application of in vivo and in vitro pharmacokinetics for ...drug in different solvents, depending...

[CANCER RESEARCH 47, 3718-3724, July 15, 1987)

Application of in Vivo and in Vitro Pharmacokinetics for Physiologically RelevantDrug Exposure in a Human Tumor Clonogenic Cell Assay1

Francis Mi-Osman, ' Jane Giblin, Dolores Dougherty, and Mark L. Rosenblum

Northwest Neuro-Oncology Research Laboratory, Department of Neurological Surgery, University of Washington, Seattle, Washington 98195 [F. A-OJ, and the BrainTumor Research Center, Department of Neurological Surgery, University of California, San Francisco, California 94143 [J. G., D. D., M. L. RJ

ABSTRACT

The biological half-lives and decay rate constants under the conditionsof a human brain tumor clonogenic cell assay were determined for sixclinically used anticancer agents. The agents studied were: l,3-bis(2-chloroethylH-nitrosourea; 3-(2-chloroethyl-3-nitrosoureido-2-deoxy-o-glucopyranose; c/j-diaminedichloroplatinum(II); 2,5-diaziridinyl-3,6-bis-(carboethoxyamino)-l,4-benzoquinone; 4-demethylepipodophylotoxin-D-thylidene glucoside; and 9-hydroxy-2-./V-methylellipticine. In vitro decay of all six drugs was found to be according to first order kinetics. Thehalf-lives of two drugs, namely, l,3-bis(2-chloroethyl-l-nitrosourea and3-{2-chloroethyl-3-nitrosoureido-2-deoxy-D-glucopyranose under the human tumor clonogenic cell assay (HTCA) conditions were found to besimilar to their terminal in vivo half-lives in humans. For the other drugs,however, there was a very large difference between their in vitro and invivo pharmacokinetics. In the case of 2,5-diaziridinyl-3,6-bis(car-boethoxyamine)-l,4-benzoquinone, we observed about an 80-fold difference between its in vitro half-life of 40.76 h and its in vivo terminal half-life of 0.52 h. We describe the principles upon which these data can beused to design clinically more relevant in vitro drug exposure protocolsin HTCAs. Since, generally, tumor cells are exposed to drugs in theHTCA either continuously or for a specified duration, e.g., 1 or 2 h, wecomputed the initial in vitro drug concentrations to which tumor cellsshould be exposed such that the resulting in vitro (c x t) after a 2-h ora continuous exposure will be within clinically achievable levels. Theapplication of these in vivo and in vitro pharmacokinetic principles willprovide for more physiological testing of patient tumor cell sensitivity toanticancer drugs in the HTCA, and is likely to result in lower rates offalse positive responses in clinical trials using clonogenic cell assays.

INTRODUCTION

HTCAs3 (1-3) have been used to study tumor biology (4), todetermine the sensitivities of tumors of patients to chemother-apeutic agents (5-10), and more recently to investigate thepotential activity of new anticancer drugs (11). The optimumdesign of drug exposure protocols for performing clinicallyrelevant drug sensitivity studies utilizing HTCAs, however,requires not only knowledge of the clinical pharmacology andpharmacokinetics of the agents being tested, but also information on drug stability and cell kill kinetics under the HTCAconditions. Such information is necessary to ensure that tumorcells are exposed to drugs at concentrations and for durationsthat are not only clinically relevant, but also are appropriatewith the pharmacology, cell cycle phase specificity, and routeof administration of the given drugs. Because data on intratu-

Received 11/7/86; revised 3/30/87; accepted 4/24/87.The costs of publication of this article were defrayed in pan by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1Supported in part by NIH Grants Ã‡A13S2S and CA 31882.In whom requests for reprints should be addressed, at Department of

Neurological Surgery (RI-20), University of Washington, Seattle, WA 98195.3The abbreviations used are: HTCA, human tumor clonogenic cell assay;

MEM, minimum essential medium; PCS, fetal calf serum; EBS, Earle's balancedsalts; BCNU (carmustine), l,3-bis(2-chloroethyl)-l-nitrosourea; CHZ (chlorozo-tocin), 3-(2-chloroethyl-3-nitrosoureido-2-deoxy-D-glucopyranose; AZQ (diazi-quone), (2,5-diaziridinyl-3,6-bis(carboethoxyamino)-l,4-benzoquinone; VM 26(teniposide), 4-demethyl-epipodophylotoxin-D-thylidene glucoside; HME (ellip-tinium), 9-hydroxy-2-/V-methylellipticine; Cis-Pt (cisplatin), os-diaminedichloroplatinum(II); IND, Investigational New Drug; HPLC, high performance liquidchromatography; Â¡.t.,intratumoral; Â¡.a.,intraarterial.

moral drug concentrations and drug decay kinetics are generallynot available the clinical pharmacokinetic parameters mostfrequently used to determine drug exposures in clonogenic cellassays have been the drug peak plasma concentration, the drugplasma terminal half-life, and the drug plasma (c x r), representing the integral of plasma drug levels over time. However,in order to use these data to compute in vitro drug exposuretimes and drug concentrations such that the resulting in vitro(c x / ) would be within clinically achievable values, one needsto know both the pharmacokinetic model that fits the dynamicsof i/i vitro drug concentration changes over time and the drugdecay rate constants under the conditions of exposure of thecells to the drug in the assay. Unfortunately, there is only littlequantitative data (12, 13) available on the chemical and/orbiological stability of chemotherapeutic agents under HTCAconditions.

In this study, we used a modification of a previously describedbioassay (41, 42) to determine the in vitro stabilities of sixantitumor agents, as reflected by their in vitro biological decayrate constants and half-lives under the conditions of a braintumor clonogenic cell assay (14). The initial drug concentrations used in the bioassay were those concentrations whichdemonstrated a significant in vitro cell kill in a cytotoxicityprescreen using the 9L rat gliosarcoma cell line.

The results show that while for the nitrosoureas in vitro half-lives under the clonogenic cell assay conditions were within theranges of the terminal half-lives observed in clinical pharmacokinetic studies, for the other drugs there was a very significantdifference between their in vivo and in vitro half-lives. Based onour data, we make suggestions as to in vitro drug concentrationswhich in (a) a 2-h treatment, and (b) a continuous exposure(greater than 5 in vitro half-lives) will result in drug exposureswithin clinically achievable levels. Using the principles described in this paper and the drug decay data we present, otherin vitro drug concentrations and exposure durations can becomputed with which more clinically relevant in vitro drugexposures can be achieved for different methods of clinical drugadministration, e.g., intraarterial or i.p.

MATERIALS AND METHODS

9L Gliosarcoma. The 9L tumor cell line is an A'-methylnitrosourea-

induced gliosarcoma of the Fisher 344 rat. It produces solid braintumors in rats within 2 weeks of intracerebral injection (IS). The cellsused for this study were maintained by weekly in vitro passages ofmonolayer cultures in MEM supplemented with EBS and 15% FCS.

Human Glioma SF 126. The human glioma cell line SF 126 wasoriginated in our laboratory from a fresh tumor biopsy obtained bycraniotomy from a patient bearing a grade IV astrocytoma. The cellswere cryopreserved in MEM containing 10% dimethyl sulfoxide and20% FCS. Prior to use in these studies, thawed cells were washed twicewith medium and cultured as monolayers in T75 flasks containingMEM, EBS, and 15% FCS. The cells were characterized by theirpositive reactivity to a monoclonal antibody, GE 2, which recognizes aglioma-associated antigen (16), using previously described methods(17).

3718

on March 16, 2021. © 1987 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

PHARMACOKINETICS FOR DRUG EXPOSURE IN HTCA

All cultures used in the experiments were screened for and shown tobe free of Mycoplasma contamination.

Drugs. The drugs studied were provided by the following suppliers:BCNU and Cis-Pt (Bristol Myers, Syracuse, NY); CHZ, VM 26, andAZQ (Drug Development Branch, National Cancer Institute, Bethesda,MD); I IMI (Sanofi-Labaz, Bordeaux, France, and kindly made available for these studies by Dr. Victor Levin).

Stock solutions were prepared by dissolving the weighed amounts ofdrug in different solvents, depending upon the solubility of the drug.Subsequent dilutions were made with Hanks' balanced salt solution.

All drug solutions were prepared just before use and were kept on iceprior to addition to the treatment flasks. The solvents used and thehighest final concentrations in culture were as follows: BCNU (ethanol;1:4,000). CHZ (dimethyl sulfoxide; 1:4,000). AZQ (A^-dimethyla-cetamide; 1:10,000). HME acetate (H2O; 1:10,000); and Cis-Pt (MEM;1:10,000). VM 26 was dissolved at 10 mg/ml in a solvent consisting of150 mg benzyl alcohol, 300 mg NJV-dimethylacetamide, 2.5 g purifiedpolyethoxylated castor oil, 300 mg TVyV-dimethylacetamide, sufficientmaleic acid to achieve a pH of 5.1, and 4.715 g absolute ethanol. Thefinal solvent concentration in culture was 1:100,000.

For each experiment a set of control flasks were set up which weretreated with solvent only at the concentrations stated above.

Cytotoxicity Prescreen. In order to determine the initial drug concentrations at which to expose cells for the bioassay, a cytotoxicity studywas performed for each drug by using the 9L gliosarcoma cell line.Peak plasma drug concentrations as reported in clinical studies wereused to determine the range of concentrations for this initial prescreening study which was performed as described below. Confluent mono-

layer cultures were trypsinized with 0.25% trypsin in 0.4% EDTAsolution in Ca2+/Mg2+-free phosphate-buffered saline and seeded in

triplicate T25 tissue culture flasks at cell numbers ranging between 50and 5000 cells (depending on the expected cell kill). MEM supplemented with EBS, 20% PCS, and 5x10" X-irradiated (4000 rads) 9Lcells were added, and the flasks were incubated at 37'C for 24 h in a

humidified atmosphere of 5% CO2/air. Fifty Â¡Aof stock solutions ofeach drug, prepared as previously described, were added to each set offlasks, the flasks were incubated for a further 24 h, and then rinsed andrefed with fresh MEM containing 20% FCS. After 2 weeks at 37'C

and 5% CO2/air, colonies were stained with a 3% crystal violet/methanol solution and counted under low power microscopy.

Bioassay. The dose at which a 1-3-log clonogenic cell kill of the 9Ltumor cell line was achieved in the prescreen was used as the initialconcentration in the bioassay, which was performed as follows: a set of4 T75 tissue culture flasks were set up, each containing 1.5 x 10s SF126 human glioma cells and 1.5 x 10' heavily irradiated 9L cells inMEM containing 20% FCS. The flasks were incubated at 37Â°Cand 5%

CO2 in a humidified incubator for 24 h to allow the cells to attach tothe culture flask surface. Stock drug solution was then added to eachflask to yield the predetermined initial drug concentration. At varioustime points, triplicate 1.5-ml aliquots of the supernatant were removedand immediately filtered through 0.4 ,/m sterile cellulose acetate filters(Millipore, Bedford, MA). The first 0.5 ml was discarded and theremaining 1 ml was added to T25 tissue culture flasks, each containingthe following: 1()â€¢'io4 9L cells (depending upon the expected cell kill),

5 x IO4 heavily X-irradiated 9L cells, and MEM supplemented with

20% FCS. The time points at which the drug solutions were aliquotedwere determined in pilot studies to be 0-60 min (BCNU, CHZ, andVM 26); 0-9 h (Cis-Pt); and 0-24 h (AZQ and HME). After 24 h at37'C and 5% CO2, these second sets of flasks were rinsed twice, fresh

MEM containing 20% FCS was added to them, and they were incubatedas previously described. Colonies were stained and counted after 2weeks.

Simultaneously, a dose-response study was performed as describedin the prescreening experiments, but using ranges of drug concentrations which included the initial concentrations used in the bioassaydescribed above.

Evaluation of Bioassay Survival Curves. For each drug, the survivingfraction, S.F. (/ ) of 9L gliosarcoma clonogenic cells at each treatmenttime point, was computed by using the equation

S.F.(f) = CFE(i)/CFE(0) (A)

whereby CFE(f) is the colony-forming efficiency of cells treated withsupernatant removed at time t and CFE(O) is the colony-formingefficiency of untreated control cells. CFE is defined as:

CFE = No. of colonies formed/no, of cells seeded (B)

The surviving fractions were plotted against the time points at whichthe supernatants were removed. Each S.F.(f) was subsequently converted to the corresponding drug concentration, C(t) by reading offfrom the simultaneously determined dose-response curve, the drugconcentration necessary to achieve an equivalent reduction in survivalin the 9L clonogenic system. The in vitro decay rate constant, A. wasthen calculated by using the equation for first order kinetics.

-dc/dt = kC(0)

which upon integration gives:

C(t) =

(C)

(D)

where C(Ã)is the drug concentration remaining in the flask after timei and ('(()) is the initial drug concentration.

Rearranging Equation D and taking logs gives

and at

k = ln[C(0)/C(/)]/Ã

rÂ«= 0.693/A

(E)

(F)

k was computed by linear regression analysis of the concentration-timecurve obtained for each drug, and using Equation F, the half-life i ofthe drug under the treatment conditions was computed.

Computation of Concentrations for Clinically Relevant Drug Exposure.In order to determine the initial concentration to which tumor cellsshould be exposed such that the resulting in vitro (c x n is equivalentto the clinically achievable (c x t), the trapezoidal approximation (18)was used:

(C X i) = '/!(Co + C,)(f, - ÃŒo)

C2)(/2 - - f.-,)(G)

where Co, C\, C2 ... C, are the drug concentrations remaining inculture after exposure times of to, t\, i2 ... iÂ»,respectively. Since wehad previously determined the in vitro decay rate constant k, it ispossible to compute the remaining drug concentration, ('(/). in culture

at any given time, t, by using Equation D.Substituting the values in Equation G allows calculation of <"Â«)).the

initial drug concentration required to achieve the clinical (c x /) afterin vitro exposure of the tumor cells for a given time, e.g., 2 h. For acontinuous drug exposure (t = Â»),the approximation used for calculating C(0) is as follows:

and rearranging gives

(c x f) = C(0)A

C(0) = Me x t)

(H)

(I)

where k is the in vitro decay rate constant of the drug and C(0) is theinitial concentration to which tumor cells would be exposed.

RESULTS

Cytotoxicity Prescreen. Fig. 1 illustrates the survival curveobtained with m-platinum in the prescreening study performedon the 9L gliosarcoma cell line. This representative curve istypical of the survival curves for the other five drugs. The resultsobtained from the survival curves for all six drugs are summa-

3719




123Cisplotinum Dose

Fig. 1. Survival curve of 9L gliosarcoma cells to cu-platinum in the in vitropre-screening assay (see "Materials and Methods"). This curve is representative

of the curves for the other five drugs studied.

Table 1 Prescreen cytotoxicily data on 9L gliosarcoma cell line

DrugAZQ

Cis-PtVM26BCNUCHZHMEIn

vitroconcentration(MM)1.0

3.00.15

30.10

10Surviving

fractionof 9Lcells1

x 10-3x 10-1 x 10-4x 10-

2.7 x 10-2x 10-

rized in Table 1. The drug concentrations selected as startingconcentrations for the subsequent drug-decay studies were thoseconcentrations that gave a 1-3-log reduction in clonogenic cellsurvival of 9L cells, depending upon the sensitivity of this cellline to the individual drugs.

In Vitro Drug Decay Kinetics. The biological in vitro decaycurves of all six drugs are shown in Fig. 2 and were found to Hifirst order kinetics as described by Equations C and D. Thehalf-lives, t>/,,and decay rate constants, k, were calculated foreach drug by using linear regression analysis of the corresponding drug-decay curve. The results of these analyses are summarized in Table 2. Of the six drugs, HME. AZQ, and VM 26were biologically the most stable under the in vitro clonogeniccell assay conditions, with decay rate constants of 0.008 h~',0.017 h'1, and half-lives of 80 h, 40.76 h, and 36.47 h, respec

tively. For HME an initial drop in concentration of approximately 30% was observed. In a previous study (17) using an11IMC technique, a similar rapid drop in HME concentrationwas observed in vitro and was shown to be the result of adsorb-ance of the drug to the tissue culture flask and not due to actualdrug decay. The relevant clinical pharmacokinetic data for allthe drugs as published in various reports are given in Table 3.

Fig. 3 shows the results of the standard dose-response studythat was performed simultaneously with the drug-decay study.The theoretical concentrations were computed, based upon theamounts of drug weighed and the volume of solvent used tomake the stock solutions. The experimental drug concentrations were those that were read off from the standard dose-

response curves, and were based upon the level of 9L clonogeniccell kill achieved at those drug concentrations. As Fig. 3 shows,there was a good correlation (r = 0.986) between the theoreticaland the experimental drug concentrations determined by bumssay. The highest variation in the two concentrations was 30%for HME. As mentioned earlier, however, this discrepancybetween the two HME concentrations was probably mainly aresult of drug adsorption onto the surface of the tissue cultureflask. Correcting for this adsorption, a variation of 9% wascomputed between theoretical and experimental HME concentrations.

Drug Concentrations for Clinically Relevant in Vitro DrugExposure. Table 4 gives the computed drug concentrationsrequired to achieve clinical (c X t) levels in the clonogenic cellassay after (a) a 2-h drug exposure, and (/>) a continuousexposure. For BCNU and CHZ these two concentrations werevery similar and were 25.2 and 24.6 /Â¿M(BCNU), and 30.2 and29.1 MM(CHZ), respectively. In contrast, for HME, AZQ, Cis-Pt, and VM 26, there were very wide differences in the in vitroconcentrations required to achieve clinically relevant (c x t)levels under the two exposure conditions stated above.

DISCUSSION

Several reports (5-10) have documented, both retrospectivelyand prospectively, that human tumor clonogenic cell assays arecapable of predicting the response of an individual patient tocancer chemotherapy. For previously untreated ovarian cancerpatients with good performance status, the assay has beenreported to perform better than the best choice of chemotherapyby the physician (6, 9). Generally, however, these studies haveshown that clinical response correlates better with in vitroresistance (greater than 90%) than for sensitivity (50-70%).Significant among the assay-related factors that could accountfor these relatively low levels of positive in vivo-in vitro correlation rates are the design of the in vitro drug exposure protocols, the method of determining /"//vitro sensitivity indices (19),

and the criteria for establishing the cut-off index for in vitrosensitivity to be used in such correlations (20). Although Albertset al. (21) have reviewed the application of pharmacologicalprinciples in HTCAs and have proposed the concept of a "cut

off concentration for determining in vitro drug sensitivity, ithas not been possible, in most cases, to apply these principlesin an optimized fashion to HTCAs. This is primarily becauselittle (12, 13) and for some drugs, no quantitative data on invitro drug-decay kinetics under HTCA conditions are available.In the absence of such information, most drug sensitivity studieswith the HTCA have been performed simply by exposing tumorcells to peak plasma drug concentrations, or a fraction (5-10%)of it, continuously or for 1-2 h.

In this report, a bioassay was used to investigate the phar-macokinetics of six clinically used anticancer agents under theconditions of a human brain tumor clonogenic cell assay. Forall the drugs investigated, in vitro decay was according to firstorder kinetics. Similar first order kinetics have been reportedfor the in vitro decay of both m-platinum (22) and BCNU (23)in human plasma. Under the HTCA conditions used in thispresent study, the two nitrosoureas, BCNU and CHZ, demonstrated in vitro half-lives of 0.54 and 0.52 h, respectively. Thesevalues are relatively close to their respective in vivo terminalhalf-lives of 0.36 and 0.4 h. as determined in human clinicalphase I trials (23, 24). Furthermore, for these two drugs,because of their relatively high in vitro decay-rate constants of1.28 h-' (BCNU) and 1.33 tr1 (CHZ) and correspondingly

3720




1001 100

BCNU

I10

10

CHZ

20 40 60 O 20 40 60

Time (Mins) Time (Mins)

100

Fig. 2. In vitro decay of BCNU, CHZ, VM26, Cis-Pt, HME, and AZQ under the conditions of the brain tumor clonogenic cell assay.All six log-linear plots show pseudo-first orderkinetic decay under the assay conditions.

10

10

VM26

20 40

Time (Mins)

60

Cis-Pt

0 2 4 6 8 10

Time (Hrs)

10

ITÂ»â€”ZLUÃœ

HME

Â§5

AZQ

6 12 18

TIME (HRS)

24 6 12 18

Time (Hrs)

24

short in vitro half-lives, //; vitro drug concentrations required toachieve clinically relevant (c x t) levels were similar for both a2 h and a continuous drug exposure. In contrast, for AZQ, Cis-Pt, VM 26, and HME, we showed in this study that widedifferences existed between their decay kinetics in the HTCAand their published human clinical plasma pharmacokinetics(17, 23-29). The in vitro half-lives of all four drugs were much

longer than their in vivo terminal half-lives. For example, Curtet al. (26) reported an in vivo plasma terminal half-life of 0.52h for AZQ compared with a half-life of 40.76 h we measuredunder our HTCA conditions. The terminal plasma disappearance rate of AZQ is therefore, almost 80 times faster than thebiological decay rate of the drug in the HTCA. This long half-life of AZQ determined by bioassay may reflect, in part, the

3721




Table 2 In vitro pharmacokinetic data as determined under the tumor clonogeniccell assay conditions

DrugAZQCis-Pt

VM26BCNUCHZHMEInitial

concentration(fM)1.03.0

0.15303010Half-life

(h)40.8

15.136.5

0.540.52

86.3Decay

rateconstant

(h-1)0.017

1.0461.0191.281.330.008

Table 3 Clinical pharmacokinetic data of drugsAll drugs were administered by the i.v. route.

DrugAZQ

Cis-PtVM26BCNUCHZHMETreatment

dose andschedule20

mg/m2100 mg/m267 mg/m2

220 mg/m2120 mg/m2

2 mg/kgMean

peakplasma

concentration ((/M)1.9

8.336.3

9.2127.8

2.5Termi

nalhalf-life

(h)0.52

15.0724.75

0.360.44.4(cxr)

(dM-h)10.4

6.5173.6

19.221.9

3.3Reference2627

28232429

|

I

40

30

20

10-

0 10 20 30 40

Theoretical Drug Concn. (\M)

Fig. 3. Correlation between drug concentrations determined by bioassay andtheoretically computed values based upon amount of drug used to prepare stocksolutions, r = 0.986.

Table 4 In vitro drug concentrations to which cells should be exposed (a) for 2 h,and (b) continuously in order to obtain clinically achievable (c x t) levels

C(0) (Â»IM)to achieve clinical (c x /)

DrugAZQCis-PtVM26BCNUCHZHME2-hexposure5.33.590.425.230.21.58Continuousexposure0.180.33.324.629.10.026

fact that hydrolysis of AZQ can result in opening of the aziri-dinyl ring to yield potentially cytotoxic intermediates (30).However, we believe this to occur minimally in our system atthe chloride, phosphate, and glucose concentrations. At the pHof 7.2 in the assay, the rate of such hydrolytic ring opening willbe relatively low; the tw for AZQ at pH 7.2 is about 80 h (30).Furthermore, if two or more major cytotoxic effects (due toparent drug and degradation products) were being assayed, thenthe resulting decay curves will be bi- or multiphasic, in contrastto the first order decay kinetics we observed for AZQ and theother drugs. Thus, for AZQ, HME, Cis-Pt, VM 26, and other

3722

agents with half-lives that are much longer in vitro than in vivo,exposure of cells to peak plasma drug concentrations (or even5 or 10% of it) will result in in vitro (c x t) levels that are manytimes higher than those achievable clinically. As a result, thereis an increased probability of a false positive response and asimultaneously higher probability of a true negative response,as has been consistently observed in in vivo-in vitro studies thathave been performed in this manner. The findings in this reportof extreme differences in the /// vitro and in vivo pharmacoki-netics of some anticancer drug suggest the need for similarstudies to be performed for other agents that are tested in theHTCA for clinical purposes. However, it is important to recognize that the biological assay used in the present study, unlikethe HPLC assays generally used in in vivo pharmacokineticstudies, does not necessarily always measure the parent drug.Indeed, in some cases one may be measuring both the drug andits intermediates if both of these have cytotoxic activity. However, for some agents, e.g., the chloroethylnitrosoureas (31), theconcentration of the drug degradation product is proportionalto that of the parent drug, and decay of the drug intermediateis the rate-limiting step in the decay of the parent drug. Theoretically, in such situations, as of course is also the case for adrug whose cytotoxic action is due to the parent drug directly,a bioassay will provide for a true measure of the parent drugdecay. Furthermore, first order kinetics, as we observed for allsix drugs, strongly suggests the existence of only one majorcytotoxic species. If more than one major cytotoxic species wereto be present, then because of the expected different decay rateconstants of these species, a bi- or multiphasic decay curvewould be obtained.

The good correlation, r = 0.986, we observed between thetheoretical and the bioassay-determined drug concentrationssuggests that for the six drugs, the bioassay as described hereallows for reasonably accurate determination of drug concentrations, and may be used in the absence of more specific assaysrequiring HPLC and other more complex techniques. However,a study comparing drug decay as measured by HPLC and bybioassay needs to be performed to confirm this for individualdrugs.

Because of the differences in complexity of different HTCAtechniques, differences are likely to exist in drug-decay kineticsunder the different assay conditions. Such assay-specific variation in in vitro drug pharmacokinetics may require that drugdecay be determined for each HTCA technique and the valuesobtained be used to optimize drug exposure in that assay. Thisproblem may become less significant as the HTCAs becomebetter optimized and standardized.

INDs pose an additional problem since there are usually nohuman pharmacokinetic data available for these agents at thetime of initial screening. In the National Cancer Institute-sponsored multilaboratory study on the application of theHTCA to IND screening (11), tumor cells were exposed todrugs continuously at a single arbitrary concentration of 10 Â¿Â¿g/ml. Active agents were then subjected to a dose-response studyin order to better define their 70% infective dose values. Thisapproach did not address any potential differences in the invitro stability of the drugs. It may be advantageous even at thisearly stage of IND screening, using a simple bioassay such aswe describe here, to determine the in vitro decay kinetics ofagents that may be going into clinical trials. The data thusobtained could be used with clinical pharmacokinetic information that will be acquired during phase I clinical trials of activeagents to better design further clinical trials of the agents withthe HTCA.




ACKNOWLEDGMENTS

The authors wish to express their sincere thanks to Dr. Victor Levinfor making available to us unpublished human BCNU pharmacokineticdata and to Karin McDougall for typing of this manuscript.

REFERENCES

9.

10.

12.

13.

In this present report, the clinical pharmacokinetic data usedto perform the computations were all derived from publishedclinical phases I and II trials, in which drug administration wasdone i.v. The peak plasma drug levels, the achievable (c x t),and the other pharmacokinetic constants thus obtained willconsequently be different from those that will be obtained withother forms of drug administrations. For example, i.t., i.p.. andi.a. drug administrations will generally result in higher i.t. druglevels than after i.v. administration (32-38). Similarly, high 8-

dose chemotherapy, which is becoming more frequent in someclinical situations, will result in drug peak plasma and (c x nlevels that are higher than those obtained by using more conventional lower dose chemotherapy protocols. However, if theclinical and //; vitro drug pharmacokinetics are known, theprinciples described in this study can be applied to ensure thedesign of more clinically relevant in vitro drug exposure protocols for these various clinical situations. Another important "

factor that must be considered when designing //; vitro drugexposure of cells as proposed in this paper is the cell cyclephase specificity of the drugs to be tested, kinetic-ally, human

tumors and consequently the cell suspensions prepared fromthem are asynchronous and only a fraction of the cell populationwill be in a given cell cycle phase at a given time (39). Consequently, exposure of the cells to agents such as vincristine,hydroxyurea, or l-/3-o-arabinofuranosylcytosine, which are cell 14

cycle phase specific in their mechanisms of action, will have tobe for a duration of at least one cell cycle for any potential is.cytotoxic effect of the drugs to be observed. However, sincegenerally cell cycle kinetic data are not available at the time of H,tumor specimen acquisition, in vitro drug exposure durationscorresponding to 4 of 5 times the in vivo terminal half-life of , 7the drug in human subjects will provide for a theoreticallyacceptable approximation. Using the methods described in thispaper, the in vitro drug concentration required to achieve clin- )8ically relevant (c x t) levels for the drug can then be easilycalculated. l9

The important role played by the pharmacology and especially the pharmacokinetics of a given drug in its therapeutic 20effectiveness requires that clinical in vitro-in vivo trials with theHTCA need to be designed based on sound knowledge of 21.differences in drug pharmacokinetics in the assay and in patients. To our knowledge, to date no retrospective or prospective 22,clinical trial with the HTCA has been performed in which invitro drug exposure was planned by using information on thein vitro pharmacokinetics of the drugs. We speculate that for 23.the reasons discussed in this paper, such inclusion of //; vitrodrug decay data in designing drug exposure protocols will 24improve the accuracy of positive correlations in clinical trialswith the HTCA. 25.

26.

27.

28.1. Hamburger, A. W., and Salmon, S. E. Primary bioassay of human tumor

stem cells. Science (Wash. D.C.), 197:461-463, 1977.2. Courtenay, V. I).. and Mills, J. An in vitro colony assay for human tumors 29.

grown in immune-suppressed mice and treated in vivowith cytotoxic agents.Br. J. Cancer, 37: 261-268, 1978.

3. Maurer, H. R., and Ali-Osman, F. Tumor stem cell cloning in agar-containing 30.capillaries. Naturwissenschaften, 68: 381-383, 1981.

4. Salmon, S. E., and Trent, J. M. (eds.). In: Human Tumor Cloning. Section

3723

I: Tumor Biology, pp. 3-149. Orlando, FL: GruÃ±e& Stratton, Inc., 1984.Salmon, S. E., Hamburger, A. W., Soehnlen, B., Durie, B. G., Alberts, D.S., and Moon, T. E. Quantitation of differential sensitivity of human tumorstem cells to anticancer drugs. N. Engl. J. Med., 298: 13-21, 1978.Alberts, D. S., Chen, H-Sg., Salmon, S. E., Surwitt, E. A., Young, L., Moon,T. E., and Meyskens, F. L. Chemotherapy of ovarian cancer directed by thehuman tumor stem cell assay. Cancer Chemother. Pharmacol., 6: 279-285,1981.Bertelsen, C. A., Sondale, V. K., Mann, B. D., Korn, E. L., and Kern, D. H.Chemosensitivity testing of human solid tumors. Cancer (Phila.), 53: 1240-1245, 1984.Park, C. H., Wiernick, P. H., Morrison, F. S., Amare, M., Van Sloten, K.,and Maloney, T. R. Clinical correlations of leukemic clonogenic cell che-mosensitivity assessed by in vitro continuous exposure to drugs. Cancer Res.,43: 2346-2349, 1983.Welander, C. E., Homesley, H. D., and Jobson, V. W. Multiple agentchemotherapy prospectively selected by the human tumor clonogenic assay(HTCA) for advanced ovarian cancers. Int. J. Cell Cloning, /.- 290-291,1983.Von Hoff, D. D., Clark, G. M., Stogdill, B. J., Sarosdy, M. F., O'Brien, M.

T., Casper, J. T., et al. Prospective clinical trial of a human tumor cloningsystem. Cancer Res., 43: 1926-1931, 1983.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. Application of a human colony-forming assay to new drug screening. Cancer Res., 45: 2145-2153, 1985.Ludwig, R., Alberts, D. S., and Salmon, S. E. Schedule dependency and invitro stability of anticancer drugs in the human tumor clonogenic assay.Proceedings of the 13th International Congress Chemotherapy, 224: 101-106, 1983.Hildebrand-Zanki, S. U., and Kern, D. H. A new bioassay for in vitro drugstability. In: S. E. Salmon and J. M. Trent (eds.), Human Tumor Cloning,pp. 451-458, Orlando, FL: GruÃ±e& Stratton, Inc., 1985.Rosenblum, M. L., Vasquez, D. A., Hoshino, T., and Wilson, C. B. Development of a clonogenic cell assay for human brain tumors. Cancer (Phila.),41: 2305-2314, 1978.Barker, M., Hoshino, T. Gurday, et al. Development of an animal braintumor model and its response to therapy with l,3-bis(2-chloroethyl)-l-nitro-sourea. Cancer Res., 33:976-986, 1873.Schnegg, J. F., Diserens, A. C., Carrell, S., Acolla, R. S., and de Tribolet, N.Human glioma-associated antigens detected by monoclonal antibodies. Cancer Res., 41: 1209-1213, 1981.Ali-Osman, F., Rosenblum, M. L., Giannini, D. D., and Levin, V. A.Potential antiglioma activity of 9-hydroxyl-2-Af-methylellipticine as determined by pharmacological and human tumor clonogenic cell studies. CancerRes., 45: 2988-2992, 1985.Gibaldi, M. Biopharmaceutics and Clinical Pharmacokinetics. pp. 315-316.Philadelphia: Lea & Febiger, 1984.Ali-Osman, F., Bier, J., and Maurer, H. R. In vitro cytostatic drug sensitivitytesting in the human tumor stem cell assay: a modified method for thedetermination of the sensitivity index. Tumor Diagn. Ther., 4: 1-6, 1983.Moon, T. E. Quantitative and statistical analysis of the association betweenin vitro and in vivo studies. In: S. E. Salmon (ed.), Cloning of Human TumorStem Cells, pp. 209-221. New York: Alan R. Liss, 1980.Alberts, D. S., Chen, H. S. G., and Salmon, S. E. In vitro drug assay:pharmacologie considerations. ///: S. E. Salmon (ed.). Cloning of HumanTumor Stem Cells, pp. 197-207. New York: Alan R. Liss, Inc., 1980.Repta, A. J., and Long, D. F. Reactions of cisplatinum with human plasmaand plasma fractions. In: A. W. Prestayko, S. T. Crooke, and S. K. Carter(eds.), Cisplatinum, Current Status and New Developments, pp. 285-304.New York: Academic Press, Inc., 1980.Levin, V. A. Clinical pharmacology of the nitrosoureas. In: A. W. Prestayko,S. T. Crooke, L. H. Baker, S. K. Carter, and P. S. Schein (eds.), Nitrosoureas,pp. 171-180. New York: Academic Press, Inc., 1981.Hoth, D., Woolley, P., Green, D., Macdonald, J., and Schein, P. Phase Istudies on chlorozotocin. Clin. Pharmacol. Ther., 23: 712-722, 1978.Alberts, D. S., and Chen, H. S. G. Tabular summary of pharmacokineticparameters relevant to in vitro drug assay. In: S. E. Salmon (ed.). Cloning ofHuman Tumor Stem Cells, pp. 197-207. New York: Alan R. Liss, Inc.,1980.Curt, G. A., Kelley, J. A., Kufta, C. V., Smith, B. H., Kornblith, P. L.,Young, R. C., and Collins, J. M. Phase II and pharmacokinetic study ofaziridinyl benzoquinon [2.5-diaziridinyl-3,6-bis(carboethoxyamino)-l,4-ben-zoquinone, diaziquone, NSC 182986) in high-grade gliomas. Cancer Res.,Â«.-6102-6105, 1983.Patton, T. F., Himmelstein, K. J., Belt, R., Bannister, S. J., Sternson, L. A.,and Repta, A. J. Plasma levels and urinary excretion of filterable platinumspecies following bolus injection and intravenous infusion of cis-bischloro-diam-inoplatinum(II) in man. Cancer Treat. Rep., 62: 1359-1362, 1978.Allen, L. M., and Creaven, P. J. Comparison of the human pharmacokineticsof VM-26 and VP-16, two antineoplastic epipodphyllotoxin glucopyranosidederivatives. Eur. J. Cancer, //: 697-707, 1975.Covette, A., Huertas, D., Drox, J-P., Rouesse, J., and Amiel, J-L. Pharmacokinetics of 2-methyl-hydroxyllipticinium acetate (NSC-264137) in cancerpatients (phase I study). Eur. J. Cancer Clin. Oncol., 18: 1285-1292, 1982.Poochikian, G. K., and Cradock, J. C. 2,5-Diaziridinyl-3,6-bis(car-boxyamino)-l,4-benzoquinone I: kinetics in aqueous solutions by high-performance liquid chromatography. J. Pharm. Sci., 70: 159-162, 1981.




31. Weinkam, R. .1.. and Deen, D. F. Quantitative dose-fesponse relations forthe cytotoxic activity of duoroethylnitrocoareas in cell culture. Cancer Res.,42:1008-1014, 1982.

32. Bier, .1..Benders, P., Wenzel, M., and Bitter, K Kinetics of "Co-bleomycin

in mice after intravenous, subcutaneous and intratumoral injection. Cancer(Phila.), 44:1194-1200, 1979.

33. \ 11( Km.m. F., Bier, .1..Bier, H. Siegel, T., Maorer, H. R., and Ohanian, S.Correlation of intralesional in vivo chemotherapy of line hepatoma with invitro drug sensitivity. Int J. Cell Cloning, /.-118-127,1983.

34. Dedrick, R. I>.. Myers, C. E., Bungay, P. M.. and DeVita, V. T. Phannaco-kinetic rational for peritoneal drug administration in treatment of ovariancancer. Cancer Treat. Rep., 62:1-11, 1978.

35. Jones, R. B., Myers, C E., Guarino, A. M.. Dedrick, R. I... Hubbarb, S. \ 1 .and DeVita, V. T. High volume intraperitoneal chemotherapy ("belly bath")

for ovarian cancer pharmacologie basis and early results. Cancer Chemother.PharnucoL, /: 161-166,1978.

36. Eckman, W. W., Patlak, C. S., and Fenstermacher, J. D. A critical evaluation

of the principles governing the advantages of intraarterial infusions. J.Phannacokinet Biopharm., 2:257-285, 1974.

37. Levin, V. A., Kabra, P. M . and Freeman-Dove, M. A. Pharmacokinetics ofintracarotid artery I4C-BCNU in the squirrel monkey. J. Neurosurg^ 4&587-593, 1978.

38. Greenberg, H. S., Ensminger, W. D., Chandler, W. F., Layton, P. D., Junck,L.. Knake, J., and Vine, A. K. Intra-arterial BCNU chemotherapy fortreatment of malignant gliomas of the central nervous system. J. Neurosurg.,61:423-429, 1984.

39. Tator, C. H., and Wassenaar, W. Intraneoptastic injection of methotrexatefor experimental brain tumor therapy. J. Neurosurg., 46:165-174,1977.

40. Steel, G. G. The growth kinetics of tumors. I .union. Oxford UniversityPress, 1977.

41. Deen, F. D., and Williams, M. E. Bioassay for estimating the decay rate ofcytoxic drugs in vitro. Proc. Amer. Assn. Cancer Res.,.'-.- 268, 1981.

42. Pulst, S. M., Levin, V. A., and Deen, D. F. In vitro cytotoxic effects ofDibromodulcitol in 9L rat brain tumor cells. Pharmaceutical Res., 3: 302-306, 1986.

3724



1987;47:3718-3724. Cancer Res Francis Ali-Osman, Jane Giblin, Dolores Dougherty, et al. Clonogenic Cell AssayPhysiologically Relevant Drug Exposure in a Human Tumor

Pharmacokinetics forin Vitro and in VivoApplication of

Updated version

http://cancerres.aacrjournals.org/content/47/14/3718

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/47/14/3718To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



application of in vivo and in vitro pharmacokinetics for ...drug in different solvents, depending...

Documents