Characterization of an Ovarian Carcinoma Cell Line Resistant to Cisplatin and Flavopiridol

Materials.

Flavopiridol and BCNU were provided by the Pharmaceuticals Resources Branch of the National Cancer Institute (Bethesda, MD). Topotecan was kindly provided by Dr. Randall K. Johnson (SmithKline Beecham, King of Prussia, PA). Oxaliplatin from Sanofi Research (Malvern, PA) was provided by Dr. Charles Erlichman (Mayo Clinic). Cytarabine, 5-fluorouracil, etoposide, paclitaxel, doxorubicin and cisplatin were purchased from Sigma Chemical Co. (St. Louis, MO). Antibodies to cyclin A, cyclin E, cdc2/CDK1, CDK2, and CDK4 (26) were kindly provided by Dr. Robert Abraham (Duke University, Durham, NC). Antibodies to the following polypeptides were purchased from the indicated suppliers: cyclins B, D₁, and D₃ from Transduction Laboratories (Lexington, KY); cyclin D₁ from Santa Cruz Biotechnology (Santa Cruz, CA); and Rb from PharMingen (San Diego, CA). All other reagents were obtained as described previously (8 , 27) .

Stock (1000-fold concentrated) solutions of flavopiridol, cytarabine, paclitaxel, topotecan, doxorubicin, etoposide, and 5-fluorouracil were prepared in DMSO and stored at −20°C prior to use. Cisplatin, oxaliplatin and BCNU were prepared immediately before use as 1000-fold concentrated solutions in DMSO. The absorption spectrum of flavopiridol was determined using a Beckman (Palo Alto, CA) DU7400 diode array spectrophotometer using 100 μl sample cells and 1% aqueous triethylamine (pH 5.0) as a “blank.”

Tissue Culture and Colony-forming Assays.

OV202 and OV202 hp cells were cultured in MEM with Earle’s salts and nonessential amino acids containing 20% (v/v) heat-inactivated fetal bovine serum, 100 units/ml penicillin G, 100 μg/ml streptomycin, and 2 mm glutamine (medium A). Cells were passaged once weekly and maintained at 37°C in an atmosphere containing 95% air-5% CO₂ (v/v). To determine population doubling times, 1 × 10⁵ cells were seeded in replicate 60-mm tissue culture plates, incubated for 24–96 h, trypsinized, and counted on a hemacytometer.

Colony-forming assays were performed as described previously (9) . In brief, subconfluent cells were released with trypsin, plated at a density of 3000 cells/plate in multiple 35-mm dishes containing 2 ml of medium A and incubated for 14–16 h at 37°C to allow cells to attach. Graded concentrations of each drug or equivalent volumes of DMSO (0.1%) were then added to triplicate plates. After a 24-h treatment, plates were washed twice with serum-free MEM and incubated in drug-free medium A for an additional 14 days. Alternatively, flavopiridol was added on the day after plating and left in the culture. In either case, the resulting colonies were stained with Coomassie Blue and counted. Diluent-treated control plates typically contained 175–225 colonies.

Cytogenetic Studies.

Routine cytogenetic analyses were performed using in situ culture and robotic harvesting as described previously (28) . Spectral karyotype analysis of previously G-banded metaphase preparations was performed by the method of Veldman et al. (29) using reagents, instrumentation, and software purchased from Applied Spectral Imaging (Carlsbad, CA).

Immunoblotting.

OV202 and OV202 hp cells grown to 50% confluence in triplicate 100-mm tissue culture dishes were washed four times in situ with 10-ml volumes of calcium- and magnesium-free PBS and solubilized in alkylation buffer [6 m guanidine hydrochloride, 250 mm Tris-HCl (pH 8.5 at 21°C), and 10 mm EDTA supplemented immediately before use with 150 mm β-mercaptoethanol and 1 mm α-phenylmethylsulfonyl fluoride]. Samples were then processed for SDS-PAGE and subsequent immunoblotting using techniques described in detail previously (27) .

Flow Cytometry.

OV202 and OV202 hp cells were grown to 30–40% confluence in 100-mm tissue culture dishes, treated with the indicated concentrations of flavopiridol, released by trypsinization, and sedimented at 200 × g for 5 min. Samples were then washed in PBS at 4°C, fixed in 50% ethanol, digested with RNase A, and stained with propidium iodide as described (8) . Flow cytometry was performed on a Becton Dickinson FACScan (San Jose, CA) using an excitation wavelength of 488 nm and an emission wavelength of 585 nm. Histograms were analyzed using ModFit software (Verity Software House, Topsham, ME).

Assessment of Cellular Platinum Accumulation and Platinum-DNA Adducts.

For the assessment of whole cell platinum accumulation, OV202 or OV202 hp cells grown to 50% confluence on quadruplicate 100-mm tissue culture dishes were treated with cisplatin or oxaliplatin (added from freshly prepared 1000-fold concentrated stocks). After incubation for the indicated period of time, cells were washed in situ four times with ice-cold PBS and solubilized by direct addition of 3 ml of 70% nitric acid to the first of each set of four plates followed by serial transfer of this solution to each additional plate in each set to solubilize the remaining cells. After a 24 h incubation at 20–22°C, aliquots were removed for estimation of protein and elemental platinum. To assess protein content, samples were diluted 1:2 or 1:4 in nitric acid and assayed for absorbance at 358 nm. Results were compared to absorbance of known amounts of BSA in the same diluent (30) . To assess platinum content, aliquots containing 1 ml of nitric acid extract were heated to 100°C for 5 min, combined with 1 ml of 30% H₂O₂, heated to 100°C for 5 min, and subjected to mass spectrometry as described below.

For the assessment of platinum-DNA adducts, OV202 or OV202 hp cells grown to 50–80% confluence on four to six 100-mm tissue culture dishes were treated with cisplatin or oxaliplatin as described above. At the completion of the incubation, cells were scraped from the plates without removal of drug-containing medium, sedimented at 200 × g for 10 min, washed four times with 10-ml volumes of ice-cold PBS, and lysed in 5 ml of TEN buffer [10 mm Tris-HCl (pH 7.4 at 21°C), 10 mm EDTA, 150 mm NaCl] containing 0.4% SDS and 1 mg/ml proteinase K. After proteinase K digestion, highly purified DNA was isolated by extraction with phenol/chloroform and chloroform, ethanol precipitation, digestion with RNase A, and extraction with phenol/chloroform and chloroform as described previously (31) .

Elemental platinum in cell lysates and purified DNA was determined by inductively coupled plasma mass spectrometry using a Perkin-Elmer Sciex Elan 6000 mass spectrometer (Norwalk, CT) as recently described in detail (32) . Platinum species at 194 and 195 a.m.u. were summed using a program that sweeps 1–263 a.m.u. 50 times/reading. Platinum standards [0–20 ng/ml in 0.6 m HCl or 1:1 (70%) nitric acid:H₂O₂ as appropriate] were used to confirm the linearity of the assay (R ≥ 0.999), and rhodium 103 served as an internal standard. Each sample was evaluated in duplicate, and the means of the duplicate results were used in all calculations. This approach was 10–100-fold more sensitive than the atomic absorption spectroscopy method used previously (31) .

Assessment of Intracellular Flavopiridol Concentrations.

To measure cellular flavopiridol concentrations, flavopiridol was released from washed cells by PCA treatment, extracted into chloroform, and subjected to HPLC. Briefly, cells grown to 50% confluence in 50–100 10-cm tissue culture dishes were treated with 9μ m flavopiridol for 24 h.5 At this time point, >98% of cells remained adherent and excluded trypan blue, consistent with the delayed death reported previously in solid tumor cells treated with flavopiridol (9) . After treatment, cells were washed rapidly in situ four times with 3-ml volumes of ice-cold PBS and scraped into 3 ml of 0.5 m PCA. The PCA extracts were combined on ice, sonicated, and sedimented at 200 × g for 10 min to remove macromolecules. The resulting supernatant (∼150 ml) was extracted four times with 25-ml volumes of chloroform. The PCA supernatant was then neutralized to pH ∼7.4 with 0.5 volumes of 0.33 m phosphate buffer (pH 7.4) that was supplemented with 1 m KOH (31) . After sedimentation to remove precipitated potassium perchlorate, the supernatant was extracted again four times with 25-ml volumes of chloroform.

The pooled chloroform extracts were evaporated to dryness under anhydrous nitrogen. The resulting residue was solubilized in 100 μl of 1% (v/v) aqueous triethylamine adjusted to pH 5.0 with acetic acid (mobile phase A), a solvent chosen because of low flavopiridol solubility at higher pH. After centrifugation at 200 × g for 10 min to remove insoluble material, 60 μl of the resulting supernatant were subjected to HPLC analysis using a Beckman 125 dual pump gradient system equipped with 507e autosampler, 168 diode array detector, and IBM 350 personal computer with Beckman Gold Nouveau software. A Brownlee MPLC Newguard C18 precolumn (3.2 mm × 15 mm × 7 μm) and a Beckman Ultrasphere ODS column (4.6 mm × 250 mm × 5 μm) preequilibrated for at least 30 min with 80% mobile phase A and 20% methanol were used for all analyses. Separation was accomplished using a flow rate of 2 ml/min and the following elution gradient: 0–10 min of 80% mobile phase A, 20% methanol; 10–20 min of linear ramp to 20% mobile phase A, 80% methanol; 20–25 min of linear ramp to 100% methanol; 25–35 min of 100% methanol.

As indicated below, the retention time of flavopiridol using this system was 18.6 min. Peak areas were measured at 311 nm, aλ _max of flavopiridol in mobile phase A. Detection at 311 nm allowed baseline chromatographic resolution in comparison to using detection at 254 nm or other wavelengths. Standard curves of peak area at 311 nm versus flavopiridol concentration generated with each experiment by assay of multiple concentrations of authentic flavopiridol dissolved in mobile phase A uniformly gave linear correlation coefficients >0.99. To provide a basis for comparison of the flavopiridol contents, protein concentrations were determined by solubilizing the PCA pellets in 6 ml of 70% nitric acid, incubating for 24 h at 20–22°C, and comparing the absorbance of the resulting solutions at 358 nm with that of known amounts of BSA solubilized under the same conditions (30) .

Statistics.

Differences between cell lines were assessed using two-sided t tests and pooled estimates of variance.

Characterization of OV202 and OV202 hp Cell Lines.

During the course of studies on cisplatin sensitivity in a series of recently described primary ovarian carcinoma lines (33) , we observed that one of the lines spontaneously developed resistance to flavopiridol and cisplatin upon prolonged passage. The present studies were designed to elucidate the mechanisms of this resistance.

In view of the recent determination that MCF-7/Adr^R cells were probably not derived from MCF-7 cells (34) , we first examined the relationship between the parental and hp cells. Cytogenetic analysis of the parental OV202 cells (studied in passage 24) and the highly passaged derivative OV202 hp (studied at passages 100–120) showed that the two lines had multiple chromosomal anomalies, including several marker chromosomes that were shared between the two lines (G-banded karyotypes not shown). Because the markers were difficult to accurately characterize by routine cytogenetic analysis, we also performed spectral karyotyping on metaphase spreads. Table 1⇓ summarizes the results of these analyses. The two lines shared several abnormalities, confirming their relationship to each other. For example, both cell lines contained a chromosome 20 derivative with a HSR derived from chromosome 19. Fluorescence in situ hybridization analysis demonstrated that this HSR contains multiple copies of the AKT2 gene (data not shown), an oncogene that is sometimes amplified in ovarian cancer (35) . In addition, both lines shared additional marker chromosomes derived by translocations involving chromosomes X and 9, 1 and X, 1 and 13, 2 and 6, 3 and 5, 3 and 6, 5 and 21, 4 and 11, 9 and 12, 3 and 13, and 5 and 18. Despite these shared marker chromosomes, several differences were also observed. OV202 hp, for example, contained an unbalanced 3;11 translocation and two different types of chromosome 5 deletions that were not observed in OV202. Additional differences are summarized in Table 1⇓ .

Analysis of the growth characteristics of these cell lines revealed that OV202 hp cells proliferated more rapidly, with a mean population doubling time about half that of OV202 cells (45 ± 14 versus 86 ± 10 h; P < 0.05). At confluence, OV202 cells formed a uniform monolayer, whereas OV202 hp cells piled up, leading to a 50% higher cell density for OV202 hp at confluence. The plating efficiencies of the two lines were essentially the same (OV202, 5.4 ± 0.2%; OV202 hp, 6.3 ± 1.1%; P > 0.2).

When cell cycle regulatory proteins were examined by immunoblotting (Fig. 1A)⇓ , levels of cyclin A, cyclin E, CDK2, CDK4, and proliferating cell nuclear antigen were indistinguishable in the OV202 and OV202 hp cells. Cyclin D₁, which was at the limit of detection with two separate antibodies, likewise appeared to be equally expressed in the two cell lines. In contrast, levels of cyclin D₃, cdc2/CDK1 and Rb were approximately 2-fold higher in the OV202 hp cells, whereas cyclin B was approximately 2-fold lower.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Comparison of OV202 and OV202 hp cells. A, immunoblot analyses of various polypeptides in OV202 and OV202 hp cells. Aliquots of whole cell lysates containing 50 (lanes 1 and 4), 25 (lanes 2 and 5), or 12.5 (lanes 3 and 6)μ g of protein were subjected to SDS-PAGE followed by immunoblotting using antibodies that recognize the indicated polypeptides. Immunoblots are representative of three independent experiments. B, cell cycle distributions of untreated and flavopiridol-treated OV202 and OV202 hp cells. Cells grown to 30% confluence were treated for 24 h with the indicated concentration of flavopiridol and harvested for flow cytometry. Data shown are representative of three independent experiments.

Despite the differences in population doubling time and certain cell cycle regulatory proteins, log phase OV202 and OV202 hp cells had similar cell cycle distributions, as assessed by flow cytometry (Fig. 1B)⇓ . However, after exposure to flavopiridol for 24 h, a treatment designed to mimic one of the flavopiridol administration schedules currently being evaluated clinically, the two lines differed (Fig. 1B)⇓ . Whereas OV202 cells demonstrated a dose-dependent increase in G₂-M population in response to flavopiridol, OV202 hp cells did not ⇓ . These observations raised the possibility that OV202 hp cells might be resistant to the effects of flavopiridol.

Sensitivity of OV202 and OV202 hp to Chemotherapeutic Agents.

The sensitivity of OV202 and OV202 hp cells to various chemotherapeutic agents was further assessed using colony-forming assays. After a 24 h flavopiridol treatment, OV202 and OV202 hp cells displayed IC₉₀s of 220 ± 23 and 1100 ± 120 nm, respectively, indicating a 5-fold decrease in flavopiridol sensitivity in OV202 hp cells (Fig. 2A⇓ ; P < 0.001). When cells were instead exposed to flavopiridol continuously, OV202 and OV202 hp displayed IC₉₀s of 71 ± 10 and 145 ± 23 nm, respectively (P < 0.02; Fig. 2B⇓ ). Additional experiments demonstrated that the diminished sensitivity was also observed when cell death was directly measured by determining uptake of the nonpermeable dye trypan blue or by assessing the area of the sub-diploid peak by FACS analysis at various times after a 24-h flavopiridol exposure.6 Likewise, after a 24-h cisplatin treatment, OV202 and OV202 hp cells displayed IC₉₀s of 850 ± 50 and 2900 ± 100 nm, respectively, in colony-forming assays, indicating a 3-fold decrease in cisplatin sensitivity in OV202 hp cells (Fig. 2C⇓ ; P < 0.001). In contrast, OV202 and OV202 hp cells exhibited indistinguishable sensitivities to oxaliplatin, paclitaxel, topotecan, BCNU, 5-fluorouracil, doxorubicin, etoposide, and cytarabine (Fig. 2, D–F⇓ , and data not shown). As discussed below, this pattern of selective resistance cannot be explained by up-regulation of P-glycoprotein, enhanced nucleotide excision repair, or a general failure of OV202 hp cells to engage the cell death machinery. These considerations prompted us to search for drug-specific factors that might account for the resistance.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Effects of antineoplastic agents on colony formation in OV202 and OV202 hp cells. Triplicate plates of cells were exposed to the indicated concentrations of flavopiridol (A), cisplatin (C), oxaliplatin (D), paclitaxel (E), or topotecan (F) for 24 h, washed, and incubated in drug-free medium until colonies were counted. Alternatively, cells were exposed to flavopiridol continuously until colonies were counted (B). Error bars, ± 1 sample SD. Data shown are representative of at least three independent experiments for each drug and each cell line.

Whole-cell Platinum Accumulation and Platinum-DNA Adducts after Treatment with Cisplatin or Oxaliplatin.

To evaluate the hypothesis that cisplatin resistance in OV202 hp cells results from decreased cisplatin-induced DNA damage, covalently bound platinum was measured by mass spectroscopy after purification of DNA from OV202 and OV202 hp cells exposed to cisplatin for 24 h, the same incubation time used in the colony-forming assays. Cisplatin treatment of OV202 hp cells consistently produced fewer platinum-DNA adducts in comparison to OV202 cells (120 ± 20 versus 200 ± 20 pg of platinum/μg of DNA; P < 0.01; Fig. 3A⇓ ). Additional experiments involving cisplatin exposure followed by incubation in drug-free medium indicated that the rates of removal of platinum-DNA adducts in the two lines were similar, with persistence of ∼75% of the covalently bound platinum in both cell lines 48 h after cisplatin removal.⁶

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Comparison of platinum accumulation and formation of platinum-DNA adducts in OV202 and OV202 hp. A, platinum-DNA adducts resulting from exposure of OV202 and OV202 hp cells to 40 μm cisplatin for 24 h. Error bars, +1 SD from four separate experiments. B, whole cell platinum accumulation in OV202 and OV202 hp cells exposed to 40 μm cisplatin for 24 h. Error bars, +1 SD from four separate experiments. C, platinum-DNA adducts resulting from exposure of OV202 and OV202 hp cells to 40 μm oxaliplatin for 24 h. Error bars, +1 SD from three separate experiments. D, comparison of the kinetics of cisplatin uptake into OV202 and OV202 hp cells. Cells were exposed to 40 μm cisplatin for the indicated length of time, washed, and analyzed for total cellular platinum content as described in “Materials and Methods.” E, comparison of efflux kinetics. Cells were exposed to 40 μm cisplatin for 5 h, washed, incubated in drug-free medium for the indicated length of time, and analyzed for total cellular platinum content as described in“ Materials and Methods.” Error bars in D and E, mean ± 1 sample SD of triplicate samples, each consisting of three plates. Each experiment was repeated two additional times with similar results.

To determine whether the lower levels of platinum-DNA adducts in cisplatin-treated OV202 hp cells might result from decreased drug accumulation, the platinum content of whole cell lysates was measured after a 24-h cisplatin treatment. Total cellular platinum content was consistently lower in OV202 hp cells than in parental cells (Fig. 3B⇓ ; 140 ± 20 versus 230 ± 50 pg of platinum/μg of protein; P < 0.02), suggesting that reduced cisplatin accumulation in OV202 hp cells might contribute to their resistance. This decreased accumulation of cisplatin by OV202 hp cells was evident at drug concentrations as low as 4μ m7 and at exposure times much shorter than 24 h (see below). In contrast, the two cell lines did not exhibit any difference in platinum-DNA adducts after treatment with oxaliplatin (Fig. 3C⇓ ; 82 ± 3 versus 86 ± 8 pg of platinum/μg of DNA; P > 0.5). This latter result is consistent with the equal sensitivities of the cell lines to oxaliplatin in clonogenic assays (Fig. 2D)⇓ and suggests that the mechanism of resistance to cisplatin does not convey resistance to all agents that form platinum-DNA adducts.

To begin to investigate potential explanations for diminished cisplatin accumulation in OV202 hp cells, the kinetics of cisplatin update (Fig. 3D)⇓ and efflux (Fig. 3E)⇓ were examined. After cells were loaded with cisplatin, the rate of efflux was lower in the OV202 hp cell line ⇓ , arguing against the possibility that overexpression of an efflux pump was responsible for the observed differences in drug accumulation. On the other hand, the amount of drug accumulated at each time point after drug addition was lower in the OV202 hp line than in the OV202 line (Fig. 3D)⇓ . Collectively, these results suggest that the rate of cisplatin transport into the OV202 hp cells is lower than into OV202 cells.

Cellular Flavopiridol Concentrations in OV202 and OV202 hp.

To evaluate the possibility that flavopiridol resistance in OV202 hp cells might also result from decreased cellular drug accumulation, flavopiridol content was assessed in OV202 and OV202 hp cells exposed to flavopiridol. For these experiments, we developed a new HPLC assay for flavopiridol. Based upon the observation that flavopiridol is much less soluble in aqueous buffers at neutral or alkaline pH, cells were lysed in PCA. Flavopiridol in these lysates was extracted into chloroform, concentrated, and subjected to reverse phase HPLC under slightly acidic conditions (Fig. 4A)⇓ . Absorbance of the flavopiridol chromophore at 311 nm (Fig. 4B)⇓ was used for detection. These conditions yielded good baseline separation of flavopiridol from the few other peaks that were present in the extracts (Fig. 4A)⇓ and permitted generation of linear standard curves with correlation coefficients uniformly >0.99.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Assessment of cellular flavopiridol concentrations in OV202 and OV202 hp cells. A, HPLC chromatogram showing a representative elution profile of flavopiridol (detection at 311 nm) for an extract of OV202 cells exposed to 9μ m flavopiridol for 24 h. B, absorption spectrum of flavopiridol (18 μm) dissolved in mobile phase A. C, cellular flavopiridol concentrations in OV202 and OV202 hp cells exposed to 9 μm flavopiridol for 24 h. Results are representative of three independent experiments.

Application of this assay (Fig. 4C)⇓ revealed that OV202 hp cells exposed to flavopiridol for 24 h consistently contained lower cellular flavopiridol concentrations than OV202 cells subjected to identical treatment. The mean flavopiridol concentration in OV202 hp cells was 49 ± 21% that in OV202 cells. These data suggest that flavopiridol resistance in OV202 hp cells, like cisplatin resistance, reflects decreased cellular drug accumulation.

Effect of Flavopiridol Pretreatment on Cisplatin Accumulation.

To assess the possibility that flavopiridol and cisplatin might compete for the same transporter in OV202 cells, cells were exposed to cisplatin in the absence or presence of flavopiridol (Fig. 5)⇓ . At molar excesses of up to 4-fold, flavopiridol consistently increased rather than decreased whole cell platinum accumulation. These data do not provide any evidence for transport of flavopiridol and cisplatin by a single transporter, but they do provide a potential explanation for the previously described synergy between flavopiridol and cisplatin (8) as discussed below.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 5.

Effect of flavopiridol on platinum uptake in OV202 and OV202 hp cells. Cells were treated the indicated concentration flavopiridol or diluent for 24 h. In the continued presence of flavopiridol or diluent, 4 μm cisplatin was then added for 24 h. At the completion of the second incubation, cells were harvested for measurement of whole cell platinum as described in “Materials and Methods.”