replenished on days 7 and 14.

After the required incubation the medium from each dish was poured off, and each dish washed twice in slow running cold tap water and once with methanol. The cells were then fixed in methanol for 10 minutes, washed in tap water and stained with 10% Giemsa (Gurr's improved R66, BDH) for 5 minutes. RT4 cells stained rapidly in Giemsa, therefore a maximum staining time of 1 minute was used. Cells were then washed in tap water to remove excess stain and left to air dry. Colonies of 50 or more cells

only were scored using a binocular disecting microscope (Nachet), and the mean colony—forming ability of drug—treated cells expressed as a percentage of the control untreated cells.

(J) Comparison of growth rate curves of RT112 cells cultured inmicrotest plate wells and 35 mm dishes

Introduction:

The population doubling time (PDT) experiments were performed using 35 mm dishes (Nunc) containing 5 ml medium. However, due to economy measures, the drug treatment of cells for clonogenic assays was performed in microtest plate wells (Nunc) using 0.2 ml medium. Equivalent cell numbers per surface area were used that achieved exponential growth 48 hours after plating, and maintained growth in a logarithmic fashion during the following 24 hours, when cells would be exposed to drugs. It was therefore necessary to ensure that the growth rate of cells grown in microtest plate wells did not vary from that achieved with cells grown in 35 mm dishes (Nunc).

Method:

The surface area of each flat—bottom well is 0.32 cm^ and that of a 35 mm dish 9.62 cm , a 30 times difference. Therefore, using the cell line RT112, a range of cell densities from 1,600 to 25,600 were plated into microtest plate wells in duplicate, and a similar range of equivalent cell densities from 48,000 to 384,000 cells plated in 35 mm dishes.

At 24 hour intervals, from 24 to 96 hours after plating, the cells from duplicate dishes and wells of each cell density were enzymatically detached and diluted in filtered PBSA(for details see this Chapter, Section 2.4 (B)). Cell counts were performed using a Coulter Counter (Coulter Electronics) and the mean reuslts of two experiments plotted on semi—logarithmic/linear graph paper (Figure 2.4).

•H a>oJ-l cd4-> CM0) CO 5-1a. u 3a» COE x E E cm 3 *hm cC r H

CO05•H rH C5a) c•o o a;ai 5-1* 3 05Q> . CM05 r-s CMCO 1 •H1 * 3C M 1 tH W t-itH OH co CM« a>4->M05 (0 ac <-t 3•fH Q. O1—1 o4-1 or H CO cdr H 0505 4-r oa 4-4 •o COa> u — r Hx a r H4-4 *H o 05E C O >MO E M-l 05cm O o 4-44-1 cd

CO 4-4 5-t r Ha> o O a> x 4-4J-r o 4-43 4-4 cd COu cd CM 05rH M0) CM cd4-4 o(d r H >, 5-1U rH X Oa> ■ 4•Hx > E4->> v O * 3 * 3o o\ 05 C5-1 M cdbo -a COc 3 COM H cd •*“> 05O " 3 X

"Is cd COc *H° c * 3CO W 05•fH 05 CM5-1 CO X Ocd o> a x 05 cdE CO > 05O -H cd 5-1O T3 X cd

-a-CM055-13bO

o oo o o TDLD CD o o o CDin cu r o o ■D r-.cu CD cu CD CD ^h\ \ \ on a) cuo o o \ cn 3o o o o Oo o o o o cn -c00 cu o o rH WCD CD cn (D CD t— t -r!on ri cn "5T CD a _ CO ^ O)

-

_ 03

■<TCU

OmCM

ooo o o o m

• • • • •© in o m C Mm CM

(000‘0I x)nan/Hsia/HaawriN n ao aaxsnrav avxoi,

TIME

hrs

Results and conclusion:

It can be seen, Figure 2.4, that the mean growth rate curves, for each equivalent cell density, did not significantly differ over the 96 hour period. Thus, the growth kinetics of cells grown in microtest plates and used for drug sensitivity assays do not differ from those in 35 mm dishes used to determine the PDT.

(K) Experiment to determine at what stage RT112 cells, at an equivalent cell density used for clonogenic assay, enter plateau—phase growth

Introduction:

In a series of experiments (Chapter 4) to compare the effect of clonogenic assay methodology on in vitro drug sensitivity data, both exponentially—growing and plateau phase RT112 cells were used. It was therefore necessary to determine the time taken for these cells to enter plateau phase growth under standard growth conditions used in these studies.

Method:

Exponentially-growing RT112 cells were plated in RPMI 1640 medium plus normal supplements at a cell density of 384,000 per 35 mm dish, equivalent to 12,800 cells per microtest plate well, and incubated at 36.5°C in a humidified atmosphere of 5% CO2 in air. At 24 hour intervals four replicate dishes were removed and the cells harvested in the normal manner (see this Chapter, section 2.4 (B)). Cell counts were made using a Coulter counter (Coulter Electronics) and the mean results of two experiments plotted on semi—logarithmic/linear graph paper. The medium in the dishes was replenished on days 7, 11 and 15. In addition, cells from separate 15 day old cultures were harvested and 50%

Figure 2.5

Growth curve

of the cel

l lin

e RT112.

Exponentially-growing

cells

were

seeded into

35mm

dishes and cel

l number counted, using

a Coulter

Counter, from

four

replicate

dishes at

24 hou

r intervals.

Results

are plotted

as tumour cel

l growth (cell

number/dish) against

time.

_ 10

CO

- <0

- cn

oK>10 CN10

(suoixttui) qsip/jaquinu nao iqiAOjS nao jnouinj, me

dium replenished

serial dilutions plated into 50 mm dishes to determine the colony forming efficiency (CFE) of these cells. These cultures were maintained for 14 days before fixing as normal (see this Chapter Section 2,4 (A)),

Results and conclusion:

Exponential growth continues until the 9th day, after which time, even with replenishment of medium, the growth curve begins to level off and peaks on day 14 after which time viable cell number per dish begins to decrease (Figure 2,5). Therefore, in experiments using plateau—phase cells the cultures were maintained for 14 days before use with the medium replenished on the 7th and 11th day. Cell samples taken from 15 day old cultures exhibit a CFE range of between 14.4 and 15.3%, which is not significantly different from the CFE results of exponentially—growing cells. Thus, variations in drug sensitivity data between plateau-phase and exponentially—growing cells cannot be attributed to differences in the CFE of these two cell populations.

(L) Experiment to determine the effect of trypsin/verseneconcentration and volume on cell viability

Introduction:

In order to determine the effect of drugs on cells using the clonogenic assay method, performed in these studies it is necessary to transfer the cells from microtest plate wells to petri dishes. This cell transfer necessitates the use of an enzyme mixture to detach the cells. Thus, the purpose of this experiment was to estimate the lowest effective concentration of the enzyme mixture. In addition, the enzyme mixture might alter the colony forming efficiency when transferred with the cells, and therefore a comparison was also made between the CFE of washed cells recovered by centrifugation, thus removing the enzyme mixture, and that of cells transferred with the enzymes using a range of concentrations and volumes.

Method:

Exponentially-growing RT112 cells were seeded into the wells of a microtest plate (Nunc) at 320 viable cells per well, in 0.2 ml RPMI 1640 medium plus 5% FBS and glutamine at a final concentration of 2 mM, and incubated at 36.5°C in a humidified atmosphere of 5% CO2 in air for 24 hours. After 24 hours incubation the medium was aspirated, cells washed once with 0.2 ml PBSA and a 1:5 mixture of 0.25% trypsin (Difco 1:250) and 0.02% versene (BDH) added to each row of 5 replicate wells at concentrations and volumes shown in Table 2.11, and incubated at 36.5°C in a humidified atmosphere of 5% CO2 in air. Each concentration and volume of the enzyme mixture was used in duplicate. The cells in suspension from one duplicate set of five replicate wells were added directly to 50 mm dishes (Nunc), containing 5 ml of pre-warmed/gassed medium. The cells in suspension from the other duplicate set of five replicate wells were dispensed into cryo tubes (Nunc) containing 1 ml of fresh medium and centrifuged at 250 g for 10 minutes. The supernatant was aspirated and 0.2 ml of fresh growth medium added. The washed cells were then added to 50 mm dishes (Nunc) containing 5 ml of pre-warmed/gassed medium. The cells in each dish were thoroughly mixed to ensure even distribution using a 10 ml syringe fitted with a 16 G 4 inch needle then incubated at 36.5°C in a humidified atmosphere of 5% CO2 in air for 14 days. The medium was changed on the seventh day. The cells were then fixed and stained with 10% Giemsa and colonies with greater than 50 cells counted using a dissecting microscope (Nachet), (see this Chapter Section 2.4 (A) for details). The number of colonies expressed as a percentage of the total viable cells seeded in each microtest plate well represent the secondary CFE %. The mean results of duplicate experiments are shown in Table 2.11.

Results and conclusion:

Three factors on the effect of trypsin/versene on the secondary CFE of RT112 cells were measured in this experiment.

These factors were: (1) the concentration of the enzyme mixture, which consisted of the undiluted (neat) 1:5 mixture of 0.25% trypsin in 0.02% versene and a 1:1 and 1:5 dilution of this neatmixture in PBSA; (2) the volume of trypsin/versene added to eachwell of the microtest plate, which ranged from 50 |il to 200 jil; (3) centrifugation of cells, whereby detached cells were washed in 1 ml fresh medium and recovered by centrifugation to remove the enzyme before plating into 50 mm dishes (Nunc).

(1) Effect of trypsin/versene concentration: From the results it could be seen that as the concentration of trypsin/versene transferred with the cells was decreased the secondary CFEpercentage increased from 14.7% to 69.1% (see rows 2, 8 and 10,Table 2.11). Therefore it can be concluded that the enzyme mixture, although diluted in 5 ml of growth medium in the petri dishes, was still causing cell damage to the transferred cells. Thus the higher the concentration, the higher the incidence of cell damage, the lower the percentage of secondary CFE.

(2) Volume of trypsin/versene: There was an increase in secondary CFE percentage when smaller volumes of neat 1:5 trypsin/versene were used for cell detachment ranging from 14.7% to 48.4%. This was probably due to the fact that the smaller the volume of trypsin/versene transferred with the cells the higher the dilution factor when added to the 5 ml of growth medium in the petri dishes. Thus, the effect of enzyme concentration in the petri dishes was again demonstrated. The use of smaller and smaller volumes of trypsin/versene for cell detachment was limited because a point is reached when not all cells were successfully detached and thorough aspiration of the well contents becomes difficult.

(3) Washing and centrifugation of cells: The secondary CFE percentage remained relatively constant, at approximately 48% when the detached cells were washed in 1 ml fresh medium and recovered by centrifugation before transfer to petri dishes, regardless of the enzyme concentration used (see rows 1, 3, 5, 7 and 9, Table 2.11).

The conclusion drawn from these results was that when diluted trypsin/versene was used a higher secondary CFE was achieved when cells were not centrifuged. This was probably because the act of centrifugation may in itself damage some cells, therefore, with the inclusion of this extra step in the transfer process, additional cell loss may occur resulting in a lower secondary CFE. It was therefore proposed, as a result of these experiments, to use 200 |il of a 1:5 dilution of the neat 1:5 mixture of trypsin/versene for the transfer of cells in experiments where the effect of drugs on cells using clonogenic assays are to be determined.

Table 2.11 Effect of various concentrations and volumes of a 1:5 trypsin/versene mixture on the secondary colony forming efficiency of RT112 cells

MicrotestPlateRow

Concentration and 1 volume of T/V 1:5 1

11

Centri­fugation

MeanColonyCount

Secondary CFE %

1 200 pi1

neat 1 1

-t- 147 45.9

2 200 pi1

neat 1 1

— 47 14.7

3 100 pi1

neat 1 1

150 46.9

4 100 pi1

. neat 1 1

— 70 21.9

5 50 pi1

neat 1 1

+ 138 43.1

6 50 pi1

neat 1 1

- 155 48.4

7 200 pi1:1 1 dilution 1

1+ 161 50.3

8 200 pi1:1 1 dilution 1

1— 173 54.1

9 200 pi1:5 1 dilution 1

1168 52.5

10 200 pi1:5 1 dilution 1

1- 221 69.1

CHAPTER 3. CHARACTERISATION OF HUMAN UROTHELIAL CELL LINES

CONTENTS

PAGE

3.1 INTRODUCTION 166

3.2 MATERIALS AND METHODS ' 168

(A) Cell culture 168(B) Enzyme polymorphism analysis 168(C) Tumorigenicity in nude mice 168(D) Cytology 168(E) Colony forming efficiency 169(F) Population doubling time 169(G) Labelling index 169(H) Monoclonal antibodies 169(I) Time lapse cinemicroscopy 169

3.3 RESULTS 170

(A) Enzyme polymorphism analysis 170(B) Tumorigenicity in nude mice 171(C) Monoclonal antibody reactivity 173(D) Growth characteristics 175

3.4 DISCUSSION 176

TABLES AND FIGURES 183

3.1 INTRODUCTION

The natural history of human tumours and their clinical management are dictated primarily by the site and histological type of disease (Hill and Tannock, 1987). Nevertheless, among tumours of any one histological type there is a broad spectrum of biological behaviour and response to particular treatment regimen (Russel et al, 1986). Consequently model systems are needed which accurately represent not only individual tumours, but also reflect the range of properties of each histological type of disease. One in vitro model system that might encompass these dual criteria is a panel of continuous cell lines derived from tumours of one histological type. The concept of using a panel of cell lines derived from tumours of one histological type has, for example, been evaluated for lung tumours (Carney et al,1983), ovarian carcinomas (Ozols et al, 1984) and melanomas (Tveit et al, 1981b). However, the purpose of the studies presented in this chapter was to investigate the extent to which these two criteria are fulfilled by a panel of 22 cell lines dervied from human urothelium. Eighteen lines were derived from transitional cell carcinomas, one from a squamous cell carcinoma, and three from normal urothelium. No previous attempt has been made systematically to examine under standardised conditions the properties of a large series of human TCC cell lines, although there are numerous reports concerning individual or small numbers of cell lines derived from this histological type of tumour (Hepburn and Masters, 1983; Grossman et al, 1984; Kyriazis et al, 1984; Lin et al, 1985).

It was not until 1967 that the first continuous cell line was established from a human bladder carcinoma (Rigby and Franks 1970). The cell line was derived from a well-differentiated transitional cell cancer and designated RT4. There are now more than seventy continuous cell lines derived from both primary and secondary human bladder cancers of all different histological grades which have been reported in the literature. Brief details

of a number of these lines have been reported previously (Fogh, 1978? Williams, 1980), and thirty-eight reviewed in more detail by Hepburn and Masters (1983).

The biological characteristics of 22 human urothelial cell lines presented in this chapter include! (i), enzyme polymorphism analysis, to determine the phenotypic identity and human origin of each cell line,(ii) transplantibility in nude mice (nu/nu), to confirm neoplastic potential and histological appearance of the resultant xenografts, (iii) cytological appearance of smears made from monolayers of exponentially growing cells and (iv) production and characterisation of monoclonal antibodies measured by direct immunofluorescence. These monoclonal antibodies were raised against antigenic determinants present on the cell surface of one of the lines, RT112, and it was determined whether these antigens were common to all the lines. These immunofluorescence studies, the data of which has been published separately (Trejdosiewicz et al, 1985), were performed jointly with a number of colleagues at the Imperial Cancer Research Fund, London.

In addition to the characterisation studies mentioned above a number of growth characteristics were also investigated. These includes (i) the proportion of cells within each line capable of colony formation from a single cell — the colony forming efficiency (CFE), (ii) the population doubling time (PDT) determined from measurements extracted from growth rate curves and (iii) the labelling index measured after a 15 minute pulse with methyl—% thymidine.

3.2 MATERIALS AND METHODS

Twenty-two cell lines were included in this study, eighteen of which were derived ostensibly from transitional cell carcinomas, one from a squamous cell carcinoma, SCaBER and three derived from what was referred to as "normal" urothelium: HCV29, HS0767 and HU609. Details concerning the site of origin, year culture established, sex of patient, prior therapy and originator of each cell line are given in Chapter 2. Materials and Methods, Table 2.9.

(A) Cell culture

Full details are given in Chapter 2. Materials and Methods,2.2 (A). Cell stocks were held in liquid nitrogen and each line was used over a restricted number of passages (Maximum of 10) before returning to this stock.

(B) Enzyme polymorphism analysis

A maximum of thirteen polymorphic enzymes were isozyme—typed in each cell line. Details of the techniques, reagents and references are given in Chapter 2. Materials and Methods, 2.4 (G).

(C) Tumorigenicity in nude mice

Details of these studies are given in Chapter 2: Materials and Methods, 2.4 (E). Tumours were excised when they had grown to 0.5 — 1.0 cm in diameter, processed using routine histological procedures and examined using light microscopy.

(D) Cytology

For details of techniques and stains see Chapter 2. Materials and Methods, 2.4 (F).

(E) Colony forming efficiency

Details of this experimental protocol are given in Chapter 2. Materials and Methods, 2.4 (A).

(F) Population doubling time

For details see Chapter 2. Materials and Methods, 2.4 (B).

(G) Labelling index

Details of cell labelling and autoradiography are described in Chapter 2. Materials and Methods, 2.4 (C,i—iii).

(H) Monoclonal antibodies

A series of monoclonal antibodies, known as the LBS series, were raised against RT112 cells. The binding pattern of this series to seventeen human urothelial cell lines was assessed.The technique used was that of indirect immunofluorescence, described fully in Chapter 2. Materials and Methods 2.4.(H).

(I) Timelapse cinemicroscopy

The growth of a number of daughter cells over a few generations of two human bladder cell lines, HT1197 and T24, were studied using timelapse cinemicroscopy. Cells were incubated in dishes in a special chamber fitted to an inverted microscope under standard conditions for cell culture. Details of these conditions and the apparatus are described in Chapter 2:Materials and Methods, 2.4 (D).

3.3 RESULTS

All the human bladder cell lines were isozyme typed and checked for PPLO contamination. These cell lines can be divided into two groups. Firstly, there are those with a unique isozyme pattern and thus derived from different individuals. This first group demonstrates the range of isozyme phenotypes encountered in cell lines derived from this histological type of tumour. Secondly, there are a group of six lines (T24, J82, MGH—Ul, MGH— U2, HU456 and HU961T) which have identical isozyme patterns and are probably cross—contaminants to T24. This second group have been cultured separately in different laboratories for many years and thus can be used to investigate the stability of biological characteristics in long-term culture. Two further lines, VM—CUB—1 and VM—CUB—3, also had identical isozyme patterns, although different from the other group of six T24 cross—contaminated cell lines. Thus, only one, VM-CUB—3, was included in the other characterisation studies. Five of the cell lines (PSI, C0L0232, J82, KK47 and J82C0T) were PPLO positive on repeated examination and thus were not studied in detail.

(A) Enzyme polymorphism analysis

Analysis of up to thirteen polymorphic enzymes was carried out on 22 cell lines investigated in this thesis. The results are shown in Table 3.1. Amongst these cell lines there were sixteen that differed at one or more allele. However, T24 and five other lines (J82, MGH-U1, MGH-U2, HU456, HU961T) were identical at each locus for all thirteen polymorphic enzymes, indicating that these almost certainly have been cross- contaminated. In addition, two further lines, VM—CUB—1 and VM- CUB—3 were also shown to have an identical isozyme pattern. J82C0T, the original stocks of one of the T24 cross—contaminated lines, J82, had an isozyme pattern identical to another line derived from the same patient and distinct from T24 (S Povey, Personal Communication). None of the lines has the same isozyme

pattern as HeLa cells, but the KK47 line differed at only one locus so the possibility of it being a HeLa cell derivative cannot be entirely excluded on these data alone.

The three cell lines said to have been derived from "normal" urothelium (HCV29, HU609 and HS0767) all showed an adenosine deaminase (ADA) of slow electrophoretic mobility (designated "tissue" in the table) which indicates that the enzyme has formed a large molecular weight complex with a glycoprotein (Swallow et al, 1977). Apart from making the ADA polymorphism untypeable this is of some interest in confirming the "untransformed" phenotype of these cells, since this form is very common in normal adult fibroblast cultures but is virtually never seen in transformed cells in culture (Inwood et al, 1980).However, the cell line HT1197, which forms tumours on xenotransplantation, also expresses this "untransformed" phenotype for ADA, in contrast with all the other tumour lines described in this study. The enzyme polymorphism analysis described was kindly performed by Dr Sue Povey, The Galton Laboratory, University College, London.

(B) Tumorigenicity in nude mice

Differences amongst the cell lines in the proportion of cell implants producing tumours and the period of time taken for these tumours to develop is shown in Table 3.2. Two tumour—derived cell lines (T24 and TCCSUP) and the three lines derived from normal urothelium (HCV29, HU609, HS0767) failed to develop tumours during a maximum follow-up period of 2 years. Some of the mice used in this study, particularly those observed over long time periods, died as a result of complications such as preputial abscess or lymph node disease, before the formation of tumours could be observed.

The histological sections of the xenografts were also examined independently by two consultant urological histopathologists based at St Paul's Hospital, London. The only information supplied to the histopathologists was that these were

xenografts of cell lines. The two morphological descriptions of each xenograft were in agreement and are summarised in Table 3.2. Replicate sections of xenografts from the same cell line, but from different mice, had a similar morphology. A firm diagnosis of transitional cell cancer was made for the cell lines RT4, RT112, 253J and the four cross—contaminated T24 sublines assessed in these tumorigenicity studies, namely: MGH—Ul, MGH—U2, HU456 and HU96IT.

The xenografts produced by RT4 and RT112 had morphologies in most respects identical to those of the patient's original biopsies taken in 1967 and 1973 respectively, although areas of squamous metaplasia were more frequent in the xenografts. These comparisons are shown in Figure 3.1. Innoculation of HT1376, HT1197 and VM—CUB—3 cells into nude mice produced poorly- differentiated tumours diagnosed as squamous cell cancers or transitional cell cancers with extensive squamous metaplasia. Nevertheless, the xenografts produced by HT1376 and HT1197 had morphologies similar to those illustrated in the original description of these cell lines (Rasheed et al, 1977), which were derived from tumours described respectively as poorly- differentiated transitional cell carcinoma and transitional cell carcinoma with nests of large pleomorphic squamous carcinoma cells. Thus, in the cases in which the pathology of the xenograft and the parent tumour could be compared, the morphologies were similar. Histological sections of all the tumours produced as a result of ihnoculating the cell lines s/c into nude mice are shown in Figure 3.2.

The cross—contaminated sublines of T24 produced xenografts with some difference in morphological appearance. Paraffin sections of tumours produced by T24 (kindly supplied by Dr L M Franks, Dept Cellular Pathology, Imperial Cancer Research Fund, London) show a spindle cell morphology which was also seen in the MGH—Ul and MGH—U2 cell lines. On the other hand,

the morphologies of the tumours produced by the sublines HU456 and HU961T differed. They produced transitional cell carcinomas with areas of squamous metaplasia. Nevertheless, glandular and squamous metaplasia and spindle cell changes can each be observed in transitional cell cancer and thus all the sublines produced tumours with a possible diagnosis of TCC.

Although the cell line KK47 was shown to be contaminated with PPLO, tumorigenicity studies were performed because there was a suspicion that this line may be HeLa. This suspicion was based upon the enzyme polymorphism analysis which only differed from HeLa at one locus. Also it has a very rapid growth rate. However, morphological examination of the tumours produced by KK47 show papillary transitional cell cancer with a high mitotic rate and some areas of squamous metaplasia.

All the cell lines had a cytological appearance consistent with an origin from transitional cell carcinoma, Table 3.2. The squamous elements seen on xenotransplantation were not apparent on cytological examination, although there was considerable variation in the degree of cellular differentiation. Therefore, the areas of squamous metaplasia, frequently observed in the xenografts, perhaps reflect differences in the hormonal environment in the mice.

(C) Monoclonal antibody reactivity

The results obtained by immunofluorescence are summarised in Table 3.3. On the basis of reactivity with the panel of continuous cell lines the antibodies were subdivided into 3 groups. Group I (LBS—1 and LBS—19) reacted with all epithelial human derived cell lines tested, but not with murine cell lines. Of the remaining 9 monoclonal antibodies, in groups II and III, none reacted with all human urothelial cell lines. Group II antibodies (LBS—2, 8, 15 and 17) tended to react with the more differentiated cell lines such as RT4, RT112 and HT1376, but

weakly or not at all with the anaplastic lines such as T24 and TCCSUP. The Group III antibodies (LBS-10, 20A, 20B, 21 and 34) were characterised by a generally weaker immunofluorescence reaction and showed no clear correlation between reactivity profile and degree of urothelial differentiation. Differences in the immunolocalisation patterns can be seen in Figure 3.3.

Generally, when an immunofluorescence reaction was observed it could be visualised on the methanol—acetone fixed cells (showing cytoplasmic localisation) as well as on unfixed or formaldehyde fixed cells (showing the cell—surface localisation). However, the Group III antibodies showed predominantly cytoplasmic localisation patterns with weak or negative cell surface reaction (see Table 3.3). The cytoplasmic localisation pattern shown by these antibodies was usually rather amorphous (Figure 3.3(a)), but occasionally a reticular localisation was observed for some cell lines, such as VM—CUB—3 (Figure 3.3(b)). The antibody LBS—20A proved to be the exception in this group showing a fine "speckly" localisation pattern on some cells (Figure 3.3(c)).

The Group I antibodies (LBS—1 and LBS—19) showed characteristic localisation patterns on all the cell lines.Bright cytoplasmic fluorescence around the nucleus in a "signet ring" distribution could be seen on the methanol—acetone fixed cells (Figure 3.3(d)), whereas the unfixed or formaldehyde fixed cells showed a punctate distribution over the cell surface (Figure 3.3(e)).

Of the Group II antibodies, both LBS—8 and 17 showed a characteristic cell "border" localisation pattern both on unfixed and methanol—acetone fixed cells (Figure 3.3(h) and (i)).However, LBS—2 and 15 showed a bright surface localisation in an "apical" distribution (Figure 3.3(f)), whereas intracellular

localisation varied with the cell type. For both KK47 and HT1376 cells these antibodies showed a cytoplasmic filamentous "cytoskeletal" localisation pattern (Figure 3.3(g)). However, a positive immunofluorescence reaction on all the other urothelial cell lines showed a more amorphous localisation pattern, similar to that observed for the Group III antibodies.

(D) Growth characteristics

The colony forming efficiency (CFE), population doubling time (PDT) and labelling index (LI) of each cell line are shown in Table 3.4. All the lines show a linear relationship between cell numbers plated and colonies formed, i.e. they do not show density dependence, and therefore only an average figure is given. The similarity between T24 and its cross-contaminated sublines, in the growth characteristics measured, contrast with the wide variation among the other distinct lines. For the CFE determinations colonies with greater than 50 cells were found within 8 days for T24, MGH—U1 and MGH—U2j 10 days for HU961T and 253J; 11 days for HU456; all without replenishment of the medium. The remaining cell lines received fresh medium every 7 days, with RT112, TCCSUP and VM-CUB-3 requiring 14 days and RT4, HT1197 and HT1376 21 days to achieve a colony size greater than 50 cells. The CFE for the distinct cell lines ranges from 3.2 to 86.8% and the PDT from 19 to 60.7 hours, whereas T24 and its cross—contaminated sublines show a much narrower range for CFE measurements from 60.5 to 86.8% and all show a similar PDT from 19 — 21.2 hours.

In general there appears to be an association between a short PDT and both high CFE and LI. An exception to this trend was observed for the cell line HT1197 which, despite having the longest PDT and lowest CFE, has a relatively high LI of 37.3%. This finding was investigated further using time-lapse cinemicroscopy, comparing the cell division times and cell loss rate of HT1197 and T24 (see Table 3.5). The apparent paradox may be explained by the high death rate of HT1197 (36%) which occurs in in vitro culture, compared with that of T24 (0.5%).

3.4. DISCUSSION

This chapter describes certain biological characteristics of a panel of continuous cell lines derived from transitional cell carcinoma (TCC) of the human bladder, maintained under standardised culture conditions. A number of these cell lines were subsequently used in the parallel studies of in vitro drug sensitivities (see Chapter 6). Together the data provide a basis on which to evaluate a panel of cell lines as a model system for tumours of this histological type.

Within the last few decades many electrophoretically distinguishable enzyme polymorphisms have been recognised. Differences in the phenotype of the loci expressed by cell lines can be distinguished using starch gel electrophoresis. Isozyme analysis is a most useful biological characteristic, the study of which can be used to identify cross—contamination between cell lines, determine the species of origin and, if tissue from the patient is available, establish that the cells are derived from that individual. This approach may be superceded by the more sensitive analysis of DNA polymorphisms (Jeffreys et al,1985a,b). Isozyme phenotypes are retained in vitro and usually remain stable over many years in culture (Wright et al, 1979).The isozyme patterns of some of the cell lines reported in this chapter have been described previously (Povey et al, 1976; Wright et al, 1981; Dracopoli and Fogh, 1983a,b; O'Toole et al, 1983) and these data are in agreement.

A number of bladder tumour cultures appear to have been cross—contaminated with T24 cells in at least three separate laboratories. Furthermore, the two VM—CUB lines both show identical isozyme phenotypes and thus have presumably become cross—contaminated at some time. Such cross—contamination is a common problem and must be excluded in studies using panels of cell lines as described in earlier publications (Povey et al, 1976; O'Toole et al, 1983).

The cell line KK47 was originally suspected of being HeLa, however it differs from strains of HeLa previously tested by Dr Sue Povey, Galton Laboratory, University College, London (personal communication) at the locus ACPI. A number of variants of the A allele for G6PD, which is fast moving on starch gel electrophoresis, have been reported in the literature, and hence, they are not easily distinguished (Yoshida and Beutler, 1983). Nevertheless, it is always conceivable that the expression of the A allele could have been lost. Loss of function of an allele resulting in increased apparent homozygosity is reported (Dracopoli and Fogh, 1983a), but it has not yet been established whether this loss occurs in the tumour in vivo or as a result of culture in vitro.

Xenograft production is another essential aspect of cell line characterisation. It demonstrates that the cells have retained a neoplastic phenotype and have a morphology compatible with that of the tumour of origin. In the largest series studied 127/162 (78%) produced tumours, the histopathology of which correlated with the tumour of origin in every case (Fogh et al, 1977b). The ability to generate a defined morphology, following xenotransplantation to an ectopic site, provides the most convincing evidence that a cell line has retained significant properties of the tumour of origin. In this study the similarities between the morphologies of xenografts produced by RT4 and RT112 cells and sections of the original biopsies, from which these lines were established in 1967 and 1973 respectively, were demonstrated by direct comparison. The morphology of the RT4 xenograft was in parts identical to that of the tumour of origin and RT112 produced a typical moderately well—differentiated transitional cell carcinoma. Amongst the xenografts, squamous metaplasia was more common and widespread than would be expected in transitional cell carcinoma, perhaps reflecting differences in the hormonal environment at a subcutaneous site in nude mice compared with the human bladder.

However, as a group, the xenografts showed a range of morphologies and histological grades compatible with those of TCC.

Tumorigenicity was highly variable among the five T24 sublines tested. A similar finding was observed by Marshall et_ al (1977). In addition, cloning of one of these T24 sublines, MGH—U1 (also designated EJ in some studies) also resulted in lines differing widely in tumorigenicity and growth characteristics (Hastings and Franks, 1983). A similar loss of tumorigenicity following in vitro culture was described by Tveit and Pihl (1981) using a series of cell lines derived from human melanomas. Specific reasons for this variability in tumorigenicity have not been fully defined. However, the T24 cell line was shown to have a much lower colony forming efficiency in agar compared to the subline MGH—U1 (EJ) (Marshall et al, 1977) which may be related to the tumorigenicity of these lines in nude mice. Furthermore, whether these differences reflect major changes in either gene function or cell surface antigen expression (altering the immunogenicity of the cells in the host mice) is a matter of speculation. Nevertheless, T24 and its sublines might provide a system for studying the factors involved in tumorigenicity, particularly as T24 was the first human cell line shown to contain transfectable oncogenes (Shih et al, 1981).

Cytological preparation of cell lines, examined in conjunction with an experienced cytopathologist, provide an additional means of characterisation. In contrast to the routine examination of cell lines under phase microscopy, the cytological preparations can be used to diagnose the histological type of the cells (Hajdu et al, 1974). All the lines examined in this study were considered to be of a transitional cell origin, but with variying degrees of differentiation, reflecting the heterogeneity seen in the clinic.

The majority of the in vitro growth characteristics of the T24 sublines were similar, contrasting with the wide variation among the other distinct cell lines. The relative stability and reproducibility of these data probably result from the use of standardised conditions. All the lines were grown in one type of tissue culture medium supplemented with a single batch of foetal bovine serum and used over a restricted passage range.Alteration in any one of these conditions can radically alter the properties of cells in vitro. In general the CFE is inversely proportional to the PDT, with the cell line having the lowest proportion of clonogenic cells capable of indefinite growth appearing to have the longest PDT. However, it should be remembered that the PDT results are for the population as a whole and may not necessarily reflect the cell cycling time of the clonogenic cells. This is particularly illustrated for HT1197 which has the lowest CFE and the longest PDT yet a relatively high LI. The explanation for this was revealed by time-lapse cinemicroscopy which showed that HT1197 had a high cell death rate and a relatively high frequency formation of giant cells containing increased nuclear material. The growth rates of all the cell lines studied are considerably higher than would be expected for human tumours, either as measured directly in patients (Shackney et al, 1978) or indirectly in biopsies in vitro (Levi et al, 1969). Similarly, the CFE on plastic (3 — 87%) is far higher than would be achieved by cells taken directly from human biopsies and grown in agar (Sarosdy et al,1982).

The monoclonal antibodies raised against RT112 cells are classified into three groups according to their reactivity with the panel of cell lines. This work has been published in detail elsewhere (Trejdosiewicz et al, 1985). Localisation of the LBS series of monoclonal antibodies is predominantly cytoplasmic. However, the biochemical analysis defining the epitopes they recognise has not been determined. Although all the antibodies are not totally urothelial specific it is thought they can be

used as subtle discriminators of phenotype of urothelium derived cell lines. None of the antibodies reacted with human or murine fibroblast cell lines or human lymphoblastoid cell lines. Group I (LBS—1 and LBS—19) reacted with all human epithelium—derived cell lines tested but not those derived from murine origin.Group II (LBS 2, 8, 15 and 17) reacted only with human urothelium—derived cells, whereas Group III (LBS 10, 20A, 20B, 21 and 34) reacted not only with urothelium derived human cell lines but also with epithelial and urothelial murine cell lines.

In addition to characterising cell lines derived from transitional cell carcinoma, preliminary studies on three lines thought to be derived from normal urothelium were included. The culture of human normal transitional cells in vitro is generally considered to be difficult (Knowles et al, 1980), although using specialised culture conditions it is possible (Reznikoff et al,1983). However, no permanent cell lines were established and these cells eventually senesced. The fact that normal cells require special growth conditions and have a limited lifespan has also been shown in other studies (Chlapowski and Haynes, 1979; Pauli et al, 1980; Kirk et al, 1985). Therefore, the three cell lines defined as "normal" in this study are by definition composed of immortalised cells and hence, are abnormal. Nevertheless, data from these studies is in support of their origin from normal tissue. All three lines expressed the untransformed phenotype of the ADA enzyme, in contrast to only 1/13 of the tumour—derived lines. In addition, none of these "normal" lines developed tumours on xenotransplantation in contrast to 7/8 of the tumour-derived lines. The untransformed phenotype is seen in virtually all normal adult fibroblast cultures, but is rare in transformed cells in culture (Inwood et al, 1980). Cell lines derived from normal urothelium could have many applications, for instance, in transformation experiments, in toxicology studies and to screen for differential cytotoxicity of potential chemotherapeutic agents. However, further characterisation will be needed to establish that these three lines are urothelial.

The ultimate goal of the studies presented in this chapter was to assess the relevance of a selected panel of human bladder tumour cell lines as a model system for transitional cell carcinoma. In order to validate their use for this purpose, three criteria need to be satisfied: firstly, each cell line used within a panel should be representative of its tumour of origin, secondly, the panel as a whole should reflect the spectrum of properties encountered clinically in tumours of that histological type and thirdly, these properties should be stable in long-term culture. The first criterion, that each cell line should retain certain properties of the tumour of origin is, in general, met since polymorphic enzyme phenotypes are usually retained and, on xenotransplantation, most lines are tumorigenic and retain the characteristic and heterogeneous morphology of the parent tumours. Thus, a combination of isozyme analysis and xenotransplantation can be used to exclude the minority of lines not retaining these fundamental properties of the tumour of origin. Nevertheless, some characteristics, such as growth rate and clonogenicity, do alter as a result of in vitro cell culture. Thus cells in culture may only be representative of the stem cell population in the parent tumour. Furthermore, these characteristics will differ depending on the culture conditions, therefore these should be standardised. The second criterion, that the cell lines should reflect the range of properties of transitional cell carcinoma is also met, in general, since the cell lines show a wide range of morphology, degree of histological differentiation, biochemical properties, growth characteristics, sensitivities to chemotherapeutic drugs (see Chapters 5 and 6), and reactivity with monoclonal antibodies (Trejdosiewicz et al, 1985). These findings confirm similar studies on cell lines derived from human lung tumours (Baylin e_t al, 1984), ovarian carcinomas (Hamilton et al, 1984a) and melanomas (Tveit and Pihl, 1981). However, the number of cell lines of each histological type required and the basis on which to select these to provide an adequate representation of each disease are yet to be defined. The third criterion, that certain properties should remain stable in long-term culture, is the most

demanding. It is also paradoxical since tumours are inherently heterogeneous and capable of generating diversity, for example by developing drug—resistant phenotypes. One means of addressing the question of stability in long-term culture was the use of the cross—contaminated T24 sublines, since the cells had been grown in various laboratories under different culture conditions, separately for periods of up to 10 years. The sublines were similar in their isozyme patterns, growth rates and in vitro drug sensitivities indicating that these features are relatively stable in long-term culture. In contrast, the sublines showed marked differences in tumorigenicity and reactivity with certain monoclonal antibodies demonstrating that some properties are not stable in long-term culture. To limit such differences and ensure inter—laboratory comparability it would be necessary to use a common stock of each cell line over a limited number of passages and the same culture conditions.

In conclusion, the studies described in this chapter have explored the concept that a group of cell lines derived from human transitional cell carcinoma might provide a useful model system for this histological type of tumour. The data provide evidence in support of this concept, but also illustrate some of the limitations of such a model system. The disadvantages include the need for rigorous standardisation of culture conditions, the requirement to characterise large numbers of cell lines and exclude cross—contamination and mycoplasma infection. Furthermore, certain characteristics are not stable in long-term culture. The advantages on the other hand include the unlimited supply of cells derived from each human tumour and the relative stability of features such as enzyme phenotypes, in vitro drug sensitivities and the ability to generate tumours histopathologically similar to the tumour of origin. In addition, the heterogeneity observed clinically within one histological type of tumour are displayed by the panel of cell lines.

Isozyme

profiles of

human

bladder

cell lines

cn

r—IXiCOH

<■<<<«(<< Q < ■< •< •<(Qe3S3e3fltgutaz<atQa«9d ea pa a oa •<< U oa

I I I I I I I IIn c m n n n n n h h h n n h

I Ih h ^ h h h m n n h m n n ^ h

a a o o aZ Z Z Z H Z

I a I a ar-t CM Z f t <n z z

a |

O a.J 3• - e> un h o•• t-tT3 •< «)e o -uto m ai coo c uZ y *hH v BO «wo CO ou <d u «>oCMCM pH pH CM CM Z CM p H CM z •* y a.« 4-1Z ooo a V <y V •vu •

Q 3 «V yCQ OQ ca po < PQ oo < Z 03 00 a a 4-iCOo oo cd «du M c 4-13 o <da u H u z<z< << ■<< < < ><< 00y 3 a.o COH coy y0) V Z CD CO oa 3 3 3 a. c o X.00 oo 00 CO CO •H z a.00 co oo co o a a.pH pH •pH pH z cd 00 yH H H a. « o 4-1e z idI4Hcd P> |-H

p H p H Ou l p HI II I

H 4 Z H “4

I IM H N

pH pH p H pH ^ 35

4-1 VO

O 0) »O u Q *-l « O

4-1 Op-o uM O ••<H « «£ O ffl4J i - t <0

I Q

I a

II I II I( M N N H N H N N H N

00 pHCM H Eh•pH PPX•pH OCMCO 1TinM3 UpHy y p»X p»O'CMCMpHCJ a CMoa=3XXXEhHzzX—-5-3OZ< U (O H HH a m h Has H cm x x

I Ii s (B 83O' N 3 DY Yz z > >

CMCO OS r -CM Cd O' O' X

r-p o 03 CM o f ' '4 , J H m > vfl o34 O CO O O 3 C/5^UOiUIXBS

ys a. V•oXpH pHcOa a.cdryCL, eVVU o4-14-1■oCL, yCOyMay <hHye y oUh4JpHa •«300ypHOT3 VcoeocpH toV a 3T>V•oo yepHy o yyz 3 y3 v zpHpH<z oa o Vua< pHCL, zT> ye yo y yyCO3 to a.•• yCx y >3apHs 4-1 4-1cypH«yyopHz yz e4JhHM oo a.oo4J•yVy a•v c«e4JpHopHoycdz4-1—

> zpHK a.oyyySh O c3uoe a •o az4-1y>>pH azpH■opH y Q pH< ay CjO yzH

Table 3.2 Cell line morphology and tumorigenicity

CELL LINE

T24

MGH—U1 (EJ)

MGH—02

HU456

HU961T

RT112

RT4

253J

HT1376

HT1197

VM—CUE—III

TCCSUP

RX47

CYTOLOGICAL APPEARANCE OF CULTURED CELLS

Differentiated TCC

Differentiated TCC

Differentiated TCC

Differentiated TCC

Differentiated TCC

Differentiated TCC

Differentiated TCC

TCC, probably differentiated

TCC, aostly differentiated

Pleoaorphlc TCC, ■lzturedifferentiated and anaplaatlc cells with squaaold change

TCC, probably differentiated

Consistent with differentiated TCC

Differentiated TCC

Abbreviations!

XENOGRAFT HISTOPATHOLOGY AND GRADE

Did not fora tumours

G3, possibly TCC with spindle cell areas and high MR

G3, possibly TCC with spindle cell areas and high MR

G2, TCC vlth squamous metaplasia

G2-3,TCC vlth squamous metaplasia

Gl-2, TCC with focal squamous and glandular metaplasia

G2, TCC vlth squaaous metaplasia and high MR

Gl-2, TCC with glandular metaplasia

G3, squaaous cell or extensive squamous metaplasia, and high MR

G3, squaaous cell ca

G3, squaaold cell ca

Did not form tumours

G2, TCC vlth squamous metaplasia and high MR

TUMORIGENICITY (TIME TAKEN FOR TUMOUR TO GROW IN NUDE MICE)

0/8 (after 24 months)

4/4 (1 month)

4/4 (1 month)

3/8 (6—20 aonths)

3/3 (5-6 aonths)

3/3 (1 month)

4/4 (2-5 aonths)

2/2 (6—12 aonths)

3/3 (1 month)

4/4 (1—20 aonths)

4/4 (2 months)

0/8 (after 24 aonths)

2/4 (2 months)

TCC — transitional cell carcinoaa MR - altotic rate

Table 3.3

Reactivity of the human bladder cell line panel with the LBS—series monoclonal antibodies

Group I Group II Group III

Cell Line 1 19 2 8 15 17 10 20A 20B 21 34

T24 A-/A- A /A -/+ +/- +/-I- -/- -/- -/- -/- -/- -/-

MGH—U1 (EJ) A-/A- A /A -/- -/- -/- -/- -/A- +/- +>/- A-/- A-/-

MGH-U2 A-/A- A / - -/- - /A t - /A t -/A- -/- -/- -/- -/-

HU456 A-/A- A /A -/- - /A t -/- - /A t -/- -/- A-/- A r /- A-/-

HU96IT A-/A- A /A -/- -/- —/■¥ - /A - -/- -/- -/- -/- A - /-

RT4 A-/A- A /A +/+ +/+ At/A- At/A- A-/A- A t / - +/- -/- A - /-

RT112 A-/A- A /A +/+ +/+ WA- Ar/A- A - /- +/- At/- +/- A-/A-

HT1376 A-/A- A /A +/+ +/+ Ar/A- A t/~ -/- -/- -/- A-/A- A-/A-

HT1197 A/A- A/A -/+ +/- +/- -/- A-/~ +/- -/A- A - /- A - /-

COLO 232 A/A A/A -/- +/- A-/~ -/- +/- +/- -/- A - /- -/-

KK47 A/A A/A W - -/- A-/A- -/- -/— A - /- -/- A-/A- +/-253J A /A A /A -/+ -/- - /A - -/- -/- -/- -/- -/- -/-TCCSUP A /A A / - -/- -/- -/- -/- -/- -/- -/- A-/~

VM—CUB—1 A /A +/+ -/- -/- A-/A- -/- -/- -/- -/- -/- A - /-

VM—CUB—3 A/A +/+ - /A r - /A t - /A - -/- -/- - /A t -/- A - /- - /A -

SCaBER A/A +/+ -/- +/+ -/- At/A- -/- -/- -/- -/- A-/—

HCV-29 +/+ +/+ -/- -/+ -/- -/- -/- -/- -/- -/- -/-

Immunofluorescence reactions are scored separately for cytoplasmic (C) and cell surface (CS) [C/CS]

Antibodies are arranged into 3 groups on the basis of their reactivity with cell line panel! Group I are human specific 'pan—epithelial'j Group II are human specific urothelial; Group III also react with mouse urothelial cell lines.

Table

3.4

Growth characteristics

of the human

bladder

cell lin

e panel

4)CJT)0)U«0eT-4a4-1cor09COOMU

o Z H pH CO O CM 00 p-* CM 00 CM cn-J • • • • • • • • • • •-J X cn CM CM CO pH O' vO CM CM p-**Cd Ed m in m in m CM cn CM cn H cn03 O3 5

Z Zt-t o x ac ^ ' ai >• O O U h Z H oo• P»s• pH m• m• 00• r*• oo• n• o O' CMtd H o in O CO m pH pH • • •>< M MZ O < O l-H J J ex a.O ChU Cd

Cd 3C Z )-»O H ">■ i-t oo H O W *

CO p «n \Q CO pH CM CM in r-* cn cn

o p H r«* ON pH p H ON 00 CM oo PN.< Z 3 • • • • • • • • • • • •J >-i O ON o O o pH m r** CO O \o OS >J s pH CM CM CM CM CM CM CM CM cn cn vO0, CQ wO =3 CL* Oa

aCdCl, c/3 O Cd S3 COcn CMp H vO o cn m oo p H \0 mOCd ■< C/3o oo as T

ONp Hln

O'1r>* fp H

sOJr

<rJ,

CMisO TONO'1CM X 1 1

Z CO Cd CM O CO m in cn p H m 00 cn cn< < aa as a. z 3 Z

Z Cd O Z M

J O J h-HCMT

p H

F-Nw

?33H!“■4 sO O' m CM

p HpH

cn1CQ3Y

cu3cna ncn

MDcnfH

p^O'pHpH

Cd CO O O CM =3 H 33 o m H H HO Cd 3C 33 33 H 33 03 > H CM 33 04 33Q

Table

3.5

Estimates

of cell los

s rat

e and lntermltotlc times

of HT1197 and T24

followed for two cell

U-lOT3H<u <v•H 4-1(4-1 CO»-iV 4-1X CO4-1 JO

3a COoM 4>(4-1 X4-14-1CO aO o»—H M

(4-1mM 4-1V c» <UaCO xfH oi-( COV 4-1u U-o0)a Mo Ocn eo• •HSs 4-1a, COo Mo 00CO •Ho auo (4-1•H oa0)c 4-1•H c-HO 3COa) a;CO Ma(0 COrH1 COa> COa•H t-i4-1 <0Sn 4-1*o •Ha)COco T?•H 0)4-1 4-1co co1—10) 3C(V CO00 4-1

a -

HlOcn

■O <4-1U O<M T>4-1 < -t 9a « vO- J <W >

OCNA

«4>CNI

CNO

?ao

41oo ee «co 41ac £

oI <-< aoi H mW O M 9H H ac oZ l-H HH a►h z H W

X oo

Figure 3.1 Comparison of original patient biopsy and xenograft of resultant cell lines RT4 and RT112

Patient biopsy, RT4 H 6c E x 210

Patient biopsy, RT112 H 6c E x 210

iis*U

Xenograft of RT112 cell lineXenograft of RT4 cell line H 6c E x 84 H 6c E x 84

A- KJ J

Figure 3.2 Histological section (H & E) of tumours formed in nude mice as a result of subcutaneous injection of each cell line

RT4

H & E x 84

RT112H & E x 84

T24H & E x 84

MGH - U1 H 6c E x 210

1

Figure 3.2 (continued)

MGH - U2 H & E x 210

H13456H 6c E x 210

w m mftAt .. . ■iHfitt@3SB

HT1197H 6c E x 210

1

Figure 3.2 (continued)

VM-CUB-3 H & E x 84

HU961T H & E x 84

253JH & E x 210

KK47H & E x 210

Figure 3.3 Immunofluorescence micrographs showingcharacteristic cell—surface and intracellular immunolocalisation patterns seen with LBS—series monoclonal antibodies on human bladder cell lines

Key:

a) LBS-34 (KK47 - methanol/acetone fixed)

b) LBS-10 (VM-CUB-3-methanol/acetone fixed)

c) LBS-20A (RT112 - methanol/acetone fixed)

d) LBS-19 (T24 - methanol/acetone fixed)

e) LBS-19 (RT112 - unfixed)

f) LBS-2 (RT112 - formalin fixed)

g) LBS-2 (HT1376 - methanol/acetone fixed)

h) LBS - 8 (RT112 - unfixed)

i) LBS - 8 (RT112 - methanol/acetone fi^ed)

1 y J

Figure 3.3 (continued)

CHAPTER 4. COMPARISION OF CLONOGENIC ASSAY PROCEDURES FOR MEASURING THE DRUG SENSITIVITIES IN VITRO OF HUMAN TUMOUR CELL LINES

CONTENTS

PAGE4.1 INTRODUCTION 196

4.2 MATERIALS AND METHODS 197(A) Cell culture 197(B) Clonogenic assay 197(C) Feeder cells 198(D) Drugs 199(E) ^n—thymidine labelling 199

4.3 RESULTS 199(A) Methotrexate 200

(i) Comparison of in situ and transfer assay 200methods using cells plated at low density

(ii) Comparison of low and high cell density 200using the transfer assay

(iii) Effect of delayed transfer 201(iv) Effect of feeder cells 201(v) Comparison of the drug sensitivity of low 201

and high density exponentially-growing cells with quiescent cells

(B) Adriamycin 202(i) Comparison of in situ and transfer assay 202

methods using cells plated at low density (ii) Comparison of low and high cell density 202

using the transfer assay(iii) Effect of delayed transfer 203(iv) Effect of feeder cells 203(v) Comparison of the drug sensitivity of low 203

and high density exponentially—growing cells with quiescent cells

(C) % —thymidine labelling 203

4.4 DISCUSSION 204

TABLES AND FIGURES 210

4.1 INTRODUCTION

There are many methods whereby cell survival in vitro may be assessed (see Chapter 1, Section 1.6 (C)). However, the majority of comparative studies indicate that only clonogenic assays accurately measure reproductive capacity (Roper and Drewinko, 1976; Bhuyan et al, 1976; Hill et al, 1981; Belfield and Simmonds, 1982; Rupniak et al, 1983b). In contrast, a review of clonogenic and non-clonogenic in vitro chemosensitivity assays (Weisenthal and Lippman, 1985) suggests an opposing view that some non—clonogenic assays may be more advantageous for estimating the general chemosensitivity of proliferating cells. Nevertheless, clonogenic assay has become the most widely used technique for quantifying, in vitro, the cytotoxic effect of anticancer drugs by measuring the reproductive cell survival of cells from established continuous cell lines (Hill, 1983a) and freshly isolated human tumour cells (Salmon et al, 1980a; Carney et al, 1980).

Despite the widespread use of clonogenic assays standard conditions and procedures have not been established. The method usually adopted is to expose exponentially—growing cells to an anticancer drug and then transfer them as a single cell suspension. The number of cells capable of developing colonies, expressed as a proportion of the untreated controls, can then be determined. A variation of this method is the introduction of a delay period between drug exposure and cell transfer. In addition, some cell lines require the use of a feeder layer to permit colony development and eliminate the non-linearity of colony formation (Eliason et al, 1985). This adds a further modification to the assay procedures. Another method, which departs furthest from the basic procedure, is to leave cells in situ to develop colonies following exposure to cytotoxic drugs.It is clear that all these various assay modifications are likely to influence reproductive cell survival. Nevertheless, although the theoretical implications have been considered (Twentyman, 1979), few comparative experimental data are available.

The purpose of the study described in this chapter was to determine the practical effects of eleven variations of the clonogenic assay procedure by comparing measurements of in vitro drug sensitivities. A continuous cell line, RT112, derived from a moderately-well differentiated transitional cell carcinoma of the human bladder was used. Comparisons were made of the drug sensitivities of cells either left in situ or transferred following drug exposure. Using the transfer assay, cells were plated at low density, comparable to that required in the in situ assay, and at a high density, using cells in exponential growth and quiescent cells. The latter are cells that have reached the plateau phase of the growth curve (see Chapter 2,2.4 (k)). In addition, the effects of feeder cells and a 24 hour delay between drug exposure and cell transfer were studied. Due to the importance of the cell cycle in measurements of drug sensitivities these assay variables were studied using both a class II agent, methotrexate, a cell cycle phase—specific drug, and a class III agent, adriamycin, a cell cycle non­specific drug (Bruce et al, 1966).

4.2 MATERIALS AND METHODS

(A) Cell culture

Described in Chapter 2, 2.2 (A). Quiescent growth was achieved by incubating cells plated at high density (12,800/well) for 14 days at 36.5°C with media changing on days 7 and 11.

(B) Clonogenic assay

The clonogenic assay procedures compared are shown in Figure4.1. For the in situ assay 1,000 exponentially—growing cells were plated in 50 mm plastic petri dishes (Nunc) containing 5 ml medium and incubated for 48 hours before drug exposure. For the transfer assay 1,000 (low density) or 12,800 (high density) exponentially—growing cells were plated in microtest plate wells

containing 0.2 ml medium and incubated for 48 hours before drug exposure. For the treatment of quiescent cells 12,800 exponentially—growing cells were plated in microtest plate wells containing 0.2 ml medium and incubated for 14 days before drug exposure. Following a 24 hour exposure to drugs the medium containing drug was aspirated and the cells washed three times with serum free medium (5 ml for the dishes and 0.2 ml for the microtest plate wells) and once with calcium and magnesium free phosphate buffered saline (PBSA). For the in situ assay the cells were replenished with fresh medium. For the transfer assay the exponentially—growing cells were detached either immediately or 24 hours later using 0.2 ml of 0.01% trypsin in 0.003% versene and plated at concentrations yielding 100 — 200 colonies per 50 mm dish. The quiescent cells were detached immediately following drug exposure and plated into dishes as described above. The medium in the dishes was replenished after 7 days and after 14 days' incubation the colonies were fixed in methanol and stained with 10% Giemsa. Colonies consisting of 50 or more cells were counted using a binocular dissecting microscope. The mean number of colonies per drug concentration was expressed as a percentage of the control value. Data for each of the eleven assay procedures were derived from a minimum of 3 experiments.

(C) Feeder Cells

The clonogenicity of the cell line used in this study (RT112) show a linear relationship between cells plated and colonies formed, thus do not require feeder cells. However, the cloning efficiency of many continuous cell lines increases with higher cell concentrations showing a "feeder effect" (Eliason et al, 1985). The use of such cell lines in drug sensitivity assays would require the addition of feeder cells to maintain a linear relationship between density of cells and colonies formed. Therefore, to determine what effects, if any, the addition of feeder cells have on the measurement of clonogenic cell survival in the assay systems, comparisons were made using cells plated on

a layer of 3T3 cells. Monolayers of Swiss 3T3K mouse fibroblasts (Todaro and Green, 1963) were obtained by plating 2 x 10^ cells per 50 mm dish. After 10 to 14 days the medium was replenished and RT112 cells plated onto the monolayer either prior to drug exposure using the in situ assay or after drug exposure using the transfer assay. Feeder only controls were incubated in each experiment to check for the incidence of spontaneous transformation (loss of contact inhibition).

(D) Drugs

Pharmaceutical preparations of methotrexate (MTX) and adriamycin (ADR) were used in these studies (see Chapter 2,2.1 (C) for details). Each 50 mg vial of MTX contained 3.2 mg methylparaben and 0.8 mg propylparaben as preservatives. Each 50 mg vial of ADR contained 250 mg lactose. These substances were added to the controls at concentrations equivalent to those present in the highest drug concentration used.

(E) 8H—thymidine labelling

Details of cell labelling are given in Chapter 2, 2.4 (C).

4.3 RESULTS

The dose—response curves, obtained by plotting the percentage clonogenic survival against drug concentration on a logarithmic scale, are shown for MTX in Figure 4.2 A—D and for ADR in Figure 4.2 E—H. Extrapolating from these dose—response curves the ID70 values (concentration required to reduce clonogenic cell survival by 70%) following a 24 hour exposure for both MTX and ADR are shown in Table 4.1. A wide spectrum of ID70 values, approximately 1000 fold, was observed using MTX, from 2.5 x10_8M (11.4 ng ml”1) to 2.5 x 10“5M (10,215 ng ml”1). In contrast those of ADR are within a narrow range, approximately 4 fold, from 1.2 xl0“8M (8.7 ng ml-1) to 6.8 xl0_8M (39.2 ng ml”1).

In general, cells treated and left In situ were less sensitive than those transferred to new dishes after drug exposure. Within the transfer assay procedures cells were less sensitive and the ID70 values more variable at the lower plating density. In addition, there was a greater than 10 fold decrease in cell sensitivity to both MTX, Figure 4.2 D, and ADR, Figure 4,2 H, using quiescent cells compared to exponentially—growing cells,

(A) Methotrexate

(i) Comparison of in situ and transfer assay methods using cells plated at low density.

The sensitivity of cells plated at low density (1000 per well or dish) assayed by the in situ and transfer methods are illustrated in Figure 4,2 A and B, Classical exponential/plateau dose—response curves for this class of agent were not clearly defined. Instead increasing cytotoxicity was observed over a wide drug concentration range from 10“ 1 to 10”'%• Using the transfer method cells were more sensitive to MTX over the whole drug concentration range compared with those used in the in situ assay, except at the highest (10“%) and lowest (10“1%) concentrations, at which point the sensitivities were similar.

(ii) Comparison of low and high cell density using the transfer assay.

Exponential/plateau dose response curves for MTX were obtained using cells plated at high density and transferred immediately into new dishes with or without feeder cells (Figure4.2 B and C). Using the other assay methods, however, the plateau region was less sharply defined. The 1D70 values for cells plated at high density ranged from 9 xl0“®M (40.9 ng ml“ ) to 1.75 xl0“7M(79.5 ng ml 1), in contrast to the greater range of sensitivity obtained for cells plated at low density, with 1D70 values ranging from 2.5 xl0“% (11.4 ng ml“l) to 9 xl0“7M (408.6 ng ml“l).

(iii) Effect of delayed transfer.

The dose—response curves for cells used at low density (Figure 4.2 B) showed a decrease in sensitivity, as a result of the 24 hour delay, following exposure to MTX concentrations up to 10“%. For concentrations greater than 10”% the dose—response curves with and without a 24 hour delay were similar. Using high density cells (Figure 4.2 C) the effect of a 24 hour delay before transfer increased cell sensitivity to MTX at concentrations greater than 10“%. For instance, the proportion of cells surviving at 10“% were reduced from 14 to 5% (p < 0.02).However, with the addition of feeder cells the dose—response curves were similar following a 24 hour delay except at lower drug concentrations where the curves diverged. For instance, at 10“% cell survival for cells transferred immediately onto feeder cells was 75% which increased to 100% after a 24 hour delay(p < 0.001).

(iv) Effect of feeder cells.

Using the in situ assay (Figure 4.2 A) the addition of feeder cells produced a significant increase in sensitivity at concentrations greater than 10“% (e.g. p < 0.001 at 10”%).Using the transfer assay for cells plated at low density the addition of feeder cells decreased cell sensitivity at low drug concentrations and increased the sensitivity at high drug concentrations (Figure 4.2 B). However, with cells plated at high density (Figure 4.2 C) feeder cells increased cell sensitivity at both high and low drug concentrations (p < 0.001 at 10“9 and 10”%). On the other hand, sensitivity, as a result of the 24 hour delay, was only increased at high drug concentrations (p < 0.02 at 10“%).

(v) Comparison of the drug sensitivity of low and high density exponentially—growing cells with quiescent cells.

The dose—response curve obtained from low density exponentially—growing cells to MTX showed increasing cytotoxicity

throughout the concentration range with a less well-defined plateau region. The high density exponentially—growing cells, in comparison, were much less sensitive at low drug concentrations (<10”7M) with a sharply defined plateau region showingapproximately 14% cell survival at drug concentrations >10“%.Quiescent cells exposed to MTX were more resistant to MTX, with almost 100% survival at 10“"% compared to 60 and 35% cell survival for exponentially—growing cells at high and low density respectively (Figure 4.2D). In addition, the dose—response curve for quiescent cells plateaued around 24% cell survival.

(B) Adriamycin

(i) Comparison of in situ and transfer assay methods usingcells plated at low density.

The sensitivity of exponentially—growing cells plated at low density (1000 per well or dish) assayed by the in situ and transfer methods are illustrated in Figure 4.2 E and F. Both assay methods show a pronounced shoulder to the otherwise classical exponential dose—response curve for this class of agent. Compared with the transfer method, cells assayed by the in situ assay (see also Table 4.1) are less sensitive to ADR at concentrations greater than 10“% (p < 0.002 at 10“7M).

(ii) Comparison of low and high cell density using the transfer assay.

Both high and low density cells, in the presence or absence of feeder cells (Figure 4.2 F and G), show the classical exponential dose—response curve, i.e. with an expected decrease in cell survival with increasing drug concentration. However, a more pronounced shoulder was obtained using low density cells.In addition, at drug concentrations greater than 2.5 xl0“% cells at high density were more sensitive than those plated at low density (p < 0.05 at 10“7M). Comparing the drug concentrations required to achieve 70% kill for each assay variable (Table 4.1) a smaller range was obtained using high density cells.

(iii) Effect of delayed transfer*

The dose—response curves for cells at low density (Figure 4*2 F) show a small but significant increase in sensitivity as a result of the 24 hour delay in the presence or absence of feeder

- Qlayers at ADR concentrations less than 5 x 10 M (e.g. without feeder layers p < 0.001 at 10”®M, with feeder layers p < 0.05 at 10"8M). However, using high density cells (Figure 4.2 G), a 24 hour delay before cell transfer made no significant difference

Q 7throughout the drug concentration range of 10 to 10 M in the presence or absence of feeder cells.

(iv) Effect of feeder cells.

In general, the addition of feeder cells (Figure 4.2 E, F and G) reduced cell sensitivity to ADR at concentrations greater

othan 10 M. For low density cells using the in situ or transfer assay methods the addition of feeder cells had little effect on sensitivity in the shoulder region of the dose response curve. However, using cells at high density the addition of feeder cells (Figure 4.2 G) increased cell sensitivity at drug concentrations less than 10~8M (p < 0.02).

(v) Comparison of the drug sensitivity of low and high density exponentially—growing cells with quiescent cells.

There was little difference in the dose—response curves obtained for exponentially—growing cells exposed to ADR for 24 hours at low or high density. However, quiescent cells show a marked decrease in sensitivity with, for example, a greater than 1 log increase in the drug concentration required to kill 70 per cent of the cell population compared to those treated in exponential growth (Figure 4.2 D, H).

(C) % —thymidine labelling.

The tritiated thymidine labelling indices of cells plated at three densities equivalent to those used for the in situ, low and

high cell density transfer assays are shown in Figure 4.3 and Table 4.2. During the first 24 hours of exposure there were marked differences in the proportions of cells labelled. For instance, after 4 hours, the percentage labelling was 21.8, 32.6 and 41.6% for cells plated at concentrations equivalent to those used in the in situ, low and high cell density transfer assays respectively. Thus, while 35% of the cells plated at the highest density were labelled within 2 hours, 5 and 9 hours elapsed before a similar proportion of cells plated at concentrations equivalent to the low cell density and in situ assays respectively were labelled. Nevertheless, within 24 hours, at least 95% of the cells were labelled at each cell density and after 48, 72 and 96 hours continuous exposure to % —thymidine no unlabelled nuclei were observed (Table 4.2).

4.4 Discussion

This study shows that the clonogenic assay procedures adopted for monolayer cultures can significantly influence measurement of drug sensitivities in vitro. These data are particularly relevant if clonogenic assays are to be used for comparing the activity of cytotoxic drugs as has been suggested by Salmon et al (1980b) and Von Hoff et al (1981) or for evaluating new drugs (Venditti, 1983; Shoemaker et al, 1985).

Clonogenic assay procedures for monolayer cultures are of two main types: either cells are left in situ or transferred following drug exposure. The in situ assay is limiting for four reasons: Firstly, low cell numbers must be plated in order toobtain discrete colonies (100 — 200 per 50 mm dish) that can be easily scored. Thus, accurate measurement of clonogenic cell survival at drug concentrations producing greater than one log cell kill are difficult to obtain. Secondly, cells may be exposed to the drug shortly after enzyme or mechanical detachment and replating. Depending on the time elapsed between exposure to enzymes such as trypsin, commonly used for cell transfer, and

drug treatment, survival values have been shown to vary (Barranco et al, 1980). Thirdly, some cell division may occur prior to drug exposure resulting in micro "colonies" of a few cells, therefore a correction factor may need to be applied (Szechter ejt al, 1978). Fourthly, some drugs may remain bound to the culture vessel after the washing procedure resulting in additional cytotoxicity.

It could be postulated that the differences in results obtained with the in situ assay, compared with the transfer assay, might in part be a result of exposing cells to drugs in the former method at a low density. Indeed, there have been a number of reports suggesting that cell density has an effect on cytotoxicity (Korbling et al, 1982; Chambers et al, 1984; Ohnuma et al, 1986). Furthermore, model systems used to study the kinetics of drug—induced lethality show that cell density affects the efficacy of cytotoxic agents to varying degrees. Clonogenic assays using low cell densities may not therefore reflect the achievable cell kill in vivo for some drugs, where tumour cell densities are probably high (Arkin et al, 1984). Therefore, in order to investigate the effect of cell density in this study the sensitivities of cells plated at low and high density were also measured using the transfer assay. The responses of cells plated at the lower density had features in common with those obtained using the in situ assay, including a pronounced shoulder at low adriamycin concentrations and a less obvious plateau at high methotrexate concentrations. Furthermore, much greater variation was observed in drug sensitivity using cells plated at low rather than high density (see Table 4.1).

From a practical standpoint, however, the in situ assay is simpler and cheaper than the transfer assay and provides a more suitable method for measuring cell survival under continuous

exposure, such as that used by Bird et al, 1985. Nevertheless, there are inherent limitations in the use of the in situ assay mentioned above, which restrict its value for generating in vitro drug sensitivity data. However, there are also limitations of the transfer assay.

Furthermore, the clonogenic assay procedure in which cells are transferred following drug exposure can, in addition, be modified by introducing a delay between drug treatment and cell transfer. This interval may permit some clonogenic cells to recover from potentially lethal damage, before receiving the additional insults of enzyme detachment and transfer (Barranco et al, 1980) which can influence the fraction of surviving clonogenic cells (Twentyman, 1979). A linear relationship was observed between survival and the length of delay up to 48 hours following UV irradiation, and similar results were obtained following exposure to X—rays, bleomycin or nitrogen mustard, although no further increase in survival ocurred after a six hour delay (Hahn et al, 1974). Similar increases in clonogenic cell survival using delayed transfer have been observed following exposure to nitrosoureas (Barranco et al, 1975). In contrast, Hahn (1976) did not observe any increase in cell survival following delayed transfer after exposure to adriamycin, a finding confirmed in this study using RT112 cells. Similarly, a 24 hour delay had little effect on the fraction of clonogenic cells surviving methotrexate exposure, at least on the exponential part of the dose—response curve. Nevertheless, from these results and those published in the literature (Twentyman, 1979; Barranco et al, 1980), the effect of delayed transfer is not the same for all drugs and therefore the introduction of a delay makes comparison of data more difficult. For this reason there would appear to be no advantage in using delayed transfer in studies restricted to the measurement of cytotoxicity.

When colony forming efficiency is not independent of the number of cells plated, meaningful drug sensitivity data cannot be obtained. While this problem is commonly overcome with the use of feeder cells, these in turn may influence clonogenic cell

survival following drug exposure, for instance by metabolic co­operation. The RT112 cells used in this study do not show a "feeder—effect" and therefore were chosen to examine any influence the addition of feeder cells may incur. The Swiss 3T3 feeder cells used in this study appeared to have little effect on the dose—response curves of tumour cells exposed to adriamycin or methotrexate. A more pronounced shoulder was obtained with low concentrations of adriamycin and the proportion of cells resistant to high methotrexate concentrations was slightly reduced. However, changes in drug effects have been observed in the presence of feeder cells (Weizsacher and Dean, 1980), but these only occurred under sub—optimal growth conditions, in other words when maximum colony forming efficiency was not attained. Thus, it appears that under optimal conditions in these assay systems, feeder cells do not have major effects on the expression of drug—related cytotoxicity.

Marked differences in tritiated thymidine labelling indices were observed between cell densities equivalent to those used for the in situ, low and high cell density transfer assays (see Table4.2, Figure 4.3). During the first few hours exposure to % — thymidine the percentage of cells labelled was greatest for those plated at high density, i.e. the proportion of those cells that would have entered and passed through S phase at any given time would be greater. These cell kinetic data may explain why, for methotrexate (a class II drug with maximum lethal activity in G]/S phase (Hill and Baserga, 1975; Drewinko and Barlogie,1984)), a wider variation in the dose—response curves was observed between the assay procedures. On the other hand, adriamycin (grouped as a class III drug) exerts its lethal activity throughout the cell cycle, therefore only minor differences were observed between the assay procedures.

Factors other than those investigated in this study could also influence measurements of drug sensitivities in vitro. In common with many previous reports it has been shown that the proportion of cells proliferating at the time of drug exposure is

significant. In agreement with earlier reports (Drewinko et alt 1981; Barranco, 1984) results from this study show that cells in logarithmic growth were much more sensitive to adriamycin and methotrexate than quiescent cells. Other important factors which can influence in vitro drug sensitivities are the length of drug exposure times, the type of medium and serum and the substrate used (e.g. agar or plastic).

The effect of different drug exposure times to which cells in vitro have been exposed have been reported (Hill et al, 1981; Calabro—Jones et al, 1982; Rupniak et al, 1983a). From these studies it could be implied that certain drugs require prolonged incubation times to exert their cytotoxicity, since negligible cell kill occurs within a 1 hour exposure commonly used. Furthermore, long-term infusions of many drugs are increasingly being adopted because of their improved therapeutic effectiveness. Therefore, 24 hour in vitro exposure times may more accurately reflect the pharmacokinetics of drug—induced cytotoxicity of such drugs (Rupniak et al, 1983b). Some media, such as Hams F12, contain thymidine and hypoxanthine at concentrations which may influence the expression of cytotoxicity of antimetabolite drugs such as methotrexate and 5—fluorouracil (see references cited by Hakala and Rustum, 1979). The possibility of salvage precursors supporting cell growth has recently been reported by Jackman et al (1984). They showed that as little as 0.1 pM thymidine can support cell growth in the presence of an inhibitor of thymidylate synthesis, such as methotrexate, provided that the initial cell inoculum is low (1 x 10~* cells/ml).

The type, concentration and batch of serum is a further factor that can influence in vitro drug sensitivity studies (Freshney, 1983a). In addition, many antitumour agents bind to serum proteins (Hahn et al, 1974; Takashashi et al, 1980; Haspel et al, 1984), and therefore the use of a defined serum—free medium may be preferred for in vitro drug sensitivity studies. A

number of reports have shown that different serum—free hormone supplemented media can support in vitro growth of various cell types including small cell lung cancer (Carney et al, 1981), human keratinocytes (Maciag et al, 1981), human bronchial epithelial cells (Lechner et al, 1982) and human bladder (Messing et al, 1982). Thus, studies which need to eliminate complex, undefined and variable elements, which may occur in serum supplemented media, are now possible.

Finally, the clonogenicity of cells differs on plastic and agar (Marshall et al, 1977) and thus different substrates can influence in vitro studies. Perhaps of greater significance, because many drug sensitivity investigations use clonogenic assays in agar, is that the concentration and type of agar has been shown to affect the percentage of clonogenic cells (Buick and Fry, 1980; Whelan and Hill, 1981). Furthermore, it appears that the type of agar assay used can significantly affect the clonogenicity of tumour cells (Tveit et al, 1981a; Walls and Twentyman, 1985; Endresen et al, 1985).

In conclusion, this study shows that clonogenic assay procedures can influence the measurement of in vitro drug sensitivities and therefore indicates that greater standardisation, both of culture conditions and assay procedure, is essential if comparative studies are to be of value.

Figure 4,1 Clonogenic assay procedures used to compare in vitro drug sensitivities of human tumour cell lines. Quiescent growth was achieved by incubating cells plated at high density (12,800/well) for 14 days at 36.5°C with media changing on days 7 and 11,

< ♦ “ Exponentially growing cells >

1000Viable cells

1000 o r 12,800 Viable cells

o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o

Feedercells

INCUBATE a t 36.5°c for <48 h ours

DRUG EXPOSURE for 24 h o u rs

WASH ( 3x medium p lus 1x PBSA )

FRESH MEDIUM Immediate t r a n s f e r to f resh d ishes24 h r delay

laefore t r a n s fe r to f resh d ishes

• I Feeder / cells

A f te r 7 d ay s - Media chanqi

After 14 d ay s incubation Fix, s ta in , c o u n t .

IN SITU TRANSFER 24 TRANSFER

Figure 4.2 Comparison of dose response curves obtained using various procedures after a 24 hour exposure to methotrexate or adriamycin

A, E In situ assay: (•) With 3T3 feeder cells; (O) Nofeeder cells.

B, F Low density cells, (1000/well during drug exposure):(O) Immediate transfer; (□) Transfer after 24 hour delay; (#) Immediate transfer onto 3T3 feeder cells; (■) Transfer onto 3T3 feeder cells after 24 hour delay.

C, G High density cells (12,800/well during drug exposure):(0) Immediate transfer; (□) Transfer after 24 hour delay; (#) Immediate transfer onto 3T3 feeder cells; (■) Transfer onto 3T3 feeder cells after 24 hour delay.

D, H Immediate transfer assay: (O) Low density cells;(□) High density cells; (a ) Quiescent cells.

Each point represents the mean of at least 3 experiments and only standard error bars in excess of 10% are included.

Figure U .9 (continued)

%Survival

100 -i100 - |

50-50 -

I Q -10 -

5

%Survival

100 - |

50-50-

1 0 -1 0 -

Methotrexate log-|Q'1M M ethotrexate log -j q"1M

Figure 4.2 (continued)

100 -j 50 '

100 -

50 -

10 -10 -

0.5 -0.5 -

0.1 -

0.05 -0.1 -

0.05 -

0.01 0.017

Survival

100 I50 “

10 -

0.5 “

0.1 -

0.05 “

0.0179 8

100 -| 50 -

10 -

0.5 -

0.05 “

0.01 J7 6-10 9 8

Adriamycin log-jo’1M Adriamycin log-|o’1M

Table 4.1 A comparison of in vitro drug concentrations required to reduce survival by 70% (ID70), depending on the assay procedure used.

1 1 1 I METHOTREXATE 1 ADRIAMYCIN 1

1 ASSAY 1 1 1 1 PROCEDURE* 1

1D70 Logi0-1M 1 (ngml“M l

1D70 Log1(p-lM l (ngml”*) 1

1 1 1 1 1 1

CELLLOU

DENSITY 1 1 HIGH 1 1 1

CELL DENSITY 1 LOU 1 HIGH 1

1 11 1 I Transfer I l l l 1

2.5 xlO”8 (11.4)

1 1 1 9.0 xlO”8 1 1 (40.9) 1 1 1

1 1 3.5 xlO-8 1 2.75 xlO-8 1 (19.6) 1 (16.0) 1

1 11 1 l Transfer 24 1 l l 1 1

8.5 at IQ”7 (385.9)

1 1 < 9.0 xlO“8 1 1 (40.9) 1 1 1

1 1 1.19 xlO”8 1 2.68 xlO”8 1

(8.7) 1 (16.0) 1 1 1

I l i In slcu l l l 1 l

2.5 xlO“5 (10215.0)

1 1

1 1 1 1

1 1 5.5 xlO*"8 1 - 1 (31.9) 1 1

1 1l l 1 Transfer -I* 1 I 3T3 1 1 1

1.25 xlO“7 (79.5)

1 1 1 1.13 xlO”7 1 I (51.1) 1 1 1

1 1 5.25 xlO”8 1 2.0 xlO-8 1 (30.2) 1 (11.6) 1

1 11 1 I Transfer 24 1 1 + 3T3 1 1 1

9.0 xlO“7 (408.6)

1 I 1 1.75 xlO”7 I 1 (79.5) 1 1 I

1 1 4.25 xlO“8 1 3.88 xlO“8 1 (24.7) 1 (22.0) 1

1 1 1 In slcu + 3T3I 1 1 1 1

6.5 xlO"7 (295.1)

1 I

1 1 1 1

1 1 6.75 xlO”8 1 - 1

(39.2) 1 1 1 1

Celia could not be placed ac Che high densicy using Che In situ assay.

* For decalls see Figure 4.1

Table 4.2

The labelling indices of cells plated at densities of 1 xlO^, 6,1 xlO^ and 7.8 xlO^/5 cm^ dish, equivalent to the cell concentrations used for the in situ, low and high cell density transfer assays respectively, and exposed to H—thymidine continuously for periods ranging from 15 minutes to 96 hours.

1 'Exposure Co l ^HTdR l (Hours)

PERCENTAGE OF CELLS LABELLED 1

In Slcu 1 Low densicy l crsnsfer l

1 High denslcy I I transfer I 1 l

1 0.25 9.9l1 12.3 1

i l 1 20.4 1 1 1

1 2 16.51l 28.1 1

1 1 l 35.3 1 1 1

1 4 21.811 32.6 1

l 1 1 41.6 1 1 1

1 8 32.111 40.6 1

1 1 1 52.2 1 1 1

1 24 95.411 97.6 1

l 1 1 100.0 1 1 1

1 48 100.011 100.0 1

1 1 1 100.0 1 1 1

1 72 100.011 100.0 1

1 1 1 100.0 1 1 1

1 96 100.011 100.0 1

! 1 1 100.0 1 1 1

Figure 4.3 Comparison of the fraction of labelled cells plotted against time of exposure to [methyl^H] thymidine, with the cell densities equivalent to (x) high density transfer, (O) low density transfer, (a) in situ assay.

llbtiid*00!

60-

4a

24

CHAPTER 5. CYTOTOXIC ACTIVITIES OF CHEMOTHERAPEUTIC DRUGS AND THE EVALUATION OF DRUG COMBINATIONS WITH SPIROGERMANIUM USING THE HUMAN BLADDER CANCER CELL LINE, RT112

CONTENTS

PAGE

5.1 INTRODUCTION 219

5.2* MATERIALS AND METHODS 221

(A) Cell culture 221(B) Clonogenic assay 221(C) Drugs 222(D) Drug combination studies 222

5.3 RESULTS 222

5.4 DISCUSSION 222

(A) Drugs as single agents 225(B) Chemosensitivity studies using human bladder 226

cell lines(C) Comparison of in vitro findings with clinical 227

studies(D) In vitro combinations of spirogermanium with 229

four "standard” anticancer agents

5.5 SUMMARY 230

TABLES AND FIGURES 233

5.1 INTRODUCTION

There are about 8000 new patients reported annually in Britain with urothelial tumours (Tate et al, 1979). Overall, about two—thirds on first presentation have superficial papillary tumours, greater than 60% of these will have a recurrence, and 10 — 20% of these recurrences will have progressed to a higher grade (Gilbert et al, 1978). While the combined therapy of pre- operative radiotherapy and cystectomy has been shown to reduce the regional recurrence rate, there is little evidence that such a combined modality therapy alters the subsequent development of distant metastases (Whitmore et al, 1977). It Is therefore not surprising, because of the high recurrence rate of distant metastases, that the 5 year survival of patients with invasive bladder carcinoma treated with radiotherapy and surgery, alone or in combination, is between 20 — 50% (deKernion, 1977). Thus, it is generally thought that systemic chemotherapy increases the potential for cure In patients presenting with locally unresectable disease or who develop recurrent superficial or disseminated metastatic bladder cancer (Traynor and Carter,1981).

There have been a number of clinical studies using single agents for the systemic treatment of patients with metastatic bladder cancer. Drugs such as adriamycin, VM—26, bleomycin (Pavone—Macaluso, 1976), methotrexate (Turner et al, 1977) and cisplatin (Yagoda et al, 1976) are active, producing response rates ranging from 16 — 56%, although the remission times are short. The drugs that have been most thoroughly assessed clinically and shown to be active are cisplatin, adriamycin and methotrexate (Yagoda, 1980; Harker and Torti, 1983). In addition, there are many new agents and analogues of existing compounds which could be of value. However, these need to be evaluated in clinical trials which are both time consuming and costly. Therefore, a need exists for an in vitro screening system to select those agents most likely to have activity against human tumours. It has been postulated (Finlay and

Baguley, 1984) that a panel of suitable cell lines could be used for such an in vitro screen, and in addition, would also indicate the nature of variation in drug sensitivity between tumour lines of the same histological type. However, in studies measuring the potential activity of chemotherapeutic drugs against human tumours it would be preferable to use human cancer cell lines rather than those of rodent origin, particularly in view of fears that experimental animal tumours may manifest drug sensitivities which are different to those of human neoplastic cells (Drewinko et al, 1980).

Continuous cell lines derived from transitional cell carcinoma of the human bladder provide a cheap, rapid and reproducible system for assessing the cytotoxicity of chemotherapeutic drugs in vitro. Although there are numerous in vitro assays for assessing the cytotoxicity of chemotherapeutic drugs (see Chapter 1, 1.6 C), the clonogenic assay has been shown to be the most reliable and quantitative method of assessing the proliferative capability of tumour cells after exposure to drugs (Roper and Drewinko, 1976; Bhuyan et al, 1976; Hill et al, 1981; Belfield and Simmonds, 1982; Rupniak et al, 1983a).

The aim of this study was to measure the range of in vitro sensitivities for twelve cytotoxic drugs, which include agents used for the treatment of this disease as well as new experimental drugs including those already entering phase II clinical trials. The continuous cell line, RT112, a line derived from a transitional cell carcinoma of the human bladder (Hastings and Franks, 1981; Hepburn and Masters, 1983) was used for this purpose. Its clonogenic capacity on plastic has been shown to be reliably reproducible (see Chapter 3) and, in addition, on xenotransplantation into immune—suppressed animals (nude mice, nu/nu) produced tumours with a morphology similar to that of the patient's original tumour, that of a moderately well- differentiated transitional epithelium of the human bladder (for details see Chapter 3).

In phase 11/111 clinical trials responses for a new class of antitumour agent, spirogermanium, has been reported in ovarian and breast cancers and in malignant lymphoma (Schein et al, 1980; Trope et al, 1981; Falkson and Falkson, 1983). In addition, experimental laboratory studies have established that the <|rug has cytotoxic activity against a range of experimental animal tumours in vitro and in vivo (Rice et al, 1974; Mulinos and Amin, 1980; Yang and Rafla, 1983). Furthermore, activity of spirogermanium has also been demonstrated against a number of continuous cell lines derived from human tumours (Hill et al, 1982; Hill et al, 1983), although its precise mechanism of action has not been established.

Spirogermanium is an azaspirane—germanium compound (Rice et al, 1974) and phase I studies have revealed a significant lack of haematologic, hepatic or renal toxicity (Schein et al, 1980; Mulinos and Amin, 1980). Therefore, spirogermanium was evaluated in this study as a useful candidate for combination chemotherapy with standard effective antitumour drugs for bladder cancer namely, cisplatin, adriamycin, methotrexate and 5—fluorouracil.

5.2 MATERIALS AND METHODS

(A) Cell culture

The cell line RT112 was used for this series of experiments over a limited passage range of 35 — 45. Culture techniques are described in Chapter 2, 2.2 (A).

(B) Clonogenic assay

The in vitro cytotoxic activities of chemotherapeutic drugs were measured using clonogenic drug sensitivity assays, described in detail in Chapter 2, 2.4 (I).

(C) Drugs

Stock solutions and dilutions were prepared just prior to use for each experiment as described in Chapter 2, 2.1 (C). The source of each drug and the initial solvent used are listed in Table 2.8.

(D) Drug combination studies

The drug spirogermanium was assessed in combination with cisplatin, 5—fluorouracil, methotrexate and adriamycin. From the dose—response curve of spirogermanium alone a concentration associated with negligible cytotoxicity, permitting a >90% cell survival, was selected for combination studies. The clonogenic assays were performed as described in Chapter 2, 2.4 (I), with spirogermanium and one of the other antitumour drugs being added simultaneously.

5.3 RESULTS

Figure 5.1 (i—xii) shows the in vitro response of RT112 cells to the twelve anticancer agents in which clonogenic cell survival is plotted on a logarithmic scale against drug concentration on both a linear and logarithmic scale. From the dose—response curves with drug concentrations plotted on a logarithmic scale, i.e. molarity, the clonogenic cell survival of RT112 cells show a similar pattern for all twelve drugs, with an initial shoulder at low drug concentrations followed by an exponential response. When drug concentrations are plotted on a linear scale three drugs, methotrexate, 4—hydroxyanisole and VP16—213, produced dose—response curves consisting of an exponential phase followed by a plateau phase. The plateau phase indicates that a fraction of the clonogenic cells were resistant to these drugs during a 24 hour exposure, the proportion being approximately 15, 10 and 23% for methotrexate, 4—hydroxyanisole and VP16—213 respectively. In contrast, exponential dose—

—response curves were seen with 5—fluorouracil and, following a small shoulder region at low drug concentrations, with adriamycin, mitomycin C, dibromodulcitol, CBDCA and mitoxantrone.Spirogermanium and cisplatin also produced exponential dose— response curves, but the shoulder region at low drug concentrations was more pronounced. Bleomycin produced a biphasic exponential dose—response curve.

The doses required to reduce clonogenic cell survival by 50% (ID50), 70% (ID70) and 90% (ID90) were derived from these dose— response curves and are listed on a weight and molarity basis in Table 5.1. The molar concentration indicates the number of molecules required to achieve this level of cell kill and thus, drugs with a high molecular weight, such as bleomycin, will appear to be relatively more effective than in comparisons made on a gram weight basis. The clonogenic cell survival data is also represented by a bar diagram in Figure 5.2. This shows the effective molar concentrations required to reduce clonogenic cell survival from 50 — 90%, which ranges from 5 xl0“" M for mitoxantrone to approximately 1 xl0“^M for 4—hydroxyanisole.

An attempt was made to relate the drug concentrations which were cytotoxic in vitro with the circulating levels that can be achieved in patients. No definitive method exists for such comparisons because numerous assumptions have to be made concerning factors such as bioavailability and exposure period to each drug both in vivo and in vitro. One approach to the problem is to compare a specific level of clonogenic cell kill such as the ID70 to either the peak plasma levels or measurements of the area under the curve derived from plots of plasma concentrations against time (CxT) following administration of each drug in clinical studies. Published information on pharmacokinetic studies in patients concerning the drugs used in this study are summarised in Table 5.2. From this table figures for the CxT were taken and divided by the ID70 for each drug (Table 5.3).

This index provides a measure of the potential cytotoxicity of concentrations achievable in patients compared to clonogenic cell kill in vitro. The larger the figure in Table 5.3 the greater is the relative efficacy of each drug, ranging from 136 for methotrexate to 0.07 for spirogermanium.

The four anticancer agents chosen to be used in combination with spirogermanium were cisplatin, methotrexate, adriamycin and 5—fluorouracil. A concentration of 1.28 pg/ml spirogermanium was selected from the shoulder region of the dose—response curve obtained following a 24 hour drug exposure, shown in Figure 5.1 ix, as being associated with negligible cytotoxicity. Figure5.3 shows the lethal effects of combinations of spirogermanium with each of the four anticancer agents where RT112 cells were exposed for 24 hours. Enhanced cell kill was noted when spirogermanium was combined, in particular with cisplatin and to a lesser extent to 5—fluorouracil, at all concentrations tested, whilst no enhanced cell kill occurred from a combination of spirogermanium with adriamycin. In contrast, spirogermanium in combination with methotrexate resulted, consistently, in a marginal increase in cell survival, perhaps indicating antagonism.

5.4 DISCUSSION

Using a continuous cell line derived from a transitional cell cancer of the human bladder, it is shown that there is a greater than 1000—fold range in the gram weights of different chemotherapeutic drugs required to achieve the same level of clonogenic cell kill. These data, derived from the clonogenic assay described in this study, are highly reproducible and provide a rapid and inexpensive method for comparing the in vitro cytotoxicities of chemotherapeutic drugs. The cell line RT112 was chosen for this study because it is shown to produce characteristic transitional cell cancer following xenotransplantation in the nude mouse (see Chapter 3).

(A) Drugs and single agents

Comparing the drugs as single agents the shapes of the dose— response curves are similar to those described previously for other human cell lines (Drewinko et alt 1981; Hill and Whelan, 1981). In addition, the patterns conform to the classification of chemotherapeutic drugs described by Bruce et al (1966) with incubation times of up to 24 hours. Cell cycle phase—specific or class II agents kill cells during a specific phase(s) of the cycle, therefore, cells which do not pass through this phase(s) during the exposure period are spared. This results in an exponential/plateau dose—response curve, as exemplified in this study by methotrexate, 4—hydroxyanisole and VP16—213. On the other hand, cell cycle specific or class III agents kill cells throughout the cycle although most are more cytotoxic to proliferating rather than resting cells (Hill, 1978a), thereby producing an exponential dose—response curve as observed with most of the remaining drugs used here. The exception being bleomycin, whose dose—response curve can be described as a 2— component exponential curve. Such a dose—response curve may be due to the fact that although this drug binds to DNA like many other agents, on release from the bound state by lysis of affected cells, active drug can kill further surviving cells (Hahn et al, 1973). It has also been suggested that the biphasic dose—response curve is due to the presence of sensitive and resistant moieties of the cell population (Roizin—Towle and Hall, 1978). However, synchronised cell populations also show the same shape curve (Drewinko et al, 1973). A combination of all these theories may account for this particular shape curve, including that of Terasima et al (1976) who suggested that bleomycin kills cells exponentially and then induces a resistant fraction. The clinical significance of these cell kinetic differences in the effects of drugs has been reviewed (Hill and Baserga, 1975; Hill and Price, 1982).

A slightly different classificiation of chemotherapeutic drugs into five groups according to the shape of their in vitro dose—response curves has been described by Drewinko et al (1979).

These have been described in more detail in Chapter 1, 1.6 E(iii). One of these groups is termed threshold exponential, and is characterised by an initial shoulder region at low drug concentrations, which was observed in this study with a number of the agents. This shoulder region could be explained by an accumulation of sublethal damage (Drewinko et al, 1979).

(B) Chemosensitivity studies using human bladder cell lines

Some work has been published previously on the sensitivities of continuous cell lines derived from human bladder cancers to chemotherapeutic drugs. Using the cell line T24 Kato et al (1979) measured uptake of ^ C —leucine while Shrivastav and Paulson (1980), using the cell line 3176, measured the uptake of radioactive precursors of DNA, RNA and protein. In addition, clonogenic assays were used by Hagen et al (1979) to measure the in vitro sensitivities of the bladder cell lines MGH—Ul, MGH—U2 and RT4, and by Hisazumi et al (1983) using cell lines KK47,KW103 and RT4. However, the data for RT4, one of the cell lines used in both investigations, differ substantially perhaps reflecting differences in culture conditions and/or the clonogenic assay procedure (see Chapter 4). Another approach using clonogenic assay to measure anticancer drug activity was reported by Niell et al, (1985). Human bladder tumour cells were cultured using a modification of the agar assay of Hamburger and Salmon (1977). The findings demonstrated a lack of correlation between in vitro and in vivo results. The reason may be related to the inability of the tumour colony assay to reflect the true drug sensitivity of some drugs, such as cisplatin and methotrexate. Furthermore, it has been suggested that the drug sensitivity of a tumour cell line may change with prolonged serial passage in culture (Berry et al, 1975; Fuskova et al, 1975). Nevertheless, it has frequently been implied that clonogenic assay is the only reliable and quantitative method of assessing the proliferative capability of tumour cells after

exposure to cytotoxic drugs (Roper and Drewinko, 1976; Drewinko et al, 1979; Hill et al, 1981; Hill, 1983a; Rupniak et al,1983a). There are however alternative methods reported (see review, Weisenthal and Lippman, 1985) that suggest more accurate results may be obtained using a non—clonogenic in vitro chemosensitivity assay.

C. Comparison of in vitro findings with clinical studies

In order to compare the in vitro cytotoxicities of the agents used in this study with clinical findings an attempt has been made to relate drug activity in vitro (Table 5.1) with the pharmacokinetics in patients (see Table 5.2). Although there is considerable variation in the published reports of both peak plasma levels and CxT, the latter, which is the same as the area under the plasma, serum or blood drug concentration time curve (AUC) from time zero to infinity, is probably the best pharmacokinetic parameter for predicting response to an anticancer drug (Alberts et al, 1980). Measurements of CxT give a more accurate representation of drug bioavailability to tumour cells in patients than peak plasma concentrations, because the period during which measureable drug levels are circulating is taken into account. The value of CxT has been emphasised before (Broder and Carbone, 1976; Powis, 1985). Measurements of the CxT derived from clinical studies were divided by the ID70 obtained in vitro to give an index with which to compare the potential cytotoxicity of each drug in vivo (see Table 5.3). On this basis, the most effective drugs in vitro against RT112 cells are methotrexate, mitoxantrone, VP16—213, adriamycin, mitomycin C and cisplatin. Mitoxantrone is an anthracenedione with structural similarities to adriamycin but, theoretically, with less cardiotoxic potential (Smith, 1983). Experimental and early preclinical studies indicate that mitoxantrone has at least comparable antitumour activity to adriamycin (Smith, 1983) and our findings are in accord with this conclusion. Another pair of

drugs with close structural similarities, cisplatin and CBDCA which is a structural analogue of cisplatin, were also studied.On the basis of the limited pharmacokinetic data available on CBDCA it would appear to have less in vitro cytotoxicity on a gram—weight basis. However, much higher concentrations of CBDCA are achievable in patients and therefore the relative activities are comparable. As CBDCA has significantly less normal tissue toxicity than cisplatin (Calvert et al, 1982) it may therefore also be of value in the treatment of bladder cancer.

Systemic chemotherapy may be used to treat superficial as well as advanced bladder cancer. In the former case different pharmacokinetic principles apply and the most effective drugs are likely to be those excreted predominantly in the urine or metabolised to active forms by the urothelium. For instance, 90% of an intravenously administered dose of methotrexate is excreted unchanged in the urine within 24 hours (Balis et al, 1983). Cisplatin also is excreted primarily in the urine, although the fraction of dose eliminated in this manner appears to depend on the length of the infusion. Following a 6 hour infusion, 75% of a total dose of cisplatin was excreted in the urine, compared with only 40% following a 15 minute infusion (Patton et al, 1978; Belt et al, 1979). In contrast to cisplatin and methotrexate, less than 15% of adriamycin is excreted in the urine (Balis et_ al, 1983). Of particular interest is 4—hydroxyanisole, which appears to be excreted in the urine as a glucuronide, and the possibility exists that the beta—glucuronidase activity in urothelium could liberate significant amounts of the active drug (Morgan et al, Personal communication).

For a variety of reasons the value of chemotherapeutic drugs has been difficult to evaluate in patients with abnormal bladder cancer, as discussed by Yagoda (1980). Consequently there is little reliable information from phase II clinical trials with large series of patients. Proven activity for single agents in

metastatic disease was ascribed only to cisplatin, adriamycin and methotrexate (Yagoda, 1980; Harker and Torti, 1983; Oliver et al,1984), three of the most effective agents shown in these in vitro studies. The other three agents with comparable in vitro cytotoxicities, mitoxantrone, VP16—213 and mitomycin C, have as yet had limited use in advanced bladder cancer.

(D) In vitro combinations of spirogermanium with four "standard" anticancer agents

Spirogermanium is a novel synthetic compound that incorporates the element germanium into a heterocyclic ring structure (Rice et al, 1974). Antitumour activity against certain cell lines has been reported in vitro (Rice et al, 1974; Woolley et al, 1979; Mulinos and Amin, 1980; Hill et al, 1982; Hill et al, 1983). A lack of acute cumulative haematologic, renal or hepatic toxicities of spirogermanium has been shown in phase I clinical studies (Mulinos and Amin, 1980; Schein et al, 1980; Budman et al, 1981) with the limiting toxicity being neurological. Hence, the possibility of using spirogermanium with its relative absence of side-effects in the treatment of human neoplastic disease is an attractive one. A combination of the apparent absence of major side—effects from spirogermanium and the shape of the in vitro dose—response curve with a large shoulder region followed by a steep exponential curve suggested it might be a suitable agent to combine with other "standard” cytotoxic drugs used to treat patients with advanced bladder cancer, in an attempt to increase selectively tumour cell kill. The in vitro drug concentrations of the four drugs selected for combination with spirogermanium in this study approximated to 1 - 10% of peak plasma levels achievable in man (Alberts and Chen, 1980). In common with measuring the effects of single agents a 24 hour drug exposure time was employed since it has been shown that for certain drugs, including methotrexate, negligible cell kill occurs in vitro following a 1 hour exposure (Hill et al, 1980; Rupniak et al, 1983b).

In this study three patterns of survival were seen when the effects of spirogermanium were assessed in combination with four "standard" anticancer drugs:

(i) Spirogermanium in combination with cisplatin and 5 fluorouracil proved synergistic at all concentrations tested, particularly marked for cisplatin, providing an approximate 2 — 5 fold enhancement of cell kill.

(ii) Spirogermanium in combination with adriamycin produced no significant change in cell survival.

(iii) Spirogermanium in combination with methotrexate appeared slightly antagonistic at all concentrations tested, although not to any level of significance.

These results are in agreement with earlier reports using human tumour lines and murine lymphoblasts (Hill et al, 1983;Hill et al, 1984), whereby combination of spirogermanium with either cisplatin or 5 fluorouracil proved superior to the single agents alone in reducing cell survival.

5.5 SUMMARY

The most cytotoxic agents tested were mitoxantrone, mitomycin C, adriamycin, methotrexate and cisplatin. These in vitro findings are compatible with the activity of these drugs given systemically as single agents in phase II clinical trials in patients with advanced bladder cancer. However, a number of studies have shown inactivity of some drugs in vitro, particularly methotrexate (Sarosdy et al, 1982; Niell et al,1985). This inactivity may be due to artifacts of the clonogenic assay conditions employed. It has been reported as long ago as 1959 (Hakala and Taylor) that thymidine, glycine and a purine source in the medium protected cells from the growth inhibitory

effect of methotrexate. It has subsequently been demonstrated that the addition of nucleosides to the medium reverses methotrexate cytotoxicity (Pinedo and Chabner, 1977; Howell et al, 1981).

Clonogenic assays, using for example, McCoy's medium or CMRL 1066 medium, both of which contain nucleosides which can reverse the block produced by methotrexate, may therefore underpredict the activity of methotrexate in vitro. The medium used in these studies, RPMI 1640 (see Table 2.1 Chapter 2), does not contain such nucleosides as thymidine and hypoxanthine, and thus such a problem does not exist. However, it has been shown that such nucleosides may be present in undialysed foetal calf serum (Umbach et al, 1984). Furthermore, they have also identified a non—dialysable, non-nucleotide substance, found in foetal calf serum that can inactivate the inhibitory effect of methotrexate on the clonal growth of granulocyte—macrophage forming units in vitro.

Any of these potential nucleoside—rich culture conditions are therefore not optimal for evaluating tumour cell drug sensitivity to methotrexate. The foetal calf serum used in this study was not dialysed and therefore could have been a potential problem in the evaluation of methotrexate cytotoxicity on RT112 cells. However, because the in vitro results show good correlation with published clinical data it could be concluded that this problem did not exist with the particular batch of foetal calf serum used. Nevertheless, this may not necessarily be true for other batches of serum.

It is impossible to test every new anticancer agent and combination which becomes available in clinical trials of bladder cancer with statistically meaningful numbers of patients. In addition, the standard experimental screening systems have many limitations and do not usually include bladder tumours. The

current possibilities for selecting agents to test in phase II clinical trials for this disease using human cells are either human bladder tumour xenografts, stem cell assays and continuous cell lines. There are advantages and limitations to each of these models, although continuous cell lines provide the most reproducible, flexible and inexpensive system.

The results from this chapter illustrate that a human tumour cell line derived from bladder cancer can be used successfully to evaluate in vitro activity of new chemotherapeutic drugs.However, since human cancers of the same type and with equal histopathology do not respond in a uniform manner to drugs (Pihl, 1986; see also Chapter 6) an appropriate selection from the available human bladder cancer continuous cell lines, as suggested by Hisazumi et al (1983), might be used as a model system for assessing drugs for the therapy of this disease. It has recently been reported that this new "disease—oriented" screening approach is now being incorporated into the NCI anticancer drug research programme (Pihl, 1986) In which human tumour cell lines will largely replace the animal leukaemia models as a predictor of drug efficacy.

Figure 5.1 (i-xii)

Clonogenic cell survival of RT112 against a range of chemotherapeutic drugs.Each diagram shows clonogenic cell survival on a logarithmic scale against a range of concentrations of each drug plotted on a linear and logarithmic scale. The concentrations of adriamycin, mitoxantrone, mitomycin C, cisplatin, methotrexate and VP16-213 are plotted in ng/ml while the remaining drugs are plotted in yg/ml. Each point is the mean of at least three experiments, and only standard error bars in excess of 5% are included.

Abbreviations for Figures 5.1 and 5.2

(i) ADRIAMYCIN (ADR)

(ii) BLEOMYCIN (BLM)

(iii) CISPLATIN (DDP)

(iv) CBDCA

(v) DIBROMODULCITOL (DBD)

(vi) 5-FLU0R0URACIL (5-FU)

(vii) 4-HYDROXYANISOLE (AOHA)

(vi i i) METHOTREXATE (MTX)

(ix) MITOMYCIN C (MIT.C)

(x) MITOXANTRONE (MX)

(xi) SPIROGERMANIUM (Sp)

(xii) VP16-213

Drug

Cone.

Molarity

CO

COin

como

o

m

0)u3b0•H

o

o

o

r S

£

3I

£

■ in

a>u3toT-lpEH

m

xmCM

o

Drug

Cone.

Molarity

ooro

Om

cn

m<DU3bO•Hfa

O

o

o

IA

Figure 5.1

(iv)n

O

n

©

O

OXm

Oruq

Cone, uq

Ml

mo

oo o

LOQ)U3bOfe

o

NOIoX

I

o

ooo

>

<D3-i3to

©X

oXCM

oX

IJI

om

CM

0)5-13tiO•H

m

ox

CM

m©S

Mbo

OOo oirv oCSI O

a>u3bO•H

oX

o

oX

1CN s

s

n

.HH-

m

oJ-l3bO*r-tfa

OHX

MGO\o

oz inCMucou

bo3O GOIo

CM

8 S IA

Drug

Cone.

Molarity

mO

CDU3to•Hfa

ox

oX

CM

K\ 1S

Figure 5.1 (xii)

O>3“tNI

OCMOun ooo

X

©mpv

oX

©o o o o

■8

CMlD<D

.5

l—* QQ) & u in

LD

uE-tHs

CMI'sDxE-*S

faQQ£CQ

DfalOQCQQ

<U <3Q ECfa CQ o

CQ U

DRUG

CONC

ENTR

ATIO

N LOG

10

Table 5.1

The concentrations of each drug, expressed as (a) pg/ml (b) molar concentations, required to reduce clonogenic cell survival by 50% (ID50), 70% (ID70) and 90% (ID90).

1 Drug t Molecular 1 l Weight I

Reduction in I Clonogenic I

Cell Survival I

Hg ml"1 Molar i Concentration I

1 'ID50 1 0.003 5.06 xlO” 9 1

l Mitoxantrone 1 517 l ID70 1 0.003 5.66 xlO"9 1ID90 1 0.004 6.85 xlO"9 1

ID50 1 0.009 2.78 xlO"8 ll Mitomycin C 1 334 1 ID70 1 0.014 4.10 xlO"8 1

ID90 1 0.022 6.68 xlO” 8 1

ID50 1 0.011 1.93 xlO” 8 1l Adriamycin 1 580 1 ID70 1 0.016 2.81 xlO” 8 1

ID90 1 0.026 4.49 xlO” 8 1

ID50 1 0.011 2.51 xlO” 8 11 Methotrexate 1 454 1 ID70 1 0.039 8.50 xlO"8 1

ID90 1 |

ID50 1 0.155 5.17 xlO"7 11 Cisplatin I 300 1 ID70 1 0.196 6.53 xlO"7 1

ID90 1 0.251 8.37 xlO"7 1

ID50 1 0.286 4.87 xlO"7 i1 VP16-213 1 588 1 ID70 I 0.399 6.80 xlO” 7 1

ID90 1 * 6.80 xlO"7 1

Table 5.1 (continued)

1 1 1 Drug 1 1 1 1 1 l 1

MolecularWeight

Reduction In Clonogenic

Cell Survival

pg al”1 Molar 1 Concentration I

1 1 ID50 0.410 2.93 xlO"7 1I Bleoaycin I 1400 ID70 1.040 7.43 xlO"7 1i l l l

ID90 2.410 1.72 xicr* I

l 1 1 1 ID50 0.850 6.54 xlO"* 11 5-Pluorouracil 1 130 ID70 1.475 1.13 xl0"5 |1 1 1 1

ZD90 2.785 2.14 xiorS |

l 1 1 1 ID50 1.500 4.87 xlO"6 11 Dlbroaodulcitoll 308 ID70 2.150 6.98 x10T6 Il 1 1 1

ID90 3.580 1.16 xior5 |

1 1 1 1 ID50 3.480 1.02 xiorS ,1 Splrogeraanlua 1 341 ID70 3.710 1.09 xlcr-3 ,1 1 1 1

ID90 3.980 1.17 xlor-3 I

l 1 1 1 ID50 7.500 2.03 xlO”5 11 CBDCA 1 371 ID70 11.800 3.19 xl0”5 |1 1 1 1

ID90 21.100 5.70 xlO"5 1

1 1 1 1 ID50 12.000 9.68 xlO"5 11 4-Bydroxy^* 1 124 ZD70 27.000 2.18 xlO"4 11 anlsole 1 1 1

ID 90 172.000 1.39 xlO”3 1

■ A luaury of s o n of Che published data daacrlblng tha pharaacoklnetlcs In paclenca of cha drugs used In this atudy, Including cha dosaa and rouCao of administration, aaaauraaancs of cha rangas of peak plasaa concentrations, concentration x tlaa characteristics snd teralnal half-lives. The abbreviations used aret lv — intravenous) la — lntra-artarlal) la — lntraauscularj NA — not available.

1 Drug DoseRoute of

AdalnlscrationPeak Plasaa

Concentrations Hg al”1

CxTRangepg hr al”*

Teralnal 1 Half-life l

hr l

i Adrlaaycin 60 eg/M2 2—10 ain infusion

75 eg/*2 15 ain infusion

lv

lv

0.22 1.51

0.54 - 1.95

28.1 l

30.1 1

I Bleoaycin 2-10 Bg/N2 la 0.13 - 0.59 0.5 - 3.4 2.3 - 3.0 1

15 ag/N2 iv 2 - 4 5.0 4.0 1

1 CBDCA 400 ag/W2 1 hour infusion 2 0 - 3 2 0 ng/M2

24 hour infusion

iv

ivNA

146.5 5.9 1

2.8 1

I Cisplatin 100 ng/M2 24 hour infusion

iv 0.3 5.4 0.4 1

l Dibroeo— l dulcltol

15 eg/M2 15.ag/1tg

oraloral

17.51.2

5.1 - 13.4 8 l 8 1

1 5PU 460-525 eg/M2 Bolus IS ag/kg

15 ain infusion

iv

lv40.0 16.3

0.7 1

0.2 1

1 40HA 57 ag/kg Bolus

la 85.0 - 97.0 38.5 0.28 1

I Mechocrexate 30 ag/M2 100 mg/M2

iv 2.754.5

5.3 7.6 110 1

1 Mlcoayein C 10-30 ag/M2 10-20 ag/M2

iviv

0.5 - 2.7 0.4 - 2.8

0.1 - 1.1 0.2 - l.l

0.15 - 0.3 1 0.9 1

l Mitoxantrone 12 eg/M2 30 ain infusion

12 ag/M2

lv

iv

0.4 - 1.0 0.3 24 1

2 1

1 Splrogeraanlua 120 ag/M2 iv 0.60 Approx 0.26 NA I

l VP16-213 .100 ag/M2 121-214 ag/M2

oral

lv bolus

iv (0.5 - lhr) infusion

4.8 - 21.638.2 - 113.9 4.4 - 9.6 1

5.7 1

Table 5.3 Comparison of in vitro concentrations required to reduce clonogenic cell survival by 70% (ID70) with C x T values obtained from clinical studies (using the data from Table 5.2). By dividing the C x T by the ID70 a measure of the relative cytotoxicity of each drug is obtained.

1 Drug ID70 pg ml”1

C x T pg hr ml-1

C x T ID70

1 Methotrexate 0.039 5.3 136

1 Mitoxantrone 0.003 0.3 100

1 VP16-213 0.40 38.2 - 113.9 95.5 - 284.8

1 Adriamycin 0.016 0.54 - 1.95 34 - 122

1 Mitomycin C 0.014 0.1 - 1.1 7 - 7 9

1 Cisplatin 0.196 5.4 28

1 CBDCA 11.800 146.5 12.4

1 5FU 1.475 16.3 11

1 Dibromo— 1 dulcitol

2.150 5.1 - 13.4 2 - 6

1 40HA 27.000 38.5 1.4

1 Bleomycin 1.040 0.5 - 5.0 0.5 - 5.0

1 Spirogermanium 3.710 Approx 0.25 Approx 0.07

Figure 5.3

Lethal effects of a 24 hour exposure to combinations of splrogermanium (1.28 pg/ml) with cisplatin, 5—fluorouracll, adrlamycln and methotrexate on RT112 cells.

Each point represents the mean of at least three determinations + standard error.

100

9 0SP

8 0

7 0

6 0

904 0

3 0+SP

20+SP +SP

+SP10

02 2 57 9 112.9 1 9 00

CISPLATIN n g /m l

100

+SP90 SP

8070

6 0\ 1 +SP

5 0

4 0

30

\ NN20+SP

10 +SP

00 162.5 1 6 2 96 9 0 13 00

9 -n .U 0 R 0 U R A .C IL n g /m l

X CLONOGENIC CELL SURVIVAL

* CLONOGENIC CELL SURVIVAL

Flexure 5.3 (continued)

100

SP90

+SP6070

60

50

+SP40

30

20

+SP10 W14.5

0295.80

ADRIAMYCIN n g/m l

110

+SP10090 SP

80

70

60

50+SP

40

+SP+SP20

10

00 2.29 113.522.9

METHOTREXATE n g /m l

CHAPTER 6. IN VITRO DRUG SENSITIVITIES OF A PANEL OFCONTINUOUS CELL LINES DERIVED FROM HUMAN BLADDER CANCER TO ADRIAMYCIN AND METHOTREXATE s RANGE AND STABILITY IN LONG-TERM CULTURE

CONTENTS

PAGE

6.1 INTRODUCTION 255

6.2 MATERIALS AND METHODS 257(A) Cell lines 257(B) Cell culture 257(C) Drugs 258(D) Clonogenic assay 258

6.3 RESULTS 258

6.4 DISCUSSION 260

TABLES AND FIGURES 268

6.1 INTRODUCTION

Phase II clinical trials are conducted to evaluate the activity of single agents against one histopathological type of tumour in patients. The successful development of cancer chemotherapy has depended on such trials and it is now possible, for instance, to cure the majority (70 — 90%) of patients with advanced testicular germ cell tumours (Chabner, et al, 1984;Frei, 1985; Ozols and Yagoda, 1986). However, long-term disease free survival for patients with transitional cell cancer of the bladder has not yet been achieved, with most clinical studies showing less than 50% survival after 18 months (Ozols and Yagoda, 1986).

Drugs which have been clinically assessed and shown to be effective as single agents for transitional cell cancer are cisplatin, adriamycin and methotrexate (Grossman, 1979; Yagoda, 1980; Harker and Torti, 1983; Oliver et al, 1984; Edsmyr et al, 1984; Ozols and Yagoda, 1986). Many new drugs and analogues of existing ones are available for assessment. However, selection of new agents for phase II trials is largely an empirical process, because usually little direct clinical information is available other than toxicological data from phase I trials and the demonstration of anticancer activity in experimental animal systems. Even within one histopathological type of tumour there is a broad range of sensitivity to individual agents (Pihl,1986) and no reliable method of predicting response.Furthermore, phase II trials are usually conducted in patients with advanced disease which has already failed conventional forms of chemotherapy and is therefore likely to be drug resistant. Because of these factors, phase II trials are difficult to mount and for many histopathological types of tumour, including many of the genito—urinary tumours, all available agents have not been adequately tested (Ozols and Yagoda, 1986). Results from clinical trials show, on the one hand, that early use of

chemotherapy can increase disease — free survival in a number of neoplasms, including Wilms' tumour (D'Angio et al, 1982), rhabdomyosarcoma (Heyn et al, 1974), osteosarcoma (Rosen andNirenberg, 1984; Link et al, 1985), Ewing's sarcoma (Nesbit et_al, 1981), Hogkin's disease (Horning et al, 1984), ovariancarcinoma (Alberts, 1984) and breast cancer (Fisher et al, 1985;Bonadonna et al, 1985). However, for a large number of tumours, including many of the common epithelial cancers of adulthood, similar efforts to date have not yielded the same positive results (Wittes, 1986). There is thus a strong case for developing model systems for selecting the agents most likely to have activity in phase II clinical trials. Current legislation demands that new drugs are extensively tested for their cytotoxicity (Freshney, 1987). This usually involves a large number of animal experiments which are very costly and raise considerable public concern. There is, therefore, greater emphasis on developing in vitro model systems which will be more acceptable for both emotional and economic reasons. Clear differences exist between in vitro and in vivo measurements, thus, it is generally accepted that in vitro data must be interpreted in conjuction with experimental studies on drug metabolism and pharmacokinetics.

Large numbers of cell lines have been developed from many different human tumours and it has been suggested (Pihl, 1986) that a panel of continuous cell lines derived from tumours of one histopathological type might provide a model system for that disease. For transitional cell cancer of the human bladder assessment of this hypothesis has become a reality as a large number of established cell lines derived from tumours of this type have been cited in the literature (for review see Hepburn and Masters, 1983). Most cell lines retain properties of their tumour of origin, the most convincing evidence being the ability, following xenotransplantation, to generate a morphology similar to that of the parent tumour, even after many years in culture

(see Chapter 3). Each cell line must be fully characterised and it is yet to be determined how many lines are required to provide a full representation of the tumours in any one histopathological type. In this thesis (see Chapter 3) 22 cell lines derived from human urothelium have been characterised in terms of their biochemical, morphological and growth characteristics and also their reactivities with a series of monoclonal antibodies isolated after the immunisation of mice with one of these cell lines, RT112 (Trejdosiewicz et al., 1985). This chapter describes the in vitro drug sensitivities of 12 of the characterised lines to adriamycin and methotrexate, two drugs shown to be clinically effective in the treatment of advanced transitional cell carcinoma (Ozols and Yagoda, 1986). The purpose was to determine the range of in vitro drug sensitivities exhibited by cell lines derived from one histopathological type of human tumour. In addition, 4 cross—contaminated sublines of T24 were included in these studies to investigate the stability of in vitro drug response of cells maintained in long-term culture.

6.2 MATERIALS AND METHODS

(A) Cell lines

Details of the cell lines, their origin and the patient's prior therapy are given in Chapter 2, Table 2.9. Further details of the four cross—contaminated sublines of T24 are also described by O'Toole et al (1983). The biological and biochemical characteristics of all the cell lines are described in Chapter 3.

(B) Cell culture

All twelve cell lines were grown as monolayer cultures and used over a limited culture period (maximum of ten passages).Full details of culture techniques are described in Chapter 2,2.2 (A).

(C) Drugs

The chemical structure, biological activity and sources of the drugs are described in Chapter 2, 2.1 (C).

(D) Clonogenic assay

The in vitro cytotoxic activities of adriamycin and methotrexate were measured using clonogenic drug sensitivity assays, described in detail in Chapter 2, 2.4 (I).

6.3 RESULTS

The dose—response curves obtained after a 24 hour exposure to adriamycin and methotrexate produced patterns of cell survival that correspond with the kinetic classification of anti- tumour agents originally proposed by Bruce et al (1966). The principal feature of any cell cycle non-specific agent, such as adriamycin, is that these agents kill cells In all phases of the cell cycle, thus producing an exponential dose—response curve.On the other hand, cell cycle phase—specific agents, such as methotrexate, will not kill all the cells in an asynchronously growing cell population. Thus, the dose—response curves show an initial decrease in cell survival with increasing drug concentrations and then plateaus. The fraction of cells In the plateau range surviving increasing concentrations of a chemotherapeutic drug is equal to the faction of resistant cells within the whole population (Gray, 1983).

The drug concentration required to reduce clonogenic cell survival by 70% (ID70) and the percentage survival at a fixed drug concentration for each of the twelve cell lines is shown in Tables 6.1 and 6.2 for adriamycin and methotrexate respectively. The similarity between the ID70 values for T24 and the four cross—contaminated sublines can be clearly seen, particularly for

methotrexate which, overall, shows a much greater variation between the other cell lines. This is further emphasised by the bar diagrams shown in Figures 6.1 and 6.2 for adriamycin and methotrexate respectively. The effective drug concentration required to reduce clonogenic cell survival between 50 — 90% for adriamycin is very narrow (approximately 1 log drug concentration) for all twelve cell lines, ranging from 7.24 x 10“9M (4.2 ng/ml) for HU456 to 7.81 x 10”8M (45.29 ng/ml) for RT4. Conversely, the effective drug concentration required for methotrexate to reduce clonogenic cell survival from 50 — 90% is much wider (greater than 4 log drug concentration), ranging from 1.28 x 10“8M (5.8 ng/ml) for MGH—U1 to in excess of 1 x 10

(45.5 |ig/ml) for all the cell lines that plateaued at a higher clonogenic cell survival than 90%.

The dose—response curves of the eight distinct cell lines to adriamycin and methotrexate are shown in Figures 6.3 and 6.4 respectively. Clonogenic cell survival is plotted on a logarithmic scale against a range of drug concentrations on a linear scale. To avoid complication the dose—response curves for the four cross—contaminated sublines of T24, which are all very similar, have been plotted separately for adriamycin and methotrexate in Figures 6.5 and 6.6 respectively.

Following a 24 hour exposure to adriamycin all the lines showed an exponential dose—response curve following a small Initial shoulder region, with the exception of 253J, which exhibited a biphasic dose—response curve and was relatively resistant at high drug concentrations. Following a 24 hour exposure to methotrexate the dose—response curves were initially exponential and then plateaued at high drug concentrations, with the exception of HT1376, which was very resistant to this drug. The fraction of clonogenic cells for each cell line surviving after a 24 hour exposure to methotrexate are indicated by the plateau phase of the dose—response curve, ranging from 1% for T24 to 50-60% for VM—CUB—3•

Concentrating on the eight distinct cell lines, the ID70 values for adriamycin (Table 6.1) show an almost three fold range from 13.5 ng/ml for T24 to 31.7 ng/ml for RT4. The percentage cell survival at 30 ng/ml range from 2% for T24 to 34% for RT4. In contrast, for methotrexate the ID70 values show a more than fifty fold range from 9 ng/ml for T24 to greater than 450 ng/ml for HT1376 and VM—CUB—3, whereas the percentage cell survival at 100 ng/ml range from 1.2% for T24 to 96% for HT1376.

No correlation could be made between the pattern of sensitivities in vitro with the stage of the tumour from which the cell lines were established, i.e. primary, secondary or recurrent (see Table 6.3). Similarly, there was little association between the population doubling times or labelling indices and sensitivity to either drug as shown by the scatter plots shown In Figures 6.7 and 6.8 for adriamycin and methotrexate respectively. Furthermore, the rank of sensitivity exhibited by the cell lines was not identical for the two drugs, except T24, which demonstrated the fastest population doubling time and the highest labelling index and was the most sensitive line to both drugs, and RT112 which ranked third most sensitive for both drugs.

6.4 DISCUSSION

This study describes the in vitro sensitivities of twelve continuous cell lines derived from one histopathological type of tumour, transitional cell cancer of the human bladder, to adriamycin and methotrexate. The purpose, together with those studies described in Chapter 3, was to investigate the concept that a panel of cell lines derived from transitional cell cancer of the human bladder can be used as a model system for that disease. It has been proposed that cell lines might be used to screen for anticancer activity (Finlay and Baguley, 1984). This would be of great value because the development of a new

chemotherapeutic agent is a complex and expensive process, involving synthesis, formulation, measurement of anticancer activity in experimental animal systems, toxicology and eventually, in less than 0.01% of cases, clinical evaluation. Therefore, in addition to determining the value of a panel of cell lines as a model system, the potential of these lines at a later stage of drug development, prior to Phase II clinical trials in which the activity of new agents is assessed against particular histopathological types of tumour, was evaluated.

For cell lines to be used for this purpose, three criteria must be fulfilled : (a) the pattern of drug sensitivities of individual cell lines should reflect those of the tumours of origin, (b) the spectrum of sensitivity and proportion of drug- sensitive cell lines should be compatible with those encountered clinically in tumours of that histopathological type and (c) drug sensitivities should be stable in long-term culture. The following discussion will focus on these issues, with particular reference to the data from this study.

To fulfill the three criteria outlined above the first question is whether cell lines retain the pattern of drug sensitivies of their tumour of origin. The ideal test would be to compare the drug sensitivities of a cell line with the clinical response of the previously untreated parent tumour to a single agent. However, there are a number of reasons why such a comparison would be extremely difficult. Continuous cell lines can be derived from only a minority of bladder tumours and most patients with advanced disease usually receive combination chemotherapy, not single agents, often some years after the primary tumour has been resected. There are, nevertheless, a number of indirect ways in which this question can be addressed.

One way is that used by Carney et al (1983). They compared the sensitivities of three small cell lung cancer cell lines derived from newly diagnosed, untreated patients and from four lines derived from patients who relapsed following multiple— drug combination chemotherapy. Using the same drugs as those administered to the patients who had relapsed, an excellent correlation between in vitro senstivity and in vivo response was obtained. From these results they concluded that such lines may be just as useful a screening system from clinical correlations as the original tumour biopsy specimens.

Another approach has been to compare the in vitro drug sensitivities of xenografts and derived cell lines. Having shown that the in vitro chemosensitivities of five human melanoma xenografts strongly correlated with their response in vivo (Tveit et al, 1980), only small changes in chemosensitivity to six drugs were seen in seven derived lines further tested as early subcultures (Tveit et al, 1981b). In contrast to these seven lines, two additional lines did show changes In chemosensitivity. One line which had lost its tumorigenic ability had concomitantly become far more sensitive than its parent tumour. The second which also lacked ability to form tumours in nude mice had also acquired an increased chemosensitivity upon cultivation. Furthermore, a separate line derived from the same xenograft as one of these non—tumorigenic lines, did retain tumorigenicity and substantially the same pattern of drug sensitivities as the parent tumour (Tveit et al, 1981b).

A third approach to this question is to compare the in vitro drug sensitivities of cell lines derived from a histopathological type of tumour that is curable using chemotherapy, such as testicular germ cell tumours, with those derived from tumours which may respond, but are currently incurable. Such an approach has been described by Walker et al (1987) who compared the

in vitro drug sensitivities of five cell lines derived from human testicular germ cell tumours with those of five human bladder cell lines used in this study. The testicular cell lines were on average five times more sensitive to adriamycin and cisplatin than the bladder cell lines.

A similar study to the one presented in this chapter was that described by Niell et al (1985). They used human bladder cancer cell lines, selecting those that were derived from patients who had not received prior chemotherapy. The assay, a modification of that used by Hamburger and Salmon (1977), was a two layer agar assay measuring colony growth, with cloning efficiencies ranging from 0.1 to 1%. Single cell suspensions were difficult to obtain in five cell lines and the activity of anticancer agents could not be evaluated for two other lines due to inadequate clonal growth. Overall their results showed a low level of activity for most drugs, as may be expected for this histopathological type of tumour. However, they also achieved poor in vitro activity to methotrexate and cisplatin, two of the most clinically active cytotoxic drugs which they failed to explain. It is possible that their less than optimum growth conditions, resulting in poor cloning efficiencies, and the type of medium used may have some influence on their results. The low activity shown with methotrexate may be due to the use of CMRL 1066 medium In the agar assay, which contains thymidine at a concentration of 10 mg/L (Parker et al, 1957). Negligible decrease in cell survival after exposure to a range of methotrexate concentrations has previously been shown to be due to such nucleosides in the medium allowing complete rescue of the cells (Hakala and Rustum 1979; Bellamy et al, 1984). The low activity shown for cisplatinum on the other hand is more difficult to explain. It may possibly be due to solubility problems or breakdown during storage at —80°C. It has been suggested that some chemotherapeutic drugs lose activity in storage even at these low temperatures (Bosanquet 1985).

Nevertheless, all these studies suggest that by careful selection and characterisation, a panel of cell lines, derived from one histopathological type of tumour, and that reflect the drug sensitivities of their parent tumour, could be used as an in vitro model for determining anticancer drug activity in that particular disease.

From clinical and experimental data it is generally accepted that a wide range of response to cytotoxic drugs is often exhibited by tumours of the same histopathological type (Yung et al, 1982; Heppner and Miller, 1983; Bertelson et al, 1983;Skipper and Schabel, 1984). This feature is also true with the panel of continuous cell lines used in this study. None of the tumours from which the lines were established had received prior chemotherapy, yet two, HT1376 and VM—CUB—3, appear to be particularly resistant to methotrexate. Probable mechanisms of resistance to this cytotoxic agent are well documented (Bertino, et al, 1981; Curt et al, 1984; Carter, 1984; Schimke et al, 1984; Diddens et al, 1984. The second issue, therefore, is whether the spectrum of these drug sensitivities in vitro accurately reflects the range of response to individual agents seen clinically in tumours of a particular histopathological type.There are a number of studies which have been published in the literature where the efficacy of cytotoxic agents has been evaluated using a single continuous cell line derived from human tumours such as bladder (Kato et al, 1979; Hepburn et al, 1985), neuroblastomas (Hill and Whelan, 1981), acute myelogenous leukemia (Koeffler et al, 1981), lymphoma (Wu et al, 1982), cervical carcinoma (Shoemaker et al, 1983) and colon adenocarcinoma (Drewinko et al, 1983). However, such studies raise doubt as to whether the response of an individual cell line, derived from a selected subpopulation of tumour cells, will be representative of the parent tumour (see Post and Greig, 1983; Skipper and Schabel, 1984). Each cell line, as shown in these studies, has a unique pattern of sensitivity to different cytotoxic agents. For

example, HT1376, a cell line derived from a primary tumour in an untreated patient, was highly resistant to methotrexate but relatively sensitive to adriamycin. Thus, individually their representation of a single histopathological tumour type, particularly for screening cytotoxic agents, is questionable.For this reason it is therefore generally agreed that in order to obtain valid drug sensitivities in vitro that could be used to compare with in vivo findings, data from a whole panel of cell lines derived from the same histopathological type of tumour need to be examined. A number of such studies have been published.Of the urological cancers these include bladder (Hagen et al, 1979; Hisazumi et al, 1983; Borden et al, 1984; Groveman et al, 1984; Niell et al, 1985), prostate (Metcalfe et al, 1983) and testicular germ cell tumours (Walker et al, 1987). Other tumour types include those from pancreas (Chang, 1983), melanoma (Tveit et al, 1981b; Courdi et al, 1983), colon and breast (Finlay and Baguley, 1984), ovary (Hamilton et al, 1984b), small cell lung (Carney et al, 1983), lung (Tsukeda et al, 1978) and neuroblastoma (Siegel et al, 1980). Nevertheless, studies using a single or even a few cell lines still remain valuable, particuarly in relation to a) evaluating the potential of improved analogues of existing agents, b) determination of drug combinations by identifying synergy or antagonism and c) identifying the mechanism of action and drug resistance.

The third and probably most difficult question to answer is whether the biological properties of cell lines are stable in long-term culture. In particular the pattern of in vitro drug sensitivities must be shown to remain stable if cell lines are to be used as model systems for screening cytotoxic drugs. The stability of both the biological properties and chemo— sensitivities were investigated by comparing these characteristics in T24 and four sublines, which had been cultured in different laboratories separately for periods of up to 10 years. T24 and these sublines had virtually identical isozyme patterns and growth rates, but differed markedly in tumorigenicity and monoclonal antibody reactivities (see Chapter

3). In addition, sensitivities to adriamycin and methotrexate, as shown in this chapter, were very similar, suggesting that this particular property has remained stable. There is supporting evidence that in vitro sensitivities can be stable in long-term culture using the melanoma xenografts described earlier. Tveit et al (1980) observed no change In drug sensitivities between early subcultures and established lines and only minor changes in drug sensitivities following transplantation of these lines into athymic mice. Similarly, no significant change in sensitivity to 6-mercaptopurine and other synthetic compounds has been observed in the KB cell line, a line derived from HeLa, over a period of 22 years, based on data from over 2000 assays in 13 separate laboratories (Shoemaker et al, 1983). Another approach is to compare different cell lines established from the same biopsy. Two such breast cancer cell lines exhibited similar cell survival curves following exposure to x—irradiation and twelve chemotherapeutic drugs (Gioanni et al, 1985).

In contrast to these observations indicating that drug . sensitivities do not change In long-term culture, there are a number of reports which suggest variation in chemosensitivity do exist in tumours of the same histological type (Meyskens et al, 1981; Courdi et al, 1983). Further observations of heterogeneity in sensitivitity have been published, however, most of these have been made on tumour sub—populations derived by cloning primary cultures or established cell lines (Hakansson and Trope, 1974; Barranco et al, 1978 a,b; Heppner et al, 1978; Kimball and Brattain, 1980; Dexter et al, 1981; Tsuro and Fidler, 1981; Yung et al, 1982; Stephens and Peacock, 1982; Siebert et_ al, 1983; Welch and Nicolson, 1983; Talmadge et al, 1984). In addition, it has been suggested that even for established cell lines it is possible that during long-term passage in culture, drug sensitivities may drift with time (Niell et al, 1985). Such evidence that the drug sensitivities of a cell line may change with prolonged serial passage in culture has been previously reported (Fuskova et al, 1977).

Furthermore, cell lines that have been successfully established from a particular histopathological type of tumour may not necessarily be representative of that particular tumour type. Therefore, cell lines used for in vitro studies may not represent the overall sensitivities of that particular histopathological tumour to anticancer drugs. In an attempt to avoid this problem considerable care must be taken in characterising cell lines and, when used in drug sensitivity assays, optimising the culture conditions employed. Furthermore, as in all the studies presented in this thesis, the cell lines should be grown under identical culture conditions over a restricted range of passage numbers. Such precautions are essential to obtain valid comparisons of drug sensitivity between and within laboratories.

Finally, there remains the question for each drug how measurements of in vitro sensitivity relate to clinical response, an important consideration if cell lines are to be used as an in vitro screen. To assist in answering such a question the in vitro data must be analysed in conjunction with results from experimental studies of drug metabolism and pharmacokinetics. Overall, there is a major requirement for many more human tumour cell lines to be established from a wider variety of histopathological tumour biopsy specimens in order that a greater choice of well characterised cell lines are available. Nevertheless, the data presented In this thesis show that continuous cell lines derived from transitional cell carcinoma of the human bladder can retain certain biological characteristics of the tumours of origin. Thus, a selected panel of such lines can be used as a model system for that disease. In addition, if the results from this chapter are taken in conjunction with published data such a panel of cell lines could be used in a drug screening program.

Table 6.1 Comparison of the dose of adriamycin required to reduce the clonogenic cell survival by 70% (ID70) of a panel of human bladder cancer cell lines after a 24 hour exposure. Also shown is the percentage clonogenic cell survival at a fixed concentration of adriamycin.

1 Cell Line 1 Designation ngml 1

1ID70 1 1 molar 1 1 1

% Survival at 1 30 ngml”! |

1 T24 13.511 2.33 x »-* o 1 00 2.0 1

1 HT1197 14.0 1 2.41

001oT—1X 8.3 11 RT112 16.3 1 2.81 xlO”8 1 6.4 11 HT1376 17.0 1 2.93 xlO”8 1 7.8 11 VM—CUB—3 18.0 1 3.10 xlO-8 1 6.8 11 TCCSUP 21.0 1 3.62 xlO”8 1 15.0 11 253J 21.0 1 3.62 xlO”8 1 15.5 11 RT4 31.7 1 5.47

1xlO”8 1

134.0 1

1 HU456 9.011 1.55

1xlO”8 1 0.7 1

1 .MGH-U2 9.5 1 1.64 xlO-8 1 0.9 11 MGH—U1 11.5 1 1.98 xlO-8 1 1.3 11 (T24) 13.5 1 2.33 xlO”8 1 2.0 11 HU961T 13.5 1 2.33

1

001or—1X 2.6 1

Table 6.2 Comparison of the dose of methotrexate required to reduce the clonogenic cell survival by 70% (ID70) of a panel of human bladder cancer cell lines after a 24 hour exposure. Also shown is the percentage clonogenic cell survival at a fixed concentration of methotrexate and the percentage survival at the plateau region of the dose response curves.

1 Cell Line 1 Designation

111 ngml”! 1

ID70 1 molar 1

1 1 % Survival at 1 % Survival 1 100 ngml“l 1 at plateau 1

1 1

1 T2411 9.0

11 1.97 xl0“8 1.2

1 1 1 1.0 1

1 HT1197 1 158.0 1 3.47 xlO”7 32.0 1 21 - 25 11 RT112 1 38.6 1 8.50 xl0“8 22.5 1 15.01 HT1376 1>450.0 l>1.00 xlO-6 96.0 1 — 11 VM—CUB—3 1 >450.0 l>1.00 xlO-6 61.0 1 50 - 60 11 TCCSUP 1 22.0 1 4.85 xl0“8 13.0 1 10 11 253J 1 50.0 1 1.10 xlO”*7 17.0 1 2 - 6 11 RT4 1 94.9

11 2.09 1

xlO”7 29.0 1 21 - 22 1 1 1

1 HU45611 11.0

11 2.42 xlO”8 3.5

1 1 1 2.0. 1

1 MGH-U2 1 16.0 1 3.52 xlO“8 6.0 1 3.0 11 MGH-U1 1 11.0 1 2.42 xlO-8 2.1 1 1.0 11 (T24) 1 9.0 1 1.97 xlO”8 2.5 1 1.0 11 HU961T 1 9.0

11 1.97 1

xlO“8 3.8 1 2.0 1 1 1

Table 6.3 Comparison of the dose of adriamycin/methotrexate required to reduce the clonogenic cell survival by 70% (ID70) with the population doubling time (PDT) and labelling index (LI) of eight distinct human bladder cancer cell lines.

1 Cell Line 1 Designation

Stage of Disease

MTXID70

ngml“l

ADRID70

ngml”!

PDT(hrs)

1 1 1 LI 1 1 % 1 1 1 1 1

1 T24 Primary 9.0 13.5 20.91 1 1 51.1 1

1 HT1197 Recurrent 158.0 14.0 60.7 1 37.3 11 RT112 Primary 38.6 16.3 24.1 1 28.8 11 HT1376 Primary >450.0 17.0 30.7 1 32.4 11 VM—CUB—3 Primary >450.0 18.0 25.9 1 41.7 11 TCCSUP Primary 22.0 21.0 27.8 1 39.2 11 253J Secondary 50.0 21.0 28.2 1 26.8 11 RT4 Recurrent 94.9 31.7 36.8 1 12.2 1

1 1

Figure 6.1 Effective adriamycin concentrations required to reduce clonogenic cell survival from 50-90%

RT112

RT4

HT1197

HT1376

253J

VM-CUB-3

TCCSUP

T24

MGH-Ul

MGH-U2

HU456

HU961T

73

DRUG CONCENTRATION LOG 10*1 M

Figu

re

6.2

Eff

ecti

ve

eeth

otre

xate

co

ncen

trat

ions

re

quir

ed

to re

duce

cl

onog

enic

ce

ll

surv

ival

f(o

a 50

-90X

I

s

DRUG

CO

NCEN

TRAT

ION

LOG

10~

Figure 6.3 Clonogenic cell survival of a panel of eight distinct human bladder cancer continuous cell lines shown on a logarithmic scale against a range of concentrations of adriamycin plotted in ngml”! on a linear scale. Each point with standard error bars is the mean of at least three experiments.

100.0

5 0 .0

25 .0

10.0

5 .0+ TCCSUP

2.5 AHT1197

RT1120 20

ADRIAMYCIN, n g /m l

100.0

25.0

10.0

2.5HT1376

1.0

□ T24

0.1400 20

ADRIAMYCIN, n g /m l

Figure 6.4 Clonogenic cell survival of a panel of eight distinct human bladder cancer continuous cell lines shown on a logarithmic scale against a range of concentrations of methotrexate plotted in ngml” on a linear scale. Each point with standard error bars is the mean of at least three experiments.

100.0

30.0 ■ O CUBIII

23.0 -HT1197

RT112

10.0%TCCSL P

0 200 400METHOTREXATE, n g /m l

100.0

50.0 -

25.0 -

ip_ i □

10.0 -

5.0 -

2.5 -

1.0

CDICD\-BO*

+ HT15 7(6

■|RT4

T24-e—200 400

METHOTREXATE, n g /m l

Figure 6.5 Clonogenic cell survival of T24 and the four cross- contaminated sublines shown on a logarithmic scale against a range of concentrations of adriamycin plotted in ngml"" - on a linear scale. Each point with standard error bars is the mean of at least three experiments.

100.0

1.0

BT24AHU961T

0.1MGHU1MGHU:

HU156

0 20 40

ADRIAMYCIN. n g /m l

Figure 6.6 Clonogenic cell survival of T24 and the four cross- contaminated sublines shown on a logarithmic scale against a range of concentrations of methotrexate plotted in ngml“l on a linear scale. Each point with standard error bars is the mean of at least three experiments.

100.0

50.0

25.0

10.0

5.0

2.5+ MGHU

1.04002000

METHOTREXATE, n g /m l

Z/ /

Figure 6.7 Scatter plots for the eight distinct human bladder cancer continuous cell lines showing the lack of correlation between LI AND PDT against the ID70 values for adriamycin.

ADR against PDT

6 9

6 0

5 5

5 0

e «£aa. 40

3 5

3 0

2 5

2013 15 17 19 21 2 3 2 5 2 7 2 9 31

A D R .lD 70.ng/m l

ADR against L!%

5 5

5 0

♦ 5

4 0

3 5M33 0

2 5

20

15

1013 15 17 19 21 2 3 2 5 2 7 2 9 31

□ t 2 4

a V M -C U B -3□ TCCSLIP

a H T 1 1 9 7

a H T 1 3 7 6

a R T 1 1 2a 2 5 3 J

RT4 at 1---- 1---- 1---- 1---- i t — I---- r— i----1---- r—i--- 1---- 1---- 1---- 1----r

□ H T 1 1 9 7

a H T 1 3 7 6

V M -C U B -3 a RT112 a

FT4 °

2 5 3 JTOCSUP

A O R.I070.ng/m 1

Figure 6.8 Scatter plots for the eight distinct human bladder cancer continuous cell lines showing the lack of correlation between LI and PDT against the ID70 values for methotrexate.

MTX against PDT

□ O T 1197

a RT4

a a 2 5 3 JTCCSUP

a R T 112

a T 2 4

H T 1 3 7 6a

aV M -C U B -3

— —i----------- ~r -■0 200

< i400

WfD(,ID70.ng/ mi

MTX aga in st Ll%

39

50

49

40a TCCSUP

a H T 1 1 9 735

HT137630

R T 1 1 2 ° 2 5 3 J

29

20

19

a RT4

104002000

UTXj070.ng/rrU

CHAPTER 7. GENERAL DISCUSSION AND SUMMARY

CONTENTS

PAGE

7.1 INTRODUCTION 281

7.2 PRELIMINARY VALIDATION OF CELL LINES AS 282A MODEL SYSTEM FOR IN VITRO DRUG SCREEING

(A) Characterisation of cell lines 283(B) Properties of cell lines 285(C) Stability in long-term culture 287

7.3 CRITERIA CHOSEN FOR DRUG SENSITIVITY 288MEASUREMENTS IN VITRO

7.4 IN VITRO DRUG SCREENING ASSAYS 290

7.5 VARIABLES WITHIN THE CLONOGENIC ASSAY; 292INFLUENCE ON CHEMOSENSITIVITY

7.6 FUTURE DIRECTIONS 294

7.7 CONCLUDING REMARKS 297

7.1 INTRODUCTION

The aim of the studies described in this thesis was to evaluate whether a panel of cell lines derived from one histological type of human tumour, transitional cell carcinoma, could be used as an in vitro model system to predict the efficacy of anticancer agents for this disease. Two in vitro drug sensitivity studies were performed using a clonogenic assay. One of these studies (Chapter 6) assessed the range of sensitivities of a panel of cell lines to two effective cytotoxic agents used in the treatment of this disease, namely methotrexate and adriamycin. From these results an attempt was made to relate drug sensitivities in vitro with the heterogeneity of clinical response. The other study measured the response of a well differentiated transitional cell carcinoma cell line to a series of cytotoxic agents (Chapter 5). The rank order of in vitro sensitivities was compared to published clinical data in order to determine whether such a panel of cell lines could be effective in selecting the ideal agent(s) for the treatment of this disease.

Currently, there is no reliable screening system other than clinical trial for selecting drugs active against tumours of a particular histological type. Nevertheless, most cytotoxic drugs in use clinically today have been studied first in animal model systems. Therefore, there is little doubt that a reliable in vitro model system for screening drugs effective against a particular histological type of tumour would be invaluable to the clinical oncologist. The majority of the earlier in vitro studies, and some more recent ones, have used murine tumour cell lines (see Mickey, 1986). However, their use instead of human cell lines may not be a fully justified choice, since species differences are expressed even on a cellular level (Harper et al,

1974). In addition, tests comparing in vitro assays and cytotoxic drugs have shown the rank order of sensitivities to differ from those using human tumour cell lines (Hakala et al, 1974; Bellamy et al, 1984). Thus, the emphasis now for in vitro studies is to use human tumour cell lines.

It is only within the last fifteen years that cell lines derived from human tumour material have become more widely available, mainly due to improved culture techniques (Smith and Dollbaum, 1981). Thus, experimental studies, including predictive chemosensitivity testing, using panels of human tumour cell lines derived from many different histological types are now possible. The human tumour types with the largest reported number of established cell lines include bladder (Hepburn and Masters, 1983), lung (Pettengill et al, 1980; Ruckdeschel et al, 1987; Duchesne et al, 1987), head and neck (Easty et al, 1981), breast (Engel and Young, 1978), colorectal (Leibovitz et al, 1976), melanoma (Tveit et al, 1981b), sarcomas (Bruland et al, 1985) and testicular germ cell tumours (Andrews et al, 1980; Walker et al, 1987).

7.2 PRELIMINARY VALIDATION OF CELL LINES AS A MODEL SYSTEM FORIN VITRO DRUG SCREENING

In order to fulfil the aims of this thesis three important questions need to be addressed.

(A) Are the cell lines of human origin and are their characteristics consistent with an origin from transitional cell carcinoma?

(B) Do the cell lines reflect the properties of human transitional cell carcinoma encountered clinically?

(C) Do these properties remain stable in long-term culture?

(A) Characterisation of Cell Lines

There are now a large number of cell lines available from many histologically different human tumours. It is therefore possible to select a panel which would be representative of a particular tumour type. However, a factor often neglected is that such a panel of cell lines should be thoroughly characterised. The literature shows no previous attempt has been made systematically to examine, under standardised conditions, the characteristics of a large panel of cell lines derived from human transitional cell carcinoma, although there are a number of reports concerning individual or small numbers of lines (see Masters et al, 1986). This is also true of other human tumour types with the exception of small cell lung cancer (Carney et al, 1983; Ruckdeschel et al, 1987; Duchesne et al, 1987).

Characterisation of cell lines was achieved by isozyme analysis and histological examination of xenografts produced in nude mice. Data, as a result of isozyme analysis, showed that although all the 22 urothelial cell lines examined were of human origin, only 16 lines exhibited differences at one or more allele, indicating each one to be unique. Five cell lines (J82, MGH-U1, MGH-U2, HU456, HU961T) were identical to T24 at each locus examined. Although no individual enzyme phenotype is uncommon, the combination seen in T24 would be expected in less than 1% of the population (O'Toole et al, 1983). In addition,T24 was established earlier than the other five cell lines and was obtained directly from the laboratory where it was originated (Bubenik et al, 1973). The results therefore suggest that these five established lines have been cross—contaminated with T24. However, original stocks of one of these lines, namely J82, designated J82COT (see Chapter 3), established in 1972 (O'Toole et al, 1978), were shown to differ from T24. Two further cell lines, namely VM—CUB—1 and VM—CUB—3, also had identical isozyme phenotypes (but different from T24) again suggesting a cross- contamination. Such cross—contamination between established cell

lines is a common problem and is not restricted to the ubiquitous HeLa cells (Nelson—Rees et al 1981; Harris et al, 1981), therefore it would be prudent to frequently monitor the isozyme patterns of cell lines.

The approach of identifying electrophorectically distinguishable enzyme polymorphisms used in this thesis is now being superceded by the more sensitive analysis of DNA fingerprinting first described by Jeffreys and co-workers,(1985a,b). A variation of the "Jeffreys" technique utilising a locus-specific mini satellite probe (Masters et al, 1988), has since confirmed the isozyme analysis results described in this theses. Characterisation of cell lines by one or other of these two methods is important if the use of cross-contaminated cell lines is to be eliminated.

Histological examination of xenografts produced in nude mice enables one to investigate whether cell lines retain a neoplastic phenotype and have a morphology compatible with that of the tumour of origin. The histopathology of resultant xenografts successfully grown from eleven urological tumour cell lines used in these studies correlated well with that described in the original literature relating to the establishment of that particular cell line. All the cell lines could be described as consistent with an origin from transitional cell carcinomas (see Pauli et al, 1983), that is, most cells had a predominant epithelial pattern being large and polygonal in shape with well defined margins. Also nuclei were large, rounded and uniform in size.

The technique of xenotransplantation has been widely used to demonstrate that most human tumour cell lines retain a neoplastic phenotype and produce xenografts with a morpology compatible with that of the tumour of origin (Povlsen et al,1980; Sharkey and Fogh, 1984). In the largest series of cell lines studied, 127 of 162 (78%) produced tumours and in every case the histopathology correlated with that of the tumour of origin (Fogh et al, 1977b). Loss of tumorigenicity, as shown by

two of the tumour cell lines used in these studies, has also been observed in other cell lines (Fogh et al, 1977b; Tveit and Pihl, 1981). However, the failure to produce tumours in an athymic host is not a final proof of non-malignancy (Don and Kieler, 1980) and may indeed reflect changes in the immunogenicity of the cells. For example, tumorigenicity was highly variable among the five T24 sublines tested, perhaps reflecting the changes observed with regard to cell surface antigen expression demonstrated with monoclonal antibodies on these cell lines (Trejdosiewicz et al, 1985).

Further evidence as to whether the selected panel of cell lines used are representative of human transitional cell carcinoma is provided by two of the lines in particular, RT4 and RT112. With these two lines direct comparisons with histological sections from paraffin blocks of biopsies of the patient's original tumour from which they were established could be made. Several xenograft samples of both lines could be paired with the patient's biopsy specimen even when examined 'blind' by three experienced pathologists.

(B) Properties of Cell Lines

Tumours are composed of multiple clonal subpopulations of cancer cells which differ among themselves in many properties, for example, karyotype, growth rate, ability to metastasise, sensitivity to cytotoxic drugs. Such tumour heterogeneity is a well known phenomenon and well documented (Siracky, 1979; Woodruff, 1983; Skipper and Schabel, 1984; Heppner, 1984). Therefore, a panel of cell lines need to be studied if they are to prove valuable as a model for a particular histopathological type of tumour. From the studies in this thesis it is shown that cell lines can be selected which are representative of transitional cell carcinoma on the basis of isozyme analysis and xenotransplantation. The isozyme analysis identifies cross—

contamination between cell lines of the same or another species, as shown in these studies identifying the T24 sublines. In addition, if tissue from the patient from which a particular line was established is available, it can be used to confirm that the cell line was derived from that individual.

Xenotransplantation demonstrates that most human tumour cell lines retain a neoplastic phenotype and produce xenografts with a morphology compatible with that of the tumour of origin. However, although xenografts grown in immune-deprived animals largely reflect the properties of the parent tumour, kinetic differences have been observed (Steel and Peckham, 1980).

Overall the results from these studies suggest that a panel of lines selected on the basis of isozyme analysis and xenotransplantation do reflect the properties of human transitional cell carcinoma encountered clinically.Nevertheless, some characteristics, such as growth rate and clonogenicity alter as a result of in vitro culture. Therefore, such changes must be taken into account when using cell lines as a model system. However, using cell lines over a restricted passage range and standardising culture conditions may be enough to prevent such changes.

The selected panel of transitional cell carcinoma cell lines used in these studies shows a wide range of morphology, degree of histological differentation, biochemical properties, growth characteristics, sensitivities to chemotherapeutic drugs and reactivities to monoclonal antibodies, reflecting the tumour heterogeneity exhibited clinically. Similar heterogeneity has been observed in other panels of cell lines derived from human melanomas (Tveit and Pihl, 1981), lung tumours (Baylin et al, 1984) and ovarian carcinomas (Hamilton et al, 1984a).

(C) Stability of cell lines in long-term culture

In order that a panel of lines can be used for any in vitro model system it is essential that the properties of the cells remain relatively stable in long-term culture. By using the T24 sublines, which have been maintained in various laboratories under different culture conditions, separately for periods of up to 10 years, it was clear in most respects these lines exhibit stable characteristics in long-term culture. Their isozyme profiles remain the same, and their morphology, growth characteristics and in vitro drug sensitivities were virtually identical. Isozyme patterns of some of the other cell lines used in these studies have been described previously (Povey et al,1976; Wright et al, 1981; Dracopoli and Fogh, 1983a,b; O'Toole et al, 1983), thus emphasising the point that phenotypes are retained and appear to remain stable after many years in culture.

However, not all characteristics of cell lines appear to remain stable. In particular tumorigenicity and reactivity to certain monoclonal antibodies were markedly different.

In these studies T24 and its sublines all showed high cloning efficiencies on plastic ranging from 57—87% and all formed tumours in nude mice with the exception of T24. MGH— Ul(EJ) and MGH—U2 were the most tumorigenic sublines, with 100% take rates. Other studies describing a number of characteristics of T24 and MGH—Ul(EJ) show differences in their cloning efficiencies (Marshall et al, 1977; Hastings and Franks, 1983), which were lower than those presented here, while one of the reports (Marshall et al, 1977) had identical results with regard to tumorigenicity in nude mice with MGH—U1 but not T24. However, T24 did produce tumours (3/7) in the studies reported by Hastings and Franks (1983).

There are other examples in the literature of cell lines derived from human tumours that were not tumorigenic in nude mice

(Giovanella et al, 1974; Fogh et al, 1977b; Tveit and Pihl,1981). These variations may be due to the culture conditions of the cell lines in different laboratories or differences in sublines of the nude mice used. The differences may also reflect changes in the immunogenicity of cells, perhaps reflected by the changes observed in the T24 sublines in cell surface antigen expression demonstrated with the LBS series of monoclonal antibodies (Trejdosiewicz et al, 1985). These differences may be due to divergence between strains or reflect true differences in the biological behaviour of T24 and its sublines.

7.3 CRITERIA CHOSEN FOR DRUG SENSITIVITY MEASUREMENTS IN VITRO

The objective of in vitro cancer chemotherapy studies is to identify drugs which will be toxic to cancer cells, usually as single agents. However, no in vitro test, including those described in this thesis, is able to provide information on selectivity because normal cells are not tested. In addition, clinically, most cancers are treated by the use of multidrug regimens. Therefore, in vitro tests really only compare the potency of drugs. These are usually defined in terms of a therapeutic index (Conners and Phillips, 1975), that is, ranking the drugs into some order of efficacy.

Many investigators relate the in vitro drug concentrations to the peak plasma levels achievable in the patient. However, the pharmacokinetic parameter relating to drug therapeutic action and toxicity, the plasma drug concentration—time product (CxT), gives a more accurate representation of drug bioavailability to tumour cells in patients (Alberts et al, 1980). This is because the period during which measurable drug levels are circulating is taken into account. The therapeutic index used in these studies

was derived by dividing published CxT values by the ID70 value for each drug. This provided a measure of the potential cytotoxicity of concentrations achievable in patients compared to clonogenic cell kill in vitro. Using this index the rank order of activity of drugs compared well with known clinical findings relating to single agent therapy.

There is no 'standard' exposure period used in in vitro drug assays, although those most commonly employed are 1 or 24 hour or continuous exposure. Roper and Drewinko (1976) suggested that an in vitro 1 hour incubation time may be representative of many exposure times in vivo because most drugs are delivered as a bolus treatment, resulting in effective plasma levels in 30—60 minutes. Furthermore, it has been shown that within well defined experimental conditions cytotoxicity is proportional to dose (ie., concentration x period of exposure) (Walker et al, 1986). However, this may be entirely inappropriate for many drugs that have variable rates of clearance for plasma and drugs that are specific for particular cell cycle phases, such as methotrexate (Rosenblum et al, 1981). Both cell cycle non-specific and cell cycle phase—specific drugs were used in the studies presented in this thesis, thus to circumvent the problems described above, and also standardise the in vitro assays, a 24 hour drug exposure period was used.

Finally, the assays described in this thesis were run in the presence of serum. It is well known that many drugs have different half-lives and stabilities in the presence of serum compared to the absence of serum (Bosanquet, 1985; 1986).However, in the clinical situation the presence of human serum is a forgone conclusion and for the purposes of these studies it was not thought necessary to eliminate serum from the assay medium. Nevertheless, it could be argued that there is a good case for examining the action of drugs on cells in the absence of serum, thus obviating any possible drug—serum interactions

(Takahashi et al, 1980; Nissen and Tanneberger, 1981). Although the growth of cells in serum—free medium is a relatively recent phenomenon, there have been a number of reports suggesting that growth of an increasing variety of cell types in such conditions is now possible. Sato and co-workers (Hayashi and Sato, 1976; Rizzino and Sato, 1978; Bottenstein et al 1979; Mather and Sato, 1979; Barnes and Sato, 1979) have demonstrated specific (and different) hormone requirements for the growth in vitro of a variety of established cell lines in serum—free media. More recently, human transitional cell cancer cell lines (Messing et al, 1982) and small cell lung cancer cell lines (Carney et al,1981) have been successfully cultured in serum—free, biochemically defined media.

7.4 IN VITRO DRUG SCREENING ASSAYS

A variety of assays with the aim of selecting cytotoxic drugs with good antitumour activity have followed the first publication in 1954 by Black and Speer. They reported a correlation between clinical response and alteration in cellular dehydrogenase activity following in vitro drug exposures. More recently many new approaches measuring the effects of cytotoxic drugs in vitro have been employed (see Chapter 1) and have been reviewed (Tanneberger et al, 1979; Von Hoff and Weisenthal, 1980; Hamburger, 1981; Livingston et al, 1980; Weisenthal, 1981;Mattern and Volm, 1982; Hill, 1983a; Carney and Winkler, 1985; Weisenthal and Lippman, 1985). These modalities measure the effects of drugs by noting changes in cellular morphology, cellular metabolic activity, particular enzymatic inhibition, or loss of cell viability. The overall evidence from these reviews is that the clonogenic assay is currently the only reliable indicator of cellular reproductive capacity, the loss of which may be a delayed process. Thus, many other approaches may grossly overestimate or underestimate the viable cell population because they are measuring the immediate effects of metabolic dysfunction.

Nevertheless, a number of arguments against clonogenic assays have been put forward by Weisenthal and Lippman (1985). These include clump artifacts, poor cloning efficiencies, difficulties in colony counting and length of time before the assays can be evaluated. Some of these arguments may be true for cells derived from primary tissues, but are no longer valid when using continuous cell lines as described in this thesis. Primary tissue cells are, for example, usually cloned in agar, thus it is difficult to detect a >2-log tumour cell kill. Therefore, the possibility of detecting the presence of resistant tumour cells is minimal. Whereas, most established cell lines can be cloned on plastic, drug treated in high numbers and transferred at lower cell densities for colony formation to new dishes, and, therefore, by back calculation tumour cell kill of many logs is possible. Furthermore, the difficulties associated with tumour disaggregation and the preparation of single cell suspensions are no longer a problem when dealing with continuous cell lines.

Two non—clonogenic assays, whose use has been widely reported and sometimes suggested as being comparable to the clonogenic assay, are those based on precurser incorporation and dye exclusion. Most labelling experiments are performed within one generation time following drug exposure and thus are subject to a number of pitfalls. These include those relating to artifactual alterations in nucleoside intracellular pool sizes (Nakata and Bader, 1969) and salvage versus de novo dTMP synthesis pathways (Wolberg, 1972; Tisman et al, 1973).Therefore, precursor incorporation may either be falsely high (Watson and Dawson, 1977) or falsely low (Twentyman, 1978). If labelling were delayed for several generation times following drug treatment it has been suggested these artifacts may be less pronounced (Weisenthal and Lippman, 1985). Conversely, it has also been suggested that prelabelling avoids such artifacts (Roper and Drewinko, 1976). Nevertheless, ^Cr—labelled cells

are known to leak some of their label even In the absence of drug Induced damage which would also lead to Interpretation problems.

It has also been reported that serious quantitation problems exist for assays measuring the failure of cells to exclude vital dyes after drug exposure (Mickey, 1986), for example, cells destined to die may exclude dye for up to 96 hours following drug exposure. During that time, a fast—growing unaffected cell population may outgrow the affected cell population so that dye uptake is greatly underestimated.

However, the use of non-clonogenic in vitro assays should not be abandoned, since the possibility remains that a non- clonogenic assay may be more appropriate in certain circumstances. If, for example, the clonogenic survival is very low, making interpretation difficult, there may be considerable advantages in using a biochemical assay, such as isotopically labelled precursor incorporation (Silvestrini et al, 1983). Furthermore, semiautomated colourimetrlc assays based on the use of the tetrazolium salt, MTT, [3—(4,5-dimethylthiazol—2—yl)—2,5— diphenylformazan bromide] (Mossmann, 1983; Carmichael et al,1987) may be valuable as a primary screening assay to measure cytotoxic drug activity. This assay has potential as a rapid method of screening for drug responsiveness of cell lines. A number of variations of the assay have been described since the original study by Mossmann (1983) and these have been recently investigated (Twentyman and Luscombe, 1987).

7.5 VARIABLES WITHIN THE CLONOGENIC ASSAY; INFLUENCE ONCHEMOSENSITIVITY

The studies described in this thesis (see Chapter 4) have shown that the measurement of cell survival following drug exposure can be greatly influenced by the clonogenic assay method employed. Basically, cells can be left in situ or transferred to

new dishes following drug exposure. The main feature with cells left in situ is that only low numbers of cells can be used in order that discrete colonies are obtained. Two problems arise from this. Firstly, it is almost impossible to detect cell kill over more than 1—log. This feature, often overlooked, is also true of some tumour cloning assays performed in agar (Chabner,1982). This may simply be inadequate for the correlations between the in vitro testing and the in vivo situation because of the problem of resistant tumour cell subpopulations. In well characterised animal tumour systems, Skipper and Schabel (1984) have found that many drugs can produce multiple—log cell kill of stem cells, yet have a negligible effect on tumour size and no effect on the survival of the animal. Similar findings have been reported by Stephens et al (1984).. Responsive but incurable tumours appear to survive chemotherapy because of the presence of tiny populations of resistant clonogenic cells (perhaps less than one cell per 10,000) (Weisenthal and Lippman, 1985). Since it is doubtful that in situ assays will be capable of detecting much more than a 1—log tumour cell kill, there is little hope that these assays may detect the presence of these resistant tumour cells. However, if cells are transferred after drug exposure, higher cell numbers can be treated because only a proportion of these cells need to be transferred into new dishes (ie. effectively diluted). Therefore, it is relatively easy to obtaindose—response curves for tumour cell kill over several logs.

The second problem, which is not restricted to in situ assays, is that relating to the growth stage of the cell population. It is well known that when the majority of the cells within the population are in the plateau phase of growth (quiescent) they are often less sensitive to the action of chemotherapeutic agents than cells rapidly dividing in exponential growth (Drewinko et al, 1981; Barranco, 1984). Thisfact was reflected in the results of this study. Both methotrexate and adriamycin were shown to be apparently much less effective as cytotoxic drugs against a quiescent cell population

compared to those growing exponentially. However, even for exponentially growing cell populations, cell density may also have an influence on the biological action of some chemotherapeutic agents (Ohnuma et al, 1986). For example, adriamycin could be progressively inactivated by increasing cell densities. Of the two drugs used in the studies comparing clonogenic assay methods (Chapter 4), methotrexate was more susceptible to changes in assay methodology than adriamycin. It is probable that the cell kinetic effects may have had a greater bearing on the sensitivity measurements than cell density for the assay variables investigated.

Further influences were also shown possible with the addition of 3T3 feeder cells, no doubt due to some form of metabolic co-operation. However, all the bladder cell lines used in these studies did not show a feeder—effect (no density dependence, see Eliason et al, 1985) thus, this complication could be avoided.

Summarising the data on clonogenic assay methods assessed in these studies with those published in the literature it can be said that when proposing to use a panel of human tumour continuous cell lines for drug evaluations, their growth conditions should be optimised and the method of clonogenic assay standardised. Ideally, the assay conditions should be standardised, both inter and intra laboratory, in order that a meaningful direct relationship between results can be made. Unfortunately, clonogenic assays reported in the literature are performed on plastic or in various agars, using different media and varying types and concentrations of sera. Thus, direct comparisons of results are difficult to achieve.

7.6 FUTURE DIRECTIONS

Judging from recent studies published in the literature it is evident that the use of human tumour cell lines will play an increasingly dominant role in in vitro investigations. However,

whilst accepting the need to establish more cell lines from a wider variety of human neoplasms, the studies presented in this thesis suggest that methodologies used for in vitro drug sensitivity testing can be improved and, more importantly, standardisation of methods should be encouraged.

One particular area in which cell lines are becoming widely used is in screening and predicting the clinical efficacy of new anticancer agents and their analogues, for example, studies such as those of Von Hoff et al (1981) and Ahman et al (1982) have shown encouraging prospective in vitro/in vivo correlations.This type of approach is now being adopted by the NCI anticancer drug research programme. It is showing a major shift away from a panel of mouse tumour lines and a human mammary xenograft towards a new 'disease—orientated' screening approach, namely, the replacement of animal leukemia model with human tumour cell lines. Due to ethical and economic reasons pharmaceutical companies, such as ICI, are also moving towards the use of human tumour cell lines for in vitro screening. However, since human cancers with similar localisation and equal histopathology do not respond in a uniform manner to drugs, it will be necessary, as shown in this thesis, to use a panel of cell lines from each histiotype.

Other related areas of investigational studies to which human tumour cell lines are being used successfully include the elucidation of mechanism(s) of drug action and the prediction of drug resistance. Such studies as these may help to understand why there is observed heterogeneity of response to drugs amongst tumour cell populations and why some cancers are clinically more susceptible to chemotherapy than others (see Bradley et al, 1987). For some tumour types the clinical differences in response have been shown to be retained in vitro (Walker et al, 1987). They showed that on average testicular germ cell tumour lines were five times more sensitive to cisplatin and adriamycin than a number of human bladder tumour cell lines.

Identifying mechanisms of drug resistance can also be investigated by using cell lines in which resistance has been induced in vitro experimentally, such as the MCF—7 tamoxifen- resistant cell line isolated by Nawata and co—workers (1981) and the vincristine—resistant and vindesine—resistant murine lymphoblast cell line established by Hill and Whelan (1982).

More recently, the usefulness of multicellular tumour spheroids has been discussed in the literature. Spheroids are structures with an intermediate level of complexity that can be considered models of tissue micro—regions and micro-metastases (Nederman, 1983). As in solid tumours the spheroids have proliferating and quiescent regions with highly different percentages of cells in GI/GO phase in the outer, middle and inner regions and the rate of oxygen and glucose consumption may differ considerably according to distance from the surface.Thus, the three-dimensional structure of the spheroid is particularly suited for addressing problems of drug accessibility and evaluating the differential responses of hypoxic versus oxic tumour cells to chemotherapeutic drugs, particularly adriamycin and CCNU and radiation. Such work was assessed by Erlichman and Vidgen (1984) who used spheroids developed from a human bladder tumour cell line.

A much more recent and exciting area of work being investigated is that concerning cellular interactions and the role of the extracellular matrix. The majority of common carcinomas in man are of epithelial origin. The typical epithelial cell is highly polarised with respect to its free and basal surfaces (Hay, 1984), the apical cell surface being enriched with enzymes and other molecules that do not occur on the basolateral surface of the same cell (Sabatini et al, 1983). Extracellular matrix molecules seem only to interact with the basal surface and it is presumed that these interactions,

between the epithelium and the mesenchyme, are among the most important since this relationship was among the first to evolve and is the most ubiquitous of metazoan heterotypic cell—cell interactions (Reid et al, 1981). It has been established that tissue specific traits are frequently lost when culturing or performing in vitro tests using cells on plastic but these can be, to some extent, retained by growth of cells on collagen or other matrix substrata. For example, modification of cell growth, polarity and morphology has been shown for mammary, gastro intestinal tract and kidney epithelieum (Handler et al, 1980; Lee et al, 1984; Oliver et al, 1987).

7.7 CONCLUDING REMARKS

The biological characteristics of the panel of cell lines derived from human transitional cell carcinomas were shown to remain stable during long-term culture. In addition, the cell lines were also shown to retain the pattern of drug sensitivities of the tissue of origin. Thus, the findings of this thesis, in conjunction with published data, support the concept that continuous cell lines derived from one histological type of tumour provide an in vitro model system for that disease in respect of biology and drug screening.

Drug sensitivities were measured by clonogenic assay, which is arguably the best parameter of.permanent loss of reproductive integrity. However, comparison of such assays demonstrated significant differences between dose—response curves, therefore it was concluded that methods should be standardised. Using standard assay conditions the studies described show, in conjunction with pharmacokinetic data, that the panel of cell lines, as an in vitro model, is able to predict drug activity by ranking single agents in order of effectiveness, reflecting that obtained in the clinic. It is envisaged that panels of human tumour cell lines could be used for screening new agents, investigating drug resistance and cross—resistance and used for

drug targeting studies. However, the assessment of responsiveness of human tumour cell lines by any in vitro method, including clonogenic assays, has inherent limitations and it is probable that no single test will alone give all the information required. Therefore, a more flexible approach should be taken with regard to the type of in vitro assay selected for a particular purpose, together with further development and refinement of those methods.

REFERENCES

Abaza, N.A., Leighton, J. and Zajac, B.A. (1978) Clinical bladder cancer in sponge matrix tissue culture procedures for collection, cultivation and assessment of viability. Cancer, 43, No. 3, 1364-1374.

Abbassian, A. and Wallace, D.M. (1966) Intracavitary chemotherapy of diffuse non-infiltrating papillary carcinoma of the bladder. J. Urol., 96, 461-165.

Achterrath, W., Niederle, N., Raettig, R. and Hilgard, P. (1982) Etoposide-chemistry, preclinical and clinical pharmacology. Cancer Treat. Rev., 9, suppl. A, 3-13.

Adamson, R.H. (1974) Daunomycin (NSC-82151) and Adriamycin (NSC-123129): a hypothesis concerning antitumour activity and cardiotoxicity. Cancer Chemother. Rep., 58, No. 3, 293-294.

Agre, P. and Williams, T.E. (1983) The human tumor cloning assay in cancer drug development. Invest. New Drugs, 1, 33-45.

Ahman, F.R.,, Meyskens, F.L., Moon, T.E., Durie, B.G.M. and Salmon, S.E. (1982) In vitro chemosensitivities of human tumor stem cells to the phase II drug 4-(9-acridinylamino) methanesulfur-m-anisidide and prospective in vivo correlations. Cancer Res., 42, 4495-4498.

Alberts, D.S. (1984) Adjuvant therapy of ovarian cancer. In: Adjuvant Therapy of Cancer IV. Eds. S.E. Jones and S.E. Salmon, Grune and Stratton, Orlando, pp507-519.

Alberts, D.S. and Chen, H-S.G. (1980) Tabular summary of pharmacokinetic parameters relevant to in vitro drug assay. In: Cloning of Human Tumor Stem Cells. Ed. S.E. Salmon. Alan R. Liss Inc., New York, pp351-359.

Alberts, D.S., Chang, S.Y., Chen, H-S.G., Moon, T.E., Evans, T.L., Furner, R.L., Himmelstein, K. and Gross, J.F. (1979) Kinetics of intravenous melphalan. Clin. Pharmacol. Ther., 26, No. 1, 73-80.

Alberts, D.S., Chen, H-S.G. and Salmon, S.E. (1980) In vitro drug assay: Pharmacologic considerations. In: Cloning of Human Tumor Stem Cells. Ed. S.E. Salmon. Alan Liss, New York., ppl97-207.

Alberts, D.S., Salmon, S.E., Chen, H-S.G., Moon, T.E., Young, L. and Surwit, E.A. (1981a) Pharmacological studies of anticancer drugs with the human tumor stem cell assay. Cancer Chemother. Pharmacol., 6, 253-264.

Alberts, D.S., Chen, H-S.G., Salmon, S.E., Surwit, E.A., Young, L., Moon, T.E. and Meyskens, F.L. (1981b) Chemotherapy of ovarian cancer directed by the human tumor stem cell assay. Cancer Chemother. Pharmacol., 6, 279-285.

Amatruda, T.T., Mulrow, P.J., Gallagher, J.C. and Sawyer, W.H. (1963) Carcinoma of the lung with inappropriate antidiuresis. Demonstration of antidiuretic-hormone-like activity in tumor extracts. N. Engl. J. Med., 269, 544-549.

Ambrose, E.J. and Easty, D.M. (1963) The use of human cells in tissue culture as a method of cancer chemotherapeutic screening.In: Cancer Progress. Ed. R.W. Raven. Butterworths and Co (publishers) Ltd., London, ppl90-202.

Ambrose, E.J., Andrews, R.D., Easty, D.M., Field, E.O. and Wylie, J.A.H. (1962) Drug assays on cultures of human tumour biopsies.The Lancet, 1, 24-25.

Andrews, P.W., Bronson, D.L., Benham. F., Strickland, S. and Knowles, B.B. (1980) A comparative study of eight cell lines derived from human testicular teratocarcinoraa. Int. J. Cancer, 26, 269-280.

Arkin, H., Ohnuma, T., Holland, J.F. and Gailani, S.D. (1984) Effects of cell density on drug-induced cell kill kinetics in vitro (inoculum effect). Proc. Am. Assoc. Cancer Res., 25, 315.

Arlett, C.F, and Harcourt, S.A. (1980) Survey of radiosensitivity in a variety of human cell strains. Cancer Res., 40, 926-932.

Arnold, A.M. and Whitehouse, J.M.A. (1981) Etoposide: A new anti-cancer agent. The Lancet, 2, 912-915.

Atkin, N.B. (1976) Cytogenetic aspects of malignant transformation. Exp. Biol. Med., 6, 1-171.

Baguley, B.C. and Wilson, W.R. (1987) Comparison of in vivo and in vitro drug sensitivities of Lewis lung carcinoma and P388 leukaemia to analogues of amsacrine. Eur. J. Cancer Clin. Oncol., 23, No. 6, 607-613.

Baker, J.R. (1944) The structure and chemical composition of the Golgi element. Q.J. Microscop. Sci., 85, part I, 2-71.

Balis, F.M., Holcenberg, J.S. and Bleyer, W.A. (1983) Clinical pharmacokinetics of commonly used anticancer drugs. Clin. Pharmacokinet., 8, 202-232.

Barfod, N.M. (1977) Flow microfluorometric estimation of G1 and G2 chalone inhibition of the JB-1 tumour cell cycle in vitro. Exp.Cell Res., 110, 225-236.

Barnes, D. and Sato, G. (1979) Growth of a human mammary tumour cell line in a serum-free medium. Nature, 281, 388-389.

Barranco, S.C. (1984) Cellular and molecular effects of adriamycin on dividing and non dividing cells. Pharmacol. Ther., 24, 303-319.

Barranco, S.C., Novak, J.K. and Humphrey, R.M. (1973) Response of mammalian cells following treatment with bleomycin and 1,3-Bis (2-chlorethyl)-l-nitrosourea during plateau phase. Cancer Res.,33, 691-694.

Barranco, S.C., Novak, J.K. and Humphrey, R.M. (1975) Studies on recovery from chemically induced damage in mammalian cells. Cancer Res., 35, 1194-1204.

Barranco, S.C., Haenelt, B.R. and Gee, E.L. (1978a) Differential sensitivities of five rat hepatoma cell lines to anti-cancer drugs. Cancer Res., 38, 656-660.

Barranco, S.C., Flournoy, D.R., Bolton, W.E. and Oka, M.S. (1978b) Cell killing and kinetic effects of diglycolaldehyde on dividing and non- dividing mammalian cells in vitro. J. Natl. Cancer Inst., 61, No.5, 1307-1310.

Barranco, S.C., Bolton, W.E. and Novak, J.K. (1980)Time-dependant changes in drug sensitivity expressed by mammalian cells after exposure to trypsin. J. Natl. Cancer Inst., 64, No.4, 913-916.

Baserga, R. (1968) Biochemistry of the cell cycle: a review. Cell Tissue Kinet., 1, 167-191.

Bates, S.E. and Longo, D.L. (1985) Tumor markers: value and limitations in the management of cancer patients. Cancer Treat. Rev., 12, 163-1207.

Baylin, S.B., Jackson, R.H., Lemnarz, W. and Shaper, J.H. (1984)The spectrum of human lung cancer cells in culture: a potential model for studying molecular determinants of tumor progression and metastasis. In: Cancer Invasion and Metastasis: Biologic and Therapeutic Aspects, Ed. G.L. Nicholson. New York, Raven Press, pp281-291.

Belfield, A. and Simmonds, A.P. (1982) Assessment of action of cis-diamminodichloro-platinum II on HeLa cells by different methods detects time-dependance. Br. J. Cancer, 46, 990-993.

Belej, M.A., Troetel, W.M., Weiss, A.J., Stambaugh, J.E. and Manthei, R.W. (1972) The adsorption and metabolism of dibromodulcitol in patients with advanced cancer. Clin. Pharmacol. Ther., 13, No.4, 563-572.

Bellamy, A.S., Whelan, R.D.H. and Hill, B.T. (1984) Studies of variation in inherent sensitivities to radiation, 5-fluorouracil and methotrexate in a series of human and murine tumor cell lines in vitro. Int. J. Radiat. Oncol. Biol. Phys., 10, 87-93.

Belt, R.J., Himmelstein, K.J., Patton, T.F., Bannister, S.J., Sternson, L.A. and Repta, A.J. (1979) Pharmacokinetics of non-protein-bound platinum species following administration of cis-dichlorodiammine platinum II. Cancer Treat. Rep., 3,1515-1521.

Benson, R.C., Tomera, K.M. and Kelalis, P.P. (1983) Transitional cell carcinoma of the bladder in children and adolescents. J.Urol., 130, 54-55.

Bentley, R. (1861) New American Remedies. I. Podophyllum pettatum. Pharmacol. J. Trans., 3, 456-464.

Benvenuto, J.A., Anderson, R.W., Kerkof, K., Smith, R.G. and Loo, T.L. (1981) Stability and compatibility of antitumor agents in glass and plastic containers. Am. J. Hosp. Pharm., 38, 1914-1918.

Berry, R.J., Laing, A.H, and Wells, J. (1975) Fresh explant culture of human tumours in vitro and the assessment of sensitivity to cytotoxic chemotherapy. Br. J. Cancer, 31, 218-227.

Bertelson, C.A., Sondak, V.K., Mann, B.D. Korn, E.L. and Kern, D.H.(1983) Chemosensitivity testing of human solid tumors. Cancer,53, 1240-1245.

Bertino, J.R., Dolnick, B.J., Berenson, R.J., Schear, D.J. and Kamen, B.A. (1981) Cellular mechanisms of resistance to methotrexate. In: Molecular Action and Targets for Cancer Chemotherapeutic Agents. Eds. A.C. Sartorell, J.S. Lazo and J.R. Bertino. Academic Press Inc., pp385-397.

Bertino, J.R., Kowal, C.D., Klein, M.E., Dombrowski, J. and Mini,E. (1984) The potential for chemotherapy and hypertension.Front. Radiat. Ther. Oncol., 18, 162-170.

Bertoncello, I., Bradley, T.R., Campbell, J.J., Day, A.J.,McDonald, I.A., McLeish, G.R., Quinn, M.A., Rome, R. and Hodgson,G.S. (1982) Limitations of the clonal agar assay for the assessment of primary human ovarian tumour biopsies. Br. J.Cancer, 45, 803-811.

Bhuyan, B.K., Loughman, B.E., Fraser, T.J. and Day, K.J. (1976) Comparison of different methods of determining cell viability after exposure to cytotoxic compounds. Exp. Cell. Res., 97, 275-280.

Bickis, I.J., Henderson, I.W.D. and Quastel, J.H. (1966) Biochemical studies of human tumors II: In vitro estimation ofindividual tumor sensitivity to anticancer agents. Cancer, 19, 103-113.

Bird, M.C., Bosanquet, A.G. and Gilby, E.D. (1985) In vitro determination of tumour chemosensitivity in haematological malignancies. Hematol. Oncol., 3, 1-10.

Black, M.M. and Spear, F.D. (1954) Further observations on the effects of cancer chemotherapeutic agents on the in vitro dehydrogenase activity of cancer tissue. J. Natl. Cancer Inst., 14, No. 5, 1147-1158.

Bleyer, W.A. (1977) Methotrexate: Clinical pharmacology, current status and therapeutic guidelines. Cancer Treat. Rev., 4, 87-101.

Bleyer, W.A. (1978) The clinical pharmacology of methotrexate. Cancer, 41, 36-51.

Blum, R.H. and Carter, S.K. (1974) Adriamycin: A new anticancer drug with significant clinical activity. Ann. Intern. Med., 80, 249-259.

Bogden, A.E., Kelton, D.E., Cobb, W.R. and Esber, H.J. (1978) A rapid screening method for testing chemotherapeutic agents against human tumor xenografts. In: Proceedings of the Symposium on the Use of Athymic Nude Mice in Cancer Research. Gustav Fischer, pp231-250.

Bonadonna, G., Valagussa, P., Rossi, A., Tancini, G., Brambilla,C., Zambetti, M. and Veronesi, U. (1985) Ten-year experience with CMF-based adjuvant chemotherapy in resectable breast cancer.Breast Cancer Res. Treat., 5, No.2, 95-115.

Bosanquet, A.G. (1985) Stability of solutions of antineoplastic agents during preparation and storage for in vitro assays. I General considerations, the nitrosoureas and alkylating agents. Cancer Chemother. Pharm., 14, 83-95.

Bosanquet, A.G. (1986) Stability of solutions of antineoplastic agents during preparation and storage for in vitro assays. II Assay methods, adriamycin and the other antitumour antibiotics. Cancer Chemother. Pharmacol., 17, 1-10.

Bosmann, H.B. (1971) Mechanism of cellular drug resistance.Nature, 233, 566-569.

Bottenstein, J.E., Sato, G.H. and Mather, J.P. (1979) Growth of neurothelial-derived cell lines in serum-free, hormone-supplemented media. In: Hormones and Cell Culture, Eds. G.H. Sato and R. Ross. Book B Cold Spring Harbor Conference, Cell Proliferation Vol. 6, pp531-544.

Bracken, R.B. and Johnson, D. (1979) Treatment of multiple superficial tumors of the urinary bladder with intravesical mitomycin C. In: Mitomycin C: Current Status and New Developments. Eds. S.K. Carter and S.T. Crocxke. Academic Press New York, pp205-211.

Bradley, G., Juranka, P.F. and Ling, V. (1988) Mechanism of multidrug resistance. Biochimica et Biophysica Acta, 948, 87-128.

Bridges, J.W. (1980) An evaluation of the present status and potential application of in vitro toxicity tests. In: Towards a Better Safety of Drugs and Pharmaceutical Products. Ed. D.D.Breimer. North - Holland, Elsevier, pp283-299.

Broder, L.E. and Carbone, P.P (1976) Pharmacokinetic considerations in the design of optimal therapeutic regimens for the treatment of breast cancer. A conceptual approach. Med. Pediatr. Oncol., 2, 11-27.

Brodeur, G.M., Green, A.A., Hayes, F.A., Williams, K.J., Williams,D.L. and Tsiatis, A.A. (1981) Cytogenic features of human neuroblastomas and cell lines. Cancer Res., 41, 4678-4686.

Brooks, R.F. (1975) The kinetics of serum-induced initiation of DNA synthesis in BHK21/C13 cells, and the influence of exogenous adenosine. J. Cell Physiol. 86, 369-378.

Brooks, R.F., Riddle, P.N., Richmond, F.W. and Marsden, J. (1983) The G1 distribution of "Gl-less" V79 Chinese hamster cells. Exp. Cell Res., 148, 127-142.

Bruce, W.R. and Valeriote, F.A. (1968) Normal and malignant stem cells and chemotherapy. In: The Proliferation and Spread of Neoplastic Cells. Vol. 21. Published for M.D. Anderson Hospital and Tumor Institute at Houston (by Williams and Wilkins Co., Baltimore) pp409-422.

Bruce, W.R., Meeker, B.E. and Valeriote, F.A. (1966) Comparison of the sensitivity of normal hematopoietic and transplanted lymphoma colony-forming cells to chemotherapeutic agents administered in vivo. J. Natl. Cancer Inst., 37, 233-245.

Bruland, 0., Fodstad, 0. and Pihl, A. (1985) The use of multicellular spheroids in establishing human sarcoma cell lines in vitro. Int. J. Cancer, 35, No. 6, 793-798.

Bubenik, J., Baresova, M., Viklicky, V., Jakoubkova, J., Sainerova,H. and Donner, J. (1973) Established cell line of urinary bladder carcinoma (T24) containing tumour-specific antigen. Int. J.Cancer, 11, 765-773.

Budman, D.R., Schulman, P., Vincigerra, V. and Degman, T. (1981) A phase I trial of spirogermanium given by infusion on a multiple dosing schedule. Proc. Am. Soc. Clin. Oncol., 22, 362.

Buick, R.N. and Fry, S.E. (1980) A comparison of human tumour-cell clonogenicity in methylcellulose and agar culture. Br. J. Cancer, 42, 933-936.

Buick, R.N. and Pollack, M.N. (1984) Perspectives on clonogenic tumor cells, stem cells and oncogenes. Cancer Res., 44, 4909-4918.

Buick, R.N., Stanisic, T.H., Fry, S.E., Salmon, S.E., Trent, J.M. and Krasovich, P. (1979) Development of an agar-methyl cellulose clonogenic assay for cells in transitional cell carcinoma of the human bladder. Cancer Res., 39, 5051-5056.

Buick, R.N., Fry, S.E. and Salmon, S.E. (1980) Effect of host-cell interactions on clonogenic carcinoma cells in human malignant effusions. Br. J. Cancer, 41, 695-704.

Burn, J.I., Johnston, I.D.A., Davies, A.J. and Sellwood, R.A.(1966) Cancer chemotherapy by continuous intra-arterial infusion of methotrexate. Br. J. Surg., 53, No. 4, 329-336.

Burrows, M.T., Burns, J.E. and Suzuki, Y. (1971) Studies on the growth of cells: the cultivation of bladder and prostatic tumors outside the body. J. Urol., 1, 3-15.

Calabro-Jones, P.M., Byfield, J.E., Ward, J.F. and Sharp, T.R. (1982) Time-dose relationships for 5-fluorouracil cytotoxicity against human epithelial cancer cells in vitro. Cancer Res., 42, 4413-4420.

V

Calvert, A.H., Harland, S.J., Newell, O.R., Siddik, Z.H., Jones,A.C., McElwain, T.J., Raju, S., Wiltshaw, E., Smith, I.E., Baker, J.M., Peckham, M.J. and Harrop, K.R. (1982) Early clinical studies with cis-diammine-1,1-cyclobutane dicarboxylate platinum II.Cancer Chemother. Pharmacol., 9, 140-147.

Calvo, F., Carney, D.N., Brower, M. and Minna, J.D. (1983) Hormone supplemented media for cloning human breast cancer: Increasedcolony formation with alteration of chemosensitivity. Br. J.Cancer, 48, 683-688.

Canagaratna, M.C.P. and Riley, P.A. (1975) The pattern of density dependant growth inhibition in murine fibroblasts. J. Cell Physiol., 85, 271-282.

Carmichael, J., DeGraff, W.G., Gazdar, A.F., Minna, J.D. and Mitchell, J.B. (1987) Evaluation of a tetrazolium-based semiautomated colorimetric assay: assessment of radiosensitivity. Cancer Res., 47, 943-946.

Carney, D.N. and Winkler, C.F. (1985) In vitro assays ofchemotherapeutic sensitivity. In: Important Advances in Oncology. Eds. V.T. De Vita, S. Heilman and S.A. Rosenberg. J.B. Lippincott Co. pp78-103.

Carney, D.N., Gazdar, A.F. and Minna, J.D. (1980) Positive correlation between histological tumor involvement and generation of tumor cell colonies in agarose in specimens taken directly from patients with small-cell carcinoma of the lung. Cancer Res., 40, 1820-1823.

Carney, D.N., Bunn, P.A., Gazdar, A.F., Pagan, J.A. and Minna, J.D.(1981) Selective growth in serum-free, hormone-supplemented medium of tumor cells obtained by biopsy from patients with small cellcarcinoma of the lung. Proc. Natl. Acad. Sci., 78, No. 5,3185-3189.

Carney, D.N., Gazdar, A.F. and Minna, J.D. (1982) In vitro chemosensitivity of clinical specimens and established cell lines of small cell lung cancer. Proc. Am. Soc. Clin. Oncol., 1, 10.

Carney, D.N., Mitchell, J.B. and Kinsella, T.J. (1983) In vitro radiation and chemotherapy sensitivity of established cell lines in human small cell lung cancer and its larger cell morphological variants. Cancer Res., 43, 2806-2811.

Carney, P.G. and Malmgren, R.A. (1967) Comparison of techniques for obtaining single cell suspensions from tumors.Transplantation, 5, No. 3, 455-458.

Carrel, A. and Burrows, M.T. (1911) Cultivation in vitro of malignant tumors. J. Exp. Med., 13, 571-575.

Carter, S.K. (1984) Some thoughts on resistance to cancer chemotherapy. Cancer Treat. Rev., 11, Suppl. A, 3-7.

Carter, S.K. and Soper, W.T. (1974) Integration of chemotherapy into combined modality treatment of solid tumors. 1. The overall strategy. Cancer Treat. Rev., 1, 1-13.

Castor, L.N. (1971) Control of division by cell contact and serum concentration in cultures of 3T3 cells. Exp. Cell Res., 68, 17-24.

Ceccarini, C. and Eagle, H. (1971) pH as a determinant of cellular growth and contact inhibition. Proc. Natl. Acad. Sci. USA, 68, No. 1, 229-233.

Chabner, B. (1982) Clinical pharmacokinetics and drug monitoring. In: Pharmacologic Principles of Cancer Treatment. Eds. B.A.Chabner and W.B. Saunders. Philadelphia. ppl00-108.

Chabner, B.A., Fine, R.L., Allegra, C.J., Yeh, G.W. and Curt, G.A. (1984) Cancer chemotherapy, progress and expectations, 1984. Cancer, 54, 2599-2608.

Chambers, S.H., Bleehan, N.M. and Watson, J.V. (1984) Effect of cell density on intracellular adriamycin concentration and cytotoxicity in exponential and plateau phase EMT6 cells. Br. J. Cancer, 49, 301-306.

Chang, B.K. (1983) Differential sensitivity of pancreatic adenocarcinoma cell lines to chemotherapeutic agents in culture. Cancer Treat. Rep., 67, 355-361.

Chlapovski, F.J. and Haynes, L. (1979) The growth and differentiation of transitional epithelium in vitro. J. Cell. Biol., 83, 605-614.

Cobb, J.P., Walker, D.G. and Wright, J.C. (1961) Comparative chemotherapy studies on primary short-term cultures of human normal, benign and malignant tumor tissues - A five year study. Cancer Res., 21, 583-590.

Cohen, B.E., Egorin, M.J., Nayar, M.S.B. and Gutierrez, P.L. (1984) Effects of pH and temperature on the stability and decomposition of N,N',NH- Triethylenethiophosphoramide in urine and buffer. Cancer Res., 44, 4312-4316.

Colvin, M., Padgett, C.A. and Fenselau, C. (1973) A biologically active metabolite of cyclophosphamide. Cancer Res., 33, 915-918.

Connors, T.A. and Phillips, B.J. (1975) Screening for anti-cancer agents, the relative merits of in vitro and in vivo techniques. Biochem. Pharmacol., 24, 2217-2224.

Connors, T.A., Cox, P.J., Farmer, P.B., Foster, A.B. and Jarman, M(1974) Some studies of the active intermediates formed in the microsomal metabolism of cyclophosphamide and isophosphamide. Biochem. Pharmacol., 23, 115-129.

Corder, M.P., McFadden, D.B. and Stump, D.C. (1984) Chemotherapy of advanced transitional cell carcinoma of the bladder. In: Bladder Cancer - Buttervorths International Medical Reviews. Urology 1, Eds. P.H. Smith and G.R. Prout. Butterworths and Co (publishers) Ltd., London, pp240-275.

Costello, A.J., Tiptaft, R.C., England, H.R. and Blandy, J.P.(1984) Squamous cell carcinoma of bladder. Urol., 23, No. 3, 234-236.

Courdi, A., Gioanni, J., Lalanne, C.M., Fischel, J.L., Schneider, M., Ettore, F. and Lambert, J.C. (1983) Establishment, characterization and response to cytotoxic and radiation treatment of three human melanoma cell lines. In Vitro, 19, No.6, 453-461.

Courtenay, D. (1983) The Courtenay clonogenic assay. In: Human Tumour Drug Sensitivity Testing in vitro. Eds. P.P. Dendy and B.T Hill. Academic Press Inc. (London) Ltd., ppl03-lll.

Courtenay, V.D. (1976) A soft agar colony assay for Lewis lung tumour and B16 melanoma taken directly from the mouse. Br. J. Cancer, 34, 39-45.

Courtenay, V.D. (1984) A replenishable soft agar colony assay for human tumour sensitivity testing. In: Recent Results in Cancer Research, Vol. 94, (Predictive drug testing on human tumour cells) Eds. V. Hofmann, M.E. Berens and G. Martz. Springer-Verlag, Berlin, ppl7-34.

Courtenay, V.D. and Mills, J. (1978) An in vitro colony assay for human tumours grown in immune-suppressed mice and treated in vivo with cytotoxic agents. Br. J. Cancer, 37, 261-268.

Courtenay, V.D., Selby, P.J., Smith, I.E., Mills, J. and Peckham, M.J. (1978) Growth of human tumour cell colonies from biopsies using two soft-agar techniques. Br. J. Cancer, 38, 77-81.

Crissman, H.A. and Tobey, R.A. (1974) Cell-cycle analysis in 20minutes. Science, 184, 1297-1298.

Crooke, S.T. and Bradner, W.T. (1976) Mitomycin C: a review.Cancer Treat. Rev., 3, 121-139.

Crooke, S.T. and Reich, S.D. (Eds), (1980) In: Anthracyclines: Current Status and New Developments. Academic Press, New York.

Cross, R.J., Glashan, R.V., Humphrey, C.S., Robinson, M.R.G., Smith, P.H. and Williams, R.E. (1976) Treatment of advancedbladder cancer with adriamycin and 5-fluorouracil. Br. J. Urol.,48, 609-615.

Curt, G.A., Grygiel, J.J., Corden, B.J., Ozols, R.F., Weiss, R.B., Tell, D.T., Myers, C.E. and Collins, J.M. (1983) A Phase I and pharmacokinetic study of diamminecyclobutane-dicarboxylatoplatinum (NSC 241240). Cancer Res., 43, 4470-4473.

Curt, G.A., Clendeninn, N.J. and Chabner, B.A. (1984) Drug resistance in cancer. Cancer Treat. Rep., 68, 87-99.

D'Angio, G.J. and the Wilms' Tumor Study Writing Committee (1982) Biology and management of Wilms' tumor. In: Cancer in the Young.Ed. A.S. Levine. Masson Publishing USA Inc., New York, pp633-662.

Daniels, J.R., Daniels, A.M., Luck, E.E., Whitman, B., Casagrande, J.T. and Skinner, D.G. (1981) Chemosensitivity of human neoplasms with in vitro clone formation. Cancer Chemother. Pharmacol., 6, 245-251.

Deen, D.F., Bartle, P.M. and Williams, M.E. (1979) Response of cultured 9L cells to spirohydantoin mustard and X-rays. Int. J. Radiat. Oncol. Biol. Phys., 5, 1663-1667.

deKernion, J.B. (1977) The chemotherapy of advanced bladder carcinoma. Cancer Res., 37, 2771-2774.

Dendy, P.P., Dawson, M.P.A., Warner, D.M.A. and Honess, D.J. (1976) Quantitative assays of in vitro drug damage. In: Human Tumours in Short Term Culture. Ed. P.P. Dendy. Academic Press Inc. (London) Ltd., ppl39-149.

Dewey, D.L., Butcher, F.V. and Galpine, A.R. (1977) Hydroxyanisole-induced regression of the Harding-Passey melanoma in mice. J. Pathol., 122, No. 3, 117-127.

Dewey, W.C. (1979) In vitro systems: Standardization of end points. Int. J. Radiat. Oncol. Biol. Phys., 5, 1165-1174.

Dexter, D.L., Spremulli, E.N., Fligiel, Z., Barbosa, J.A., Vogel, R., VanVoorhees, A. and Calabresi, P. (1981) Heterogeneity of cancer cells from a single human colon carcinoma. Am. J. Med., 71, 949-956.

Dickson, C., Elkington, J., Hales, A. and Weiss, R. (1980) Antibody-mediated neutralization of virus is abrogated by mycoplasma. Infect. Immun., 28, No, 3, 649-653.

Dickson, J.A. and Suzanger, M. (1976) 1. In vitro sensitivitytesting of human tumour slices to chemotherapeutic agents - its place in cancer therapy. In: Human Tumours in Short Term Culture. Ed. P.P. Dendy. Academic Press Inc. (London) Ltd., ppl07-138.

Diddens, H., Niethammer, D. and Jackson, R.C. (1984) Resistance of human tumor cell lines to antifolates. Cancer Treat. Rev., 11, Suppl. A, 37-41.

DiMarco, A. (1975) Adriamycin (NSC-123127): Mode and mechanism ofaction. Cancer Chemother. Rep., Part 3, 6, No. 2, 91-106.

DiPaolo, J.A. and Moore, G.E. (1958) The use of ascites tumors in solid agar for screening of cancer chemotherapeutic agents. Ann. New York Acad. Sci., 76, 870-881.

Don, P. and Kieler, J. (1980) Cultivation of human bladder epithelial cells and studies of their invasiveness in vitro as a criterion of malignant alteration. In: Biology of the Cancer Cell. Proceedings of the Fifth Meeting of the European Association for Cancer Research, Kugler Publications, pp327-346.

Dorr, R.T. and Alberts, D.S. (1982) Pharmacology of doxorubicin. In: Current Concepts in the use of Doxorubicin Chemotherapy. Ed. S.E. Jones. Farmitala Carlo Erba, Milan, pp3-20.

Double, J.C. and Brown, J.R. (1975) The interaction of aminoalkylaminoanthraquinones with deoxyribonucleic acid. J.Pharm. Pharmacol., 27, 502-507.

Douple, E.B. and Richmond, R.C. (1979) A review of platinum complex biochemistry suggests a rationale for combined platinum-radiotherapy. Int. J. Radiat. One. Biol. Phys., 5, 1335-1339.

Dracopoli, N.C. and Fogh, J. (1983a) Loss of heterozygosity in cultured human tumor cell lines. J. Natl. Cancer Inst., 70, 83-87.

Dracopoli, N.C. and Fogh, J. (1983b) Polymorphic enzyme analysis of cultured human tumor cell lines. J. Natl. Cancer Inst., 70, 469-476.

Drago. J.R. (1984) The noble rat bladder cancer model. Cancer,53, 1093-1099.

Drewinko, B. and Barlogie, B. (1984) Cell cycle pertubation effects. In: Antitumor Drug Resistance. Eds. B.W. Fox and M. Fox. Springer-Verlag, Berlin. pplOl-141.

Drewinko, B., Brown, B.W. and Gottlieb, J.A. (1973) The response of synchronized human lymphoma cells to bleomycin and l,3-Bis(2-chlororoethyl)-l-nitrosourea. Cancer Res., 33,2732-2736.

Drewinko, B., Roper, P.R. and Barlogie, R. (1979) Patterns of cell survival following treatment with antitumor agents in vitro. Eur. J. Cancer, 15, 93-99.

Drewinko, B., Yang, L-Y., Ho, D.H.W., Benvenuto, J., Loo, T.L. and Freireich, E.J. (1980) Treatment of cultured human colon carcinoma cells with fluorinated pyrimidines. Cancer, 45, 1144-1158.

Drewinko, B., Patchen, M., Yang, L-Y. and Barlogie, B. (1981) Differential killing efficacy of twenty anti tumor drugs on proliferating and non-proliferating human tumor cells. Cancer Res., 41, 2328-2333.

Drewinko, B., Yang, L-Y, Barlogie, B. and Trajillo, J.M. (1983) Comparative cytotoxicity of bisantrene, mitoxantrone, ametantrone, dihydroxyanthracenadione, dihydroxyanthracenedione diacetate and doxorubicin on human cells in vitro. Cancer Res., 43, 2648-2653.

Drury, R.A.B. and Wallington, E.A. (Eds), (1980) Carleton's Histological Technique, 5th Ed. Oxford University Press, Oxford.

Duchesne, G.M., Eady, J.J., Peacock, J.H. and Pera, M.F. (1987) A panel of human lung carcinoma lines: Establishment, properties andcommon characteristics. Br. J. Cancer, 56, 287-293.

Ji.O

Duckworth, D.A. (1950) The treatment of papillomatosis with podophyllin. J. Urol., 64, 740.

Dulbecco, R. and Vogt, M. (1954) Plaque formation and isolation of pure lines with poliomyelitis viruses. J. Expt. Med., 99, 167-182.

Durand, R.E. (1980) Variable radiobiological responses of spheroids. Radiat. Res., 81, 85-99.

Durkin, W.J., Ghanta, V.K., Balch, C.M., Davis, D.W. and Hiramoto, R.N. (1979) A methodological approach to the prediction of anticancer drug effect in humans. Cancer Res., 39, 402-407.

Durr, F.E., Wallace, R.E. and Citarella, R.V. (1983) Molecular and biochemical pharmacology of mitoxantrone. Cancer Treat. Rev., 10, Suppl. B, 3-11.

Eagle, H. (1955) Propagation in a fluid medium of a human epidermoid carcinoma, strain KB (21811). Proc. Soc. Exp. Biol. Med., 89, 362-364.

Eagle, H. (1973) The effect of environmental pH on the growth ofnormal and malignant cells. J. Cell Physiol., 82, 1-8.

Eagle, H. and Foley, G.E. (1956) The cytotoxic action of carcinolytic agents in tissue culture. Am. J. Med., 21, 739-749.

Eagle, H. and Foley, G.E. (1958) Cytotoxicity in human cell cultures as a primary screen for the detection of anti-tumor agents. Cancer Res., 18, No. 7, 1017-1025.

Easty, D.M. and Wylie, J.A.H. (1963) Screening of 12 gliomataagainst chemotherapeutic agents in vitro. Br. Med. J., 1, 1589-1592.

Easty, D.M., Easty, G.C., Carter, R.L., Monaghan, P. and Butler, L.J. (1981) Ten human squamous carcinoma cell lines derived from squamous carcinomas of the head and neck. Br. J. Cancer, 43, 772-785.

Easty, G.C. (1970) Organ culture methods In: Methods in Cancer Research. Ed. Harris Busch. Academic Press Inc. (London) Ltd.,Vol. 5, ppl-43.

Edsmyr, F. (1981) Radiotherapy in the management of bladder cancer. In: Bladder Cancer: Principles of Combination Therapy.Eds. R.T.D. Oliver, V.F. Hendry and H.J.G. Bloom. Butterworths and Co (Publishers) Ltd., London ppl39-149.

Edsmyr, F., Andersson, L. and Esposti, P-L. (1984) Chemotherapy in advanced urinary bladder cancer. Urol., 23, No. 3, Suppl. 51-53.

Ekvall, B. (1983) Screening of toxic compounds in mammalian cell cultures. Ann. New York Acad. Sci., 407, 64-77.

Eliason, J.F., Aapro, M.S., Decrey, D. and Brink-Petersen, M.(1985) Non-linearity of colony formation by human tumour cells from biopsy samples. Br. J. Cancer, 52, 311-378.

Elkind, M.M. and Whitmore, G.F. (1967) In vitro survival curves. In: The Radiobiology of Cultured Mammalian Cells. Gordon and Breach Science Publishers Inc., New York, pp53-115.

Elliot, J.S. (1962) Treatment of bladder cancer with amethoptein. Cancer Chemother. Rep., 16, No, 20, 147.

Endresen, L., Tveit, K.M., Rugstad, H.E. and Pihl, A. (1985) Chemosensitivity measurements of human tumour cells by soft agar assays are influenced by the culture conditions. Br. J. Cancer,51, 843-852.

Engel, L.W. and Young, N.A. (1978) Human breast carcinoma cells in continuous culture. Cancer Res., 38, 4327-4339.

Engel, L.W., Young, N.A., Tralka, T.S., Lippman, M.E. O'Brien, S.J. and Joyce, M.J. (1978) Establishment and characterization of three new continuous cell lines derived from human breast carcinomas. Cancer Res., 38, 3352-3364.

Engelholm S.A., Spang-Thornsen, M. and Vindelov, L.L. (1983) A short-term in vitro test for tumour sensitivity to adriamycin based on flow cytometric DNA analysis. Br. J. Cancer, 47, 497-502.

Erlichman, C. and Vidgen, D. (1984) Cytotoxicity of adriamycin in MGH-UI cells grown as monolayer cultures, spheroids and xenografts in immune-deprived mice. Cancer Res., 44, 5369-5375.

Estape, J. and Milla, A. (1982) Review of the clinical activities of VP-16. In: New Approaches in Cancer Therapy, Vol II. Eds. H. Cortes Funes and M. Rozeniveig. Raven Press, New York, ppl5-30.

Falkson, G. ad Falkson, H.C. (1983). Phase II trial ofspirogermanium for treatment of advanced breast cancer. CancerTreat. Rep., 67, no.2. 189-190.

Falkson, G., Pretorius, L. and Falkson, H.C. (1982) Dibromodulcitol in the treatment of breast cancer. Cancer Treat. Rev., 9, 261-266.

Fell, H.B. and Robison, R. (1929) The growth, development and phosphatase activity of embryonic avian femora and limb-buds cultivated in vitro. Biochem. J., 23, 767-784.

Finlay, G.J. and Baguley, B.C. (1984) The use of human cancer cell lines as a primary screening system for antineoplastic compounds.Eur. J. Cancer Clin. Oncol., 20, no. 7, 947-954.

Fisher, B., Redmond, C. and Fisher, E.R. (1985) A summary of findings from the national surgical adjuvant breast and bowel project trials of adjuvant therapy. In: Adjuvant Therapy ofCancer, Vol.4, Eds, S.E. Jones and S.E. Salmon. Grune and Stratton Inc., Orlando, ppl85-194.

Flanagan, S.P. (1966) "Nude", a new hairless gene with pleiotropic effects in the mouse. Genet. Res., Camb., 8, 295-309.

Fogh, J. (1978) Cultivation, characterization and identification of human tumor cells with emphasis on kidney, testis and bladder tumors. Natl. Cancer Inst. Mongr., 49, 5-9.

Fogh, J. and Fogh, H. (1964) A method for direct demonstration of pleuropneumonia-like organisms in cultured cells. Proc. Soc. Exp. Biol. Med., 117, 899-901.

Fogh, J. and Trempe, G. (1975) New human tumor cell lines. Ini Human Tumour Cells In Vitro. Ed. J. Fogh. Plenum Press, New York and London, ppll5-159.

Fogh, J., Wright, W.C. and Loveless, J.D. (1977a) Absence of HeLa cell contamination in 169 cell lines derived from human tumors. J. Natl. Cancer Inst., 58, 209-214.

Fogh, J., Fogh, J.M. ad Orfeo, T. (1977b) One hundred and twenty-seven cultured human tumor cell lines producing tumors in nude mice. J. Natl, Cancer Inst., 59, 221-225.

Fogh, J., Dracopoli, N., Loveless, J.D. and Fogh, H. (1982)Cultured human tumor cells for cancer research: assessment of variation and stability of cultural characteristics. Prog. Clin. Biol. Res., 89, 191-223.

Foley, G.E. and Epstein, S.S. (1964) Cell culture and cancer chemotherapy. In: Advances in Chemotherapy. Vol.l. Eds. A Goldin and F. Hawking. Academic Press, New York. ppl75-353.

Formina, I.P., Terentieva, T.G., Navashin, S.M. and Gibadullin,R.A. (1970) Primary explants of human tumors as a model for studies on the effect of antitumor preparations. In: Proceedings of the 6th International Congress of Chemotherapy, Vol. 2, University Park press, Baltimore, pp340-343.

Fossa, S.D. and Gudmundsen, T.E. (1981) Single-drug chemotherapy with 5-Fu and adriamycin in metastatic bladder carcinoma. Br. J. Urol., 53, 320-323.

Fox, M. (1965) The effect of cyclophosphamide on some urinary tract tumours. Br. J. Urol., 37, 399-409.

Franco, R., Kraft, T., Miller, T., Popp, M. and Martelo, 0. (1984) Storage of chemotherapy drugs for use in the human tumor stem cell assay. Int. J. Cell Cloning, 2, 2-8.

Frank, W. and Osterberg. A.E. (1960) Mitomycin C (NSC-26980) - An evaluation of the Japanese reports. Cancer Chemother. Rep., 9, 114-119.

Freedman, V.H. and Shin. S, (1974) Cellular tumorigenicity in nude mice. Correlation with cell growth in semi-solid medium. Cell, 3, 355-359.

Freedman, V.H. and Shin, S. (1978) In: The Nude Mouse in Experimental and Clinical Research. Eds. J. Fogh and B.Giovanella. Academic Press, New York, pp353-384.

Frei, E. (1985) Curative cancer chemotherapy. Cancer Res., 45, 6523-6537.

Frei, E., Jaffe, N., Tattersall, M.H.N., Pitman, S. and Parker, L.(1975) New approaches to cancer chemotherapy with methotrexate. New Engl. J. Med., 292, 846-851.

Freshney, R.I. (1983a) Maintenance of the culture - cell lines. In: Culture of Animal Cells. Alan R. Liss Inc., New York, ppll9-128.

Freshney, R.I. (1983b) Specialized techniques I. In: Culture of Animal Cells. Alan R. Liss Inc., New York, pp225-243.

Freshney, R.I. (1987) Measurement of cytotoxicity and viability. In: Culture of Animal Cells, 2nd Edition. Alan R. Liss Inc., New York, pp245-256.

Friedman, H.M. and Glaubiger, D.L. (1982) Assessment of in vitro3drug sensitivity of human tumor cells using [ H] thymidine

incorporation in a modified human tumor stem cell assay. Cancer Res., 42, 4683-4689.

Fuskova, A., Fuska, J., Ivanitskaya, L.P. and Makuko L.V. (1975) Different sensitivity to antibiotics of the etablished intermitent in vivo - in vitro tumor cell lines in comparison with the primary cultures. Neoplasma, 24, No.2, 239-242.

Gartler, S.M. (1967) Genetic markers as tracers in cell culture. Natl. Cancer Inst. Monogr., 26, 167-195.

Gazdar, A.F., Carney, D.N., Russell, E.K., Sims, H.L., Bayliss, S.B., Bunn, P.A., Guccion, J.G. and Minna, J.D. (1980) Establishment of continuous, clonable cultures of small-cell carcinoma of the lung which have amine precursor uptake and decarboxylation cell properties. Cancer Res., 40, 3502-3507.

German, J. (1974) Chromosomes and Cancer. New York, John Wiley and Sons Inc., pp375-497.

Gerner, E.W. and Schneider, M.J. (1975) Induced thermal resistance in HeLa cells. Nature, 256, 500-502.

Gerweck, L.E. (1977) Modification of cell lethality at elevated temperature. Radiat. Res., 70, 224-235.

Gey, G.O., Coffman, W.D. and Kubicek, M.T. (1952) Tissue culture studies of the proliferative capacity of cervical carcinoma and normal epithelium. Cancer Res., 12, 264-265.

Gibaldi, M. and Perrier, D. (Eds), (1982) Pharmacokinetics.Marcel Dekker Inc., New York.

Gibert, F. (1983) Chromosomes, genes, and cancer: a classification of chromosome abnormalities in cancer. J. Natl. Cancer Inst., 71, 1107-1114.

Gilbert, H.A., Logan, J.L., Kagan, A.R., Friedman, H.A., Cove,J.K., Fox, M., Muldoon, T.M., Lonni, Y.W., Rowe, J.H., Cooper,J.F., Nussbaum, H., Chan, P., Rao, A. and Starr, A. (1978) The natural history of papillary transitional cell carcinoma of the bladder and its treatment in an unselected population on the basis of histologic grading. J. Urol., 119, No. 4, 488-492.

Gioanni, J., Courdi, A., Lalanne, C.M., Fischel, J-L, Zanghellini,E., Lambert, J-C., Ettore, F. and Namer, M. (1985) Establishment, characterization, chemosensitivity and radiosensitivity of two different cell lines derived from a human breast cancer biopsy. Cancer Res., 45, 1246-1258.

Giovanella, B.C., Stehlin, J.S., Williams, L.J., Lee, S.S. and Shepard, R.C. (1978) Heterotransplantation of human cancers into nude mice: a model system for human cancer chemotherapy. Cancer, 42, 2269-2281.

Glen, J.F., Hunt, L.D. and Lathem, J.E. (1963) Chemotherapy of bladder carcinoma with 5-flourouracil. Cancer Chemother. Rep., 17, No. 27, 67-69.

Gold, P., Shuster, J. and Freedman, S.O. (1978) Carcinoembryonic antigen (CEA) in clinical medicine. Historical perspectives, pitfalls and projections. Cancer, 42, No. 3, Suppl, 1399-1405.

Goldenburg, D.M. and Witte, S. (1967) Chemotherapy of two morphologically similar human tumours growing in the cheek pouch of the golden hamster: H. Ad. No. 1 and GW-39. Eur. J. Cancer, 3, 95-101.

Goldin, A., Venditti, J.M., Macdonald, J.S., Muggia, F.M., Henney, J.E. and DeVita, V.T. (1981) Current results of the screening program at the division of cancer treatment, National Cancer Institute. Eur. J. Cancer, Yj_9 129-142.

Gordon, M.Y., Blackett, N.M. and Douglas, I.D.C. (1975) Colony formation by human haemopoietic precursor cells cultured in semi-solid agar in diffusion chambers. Br. J. Haematol., 31, 103-110.

Gorham, L.W. and Waymouth, C. (1965) Differentiation in vitro of embryonic cartilage and bone in a chemically-defined medium. Proc. Soc. Exp. Biol. Med., 119, 287-290.

Grady, J.E., Lummis, W.L. and Smith, C.G. (1960) An improved tissue culture assay. Ill Alternative methods for measuring cell growth. Cancer Res., 20, No. 7, 1114-1117.

Grail, A., Bates, G., Ward, A.M., Jones, W.G. and Hancock, B.W.(1982) Serum ferritin as a third marker in germ cell tumors. Eur. J. Cancer Clin. Oncol., 18, No. 3, 261-269.

Gray, J.W. (1983) Quantitative cytokinetics: cellular response to cell cycle specific agents. Pharmac. Ther., 22, 163-197.

Greene, H.S.N. (1952) The significance of the heterologous transplantability of human cancer. Cancer, 5, No. 1, 24-44.

Grossman, H.B. (1979) Current therapy of bladder carcinoma. J. Urol., 121, 1-7.

Grossman, H.B., Wedemeyer, G. and Ren, L. (1984) UM-UC-1 and UM-UC-2: characterization of two new human transitional cell carcinoma lines. J. Urol., 132, 834-837.

Gupta, V. and Eberle, R. (1984) Modulation of tumour cell colony growth in soft agar by oxygen and its mechanism. Br. J. Cancer,49, 587-593.

Hagen, K., Daly, J.J., Kamali, H.M., Lin, J.C., Yu, S-C. and Prout, G.R. (1979) New assay for cytotoxic agents in human bladder cancer. Surg. ForUm, 30, 560-562.

Hahn, G.M. (1974) Metabolic aspects of the role of hyperthermia in mammalian cell inactivation and their possible relevance to cancer treatment. Cancer Res., 34, 3117-3123.

Hahn, G.M. (1976) Recovery of cells from induced, potentially lethal damage. Cancer Treat. Rep., 60, 1791-1798.

Hahn, G.M. (1984) Some responses of cells and tumours to elevated temperatures. Front. Radiat. Ther. Oncol., 18, 23-28.

Hahn, G.M. and Little, J.B. (1972) Plateau-phase cultures of mammalian cells. An in vitro model for human cancer. Curr. Top. Radiat. Res., 8, 39-83.

Hahn, G.M. and Li, G.C. (1982) Interactions of hyperthermia and drugs: Treatments and probes. Natl. Cancer Inst. Monogr., 61, 317-323.

Hahn, G.M. and Shin, E.C. (1983) Effect of pH and elevated temperatures on the cytotoxicity of some chemotherapeutic agents on Chinese hamster cells in vitro. Cancer Res., 43, 5789-5791.

Hahn, G.M., Ray, G.R., Gordon, L.F. and Kallman, R.F. (1973) Response of solid tumor cells exposed to chemotherapeutic agents in vivo: Cell survival after 2- and 24-hour exposure. J. Natl.Cancer Inst., 50, 529-533.

Hahn, G.M., Gordon, L.F. and Kurkjian, S.D. (1974) Responses of cycling and non-cycling cells to l,3-Bis(2-chloroethyl)-l- nitrosourea and bleomycin. Cancer Res., 34, 2373-2377.

Hajdu, S.I., Bean, M.A., Fogh, J., Hajdu, E.O. and Ricci, A. (1974) Papanicolaou smear of cultured human tumor cells. Acta Cytologica, 18, 327-332.

Hakala, M.T. and Taylor, E. (1959) The ability of purine and thymine derivatives and of glycine to support the growth of mammalian cells in culture. J. Biol. Chem., 234, 126-128.

Hakala, M.T. and Rustum, Y.M. (1979) The potential value of in vitro screening. In: Methods in Cancer Research, vol. 16, Cancer Drug Development, Part A. Eds. V.T. DeVita and H. Busch, Academic Press, New York, pp247-287.

Hakala, M.T., Suolinna, E-M.R., Kenny, L.N. and Tritsch, G.L.(1974) Asparagine metabolism in mouse sarcoma cells. I.Transport, metabolism and pools of dicarboxylic amino acids and their amides. Biochim. Biophys. Acta., 338, 1-20.

Hakansson, L. and Trope, C. (1974) On the presence within tumours of clones that differ in sensitivity to cytostatic drugs. Acta Pathol. Microbiol. Scand., 82, 35-40.

Hall, E.J. (1978) Cell, tissue and tumor kinetics. In: Radiobiology for the Radiobiologist. 2nd Edition. Ed. E.J. Hall. Harper and Row Publishers Inc., pp246-272.

Ham, R.G. (1965) Clonal growth of mammalian cells in a chemically defined, synthetic medium. Proc. Natl. Acad. Sci., 53, 288-293.

Hamburger, A.W. (1981) Use of in vitro tests in predictive cancerchemotherapy. J. Natl. Cancer Inst., 66, 981-988.

Hamburger, A.W. (1983) The Salmon-Hamburger 11 stem” cell assay.In: Human Tumour Drug Sensitivity Testing in vitro. Eds. P.P.Dendy and B.T. Hill. Academic Press Inc. (London) Ltd., ppll3-119.

Hamburger, A.W. and Salmon, S.E. (1977) Primary bioassay of humantumor stem cells. Science, 197, 461-463.

Hamburger, A.W. and White, C.P. (1981) Interaction between macrophages and human tumor clonogenic cells. Stem Cells, 1, 209-223.

Hamburger, A.W., Salmon, S.E., Kim, M.B., Trent, J.M., Soehnlen,B.J., Alberts, D.S. and Schmidt, H.J. (1978) Direct cloning of human ovarian carcinoma cells in agar. Cancer Res., 38, 3438-3444.

Hamburger, A.W., White, C.P. and Dunn, F.E. (1983) Modulation of tumour colony growth by irradiated accessory cells. Br. J. Cancer, 48, 675-682.

Hamilton, T.C., Young, R.C., Louie, K.G., Behrens, B.C., McKoy, W.M., Grotzinger, K.R. and Ozols, R.F. (1984a) Characterization of a xenograft model of human ovarian carcinoma which produces ascites and intra-abdominal carcinomatosis in mice. Cancer Res., 44,No.11, 5286-5290.

Hamilton, T.C., Young, R.C. and Ozols, R.F. (1984b) Experimental model system of ovarian cancer: applications to the design and evaluation of new treatment approaches. Semin. Oncol., 11,285-298.

Hamlyn, P. and Sikora, K. (1983) Oncogenes. The Lancet, 2, 326-330.

Hanauske, A-R, Hanauske, U. and Von Hoff, D.D. (1985) The human tumor cloning assay in cancer research and therapy. Curr. Probl. Cancer, 9, 1-50.

Hanauske, A-R., Hanauske, U. and Von Hoff, D.D. (1987) Recent improvements in the human tumor cloning assay for sensitivity testing of antineoplastic agents. Eur. J. Cancer Clin. Oncol., 23, No. 6, 603-605.

Handler, J.S., Perkins, F.M. and Johnson, J.P. (1980) Studies of renal cell function using cell culture techniques. Am. J.Physiol., 238, F1-F9.

Harker, W.G. and Torti, F.M. (1983) The chemotherapy of bladder carcinoma: systemic therapy. In: Recent Results in Cancer Research. Ed. F.M. Torti, 85, Berlin, pp37-49.

Harnden, D.G. and Taylor, A.M.R. (1979) Chromosomes and cancer. In: Advances in Human Genetics. Eds. H. Harris and K. Hirschhorn, Plenum Press, New York, ppl-54.

Harper, R.A., Flaxman, B.A. and Chopra, D.P. (1974) Mitotic response of normal and psoriatic keratinocytes in vitro to compounds known to affect intracellular cyclic AMP. J. Invest. Dermatol., 62, No. 4, 384-387.

Harris, H. and Hopkinson, D.A. (1976) Handbook of enzyme electrophoresis in human genetics, North-Holland, Amsterdam.

Harris, L.W. and Griffiths, J.B. (1974) An assessment of methods used for measuring the recovery of mammalian cells from freezing and thawing. Cryobiology, ljL, 80-84.

Harris, N.L., Gang, D.L., Quay, S.C., Poppema, S., Zamecnik, P.C., Nelson-Rees, W.A. and O'Brien, S.J. (1981) Contamination of Hodgkin's disease cell cultures. Nature, 22, No. 289, 228-230.

Harrison, R.G. (1907) Observations on the living developing nerve fiber. Proc. Soc. Exp. Biol. Med., 4, 140-143.

Haspel, H.C., Ren, Y-F., Watanabe, K.A., Sonenberg, M. and Corin, R.E. (1984) Cytocidal effect of gossypol on cultured murine erythroleukemia cells is prevented by serum protein. J. Pharmacol. Exp. Ther., 229, No. 1, 218-225.

Hastings, R.J. and Franks, L.M. (1981) Chromosome pattern, growth in agar and tumorigenicity in nude mice of four human bladder cell lines. Int. J. Cancer, 27, 15-21.

Hastings, R.J. and Franks, L.M. (1983) Cellular heterogeneity in a tissue culture cell line derived from a bladder carcinoma. Br. J. Cancer, 47, 233-244.

Hay, E.D. (1984) Cell-matrix interaction in the embryo: Cellshape, cell surface, cell skeletons and their role in differentiation. In: The Role of Extracellular Matrix in Development. Ed. R.L. Trelstad. Alan R. Liss Inc., New York, ppl-31.

Hayashi, I. and Sato, G.H. (1976) Replacement of serum by hormones permits growth of cells in a defined medium. Nature, 259, 132-134.

Hayes, D.M., Cvitkovic, E., Golbey, R.B., Schneiner, E., Helson, L. and Krakkoff, I.H. (1977) High dose cis-platinum diammine dichloride. Amelioration of renal toxicity by mannitol diuresis. Cancer, 39, 1372-1381.

Heidelberger, C. (1975) Fluorinated pyrimidines and their nucleosides. In: Handbook of Experimental Pharmacology, vol. 38, part II, Eds. A.C. Sartorelli and D.G. Johns. Springer-Verlag. Berlin, ppl93-231.

Heidelberger, C., Griesbach, L., Montag, B.J., Mooren, D., Cruz,0., Schnitzer, R.J. and Grunberg, E. (1958) Studies on fluorinated pyrimidines. II. Effects on transplanted tumors. Cancer Res., 18, 305-317.

Henderson, B.M., Dougherty, W.J., James, V.C., Tilley, L.P. and Noble, J.F. (1982) Safety assessment of a new anticancer compound, mitoxantrone, in Beagle dogs: comparison with doxorubicin. I. Clinical observations. Cancer Treat. Rep., 66, No.5, 1139-1143.

Hepburn, P.J. and Masters, J.R.W. (1983) The biological characteristics of continuous cell lines derived from human bladder. In: The Pathology of Bladder Cancer. Vol 2, Eds.G.T.Bryan and S.M.Cohen. CRC Press Inc. Boca Raton, Florida, pp213-227.

Hepburn, P.J. and Masters, J.R.W. (1986) Influence of clonogenic assay methodology on measurement of drug sensitivities in vitro. Cancer Chemother. Pharmacol., 16, 80-82.

Hepburn, P.J., Oliver, R.T.D., Riley, P.A., Hill, B.T. and Masters, J.R.W. (1985) Comparison of the cytotoxic activities of chemotherapeutic drugs using a human bladder cancer cell line.Urol. Res., 13, 27-34.

Heppner, G.H. (1984) Tumor Heterogeneity. Cancer Res., 44, 2259-2265.

Heppner, G.H. and Miller, B.E. (1983) Tumor heterogeneity: biological implications and therapeutic consequences. Cancer Metastasis Rev., 2, 5-23.

Heppner, G.H., Dexter, D.L., DeNucci, T., Miller, F.R. and Calabresi, P. (1978) Heterogeneity in drug sensitivity among tumor cell subpopulations of a single mammary tumor. Cancer Res., 38, 3758-3763.

Herring, H.T. (1903) The treatment of vesical papilloma by injection. Br. Med. J. 2, 1398.

Heyn, R.M., Holland, R., Newton, W.A., Tefft, M., Breslow, N. and Hartmann, J.R. (1974) The role of combined chemotherapy in the treatment of rhabdomycosarcoma in children. Cancer, 34, No.6, 2128-2142.

Hicks, R.M. and Wakefield, J.St.J. (1972) Rapid induction of bladder cancer in rats with N-methyl-N-nitrosoureu 1. Histology. Chem. Biol. Interact., 5, 139-152.

Hill, B.T. (1978a) Cancer chemotherapy. The relevance of certain concepts of cell cycle kinetics. Biochim. Biophys. Acta, 516, 389-417.

Hill, B.T. (1978b) The management of human "solid” tumours some observations on the irrelevance of traditional cell cycle kinetics and the value of certain recent concepts. Cell Biol. Int. Rep., 2, No. 3, 215-230.

Hill, B.T. (1983a) Use of continuous human tumour cell lines to evaluate drugs by clonogenic assays. In: Human Tumour Drug Sensitivity Testing In Vitro. Eds. P.P. Dendy and B.T. Hill. Academic Press Inc. (London) Ltd.,ppl29-146.

Hill, B.T. (1983b) An overview of correlations between laboratory tests and clinical responses. In: Human Tumour Drug Sensitivity Testing In Vitro. Eds. P.P. Dendy and B.T. Hill. Academic Press Inc. (London) Ltd.,p235-249.

Hill, B.T. (1983c) An overview of clonogenic assays for human tumour biopsies. In: Human Tumour Drug Sensitivity Resting In Vitro. Eds., P.P. Dendy and B.T. Hill. Academic Press Inc. (London) Ltd., pp91-102.

Hill, B.T. and Baserga, R. (1975) The cell cycle and its significance for cancer treatment. Cancer Treat. Rev., 2, 159-175.

Hill, B.T. and Whelan, R.D.H. (1981) Assessments of the sensitivities of cultured human neuroblastoma cells to antitumour drugs. Pediatr. Res., 15, 1117-1122.

Hill, B.T. and Price, L.A, (1982) An experimental biological basis for increasing the therapeutic index of clinical cancer therapy.In: Cell Proliferation, Cancer and Cancer Therapy, Ed. R. Baserga, Annals of the New York Academy of Sciences, New York, Vol. 397, pp71-87.

Hill, B.T. and Whelan, R.D.H. (1982) Establishment of vincristine-resistant and vindesine-resistant lines of murine lymphoblasts in vitro and a characterization of their patterns of cross-resistance and drug sensitivities. Cancer Chemother. Pharmacol., 8, 163-169.

Hill, B.T. and Whelan, R.D.H. (1983) Attempts to optimise colony-forming efficiencies using three different survival assays and a range of human tumour continuous cell lines. Cell Biol. Int. Rep., 7, No. 8, pp617-624.

Hill, B.T., Whelan, R.D.H., Rupniak, H.T. and Rosholt, M.N. (1980) The value of an in vitro culture system for predicting drug sensitivities and their questionable clinical relevance. Br. J. Cancer, 41 Suppl. IV, 203.

Hill, B.T., Whelan, R.D.H., Rupniak, H.T., Dennis, L.Y. and Rosholt, M.A. (1981) A comparative assessment of the in vitro effects of drugs on cells by means of colony assays or flow microfluorimetry. Cancer Chemother. Pharmacol., 7, 21-26.

Hill, B.T., Whatley, S.A., Bellamy, A.S., Jenkins, L.Y. and Whelan, R.D.H. (1982) Cytotoxic effects and biological activity of 2-Aza-8-germanspiro[4,5]-decane-2-propanamine-8,8-diethyl-N,N- dimethyl dichloride (NSC 192965; spirogermanium) in vitro. Cancer Res., 42, 2852-2856.

Hill, B.T., Metcalfe, S., Hepburn, P.J., Masters, J.R.W. and Whelan, R.D.H. (1983) Evaluation of spirogermanium in murine and human tumour lines in vitro and demonstration of synergy with cis-platin or 5-fluorouracil. Proc. 13th Int. Congr. Chemother., 30-33.

Hill, B.T., Bellamy, A.S., Metcalfe, S., Hepburn, P.J., Masters, J.R.W. and Whelan, R.D.H. (1984) Identification of synergistic combinations of spriogermanium with 5-fluorouracil or cisplatin using a range of human tumour cell lines in vitro. Invest. New Drugs, 2, 29-33.

Hill, R.P. and Tannock, I.F. (1987) Cancer as a cellular disease. In: The Basic Science of Oncology. Eds., I.F. Tannock and R.P.Hill. Pergaraon Press, Oxford, ppl-4.

Hisazumi, H., Nakajima, K. and Uchibayashi, T. (1983) Chemosensitivity of established human bladder cell lines in vitro. Gann, 74, 176-182.

Hodges, G.M. (1976) An overview of tissue culture procedures in tumour biopsy studies. In: Human Tumours in Short Term Culture.Ed. P.P. Dendy. Academic Press Inc. (London) Ltd., pp3-14.

Hodges, G.M. (1981) Growth differentiation and function of tumours in organ culture. In: Regulation of Growth in Neoplasia. Ed. G.V. Sherbet. Karger, Basel, pp52-130.

Hoffbrand, A.V. and Tripp, E. (1972) Unbalanceddeoxyribonucleotide synthesis caused by methotrexate. Br. Med. J., 2, 140-142.

Hollenberg, M.D. and Cuatrecasas, P. (1975) Insulin and epidermal growth factor. J. Biol. Chem., 250, No. 10, 3845-3853.

Holley, R.W. and Kiernan, J.A. (1968) "Contact inhibition" of cell division in 3T3 cells. Proc. Natl. Acad. Sci,. 60, 300-304.

Holmes, H.L. and Little, J.M. (1974) Tissue-culture microtest for predicting response of human cancer to chemotherapy. The Lancet,2, 985-987.

Holzer, H. (1964) Action of alkylating substances on glycolysis. In: Chemotherapy of Cancer, Ed. P.A. Plattner. Elsevier Publishing Co., Amsterdam, pp44-50.

Horning, S.J., Hoppe, R.T. and Rosenberg, S.A. (1984) The Stanford Hodgkin's disease trials 1968-1974: results of combined modality therapy in PSI I, II and III. In: Adjuvant Therapy of Cancer IV. Eds. S.E.Jones and S.E.Salmon. Grune and Stratton, Orlando, pp633-642.

Horvath, I.P., Csetenyi, J., Hindy, I., Kerpel-Fronius, S. Institoris, L., Hegedus, L. and Eckhardt, S. (1982) Metabolism and pharmacokinetics of dibromodulcitol (DBD, NSC-104800) in man - II Pharmacokinetics of DBD. Eur. J. Cancer Clin. Oncol., 18, No. 11, 1211-1219.

Hoskins, J.M., Meynell, G.G. and Sanders, F.K. (1956) A comparison of methods for estimating the viable count of a suspension of tumour cells. Expt. Cell Res., 11, 297-305.

Houghton, J.A. and Houghton, P.J. (1983) The xenograft as an intermediate model system. In: Human Tumour Drug Sensitivity Testing In Vitro. Eds. P.P. Dendy and B.T. Hill. Academic Press Inc. (London) Ltd., ppl79-200.

Howard, A. and Pelc, S.R. (1953) Synthesis of deoxyribonucleic acid in normal and irradiated cells and its relation to chromosome breakage. Heredity, 6, suppl., 261-273.

Howell, S.B., Mansfield, S.J. and Taetle, R. (1981) Thymidine and hypoxanthine requirements of normal and malignant human cells for protection against methotrexate cytotoxicity. Cancer Res., 41, 945-950.

Hueper, W.C. and Wolfe, H.D. (1937) Experimental production of aniline tumors of the bladder in dogs. Am. J. Pathol., 13,656-657.

Hurley, J.D. and Yount, L.J. (1965) Selection of anticancer drug for palliation using tissue culture sensitivity studies. Am. J. Surg., 109, 39-42.

Inwood, M., Povey, S. and Delhanty, J.D.A. (1980) Comparison of isozymes in fetal, adult and transformed fibroblasts. Cell Biol. Int. Rep., 4, 327-335.

Iyer, V.N. and Szybalski, W. (1963) A molecular mechanism of mitomycin actin: Linking of complementary DNA strands. Proc.Natl. Acad. Sci., 50, 355-362.

Iyer, V.N. and Szybalski, W. (1964) Mitomycins and porfiromycin: chemical mechanism of activation and cross-linking of DNA.Science, 145, 55-58.

Jackman, A.L., Taylor, G.A., Calvert, A.H. and Harrap, K.R. (1984) Modulation of anti-metabolite effects. Effect of thymidine on the efficacy of the quinazoline-based thymidylate synthetase inhibitor, CB3717. Biochem. Pharmacol., 33, 3269-3275.

Jeffreys, A.J., Wilon, V. and Thein, S.L. (1985a) Hypervariable 'minisatellite' regions in human DNA. Nature, 314, 67-73.

Jeffreys, A.J., Wilson, V. and Thein, S.L. (1985b) Individual-specific 'fingerprints' of human DNA. Nature, 316, 76-79.

Jewett, H.J. and Strong, G.H. (1946) Infiltrating carcinoma of the bladder: Relation of depth of penetration of the bladder wall toincidence of local extension and metastases. J. Urol., 55,366-372.

Jones, H.C. and Swinney, J. (1961) Thiotepa in the treatment of tumours of the bladder. The Lancet, 2, 615-618.

Johnson, R.K. and Goldin, A. (1975) The clinical impact of screening and other experimental tumor studies. Cancer Treat.Rev., 2, 1-31.

Johnson, R.K., Zee-Cheng, R.K-Y., Lee, W.W., Acton, E.M., Henry, D.W. and Cheng, C.C. (1979) Experimental antitumour activity of aminoanthraquinones. Cancer Treat. Rep., 63, No. 3, 425-439.

Joseph, E. (1919) Eine neue methode zur behandlung der blasengeschwulste, vorlaufige mitteilung. Zentralbl Chir., 46,931.

Kakati, S., Hayata, I. and Sandberg, A.A. (1976) Chromosomes and causation of human cancer and leukaemia. XIV. Origin of a large number of markers in a cancer. Cancer, 37, 776-782.

Kaplan, I.W. (1942) Condylomata acuminata. New Orleans, Med.Surg. J., 94, 388-390.

Kato, T., Nemoto, R., Nishimoto, T., Kumagi, I. and Miura, K.(1979) Chemosensitivity of bladder cancer cells in long-term culture and clinical responses to the selected anticancer drug. Cancer, 44, 58-63.

Kaufmann, K., Mattern, J. and Volm, M. (1975) In vitro testing of cyclophosphamide on tumors. Naturwissenchaften, 62, 446-447.

Kelman, A.P., Peresie, H.J. and Stone, P.J. (1977) An analysis of the modes of binding of the anti tumor platinum complexes to DNA.J. Clin. Haemat. Oncol., 7, 440-451.

Kennedy, K., Teicher, B.A., Rockwell, S. and Sartorelli, A.C. (1980a) The hypoxic tumor cell: a target for selective cancer chemotherapy. Biochem. Pharmacol., 29, 1-8.

Kennedy, K., Rockwell, S. and Sartorelli, A.C. (1980b)Preferential activation of mitomycin C to cytotoxic metabolites by hypoxic tumor cells. Cancer Res., 40, 2356-2360.

Kern, D.H., Tanigawa, N., Bertelsen, C.A., Sondak, V.K. and Morton, D.L. (1984) Heterogeneity of chemosensitivity response of human tumors. In: Human Tumor Cloning. Eds. S.E. Salmon and J.M. Trent. Grune and Stratton, Inc., ppl73-182.

Kessel, D. and Bosmann, H.B. (1970) On the characteristics of actinomycin D resistance in L5178Y cells. Cancer Res., 30, 2695-2701.

Kimball, P.M. and Brattain, M.G. (1980) Isolation of a cellular subpopulation from a human colonic carcinoma cell line. Cancer Res., 40, 1574-1579.

Kirk, D., Kagawa, S., Vener, G., Narayan, K.S., Ohnuki, Y. and Jones, L.W. (1985) Selective growth of normal adult human urothelial cells in serum-free medium. In vitro Cell Dev. Biol., 21, 165-171.

Kirkels, W.J., Debruyne, F.M.J., Pelgrim, O.E., Vooys, G.P. and Herman, C.J. (1983) Soft-agar culture of transitional cell carcinma colonies from urine, irrigation fluids and tumor samples. Eur. Urol.,9, 300-304.

Klevjar-Anderson, P. and Buehring, G.C. (1980) Effect of hormones on growth rates of malignant and non-malignant human mammary epithelia in cell culture. In vitro, lj>, No. 6. 491-501.

Knock, F.E. (1965) Sensitivity tests for cancer chemotherapy. Arch. Surg., 91, 376-385.

Knock, F.E. (1971) Qualitative and quantitative in vitro sensitivity tests for cancer chemotherapy. Natl. Cancer Inst. Monogr., 34, 247-250.

Knock, F.E., Galt, R.M., Oester, Y.T. and Sylvester, R. (1974) In vitro estimate of sensitivity of individual human tumors to antitumor agents. Oncol., 30, 1-22.

Knowles, M.A., Hicks, R.M., Berry, R.J. and Milroy, E. (1980)Organ culture of normal human bladder: choice of starting material and culture characteristic. Methods Cell Biol., 21B, 257-285.

Koeffler, H.P., Yen, J. and Lowe, L. (1981) An in vitro model for acute myelogenous leukemia chemotherapy. Cancer, 48, 1958-1963.

Kondo, T. (1971) Prediction of response of tumor and host to cancer chemotherapy. Natl. Cancer Inst. Monogr., 34, 251-256.

Kondo, T., Imamura, T. and Ichihashi, H. (1966) In vitro test for sensitivity of tumor to carcinostatic agents. Gann, 57, 113-121.

Korbling, M., Hess, A.D., Tutschka, P.J., Kaizer, H., Colvin, M.O. and Santos, G.W. (1982) 4-hydroperoxycyclophosphamide: a model for eliminating residual human tumour cells and T-lymphocytes from the bone marrow graft. Br. J. Haematol., 52, 89-96.

Koss, L.G., (1979) Diagnostic cytology and its histological bases. Vol II, 3rd Edition. J.B. Lippincott Co. Philadelphia.

Kovnat, A., Buick, R.N., Connolly, J.G., Jewett, M.A., Keresteci, A.G. and Tannock, I.F. (1984) Comparison of growth of human bladder cancer in tissue culture or as xenografts with clinical and pathological characteristics. Cancer Res., 44, 2530-2533.

Kraemer, P.M. (1971) Complex carbohydrates of animal cells: Biochemistry and physiology of the cells periphery. In: Biomembranes, Vol. I, Ed. L.A. Manson. Plenum Press, New York, pp67-190.

Krishan, A. (1975) Rapid flow cytofluorometric analysis of mammalian cell cycle by propidium iodide staining. J. Cell Biol., 66, 188-193.

Krishan, A., Paika, K. and Frei, E. (1975) Cytofluorometric studies on the action of podophyllotoxin and epipodphyllotoxins (VM-26, VP-16-213) on the cell cycle traverse of human lymphoblasts. J. Cell Biol., 66, 521-530.

Kyriazis, A.A., Kyriazis, A.P., McCombs, W.B. and Peterson, W.D. (1984) Morphological, biological and biochemical characteristics of human bladder transitional cell carcinomas grown in tissue culture and in nude mice. Cancer Res., 44, 3997-4005.

Laboisse, C.L., Augeron, C. and Potet, F. (1981) Growth and differentiation of human gastrointestinal adenocarcinoma stem cells in soft agarase. Cancer Res., 41 310-315.

Lahiri, S.K. (1973) Response of mouse bone marrow colony forming units in different stages of the cell cycle to in vitro incubation with mitomycin-C. Cell Tissue Kinet., 6, 509-514.

Lamerton, L.F. (1974) Basic concepts of cell population kinetics. In: Antineoplastic and Immunosuppressive Agents, Part I. Eds. A.C. Sartorelli and D.G. Johns. Springer-Verlag, Berlin, ppl42-155.

Larionov, L.F. (1965) Process of animal tumours in chemotherapy. In: Cancer Chemotherapy, Pergamon Press, London, ppl50-174.

Law, L.W., Dunn, T.B. and Boyle, P.J. (1949) Observations on the effect of a folic-acid antagonist on transplantable lymphoid leukaemias in mice. J. Natl. Cancer Inst., 10, 179-192.

Lechner, J.F., Haugen, A., McClendon, I.A. and Ward Pettis, F.(1982) Clonal growth of normal adult human bronchial epithelial cells in a serum-free medium. In Vitro, 18, No. 7, 633-642.

Lee, E.Y-H., Parry, G. and Bissell, M.J. (1984) Modulation of secreted proteins of mouse mammary epithelial cells by the extracellular matrix. J. Cell Biol., 98, 146-155.

Lee, Y.C. and Byfield, J.E. (1976) Induction of DNA degradation in vivo by adriamycin. J. Natl. Cancer Inst., 57, 221-224.

Leibovitz, A., Stinson, J.C., McCombs, W.B., McCoy, C.E., Mazur, K.C. and Mabry, N.D. (1976) Classification of human colorectal adenocarcinoma cell lines. Cancer Res., 36, 4562-4569.

Leive, L. (1965) A nonspecific increase in permeability in Escherichia coli produced by EDTA. Proc. Natl. Acad. Sci., 53, 745-750.

Lemcke, R.M. (1965) Media for the mycoplasmataceae. Lab. Pract., 14, 712-715.

Levi, F.A., Blum, J.P., Lemaigre, G., Bourut, C., Reinberg, A. and Mathe, G. (1984) A four-day subrenal capsule assay for testing the effectiveness of anticancer drugs against human tumors. Cancer Res., 44, 2660-2667.

Levi, P.E., Cooper, E.H., Anderson, C.K. and Williams, R.E. (1969) Analyses of DNA content, nuclear size and cell proliferation of transitional cell carcinoma in man. Cancer, 23, 1074-1085.

Levin, V.A. (1986) Clinical anticancer pharmacology: Somepharmacokinetic considerations. Cancer Treat. Rev., 13, No. 2, 61-76.

Liao, S. (1975) Cellular receptors and mechanisms of action of steroid hormones. Int. Rev. Cytol., 41, 87-172.

Liberman, D.F., Roti Roti, J.L. and Lange, C.S. (1973) Specific G2 phase toxicity of chloramphenicol on mouse leukemic cells (L5178Y). Exp. Cell Res., 77, 351-355.

Lieber, M.M. and Kovach, J.S. (1982) Soft agar colony formation assay for chemotherapy sensitivity testing of human solid tumors. Mayo Clin. Proc., 57, 527-528.

Lin, A.J., Cosby, L.A. and Sartorelli, A.C. (1976) Potential bioreductive alkylating agents. In: Cancer Chemotherapy. Ed. A.C. Sartorelli. American Chemical Society, Washington, pp71-86.

Lin, C-W., Lin, J.C. and Prout, G.R. (1985) Establishment and characterisation of four human bladder tumor cell lines and sublines with different degrees of malignancy. Cancer Res., 45,No.10, 5070-5079.

Link, M., Goorin, A., Miser, A., Green, A., Pratt, C., Belasco, J., Pritchard, J., Baker, A., Kirkpatrick, J. and Ayala, A. (1985) The role of adjuvant chemotherapy in the treatment of osteosarcoma of the extremity: Preliminary results of the multi-institutional osteosarcoma study. Proc. ASCO., 4, 237.

Lippman, M.E., Allegra, J.C. and Strobl, J.S. (1979) Growth requirements of a human breast cancer cell line in serum-free medium. In: Cold Spring Harbor Conference - Cell proliferation, Vol. 6, Book B. Eds., G.H. Sato and R. Ross, pp545-558.

Livingston, R.B., Titus, G.A. and Heibrun, L.K. (1980) In vitro effects on DNA synthesis as a predictor of biological effect from chemotherapy. Cancer Res., 40 2209-2212.

Llopis, B., Vera-Donoso, C.D., Gallego, J., Boronat, F. and Jimenez-Cruz, J.F. (1989) Immunoprophylaxis with Calmette-Guerin Bacillus in recurrent superficial bladder tumors. Eur. Urol., 16, 28-30.

Logothetis, C.J., Dexeus, F.H., Chong, C., Sella, A., Ayala, A.G., Ro, J.Y. and Pilat, S. (1989) Cisplatin, cyclophosphamide and doxorubicin chemotherapy for unresectable urothelial tumors: TheM.D. Anderson experience. J. Urol., 141, 33-37.

Loike, J.D. and Harwitz, S.B. (1976) Effects of podophyllotoxin and VP-16-213 on microtubule assembly in vitro and nucleoside transport in HeLa cells. Biochem., 15, No. 25, 5435-5442.

Ludwig, R. and Alberts, D.S. (1984) Chemical and biological stability of anticancer drugs used in a human tumor clonogenic assay. Cancer Chemother. Pharmacol., 12, 142-145.

Ludwig, C.U., Peng, Y.M., Beaudry, J.N. and Salmon, S.E. (1984a) Cytotoxicity of mitomycin C on clonogenic human carcinoma cells is not enhanced by hypoxia. Cancer Chemother. Pharmacol., 12,146-150.

Ludwig, R., Alberts, D.S., Miller, T.P. and Salmon, S.E. (1984b) Evaluation of anticancer drug schedule dependancy using an in vitro human tumor clonogenic assay. Cancer Chemother. Pharmacol., 12 135-141.

Lum, B.L. (1983) Intravesical chemotherapy of superficial bladder cancer. In: Recent Results in Cancer Research, Vol. 85. Ed. F.M. Torti. Berlin, pp3-36.

Maciag, T., Nemore, R.E., Weinstein, R. and Gilchrest, B.A. (1981) An endocrine approach to the control of epidermal growth:Serum-free cultivation of human keratinocytes. Science, 211, 1452-1454.

MacKintosh, F.R., Evans, T.L. and Sikic, B.I. (1981) Methodologic problems in clonogenic assays of spontaneous human tumors. Cancer Chemother. Pharmacol., 6, 205-210.

Malkovsky, M. and Bubenik, J. (1977) Human urinary bladder carcinoma cell line (T24) in long-term culture: chromosomal studies on a wild population and derived sublines. Neoplasraa, 24, No. 3, 319-326.

Manfredi, F. (1963) Calculation of total body intracellular pH in normal human subjects from the distribution of 5,5-dimethyl-2,4-oxazolidimedione (DM0). J. Lab. Clin. Med., 61, 1005-1014.

Marshall, C.J., Franks, L.M. and Carbonell, A.W. (1977) Markers of neoplastic transformation in epithelial cell lines derived from human carcinomas. J. Natl. Cancer Inst. 58, 1743-1751.

Marshall, V.F. (1952) The relation of the preoperative estimate to the pathologic demonstration of the extent of vesical neoplasmas. J. Urol., 68, No. 4, 714-723.

Marsoni, S. and Wittes, R. (1984) Clinical development of anticancer agents - A National Cancer Institute Perspective.Cancer Treat. Rep., 68, No.l, 77-85.

Masters, J.R.W. (1983) An overview of the role of organ culture for human tumour biopsies. In: Human Tumour Drug Sensitivity Testing In Vitro. Eds. P.P. Dendy and B.T. Hill, Academic Press Inc. (London) Ltd., ppl63-177.

Masters, J.R.W., Hepburn, P.J., Walker, L., Highman, W.J., Trejdosiewicz, L.K., Povey, S., Parkar, M., Hill, B.T., Riddle,P.R. and Franks, L.M. (1986) Tissue culture model of transitional cell carcinoma: Characterization of twenty-two human urothelialcell lines. Cancer Res., 46, 3630-3636.

Masters, J.R.W., Bedford, P., Kearney, A., Povey, S. and Franks, L.M. (1988) Bladder cancer cell line cross-contamination: Identification using a locus-specific minisatellite probe. Br. J. Cancer, 57, 284-286.

Mather, J.P. and Sato, G.H. (1979) The growth of mouse melanoma cells in hormone-supplemented, serum-free medium. Exp. Cell Res., 120, 191-200.

Mattern, J. and Volm, M. (1982) Clinical relevance of predictive test for cancer chemotherapy. Cancer Treat. Rev., 9, 267-298.

Mattern, J., Haag, D., Wayss, K. and Volm, M. (1981) Growth kinetics of human lung tumours in nude mice. Exp. Cell Biol., 49 34-40.

Mauro, F., Falpo, B., Briganti, G., Elli, R. and Zupi, G. (1974a) Effects of antineoplastic drugs on plateau-phase cultures of mammalian cells. II. Bleomycin and hydroxyurea. J. Natl. Cancer Inst., 52, No.3, 715-722.

Mauro, F., Falpo, B., Briganti, G., Elli, R. and Zupi, G. (1974b) Effects of antineoplastic drugs on plateau-phase culture of mammalian cells. I. Description of the plateau-phase system. J. Natl. Cancer Inst., 52, No. 3, 705-713.

Mauro, F., Spano, M., Teodori, L., Drewinko, B. and Barlogie, B.(1983) Cell cycle and effect of antineoplastic drugs: facts and pitfalls. 13th International Cancer Congress, Part C. Biology of cancer (2) Alan R. Liss Inc., New York, pp 199-209.

McAllister, R.M., Grunmeier, P.W., Coriell, L.L. and Blakemore, W.S. (1959) A quantitative technique for the measurement of in vitro cytotoxicity of 5-fluoro-2-deoxyuridine (FUDR), nitrogen mustard (HN2) and other anti tumor agents. Cancer, 12, No. 5, 938-943.

McLimans, W.F., Blumenson, L.E, and Tunnah, K.V. (1968) Kinetics of gas diffusion in mammalian cell culture systems II. Theory. Biotech. Bioengin., 10, 741-763.

Medoff, J., Medoff, G., Goldstein, M.N., Schlessinger, D. and Kobayashi, G.S. (1975) Amphotericin B-induced sensitivity to actinomycin D in drug-resistant HeLa cells. Cancer Res., 35, 2548-2552.

Mendlesohn, M.L. (1962) Autoradiographic analysis of cell proliferation in spontaneous breast cancer of C3H mouse. III. The growth fraction. J. Natl. Cancer Inst., 28, No. 5, 1015-1029.

Messing, E.M., Fahey, J.L., deKernion, J.B., Bhuta, S.M. and Bubbers, J.E. (1982) Serum-free medium for the in vitro growth of normal and malignant urinary bladder epithelial cells. Cancer Res., 42, 2392-2397.

Metcalfe, S.A., Whelan, R.D.H., Masters, J.R.W. and Hill, B.T.(1983) In vitro responses of human prostate tumour cell lines to a range of anti tumour agents* Int. J. Cancer, 32, 351-358.

Metelmann, H-R., and Von Hoff, D.D. (1983) Application of a microsomal drug activation system in a human tumor cloning assay. Invest. New Drugs, 1, 27-32.

Meyskens, F.L., Moon, T.E., Dana, B., Gilmartin, E., Casey, W.J., Chen, H-S. G., Hood Franks, D., Young, L. and Salmon, S.E. (1981) Quantitation of drug sensitivity by human metastatic melanoma colony-forming units Br. J. Cancer, 44, 787-797.

Mickey, D.D. (1986) Use of mouse bladder cancer cell lines for the evaluation of chemotherapeutic agents. In: In Vitro Models for Cancer Research. Vol. IV, Carcinomas of the Urinary Bladder and Kidney. Eds. M.M. Webber and L.I. Sekely. CRC Press Inc., Boca Raton, Florida, pp201-229.

Mikkelsen, R.B., Lin, P-S., Wallach, D.F.H. (1977) Interaction of adriamycin with human red blood cells: a biochemical and morphological study. J. Mol. Med., 2, 33-40.

Millar, B.C., Fielden, E.M. and Jinks, H.S. (1982) Further studies on the nature of the biphasic radiation survival response of Chinese hamster cells V-79-753B to molecular oxygen. Adv. Exp. Med. Biol., 157, 109-122.

Mishina, T., Oda, K., Murata, S., Ooe, H., Mori, Y. and Takahashi, T. (1975) Mitomycin C bladder instillation therapy for bladder tumors. J. Urol., 114, 217-219.

Miyamura, S., (1956) A determination method for anticancer action of antibiotics by the agar plate diffusion technique. J. Antibiotics Chemother., 6, 280-282.

Montagnier, L. (1971) Factors controlling the multiplication of untransformed and transformed cells under variousenvironmental conditions. In: Growth Control in Cell Cultures,Eds. G.E.W. Wolstenholme and J. Knight. Churchill Livingstone, Edinburgh, London, pp33-43.

Moore, G.E., Gerner, R.E. and Franklin, H.A. (1967) Culture of normal human leukocytes. J. Am. Med. Assoc., 199, No.8,519-524.

Morales, A., Eidinger, D. and Bruce, A.W. (1976) Intracavitary bacillus Calmette-Guerin in the treatment of superficial bladder tumors. J. Urol., 116, 180-183.

Morasca, L., Balconi, G., Erba, E., D'Incalci, M., Ottalenghi, L., Salmona, A. and Garattini, S. (1980) Pharmacokinetic approach to in vitro testing of ovarian cancer cell sensitivity. Oncol., 37, 169-173.

Morgan, B.D.G., O'Neill, T., Dewey, D.L., Galpine, A.R. and Riley, P.A. (1981) Treatment of malignant melanoma by intravascular4-hydroxyanisole. Clin. Oncol., 7, 227-234.

Morgan, D., Freshney, R.I., Darling, J.L., Thomas, D.G.T. and Celik, F. (1983) Assay of anticancer drugs in tissue culture: cell cultures of biopsies from human astracytoma. Br. J. Cancer, 47, 205-214.

Mossmann, T. (1983) Rapid colorimetric assay for cellular growth and survival: Application to proliferation and cytotoxicityassays. J. Immunol. Methods, 65, 55-63.

Muggia, F.M. and Rozencweig, M. (1978) Platinum compounds in clinical oncology: Review and future prospects. In: Advances in Cancer Chemotherapy. Ed. Umezawa, H. Japan Science Society Press, Tokyo/University Park Press, Baltimore, ppl41-154.

Mulcahy, R.T. (1984) Effect of oxygen on misonidazole chemosensitization and cytotoxicity in vitro. Cancer Res., 44, 4409-4413.

Mulcahy, R.T., Dembs, N.L. and Ublacker, G.A., (1984) Enhancement of nitrosourea cytotoxicity by misonidazole in vitro: Correlation with carbamoylating potential. Br. J. Cancer, 49, 307-313.

Mulinos, M.G. and Amin, P. (1980) Toxicology and Phase I study of a new anticancer agent, spirogermanium HCL (SG). Fed., Proc., 39, 747.

Muller, W.E.G., Totsuka, A., Nusser, I., Zahn, R.K. and Umezawa, H. (1975) Bleomycin inhibition of DNA synthesis in isolated enzyme systems and in intact cell systems. Biochem. Pharmacol., 24, 911-915.

Munchausen, L.L. and Rahn, R.O. (1975) Physical studies on the binding of cis-dichlorodiammine platinum (II) to DNA and hemopoly- nucleotides. Biochim. Biophys. Acta, 414, 242-255.

Murdock, K.C., Child, R.G., Fabio, P.F. and Angier, R.B. (1979) Antitumor agents I. l,4-bis[(aminoalkyl)amino]-9,10-anthra- cenediones. J. Med. Chem., 22, No. 9, 1024-1030.

Nakagawa, S., Nakao, M., Toyoda, K., Nukui, M., Takada, H., Ebisui, K. and Watenabe, H. (1988) Combination chemotherapy for advanced urothelial cancer with methotrexate, vinblastine, adriamycin and cisplatin (M-VAC). Nippon Hinyokika Gakka Zasshi, 79, part 9, 1510-1515.

Nakata, Y. and Bader, J.P. (1969) The uptake of nucleosides by cells in culture. II. Inhibition by 2-mercapto-l-(0-4-pyridethyl) benzimidazole. Biochem. Biophys. Acta, 130, 250-256.

Navata, H., Bronzert, D. and Lippman, M.E. (1981) Isolation and characterization of a tamoxifen-resistant cell line derived from MCF-7 human breast cancer cells. J. Biol. Chem., 256, 5016-5021.

Nederman, T. (1983) Growth of tumour cells as multicellular spheroids and antitumour drug evaluation. In: Human Tumour Drug Sensitivity Testing in vitro. Eds. P.P. Dendy and B.T. Hill, Academic Press Inc. (London) Ltd., ppl47-161.

Nederman, T. (1984) Effects of vinblastine and 5-fluorouracil on human glioma and thyroid cancer cell monolayers and spheroids. Cancer Res., 44, 254-258.

Nederman, T., Carlsson, J. and Malmqvist, M. (1981) Penetration of substances into tumor tissue: A methodological study on cellular spheroids. In vitro, 17, No.4, 290-298.

Nelson-Rees, W.A., Daniels, D.W. and Flandermeyer, R.R. (1981) Cross-contamination of cells in culture. Science, 212, 446-452.

Nesbit, M.E., Perez, C.A., Tefft, M., Burgert, E.O., Veitti,T.J.,Kissane, J., Pritchard, D.J. and Gehan, E.A. (1981)Multimodal therapy for the management of primary, nonmetastatic Ewing's sarcoma of bone: An intergroup study. Natl. Cancer Inst. Monogr., 56, 255-262.

Neuwald, P.D., Anderson, C., Salivar, W.O., Aldenderfer, P.H., Dermody, W.C., Weintraub, B.D., Rosen, S.W., Nelson-Rees, W.A. and Ruddon, R.W. (1980) Expression of oncodevelopmental gene products by human tumor cells in culture. J. Natl. Cancer Inst., 64, 447-459.

Niell, H.B. and Soloway, M.S. (1983) Use of the tumour colon assay in the evaluation of patients with bladder cancer. Br. J. Urol., 55, 271-274.

Niell, H.B., Webster, K.C. and Smith, E.E. (1985) Anticancer drug activity in human bladder tumor cell lines. Cancer, 56, 1039-1044.

Nissen, E. and Tanneberger, St. (1981) Influence of pH and serum on the effectivity of antineoplastic agents in vitro. Arch. Geschwulstforsch, 51, 152-156.

O'Brien, S.J., Shannon, J.E. and Gail, M.H. (1980) A molecular approach to the identification and individualization of human and animal cells in culture: Isozyme and allozyme genetic signatures.In vitro, 16, 119-135.

Office of population censuses and surveys. Mortality statistics. Review of the Registrar General on deaths in England and Wales. HMSO, London. 1982.

Ohnuma, T., Arkin, H. and Holland, J.F. (1986) Effects of cell density on drug-induced cell kill kinetics in vitro (inoculum effect). Br. J. Cancer, 54, No. 3, 415-421.

Oleson. J.R. (1984) Overview and future directions of hyperthermia. Front. Radiat. Ther. Oncol. 18, 183-186.

Oliver, C., Waters, J.F., Tolbert, C.L. and Kleinman, H.K. (1987) Growth of exocrine acinar cells on a reconstituted basement membrane gel. In Vitro Cell Dev. Biol., 23, No. 7, 465-473.

Oliver, R.T.D., England, H.R., Risdon, R.A. and Blandy, J.P. (1984)Methotrexate in the treatment of metastatic and recurrent primarytransitional cell carcinoma. J. Urol., 131, 483-485.

Oosterhuis, J.W., Andrews, P.W., Knowles, B.B. and Damjanov, I.(1984) Effects of cisplatinum on embryonal carcinoma cell lines invitro. Int. J. Cancer, 34, 133-139.

O'Toole, C., Price, Z.H., Ohnuki, Y. and Unsgaard, B. (1978) Ultrastructure, karyology and immunology of a cell line originated from a human transitional-cell carcinoma. Br. J. Cancer, 38,64-76.

O'Toole, C.M., Povey, S., Hepburn, P. and Franks, L.M. (1983) Identity of some human bladder cancer cell lines. Nature (Lond), 301, 429-430.

Ozols, R.F. and Yagoda, A. (1986) Genitourinary cancer.Testicular nonseminoraas and extragonadal germ cell tumors. In: Cancer Chemotherapy. Vol. 8. Eds. H.M. Pinedo and B.A. Chabner, Elsevier Science, pp395-420.

Ozols, R.F., Speyer, J.L., Jenkins, J. and Myers, C.E. (1984)Phase II trial of 5-fluorouracil administered intraperitoneally to patients with refractory ovarian cancer. Cancer Treat. Rep., 68, No. 10, 1229-1232.

Pantelouris, E.M. (1968) Absence of thymus in a mouse mutant. Nature, 217, 370-371.

Park, C.H., Bergsagel, D.E. and McCulloch, E.A. (1971) Mouse myeloma tumor stem cells: A primary cell culture assay. J. Natl.Cancer Inst., 46, No. 2, 411-422.

Parker, R.C., Castor, L.N. and McCulloch, E.A. (1957) Altered cell strains in continuous culture: A general survey. In: Cellular Biology Nucleic Acids and Viruses. Eds. O.V. Whitelock and T.M. Rivers. Spec. Publ. New York Acad. Sci., Vol 5, pp303-313.

Patton, T.F., Himmelstein, K.J., Belt, R., Bannister, S.J., Sternson, L.A. and Repta, A.J. (1978) Plasma levels and urinary excretion of filterable platinum species following bolus injection and IV infusion of cis-dichlorodiammineplatinum II in man. Cancer Treat. Rep., 62, 1359-1362.

Paul, J. (Ed), (1965) The growth and metabolism of cultured cells. In: Cell and Tissue Culture, 3rd Edition. Livingstone Ltd.,Edinburgh and London, pp42-59.

Paul, J. (Ed), (1975) Cell and tissue culture. Churchill Livingstone, Edinburgh, ppl72-184.

Pauli, B.U., Anderson, S.N., Memoli, V.A. and Kuettner, K.E. (1980) The isolation and characterization in vitro of normal epithelial cells, endothelial cells and fibroblasts from rat urinary bladder. Tissue Cell, 12, 419-435.

Pauli, B.U., Alroy, J. and Weinstein, R.S. (1983) The ultrastructure and pathobiology of urinary bladder cancer. In: The Pathology of Bladder Cancer. Vol. 2, Eds. G.T. Bryan and S.M.Cohen. CRC Press Inc. Boca Raton, Florida, pp41-140.

Pavelic, Z.P., Slocum, H.K., Rustum, Y.M., Creaven, P.J., Nowak, N.J., Karakousis, C., Takita, H. and Mittelman, A. (1980) Growth of cell colonies in soft agar from biopsies of different human solid tumors. Cancer Res., 40, 4151-4158.

Pavone-Macaluso, M., Ingargiola, G.B., Tripi, M. and Lamartina, M.(1984) Intravesical instillation of adriamycin in treatment of bladder cancer. Urol., 23, No. 3, suppl., 40-47.

Penso, G. and Balducci, D., (Eds), (1963) Tissue cultures in pharmacology. In: Tissue Cultures in Biological Research.Elsevier, Publishing Co., Amsterdam. pp360-383.

Pettengill, O.S., Sorenson, G.D., Wurster-Hill, D.H., Curphey,T.J., Noll, W.W., Cate, C.C. and Maurer, L.H. (1980) Isolation and growth characteristics of continuous cell lines from small-cell carcinoma of the lung. Cancer, 45, 906-918.

Phillips, B.J. (1974) A simple, small scale cytotoxicity test and its uses in drug metabolism studies. Biochem. Pharmacol., 23, 131-138.

Phillips, H.J. and Terryberry, J.E. (1957) Counting actively metabolising tissue cultured cells. Exptl. Cell Res., 13, 341-347.

Pihl, A. (1986) UICC study group on cheraosensitivity testing of human tumors. Problems - applications - future prospects. Int. J. Cancer, 37, 1-5.

Pinedo, H.M. and Chabner, B.A. (1977) Role of drug concentration, duration of exposure and endogenous metabolites in determining methotrexate cytotoxicity. Cancer Treat. Rep., 61t 709-715.

Piper, A.A. and Fox, R.M. (1982) Biochemical basis for the differential sensitivity of human T- and B-lymphocyte lines to5-fluorouracil. Cancer Res., 42, 3753-3760.

Post, G. and Greig, R. (1983) The experimental and clinical implications of cellular heterogeneity in malignant tumors. J. Cancer Res. Clin. Oncol., 106, No. 3, 159-170.

Povey, S., Hopkinson, D.A., Harris, H. and Franks, L.M. (1976) Characterization of human cell lines and differentiation from HeLa by enzyme typing. Nature, 264, 60-63.

Povey, S., Jeremiah, S., Arthur, E., Steel, M. and Klein, G. (1980) Differences in genetic stability between human cell lines from patients with and without lymphoreticular malignancy. Ann. Hum. Genet. 44, 119-133.

Povlsen, C.O., Spang-Thomsen, M., Rygaard, J. and Visfeldt, J.(1980) Heterotransplantation of human malignant tumours to athymic nude mice. In: Immunodeficient Animals for Cancer Research. Ed.S. Sparrow. Macmillan Press (London), pp95-103.

Powis, G. (1985) Anticancer drug pharmacodynamics. Cancer Chemother. Pharmacol., 14, 177-183.

Pragner, M.D. (1971) Prediction of response to L-asparaginase therapy. Natl. Cancer Inst. Monogr., 34, 257-265.

Pratt, W.B. and Ruddon, R.W. (1979a) The alkylating agents. In: The Anticancer Drugs. Oxford University Press Inc., New York, pp64-97.

Pratt, W.B. and Ruddon, R.W. (1979b) The antibiotics. In: The Anticancer Drugs. Oxford University Press Inc., New York, ppl48-194.

Prout. G.R. (1984) Superficial bladder cancer. In: Bladder Cancer. Eds. P.H. Smith and G.R. Prout. Butterworths Co (Publishers) Ltd., London. ppl51-171.

Prout, G.R. and Hagen, I.K. (1982) Chemotherapy of superficial bladder carcinoma. In: Genitourinary malignancy I, Vol. 6, Cancer Treatment and Research. Ed. E. Paulson. Martinus Nijhoff Publishers. B.V. The Hague ppl43-157.

Puck, T.T. and Marcus, P.I. (1955) A rapid method for viable cell titration and clone productions with HeLa cells in tissue culture : The use of x-irradiated cells to supply conditioning factors.Proc. Ntl. Acad. Sci., 41, 432-437.

Puck, T.T., Marcus, P.I. and Cieciura, S.J. (1956) Clonal growth of mammalian cells in vitro. J. Expl. Med., 103, 273-283.

Pugh, R.C.B. (1981) Histological staging and grading of bladder tumours. In: Bladder Cancer : Principles of Combination Therapy. Eds. R.T.D Oliver, W.F. Hendry and H.J.G. Bloom. Butterworths and Co (Publishers) Ltd., London, pp3-8.

Rajewski, M.F. (1966) Zell proliferation in normalen und malignen3Geweben: H-thymidin-einbau in vitro unter standardbedingungen.

Biophysik, 3, 65-93.

Ramachandran, J. and Suyama, A.T. (1975) Inhibition of replication of normal adrenocortical cells in culture by adrenocorticotropin. Proc. Natl. Acad. Sci. USA, 72, No. 1, 113-117.

Rasheed, S., Gardner, M.B., Rongey, R.W., Nelson-Rees, W.A. and Arnstein, P. (1977) Human bladder carcinoma: Characerization of two new tumor cell lines and search for tumor viruses. J. Natl. Cancer Inst., 58, 881-890.

Ray, G.R., Hahn, G.M., Bagshaw, M.A. and Kurkjian, S. (1973) Cell survival and repair of plateau-phase cultures after chemotherapy - relevance to tumor therapy and to the in vitro screening of new agents. Cancer Chemother. Rep. Part 1, 57, No. 4, 473-475.

Reid, L., Morrow, B., Jubinsky, P., Schwartz, E. and Gatmaitan, Z.(1981) Regulation of growth and differentiation of epithelial cells by hormones, growth factors and substrates of extracellular matrix. Ann. New York Acad. Sci., 372, 354-370.

Reznikoff, C.A., Johnson, M.D., Norback. D.H. and Bryan, G.T. (1983) Growth and characterization of normal human urothelium iji vitro. In Vitro, 19, No. 4, 326-343.

Rice, L.M., Wheeler, J.W. and Geschickter, C.F. (1974) Spirans XXII. Synthesis of 4,4-dialky1-4-germacyclohexanone and 8,8-dialkyl-8-germaazaspiro (4.5) decanes. J. Heterocyclic Chem., 11, 1041-1046.

Richter, A., Sanford, K.K. and Evans, V.J. (1972) Influence of oxygen and culture media on plating efficiency of some mammalian tissue cells. J. Natl. Cancer Inst., 49, 1705-1712.

Riddle, P.N. (1979) Specimen environments. In: Time-lapse Cinemicroscopy. Ed. P.N. Riddle, Academic Press Inc., (London) Ltd., pp23-35.

Rigby, C.C., Franks, L.M. (1970) A human tissue cell line from a transitional cell tumour of the urinary bladder: Growth, chromosome pattern and ultrastructure. Br. J. Cancer, 24, 746-754.

Riley, P.A. (1969a) Hydroxyanisole depigmentation: In vitroStudies. J. Pathol., 97, 185-191.

Riley, P.A. (1969b) Hydroxyanisole depigmentation: In vitroStudies. J. Pathol., 97, 193-206.

Riley, P.A. (1970) Mechanism of pigment-cell toxicity produced by hyroxyanisole. J. Pathol., 101, 163-169.

Riley, P.A., Sawyer, B. and Wolff, M.A. (1975) The melanocytotoxic action of 4-hydroxy-anisole. J. Invest. Dermatol., 64, No. 2, 86-89.

Riley, P.A. (1981) Control of proliferation of normaland neoplastic cells in culture. In: Regulation of Growth in Neoplasia. Ed. G.V. Sherbet. Karger, Basal, ppl31-198.

Rinaldini, L.M.J. (1958) The isolation of living cells from animal tissues. In: International Review of Cytology. Vol. 7, Eds. G.H. Bourne and J.F. Danielli. Academic Press Inc., New York. pp587-647.

Rizzino, A. and Sato, G. (1978) Growth of embryonal carcinoma cells in serum-free medium. Proc. Natl. Acad. Sci., 75, No. 4, 1844-1848.

Roberts, J.T. and Pascoe, J.M. (1972) Cross-linking of complementary strands of DNA in mammalian cells by antitumour platinum components. Nature, 235, 282-284.

Roblin, R., Albert, S.O., Gelb, N.A. and Black, P.H. (1975) Cell surface changes correlated with density-dependant growth inhibition. Glycosaminoglycan metabolism in 3T3, SV3T3 and Con A selected revertant cells. Biochera., 14, No. 2, 347-357.

Roizin-Trowle, L. and Hall, E.J. (1978) Studies with bleomycin and misonidazole on aerated and hypoxic cells. Br. J. Cancer, 37, 254-260.

Roper, P.R. and Drewinko, B. (1976) Comparison of in vitro methods to determine drug-induced cell lethality. Cancer Res., 36, 2182-2188.

Rosen, G. and Nirenberg, A. (1984) Chemotherapy of primary osteogenic sarcoma: Ten-year evolution and current status of preoperative chemotherapy. In: Adjuvant Therapy of Cancer IV. Eds. S.E. Jones and S.E. Salmon. Grune and Stratton, Orlando, pp593-600.

Rosenberg, B., VanCamp, L. and Krigas, T. (1965) Inhibition of cell division in Escherichia coli by electrolysis products from a platinum electrode. Nature, 205, 698-699.

Rosenberg, B., VanCamp, L., Trosko, J.E. and Mansour, V.H. (1969) Platinum compounds: A new class of potent anti tumour agents.Nature, 222, 385-386.

Rosenblum M.L., Dougherty, D.A., Brown, J.M., Barker, M., Deen,D.F. and Hoshino, T. (1980) Improved methods for disaggregating single cells from solid tumors. Cell Tissue Kinet., 13, 667.

Rosenblum M.L., Dougherty, D.V., Reese, C. and Wilson, C.B. (1981) Potentials and possible pitfalls of human stem cell analysis.Cancer Chemother. Pharmacol., 6, 227-235.

Ross, W.E., Zwelling, L.A. and Kohn, K.W. (1979) Relationship between cytotoxicity and DNA strand breakage produced by adriamycin and other intercalating agents. Int. J. Radiat. Oncol. Biol.Phys., 5, 1221-1224.

Rubin, H. (1971) pH and population density in the regulation of animal cell multiplication. J. Cell Biol., 5 , 686-702.

Ruckdeschel, J.C., Carney, D.N., Oie, H.K., Russell, E.K. and Gazdar, A.F. (1987) In vitro chemosensitivity of human lung cancer cell lines. Cancer Treat. Rep., 71, No. 7-8, 697-704.

Rupniak, H.T. and Hill, B.T. (1980) The poor cloning ability in agar or human tumour cells from biopsies of primary tumours. Cell Biol. Int. Rep., 4, No. 5, 479-486.

Rupniak, H.T., Dennis, L.Y. and Hill, B.T. (1983a) An intercomparison of in vitro asays for assessing cytoxocity after a 24 hour exposure to anticancer drugs. Tumori, 69, 37-42.

Rupniak, H.T., Whelan, R.D.H. and Hill, B.T. (1983b) Concentration and time-dependant inter-relationships for antitumour drug cytotoxicities against tumour cells in vitro. Int. J. Cancer, 32, 7-12.

Russell, P.J., Raghaven, D., Gregory, P., Philips, J., Wills, E.J., Jelbart, M., Wass, J., Zbroja, R.A. and Vincent, P.C. (1986)Bladder cancer xenografts: A model of tumor cell heterogeneity. Cancer Res., 46, 2035-2040.

Rutman, R.J., Cantarov, A. and Paschkis, K.E. (1954) Studies in2-acetylaminofluorene carcinogenesis. Ill The utilization of

14uracil-2-C by preneoplastic rat liver and rat hepatoma. Cancer Res., 14, 119-123.

Rygaard, J. ad Povlsen, C.O. (1969) Heterotransplantation of a human malignant tumour to "nude” mice. Acta Pathol. Microbiol. Scand., 77, 758-760.

Sabatini, D.D., Griepp, E.B., Rodriguez-Boulan, E.J., Dolan, V.J., Robbins, E.S., Papadopoulos, S., Ivanov, I.E. and Rindler, M.J.(1983) Biogenesis of epithelial cell polarity. In: Spatial Organization of Eukaryotic Cells. Ed. J.R. McIntosh. Alan R. Liss, Inc., New York, pp419-450.

Salmon, S.E. (Ed), (1980) Cloning of Human Tumor Stem Cells. New York, Alan R. Liss.

Salmon, S.E. (1982) Studies of clonogenic human stem cells. In: Recent Results in Cancer Research. Eds. G. Mathe, G. Bonadonna and S. Salmon. Springer-Verlag, Berlin, Heidelberg, New York. Vol. 80, pp317-322.

Salmon, S.E., Hamburger, A.W., Soehnlen, B., Durie, B.G.M.,Alberts, D.S. and Moon, T.E. (1978) Quantitation of differential sensitivity of human-tumor stem cells to anticancer drugs. New Engl. J. Med., 298, No.24, 1321-1327.

Salmon, S.E., Alberts, D.S., Durie, B.G.M., Meyskens, F.L., Jones, S.E., Soehnlen, B., Chen, H-S.G. and Moon,T. (1980a) Clinical correlations of drug sensitivity in the human tumor stem cell assay. In: Recent Results in Cancer Research. Eds. G. Mathe andF.M. Muggia. Springer-Verlag, Berlin, Heidelberg, New York, Vol. 74, pp300-305.

Salmon, S., Alberts, D. Meyskens, F., Durie, B., Jones. S., Soehnlen, B. and Young, L. (1980b) A new concept: In vitro Phase II trial with the human tumor stem cell assay (HSTCA) Proc. Am.Ass. Cancer Res., 21, 329.

Sapareto, S.A., Hopwood, L.E., Dewey, W.C., Rajn, M.R. and Gray, J.W. (1978) Effects of hyperthermia on survival and progression of Chinese hamster ovary cells. Cancer Res., 38, 393-400.

Sarosdy, M.F., Lamm, D.L., Radwin, H.M. and Von Hoff, D.D. (1982) Clonogenic assay and in vitro chemosensitivity testing of human urologic malignancies. Cancer, 50, 1332-1338.

Schaeffer, W.I. (1984) Usage of vertebrate, invertebrate and plant cell, tissue and organ culture terminology. In vitro, 20, No.l, 19-24.

Schein, P.S., Slavik, M., Smythe, T., Hoth, D., Smith, F., Macdonald, J.S. and Woolley, P.V. (1980) Phase I clinical trial of spirogermanium. Cancer Treat. Rep., 64, No. 10-11, 1051-1056.

Scher, H.I., Yagoda, A., Herr, H.W., Sternberg, C.N., Bosl, G., Morse, M.J., Sogani, P.C., Watson, R.C., Dershaw, D.D., Reuter, V., Geller, N., Hallander, P.S., Darracott Vaughan, E., Whitmore, W.F. and Fair, W.R. (1988) Neoadjuvant M-VAC (methotrexate, vinblastine, doxorubicin and cisplatin) effect on the primary bladder lesion. J. Urol, 139, 470-474.

Schimke, R.T., Beverley, S., Brown, P., Cassin, R., Federspiel, N., Gasser, C., Hill, A., Johnston, R., Mariani, B., Mosse, E., Rath,H., Smouse, D. and Tlsty, T. (1984) Gene amplification and drug resistance in cultured animal cells. Cancer Treat. Rev., 11,Suppl. A, 9-17.

Schlag, P. and Flentje, D. (1984) Chemosensitivity testing of human neoplasms using the soft agar colony assay. Cancer Treat. Rev., 11, Suppl. A, 131-137.

Schrek, R. (1936) A method for viable cells in normal and in malignant cell suspensions. Am. J. Cancer, 28, 389-392.

Schutz, L. and Mora, P.T. (1968) The need for direct cell contact in 'contact' inhibition of cell division in culture. J. Cell Physiol., 71, 1-6.

Schuurmans, D.M., Duncan, D.T. and Olson, B.H. (1961) Variations in drug sensitivity among five mammalian cell lines used in agar plate assay. Cancer Res., 21, No.l, 773-777.

Schwartz, H.S., Sternberg, S.S. and Philips, F.S. (1963) Pharmacology of mitomycin C. IV Effects in vivo on nucleic acid synthesis; comparison with actinomycin D. Cancer Res., 23, 1125-1136.

Sealy, R., Korrubel, J., Cridland, S. and Blekkenhorst, G. (1984) Interstitial misonidazole. A preliminary report on a new perspective in clinical radiation sensitization and hypoxic cell chemotherapy. Cancer, 54, 1535-1540.

Selby, P.J. and Steel, G.G. (1982) Use of the agar diffusion chamber for the exposure of human tumor cells to drugs. Cancer Res., 42, 4758-4762.

Selby, P., Buick, R.N. and Tannock, I. (1983) A critical appraisal of the 'Human tumor stem-cell assay'. New Engl. J. Med., 308, No. 3, 129-134.

Sellei, C., Eckhardt, S., Horvath, I.P., Kralovanszky, J. and Institoris, L. (1969) Clinical and pharmacologic experience with dibromodulcitol (NSC-104800), a new anti tumor agent. Cancer Chemother. Rep., Part 1, 53, No.6, 377-384.

Semple, J.E. (1948) Papillomata of the bladder treated with podophyllin. Br. Med. J., 1, 1235.

Senn, J.S., McCulloch, E.A. and Till, J.E. (1967) Comparison of colony-forming ability of normal and leukaemic human marrow in cell culture. The Lancet, 2, 597-598.

‘Shackney, S.E., McCormack, G.W. and Cuchural, G.J. (1978) Growth rate patterns of solid tumors and their relation to responsiveness to therapy. Ann. Int. Med., 89, 107-121.

Sharkey, F.E. and Fogh, J. (1984) Considerations in the use of nude mice for cancer research. Cancer Metastasis Rev., 3, 341-360.

Shih, C., Padhy, L.C., Murray, M. and Weinberg, R.A. (1981) Transforming genes of carcinomas and neuroblastomas introduced into mouse fibroblasts. Nature, 290, 261-264.

f

Shoemaker, R.H., Abbott, B.J., Macdonald, M.M., Mayo, J.G., Venditti, J.M. and Wolpert-DeFilippes, M.K. (1983) Use of the KB cell line for in vitro cytotoxicity assays. Cancer Treat. Rep.,67, No.1, 97.

Shoemaker, R.H., Wolpert-DeFilippes, M.K., Kern, D.H., Lieber,M.M., Makuch, R.W., Melnick, N.R., Miller, W.T., Salmon, S.E., Simon, R.M., Venditti, J.M. and Von Hoff, D.D. (1985) Application of a human tumor colony-forming assay to new drug screening.Cancer Res., 45, 2145-2153.

Shrivastav, S. and Paulson, D.F. (1980) In vitro chemotherapy testing of transitional cell carcinoma. Invest. Urol., 17,395-400.

Siebert, K., Shafie, S.M., Triche, T.J., Whang-Peng, J.J., O'Brien, S.J., Toney, J.H., Huff, K.K. and Lippman, M.E. (1983) Clonal variation of MCF-7 breast cancer cells in vitro and in athymic nude mice. Cancer Res., 43, 2223-2239.

Siegel, M.M., Chung, H.S., Rucker, N., Siegel, S.E., Seeger, R.C., Hart, I. and Benedict, W.F. (1980) In vitro and in vivo preclinical chemotherapy studies of human neuroblastoma. Cancer Treat. Rep., 64, No.8-9, 975-979.

Sikic, B.I. and Taber, R.L. (1981) Human tumor clonogenic assays. An overview. Cancer Chemother. Pharmacol., 6, 201-203.

Silverberg, E. and Lubera, J. (1987) Cancer Statistics. CA, 37,No.1, 2-19.

Silvestrini, R. Sanfilippo, 0. and Daidone, M.G. (1983) An attempt to use measurement of radioactive precursor incorporation to predict clinical response to treatment. In: Human Tumor Drug Sensitivity Testing In Vitro. Eds. P.P. Dendy and B.T. Hill. Academic Press Inc. (London) Ltd., pp281-290.

Sinkovics, J.G., Gyorkey, F., Kusyk, C. and Siciliano, M.J. (1978) Growth of human tumor cells in established culture. Methods Cancer Res., 14, 243-323.

Siracky, J. (1979) An approach to the problem of heterogeneity of human tumour-cell populations. Br. J. Cancer, 39, 570-577.

Siskin, J.E. and Kinosita, R. (1961) Timing of DNA synthesis in the mitotic cycle in vitro. J. Biophys. Biochem. Cytol., 9,509-518.

Skipper, H.E. and Schabel, F.M. (1984) Tumor stem cell heterogeneity: Implications with respect to classification of cancer by chemotherapeutic effect. Cancer Treat. Rep., 68, 43-61.

Sladek, N.E. (1971) Metabolism of cyclophosphamide by rat hepatic microsomes. Cancer Res., 31, 901-908.

Slocum, H.K., Pavelic, Z.P. and Rustum Y.M. (1980) An enzymatic method for the disaggregation of human solid tumors for studies of clonogenicity and biochemical determinants of drug action. In: Cloning of Human Tumor Stem Cells. Ed. S.E. Salmon. Alan R. Liss Inc., New York, pp339-343.

Slocum, H.K., Pavelic, Z.P., Kanter, P.M., Nowak, N.J. and Rustum, Y.M. (1981a) The soft agar clonogenicity and characteristion of cells obtained from human solid tumors by mechanical and enzymatic means. Cancer Chemother. Pharmacol., 6, 219-225.

Slocum, H.K., Pavelic, Z.P., Rustum, Y.M., Creaven, P.J., Karakousis, C., Takita, H. and Greco, W.R. (1981b)Characterisation of cells obtained by mechanical and enzymatic means from human melanoma, sarcoma and lung tumors. Cancer Res., 41, 1428-1434.

Smalley, R.V., Bartolucci, A.A., Hemstreet, G. and Hester, M.(1981) A phase II evaluation of a 3-drug combination of cyclophosphamide, doxorubicin and 5-fluorouracil and of 5-fluorouracil in patients with advanced bladder carcinoma or stage D prostatic carcinoma. J. Urol., 125, No.2, 191-195.

Smith, H.S. and Dollbaum, Ch.M. (1981) Growth of human tumours inculture. In: Tissue Growth Factors. Ed. R. Baserga. Springer-Verlag, Berlin, Heidelberg, New York, pp 451-490.

Smith, H.S., Hackett, A.J., Lan, S. and Stampfer, M.R. (1981) Use of an efficient method for culturing human mammary epithelial cells to study adriamycin sensitivity. Cancer Chemother. Pharmacol., 6, 237-244.

Smith, I.E. (1983) Mitoxantrone (novantrone): A review ofexperimental and early clinical studies. Cancer Treat. Rev., 10,103-115.

Smith, I.E., Courtenay, V.D. and Gordon, M.Y. (1976) A colony- forming assay for human tumour xenografts using agar in diffusion chambers. Br. J. Cancer, 34, 476-483.

Smith, P.H. (1981) Chemotherapy of bladder cancer. Cancer Treat. Rep., 65, (Suppl.1), 165-173.

Smith, P.H. (1984) Invasive bladder cancer - a lethal enigma. In: Bladder Cancer (Buttervorths International Medical Reviews, Urology 1). Eds. P.H. Smith and G.R. Prout. Butterworths Co. (Publishers) Ltd., London, pp 223-239.

Sobrero, A.F. and Marsh, J.C. (1984) Chemosensitivity of human tumor clonogenic cells simultaneously assayed in agar diffusion chambers and in a two-layer agar culture system. Cancer Treat. Rep., 68, No.4, 615-624.

Soloway, M. (1975) Animal models used for experimental chemotherapy of genitourinary tract tumors. In: Chemotherapy of Urogenital Tumors. Eds. G.P. Murphy and A. Mittelman.Springfield: Charles C. Thomas, ppl88-239.

Soloway, M.S. (1988) Incorporation of M-VAC combination chemotherapy in the treatment of patients with invasive and metastatic bladder cancer. Eur. Urol., 14, Suppl. 1, 42.

Soloway, M.S. and Murphy, W.M. (1979) Experimental chemotherapy of bladder cancer - systemic and intravesical. Semin. Oncol., 6,No.2, 166-183.

Soloway, M.S. and Murphy, W.M. (1981) Chemotherapy of murine bladder cancer. In: Bladder Cancer: Principles of Combination Therapy. Eds. R.T.D. Oliver, W.F. Hendry, H.J.G. Bloom. Butterworths and Co (Publishers) Ltd., pp37-55.

Soloway, M.S. and Perry, A. (1987) Bacillus Calmette-Guerin for treatment of superficial transitional cell carcinoma of the bladder in patients who have failed thiotepa and/or mitomycin C. J. Urol., 137, 871-873.

Soloway, M.S., deKernion, J.B., Rose, D. and Persky, L. (1973) Effect of chemotherapeutic agents on bladder cancer: A new animalmodel. Surg. Forum, 24, 542-544.

Sparano, B.M., Gordon, G., Hall, C., Iatropoulos, M.J. and Noble, J.F. (1982) Safety assessment of a new anticancer compound, mitoxantrone, in Beagle dogs: Comparison with doxorubicin II. Histologic and ultrastructural pathology. Cancer Treat. Rep., 66, No.5, 1145-1158.

Springer, E.L. (1980) Comparative study of the cytoplasmic organelles of epithelial cell lines derived from human carcinomas and non-malignant tissues. Cancer Res., 40, 803-817.

Spooner, M.E. and Cooper, E.H. (1972) Chromosome constitution of transitional cell carcinoma of the urinary bladder. Cancer, 29, 1401-1412.

Stahelin, H. (1973) Activity of a new glycosidic lignan derivative (VP16-213) related to podophyllotoxin in experimental tumors. Eur. J. Cancer, 9, 215-221.

Stanisic, T.H. and Buick, R.N. (1980) An in vitro clonal assay for bladder cancer: Clinical correlation with the status of the urothelium in 33 patients. J. Urol., 124, 30-33.

Stanisic, T.H., Buick, R.N., Trent, J.M., Fry, S.E. and Salmon,S.E. (1981) An in vitro clonal assay for bladder cancer: Studies of the biologic potential of the urothelium and determination of in vitro sensitivity to cytotoxic agents. J. Surg. Oncol., 18, 67-72.

Steel, G.G. (Ed.) (1977) Growth and survival of tumour stem-cells. In: Growth Kinetics of Tumours. Clarendon Press, Oxford, pp217-267.

Steel, G.G. (1978) The growth and therapeutic response of human tumours in immune deficient mice. Bull. Cancer (Paris), 65, 465-472.

Steel, G.G. and Peckham, M.J. (1980) Human tumour xenografts: Acritical appraisal. Br. J. Cancer, 41, Suppl. 4, 133-141.

Steg, A., Lelea, C., Debre, B., Boccon-Gibod, L. and Sicard, D. (1989) Systemic Bacillus Calmette-Guerin infection, 'BCGitis', in patients treated by intravesical Bacillus Calmette-Guerin therapy for bladder cancer. Eur. Urol. 16, 161-164.

Stephens, T.C. and Peacock, J.H. (1982) Clonal variation in the sensitivity of B16 melanoma to m-AMSA. Br. J. Cancer, 45, 821-829.

Stephens, T.C., Adams, K. and Peacock, J.H. (1984) Identification of a subpopulation of MeCCNU resistant cells in previously untreated Lewis lung tumors. Br. J. Cancer, 50, 77-83.

Sternberg, C.N., Yagoda, A., Scher. H.I., Watson, R.C., Herr, H.W., Morse, M.J., Sogani, P.C., Darracott Vaughan, E., Bander, N., Weiselberg, L.R., Geller, N., Hollander, P.S., Lipperman, R., Fair, W.R. and Whitmore, W.F. (1988) M-VAC (methotrexate, vinblastine, doxorubicin and cisplatin) for advanced transitional cell carcinoma of the urothelium. J. Urol, 139, 461-469.

Stevens, K.M. (1965) Oxygen requirements for liver cells in vitro. Nature, 206, 199.

Stiles, C.D., Desmond, W., Chuman, L.M., Sato, G. and Saier, M. (1976) Growth control of heterologous tissue culture cells in the congenitally athymic nude mouse. Cancer Res., 36, 1353-1360.

Stratford, I.J., Williamson, C. and Adams, G.E. (1980) Combination studies with misonidazole and a cis-platinum complex: Cytotoxicity and radiosensitization in vitro. Br. J. Cancer, 41, 517-522.

Stone, P.J., Kelman, A.D., Sinex, F.M., Bhargava, M.M. and Halvorsen, H.O. (1976) Resolution of a, 0 and y DNA of saccharomyces cerevisiae with the anti tumor drug Cis-Pt (NH^^C^* Evidence for preferential drug binding by GpG sequences of DNA. J. Mol. Biol., 104, 793-801.

Sutherland, R.M. (1980) The multi cellular spheroid system as a tumor model for studies of radiation sensitizers. Pharmacol. Ther. 8, 105-123.

Sutherland, R.M. and Durand, R.E. (1976) Radiation response of multicell spheroids: an in vitro tumour model. Curr. Top. Radiat. Res., 11, 87-139.

Sutherland, R.M., Carlsson, J., Durand, R. and Yuhas, J. (1981) Spheroids in cancer research. Cancer Res., 41, 2980-2984.

Swallow, D.M., Evans, L. and Hopkinson, D.A. (1977) Several of the adenosine deaminase isozymes are glycoproteins. Nature, 269, 261-262.

Szechter, A., Kowalsky, W.P. and Schwarz, G. (1978) Average versus viable cell multiplicity in survival determination for mammalian cells in vitro. Radiat. Res., 75, 432-438.

Taetle, R., Howell, S.B., Giuliani, F.C., Koziol, J. and Koessler,A. (1982) Comparison of the activity of doxorubicin analogues using colony-forming assay and human xenografts. Cancer, 50, 1455-1461.

Takahashi, I., Ohnuma, T., Kavy, S., Bhardwaj, S. and Holland, J.F. (1980) Interaction of human serum albumin with anticancer agents in vitro. Br. J. Cancer, 41, 602-608.

Talmadge, J.E., Benedict, K., Madsen, J. and Fidler, I.J. (1984) Development of biological diversity and susceptibility to chemotherapy in murine cancer metastases. Cancer Res., 44, 3801-3805.

Tanigawa, N., Kern, D.H., Hikasa, Y. and Morton, D.L. (1982) Rapid assay for evaluating the chemosensitivity of human tumors in soft agar culture. Cancer Res., 42, 2159-2164.

Tanneberger, S. and Nissen, E. (1983) Predicting response of human solid tumors to chemotherapy. Cancer Treat. Rev., 10, 203-219.

Tanneberger, S., Nissen, E. and Schalicke, W. (1979) Prediction of drug efficacy potentialities and imitations. In: Advances in Medical Oncology, Research and Education, Vol. 5. Ed. B.W. Fox, Pergamon Press, Oxford, ppl97-211.

Tannock, I. and Guttman, P. (1981) Response of Chinese hamsterovary cells to anticancer drugs under aerobic and hypoxic conditions. Br. J. Cancer, 43, 245-248.

Tannock, I., Gospodarowicz, M., Connolly, J. and Jewett, M. (1989)M-VAC (methotrexate, vinblastine, doxorubicin and cisplatin) chemotherapy for transitional cell carcinoma: The Princess Margaret Hospital experience. J. Urol., 142, No. 2, 289-292.

Tate, H.C., Rawlinson, J.B. and Freedman, L.S. (1979) Randomised comparative studies in the treatment of cancer in the United Kingdom: room for improvement? The Lancet, 2, 623-625.

Taub, M., Chuman, L., Saier, M.H. and Sata, G. (1979) Growth of Madin-Darby canine kidney epithelial cell (MDCK) line in hormone supplemented, serum free medium. Proc. Natl. Acad. Sci., USA, 76, No. 7, 3338-3342.

Taylor, I.W. (1980) A rapid single step staining technique for DNA analysis by flow microfluorimetry. J. Histochem. Cytochem., 28,No. 9, 1021-1024.

32Tchao, R., Easty, G.C. and Ambrose, E.J. (1967) The use of P labelled orthophosphate for the assay of chemotherapeutic agents on tumours maintained in organ culture. Br. J. Cancer, 21, 821-827.

Teicher, B.A., Lazo, J.S. and Sartorelli, A.C. (1981)Classification of antineoplastic agents by their selective toxicities towards oxygenated and hypoxic tumor cells. Cancer Res., 41, 73-81.

Tennant, J.R. (1974) Evaluation of the trypan blue technique for determination of cell viability. Transplantation, 2, No. 6, 685-694.

Terasima, T. and Umezawa, H. (1970) Lethal effect of bleomycin on cultured mammalian cells. J. Antibiot., 23, No. 6, 300-304.

Terasima, T., Yasukawa, M. and Umezawa, H. (1970) Breaks and rejoining of DNA in cultured mammalian cells treated with bleomycin. Gann., 61, 513-516.

Terasima, T., Takabe, Y., Watanabe, M. and Katsumata, T. (1976) Effect of bleomycin on cell survival and some implications for tumor therapy. Prog. Biochem. Pharmacol., 11, 68-77.

Thirwell, M.P., Livingston, R.B., Murphy, W.K. and Hort, J.S.(1976) A rapid in vitro labelling index method for predicting response of human solid tumors to chemotherapy. Cancer Res., 36, 3279-3283.

Thomson, J.E. and Rauth, A.M. (1974) An in vitro assay to measure the viability of KHT tumor cells not previously exposed to culture conditions. Radiat. Res., 58, 262-276.

Thuning, C.A., Fanshaw, M.S. and Warren, J. (1983) Mechanisms of the synergistic effect of oral dimethyl sulfoxide on antineoplastic therapy. Ann. N.Y. Acad. Sci., 411, 150-160.

Tilgen, W., Boukamp, P., Breitkreutz, D., Dzarlieva, R.T.,Engstner, M., Haag, D. and Fusenig, N.E. (1983) Preservation of morphological, functional and karyotypic traits during long-term culture and in vivo passage of two human skin squamous cell carcinomas. Cancer Res., 43, 5995-6011.

Tisman, G., Herbert, V. and Edlis, H. (1973) Determination of therapeutic index of drugs by in vitro sensitivity tests using human host and tumor cell suspensions. Cancer Chemother. Rep., 57, 11-19.

Toburen, D. (1981) Different survival of cell subpopulations in spheroids after treatment with bleomycin. J. Cancer Res. Clin. Oncol., 102, 189-193.

Todaro, G.J. and Green, H. (1963) Quantitative studies of the growth of mouse embryo cells in culture and their development into established lines. J. Cell Biol., L7 299-313.

Todaro, G.J., Lazar, G.K. and Green, H. (1965) The initiation of cell division in a contact-inhibited mammalian cell line. J. Cell Comp. Physiol., 66, 325-334.

Traynor, O.J. and Carter, A.E. (1981) Bladder Cancer. Br. Med. J., 282, 2136-2137.

Trejdosiewicz, L.K., Southgate, J., Donald. J.A., Masters, J.R.W., Hepburn, P.J. and Hodges, G.M. (1985) Monoclonal antibodies to human urothelial cell lines and hybrids: Production and characterization. J. Urol., 133, 533-538.

Tritton, T.R. and Lee, G. (1982) The anticancer agent adriamycin can be actively cytotoxic without entering cells. Science, 217, 248-250.

Trope, C., Mattsson, W., Gynning, J., Johnson, J-E., Sigurdsson, K. and Orbert, B. (1981) Phase II study of spirogermanium in advanced ovarian malignancy. Cancer Treat. Rep., 65, No. 1-2, 119-120.

Trowell, O.A. (1959) The culture of mature organs in a synthetic medium. Exp. Cell Res., 16, 118-147.

Trujillo, J.M., Drewinko, B. and Ahearn, M.J. (1972) The ability of tumor cells of the lymphoreticular system to grow in vitro. Cancer Res., 32, 1057-1065.

Tsao, M.C., Walthall, B.J. and Ham, R.G. (1982) Clonal growth of normal human epidermal keratinocytes in a defined medium. J. Cell Physiol., UO, 219-229.

Tsukeda, H., Mizumo, S. and Nitta, K. (1978) Susceptibilities of normal and malignant human lung lines in culture to the cytocidal action of antitumor agents. Cancer Res., 38, 2529-2532.

Tsuruo, T. and Fidler, I.J. (1981) Differences in drug sensitivity among tumor cells from parental tumors, selected variants and spontaneous metastases. Cancer Res., 41, 3058-3064.

Tubiana, M. and Malaise, E.P. (1976) Growth rate and cell kinetics in human tumours: some prognastic and therapeutic implications.In: Scientific Foundations of Oncology. Eds. T. Symington and R.L. Carter. William Heinemann Medical Books Ltd., ppl26-136.

Turner, A.G., Hendry, W.F., Williams, G.B. and Bloom, H.J.G. (1977) The treatment of advanced bladder cancer with methotrexate. Br. J. Urol., 49, 673-678.

Tveit, K.M. and Pihl, A. (1981) Do cell lines in vitro reflect the properties of the tumours of origin? A study of lines derived from human melanoma xenografts. Br. J. Cancer, 44, 775-786.

Tveit, K.M., Fodstad, 0., Olsnes, S. and Pihl, A. (1980) In vitro sensitivity of human melanoma xenografts to cytotoxic drugs. Correlation with in vivo chemosensitivity. Int. J. Cancer, 26, 717-722.

Tveit, K.M., Endresen, L., Rugstad, H.E., Fodstad, 0. and Pihl, A. (1981a) Comparison of two soft-agar methods for assaying chemosensitivity of human tumours in vitro: Malignant melanomas.Br. J. Cancer, 44, 539-544.

Tveit, K.M., Fodstad, 0. and Pihl, A. (1981b) The usefulness of human tumor cell lines in the study of chemosensitivity: a study of malignant melanomas. Int. J. Cancer, 28, 403-408.

Twentyman, P.R. (1978) Sensitivity to 1,3-bis(2-chlorethyl)-l- nitrosourea and l-(2-chloroethyl)-3-(4-methylcyclohexyl)-l- nitrosourea of the EMT6 tumor in vivo as determined by both tumor volume response and in vitro plating assay. Cancer Res., 38, 2395-2400.

Twentyman, P.R. (1979) Timing of assays: an important consideration in the determination of clonogenic cell survival both in vitro and in vivo. Int. J. Radiat. Oncol. Biol. Phys., 5, 1213-1220.

Twentyman, P.R. (1980) Experimental chemotherapy studies: Intercomparison of assays. Br. J. Cancer, 41, Suppl. IV, 279-287.

Twentyman, P.R. and Bleehen, N.M. (1973) The sensitivity of cells in exponential and stationary phases of growth to bleomycin and tol,3-Bis(2-chloroethyl)-l-nitrosourea. Br. J. Cancer, 28, 500-507.

Twentyman, P.R. and Bleehen, N.M. (1975) Changes in sensitivity to cytotoxic agents occurring during the life history of monolayer cultures of a mouse tumour cell line. Br. J. Cancer, 31, 417-423.

Twentyman, P.R. and Luscombe, M. (1987) A study of some variables in a tetrazolium dye (MTT) based assay for cell growth and chemosensitivity. Br. J. Cancer, 56, 279-285.

UICC (International Union Against Cancer), (1978) TNM Classification of malignant tumours. 3rd Edition. Ed. M.H. Harmer, Geneva.

UICC (International Union Against Cancer), (1981) Technical report: Bladder Cancer, No. 13, Vol.60, Eds. P. Skrabanak and A. Walsh. Geneva, pp74-95.

Uitendaal, M.P., Hubers, H.A.J.M., McVie, J.G. and Pinedo, H.M. (1983) Human tumour clonogenicity in agar is improved by cell-free ascites. Br. J. Cancer, 48, 55-59.

Umbach, G.E., Spitzer, G., Ajani, J.A., Thames, H., Hug, V.M., Rudolph, F. and Drewinko, B. (1984) Factors determining methotrexate cytotoxicity in human bone marrow progenitor cells: Implications for in vitro drug testing of human tumors. In: Human Tumor Cloning. Eds. S.E. Salmon and J.M. Trent. Grune and Stratton, New York, pp443-450.

Umezawa, H. (1974) Chemistry and mechanism of action of bleomycin. Fed. Proc., 33, No. 11, 2296-2302.

Valente, R., Marini, F., Signori, G.B. and Martino, F. (1989) Quality of life in patients with advanced bladder cancer treated by tumorectomy and systemic adjuvant polichemotherapy (M-VAC). Our experience in 14 cases. Arch. Esp de Urol., 42, No.l, 87-89.

Valeriote, F.A. and Tolen, S.J. (1972) Survival of hematopoietic and lymphoma colony-forming cells in vivo following the administration of a variety of alkylating agents. Cancer Res., 32, 470-476.

Valeriote, F. and van Putten, L. (1975) Proliferation -dependant cytotoxicity of anticancer agents: A review. Cancer Res., 35, 2619-2630.

Van Dilla, M.A., Trujillo, T.T., Mullaney, P.F. and Coulter, J.R. (1969) Cell microfluorometry: A method for rapid fluorescence measurement. Science, 163, 1213-1214.

van Hazel, G.A. and Kovach, J.S. (1982) Pharmacokinetics of mitomycin C in rabbit and human. Cancer Chemother. Pharmacol., 8, 189-192.

van Maanen, J.M.S., deVries, J., van Oort, W.J. and Pinedo, H.M.(1982) VIII 7d. Studies on the metabolism and mechanism of action of antitumor podophyllotoxin derivative VP16-213. Neth. Cancer Inst. Ann. Rep., 162-164.

Veenema, R.J., Dean, A.L., Roberts, M., Fingerhut, B., Chowhury,B.K. and Torassoly, H. (1962) Bladder carcinoma treated by direct instillation of thio-tepa. J. Urol., 88, 60-63.

Venditti, J.M. (1983) The National Cancer Institute antitumor drug discovery program, current and future perspectives: A commentary.Cancer Treat. Rep., 67, 767-772.

Volm, M., Wayss, K., Kaufman, M. and Mattern, J. (1979) Pretherapeutic detection of tumour resistance and the results of tumour chemotherapy. Eur. J. Cancer, 15, 983-993.

Von Hoff, D.D. and We'isenthal, L.M. (1980) In vitro methods to predict patient response to chemotherapy. Adv. Pharmacol. Chemother., 7, 133-156.

Von Hoff, D.D. and Clark, G.M. (1984) Drug sensitivity of primary versus metastasis. In: Human Tumor Cloning. Eds. S.E. Salmon and J.M. Trent. Grune and Stratton, Inc., ppl83-196.

Von Hoff, D.D., Rozencweig, M. and Muggia, F.M. (1979) The evaluation of cytotoxic drugs. Cancer Treat. Rev., 6, Suppl., 1-8.

Von Hoff, D.D., Cowen, J., Harris, G. and Reisdorf, G. (1981)Human tumor cloning: Feasibility and clinical correlations. Cancer Chemother. Pharmacol., 6, 265-271.

Wakaki, S., Marumo, H., Tomioka, K., Shimizu, G., Kato, E., Kamada, H., Kudo, S. and Fujimoto, Y. (1958) Isolation of new fractions of antitumor mitomycins. Antibiot. Chemother., 8, No.5, 228-240.

Walker, L., Walker, M.C., Parris, C.N. and Masters, J.R.W. (1988) Intravesical chemotherapy: Combination with dimethyl sulfoxidedoes not enhance cytotoxicity in vitro. Urol. Res., 16, 329-331.

Walker, M.C., Masters, J.R.W., Parris, C.N., Hepburn, P.J. and English, P.J. (1986) Intravesical chemotherapy: In vitro studieson the relationship between dose and cytotoxicity. Urol. Res., 14, 137-140.

Walker, M.C., Parris, C.N. and Masters, J.R.W. (1987) Differential sensitivities of human testicular and bladder tumor cell lines to chemotherapeutic drugs. J. Natl. Cancer Inst., 79, 213-216.

Walls, G.A. and Twentyman, P.R. (1985) Cloning of human lung cancer cells. Br. J. Cancer, 52, 505-513.

Wang, N., Trend, B., Bronson, D.L. and Fraley, E.E. (1980) Non-random abnormalilities in chromosome 1 in human testicular cancers. Cancer Res., 40, 796-802.

Wang, N., Perkins, K.L., Bronson, D.L. and Fraley, E.E. (1981) Cytotogenetic evidence for premeiotic transformation of human testicular cancers. Cancer Res., 41, 2135-2140.

Warren, J., Sacksteder, M.R., Jarosz, H., Wasserman, B. and Andreiotti, P.E. (1975) Potentiation of antineoplastic compounds by oral dimethyl sulfoxide in tumour-bearing rats. Ann. N.Y. Acad. Sci., 243 194-208.

Warters, R.L. and Hofer, K.G. (1974) The in vivo reproductive potential of density separated cells. Expt Cell Res., 87, 143-151.

Watson, P.G. and Dawson, M.P.A. (1977) Tissue culturing of malignant melanomas in the presence of cytotoxic agents. Trans. Opthalmol. Soc. 97, 440-441.

Waymouth, C. (1974) To disaggregate or not to disaggregate, injury and cell disaggregation transient or permanent. In vitro, 10,Nos. 1 and 2, 97-111.

Webb, J.N. (1984) The pathology of bladder cancer. In: Bladder Cancer (Butterworths International Medical Reviews, Urology 1). Eds., P.H. Smith and G.R. Prout. Butterworths and Co. (Publishers) Ltd., London, ppl04-124.

Weisenthal, L.M. (1981) In vitro assays in preclinical antineoplastic drug screening. Semin. Oncol., 8, 362-376.

Weisenthal, L.M. and Lippman, M.E. (1985) Clonogenic and nonclonogenic in vitro chemosensitivity assays. Cancer Treat.Rep., 69, No.6, 615-632.

Weisenthal, L.M., Marsden, J.A., Dill, P.L. and Macaluso, C.K. (1983a) A novel dye exclusion method for testing in vitro chemosensitivity of human tumors. Cancer Res., 43, 749-757.

Weisenthal, L.M., Dill, P.L., Kurnick, N.B. and Lippman, M.E. (1983b) Comparison of dye exclusion assays with a clonogenic assay in the determination of drug-induced cytotoxicity. Cancer Res.,43, 258-264.

Weizsaecker, M. and Deen, D.F. (1980) Effect of feeder cell-released substances on the survival of clonogenic 9L cells after treatment with antimetabolites. Cancer Res.,40, 3202-3205.

Welch, D.R. and Nicolson, G.L. (1983) Phenotypic drift and heterogeneity in response of metastatic mammary adenocarcinoma cell clones to adriamycin, 5-fluoro-2'-deoxyuridine and methotrexate treatment in vitro. Clin. Exp. Metatasis, 1, No.4, 317-325.

Werrlein, R.J. and Glinos, A.D. (1974) Oxygen microenvironment and respiratory oscillations in cultured mammalian cells. Nature, 251, 317-319.

Wheeler, K.T., Tel, N., Williams, M.E., Sheppard, S., Levin, V.A. and Kabra, P.M. (1975) Factors influencing the survival of rat brain tumor cells after in vitro treatment with1,3-Bis(2-Chloroethyl)-l-nitrosourea. Cancer Res., 35, 1464-1469.

Whelan, R.D.H. and Hill, B.T. (1981) The influence of agarose concentration on the cloning efficiency of a series of established human cell lines. Cell Biol. Int. Rep., 5, No.12, 1137-1142.

Whitemore, W.F., Batata, M.A., Ghoneim, M.A., Grabstald, H. and Unal, A. (1977) Radical cystectomy with or without prior irradiation in the treatment of bladder cancer. J. Urol., 118, 184-187.

Wibe, E. (1980) Resistance to vincristine of human cells grown as muticellular spheroids. Br. J. Cancer, 42, 937-941.

Williams, R.D. (1980) Human urologic cancer cell lines. Invest. Urol., 17, 359-363.

Williams W.M., Chen, T.S. and Huang, K.C. (1984) Effect of penicillin on the renal tubular secretion of methotrexate in the monkey. Cancer Res., 44, 1913-1917.

Wilson, W.L. (1960) Chemotherapy of human solid tumors with 5-fluorouracil. Cancer, 13, 1230-1239.

Wittes, R.E. (1986) Adjuvant chemotherapy-clinical trials and labortory models. Cancer Treat. Rep., 70, No.l, 87-103.

Wolberg, W.H. (1972) Response of DNA thymidine synthesis in human tumor and normal tissue to 5-fluorouracil. Cancer Res., 32,130-132.

Woodruff, M.F.A. (1983) Cellular heterogeneity in tumours. Br. J Cancer, 47, 589-594.

Woolley, P.V., Maguire, P., Fox, P., Hoth, D., Slavik, M. and Schein, P. (1979) Studies with the antitumor agent spirogermanium Proc. Am. Ass. Cancer Res., 20, 436.

Wright, J.C., Cobb, J.P., Gumport, S.L., Golomb, F.M. and Safadi, D. (1957) Investigation of the relation between clinical and tissue-culture response to chemotherapeutic agents on human cancer

Wright, J.C., Cobb, J.P., Gumport, S.L., Safadi, D., Walker, D.G. and Golomb, F.M. (1962) Further investigation of the relation between the clinical and tissue culture response to chemotherapeutic agents on human cancer. Cancer, 15, No.2, 284-293.

Wright, W.C., Daniels, W.P. and Fogh, J. (1979) Stability of polymorphic enzyme phenotypes in human tumor cell lines. Proc. Soc. Exp. Biol. Med., 162, 503-509.

Wright, W.C., Daniels, W.P. and Fogh, J. (1981) Distinction of seventy-one cultured human tumor cell lines by polymorphic enzyme analysis. J. Natl. Cancer Inst., 66, 239-247.

Wu, P-C, Ozols, R.F., Hatanka, M. and Boone, C.W. (1982)Anticancer drugs: effect on the cloning of Raji lymphoma cells in soft agar. J. Natl. Cancer Inst., 68, 115-121.

Yagoda, A. (1980) Chemotherapy of metastatic bladder cancer. Cancer, 45, 1879-1888.

Yagoda, A., Watson, R.C., Gonzalez-Vitale, J.C., Grabstald, H. and Whitmore, W. (1976) Cis-dichlorodiammine platinum II in advanced bladder cancer. Cancer Treat. Rep., 60, 917-923.

Yalow, R.S., Eastridge, C.E., Higgins, G. and Wolf, J. (1979) Plasma and tumor ACTH in carcinoma of the lung. Cancer, 44, Part 5, 1789-1792.

Yang, S-J. and Rafla, S. (1983) Effect of spirogermanium on V79 Chinese hamster cells. Am. J. Clin. Oncol., 5, 331-337.

Yang, L-Y. and Drewinko, B. (1983) The stability of the lethal efficacy of antitumor drugs. Proc. Am. Assoc. Cancer Res., 24, 315.

Yang, L-Y. and Drewinko, B. (1985) Cytotoxic efficacy of reconstituted and stored antitumor agents. Cancer Res., 45, 1511-1515.

Yeh, J. and Fisher, H.W. (1969) A diffusible factor which sustains contact inhibition of replication. J. Cell Biol., 40, 382-388.

Yoshida, A. and Beutler, E. (1983) G-6-PD variants: another update. Ann. Hum. Genet., 47, 25-38.

Yuhas, J.M., Toya, R.E. and Pazmino, N.H. (1974) Neuraminidase and cell viability: Failure to detect cytotoxic effects withdye-exclusion techniques. J. Natl. Cancer Inst., 53, 465-468.

Yung, W-K. A., Shapiro, J.R. and Shapiro, W.R. (1982)Heterogeneous chemosensitivities of subpopulations of human glioma cells in culture. Cancer Res., 42, 992-998.

Zietz, S. and Nicolini, C. (1978) Flow microfluorometry and cell kinetics: a review. In: Biomathematics and Cell Kinetics. Eds. A.J. Valleron and P.D.M. MacDonald. Elsevier/North-Holland Biomedical Press, pp357-394.

Zucker, R.M. and Helfman, D. (1976) Electronic volume analysis of L1210 chemotherapy. Cancer Res., 36, 4434-4438.