[CANCER RESEARCH 47, 2604-2608, May 15, 1987]

Generation of Drug-resistant Variants in Metastatic B16 MouseMelanoma Cell Lines1

C. Cilio,2 J. E. Dick, V. Ling, and R. P. Hill3

Ontario Cancer Institute [C. C., V. L., K. P. H.J, Mount Sinai Hospital Research Institute [J. E. D.J, and Department of Medical Biophysics [C. C., J. E. D., V. L., R.P. H.J, University of Toronto, Toronto, Ontario, Canada M4X1K9

ABSTRACT

Genetic instability is recognized as an important aspect of the development of tumor heterogeneity and malignancy. In a previous study [Hillet al. Science (Wash. DC), 244: 998-1001, 1984), we demonstrated thatmetastatic variants are generated at a more rapid rate in the highlymetastatic B16F10 mouse melanoma cell line than in the less metastaticB16F1 cell line. The metastatic variants were phenotypically unstable,being generated and lost at high rates; consequently, we proposed adynamic heterogeneity model of tumor metastasis which describes theseproperties quantitatively. As an extension of this work, we have examinedthe ability of these two melanoma cell lines to generate variants resistantto the drugs methotrexate and .V-(phnsphnnacctyl)-i.-a$partate. We observed that the highly metastatic B16F10 cell line generated variantsresistant to a given concentration of methotrexate or ,V-(phosplionace(yl)-L-aspartate at higher rates than the B16F1 cell line. We conclude thatB16F10 cells are genetically less stable than B16F1 cells and sinceresistance to methotrexate and jV-(phosphonacetyl)-L-asparate usuallyresults from gene amplification that B16F10 cells possess increasedability to amplify DNA. This higher rate of generation of drug-resistantvariants corresponds to the higher rate of generation of metastaticvariants we observed previously and suggests that a gene amplificationmechanism may be involved in the generation of a metastatic phenotypein B16 melanoma cells.

INTRODUCTION

Genetic instability in malignant cells has been postulated (1)as an important mechanism in the process of tumor evolutionto increasing malignancy which is termed tumor progression(2). Highly metastatic tumor cell populations have sometimesbeen found to be characterized by an increased rate of mutationto drug resistance. Cifone and Fidler (3) have reported a 3- to7-fold increase in the rate of mutation occurring at two independent genetic loci (6-thioguanine and ouabain resistance) inhigh versus low metastatic UV-2237 fibrosarcoma clones. Similar results were obtained by Fidler (4) using clonal populationsderived from K-1735 melanoma and by Cifone and Fisher (5)using SF-19 fibrosarcoma cells. In contrast, Yamashina andHeppner (6) found no correlation between metastatic abilityand the rate of spontaneous mutation in three mouse mammarytumor cell lines and Elmore et al. (7) were unable to detectincreased mutation rates in chemically transformed versus normal skin fibroblasts, despite using the same genetic loci.

The above studies used drugs to which resistance is thoughtto occur through point mutation and deletion, giving rise tostable mutants which are generated at characteristic rates between 10~6 and 10~8/cell/generation (8). While much of the

instability observed in tumor cell populations could well be aresult of such processes, other mechanisms responsible for drug

Received 7/7/86; revised 9/18/86, 1/2/87; accepted 2/12/87.The costs of publication of this article were defrayed in part 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 by the Medical Research Council of Canada, The National CancerInstitute of Canada, and the Ontario Cancer Treatment and Research Foundation.

2Present address: Istituto Internazionale di Genetica e Biofisica. C.P. 3061,80100 Napoli, Italy.

3To whom requests for reprints should be addressed, at the Ontario CancerInstitute, 500 Sherbourne Street, Toronto, Ontario, Canada M4X IK').

resistance have been identified, e.g., gene rearrangement, move-able genetic elements, gene amplification, and gene modification (e.g., methylation). In these cases unstable drug-resistantvariants are often generated at much higher rates of 10~3-10~4/

cell/generation (9).In the present paper we report on studies of the rate of

generation of drug-resistant variants in the highly metastaticB16F10 cell line and in the less metastatic B16F1 cell line (10).The drugs MTX4 and PALA were chosen for the study because

resistance to these drugs usually arises as a result of geneamplification. The B16F1 and B16F10 cell lines were examinedbecause we have shown previously that metastatic variant cellsare generated stochastically in these cell populations at a rapidrate. In fact we demonstrated that the highly metastatic B16F10cell line generates metastatic variants at a higher effective rate(~5 x 10~5/cell/generation) than the poorly metastatic B16F1cell line (~1 x 10~5/cell/generation) (H). Such variants have

been found to express an unstable metastatic phenotype (12,13).

MATERIALS AND METHODS

Cell Lines. The origin and characteristics of the B16 melanoma celllines (obtained from Dr. I. J. Fidler) have been reported elsewhere (10).B16 melanoma cells were routinely grown in a-MEM plus 10% fetalcalf serum but for treatment with MTX they were adapted to grow inlow folie acid (0.1 mg/liter) nucleoside free «-MEMplus 10% dialyzedFCS. The cells were also grown in nucleoside free a-MEM plus 10%dialyzed FCS for treatment with PALA. The cell lines were tested fortheir metastatic efficiency using the experimental metastasis assay (14).C57BL/6J mice were given injections i.v. of 5 x IO4cells in 0.2 ml of

medium, and macroscopic lung colonies were counted 3 weeks laterunder a dissecting microscope. The B16F10 cells gave approximately5 times more lung métastasesthan the B16F1 cell line. Growth in lowfolate medium did not influence the metastatic effkency of either cellpopulation (data not shown).

Drugs. For our experiments we have used a commercial source ofmethotrexate (Lederle). PALA was obtained from the U. S. NationalCancer Institute (courtesy of Dr. Narayanan, Drug Synthesis andChemistry Branch).

The sensitivity of the B16 melanoma cells to the drugs was tested byplating the cells (up to IO4 cells/plate) in the continuous presence of

the drugs. The resistant colonies were counted after 10 days of incubation following méthylèneblue staining. Preliminary experimentsdemonstrated that the frequency of resistant colonies was linearlycorrelated with the number of cells plated up to approximately 2 x IO4cells/100-mm plate.

Growth of Clonal Populations. Clonal populations of B16 melanomacells were obtained by limiting dilution into 24-well Linbro tissueculture clusters. In some experiments the parental population wascloned directly (single cloning); in others, an individual clone of aboutIO3 cells was picked and subcloned (doublecloning). When a clonereached a size of approximately 10' cells (directly estimated by counting

the number of the cells in the colonies) it was trypsinized and left inthe well to grow to a size of 10s cells. This "spreading" process resulted

4The abbreviations used are: MTX, methotrexate; PALA, A'-(phosphonacetyl)-L-aspartate; a-MEM, n minimal essential medium; FCS, fetal calf serum; DHFR,dihydrofolate reducíase.

2604

Research. on December 11, 2020. © 1987 American Association for Cancercancerres.aacrjournals.org Downloaded from

DRUG RESISTANCE AND METASTASIS

in more uniform growth in the well. At the size of IO5cells (estimatedby eye and confirmed by hemocytometer count) the clones were tryp-sinized, and the cells were transferred to tissue culture flasks to expandthe clone to 1.5 x IO7cells. In some experiments the same clones were

studied at two different sizes of splitting them into two parts when theyreached approximately 2 x IO5 cells, and expanding one-half of eachclone to approximately 1.5 x IO7cells.

Determination of the Rate of Generation of Drug-resistant Variants.Parallel clonal populations grown to three different sizes (~103 cells/clone is a very small clone; 10s cells/clone is a small clone; and 1.5 xIO7cells/clone is a large clone) were studied for the frequency of drug-

resistant variants. Either the whole clone (very small size) or replicatesamples of IO4 cells/plate (small and large size) for each clone were

plated in the continuous presence of 30 nM MTX or 12 n\i PALA. Inall the experiments, IO2cells/clone were plated in the absence of drugs

in order to determine the plating efficiency of the cells in the clone.The rates of generation of drug-resistant variants in the clonal populations were calculated using Luria-Delbruck fluctuation analysis (IS).As described previously (16) the data from analysis of multiple parallelclones was fitted to the formula

n\n (3.46 uNQ - I—I In 2 = 0

where ¿iis the apparent mutation rate (per cell per generation), N is thenumber of cells per parallel clonal population, C is the number ofparallel clonal populations, and M is the mean number of drug-resistantvariants per parallel clonal population.

Southern Blot Analysis. Amplification of the gene coding for DHFRwas examined by Southern blot analysis (17). DNA (5 or 10 Mg)fromparental populations and from MTX-resistant cells was digested withthe restriction endonuclease EcoRl, electrophoresed through a 0.6%agarose gel, transferred to a nitrocellulose filter, hybridized with nick-translated 32P-labeled plasmili bearing murine DHFR complementaryDNA (pSVS-DHFR obtained from Bethesda Research Laboratories,Inc. (Gaithersburg, MD); for details, see Ref. 18) and exposed to X-rayfilm. The filter was then washed, rehybridized with \-mos complementary DNA (1.2-kilobase Hindlll fragment of \-mos) and reexposed toX-ray film.

RESULTS

Survival curves for B16F1 and B16F10 cell populationsplated in the presence of varying concentrations of methotrex-ate are shown in Fig. 1. The highly metastatic B16F10 cell linewas consistently found to be more resistant to MTX than thepoorly metastatic B16F1 cell line; the difference between thesurvival curves is about a factor of 10 in survival.

Evidence that the colonies which grew in the drug were truly

1015 30 45Methotrexate dose (nM)

Fig. 1. Survival curves for B16F1 and B16F10 cells plated in the continuouspresence of different concentrations of MTX. Error limits are SEs (bars) calculated for three or more separate experiments.

resistant to MTX was obtained from replating experiments.Colonies of B16F1 and B16F10 cells which grew in 30 nMMTX were trypsinized after 10 days of growth, and the cellsrecovered were replated in the presence or in the absence of thesame concentration of the drug. The number of colonies growing in the presence of MTX was from 35 to 70% of the numberof colonies growing in medium alone. These percentages didnot change significantly if the cells were passaged once or twicein normal medium before replating in 30 HMMTX.

Methotrexate is an analogue of folie acid, the substrate ofthe enzyme DHFR, and inhibits DHFR by competitive bindingto its catalytic site (19). Resistance to MTX in vitro has beenfound to occur by a number of mechanisms (20-22), but overproduction of DHFR is the most frequent mechanism andresults from the amplification of the DHFR gene (23). Toexamine whether amplification of the DHFR gene had occurredin the clones of B16F1 and B16F10 which grew in 30 HMMTX,we picked colonies, expanded them in the same concentrationof the drug, and analyzed their DNA by the Southern blottechnique. Fig. 2, top, shows the results obtained when the filterwas hybridized with DHFR complementary DNA. For comparison of signal intensity two dilutions (0.5 parental andparental) of each parental DNA were examined. In Fig. 2,bottom, are the results when the filter was washed and rehybridized with \-mos complementary DNA. The intensity of thesingle band of the c-mos gene allows a comparison between thedifferent lanes indicating where there was variation in theamount of DNA loaded and allowing the signals of the DHFRgene to be normalized. Taken together the results presented inFig. 2 indicate that the DNA from MTX-resistant clones ofboth B16F1 and B16F10 cells demonstrates an increase in the

BI6FI BI6FIO

1/2 p p 2 i/2p p l 2 3

F4M ••

kb-23.0

9.5

- 6.6

- 4.3

DHFR

- 23.0

C-MOS

Fig. 2. Southern blot analysis of the DHFR (top) and c-mos (bottom) genesfrom parental (p) and MTX-resistant cells. Lane 1,0.5 dilution of B16FI parentalDNA; Lane 2, parental B16F1 DNA; Lanes 3 and 4, DNA from colonies whichwere obtained from BI6F1 parental cells plated in 30 nM MTX (expanded to IO7cells in the presence of the drug); Lane 5, 0.5 dilution of B16F10 parental DNA;Lane 6, parental B16F10 DNA; Lanes 7-9, DNA from colonies derived fromB16FIO parental cells plated in 30 nM MTX (expanded to 10' cells in the presenceof drug), kh. kilobase.

2605

Research. on December 11, 2020. © 1987 American Association for Cancercancerres.aacrjournals.org Downloaded from

DRUG RESISTANCE AND METASTASIS

intensity of the bands which correspond to the DHFR gene,relative to their parental cells. Thus, gene amplification wasmost likely responsible for the majority of MTX-resistant variants observed in these studies.

The B16F1 and B16F10 cell populations were next plated inthe continuous presence of different concentrations of the drugPALA. This drug specifically inhibits the aspartate transcar-bamylase activity of the trifunctional protein carbamyl phosphate synthetase-aspartate transcarbamylase-dihydroorotase,which catalyzes the first steps of UMP biosynthesis (24-25).The only mechanism which has been described for the development of resistance to PALA, at the drug concentrations usedin our experiments, is the amplification of the gene coding forCAD (26). The survival curves for PALA treatment are shownin Fig. 3. It can again be observed that the highly metastaticB16F10 cell line is substantially more resistant to PALA thanB16F1 cells at any given drug concentration.

In another series of experiments we studied whether variantsresistant to either MTX or PALA were actually being generatedduring the growth of populations of BI6F1 and B16F10 cells.Parallel clonal populations of cells were grown to three differentsizes (very small, ~103 cells/clone; small, ~1 x IO5cells/clone;or large, ~1.5 x IO7 cells/clone) and then tested for the frequency of drug-resistant variants which they contained. Theresults of experiments using B16F1 and B16F10 cells and thedrug MTX are shown in Figs. 4 and 5. Some of the very small(B16F10) or small (B16F1) clones were found to contain noresistant variants, and on average the clones grown to a largerpopulation size had a higher frequency of drug-resistant variants. This latter result was obtained both when individualclones were grown to different sizes before testing and whenclonal populations of different size were obtained from thesame original selected cell. Similar results were obtained inexperiments involving either single- or double-cloning procedures (see "Materials and Methods" for details). These findings

demonstrate that the clones were derived from cells which werenot drug resistant at the time of their selection, that drug-resistant variants were being generated as the clonal populationsexpanded in size, and that the large variability in the frequencyof drug-resistant variants between the clones in any one sizegroup was not due to preexisting heterogeneity in the cellpopulation. Similar results were obtained when clones weretested with the drug PALA (data not shown).

To examine whether or not a common metabolic mechanismmight account for the generation of resistance to PALA andMTX, we tested the same clonal populations of B16F1 andB16F10 cells by plating one-half of each clone in the presenceof MTX and the other half in the presence of PALA. The

10s BI6FI cells + 30nM MTX

iff2

IO-

IO'4•BI6FI cells

•BI6FIO cells

IO 15

TOLA dose Õ/J.M}

20

Fig. 3. Survival curves for B16F1 and B16F10 cell lines plated in the continuous presence of different concentrations of PALA. Error limits are SEs (bars)calculated for three or more separate experiments.

. •N=1.5x IO5

fio9}

O -

(a (b.0 N =1.5 x IO7

123456789 123456789

Clonol population (rank order)

Fig. 4. Number of MTX (30 nM)-resistant colonies observed from clones ofB16F1 grown to two sizes, 10' cells/clone (small, •)and 1.5 x IO7 cells (large,O). The arrangement of the clones goes from the one showing the smallest numberof MTX-resistant variants/IO4 cells to the clone with the highest number ofMTX-resistant variants/10* cells (rank order).

BI6FIO cells »30nM MTX

-'

•N'lO1

b) "cl

DN*I.5»10'

20 0 2 4 6 e OClonal population ( rank order )

Fig. 5. Number of MTX (30 niu)-resistant colonies observed from clones ofB16FIO grown to three sizes, IO3cells/clone (a, very small, 8), 10* cells/clone(A, small, •).and 1.5 x IO7cells/clone (c, large, Q). Clones are arranged in the

same way as described in Fig. 4.

Table 1 Generation of MTX- and PALA-resistant variants in the same cloneMTX- and PALA-resistant variants obtained from the same clones of B16F10

cell line each grown to the size of 1.5 x IO7cells.

CloneNo.12345678910MTX(30 nM)

variants/104cells256.7281.6358.2421.3268.9137.523.681.81181.832.6PALA(12 KM)

variants/ 10*cells011.734.34.23.416.78.67.89.130.2

2606

Research. on December 11, 2020. © 1987 American Association for Cancercancerres.aacrjournals.org Downloaded from

DRUG RESISTANCE AND METASTASIS

results of such an experiment, reported in Table 1, demonstrateno correlation between different clones in the frequency ofMTX- and PALA-resistant variants observed, indicating thatthese resistant variants arise independently and that resistanceto one drug does not automatically confer resistance to theother.

The variability in the frequency of drug-resistant variantsbetween the different clonal populations was further investigated (see Table 2). When repeated samples (8 or 9 plates) wereplated from a single clone, the ratio of the variance to the meanwas close to unity, as would be predicted for sampling error bythe Poisson distribution. In contrast the variance to mean ratiowas nearly 20 when the variance between the different cloneswas analyzed. The results are similar to those obtained by Luriaand Delbruck (15) in their study of bacterial mutants. Theydemonstrated that wide clonal variation, as observed here, wasconsistent with the generation of variants by a stochastic process and not consistent with the supposition that the variantswere induced by the testing process. Our results thus fulfill thecriteria applied by Luria and Delbruck (IS) to indicate that thegeneration of variants is a stochastic process. These authorsdeveloped a statistical method of analyzing such stochastic (orrandom) events (Luria-Delbruck fluctuation analysis), whichhas been widely used to study the development of both stableand unstable drug-resistant variants in mammalian cells (27).We have applied this analysis to calculate the rate of generationof the drug-resistant variants in the B16F1 and B16F10 clonalpopulations. In Table 3 are reported the effective rates ofgeneration of variants resistant to 30 HMMTX or 12¿¿MPALA,calculated for parallel clonal populations of B16F1 and B16F10cells. For both the drugs used in our study the effective rate ofgeneration of resistant variants is substantially higher in thehighly metastatic B16F10 cell population compared to thepoorly metastatic B16F1 cell population. In the last column ofTable 3 are given, for comparison, the effective rate of genera-

Table2 TreatmentofclonesofBloFlOcellswithmethotrexate(30nu)Clones of B16F10 cells were grown to 10s cells and the cells plated (8 or 9

plates at IO4cells/plate) in the presence of methotrexate (30 HM).The plates were

incubated for 10 days, stained, and the colonies counted. The mean to varianceratio was then calculated within individual clones and between the clones asindicated.

CloneNo.51525354555657585960Meancolonies/

IO4cells20.331.042.7517.252.621.655.3758.01.870.71Variance61.6111.535.3535.073.421.9914.29110.671.250.58Meanivarianceratio3.033.590.832.031.310.652.661.910.670.82

18.15 400.80 22.08

Table 3 Effective rates of generation of drug resistant variantsThe rate of generation of drug-resistant variants was obtained by applying the

Luria-Delbruck fluctuation analysis to at least nine parallel clones for eachdetermination. The size to which the clones were grown (before testing) did notsignificantly affect the calculated rate.

Rate* ofgeneration of drug-

No, of deter- resistant variantsCells Drug minations(xlO4)B16F1

MTX(30nM)PALA (12MM)B16F10

MTX(30nM)PALA(12 MM)8

5931.3

±0.38°

1.4 ±1.1414.8

±2.744.4 ±3.31Rate*

of generation

of metastatic variants (xlO5) (from

Ref.11)1.3

±0.88

5.2 ±0.69

' Mean ±SE.* Per cell per generation.

tion of metastatic variants previously determined (11) for thetwo cell lines. It should be noted that the numerical values ofthe rates apply only for the drug concentrations studied andthat because of the instability of the phenotypes, these effectiverates could be underestimates of the real rate.

DISCUSSION

The results of the experiments reported here indicate thatthere is a stochastic generation of variant cells resistant toMTX or PALA in the B16 melanoma cell lines studied. Furthermore the finding that some of the clones contained no drug-resistant variants when grown to small (B16F1) or very small(B16F10) size (see Figs. 4 and 5) indicates that the clones werederived largely from non-drug-resistant cells and that the drug-resistant variants arose during the growth of the clonal populations. The results presented in Fig. 2 indicate that geneamplification was most likely responsible for the majority ofresistant variants observed in these studies, since all five MTX-resistant colonies studied demonstrated amplification of theirDHFR gene. Overall these results therefore suggest that geneamplification can occur stochastically at a rapid rate in the B16melanoma cell populations.

The survival curves in Figs. 1 and 3 do not demonstrate atail or plateau indicating the presence of a discrete resistantsubpopulation of cells despite the evidence discussed above thatresistant variants are generated as the cell populations grow.The probable explanation for the shape of the curves is providedby the observation (28-30) that in cell populations undergoinggene amplification, individual cells demonstrate varying degreesof amplification and presumably different levels of drug resistance. Thus as the drug concentration of selection increases asmaller and smaller subpopulation of the cells exhibits resistance. It should be noted that the cells growing to form coloniesin the presence of drug are indeed drug resistant and are notprobabilistic survivors, since when cells were recovered fromthe colonies and replated in the same concentration of drugthey demonstrated a resistant phenotype.

Although the difference between the survival curves forB16F1 and B16F10 cells might superficially suggest that theB16F10 cells are innately more resistant to the drug, the curvesare more likely to be explained by the gene amplificationoccurring in the cell populations. The cloning studies demonstrated that growing B16F10 cell populations generate variantsresistant to 30 HM MTX at a higher rate than B16F1 cellpopulations. It might thus be expected that B16F10 cell populations would contain a higher frequency of such variants, withthe result that the population as a whole would demonstrate ahigher level of survival following treatment with the drug.Following the argument from the paragraph above, this effectwould result in a higher level of survival at all drug concentrations. Survival curves of similar shape to those seen in Figs. 1and 3 have been reported for other cell lines in which anincreased amplification of the DHFR gene was deliberatelyinduced by treatment with hydroxyurea, UV irradiation, andhypoxic exposure (31-33). A higher innate level of resistanceto drug in B16F10 cells is also unlikely because (a) the two celllines were derived from the same population in the absence ofdrug selection (10), and (/>)in the cloning studies (Figs. 4 and5), clones with no cells resistant to 30 nM MTX were observedfor both populations, indicating that the clones are derivedprimarily from drug-sensitive cells and that the drug-resistantvariants arise during clonal growth.

Determination of rates of generation of variants resistant todrugs in tumor cell populations has been used as a sensitiveand quantitative measure of genetic instability (9, 27). Such

2607

Research. on December 11, 2020. © 1987 American Association for Cancercancerres.aacrjournals.org Downloaded from

DRUG RESISTANCE AND METASTASIS

studies have supported Nowell's original hypothesis (1), that

during tumor progression, increased genetic instability leads toincreased malignancy. Many of the earlier studies have usedmarkers such as thioguanine or ouabain resistance which arethought to arise as a result of point mutations or deletions (27,35). Recently, Sager et al. (34) have used the methotrexateresistance marker to measure the ability of hamster cells toamplify their DHFR gene. They noted that two tumorigeniclines were able to amplify their DHFR gene more readily thana nontumorigenic line and concluded that gene amplificationmay play an important role in malignant progression.

In the present study we have taken a similar approach,examining the ability of B16F1 and B16F10 cells to generatevariants resistant to MTX or PALA. We observed that variantcells resistant to either of these drugs were generated stochastically at rapid rates during the growth of clonal populationscomprised largely of drug-sensitive cells. Moreover, the B16F10line was found to generate the drug-resistant variants at anappreciably higher rate than the B16F1 line (Table 3).

In our previous studies (11-13, 16) we demonstrated that, inB16 melanoma and Kill sarcoma cell lines, metastatic variantswere spontaneously generated but were phenotypically unstableand lost at high rates. The metastatic ability of a tumor cell linewas therefore dependent on the dynamics of generation andloss of such variants, and a dynamic heterogeneity model oftumor metastasis was proposed to describe these features in aquantitative fashion. It was postulated that the increased metastatic ability observed in the 15161Id cell line compared withthe B16F1 cell line was due to its increased rate of generationof metastatic variants (see Table 3). This higher rate of generation of metastatic variants corresponds to the higher rates ofgeneration of drug-resistant variants observed in the presentstudy. This correspondence suggests that a gene amplificationmechanism may be involved in expression of the metastaticphenotype.

The possibility that a common genetic mechanism may beresponsible for the generation of metastasis and for resistanceto anticancer drugs allows speculation that differences betweenearly and late stages of tumor progression might involve anincrease in the rate of gene amplification. Putative DNA sequences able to control the rate of the amplification processmight thus play an important role during tumor development.If métastaseshave a high probability of being initiated by cellswhich have high rates of amplification, this could explain whymétastasesoften are resistant to chemotherapy.

ACKNOWLEDGMENTS

The authors wish to acknowledge the technical assistance of R. Kubaand the secretarial assistance of Anne Marie Watson.

REFERENCES

1. Nowell, P. C. The clonal evolution of tumor cell populations. Science (Wash.DC), 194: 23-28, 1976.

2. Foulds. L. The experimental study of tumor progression: a review. CancerRes.. ;* 327-339, 1954.

3. Cifone, M. A., and Fidler. I. J. Increasing metastatic potential is associatedwith increasing genetic instability of clones isolated from murine neoplasms.Proc. Nati. Acad. Sci. USA, 78: 6949-6952, 1981.

4. Fidler, I. J. General considerations for studies of experimental cancer metastasis. Methods Cancer Res., 15: 399-439, 1978.

5. Cifone, M. A., and Fisher, M. S. Increased metastatic potential and mutationrate in a spontaneous murine fibrosarcoma following in vitro irradiation. J.Cell Bio}., 87: 290 (Abst.), 1980.

6. Yamashina, K., and Heppner, G. H. Correlation of frequency of induced

mutation and metastatic potential in tumor cell lines from a single mousemammary tumor. Cancer Res., 45:4015-4019, 1985.

7. Elmore, E., Kakunaga, T., and Barrett, J. C. Comparison of spontaneousmutation rates of normal and clinically transformed human skin fibroblasts.Cancer Res., 43: 1650-1655, 1983.

8. Thompson, L. H., and Baker, R. M. Isolation of mutants of culturedmammalian cells. In: D. M. Prescott (ed.), Methods in Cell Biology, Vol. 6,pp. 209-223. New York: Academic Press, Inc., 1973.

9. Ling, V., Chambers, A. F., Harris, J. F., and Hill, R. P. Quantitative geneticanalysis of tumor progression. Cancer Metastasis Rev., 4:173-194, 1985.

10. Fidler, I. J. Selection of successive tumour lines for métastases.Nat. NewBiol., 242: 148-149, 1973.

11. Hill, R. P., Chambers, A. F., Ling, V., and Harris, J. F. Dynamic heterogeneity: rapid generation of metastatic variants in mouse B16 melanoma cells.Science (Wash. DC), 224: 998-1001, 1984.

12. Chambers, A. F., Harris, J. F., Ling, V., and Hill, R. P. Rapid phenotypevariations in cells derived from lung métastasesof KHT fibrosarcoma.Invasion Metastasis, 4: 225-237, 1984.

13. Young, S. I).. and Hill, R. P. Dynamic heterogeneity: isolation of murinetumor cell populations enriched for metastatic variants and quantification ofthe unstable expression of the phenotype. Clin. Exp. Metastasis, 4:153-176,1986.

14. Hill, R. P., and Bush, R. S. A lung-colony assay to determine the radiosen-sitivity of the cells of a solid tumor. Int. J. Radiât.Biol. Relat. Stud. Phys.Chem. Med., 15: 435-444, 1969.

15. Luria, S. E., and Delbruck, M. Mutations of bacteria from virus sensitivityto virus resistance. Genetics, 28:491-511, 1943.

16. Harris, J. F., Chambers, A. F., Hill, R. P., and Ling, V. Metastatic variantsare generated spontaneously at a high rate in mouse KHT tumor. Proc. Nati.Acad. Sci. USA, 79: 5547-5551, 1982.

17. Southern, E. M.: Detection of specific sequences among DNA fragmentsseparated by gel electrophoresis. J. Mol. Biol., 98: 503-517, 1975.

18. Subrami, S., Mulligan, R., and Berg. P. Expression of the mouse dihydrofo-late reducíasecomplementary deoxyribonucleic acid in simian virus 40 vectors. Mol. Cell. Biol., /: 854-864, 1981.

19. Chabner, B. A. Methotrexate. In: B. A. Chabner (éd.),Pharmacologie Principles of Cancer Treatment, pp. 229-255. Philadelphia: W. B. Saunders Co.,1982.

20. Sirotnak, F. M., Moccio, D. M., Kelleher, L. E., and Goulas, L. J. Relativefrequency and kinetic properties of transport-defective phenotypes amongmethotrexate-resistant 11210 clonal cell lines derived /;/ vivo. Cancer Res.,41:4447-4452, 1981.

21. Flintoff, W. F., Davidson, S. V., and Siminovitch, L. Isolation and partialcharacterization of three methotrexate-resistant phenotypes from Chinesehamster ovary cells. Somatic Cell Genet., 2: 245-261. 1976.

22. Alt, F. W., Kellems, R. E., and Schimke, R. T. Synthesis and degradation offolate reducíasein sensitive and methotrexate-resistant lines of S-180 cells.J. Biol. Chem., 251: 3063-3074, 1976.

23. Alt, F. W., Kellems, R. E., Berlino, J. R., and Schimke, R. T. Selectivemultiplication of dihydrofolate reducíasegenes in methotrexate-resistantvariants of cultured murine cells. J. Biol. Chem., 253: 1357-1370, 1978.

24. Collins, K. I'., and Stark, G. R. Aspanate transcarbamylase: Interactionswith the transition state analogue JV-(phosphonacetyl-asparate). J. Biol.Chem., 246:6599-6605, 1971.

25. Swyrd, E. A., Seaver, S. S., and Stark, G. R. (Phosphonacetyl)-L-aspartate,a potent transition state analog inhibitor of aspartate transcarbamylase,blocks proliferation of mammalian cells in culture. J. Biol. Chem., 249:6945-6950, 1974.

26. Wahl, G. M., Padgett, R. A., and Stark, G. R. Gene amplification causesoverproduction of the first three enzymes of UMP synthesis in AXphosphon-acetyl)-L-asparate-resistant hamster cells. J. Biol. Chem., 254: 8679-8689,1979.

27. Ling, V. Genetic basis of drug resistance in mammalian cells. In: N. Bru-chovsky and J. H. Goldie (eds), Drugs and Hormone Resistance in Neoplasia,pp. 1-19. Miami: CRC Press, 1982.

28. Johnston, R. N., Beverley, S. M., and Schimke, R. T. Rapid spontaneousdihydrofolate reducíasegene amplification shown by fluorescence activatedcell sorting. Proc. Nati. Acad. Sci. USA, 80: 3711-3715, 1983.

29. Zieg, J., Clayton, C. E., Ardeshir, F., Giubilo, E., Swyrd, E. A., and Stark,G. R. Properties of single-step mutants of Syrian hamster cell lines resistantto /V-íphosphonacetyO-L-aspartate. Mol. Cell. Biol., 3: 2089-2098, 1983.

30. Hill, A. B., and Schimke, R. T. Increased gene amplification in L5178Ymouse lymphcma cells with hydroxyurea-induced chromosomal aberrations.Cancer Res., 45: 5050-5057, 1985.

31. Brown, P. C., Tlsty, T. D., and Schimke, R. T. Enhancement of methotrexateresistance and dihydrofolate reducíasegene amplification by treatment ofmouse 3T6 cells with hydroxyurea. Mol. Cell. Biol., 3: 1097-1107, 1983.

32. Tlsty, T. D., Brown, P. C., and Schimke, R. T. UV radiation facilitatesmethotrexate resistance and amplification of the dihydrofolate reducíasegene in cultured 3T6 mouse cells. Mol. Cell. Biol., 4: 1050-1056, 1984.

33. Rice, G. C., Hoy, C.. and Schimke, R.T. Transient hypoxia enhances thefrequency of dihydrofolate reducíasegene amplification in Chinese hamsterovary cells. Proc. Nail. Acad. Sci. USA, 83: 5978-5982, 1986.

34. Sager, R., Gadi, I. K., Stephens, L., and Grabowy, C. T.: Gene amplification:an example of accelerated evolution in tumorigenic cells. Proc. Nail. Acad.Sci. USA, 82: 7015-7019. 1985.

35. Baker, R. M., Brunette, D. M., Mankovitz, R., Thompson, L. H.. Whitmore,G. F., Siminovitch, L., and Till, J. E. Ouabain-resistant mutants of mouseand hamster cells in culture. Cell, /: 9-21, 1974.

2608

Research. on December 11, 2020. © 1987 American Association for Cancercancerres.aacrjournals.org Downloaded from

1987;47:2604-2608. Cancer Res   C. Cillo, J. E. Dick, V. Ling, et al.   Melanoma Cell LinesGeneration of Drug-resistant Variants in Metastatic B16 Mouse

  Updated version

  http://cancerres.aacrjournals.org/content/47/10/2604

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

  .pubs@aacr.orgDepartment 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/47/10/2604To request permission to re-use all or part of this article, use this link

Research. on December 11, 2020. © 1987 American Association for Cancercancerres.aacrjournals.org Downloaded from