Membrane-Mediated Drug Resistance and Phenotypic Reversion to Normal Growth Behavior of Chinese Hamster Cells 1, 2 June l. Biedler,3 Hansjorg Riehm,3,
4
Robert H.
F. Peterson,3 and Barbara A. Spengler 3,5
SUMMARY-Development of resistance to actinomycin D, dau· nomycin, or vincristine in Chinese hamster cells growing in vitro resulted in reversion to or retention of normal phenotypes in comparison to spontaneously transformed drug-sensitive parent populations. Sublines resistant or cross-resistant to actinomycin D showed reduced uptake of antibiotic in proportion to degree of resistance. The cells with acquired resistance were either weakly tumorigenic or nontumorigenic when tested in the cheek pouches of weanling Syrian hamsters treated with cortisone. In contrast, parent cells and several amethopterin (methotrexate)-resistant sublines produced many tumors. Antibiotic- and Vinca-alkaloid-resistant cell lines showed oriented growth patterns often associated with the behavior of normal cells in culture; antibiotic-sensitive, tumorigenic lines had morphologic characteristics of malignant cells. The altered cell membrane properties that accompanied development of resistance or cross-resistance to actinomycin D appeared to account also for the lower oncogenic potential and greater cell adhesiveness of resistant cells relative to their malignant counterparts.-J Natl Cancer Inst 55: 671-{)80, 1975.
In Chinese hamster cells exposed to and resistant to actinomycin D, daunomycin (daunorubicin), or vincristine, resistance was most likely a consequence of cell membrane alteration(s) which made them less drugpermeable (1-4). These sublines exhibited increases in resistance or cross-resistance to actinomycin D of approximately 80-fold to 2,500-fold. We have noted striking differences between drug-sensitive progenitor cells and resistant derivatives in respect to cell morphology and cell growth patterns in vitro. For spontaneously transformed parent DC-3F cells in particular, development of antibiotic or Vinca-alkaloid resistance was accompanied by reversion of morphologic characteristics associated with malignant cells in culture toward those associated with normal cells. In contrast, a large series of experimentally derived amethopterin (methotrexate)-resistant sublines (5) continued to resemble the parent DC-3F population even after prolonged cultivation at high concentrations of drug. Therefore, we systematically examined the apparent correlation between resistance or cross-resistance to actinomycin D and manifestation of normal cell phenotype in a larger group of drug-resistant Chinese hamster sublines. Ability to form tumors in the cheek pouches of cortisone-treated Syrian hamsters served as the primary criterion of malignancy; correlations were made with morphology in culture and degree of densitydependent growth inhibition. Results were partially summarized in a preliminary report (6). Tumor cell populations established in vitro may, after long periods in culture, lose the ability to produce tumors in the host of origin (7-11). Methods have been devised for experimental selection of "revertant" cells that have lost one or more of the properties characteristic of the transformed population (12). For example, variants with reduced tumorigenicity and increased growth control in vitro were indirectly selected with 5-fluoro-2'deoxy-,B-uridine (FUDR) from polyoma virus and simian
virus 40 (SV40)-transformed mouse and hamster fibroblasts, as described by Pollack et a!. (13). Rabinowitz and Sachs (14, 15) obtained variant clones from polyomaor dimethylnitrosamine-transformed hamster cells grown on glutaraldehyde-fixed normal cell monolayers. The variants showed reversion of properties of transformed cells, including decreases in tumorigenicity and saturation density. In still another system comprised of clonal hydrocarbon-transformed mouse fibroblasts, Mondal et a1. (16) obtained variant clones of reduced malignancy accompanied by marked lowering of saturation density. However, no clear differences in morphology were observed. We now describe Chinese hamster cells with acquired resistance to a variety of cancer-chemotherapeutic agents. Development of resistance associated with altered membrane permeability resulted in production or retention of cell populations with a less malignant phenotype than that of their spontaneously transformed counterparts. We refer to these drug-resistant cells as permeability "variants" or as "revertants," without implication as to pathway or mechanism of variant or revertant production. In the case of the twice-cloned, spontaneously transformed, fibroblastic DC-3F cells, revertants emerged after prolonged cultivation in progressively higher concentrations of actinomycin D, daunomycin, or vincristine. In the once-cloned, fibroblastic CLM-7 cell line established in our laboratory, resistant variants selected by actinomycin D or amethopterin (methotrexate) retained a normal phenotype (though it differed somewhat from that of the non transformed population from which they were derived). In contrast, drug-sensitive, parent CLM-7 cells underwent spontaneous neoplastic and morphologic transformation. MATERIALS AND METHODS
Cells and culture medium.-Parent cell lines DC-3F and CLM-7 were established in vitro from lung and bone marrow, respectively, of normal Chinese hamsters; they grow in monolayer culture and have chromosome numbers in the diploid range (1). DC-3F is a "spontaneously" transformed line, as determined from tumorigenicity testing. CLM-7 cells found non tumorigenic during the first 12 months after establishment were designated CLM7N ("N" denotes normal). After 18 months in continuous culture, cells underwent malignant transformation, as determined from tumorigenicity assays, and were designated CLM-7T ("Tn for transformed). CLM-7N cells frozen in liquid nitrogen after 4 months in culture and Received March 7, 1975; accepted May 9, 1975. Supported in part by Public Health Service research grant CA08748 from the National Cancer Institute. 3 Memorial Sloan-Kettering Cancer Center, 410 East 68th St., New York, N.Y. 10021. 4 Present address: Kinderklinik der Freien Universitat, Heubnerweg 6, 1 Berlin 19, Germany. 5 We thank Dr. Dorris J. Hutchison for her support and Marion Linczer Katz, Daryl D. Hill, Roberta A. Devaney, David Kessler, and Judith A. O'Neil for their participation in this study. 1
2
JOURNAL OF THE NATIONAL CANCER INSTITUTE, VOL. 55, NO.3, SEPTEMBER 1975
671
672
BIEDLER, RIEHM, PETERSON, AND SPENGLER
later reestablished in vitro again underwent malignant transformation. This branch was also designated CLM7T; data obtained from both branches were combined. Drug-resistant sublin~s wer~ developed ?y exposure of mass populations to mcreasmg concentratIons of drug and were maintained in the presence of drug at the final concentrations listed in table 1. Most resistant sublines were described previously with respect to response to selective agent, cross-resistance, karyotype, and me~ha nisms of resistance (1-5). The CLM-7/BUDR subhn~, able to grow in the presence of 500 pg 5-bromodeoxyundine (BUDR)/ml, was maintained at 50 p.g/ml. Cell lines were propagated in Eagle's minimum essential medium supplemented with 10% fet~l bovine serum, penicillin (100 IU Iml), and streptomYCIn (100 p.g/ml). Drug-treated lines were routinely grown in the absence of drug for 10-15 days with at least two culture transfers before all experimental determinations were made. Highly resistant DC-3F I AD X cells, after maintenance at 10 p.g actinomycin D /ml for approximately 9 mont~s, were subdivided and grown in drug-free medium III parallel with a treated population. The subline maintained in the absence of drug was designated DC-3F I AD X-U and was studied for about 3 years. Resistance to actinomycin D was assayed at 3- to 6-week intervals. A plot of these data shows the response to antibiotic that corresponds to the number of days in drug-free medium. at the time of heterotransplantability testing. Drug sensitivity assay.-Response of cell lines to chemical agents was determined in a 72-hour assay based on relative cell counts in control and drug-treated cultures (1). Drug response was specified by a mean effective dose (ED 50) value: the drug concentration permitting growth to 50% of control cell numbers. Degree of resistance is expressed as the ratio of EDSO values for resistant cells to paren t cells.
Plating efficiency, population .doubling time, and saturation density.-For estimates of plating efficiency, 200±20 cells were plated in 5 ml medium in 60-mm glass plates. Colonies were fixed and stained after 7 to 10 days. Only colonies of more than 5 cells were scored. Population doubling time and saturation density estimates were obtained within the same experiment, with an inoculum of 1 X 10 5 cells in 5 ml medium/60-mm glass plate. Medium was replaced twice a week. Duplicate plates w~re counted at appropriate intervals during the exponentIal growth phase to obtain the average population doubling time and during the stationary growth phase to obtain the maximum cell number /20-cm2 plate. All values are based on 2--4 replicate experiments. Heterotransplantation tests.-For determination of the tumor-producin'4 capacities of parent and drug-resistant sublines, I million, or occasionally 3 million, cells were inoculated into cheek pouches of 19- to 22-day-01d female weanling Syrian hamsters as previously described (17). Animals received 2.5 mg cortisone acetate sc at the time of cell inoculation, 3 days later, and twice a week thereafter. Cheek pouches were examined once a week for 4 weeks and tumors were measured in 3 dimensions. Samples were excised for histopathologic examination. Actinomycin D uptake determination.-Cells were exposed to tritiated actinomycin D [2.0 p-Ci/ml, 2.0p-g/ml (1.6 X 10-6 M), total external concentration] for 1 hour and prepared for radioactivity measurement as detailed in (4). Samples were counted in a Packard Liquid Scintillation Spectrometer. Tritiated actinomycin D (specific
activity, 8.4 Ci/mmole) was obtained from Schwarz/ Mann, Orangeburg, New York. RESULTS Drug Resistance
Table I lists all cell lines. selective concentrations of drugs, experimentally determined increases in resistance, and degree ~f cross-resistance to actinomy~i~ ~ for sublines not dIrectly selected by that antibIOtIC. Doseresponse data for most lines have been reported elsewhere (1,2,4,5). The BUDR-resistant subline CLM-7/BUDR exhibited no cross-resistance to actinomycin D; the ED 50 of the antibiotic was 0.0016 ,ug/ml as compared to 0.00~5 ,.g/ml for CLM-7 itself. Similarly, the amethoptenn (methotrexate)-resistant sublines DC-3F I A XXVII!, DC3F/AS, and CLM-7 jA XVIII showed no cross-reSIStance to actinomycin D; they had ED50 values of O.OOI~, 0.00l.2, and 0.0011 fJ.g/ml, respectively. The ED50 of actillomycm D for DC-SF was 0.0024 fJ.g/ ml (1). Results of uptake studies with tritiated (3H) actinomycin D by parent, sensitive DC-3F cells and experim~ntally derived sublines with either resistance or cross-reSIstance to actinomycin D were described. in (4). All sublines showed reduced uptake of the antibiotic. In a comparison of drug response in terms of ED50 and drug upta~e [counts per minute (cpm)fl06 cellsjl hr] as shown m text-figure I, there was inverse proportionality betwee? uptake and resistance. These r~sults confirmed our preVious radioautographic observatIons (1). Cell Morphology and Growth Pattern
During the development of actinomycin D-, da,unomycin-, and vincristine-resistant sublines, marked .differences in cell morphology and patterns of growth ill culture became apparent. The parent DC-~F ]ine gr~w as a loose, disordered network of fibroblastIc cells whIch, as cultures aged, formed small, discrete clumps of rounded cells and then large aggregates of loosely adhering cells (fig. I). In contrast, DC-3F / AD II (fig. 2) was composed
I(},OOO
1000
o
•
100L-~LL~~~~~~__~~~L-~~~~~
0.001
0.1
ED50 of Actinomycin
1.0
100
0
I.-Relationship between uptake of 3H-actinomycir D and response to actinomycin D in tenus of ED50 (see "Ma terials and Methods"). A, DC·SF; 0, DC-SF I AD II; ., DC-SF I AI IV: !:::.. DC·SF/AD X: ., DC-Sl'-/VCRd; and 0, DC.SF/DM XX.
TEXT-FIGURE
673
DRUG RESISTANCE AND PHENOTYPIC REVERSION TABLE
Cell line
DC-3F ________________________________ _ DC-3F I AD II_ - _____ - _________________ _ DC-3F I AD IV _________________________ _ DC-3F I AD X _____________________ - ____ _ DC-3F lAD X-U c ______________________ _ DC~3F IDM XX _______________________ _ DC-3F IVCRd _________________________ _ DC-3F IA XXVIII _____________________ _ DC-3F IA3 __-: _________________________ _ CLM-7N ______________________________ _ CLM-7T ______________________________ _ CLM 77 I AD III _______________________ _ CLM-7 I AD XV ________________________ _ CLM-7 I A XVIIL _______________ ---- __ _ CLM-7 IBUDR ________________________ _
I.-Properties of parent and drug-resistant cell lines
Drug maintenance concentration G (!,g/ml)
Increase in resistance b
0.1 1.0 10.0
81 376 2,450 48 1,269 650 47,600 108,400
Crossresistance to actinomycin D
(%)
83 415 32,200 >70
Saturation density (No. of cells XI04/cm 2)
279 125 0.5 0.5
13.0 13.5 14.5 15.0 13.0 18.0 24.5 19.0 13.5
51 46
91 80 73 67 66 23 31 8 99
0.7 1.1
11.5 11.5 16.5 16.5 16.0 13.5
55 53 63 40 72 55
54 69 53 24 52 81
1 1
0.1 1.0 50.0 50.0
Plating efficiency
(hr)
1
10.0 10.0 25.0 50.0
Population doubling time
54 57 69 45 42
• Refers to concentration of the selective agent indicated by cell-line designation. AD, actinomycin D; DM, daunomycin; VCR, vincristine; A, amethopterin (methotrexate); BVDR, 5-bromodeoxyuridine. • Resistance is expressed as ratio of ED50 values for resistant:drug-sensitive parent line, with respect to the selective agent for that subline. • DC-3F/ AD X-V had been maintained for 36-37 months without actinomycin D when determinations of resistance, doubling time, and saturation density were made.
of oriented arrays of elongated fibroblast-like cells. Only in old, crowded cultures was there some aggregate formation. The more actinomycin D-resistant DC-3F I AD IV (fig. 3) displayed an even more oriented pattern of growth. Cells were also more flattened, grew in compact arrays, and finally formed dense layers of contiguous cells. The highly resistant DC-3F I AD X subline was the most striking in its display of cell flattening and compact arrangement of cells indicative of increased cellsubstrate and cell-cell adhesiveness (fig. 4). Cells grew in islands until a dense, compact multilayer was formed. Of all the sublines it was the most difficult to detach from the glass or plastic substrate and thus required the longest exposure to agents such as trypsin or EDTA. After the subline had been maintained for 9 months in the presence of IO pg actinomycin D Iml, a subculture was made into dru~-free medium. This branch, designated DC-3F I AD X-U, was maintained for more than 3 years in the absence of antibiotic. The untreated subline gradually changed in morphology until it assumed an intermediate pattern between DC-3F and DC-3F I AD X (fig. 5). The cell line highly resistant to daunomycin, DC-3F ID M XX, resembled DC-3F I AD IV in degree of cell contiguity and flattening on the glass substrate (fig. 6); they were similarly resistant to actinomycin D (table I). Subline DC-3F IVCRd, approximately IOO-fold crossresistant to actinomycin D, grew as somewhat separated, flattened cells (fig. 7). Sublines DC-3F / A XXVIII and DC-3F I A3, after prolonged exposure to (and with high resistance to) amethopterin, associated with markedly elevated levels of dihydrofolate reductase activity (5), were compared to antibiotic-resistant sublines of DC-3F. Although there was a slight increase in cell flattening, cells grew in loose, disoriented arrays, and the DC3F I A3 subline showed focal mounding as the culture aged (figs. 8, 9). Thus sublines resistant or cross-resistant to actinomycin D showed the parallel orientation and/or cell flattening that often characterizes normal cells in culture, whereas parent DC-3F and its antifolate-resistant sublines had morphologic characteristics commonly associated with malignant cells. Drug-resistant sublines in the CLM-7 series were also
examined for morphologic differences. The CLM-7N line, during its first year of in vitro cultivation, had flattened, oriented cells that formed a compact monolayer (fig. 10). CLM-7T cells, after 18 months or more of continu'ous cultivation, were markedly altered and comprised loosely adhering, rounded, separated cells that formed focal aggregates as the culture aged (fig. II). Resistant sublines were selected from nontransformed CLM-7N cells about 4 months after establishment. The subline of lower (83-fold) resistance to actinomycin D grew as highly oriented, contiguous fibroblast-like cells (fig. 12), whereas independently derived CLM-7 lAD XV was composed of epithelial-like cells forming a flattened monolayer (fig. 13). Comparisons were made with sublines resistant to amethopterin and BUDR. CLM-7/A XVIII, highly resistant to antifolate, grew as a flattened layer (fig. 14). Unlike the amethopterin-resistant sublines of DC-3F, this subline had only slightly elevated dihydrofolate reductase activity; the primary determinant of resistance is probably altered transport (5). The BUDR-resistant subline (fig. 15) manifested the disorientation, separateness, and piling up of malignant cells. The trend toward normal growth patterns exhibited by antibiotic-resistant cell lines was further revealed by observations of colony morphology. The DC-3F series is shown in figures 16-24. Indication of strong intercellular adhesiveness was most marked for the most actinomycinresistant subline, DC-3F I AD X (fig. 19). The untreated line DC-3F I AD X-U was intermediate (fig. 20), somewhat resembling DC-3F I AD II (fig. 17). The daunomycinresistant subline (fig. 21) was strongly adherent to substrate, as was DC-3F IVCRd (fig. 22), though the latter appeared to have less cell-cell adhesiveness and grew as a spread-out, somewhat epithelioid colony. Colonies of the amethopterin-resistant lines (figs. 23, 24) consisted of loosely attached, randomly oriented cells. In the CLM-7 series, parent CLM-7N, the two actinomycin D-resistant sublines, and the transport variant subline CLM-7/A XVIII showed highly oriented growth patterns (figs. 25, 27-29), whereas CLM-7/BUDR resembled CLM-7T in its transformed-type colony morphology (figs. 26, 30).
674
BIEDLER, RIEHM, PETERSON, AND SPENGLER
Tumorigenicity
Drug-sensitive and drug-resistant cell lines were tested for their capacity to form tumors in the cheek pouches of conditioned Syrian hamsters (table 2). Parent DC-3F cells were tumorigenic, with an 82% tumor frequency over a 3-year test period. In contrast, the three actinomycin D-resistant sublines showed reduced or abrogated tumorigenicity according to degree of resistance. DC-3F j AD X cells, with a greater than 2,OOO-fold increase in resistance to the antibiotic, produced no tumors even when the inoculum level was raised to 3 million cells/ pouch. DC-SF jDM XX and DC-3F jVCRd, exhibiting cross-resistance to actinomycin D of 279-fold and 125-fold, respectively, produced few tumors. The two amethopterin-resistant sublines were as tumorigenic as parent DC-3F cells even after long exposure to high concentrations of anti folate. CLM-7T cells, considered transformed when tumorigenic after 1.5 years of cultivation in vitro, produced tumors at a frequency of 52% in seven experiments. ~'Spontaneous" traI?-sformation occurred independently m two subpopulatlons of CLM-7 (see "Materials and Methods"). All drug-resistant sublines were selected from CLM-7N cells before malignant transformation. The actinomycin D-selected subline with the higher level of resistance was weakly tumorigenic at a time when parent cells maintained in parallel had already transformed and were producing tumors at a substantially higher frequency. Subline CLM-77 lAD III, exhibiting an 83fold increase in resistance, produced tumors with a frequency similar to that of transformed parent cells. The BVDR-resistant sub line was weakly tumorigenic, and amethopterin-resistant cells produced no tumors. The comparativelv small size and late appearance ot CLM-77 lAD III tumors prompted assessment of cell lines for tumor growth potential in addition to tumor frequency (text-fig. 2). Apparently, tumors produced by actinomycin D-resistant sublines derived from either DC3F ?~ CLM-7N cells grew much less than did drugsensItIve cells, whereas tumors produced by amethopterinresistant sublines of DC-3F more closely resembled parent cell tumors in size. The decrease in tumor size at week 4, indicated for DC-3F and DC-3F I AS cells, is due in part to tumor necrosis and/or rejection and in part to
TABLE
death of animals with very large tumors before the fourth measurement. Correlation Between Drug Resistance and Tumorigenicity
. Since tumorigenicity determinations of DC-SF cell lines mdicated a positive correlation between increase in resistance to actinomycin D and decrease in oncogenic potential, the relationship between these parameters was further investigated by a different approach. DC-SF I AD X-V cells, derived from the non tumor-producing DC SF I AD X subline by maintenance in drug-free medium, were tested at intervals for response to actinomycin D and for tumorigenic capacity (table 3) . With decrease in r~sistance to drug, there was increase in tumor frequency. SlI'~ce the number of tumorgenici ty tests at each timepomt was small, we could not make precise correlations with degee of resistance. During the plateau period between .622 and 1,155 ?~ys of growth without drug, when resIstance had stabIlIzed at a level approximately 55-fold to 80-fold higher than parent DC-SF cells, the overall frequency of tumor production was 40% (27 tl~morsj68 cheek pouches). This frequency is in accord WIth the expected, based on comparable results with DC-S!, I AD ~I cells that exhibited a 27% tumor frequency assoCIated wIth an 80-fold increase in resistance to actinomycin D. Similarly, oncogenic potential of DC-SF j AD X-V cells, as mamfested by tumor size attained within the experimental period, resembled that of the DCSF I AD II subline (test-fig. 2) . Even after DC-3F / AD X-V cells were grown for m
4
Robert H.
F. Peterson,3 and Barbara A. Spengler 3,5
SUMMARY-Development of resistance to actinomycin D, dau· nomycin, or vincristine in Chinese hamster cells growing in vitro resulted in reversion to or retention of normal phenotypes in comparison to spontaneously transformed drug-sensitive parent populations. Sublines resistant or cross-resistant to actinomycin D showed reduced uptake of antibiotic in proportion to degree of resistance. The cells with acquired resistance were either weakly tumorigenic or nontumorigenic when tested in the cheek pouches of weanling Syrian hamsters treated with cortisone. In contrast, parent cells and several amethopterin (methotrexate)-resistant sublines produced many tumors. Antibiotic- and Vinca-alkaloid-resistant cell lines showed oriented growth patterns often associated with the behavior of normal cells in culture; antibiotic-sensitive, tumorigenic lines had morphologic characteristics of malignant cells. The altered cell membrane properties that accompanied development of resistance or cross-resistance to actinomycin D appeared to account also for the lower oncogenic potential and greater cell adhesiveness of resistant cells relative to their malignant counterparts.-J Natl Cancer Inst 55: 671-{)80, 1975.
In Chinese hamster cells exposed to and resistant to actinomycin D, daunomycin (daunorubicin), or vincristine, resistance was most likely a consequence of cell membrane alteration(s) which made them less drugpermeable (1-4). These sublines exhibited increases in resistance or cross-resistance to actinomycin D of approximately 80-fold to 2,500-fold. We have noted striking differences between drug-sensitive progenitor cells and resistant derivatives in respect to cell morphology and cell growth patterns in vitro. For spontaneously transformed parent DC-3F cells in particular, development of antibiotic or Vinca-alkaloid resistance was accompanied by reversion of morphologic characteristics associated with malignant cells in culture toward those associated with normal cells. In contrast, a large series of experimentally derived amethopterin (methotrexate)-resistant sublines (5) continued to resemble the parent DC-3F population even after prolonged cultivation at high concentrations of drug. Therefore, we systematically examined the apparent correlation between resistance or cross-resistance to actinomycin D and manifestation of normal cell phenotype in a larger group of drug-resistant Chinese hamster sublines. Ability to form tumors in the cheek pouches of cortisone-treated Syrian hamsters served as the primary criterion of malignancy; correlations were made with morphology in culture and degree of densitydependent growth inhibition. Results were partially summarized in a preliminary report (6). Tumor cell populations established in vitro may, after long periods in culture, lose the ability to produce tumors in the host of origin (7-11). Methods have been devised for experimental selection of "revertant" cells that have lost one or more of the properties characteristic of the transformed population (12). For example, variants with reduced tumorigenicity and increased growth control in vitro were indirectly selected with 5-fluoro-2'deoxy-,B-uridine (FUDR) from polyoma virus and simian
virus 40 (SV40)-transformed mouse and hamster fibroblasts, as described by Pollack et a!. (13). Rabinowitz and Sachs (14, 15) obtained variant clones from polyomaor dimethylnitrosamine-transformed hamster cells grown on glutaraldehyde-fixed normal cell monolayers. The variants showed reversion of properties of transformed cells, including decreases in tumorigenicity and saturation density. In still another system comprised of clonal hydrocarbon-transformed mouse fibroblasts, Mondal et a1. (16) obtained variant clones of reduced malignancy accompanied by marked lowering of saturation density. However, no clear differences in morphology were observed. We now describe Chinese hamster cells with acquired resistance to a variety of cancer-chemotherapeutic agents. Development of resistance associated with altered membrane permeability resulted in production or retention of cell populations with a less malignant phenotype than that of their spontaneously transformed counterparts. We refer to these drug-resistant cells as permeability "variants" or as "revertants," without implication as to pathway or mechanism of variant or revertant production. In the case of the twice-cloned, spontaneously transformed, fibroblastic DC-3F cells, revertants emerged after prolonged cultivation in progressively higher concentrations of actinomycin D, daunomycin, or vincristine. In the once-cloned, fibroblastic CLM-7 cell line established in our laboratory, resistant variants selected by actinomycin D or amethopterin (methotrexate) retained a normal phenotype (though it differed somewhat from that of the non transformed population from which they were derived). In contrast, drug-sensitive, parent CLM-7 cells underwent spontaneous neoplastic and morphologic transformation. MATERIALS AND METHODS
Cells and culture medium.-Parent cell lines DC-3F and CLM-7 were established in vitro from lung and bone marrow, respectively, of normal Chinese hamsters; they grow in monolayer culture and have chromosome numbers in the diploid range (1). DC-3F is a "spontaneously" transformed line, as determined from tumorigenicity testing. CLM-7 cells found non tumorigenic during the first 12 months after establishment were designated CLM7N ("N" denotes normal). After 18 months in continuous culture, cells underwent malignant transformation, as determined from tumorigenicity assays, and were designated CLM-7T ("Tn for transformed). CLM-7N cells frozen in liquid nitrogen after 4 months in culture and Received March 7, 1975; accepted May 9, 1975. Supported in part by Public Health Service research grant CA08748 from the National Cancer Institute. 3 Memorial Sloan-Kettering Cancer Center, 410 East 68th St., New York, N.Y. 10021. 4 Present address: Kinderklinik der Freien Universitat, Heubnerweg 6, 1 Berlin 19, Germany. 5 We thank Dr. Dorris J. Hutchison for her support and Marion Linczer Katz, Daryl D. Hill, Roberta A. Devaney, David Kessler, and Judith A. O'Neil for their participation in this study. 1
2
JOURNAL OF THE NATIONAL CANCER INSTITUTE, VOL. 55, NO.3, SEPTEMBER 1975
671
672
BIEDLER, RIEHM, PETERSON, AND SPENGLER
later reestablished in vitro again underwent malignant transformation. This branch was also designated CLM7T; data obtained from both branches were combined. Drug-resistant sublin~s wer~ developed ?y exposure of mass populations to mcreasmg concentratIons of drug and were maintained in the presence of drug at the final concentrations listed in table 1. Most resistant sublines were described previously with respect to response to selective agent, cross-resistance, karyotype, and me~ha nisms of resistance (1-5). The CLM-7/BUDR subhn~, able to grow in the presence of 500 pg 5-bromodeoxyundine (BUDR)/ml, was maintained at 50 p.g/ml. Cell lines were propagated in Eagle's minimum essential medium supplemented with 10% fet~l bovine serum, penicillin (100 IU Iml), and streptomYCIn (100 p.g/ml). Drug-treated lines were routinely grown in the absence of drug for 10-15 days with at least two culture transfers before all experimental determinations were made. Highly resistant DC-3F I AD X cells, after maintenance at 10 p.g actinomycin D /ml for approximately 9 mont~s, were subdivided and grown in drug-free medium III parallel with a treated population. The subline maintained in the absence of drug was designated DC-3F I AD X-U and was studied for about 3 years. Resistance to actinomycin D was assayed at 3- to 6-week intervals. A plot of these data shows the response to antibiotic that corresponds to the number of days in drug-free medium. at the time of heterotransplantability testing. Drug sensitivity assay.-Response of cell lines to chemical agents was determined in a 72-hour assay based on relative cell counts in control and drug-treated cultures (1). Drug response was specified by a mean effective dose (ED 50) value: the drug concentration permitting growth to 50% of control cell numbers. Degree of resistance is expressed as the ratio of EDSO values for resistant cells to paren t cells.
Plating efficiency, population .doubling time, and saturation density.-For estimates of plating efficiency, 200±20 cells were plated in 5 ml medium in 60-mm glass plates. Colonies were fixed and stained after 7 to 10 days. Only colonies of more than 5 cells were scored. Population doubling time and saturation density estimates were obtained within the same experiment, with an inoculum of 1 X 10 5 cells in 5 ml medium/60-mm glass plate. Medium was replaced twice a week. Duplicate plates w~re counted at appropriate intervals during the exponentIal growth phase to obtain the average population doubling time and during the stationary growth phase to obtain the maximum cell number /20-cm2 plate. All values are based on 2--4 replicate experiments. Heterotransplantation tests.-For determination of the tumor-producin'4 capacities of parent and drug-resistant sublines, I million, or occasionally 3 million, cells were inoculated into cheek pouches of 19- to 22-day-01d female weanling Syrian hamsters as previously described (17). Animals received 2.5 mg cortisone acetate sc at the time of cell inoculation, 3 days later, and twice a week thereafter. Cheek pouches were examined once a week for 4 weeks and tumors were measured in 3 dimensions. Samples were excised for histopathologic examination. Actinomycin D uptake determination.-Cells were exposed to tritiated actinomycin D [2.0 p-Ci/ml, 2.0p-g/ml (1.6 X 10-6 M), total external concentration] for 1 hour and prepared for radioactivity measurement as detailed in (4). Samples were counted in a Packard Liquid Scintillation Spectrometer. Tritiated actinomycin D (specific
activity, 8.4 Ci/mmole) was obtained from Schwarz/ Mann, Orangeburg, New York. RESULTS Drug Resistance
Table I lists all cell lines. selective concentrations of drugs, experimentally determined increases in resistance, and degree ~f cross-resistance to actinomy~i~ ~ for sublines not dIrectly selected by that antibIOtIC. Doseresponse data for most lines have been reported elsewhere (1,2,4,5). The BUDR-resistant subline CLM-7/BUDR exhibited no cross-resistance to actinomycin D; the ED 50 of the antibiotic was 0.0016 ,ug/ml as compared to 0.00~5 ,.g/ml for CLM-7 itself. Similarly, the amethoptenn (methotrexate)-resistant sublines DC-3F I A XXVII!, DC3F/AS, and CLM-7 jA XVIII showed no cross-reSIStance to actinomycin D; they had ED50 values of O.OOI~, 0.00l.2, and 0.0011 fJ.g/ml, respectively. The ED50 of actillomycm D for DC-SF was 0.0024 fJ.g/ ml (1). Results of uptake studies with tritiated (3H) actinomycin D by parent, sensitive DC-3F cells and experim~ntally derived sublines with either resistance or cross-reSIstance to actinomycin D were described. in (4). All sublines showed reduced uptake of the antibiotic. In a comparison of drug response in terms of ED50 and drug upta~e [counts per minute (cpm)fl06 cellsjl hr] as shown m text-figure I, there was inverse proportionality betwee? uptake and resistance. These r~sults confirmed our preVious radioautographic observatIons (1). Cell Morphology and Growth Pattern
During the development of actinomycin D-, da,unomycin-, and vincristine-resistant sublines, marked .differences in cell morphology and patterns of growth ill culture became apparent. The parent DC-~F ]ine gr~w as a loose, disordered network of fibroblastIc cells whIch, as cultures aged, formed small, discrete clumps of rounded cells and then large aggregates of loosely adhering cells (fig. I). In contrast, DC-3F / AD II (fig. 2) was composed
I(},OOO
1000
o
•
100L-~LL~~~~~~__~~~L-~~~~~
0.001
0.1
ED50 of Actinomycin
1.0
100
0
I.-Relationship between uptake of 3H-actinomycir D and response to actinomycin D in tenus of ED50 (see "Ma terials and Methods"). A, DC·SF; 0, DC-SF I AD II; ., DC-SF I AI IV: !:::.. DC·SF/AD X: ., DC-Sl'-/VCRd; and 0, DC.SF/DM XX.
TEXT-FIGURE
673
DRUG RESISTANCE AND PHENOTYPIC REVERSION TABLE
Cell line
DC-3F ________________________________ _ DC-3F I AD II_ - _____ - _________________ _ DC-3F I AD IV _________________________ _ DC-3F I AD X _____________________ - ____ _ DC-3F lAD X-U c ______________________ _ DC~3F IDM XX _______________________ _ DC-3F IVCRd _________________________ _ DC-3F IA XXVIII _____________________ _ DC-3F IA3 __-: _________________________ _ CLM-7N ______________________________ _ CLM-7T ______________________________ _ CLM 77 I AD III _______________________ _ CLM-7 I AD XV ________________________ _ CLM-7 I A XVIIL _______________ ---- __ _ CLM-7 IBUDR ________________________ _
I.-Properties of parent and drug-resistant cell lines
Drug maintenance concentration G (!,g/ml)
Increase in resistance b
0.1 1.0 10.0
81 376 2,450 48 1,269 650 47,600 108,400
Crossresistance to actinomycin D
(%)
83 415 32,200 >70
Saturation density (No. of cells XI04/cm 2)
279 125 0.5 0.5
13.0 13.5 14.5 15.0 13.0 18.0 24.5 19.0 13.5
51 46
91 80 73 67 66 23 31 8 99
0.7 1.1
11.5 11.5 16.5 16.5 16.0 13.5
55 53 63 40 72 55
54 69 53 24 52 81
1 1
0.1 1.0 50.0 50.0
Plating efficiency
(hr)
1
10.0 10.0 25.0 50.0
Population doubling time
54 57 69 45 42
• Refers to concentration of the selective agent indicated by cell-line designation. AD, actinomycin D; DM, daunomycin; VCR, vincristine; A, amethopterin (methotrexate); BVDR, 5-bromodeoxyuridine. • Resistance is expressed as ratio of ED50 values for resistant:drug-sensitive parent line, with respect to the selective agent for that subline. • DC-3F/ AD X-V had been maintained for 36-37 months without actinomycin D when determinations of resistance, doubling time, and saturation density were made.
of oriented arrays of elongated fibroblast-like cells. Only in old, crowded cultures was there some aggregate formation. The more actinomycin D-resistant DC-3F I AD IV (fig. 3) displayed an even more oriented pattern of growth. Cells were also more flattened, grew in compact arrays, and finally formed dense layers of contiguous cells. The highly resistant DC-3F I AD X subline was the most striking in its display of cell flattening and compact arrangement of cells indicative of increased cellsubstrate and cell-cell adhesiveness (fig. 4). Cells grew in islands until a dense, compact multilayer was formed. Of all the sublines it was the most difficult to detach from the glass or plastic substrate and thus required the longest exposure to agents such as trypsin or EDTA. After the subline had been maintained for 9 months in the presence of IO pg actinomycin D Iml, a subculture was made into dru~-free medium. This branch, designated DC-3F I AD X-U, was maintained for more than 3 years in the absence of antibiotic. The untreated subline gradually changed in morphology until it assumed an intermediate pattern between DC-3F and DC-3F I AD X (fig. 5). The cell line highly resistant to daunomycin, DC-3F ID M XX, resembled DC-3F I AD IV in degree of cell contiguity and flattening on the glass substrate (fig. 6); they were similarly resistant to actinomycin D (table I). Subline DC-3F IVCRd, approximately IOO-fold crossresistant to actinomycin D, grew as somewhat separated, flattened cells (fig. 7). Sublines DC-3F / A XXVIII and DC-3F I A3, after prolonged exposure to (and with high resistance to) amethopterin, associated with markedly elevated levels of dihydrofolate reductase activity (5), were compared to antibiotic-resistant sublines of DC-3F. Although there was a slight increase in cell flattening, cells grew in loose, disoriented arrays, and the DC3F I A3 subline showed focal mounding as the culture aged (figs. 8, 9). Thus sublines resistant or cross-resistant to actinomycin D showed the parallel orientation and/or cell flattening that often characterizes normal cells in culture, whereas parent DC-3F and its antifolate-resistant sublines had morphologic characteristics commonly associated with malignant cells. Drug-resistant sublines in the CLM-7 series were also
examined for morphologic differences. The CLM-7N line, during its first year of in vitro cultivation, had flattened, oriented cells that formed a compact monolayer (fig. 10). CLM-7T cells, after 18 months or more of continu'ous cultivation, were markedly altered and comprised loosely adhering, rounded, separated cells that formed focal aggregates as the culture aged (fig. II). Resistant sublines were selected from nontransformed CLM-7N cells about 4 months after establishment. The subline of lower (83-fold) resistance to actinomycin D grew as highly oriented, contiguous fibroblast-like cells (fig. 12), whereas independently derived CLM-7 lAD XV was composed of epithelial-like cells forming a flattened monolayer (fig. 13). Comparisons were made with sublines resistant to amethopterin and BUDR. CLM-7/A XVIII, highly resistant to antifolate, grew as a flattened layer (fig. 14). Unlike the amethopterin-resistant sublines of DC-3F, this subline had only slightly elevated dihydrofolate reductase activity; the primary determinant of resistance is probably altered transport (5). The BUDR-resistant subline (fig. 15) manifested the disorientation, separateness, and piling up of malignant cells. The trend toward normal growth patterns exhibited by antibiotic-resistant cell lines was further revealed by observations of colony morphology. The DC-3F series is shown in figures 16-24. Indication of strong intercellular adhesiveness was most marked for the most actinomycinresistant subline, DC-3F I AD X (fig. 19). The untreated line DC-3F I AD X-U was intermediate (fig. 20), somewhat resembling DC-3F I AD II (fig. 17). The daunomycinresistant subline (fig. 21) was strongly adherent to substrate, as was DC-3F IVCRd (fig. 22), though the latter appeared to have less cell-cell adhesiveness and grew as a spread-out, somewhat epithelioid colony. Colonies of the amethopterin-resistant lines (figs. 23, 24) consisted of loosely attached, randomly oriented cells. In the CLM-7 series, parent CLM-7N, the two actinomycin D-resistant sublines, and the transport variant subline CLM-7/A XVIII showed highly oriented growth patterns (figs. 25, 27-29), whereas CLM-7/BUDR resembled CLM-7T in its transformed-type colony morphology (figs. 26, 30).
674
BIEDLER, RIEHM, PETERSON, AND SPENGLER
Tumorigenicity
Drug-sensitive and drug-resistant cell lines were tested for their capacity to form tumors in the cheek pouches of conditioned Syrian hamsters (table 2). Parent DC-3F cells were tumorigenic, with an 82% tumor frequency over a 3-year test period. In contrast, the three actinomycin D-resistant sublines showed reduced or abrogated tumorigenicity according to degree of resistance. DC-3F j AD X cells, with a greater than 2,OOO-fold increase in resistance to the antibiotic, produced no tumors even when the inoculum level was raised to 3 million cells/ pouch. DC-SF jDM XX and DC-3F jVCRd, exhibiting cross-resistance to actinomycin D of 279-fold and 125-fold, respectively, produced few tumors. The two amethopterin-resistant sublines were as tumorigenic as parent DC-3F cells even after long exposure to high concentrations of anti folate. CLM-7T cells, considered transformed when tumorigenic after 1.5 years of cultivation in vitro, produced tumors at a frequency of 52% in seven experiments. ~'Spontaneous" traI?-sformation occurred independently m two subpopulatlons of CLM-7 (see "Materials and Methods"). All drug-resistant sublines were selected from CLM-7N cells before malignant transformation. The actinomycin D-selected subline with the higher level of resistance was weakly tumorigenic at a time when parent cells maintained in parallel had already transformed and were producing tumors at a substantially higher frequency. Subline CLM-77 lAD III, exhibiting an 83fold increase in resistance, produced tumors with a frequency similar to that of transformed parent cells. The BVDR-resistant sub line was weakly tumorigenic, and amethopterin-resistant cells produced no tumors. The comparativelv small size and late appearance ot CLM-77 lAD III tumors prompted assessment of cell lines for tumor growth potential in addition to tumor frequency (text-fig. 2). Apparently, tumors produced by actinomycin D-resistant sublines derived from either DC3F ?~ CLM-7N cells grew much less than did drugsensItIve cells, whereas tumors produced by amethopterinresistant sublines of DC-3F more closely resembled parent cell tumors in size. The decrease in tumor size at week 4, indicated for DC-3F and DC-3F I AS cells, is due in part to tumor necrosis and/or rejection and in part to
TABLE
death of animals with very large tumors before the fourth measurement. Correlation Between Drug Resistance and Tumorigenicity
. Since tumorigenicity determinations of DC-SF cell lines mdicated a positive correlation between increase in resistance to actinomycin D and decrease in oncogenic potential, the relationship between these parameters was further investigated by a different approach. DC-SF I AD X-V cells, derived from the non tumor-producing DC SF I AD X subline by maintenance in drug-free medium, were tested at intervals for response to actinomycin D and for tumorigenic capacity (table 3) . With decrease in r~sistance to drug, there was increase in tumor frequency. SlI'~ce the number of tumorgenici ty tests at each timepomt was small, we could not make precise correlations with degee of resistance. During the plateau period between .622 and 1,155 ?~ys of growth without drug, when resIstance had stabIlIzed at a level approximately 55-fold to 80-fold higher than parent DC-SF cells, the overall frequency of tumor production was 40% (27 tl~morsj68 cheek pouches). This frequency is in accord WIth the expected, based on comparable results with DC-S!, I AD ~I cells that exhibited a 27% tumor frequency assoCIated wIth an 80-fold increase in resistance to actinomycin D. Similarly, oncogenic potential of DC-SF j AD X-V cells, as mamfested by tumor size attained within the experimental period, resembled that of the DCSF I AD II subline (test-fig. 2) . Even after DC-3F / AD X-V cells were grown for m
