- Email: [email protected]
Cytogenetic observations of a human ovarian carcinoma clinically resistant to therapy
Cytogenetic observations of a human ovarian carcinoma, clinically resistant to therapy Jacob Rotmensch, MD; Thomas E. Turkula, MS," Ralph R. Weichselhaum, MD," and Jeffrey L. Schwartz, PhDh Chicago, Illinois Resistance of a cancer cell to therapy represents an important problem in tumor biology. Hypotheses concerning the mechanism of resistance have included genetic instability, gene amplification, aneuploidy, and altered cell growth kinetics. Cytogenetic assays allow for the analysis of each of these parameters and provide important information concerning tumor heterogeneity. In using cytogenetic analysis, we have analyzed a human ovarian carcinoma clinically resistant to therapy. The ovarian carcinoma had a complex but stable karyotype with a mean chromosome number of 61.7 chromosomes/cell. Approximately 50% of the metaphase preparations examined had double-minute chromosomes, which have sometimes been associated with gene amplification. The frequency of double minutes varied from one pair to hundreds of pairs per cell. Cell cycle kinetic analysis revealed an average generation time of 38 hours. The cells were relatively resistant to doxorubicin in vitro, and the baseline sister chromatid exchange frequency, a measure of genetic instability, was elevated. These results suggest that cytogenetic assays have potential as predictive assays of tumor chemoresistance and may provide information regarding the biologic aggressiveness encountered clinically. (AM J OssTET GYNECOL 1988;159:1099-103.)
Key words: Doxorubicin, human ovarian tumor cells, sister chromatid exchanges
Resistance of cancer cells to cytotoxic therapy remains an' important problem in cancer biology. The reason for chemotherapy or radiotherapy failure is unknown. Clinical studies have shown that tumor stage, grade, and cell type are important prognostic parameters, but even within a specific class of tumors, there is much variability in response. 1 Although evidence has accumulated to suggest that the presence of chemoresistant cells in pretreatment biopsy specimens will predict therapeutic failure, the presence of chemosensitive cells in a pretreatment biopsy specimen will not predict successful treatment. 2 · " Therefore investigations into other possible parameters associated with tumor resistance are being examined, such as deoxyribonucleic acid (DNA) content and tumor growth rates. Cytogenetic analysis affords an investigator the means to examine multiple parameters relatively quickly using the same material. One can determine relative in vitro chemosensitivity, cell generation time,
the degree of aneuploidy, and whether large-scale gene amplification has occurred. Also, with regard to all these parameters, because individual cells are analyzed, tumor heterogeneity can be determined. In this study we have examined a human ovarian epithelial carcinoma cytogenetically and compared it with normal, nontransformed, human fibroblasts. Each cell line was karyotyped and the degree of aneuploidy was determined. Cells were then analyzed for the baseline and doxorubicin-induced sister chromatid exchange frequency. Sister chromatid exhanges represent reciprocal exchanges between homologous daughter chromatids. Although the mechanism and significance of sister chromatid exchange formation remain unknown, they have been found to be sensitive and reliable markers of genetic instability and cellular sensitivity to cytotoxic agents." 8 Finally, the average generation time was determined in these cells and used as a measure of tumor cell growth rate.
From the Division of Gynecologic Oncology, Department of Obstetrics and Gynecology" and the Department of Radiation Oncology,' Pritzker School of Medicine, The University of Chicago. Supported by American Cancer Society Career Development Awards U- R. and]. L. S.), the Cancer Research Foundation U· R.), and Grant CA 42596 U· L. S. and R.R. W.) from the National Cancer Institute. Received for publication July 30, 1987; revised January 8, 1988, and May 4, 1988; accepted May 31, 1988. Reprint requests: Jacob Ronnensch, MD, Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, The University of Chicago, 5841 S. Maryland Ave., Chicago, IL 60637.
Material and methods Patient characteristics. A 48-year-old female who had a right pleural effusion and pelvic mass extending to the umbilicus underwent a thoracentesis and exploratory laparotomy in August 1984 for a stage IV, grade 2, serous cystadenocarcinoma of the ovary. After cytoreductive surgery, she had five courses of hexamethylmelamine, cytoxan, doxorubicin (Adriamycin), and cisplatin. As a result of thrombocytopenia, hexamethylmelamine and doxorubicin were discontinued, 1099
1100
Rotmensch et al.
Am
J
November 1988 Obstet Gynecol
c Fig. I. Fluorescence plus Giemsa-stained preparations of cells that went through one (A), two (B), or three (C) DNA replications.
and seven additional courses of cytoxan and cisplatin were given. After disease progression, cisplatin and a continuous infusion of 5-fluorouracil were given. With the development of ascites, a 500 ml aliquot was removed by paracentesis, and the carcinoma cells were cultured in vitro. In vitro establishment of ovarian carcinoma and maintenance of carcinoma and normal fibroblast cell lines. Two liters of ascites were centrifuged at 500 g. The pellet was removed and suspended in 5 ml of 5% serum culture medium. The culture medium consisted of 72.5% Dulbecco's modification of Eagle's medium, 22.5% Ham's nutrient mixture F-12, 5% fetal bovine serum, 20 ng/ml of epidermal growth factor (added the third day) after plating, 5 µg/ml of transferrin, 2 x 10- 11 mol/L, 3,3',5 triiodo-L-thyron ine, 10- 10 mol/L cholera toxin, 1.8 x 10- 4 mol/L adenine, 0.4 µg I ml of hydrocortisone, 50 U I ml of my cos ta tin, which was omitted after the second passage, 100 U I ml of penicillin, and 100 µg/ml of streptomycin. Lethally irradiated embryonic 3T3 fibroblasts were added at a density 2 x 104 cells/cm 2 as a feeder layer when necessary and selectively removed by 15- to 30-second incubations in 0.02% ethylenediamine tetra-acetate. After primary culture, the malignant epithelial ovarian carcinoma cells were separated, characterized, and in-
dividually plated. The epithelial cells were subcultured, and after 2 to 3 weeks, individual colonies were obtained with40diameter of 0.5 to 1.0 cm. Human fibroblasts (AG 1522) were grown in tissue 4 5 culture flasks at stock concentrations of 10 to 10 cells in modified Eagle's medium with 20% fetal calf serum, 1% L-glutamine, 1% penicillin and streptomycin, and 0.5% supplement-B. Supplement-B contained 9 gm of o-glucose and 66 mg of sodium pyruvate/ dl of Earle's balanced salt solution. The fibroblasts (AG 1522) formed a monolayer on tissue culture surfaces and were harvested by scraping or using trypsinethylenediamine tetra-acetate. Cytogenetic analysis. Five parameters were examined for each cell line: (1) chromosome number per cell, (2) the proportion of cells with evidence of gene amplification, (3) the average generation time of each cell line, (4) the baseline sister chromatid exchange frequency, and (5) the doxorubicin-sen sitivity of each cell line. For analysis, exponentially growing cells were plated at a density of 5 x 105 cells in 15 ml of complete medium in a T-75 flask. Twenty-four hours later, cells were exposed from 0 to 2 x 10- 6 mol/L doxorubicin for 1 hour and cultured for 48 to 72 hours in complete medium supplemented with 10-s mol/L bromodeoxy-
Volume 159 Number 5
Cytogenetic markers of human ovarian carcinoma
1101
25
45
50
55
60
65
70
75
80
85
90
>95
NUMBER OF CHROMOSOMES/CELL
Fig. 2. Frequency of chromosome number in the ovarian carcinoma OVC-1.
uridine. During the last 2 hours, 2 x 10- 1 mol/L colcemid was added to each culture. Cells were removed from the flasks by trypsin treatment, incubated with 0.075 mol/L KC! for 15 minutes, and then fixed in 3: 1 methanol to acetic acid. Cells were placed onto slides, and the slides were then dried overnight. The degree of aneuploidy and proportion of cells with evidence of gene amplification were determined in G-banded cells. Cells were G-banded with 0.1 % trypsin and 2% Gurr-buffered Giemsa. 9 One hundred cells were anlayzed for chromosome number and the presence of either double minutes or homogeneous staining regions. Double minutes homogenous staining regions are both cytogenetic correlates of gene amplification. Double minutes are small, paired, spherical chromosomes, whereas homogenous staining regions are chromosome regions that fail to exhibit the usual differential banding pattern.' Average generation time and baseline and induced sister chromatid exchange frequency were determined in cells stained with a fluorescence plus Giemsa technique.10 Briefly, slides were stained for 15 minutes in Hoechst 33258 (0.5 µg/ml), mounted in 0.067 mol/L Sorensen's buffer (pH 6.8), placed on a slide warmer at 55° C, and exposed to light from two General Electric 15 W black lights for 4 to 10 minutes. They were rinsed in buffer and stained in 3% Giemsa in buffer. The average generati~n time was estimated from the proportion of cells that had gone through one (P 1), two
Table I. Cytogenetic characteristics of ovarian carcinoma and normal cell line Ovarian carcinoma
Mean chromosome No./ cell Modal chromosome No./cell Chromosome No. range % cells with double minutes Average generation time
Normal cell line (1522)
61.7 62 46-95 48 38.l
46 46 46 0 24.0
0.28 ± 0.01
0.15 ± 0.02
(hr) Baseline sister chromatid exchange frequency I cell Rate of doxorubicininduced sister chromatid exchange/ chromosom (per µmol/L experiment)
0.12
0.62
(P2 ), or three (P,) rounds of DNA replication. Average
generation time is defined as follows: AGT = time in culture/P 1 + 2 (P 2 ) + 3 (P,). The proportion of cells that had gone through one, two, or three rounds of DNA replication was determined on the fluorescence plus Giemsa-stained slides, as described by Tice et al. 11 First-division cells contain chromosomes, with both sister chromatids stained uniformly dark. Second-division cells have chromosomes with one chromatid darkly stained and its sister lightly stained. Third-division cells contain some chromosomes, with one chromatid darkly stained and its sister lightly stained, whereas the rest of the chromosomes are uniformly stained lightly (Fig. 1). One hundred cells were anlayzed for average generation time; 50 cells
1102 Rotmensch et al.
November 1988 Am J Obstet Gynecol
0.5
0.4 Q)
E
0
Ill
0
E
0.3
0 .....
.r:; ()
......
w
() (/)
o Ovarian e Normal 0
2
4
Carcinoma
6
8
10
Dose (x 10- 7 m) Fig. 4. Dose response of tumor (O) and normal(•) cells to sister chromatid exchange (SCE) induction by doxorubicin.
Fig. 3. Partial meta phases of an ovarian cell line show multiple (A) and single double-minutes (DM) (B).
were analyzed for baseline and doxorubicin-induced sister chromatid exchange frequency. Results
Chromosome number and constitution. The ovarian carcinoma cells had an aneuploid distribution with a variable chromosome number per cell. In 100 ovarian carcinoma cells analyzed, the chromosome number ranged from 46 to 98 chromosomes/cell with a modal number of 62 and a mean number of 61.7 chromosomes/ cell (Fig. 2). Table I compares the cytogenetic characteristics of the ovarian carcinoma to benign cells. The karyotype was complex, with numerous rearrangement and duplications. In the metaphase preparations examined, there were variable numbers of double minutes. The number of double minutes in each cell preparation ranged from one pair to numerous pairs (Fig. 3), with greater than 10 pairs in 48% of the 100 cells examined. In subsequent in vitro passages, the frequency of double minutes declined. No homogenous staining regions were found. Average generation time, sister chromatid exchange frequency, and sensitivity to doxorubicin. The ovarian tumor line grew well and had a plating efficiency of 10%. Cell doubling times, estimated by av-
erage generation time measurements, were 38 hours for the ovarian cells and 24 hours for the normal fibroblasts. The baseline frequency of sister chromatid exchanges was significantly higher in the ovarian cell line (0.28 ± 0.01 sister chromatid exchange/chromosome) compared with the fibroblast cells (0.15 ± 0.02 sister chromatid exchange/ chromsome). The ovarian cells were also more resistant to the induction of sister chromatid exchanges by doxorubicin (Fig. 4). Sister chromatid exchanges were' induced at a rate of 0.12 sister chromatid exchange/µmol/L exposure of doxorubicin in ovarian carcinoma cells and at a rate of 0.62 sister chromatid exchange I µmoll L exposure of doxorubicin in 1522 cells. There was no evidence of heterogeneity in either baseline or induced sister chromatid exchange responses. Comment
The mechanisms for resistance of tumor cells to cancer therapy remain unknown. Many believe that resistance is caused by inherent genetic changes. By understanding those genetic events and the conditions that predispose them to those events, more rationale approaches to the treatment of cancer should develop. In an attempt to correlate in vitro cellular findings with tumor behavior and clinical outcome, we examined an ovarian tumor clinically resistant to cancer therapy for the following: ( 1) chromosome constitution, (2) evidence of gene amplification, (3) average growth rate, (4) baseline sister chromatid exchange frequency, and (5) doxorubicin sensitivity. Chromosomal number has been reported to be of
Volume 159 Number 5
prognostic significance 12 ; however, the exact relationship of polyploidization to prognosis has not been well established. It has been suggested that aneuploid tumors have a worse prognosis than near-diploid or tetraploid tumors. 12· 15 In the ovarian carcinoma cells studied, there were numerous complex chromosomal rearrangements, deletions, and cell-to-cell variability in chromosome number. Thus our results support those of others regarding the relationship between prognosis ·and chromosome constitution. 12· 15 We assume that the aneuploid cells analyzed were present in the initial tumor. However, because the analysis was done after progressive treatment with a variety of antineoplastic agents, it can only be inferred that the tumor was aneuploid initially. Although drug resistance might be associated with a stable mutagenic event or transient stimulation in DNA repair, experimentally, drug resistance has often been found to be related to gene amplification. 3 In fact, there are about 12 resistant genes that have been reported to be amplified, including those associated with doxorubicin resistance. 3 In the cells analyzed, variable numbers of double minutes were found, suggesting that gene amplification had occurred. However, these double minutes need not represent gene amplification, and in the absence of any other evidence, no conclusions can be made. If the double minutes are amplified genes, they probably are not drug-resistant genes, since the heterogeneous distribution of double minutes per cell did not appear to correlate with doxorubicin sensitivity. The average generation time of a cancer cell in vitro may reflect the doubling time of the tumor in vivo. Although the generation time of the ovarian carcinoma was slower than that of human fibroblasts, it was more rapid than that previously reported for other lines. 8 Because the analysis was done in a single tumor, no conclusions can be made concerning in vivo resistance and growth rate. Sister chromatid exchange represents DNA exchanges between homologous chromatids and has been proposed as an indicator of genetic stability and cellular sensitivity to cytotoxic agents.4·" The baseline sister chromatid exchange frequency in the ovarian carcinoma was significantly greater than that for human fibroblasts. This increase in baseline level of sister chromatid exchange frequency may have been caused by the prior exposure of cells to carcinogenic agents in vivo or may reflect some inherent genetic instability. In this study, the ovarian tumor cells were also more resistant to doxorubicin-induced sister chromatid ex-
Cytogenetic markers of human ovarian carcinoma
1103
changes. This resistance to doxorubicin was similar to that found with clonogenic survival assays in which the D37 (dose which reduces survival to 37%) for the ovarian tumor cells was 0.6 µmol/L doxorubicin, whereas the D37 for 1522 cells was 0.05 µmol/L (unpublished observations). In conclusion, we have analyzed a human ovarian tumor that is clinically resistant to therapy and found that it is also resistant in vitro to doxorubicin and has a number of hallmarks of tumor resistance. Furthermore, we have shown that cytogenetic analysis of tumor cells can yield a wealth of information and provide insights into the mechanism of resistance and the prediction of success of therapy. REFERENCES 1. Goldie JH, Coldman AJ. The genetic origin of drug resistance in neoplasms: implications for systemic therapy. Cancer Res I 984;44:3643-53. 2. Hernandez E, Rosenshein NB, Bhagavan B, Parmley T. Tumor heterogeneity and histopathology in epithelial ovarian carcinoma. Obstet Gynecol I 984;63:330-4. 3. Schimke RT. Gene amplification, drug resistance, and cancer. Cancer Res 1984;44: 1735-42. 4. Wolff S. Sister chromatid exchange. Ann Rev Gene 1977; 11: 183-201. 5. Wolff S. Chromosome aberrations, sister chromatid exchange, and the lesions that produce them. In: Wolff S, ed. Sister chromatid exchanges. New York: WileyInterscience Publications, 1982;4 l-5 7. 6. Abe S, Sasaki M. Chromosome aberrations and sister chromatid exchanges in Chinese hamster cells exposed to various chemicals. J Natl Cancer Inst 1972;58: 1635-41. 7. German J, Schonbberg S, Lovie E, Changenti RSK. Bloom's syndrome. IV. Sister chromatid exchanges in lymphocytes. Am J Hum Genet 1977;29:248-55. 8. Tofilon PH, Basic I, Milas L. Prediction of in vivo sister chromatid exchange assay. Cancer Res I 985;45:2025-30. 9. Sumner AJ, Evans HJ, Burkland RA. New technique for distinguishing between human chromosomes. Nature New Biol 1971;232:31-3. 10. Schwartz JL, Banda MJ, Wolff S. 12-0-tetradecanoylphorbol-13-acetate (TPA) induces sister chromatid exchanges and delays in cell progression in Chinese hamster ovary and human cell lines. Mutat Res I 982;92:393-409. 11. Tice R, Schneider EL, Perry JM. The utilization of bromodeoxyuridine-incorporation into DNA for the analysis of cellular kinetics. Exp Cell REs 1976; 102:232-6. 12. Atkin NB. Modal DNA value and chromosome number in ovarian neoplasia: a clinical and histopathological assessment. Cancer 1971;27:1064-73. 13. Panani A, Ferti-Passantonopoulou A. Common marker chromosomes in ovarian cancer. Cancer Genet r~ytogenet 1981 ;3:279-91. 14. Mitelman F, Levan G. Clustering of aberrations to specific chromosomes in human neoplasms IV. A survey of 1871 cases. Hereditas 1981;95:79-139. 15. Gebhart E, Bruderlein S, Tulusan A, Maillot K, Birkmann J. Incidence of double minutes, cytogenetic equivalents of gene amplification in human carcinoma cells. Intj Cancer I 984;34:369-73.
Key words: Doxorubicin, human ovarian tumor cells, sister chromatid exchanges
Resistance of cancer cells to cytotoxic therapy remains an' important problem in cancer biology. The reason for chemotherapy or radiotherapy failure is unknown. Clinical studies have shown that tumor stage, grade, and cell type are important prognostic parameters, but even within a specific class of tumors, there is much variability in response. 1 Although evidence has accumulated to suggest that the presence of chemoresistant cells in pretreatment biopsy specimens will predict therapeutic failure, the presence of chemosensitive cells in a pretreatment biopsy specimen will not predict successful treatment. 2 · " Therefore investigations into other possible parameters associated with tumor resistance are being examined, such as deoxyribonucleic acid (DNA) content and tumor growth rates. Cytogenetic analysis affords an investigator the means to examine multiple parameters relatively quickly using the same material. One can determine relative in vitro chemosensitivity, cell generation time,
the degree of aneuploidy, and whether large-scale gene amplification has occurred. Also, with regard to all these parameters, because individual cells are analyzed, tumor heterogeneity can be determined. In this study we have examined a human ovarian epithelial carcinoma cytogenetically and compared it with normal, nontransformed, human fibroblasts. Each cell line was karyotyped and the degree of aneuploidy was determined. Cells were then analyzed for the baseline and doxorubicin-induced sister chromatid exchange frequency. Sister chromatid exhanges represent reciprocal exchanges between homologous daughter chromatids. Although the mechanism and significance of sister chromatid exchange formation remain unknown, they have been found to be sensitive and reliable markers of genetic instability and cellular sensitivity to cytotoxic agents." 8 Finally, the average generation time was determined in these cells and used as a measure of tumor cell growth rate.
From the Division of Gynecologic Oncology, Department of Obstetrics and Gynecology" and the Department of Radiation Oncology,' Pritzker School of Medicine, The University of Chicago. Supported by American Cancer Society Career Development Awards U- R. and]. L. S.), the Cancer Research Foundation U· R.), and Grant CA 42596 U· L. S. and R.R. W.) from the National Cancer Institute. Received for publication July 30, 1987; revised January 8, 1988, and May 4, 1988; accepted May 31, 1988. Reprint requests: Jacob Ronnensch, MD, Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, The University of Chicago, 5841 S. Maryland Ave., Chicago, IL 60637.
Material and methods Patient characteristics. A 48-year-old female who had a right pleural effusion and pelvic mass extending to the umbilicus underwent a thoracentesis and exploratory laparotomy in August 1984 for a stage IV, grade 2, serous cystadenocarcinoma of the ovary. After cytoreductive surgery, she had five courses of hexamethylmelamine, cytoxan, doxorubicin (Adriamycin), and cisplatin. As a result of thrombocytopenia, hexamethylmelamine and doxorubicin were discontinued, 1099
1100
Rotmensch et al.
Am
J
November 1988 Obstet Gynecol
c Fig. I. Fluorescence plus Giemsa-stained preparations of cells that went through one (A), two (B), or three (C) DNA replications.
and seven additional courses of cytoxan and cisplatin were given. After disease progression, cisplatin and a continuous infusion of 5-fluorouracil were given. With the development of ascites, a 500 ml aliquot was removed by paracentesis, and the carcinoma cells were cultured in vitro. In vitro establishment of ovarian carcinoma and maintenance of carcinoma and normal fibroblast cell lines. Two liters of ascites were centrifuged at 500 g. The pellet was removed and suspended in 5 ml of 5% serum culture medium. The culture medium consisted of 72.5% Dulbecco's modification of Eagle's medium, 22.5% Ham's nutrient mixture F-12, 5% fetal bovine serum, 20 ng/ml of epidermal growth factor (added the third day) after plating, 5 µg/ml of transferrin, 2 x 10- 11 mol/L, 3,3',5 triiodo-L-thyron ine, 10- 10 mol/L cholera toxin, 1.8 x 10- 4 mol/L adenine, 0.4 µg I ml of hydrocortisone, 50 U I ml of my cos ta tin, which was omitted after the second passage, 100 U I ml of penicillin, and 100 µg/ml of streptomycin. Lethally irradiated embryonic 3T3 fibroblasts were added at a density 2 x 104 cells/cm 2 as a feeder layer when necessary and selectively removed by 15- to 30-second incubations in 0.02% ethylenediamine tetra-acetate. After primary culture, the malignant epithelial ovarian carcinoma cells were separated, characterized, and in-
dividually plated. The epithelial cells were subcultured, and after 2 to 3 weeks, individual colonies were obtained with40diameter of 0.5 to 1.0 cm. Human fibroblasts (AG 1522) were grown in tissue 4 5 culture flasks at stock concentrations of 10 to 10 cells in modified Eagle's medium with 20% fetal calf serum, 1% L-glutamine, 1% penicillin and streptomycin, and 0.5% supplement-B. Supplement-B contained 9 gm of o-glucose and 66 mg of sodium pyruvate/ dl of Earle's balanced salt solution. The fibroblasts (AG 1522) formed a monolayer on tissue culture surfaces and were harvested by scraping or using trypsinethylenediamine tetra-acetate. Cytogenetic analysis. Five parameters were examined for each cell line: (1) chromosome number per cell, (2) the proportion of cells with evidence of gene amplification, (3) the average generation time of each cell line, (4) the baseline sister chromatid exchange frequency, and (5) the doxorubicin-sen sitivity of each cell line. For analysis, exponentially growing cells were plated at a density of 5 x 105 cells in 15 ml of complete medium in a T-75 flask. Twenty-four hours later, cells were exposed from 0 to 2 x 10- 6 mol/L doxorubicin for 1 hour and cultured for 48 to 72 hours in complete medium supplemented with 10-s mol/L bromodeoxy-
Volume 159 Number 5
Cytogenetic markers of human ovarian carcinoma
1101
25
45
50
55
60
65
70
75
80
85
90
>95
NUMBER OF CHROMOSOMES/CELL
Fig. 2. Frequency of chromosome number in the ovarian carcinoma OVC-1.
uridine. During the last 2 hours, 2 x 10- 1 mol/L colcemid was added to each culture. Cells were removed from the flasks by trypsin treatment, incubated with 0.075 mol/L KC! for 15 minutes, and then fixed in 3: 1 methanol to acetic acid. Cells were placed onto slides, and the slides were then dried overnight. The degree of aneuploidy and proportion of cells with evidence of gene amplification were determined in G-banded cells. Cells were G-banded with 0.1 % trypsin and 2% Gurr-buffered Giemsa. 9 One hundred cells were anlayzed for chromosome number and the presence of either double minutes or homogeneous staining regions. Double minutes homogenous staining regions are both cytogenetic correlates of gene amplification. Double minutes are small, paired, spherical chromosomes, whereas homogenous staining regions are chromosome regions that fail to exhibit the usual differential banding pattern.' Average generation time and baseline and induced sister chromatid exchange frequency were determined in cells stained with a fluorescence plus Giemsa technique.10 Briefly, slides were stained for 15 minutes in Hoechst 33258 (0.5 µg/ml), mounted in 0.067 mol/L Sorensen's buffer (pH 6.8), placed on a slide warmer at 55° C, and exposed to light from two General Electric 15 W black lights for 4 to 10 minutes. They were rinsed in buffer and stained in 3% Giemsa in buffer. The average generati~n time was estimated from the proportion of cells that had gone through one (P 1), two
Table I. Cytogenetic characteristics of ovarian carcinoma and normal cell line Ovarian carcinoma
Mean chromosome No./ cell Modal chromosome No./cell Chromosome No. range % cells with double minutes Average generation time
Normal cell line (1522)
61.7 62 46-95 48 38.l
46 46 46 0 24.0
0.28 ± 0.01
0.15 ± 0.02
(hr) Baseline sister chromatid exchange frequency I cell Rate of doxorubicininduced sister chromatid exchange/ chromosom (per µmol/L experiment)
0.12
0.62
(P2 ), or three (P,) rounds of DNA replication. Average
generation time is defined as follows: AGT = time in culture/P 1 + 2 (P 2 ) + 3 (P,). The proportion of cells that had gone through one, two, or three rounds of DNA replication was determined on the fluorescence plus Giemsa-stained slides, as described by Tice et al. 11 First-division cells contain chromosomes, with both sister chromatids stained uniformly dark. Second-division cells have chromosomes with one chromatid darkly stained and its sister lightly stained. Third-division cells contain some chromosomes, with one chromatid darkly stained and its sister lightly stained, whereas the rest of the chromosomes are uniformly stained lightly (Fig. 1). One hundred cells were anlayzed for average generation time; 50 cells
1102 Rotmensch et al.
November 1988 Am J Obstet Gynecol
0.5
0.4 Q)
E
0
Ill
0
E
0.3
0 .....
.r:; ()
......
w
() (/)
o Ovarian e Normal 0
2
4
Carcinoma
6
8
10
Dose (x 10- 7 m) Fig. 4. Dose response of tumor (O) and normal(•) cells to sister chromatid exchange (SCE) induction by doxorubicin.
Fig. 3. Partial meta phases of an ovarian cell line show multiple (A) and single double-minutes (DM) (B).
were analyzed for baseline and doxorubicin-induced sister chromatid exchange frequency. Results
Chromosome number and constitution. The ovarian carcinoma cells had an aneuploid distribution with a variable chromosome number per cell. In 100 ovarian carcinoma cells analyzed, the chromosome number ranged from 46 to 98 chromosomes/cell with a modal number of 62 and a mean number of 61.7 chromosomes/ cell (Fig. 2). Table I compares the cytogenetic characteristics of the ovarian carcinoma to benign cells. The karyotype was complex, with numerous rearrangement and duplications. In the metaphase preparations examined, there were variable numbers of double minutes. The number of double minutes in each cell preparation ranged from one pair to numerous pairs (Fig. 3), with greater than 10 pairs in 48% of the 100 cells examined. In subsequent in vitro passages, the frequency of double minutes declined. No homogenous staining regions were found. Average generation time, sister chromatid exchange frequency, and sensitivity to doxorubicin. The ovarian tumor line grew well and had a plating efficiency of 10%. Cell doubling times, estimated by av-
erage generation time measurements, were 38 hours for the ovarian cells and 24 hours for the normal fibroblasts. The baseline frequency of sister chromatid exchanges was significantly higher in the ovarian cell line (0.28 ± 0.01 sister chromatid exchange/chromosome) compared with the fibroblast cells (0.15 ± 0.02 sister chromatid exchange/ chromsome). The ovarian cells were also more resistant to the induction of sister chromatid exchanges by doxorubicin (Fig. 4). Sister chromatid exchanges were' induced at a rate of 0.12 sister chromatid exchange/µmol/L exposure of doxorubicin in ovarian carcinoma cells and at a rate of 0.62 sister chromatid exchange I µmoll L exposure of doxorubicin in 1522 cells. There was no evidence of heterogeneity in either baseline or induced sister chromatid exchange responses. Comment
The mechanisms for resistance of tumor cells to cancer therapy remain unknown. Many believe that resistance is caused by inherent genetic changes. By understanding those genetic events and the conditions that predispose them to those events, more rationale approaches to the treatment of cancer should develop. In an attempt to correlate in vitro cellular findings with tumor behavior and clinical outcome, we examined an ovarian tumor clinically resistant to cancer therapy for the following: ( 1) chromosome constitution, (2) evidence of gene amplification, (3) average growth rate, (4) baseline sister chromatid exchange frequency, and (5) doxorubicin sensitivity. Chromosomal number has been reported to be of
Volume 159 Number 5
prognostic significance 12 ; however, the exact relationship of polyploidization to prognosis has not been well established. It has been suggested that aneuploid tumors have a worse prognosis than near-diploid or tetraploid tumors. 12· 15 In the ovarian carcinoma cells studied, there were numerous complex chromosomal rearrangements, deletions, and cell-to-cell variability in chromosome number. Thus our results support those of others regarding the relationship between prognosis ·and chromosome constitution. 12· 15 We assume that the aneuploid cells analyzed were present in the initial tumor. However, because the analysis was done after progressive treatment with a variety of antineoplastic agents, it can only be inferred that the tumor was aneuploid initially. Although drug resistance might be associated with a stable mutagenic event or transient stimulation in DNA repair, experimentally, drug resistance has often been found to be related to gene amplification. 3 In fact, there are about 12 resistant genes that have been reported to be amplified, including those associated with doxorubicin resistance. 3 In the cells analyzed, variable numbers of double minutes were found, suggesting that gene amplification had occurred. However, these double minutes need not represent gene amplification, and in the absence of any other evidence, no conclusions can be made. If the double minutes are amplified genes, they probably are not drug-resistant genes, since the heterogeneous distribution of double minutes per cell did not appear to correlate with doxorubicin sensitivity. The average generation time of a cancer cell in vitro may reflect the doubling time of the tumor in vivo. Although the generation time of the ovarian carcinoma was slower than that of human fibroblasts, it was more rapid than that previously reported for other lines. 8 Because the analysis was done in a single tumor, no conclusions can be made concerning in vivo resistance and growth rate. Sister chromatid exchange represents DNA exchanges between homologous chromatids and has been proposed as an indicator of genetic stability and cellular sensitivity to cytotoxic agents.4·" The baseline sister chromatid exchange frequency in the ovarian carcinoma was significantly greater than that for human fibroblasts. This increase in baseline level of sister chromatid exchange frequency may have been caused by the prior exposure of cells to carcinogenic agents in vivo or may reflect some inherent genetic instability. In this study, the ovarian tumor cells were also more resistant to doxorubicin-induced sister chromatid ex-
Cytogenetic markers of human ovarian carcinoma
1103
changes. This resistance to doxorubicin was similar to that found with clonogenic survival assays in which the D37 (dose which reduces survival to 37%) for the ovarian tumor cells was 0.6 µmol/L doxorubicin, whereas the D37 for 1522 cells was 0.05 µmol/L (unpublished observations). In conclusion, we have analyzed a human ovarian tumor that is clinically resistant to therapy and found that it is also resistant in vitro to doxorubicin and has a number of hallmarks of tumor resistance. Furthermore, we have shown that cytogenetic analysis of tumor cells can yield a wealth of information and provide insights into the mechanism of resistance and the prediction of success of therapy. REFERENCES 1. Goldie JH, Coldman AJ. The genetic origin of drug resistance in neoplasms: implications for systemic therapy. Cancer Res I 984;44:3643-53. 2. Hernandez E, Rosenshein NB, Bhagavan B, Parmley T. Tumor heterogeneity and histopathology in epithelial ovarian carcinoma. Obstet Gynecol I 984;63:330-4. 3. Schimke RT. Gene amplification, drug resistance, and cancer. Cancer Res 1984;44: 1735-42. 4. Wolff S. Sister chromatid exchange. Ann Rev Gene 1977; 11: 183-201. 5. Wolff S. Chromosome aberrations, sister chromatid exchange, and the lesions that produce them. In: Wolff S, ed. Sister chromatid exchanges. New York: WileyInterscience Publications, 1982;4 l-5 7. 6. Abe S, Sasaki M. Chromosome aberrations and sister chromatid exchanges in Chinese hamster cells exposed to various chemicals. J Natl Cancer Inst 1972;58: 1635-41. 7. German J, Schonbberg S, Lovie E, Changenti RSK. Bloom's syndrome. IV. Sister chromatid exchanges in lymphocytes. Am J Hum Genet 1977;29:248-55. 8. Tofilon PH, Basic I, Milas L. Prediction of in vivo sister chromatid exchange assay. Cancer Res I 985;45:2025-30. 9. Sumner AJ, Evans HJ, Burkland RA. New technique for distinguishing between human chromosomes. Nature New Biol 1971;232:31-3. 10. Schwartz JL, Banda MJ, Wolff S. 12-0-tetradecanoylphorbol-13-acetate (TPA) induces sister chromatid exchanges and delays in cell progression in Chinese hamster ovary and human cell lines. Mutat Res I 982;92:393-409. 11. Tice R, Schneider EL, Perry JM. The utilization of bromodeoxyuridine-incorporation into DNA for the analysis of cellular kinetics. Exp Cell REs 1976; 102:232-6. 12. Atkin NB. Modal DNA value and chromosome number in ovarian neoplasia: a clinical and histopathological assessment. Cancer 1971;27:1064-73. 13. Panani A, Ferti-Passantonopoulou A. Common marker chromosomes in ovarian cancer. Cancer Genet r~ytogenet 1981 ;3:279-91. 14. Mitelman F, Levan G. Clustering of aberrations to specific chromosomes in human neoplasms IV. A survey of 1871 cases. Hereditas 1981;95:79-139. 15. Gebhart E, Bruderlein S, Tulusan A, Maillot K, Birkmann J. Incidence of double minutes, cytogenetic equivalents of gene amplification in human carcinoma cells. Intj Cancer I 984;34:369-73.