Ovarian cancer cisplatin-resistant cell lines: Multiple changes including collateral sensitivity to Taxol

Annals of Oncology 9: 423-430. 1998. © 1998 Kluwer Academic Publishers. Primed in the Netherlands.

Original article Ovarian cancer cisplatin-resistant cell lines: Multiple changes including collateral sensitivity to Taxol P. Perego,1 S. Romanelli,1 N. Carenini,1 I. Magnani,2 R. Leone,3 A. Bonetti,3 A. Paolicchi4 & F. Zunino1 'Oncologia Spenmentale B, Istituto Nuzionale per to Studio eh Cura dei Tumori, Milan: 2Universita degli Studi di Milano, Dipartimemo di Biologia e Genetica, School of Medicine, Milan; iIstttutodi Farmacologia, Universita'di Verona; AUniversita degh Studi di Pisa, Pisa, Italy

Summary

level, thus suggesting a complex regulation of cellular glutathione content. In the resistant cells with the highest glutaBackground: Alteration in apoptosis pathways (in particular thione content, a reduced level of cisplatin-induced cross-link mutations of p53 gene) may result in resistance of ovarian was found. Analysis of DNA platination revealed a slight carcinoma to cisplatin. However, cisplatin resistance is likely decrease of DNA-bound platinum only in IGROV-1/Ptl cells. to be multifactorial. An understanding of the molecular alter- Again, this reduction is consistent with a protective role for ations associated with the development of resistance may be of glutathione. The expression of metallothionein I la was inconsiderable relevance in an attempt to optimize the therapeu- creased in both resistant variants. tic approach. Conclusions: Multiple changes are involved in acquired Study design: Two cisplatin-resistant sublines (IGROV-1/ resistance of ovarian carcinoma cells including reduced susPtO.5 and IGROV-1/Ptl), both characterized by mutant p53 ceptibility to apoptosis as consequence of inactivation of p53 (Cancer Res 1996; 56: 556-62), but with different degree of and expression of defence mechanisms. The relative contriburesistance were studied in terms of pattern of cross-resistance, tion is related to the degree of drug resistance. In particular, susceptibility to drug-induced apoptosis, expression of glutha- the glutathione-dependent system could have a role only in the thione-dependent system, cellular pharmacokinetics, drug- development of a high degree of resistance. Finally, the finding induced DNA damage. The resistance index (ratio between the that Taxol was very effective in inducing apoptosis in resistant IC50 of resistant and sensitive cells) after a 96-hour drug expo- sublines wilhp53 mutation supports the expression of an intact sure was 10 for IGROV-1/Pt0.5 and 14 for IGROV-1/Ptl cells. p53-independent pathway of apoptosis and suggests the pharResults: Resistant cells were cross-resistant to DNA-dam- macological interest of Taxol in the treatment of/?55-mutated aging agents and, interestingly, they had a collateral sensitivity tumors. to Taxol. The cellular response to Taxol paralleled the drug ability to induce apoptosis. The intracellular glutathione level was significantly increased in IGROV-1/Pt cells compared to Key words: cisplatin, resistance, ovarian carcinoma the sensitive counterpart. In contrast, glutathione S-transferase level was consistently reduced in both sublines. y-Glutamyl Abbreviations: GSH - glutathione; BSO - D,L-buthioninetranspeptidase activity, which was lower in resistant than in S,R-sulfoximine; SDS - sodium dodecylsulphate; ICL - intersensitive cells, was not directly correlated with glutathione strand cross-links

Introduction Platinum-based compounds are used in the therapy of several human tumors, particularly ovarian cancer [1,2]. Despite the efficacy of platinum-based therapies, drug resistance often occurs [3, 4]. The cellular and molecular basis of cisplatin resistance remains controversial. Therefore, selection of resistant cell lines could provide new cellular models useful for further studies. The mechanisms determining the cisplatin-resistant phenotype are still uncertain. They include decreased drug accumulation [5, 6], an increased level of intracellular thiols [7, 8], and an increased DNA repair [9, 10]. Among the cellular alterations resulting in reduction of the amount of drug bound to DNA, a major role

has been attributed to GSH, which could reduce cisplatin cytotoxicity through the inactivation of platinum complexes, thereby interfering with the formation of cisplatin-DNA adducts and reducing the frequency of cross-linking [11]. The aim of the present study was to characterize specific parameters of two in vitro selected resistant variants (IGROV-1/Pt0.5 and IGROV-1/Ptl) generated after prolonged exposure of the ovarian carcinoma IGROV-1 cell line to cisplatin. Specifically, the study was directed to examine a) the expression of defense mechanisms including GSH, glutathione-dependent enzymes, and metallothioneins, and b) the pharmacological characteristics of IGROV-1 and IGROV-1/Pt cells, with particular reference to the cross-resistance pattern. A mo-

424 lecular characterization of the resistant sublines indicated a loss of p53 function following mutation and a consequent reduced susceptibility to cisplatin-induced apoptosis [12]. A detailed cellular and pharmacological characterization of resistant sublines indicated that the emergence of cisplatin resistance is accompanied by multiple changes which could contribute to the degree of resistance. Materials and methods Cell lines and growth conditions The IGROV-1 cell line, established from a patient with an ovarian adenocarcinoma [13], was used in the study. It was obtained from an in v/v0-growing IGROV-1 tumor kindly given by Dr. Benard (Institut Gustave Roussy, Villejuif. France). IGROV-1 cells grew as monolayers. The two cisplatin-resistant variants, IGROV-1/PtO.5 and IGROV-1/ Ptl, were generated by prolonged continuous exposure to increasing concentrations of cisplatin. The cell lines were maintained in RPMI1640 medium supplemented with 10% fetal calf serum and 1% glutamine. The resistant sublines were cultured in medium with cisplatin, 0.5 (IGROV-1 /PtO.5) and 1 ug/ml (IGROV-1/Ptl); when grown in the absence of cisplatin, no modification of the response to the drug was observed for six months. IGROV-1 exhibited a wild-type p53, while both resistant variants carried a mutated p53 [12]. IGROV-1/PtO.5 and IGROV-1/Ptl cells had levels of bax expression below those of IGROV-1 cells [14]. For the growth curves, cells from maintenance cultures were counted and inoculated into six-well plates (9.6 cm 2 , Costar, Pleasanton, CA, USA). Duplicate samples were harvested daily and counted with a Coulter counter (PBI Electronics, Luton, UK), and the cell number was averaged for each interval. The doubling times were calculated in the logarithmic phase of growth.

Cytogenetic analysis Chromosomal spreads set up from IGROV-1 and IGROV-1/Pt cells were banded using Trypsin-Giemsa (Sigma, St. Louis, MO, USA) according to standard procedures [15]. The karyotypes were determined according to the International System for Human Cytogenetic Nomenclature (1991) scheme [16]. At least 20 karyotypes per cell line were examined in three independent experiments.

Cytofluorimetric analysis Exponentially growing cells were harvested and fixed in 70% ethanol. They were incubated with phosphate-buffered saline containing RNase (40 ug/ml) for 15 min on ice and subsequentely stained with propidium iodide (25 ug/ml) in phosphate-buffered saline for 15 min. The stained cells were analyzed by flow cytometry using an Epics C Instrument (Coulter Electronics, Hialeah. FL. USA). For each experiment, 104 cells were analyzed.

Drugs The drugs used in the study were obtained as follows: cisplatin (Platinex), etoposide (Vepesid) and Taxol from Bristol-Myers Squibb (Rome. Italy): D-actinomycin (Cosmegen) from Merck Sharp and Dohme (Rome. Italy): camptothecin from Sigma Chemicals (St. Louis, MO. USA): carmustine (Nitrumon) from Sintesa S.A. (Bruxelles, Belgium): melphalan (Alkeran) from the Wellcome Foundation (Wellcome Italia S.p.A.. Pomezia. Italy); and vinblastine (Velbe) from Eli Lilly (Indianapolis, IN. USA). Cadmium chloride and BSO were

purchased from Sigma. Doxorubicin was kindly supplied by Pharmacia Upjohn (Milan, Italy); and reduced GSH was from Boehringer Mannheim (Milan, Italy). Immediately before use, all of the drugs were prepared as suggested by the suppliers and then diluted in saline, with the exception of camptothecin, which was dissolved in dimethyl sulfoxide at 10 mM and diluted in water.

GSH level and glutathione-dependent enzymatic activities Exponentially growing cells were harvested, weighed and immediately homogenized in ice-cold 5% trichloroacetic acid. After centrifugation at 5000 x g for 10 min to remove protein, the GSH content was determined according to the method of Ellman [17]. All the enzymatic activities were evaluated as previously described [18]. Briefly, cell pellets were homogenized on ice in 0.05 M Tris-HCl buffer, pH 7.4, by 20 strokes of a tight-fitting Dounce homogenizer. Aliquots of crude homogenates were used for measuring y-glutamyl transpeptidase activity with y-glutamyl-p-nitroanihde and glycilglycine as substrates. The remaining cell homogenate was centrifuged at 105,000 x g for 60 min, and the supernatants were used for the determination of GST. All of the results are expressed as nmol/min/mg protein. Protein content was evaluated by using the BCA system (Pierce, Rockford, IL, USA).

Cytotoxic assay and apoptosis studies Cell survival after a 96-hour exposure to cisplatin and a range of antitumor drugs was assessed by the tetrazolium dye assay [19]. Preliminary experiments were performed to determine the appropriate seeding number of cells (4000 cells/well), after confirming the linear relationship between the absorbance and number of cells in the growth curve of each line. Alternatively, the antiproliferative effect of drugs was also assessed by the growth inhibition assay after a one-hour or 72-hour exposure, as previously described [12]. Briefly, cells in the logarithmic phase of growth were harvested and seeded in duplicate into six-well plates. Twenty-four hours after seeding, the drug was added to the medium and cells were incubated for the appropriate time. Cells were harvested 72 hours after the beginning of exposure and counted with a Coulter counter. In the case of BSO treatment cells were preincubated 24 hours, washed with saline, exposed to cisplatin for one hour and, after washing, incubated for 72 hours and then harvested. IC 50 is defined as the inhibitory drug concentration causing a 50% reduction of absorbance at 550 nm/decrease of cell growth over that of untreated control; the IC50 values were calculated from linear regression analysis of individual experiments. Apoptosis was assessed by fluorescent microscopy as previously described, with minor modifications [12]. After a 24-hour exposure to Taxol, cells were washed and incubated in drug-free medium for 48 hours. Floating and adherent cells were collected and fixed in 70% icecold ethanol. Cells were then stained in propidium iodide (20 ug/ml) in the presence of RNase (66 U/ml) for 30 min in the dark, and then analyzed for nuclear morphology changes. The percentage of apoptotic cells was determined from the total cell number (floating + adherent cells).

ONA platination For DNA platination studies 10 x 106 cells were exposed to cisplatin for one hour, 48 hours after seeding in 150 cm 2 flasks. Cells were then harvested and processed for DNA extraction. Briefly, cells were washed twice with saline and suspended in 500 ul of 10 mM Tris HC1, pH 8.100 mM NaCl, and 1 mM EDTA. Twenty-five ul of 20% SDS and 4 ul of 25 mg/ml proteinase K were added to each sample, and samples were incubated overnight at 37 C. DNA was phenol-extracted twice and alcohol-precipitated. After washing in 70% ice-cold ethanol, DNA was dried and dissolved in 10 mM Tris HC1, pH 8, and 1 mM EDTA. and its concentration was spectrophotometrically determined. Platinum content was measured by inductively coupled plasma mass spectroscopy [20].

425 Table I. Biological and biochemical features of IGROV-1. IGROV-1/PtO. 5 and IGROV-1/Ptl cells. Cell line

IGROV-1 IGROV-1/Pt0.5 IGROV-1/Ptl

Doubling time (hours)3

28 ± 1 36 1 2.2 33 ± 2.0

Cell cycle phase distribution1" G 0 G,

G2M

S

50.5 ± 2.5 53.5 10.5 49.012.0

32.01 1.0 29.5 + 1.5 28.01 1.6

17. 5 1 1.5 17.0 1 2 . 1 23. 0 1 1.2

GSHC

yGT d

GST d

2.88 ± 0.22 6.64 1 0.02 7.92 + 0.17

9.39 ±0.08 1.13 ±0.01 2.77 ±0.03

347.22 1 2.5 160.981 1.0 150.83 1 1.2

a

Values (mean 1 SD. n - 3) were calculated in the logarithmic phase of growth. Calculated in exponential phase of growth; the reported values (mean 1 SD. n = 3) are expressed as percentage of total cells. c Determined spectrophotometrically on lysates prepared from exponentially growing cells: values (mean 1 SD, /; = 3) are expressed as mmol/g; GSH. reduced glutathione. d Values (mean 1 SD, n — 3) are expressed as nmol/min/mg; y-GT, y-glutamyltranspeptidase; GST, glutathione S-transferase. b

Alkaline elution Cisplatin-induced ICL in IGROV-1 and IGROV-1 /Pt were determined by the alkaline elution method [21]. Cellular DNA was labeled with 0.08 mCi/ml [2-14C]thymidine (Amersham, Little Chalfont, UK) for 24 hours, and the labeled nucleoside precursor was removed 24 hours before exposure to drug. Cells were exposed to cisplatin for one hour, incubated in fresh medium for five hours, and then processed. To produce a known frequency of y-ray-induced DNA single-strand breaks before analysis, approximately 5 x 105 cells were irradiated with a cesium-137 source (1000 rad). Comparison of the frequency of ICL in sensitive and resistant cells was possible because the same dose of y-irradiation caused a similar frequency of DNA single-strand breaks in the three cell lines. Cells were deposited on a 2.0-mm-pore polycarbonate filter (25 mm diameter, Nucleopore, Pleasanton, CA, USA) and lysed with 2% SDS, 0.02 M disodium EDTA (pH 10) and 0.5 mg/ml protcinase K. The DNA was then eluted at a flow rate of 0.035 ml/min with a 0.02 M EDTA-0.1% SDS solution adjusted to pH 12.15 with tetrapropylammonium hydroxide. During a 15-hour elution, the lysate was collected in fractions and counted by liquid scintillation. DNA ICL frequency was calculated by the following formula [22]:

where r0 is the fraction of [[4C]DNA remaining on the filter in irradiated control cells and r is the fraction of [14C]DNA remaining on the filter in drug-exposed, irradiated cells, calculated after 12 hours of elution.

Northern blot hybridization analysis About 20 tig of total RNA prepared by the Li-Cl guanidine monothiocianate method [23] was electrophoresed on a formaldehyde-containing 1% agarose gel and transferred onto a nylon membrane. DNA probes were 32P-labeled with a random primer kit (Amersham, Little Chalfont, UK). Hybridization was carried out as previously described [24]. The expression level was evaluated by densitometric analysis of autoradiograms and compared with the expression level of the control (5-actin gene The probes used were: topoisomerases I and II cDNA fragments kindly provided by Dr. L. Liu (Baltimore, MD, USA); GST. metallothioneins Ha and DNA polymerase [5 obtained, respectively, from Dr. J. A. Moscow (Bethesda, MD, USA). Dr. J. S. Lazo (University of Pittsburg. USA) and Dr. S. H. Wilson (Bethesda, MD, USA).

Results Biological and biochemical characterization Relevant biological and biochemical features of a human ovarian carcinoma cell line (IGROV-1) and two

sublines selected for resistance to cisplatin, designated IGROV-1/Pt0.5 and IGROV-1/Ptl, are shown in Table 1. The doubling times were slightly augmented in the resistant cells, and cytofluorimetric analysis revealed similar cellular distributions through the cell cycle for IGROV-1, IGROV-1 /Pt0.5 and IGROV-1/Ptl cells. The intracellular level of GSH was increased in the resistant cell lines over that in sensitive cells (P < 0.05, ANOVA): a 2.9-fold increase was observed for the less resistant subline IGROV-1 /PtO.5, whereas a five-fold increase was found for the most resistant. Analysis of GSH-dependent enzyme activities revealed that IGROV-1/Pt cells had GSH S-transferase activities lower than those of parental cells. Low activity was also found in resistant cells for yGT, indicating a lack of correlation between intracellular GSH level and the level of a relevant enzyme involved in GSH homeostasis. To determine whether additional karyotypic alterations were associated with resistant phenotype the cytogenetic features were compared (Figure 1). The karyotype of the IGROV-1 cell line was similar to that described by Benard et al. (1985), except that 98% of cells were diploid. G banding analysis showed no homogeneous staining regions or double minutes in the resistant sublines in which a cellular population similar to IGROV-1 cells was observed. In IGROV-1/Pt0.5, another population (B) had additional rearrangements of chromosomes 3, 5 and 21. IGROV-1/Ptl cells were not substantially different from the population found in IGROV-1 cells, except for a rearrangement of chromosome 4. The modal karyotype for the cell lines was as follows: IGROV-1: 46, XX, der(2), t(2;5) (q35;q22), inv(3) (p21;p25), add(4) (pl6), add(9) (q34), add(10) (q26), del(20) (ql3.1), del(22) (ql3.1). IGROV-1/Pt0.5, population A: 47, X, add(X) (q28), der(2), t(2;5) (q35;q22), inv(3) (p21;p25), add(4) add(9) (q34), add(10) (q26), del(20) (ql3.1), del(22) , + 1 mar. IGROV-1/Pt0.5, population B: 46, XX, der(2), t(2;5) (q35;q22), inv(3) (P21;p25), del(3) (q21), add(4) add(5) (q22), add(9) (q34), add(10) (q26), del(20) der(21;21)(qlO;qlO), del(22) (ql3.1). IGROV-1/Ptl: 47, X, - X , der(2), t(2.5) (q35;q22), inv(3) ( P 21;p25), del(3) (q21), add(4) (pl6), add(5) (q22), add(9) (q34), add(10) (q26), del(20) (ql3.1), der(21;21) (qlO;qlO), del(22) + 2 mar.

426 2

3

4

5

9

10

20

21

22

X

4

4

Figure 1. Karyotypic alterations of IGROV-1, IGROV-1 /PtO.5 and IGROV-1 /Ptl cells.

Pattern of response to a range of cytotoxic agents Table 2 summarizes the pattern of cellular response to cisplatin and to a number of antitumor drugs with different mechanisms of action. The resistance index for cisplatin, denned as the ratio between the IC50 of the resistant and sensitive cell lines, was 10 and 14 for IGROV-1 /PtO.5 and IGROV-1/Ptl cells, respectively, and the IC50 of IGROV-1 cells was significantly different from that of the two resistant sublines (P < 0.05, ANOVA). Resistant sublines were cross-resistant to melphalan, a typical akylating agent, and only slightly less sensitive to all of the agents tested except for Taxol,

for which a collateral sensitivity was observed in IGROV-1/Ptl cells CP<0.05, ANOVA). A more detailed evaluation of response to Taxol was performed by using the growth inhibition assay. IGROV-1/PtO. 5 and IGROV-1/Ptl presented this special sensitivity. Statistical analysis (ANOVA) indicated that the IC50 of IGROV-1 cells was significantly different (P < 0.05) from that of the cisplatin-resistant sublines. An analysis of apoptosis induction was undertaken in IGROV-1 and IGROV-1 /Ptl cells. A 24-hour exposure to Taxol induced dose-dependent apoptosis in IGROV-1 cells. IGROV-1/ Ptl hypersensitivity to Taxol paralleled an increased cell susceptibility to Taxol-induced apoptosis. Apoptosis

Table 2. Pattern of cross-resistance of IGROV-1 /PtO.5 and IGROV-1 /Ptl cells to different cytotoxic agents." Drug

Cisplatin Doxorubicin Etoposide D-actinomycin Camptothecin Carmustine Melphalan Taxol" Taxolh Vinblastine Cadmium chloride a

IC5(, (ug/ml) IGROV-1

IGROV-1/Pt 0.5

RIC

IGROV-1/Ptl

RIC

0.22 ± 0.037 0.036 + 0.021 0.42 ±0.071 0.0098 ±0.0021 0.0045 ± 0.0002 11.54 ±2.041 0.85 ± 0.25 0.074 ± 0.037 0.11 ±0.01 0.0038 ±0.0018 3.61 ± 0.54

2.31 ±0.27 0.075 ± 0.024 1.38 + 0.24 0.024 ± 0.0065 0.0097 ± 0.0029 11.88 ±0.24 7.99 ±0.72 n.d. 0.029 ±0.01 0.014 ±0.0017 5.76 ±0.56

(10.5) (2.1) (3.3) (2.4) (2.1) (1.02) (9.4) n.d. (0.26) (3.7) (1.6)

3.054 ± 0.29 0.098 ±0.012 0.84 ± 0.047 0.015 ±0.0098 0.015 ±0.007 12.75 ±0.85 15.35 ±3.56 0.009 ± 0.006 0.027 ±0.02 0.0087 ± 0.0032 6.22 ± 0.62

(13.9) (2.7) (2.0) (1.5) (3.3) (11) (18.0) (0.12) (0.24) (2.3) (1.7)

Cytotoxicity was assessed by the tetrazolium dye assay on cells exposed to the drug for 96 hours. The IC 5 0 values are the means ± SD of four to six experiments: n.d. = not determined. b Cytotoxicity was assessed by the growth inhibition assay after 72 hours of drug exposure. Values are the means ± SD of three experiments. c RI = resistance index; ratio between ICSo of resistant and sensitive cells.

427 350 i

100-1

300-

80IGROV-1

250-

IGROV-1/Pt1

CO

co O

60 200-

t

<2

a:

40-

^

150

& < 100-

50-

0.05

100 TAXOL (ug/ml)

F/gHrr 2 Level of apoptosis in IGROV-1 and IGROV-1 /Ptl cells after exposure toTaxol. IGROV-1 cells were exposed to 0.05 (IC50) and 0.2 Hg/ml (IC50) Taxol and IGROV-1/Ptl cells to 0.05 |ig/ml Taxol (IC80) for 24 hours, and apoptotic cells were counted among propidium iodide-stained cells. Values represent the percentage of apoptotic cells (±SD, n = 3) in the whole population.

200

300

400

CISPI ATIN (Mg/ml)

Figure 3. DNA-bound platinum in cisplatin-sensitive and resistant cells after a one-hour exposure to cisplatin as measured by inductively coupled plasma mass spectroscopy spectrometry. IGROV-1 cells, squares; IGROV-1/Pt0.5 cells, circles; IGROV-1/Ptl cells, triangles. Table 3. Frequency of interstrand cross-links in IGROV-1 and IGROV-1/Ptl cells.

levels were similar when cells were exposed to equitoxic concentrations (Figure 2). In particular, equimolar Taxol concentrations (0.05 ug/ml) induced levels of apoptosis in IGROV-1 (13% ± 1%) lower than in IGROV-1/Ptl (26% ± 3%) cells. DNA platination and interstrand cross-links

Cell line

IGROV-1 IGROV-1/Ptl

ICL frequency3 30 (ug/ml)

300 (ug/ml)

0.0422-0.0343

0.205^0.176 0.073-0.074

a

Interstrand cross-link (ICL) frequency was measured by the alkaline elution technique. Ranges for two to three independent experiments are reported.

Since DNA is the major cellular target of cisplatin, analysis of the level of DNA-bound platinum was undertaken by using inductively-coupled plasma mass spectrometry [20]. The technique is characterized by a par- Ptl), a higher ICL frequency was found for the resistant ticularly high sensitivity, which allows measurment of cells, suggesting that an increased tolerance to DNA DNA-bound platinum when cells are exposed to drug damage can contribute to the resistant phenotype of the concentrations as low as 3 ug/ml cisplatin for one hour. variant. DNA platination after a one-hour exposure to cisplatin of IGROV-1 cells and cisplatin-resistant sublines ap- Gene expression peared to be linear up to 300 ug/ml cisplatin in all three lines (Figure 3). An appreciable decrease of DNA-bound A Northern blot analysis was performed to determine platinum was found in IGROV-1/Ptl cells at all of the whether specific genes were expressed in the resistant tested concentrations. Analysis of the ICL frequency sublines (IGROV-1/Pt0.5 and IGROV-1/Ptl) as comwas undertaken in IGROV-1 cells and in IGROV-1/Ptl pared to the parental cell line (Figure 4), and a slight cells by alkaline elution analysis (Table 3). Exposure to a increase in metallothionein Ha expression was found highly cytotoxic cisplatin concentration (300 ug/ml) (3.8-fold in IGROV-1/PtO.5 and 2.8-fold in IGROV-1/ resulted in a lower ICL frequency in IGROV-1/Ptl cells, Ptl cells). This observation suggests that cellular ability as expected on the basis of the inductively coupled to detoxify heavy metals may be a resistance factor in plasma mass spectroscopy analysis. When the two cell this cellular model. The mRNA levels of enzymes which lines were exposed to equitoxic cisplatin concentrations could be implicated in the mechanism of resistance to (ICM), 30 ug.'ml for IGROV-1. 300 ug'ml for IGROV-1 ' DNA-damaging agents, particularly to cisplatin and

428 Table 4. Effect of the combination of the glutathione-depleting agent buthionine-sulfoximine and cisplatin on IGROV-1 and IGROV-1/Ptl cells.

1 2 3 GSTTT

IGROV-1

(3-act in

GSH (umol/g) a

(Hg/ml)

GSH (nmol/g) a

IC 50 (ug/ml)

1.32 ±0.008

7.21 ±0.08

5.63 ± 0.2

110 ±0.7

0

6.0 ±0.05

0.59 ± 0.02

70 + 0.8

Untreated cells BSO-treated cells

IGROV-1/Ptl IC50

a

(3-act in

TOPO II a (3-actin ' — TOPO I (3-actin Figure 4. Northern blot analysis of selected genes in IGROV-1 (1), IGROV-1/Pt0.5 (2) and IGROV-1/Ptl (3) cells. Twenty ng of total RNA was fractionated in a 1% agarose-formaldehyde gel, transferred to a nylon filter and hybridized with the indicated probes: glutathione S-transferase 4 (GST 4), metallothionein Ma (MT IIa), DNA polymerase p (p-pol), DNA topoisomerases (topo), and P actin.

alkylating agents (DNA topoisomerases, GSH S-transferase, DNA polymerase p) [24, 25, 10], were similar.

Glutathione (GSH) depletion after exposure to buthionine-sulfoximine (BSO) was determined spectrophotometrically on lysates prepared from exponentially growing cells. Values are the means (±SD) of three independent experiments. b Cytotoxicity was assessed by the growth inhibition test after onehour-exposure to cisplatin; cells were preincubated for 24 hours with drug-free or BSO-containing medium. The concentration of BSO (22.4 ug/ml) was completely nontoxic. Values are the means (±SD) of three independent experiments.

of modulation suggested a marginal contribution of GSH to the resistant phenotype of IGROV-1 /Pt cells, since only a slight modification of IC50 was observed. Since, in contrast to cisplatin, Taxol is known to be a phase-specific cytotoxic agent, a different exposure to the drug was used in the Taxol cytotoxicity experiments. Apparently, GSH depletion caused a marginal reduction of Taxol cytotoxicity in IGROV-1 cell lines exposed for 72 hours to Taxol. In fact, Taxol IC50 was 0.095 ± 0.014 ug/ml (mean ± SD, n - 3) in IGROV-1 cells and 0.0425 ± 0.006 ug/ml (mean ± SD, n = 3) in IGROV-1/ Ptl cells, and, after GSH depletion by BSO (22.4 ug/ml), IC50 was 0.142 ± 0.038 ug/ml and 0.0525 ± 0.0007 ug/ml (mean ± SD, n - 3), respectively.

Discussion

The cell systems selected after prolonged exposure were Effect of GSH depletion on cisplatin or taxol cytotoxicity previously characterized with respect to the expression of proteins involved in the metabolic pathway leading to Since modulation of cellular determinants of resistance apoptosis, and an association between the mutation of represents a specific approach by which to overcome p53, and reduction of bax expression and cisplatin redrug resistance, we evaluated the possibility of modify- sistance has been reported [12]. Since several factors ing response to cisplatin by using the GSH-depleting probably contribute to the cisplatin-resistant phenotype agent BSO. For this purpose, the most resistant subline of cell lines with different levels of resistance, the study was used. Table 4 shows the response of IGROV-1 and was undertaken to examine the role of the factors proIGROV-1/Pt cells in terms of IC50 values obtained from posed as mechanisms of defense to this drug. dose-response curves after depletion of GSH with 22.4 Biological characterization of the three cell lines ug/ml BSO. A 24-hour exposure to the modulator had showed a slight increase in the doubling times of the no cytotoxic effect and caused a complete depletion of resistant variants. The observation is in agreement with cellular glutathione in IGROV-1 cells and a 90% deple- that described in other in vitro cisplatin-resistant cell tion in IGROV-1/Ptl cells. When the same exposure was lines [27, 28]. It is obvious that the resistant phenotype followed by a one-hour treatment with cisplatin, an of the cellular system could not be attributed to a increase in cisplatin cytotoxicity was observed only in marginal difference in the proliferative ability. Since cisIGROV-1/Ptl cells. This finding suggests that intracellu- platin cytotoxic action is somewhat independent of prolar GSH may influence cisplatin cytotoxicity when the liferative status [29], the slight differences found in the thiol is present at high concentrations, but not under cell cycle-phase distribution of the resistant variants do physiological (< 3 raM) conditions. However, the degree not appear relevant in the acquisition of resistance.

429 ity, it has been suggested that changes in GSH content could contribute to Taxol sensitivity [32]. However, this does not appear to be relevant to our cellular model in which a marginal modification of Taxol IC50s after GSH depletion was found in IGROV-1 and IGROV-1/Ptl cells. Although the basis of the collateral sensitivity have not yet been clarified, it may have important implications in clinical therapy with Taxol, which is now employed in combination with cisplatin or after chemotherapy with platinum compounds for treatment of ovarian carcinoma. Clinical studies do not unequivocally support the observation of enhanced sensitivity of cisplatin-resistant cells to Taxol since Taxol shows a poor response rate in ovarian cancer patient progressing on cisplatin treatment [33]. It is likely that the pattern of sensitivity is dependent on the mechanism of drug resistance. In ovarian carcinoma, p53 mutations may confer resistance to DNA-damaging agents as a consequence of reduced cell ability to undergo apoptosis [34]. This alteration is not expected to affect cell sensitivity to Taxol. Indeed, a different pattern of tumor response in p53 mutation tumors following Taxol-based chemotherapy suggests a collateral sensitivity to first-line therapy (i.e., in conditions in which drug-induced alternative defence mechanisms are less likely to be relevant) [35]. In conclusion, the reduced ability of resistant ovarian carcinoma cells to activate apoptosis in response to cisplatin treatment reflects not only cell alterations in specific pathways of the process [12] but also the extent of 'cytotoxic' DNA lesions (that may be influenced by defense mechanisms) and/or an increased tolerance to drug-induced DNA damage. Multiple cellular alterations are probably implicated in the development of cisplatin resistance. Their contribution is related to the Of multiple factors that contribute to cellular resist- degree of resistance. The specific mechanisms of cisplaance to cisplatin, a reduction of DNA platination and tin resistance in our cell system are probably involved in an increased tolerance to DNA damage have been pro- a collateral sensitivity to Taxol, emphasizing the lack of posed [5, 6, 9, 10]. However, we observed an appreciable cross-resistance in the clinical setting [35] and thus a reduction in DNA-bound platinum only in IGROV-1/ pharmacological benefit for the combination of platiPtl cells. Since the frequency of ICL observed in sensi- num compounds with Taxol, as suggested in preliminary tive and resistant cells following exposure to equitoxic clinical studies [36]. cisplatin concentrations was higher in IGROV-1 /Pt cells than in IGROV-1 cells, it is likely that an increased tolerance to DNA damage contributes to the cisplatin- Acknowledgements resistance phenotype of the cells. An interesting finding of the study was the collateral This work was partially supported by the Consiglio sensitivity to Taxol, an antitumor agent that interacts Nazionale delle Ricerche (finalized project 'Applicazioni with tubulin-stabilizing microtubules [30]. Such an ob- Cliniche della Ricerca Oncologica'), by the Associaservation is particularly intriguing since the IGROV-1/Pt zione Italiana per la Ricerca sul Cancro and by the cells have a mutant p53 [12]. Indeed, it has been postu- Ministero della Sanita'. We wish to thank L. Zanesi for lated that p53 status might influence sensitivity to Taxol editorial assistance. resulting from increased susceptibility to apoptosis after persistent G2M arrest [31]. Since p53 inactivation is associated with upregulation of bcl-2 and downregulation of bax [14], it is possible that Taxol treatment may References modulate the relative levels of these apoptosis regulators 1. Ozols RF, Young RC. Chemotherapy of ovarian cancer. Semin (e.g., bcl-2 phosphorylation). Based on the observation Oncol 1991; 18:222-32. that glutathione depletion antagonizes Taxol cytotoxic2. Thigpen JT, Blessing JA, Vance RB, Lambuth BW. Chemotherapy Postulated mechanisms of cellular protection against cisplatin include decreased drug accumulation and increased levels of intracellular thiols. Both events are expected to reduce drug binding to DNA. Thus, in this study, we measured the DNA-bound platinum as a parameter more closely related to drug effects than the total cellular content of platinum. An appreciable reduction of DNA-bound platinum was found only in the subline with the highest degree of resistance. The decrease was consistent with a lesser degree of ICL formation. It remains to be determined whether an accumulation defect contributes to such a change. However, since a marked increase in the cellular levels of GSH was observed in IGROV-1/Ptl over that of the parental cell line, it is likely that the thiol is effective in reducing the ability of the drug to form DNA cross-links. Indeed, depletion of GSH produces a marginal sensitization to cisplatin. However, a substantial contribution of GSH to response to cisplatin seems unlikely, since depleted cells retain a high degree of resistance. Again, the role of a glutatione-dependent system in the cells' ability to inactivate reactive forms of cisplatin appears questionable, considering the consistent reduction of GST implicated in detoxification of cytotoxic agents. The decreased expression paralleled a similar observation in ovarian cancer patients who were unresponsive to high-dose cisplatin therapy (manuscript in preparation). However, the defence capability of the resistant variants appeared to be increased by overexpression of the metallothionein Ila gene (Figure 3). A mechanism of resistance, at least partially related to intracellular thiol content, is consistent with the pattern of cross-resistance of the IGROV-1/Ptl subline (i.e., a high level of resistance to the bifunctional alkylating agent, melphalan) and with the lower sensitivity of cells to cadmium chloride.

430

3. 4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15. 16. 17. 18.

19.

20.

21.

in ovarian carcinoma. Present role and future prospects. Semin Oncol 1989; 16: 58-65. Pinto AL, Lippard SJ. Binding of the antitumor drug cisplatin to DNA. Biochem Biophys Acta 1985; 780: 167-80. Lippard SJ, Hoeschele JD. Binding of cis- and trans-dichlorodiammine platinum (II) to the nucleosome core. Proc Natl Acad Sci USA 1979; 76: 6091-5. Loh SY, Mistry P, Kelland LR. Reduced drug accumulation as a major mechanism of acquired resistance to cisplatin in a human ovarian carcinoma cell line: Circumvention studies using novel platinum(II) and (IV) ammine/amine complexes. Br J Cancer 1992; 66: 1109-15. Mistry P, Kelland LR, Loh SY. Comparison of cellular accumulation and cytotoxicity of cisplatin with that of tetraplatin and amminedibutyratodichloro(cyclohexylamine)platinum (IV) (JM22I) in human ovarian carcinoma cell lines. Cancer Res 1992; 52: 6188-93. Mistry P, Kelland LR, Abel G. The relationships between glutathione, glutathione-S-transferase and cytotoxicity of platinum drugs and melphalan in eight human ovarian carcinoma cell lines. BrJ Cancer 1991: 64: 215-20. Fram RJ, Woda BA, Wilson JM, Robichaud N. Characterization of acquired resistance to cis-diamminedichloroplatinum(II) in human colon carcinoma cells. Cancer Res 1990; 50: 72-7. Kelland LR, Mistry P, Abel G et al. Establishment and characterization of an in viiro model of acquired resistance to cisplatin in a human testicular non-seminomatous germ cell line. Cancer Res 1992; 52: 1710-6. Eastman A, Schulte N. Enhanced DNA repair as a mechanism of resistance to cis-diamminodichloroplatinum (II). Biochemistry 1988; 27. 4730-4. Lempers ELM, Reedijk J Interaction of platinum amine compounds with sulfur-containing biomolecules and DNA fragments. Inorganic Chem 1991; 37: 176-217. Perego P, Giarola M. Righetti SC et al. Association between cisplatin resistance and mutation of p53 gene and reduced bax expression in ovarian carcinoma cell systems. Cancer Res 1996; 56: 556-62. Benard JB, Da Silva J, De Blois MC et al. Characterization of a human ovarian adenocarcinoma line, IGROV-1, in tissue culture and in nude mice. Cancer Res 1985, 45. 4970-9 Perego P, Righetti SC, Supino R et al. Role of apoptosis and apoptosis-related proteins in the cisplatin-resistant phenotype of human tumor cell lines. Apoptosis 1997. 2: 540-8. Seabnght M. A rapid banding technique for human chromosomes. Lancet 1971; 2: 971-2. Hardner DG, Klinger HP ISNC 1991: An International System for Human Cytogenetic Nomenclature. Basel: Karger 1991; 2-47. Ellman GE. Tissue sulphydryl groups. Arch Biochem Biophys 1959: 82: 70-7. Pratesi G, Dal Bo L, Paolicchi A et al. The role of glutathionedependent system in tumor sensitivity to cisplatin. A study of human tumor xenografts. Ann Oncol 1995: 6: 283-9. Alley MC, Scudiero DA. Monks A et al. Feasability of drug screening with panels of human tumor cell lines using a microculture tetrazolium assay. Cancer Res 1988: 48: 589-601. Bonetti A. Apostoli P. Zaninelli M et al. Inductively coupled plasma mass spectroscopy quantitation of platinum-DNA adducts in peripheral blood leukocytes of patients receiving cisplatin- orcarboplatin-based chemotherapy. Clin Cancer Res 1996; 2: 1829-35. Kohn KW. DNA as a target in cancer chemotherapy: Measure-

22.

23.

24.

25.

26. 27.

28

29.

30.

31.

32.

33. 34.

35. 36.

ment of macromolecular DNA damage produced in mammalian cells by anticancer agents and carcinogens. Meth Cancer Res 1979: 16. 291-8. Kohn KW, Ewig RAG, Erickson LC. Zwelling LA. Measurement of strand breaks and cross-links by alkaline elution. In Friedberg EC, Hanawalt PC (eds): DNA Repair. A Laboratory Manual of Research Procedures. Vol. 1. Part B. New York: Marcel Dekker 1981; 370-401. ManiatisT. Fritsh EF, Sambrook J. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor. NY: Cold Spring Harbor Laboratory 1982. Perego P, Capranico G, Supino R, Zunino F. Topoisomerase I gene expression and cell sensitivity to camptothecin in human cell lines of different tumor types. Anti-Cancer Drugs 1994; 5: 645-9. Tan KB, Mattern MR, Boyce RA, Schein PS. Elevated DNA topoisomerase II activity in nitrogen mustard-resistant human cells. Proc Natl Acad Sci USA 1987; 84: 7668-71. Morrow CS, Cowan KH. Glutathione S-transferases and drug resistance. Cancer Cells 1990; 2: 15-22. Hong WS, Saijo N, Sasaki Yet al. Establishment and characterization of cisplatin-resistant sublines of human lung cancer cell lines. Int J Cancer 1988; 41- 462-7. Supino R, Prosperi E, Formelli F et al. Characterization of a doxorubicin-resistant murine melanoma line: Studies on crossresistance and its circumvention. Br J Cancer 1986; 54: 33-42. Ozawa S, Sugiyama Y, Mitsuhashi J, Inaba M. Kinetic analysis of cell killing effect induced by cytosine arabinoside and cisplatin in relation to cell cycle phase specificity in human colon cancer and Chinese hamster cells. Cancer Res 1989; 49: 3823-8 Rowinsky EK, Donehower RC, Jones RJ, Tucker RW Microtubule changes and cytotoxicity in leukemic cell lines treated with Taxol. Cancer Res 1988; 48: 4093-100. Wahl AF, Donaldson KL, Fairchild G et al. Loss of normal p53 function confers sensitization to Taxol by increasing G2/M arrest and apoptosis. Nature Med 1996; 2: 724-5. Liebmann JE, Hahn SM, Cook JA. Glutathione depletion by L-buthionine sulfoximine antagonizes Taxol cytotoxicity. Cancer Res 1993; 53: 2066-70. Kaufmann M., von Minckwitz G. Gemcitabine in ovarian cancer: An overview of safety and efficacy. Eur J Cancer 1997; 33: 31-3. Righetti SC, Delia Torre G, Pilotti S et al. A comparative study of p53 gene mutations, protein accumulation, and response to cisplatin-based chemotherapy in advanced ovarian carcinoma. Cancer Res 1996; 56: 689-93. Lavarino C. Delia D, Di Palma S et al. P53 in drug resistance in ovarian cancer. Lancet 1997; 349: 1556. McGuire WP. Hoskins WJ, Brady MF et al. Cyclophosphamide and cisplatin compared with paclitaxel and cisplatin in patients with stage III and stage IV ovarian cancer. N Engl J Med 1996; 334: 1-6.

Received 7 July 1997; accepted 8 January 1998.

Correspondence to: Dr. Paola Perego Istituto Nazionale Tumori Via Venezian 1 20133 Milan Italy E-mail: peregoC« istitutotumori.mi.it.