Doxorubicin (adriamycin) and epirubicin (epidoxorubicin) (Farmitalia Carlo Erba, Milan, Italy), cis-diaminedichloroplatinum (II) (Lennon, South Africa), gentamicin sulphate (Roussel Laboratories, South Africa), vincristine sulphate, 3- [4,5-dimethylthiazole-2-yl]-2,5-diphenyltetrazolium bromide (MTT) (Sigma Chemical Co., St. Louis, MO, USA), trypsin 1:250 (Difco Laboratories, Detroit, MI, USA), PBS Dulbecco 'A' (Oxoid, UK), RPMI-1640 and McCoy's 5A culture medium (Gibco, UK), tunicamycin, penicillin G (Boehringer Mannheim, Germany), EDTA (Merck Chemicals, Germany), radioactive isotope precursors and drugs (Amersham Biosciences, UK), Bio-Rad protein assay dye reagent concentrate (Bio-Rad Laboratories, UK) were used in this study. All other reagents were of the highest analytical grade and were obtained from either Sigma Chemical Co. or Merck Chemicals. P-glycoMab (consisting of lyophilized C219 monoclonal antibody and isotype-matched negative antibody, biotinylated anti-mouse antibody, avidin and biotinylated horseradish peroxidase visualization system) was purchased from Centocor Diagnostics, Tongeren, Belgium.

The UWOV2 ovarian carcinoma cell line, derived from a cystadenocarcinoma in a patient who had not responded to combination chemotherapy with actinomycin D, vincristine, cisplatin and doxorubicin 75, was maintained in RPMI-1640 medium supplemented with 5% heat-inactivated foetal calf serum (HIFCS), penicillin G (100 U/ml) and streptomycin sulphate (100 μg/ml) or gentamicin sulphate (50 μg/ml) at 37°C in 5% CO₂:air and 85% relative humidity. The expression of Pgp in UWOV2 cells was demonstrated previously using the monoclonal antibody, C219, and the avidin-biotin-immunoperoxidase P-glycoCHEK diagnostic kit with the drug-sensitive human acute lymphoblastic leukemia cell line, CCRF-CEM (Pgp-negative) and its drug-resistant derivative, CEM-VLB100 (strongly Pgp-positive) as controls 76. The BG-1 ovarian carcinoma cell line and its adriamycin-resistant derivative, BG-1/ADR were a gift from Dr C.A. Wallen (Bowman Gray School of Medicine, North Carolina, USA) and grown in McCoy's 5A medium containing 10% HIFCS, 100 U/ml insulin, 200 mM L-glutamine and antibiotics. Cells were routinely subcultured with trypsin-EDTA (0.25%–0.02%, w/v) in Ca⁺²- and Mg⁺²-free PBS and maintained in the logarithmic phase of growth. Cells were periodically tested and found to be free of mycoplasma contamination.

The effects of TM on the incorporation of radiolabelled precursors [³⁵S]methionine (>1000 Ci/mmole) and [³H]mannose (30–60 Ci/mmole) or [³H]glucosamine (20–40 Ci/mmole) into trichloroacetic acid-insoluble macromolecules (proteins and glycoproteins) were measured in UWOV2 cells as described previously 29.

To determine the effect of TM on the cytotoxicity of DXR, EXR, VCR and CDDP, preconfluent cells from stock cultures were detached with trypsin-EDTA (0.25%–0.02%, w/v) in PBS, washed twice with PBS and resuspended in complete culture medium to obtain single-cell suspensions. Standardization of cell numbers in individual wells of a 96-well flat-bottom microtiter plate was achieved by a linear correlation (r = 0.98) between cell number and absorbance up to a maximum density of 3.5–4.5 × 10⁴ cells/well. Cells were seeded at a density of 3 × 10³ cells/well in a total volume of 200 μl as follows: After trypsinization, cells were rinsed twice with PBS, resuspended in 10 ml complete culture medium and repeatedly pipetted to ensure a homogeneous mixture during dispensing of 100-μl aliquot/well. The cells were then allowed to attach and grow for 48–72h. Cytotoxic drugs were dissolved in PBS and sterilized through 0.22-μm disposable filters (Millipore, Millex-GV). The drugs were diluted in culture medium free of phenol red (RPMI-1640-selectamine kit) to avoid interference with spectrophotometric assays. Tunicamycin stock solutions were prepared by dissolving the contents of a 10-mg vial in 25 mM NaOH and then diluting to 0.8 mg/ml TM and 10 mM NaOH with pyrogen-free distilled-deionized water. The solution was sterilized by passing through a 0.22-μm disposable filter and diluted to final concentrations in culture medium. After the addition of drugs and TM to octuplicate wells, cells were incubated for a further 72h. The cytotoxicity of drug in the TM-treated (combination) and TM-free cultures (control) was determined by the MTT (3–4,5-dimethylthiazol-2,5-diphenyltetrazolium bromide) assay 77. In this assay, 20 μl of MTT (5 mg/ml in sterile PBS) were added to each well and the plates incubated for 5h at 37°C. Plates were then centrifuged at 400 × g for 5 min to pellet any floating cell aggregates. The supernatant was aspirated and the formazan crystals dissolved in DMSO (200 μl/well). Absorbances were read by a microplate reader (Titertek Multiskan model MCC/340) at a sample wavelength of 540 nm and a reference wavelength of 630 nm. The EC₅₀ for TM, DXR, EXR, VCR and CDDP were determined by non-linear regression of sigmoidal dose-response curves using Graphpad Prism (Version 4.03, GraphPad Software, San Diego California USA, http://www.graphpad.com). The best-fit EC₅₀ (concentration that produces 50% of the maximal response) values for each drug alone or in combination with a fixed concentration of 5 μg/ml TM were subjected to statistical evaluation as described in "Data analysis".

To assay VCR uptake and efflux, cells were seeded at a density of 5 × 10⁴ cells/ml in 24-well plates and allowed to grow for 48h under standard conditions. Cells were pretreated with 5 μg/ml TM for 16h to suppress de novo protein and glycoprotein synthesis. Parallel controls were set up. Total cellular accumulation of VCR was determined by exposing cells in quadruplicate wells to [G-³H]VCR sulphate (30 nM or specific activity 9.48 cpm/pmol) in the continued absence or presence of 5 μg/ml TM in a final volume of 0.5 ml for various incubation times. At the end of each incubation period, cells were washed three times with 1 ml ice-cold PBS and solubilized in 0.5 ml of 1% SDS/0.3 M NaOH. One aliquot (0.4 ml) was neutralized by the addition of 0.2 ml of 2 M acetic acid and mixed with 10 ml scintillation fluid (Beckman Ready-Solv EP) and counted in a Beckman scintillation spectrometer. Intracellular drug at each time point was determined by subtracting the value for non-specific/surface-bound drug obtained by incubation with 100 μM unlabelled VCR for 30s at 4°C from the value for total drug. The other aliquot (0.1 ml) was assayed for total cellular protein. Vincristine efflux was measured by loading control and TM-pretreated cells with [³H]VCR for 60 min (0-time value for efflux) followed by washing preloaded cells three times with ice-cold PBS and subsequently incubating at 37°C in serum- and antibiotic-free medium (2 ml) for various time intervals. The absence or presence of TM was maintained throughout the post-incubation periods. Cells were harvested as described for uptake studies. A large volume ratio (i.e., preloading volume/post-incubation volume of 4) was ensured during efflux and retention measurements to minimize reutilization of extruded drug.

The specific binding of [³H]AZD (46 Ci/mmol) and [¹⁴C]DXR (50–62 mCi/mmol) to UWOV2 cells in the absence (control) or presence of TM (TM-treated) was measured by a modification of the procedure described earlier 78. Cells were seeded at a density of 5 × 10⁴ cells/ml in 24-well plates. Confluent cell monolayers were cooled by placing the plates on ice for 10 min, washed four times with 1 ml cold PBS, pH 7.4 (to remove culture medium serum glycoproteins) and maintained for 60 min at 4°C with binding buffer (10 mM glucose, 3 mM ATP and 5 mM MgCl₂ in 10 mM Tris-HCl, pH 7.4) containing [³H]AZD or [¹⁴C]DXR in serial dilutions of 10–80 nM. Following this incubation period, the cells were washed five times with cold PBS to remove unbound [³H]AZD and [¹⁴C]DXR. Cells were then solubilized in 0.1 M NaOH and samples were counted to determine the amount of AZD or DXR bound to the cells, and aliquots were removed for protein determination. Specific binding was distinguished from non-specific binding to cells and plastic wells by dilution with excess (100 μM) unlabelled drug. To examine competitive binding between DXR and AZD, cells were exposed to equimolar concentrations of both compounds (unlabelled DXR and labelled AZD) and the specific binding of AZD determined as described above. The specific activity of the UWOV2 cell surfaces/receptors (B_max) and the binding affinity constant (K_d) for DXR and AZD were determined by non-linear regression (one-site binding hyperbola) of specific binding data, using GraphPad Prism http://www.graphpad.com.

The total cellular protein content was estimated by utilizing the Bio-Rad protein dye-binding assay kit.

Statistical analysis was performed on the variables in this study using either the Student's two-tailed t-test or one-way ANOVA. The level of significance was set at p < 0.05. Values are representative of the means ± SEM of 3 experiments (n = 8 for each experiment), unless indicated otherwise. Actual p values are presented. An experimental design for comparison of a single anticancer drug dose-response relation with that of the same anticancer drug in combination with a fixed concentration of 5 μg/ml TM was used and the measured responses compared to both the Loewe additivity and Bliss independence reference models of synergy 737980 using the CombiTool computer programme (version 2.001, IMB Jena Biocomputing Group, http://www.imb-jena.de). The drug interaction index (I_x) was calculated according to the method of Chou and Talalay 81 using CombiTool. I_x values are geometric means ± SEM (standard error of the mean) of multiple effect levels (fraction affected, F_a) 0.1, 0.2, 0.3,..., 0.99 (i.e., EC₁₀, EC₂₀, EC₃₀,..., EC₉₉) for each drug in the dose range 10^-4 to 10¹ μg/ml in the presence of a fixed concentration of 5 μg/ml TM: I_x < 1 ⇒ synergy; I_x = 1 ⇒ additivity and I_x>1 ⇒ antagonism. The Loewe dose additivity model is defined by the equation dx/Dx + dy/Dy = 1, where Dx and Dy represent the concentrations of individual drugs required to exert the same effect as concentrations dx and dy used in combination. In this case, dx would be the effective inhibitory concentration (EC) of the drug used in combination with TM and Dx the EC of the drug alone. Likewise, dy would be the fixed concentration of TM used in combination with the drug and Dy the EC of TM (determined from the individual dose-response relation for TM) to produce the same effect. Student t-tests were performed to evaluate significant differences in I_x compared to a null hypothesized I_x value of 1. The Bliss independence model is defined by the equation: E_xy = E_x + E_y - E_xE_y for 0 < E < 1 and where E_xy is the additive effect of drugs x and y as predicted by their individual effects E_x and E_y. In this case, E_xy would be the effect (fractional survival) of the drug used in combination with TM, and E_x and E_y the fractional survival of cells exposed to the drug alone and to the drug in combination with TM, respectively. EC₅₀ data are best-fit values obtained from non-linear regression analysis of the sigmoidal dose-response relation for each drug alone or in combination with TM. The potency ratio and associated 95% CI (confidence interval) were computed according to the method of Fieller 82 by subtracting the log EC₅₀ of drug in combination with TM from the log EC₅₀ of drug alone and back-transformation (antilogarithm) of data. GraphPad QuickCalcs, Graphpad Prism (Version 4.03, GraphPad Software, San Diego California USA, http://www.graphpad.com), SigmaPlot (version 9.01) and SigmaStat (Version 3.11) (Systat Software, Inc. 501 Canal Blvd, Suite E, Point Richmond, CA 94804-2028, USA, http://www.systat.com) were used for data and graphic analysis.

The inhibitory effects of TM on UWOV2 cells were measured by following the incorporation of radiolabelled precursors into proteins and glycoproteins. In the presence of TM, a marked inhibition of protein (Figure 1A) and glycoprotein (Figure 1B) synthesis was observed in UWOV2 cells. The incorporation of [³⁵S]methionine into cellular protein was greatly diminished at all incubation times, except at 4h (Figure 1A). The concentration-dependent effect of TM on [³H]glucosamine incorporation by UWOV2 cells is shown in Figure 1B. At concentrations less than 0.05 μg/ml, TM had no inhibitory activity, but at higher concentrations (0.5–50 μg/ml) the antibiotic significantly impaired glycoprotein synthesis compared to control. The inhibitory effect of TM was verified by similar analysis of the incorporation of [³H]mannose into glycoprotein in the absence (control) or presence of 5 μg/ml TM (TM-treated) at different time intervals following an initial 16h pre-incubation with the antibiotic (Figure 2).

The effects of TM on the viability of UWOV2 cells are depicted in Figure 3. TM did not affect cell viability in the concentration range 0.0001–5 μg/ml and survival was consistently greater than 95% or similar to control (i.e., cells not treated with TM). The best-fit estimate of the EC₅₀ for TM in UWOV2 cells was 23.6 μg/ml (95% CI: 10.34 to 53.96), compared to a reference ovarian cancer cell line BG-1 (EC₅₀, TM = 16.81 μg/ml; 95% CI: 5.87 to 48.17) and its adriamycin (DXR)-resistant variant, BG-1/ADR (EC₅₀, TM = 64.84 μg/ml; 95% CI: 18.98 to 221) (survival curves of TM for BG-1 and BG-1/ADR are not shown). Dose-response curves for UWOV2 cells treated with different drugs alone or in combination with 5 μg/ml TM are presented in Figure 3 and the results obtained from non-linear regression of the sigmoidal dose-response relation for each drug alone or in combination with TM are summarized in Table 1. The cells displayed a high degree of resistance to VCR (EC₅₀ = 23.2 μg/ml; 95% CI: 9.11 to 59.07) relative to the other drugs: 5.86-fold > DXR (95% CI: 2.46 to 14), 8-fold > EXR (95% CI: 1.86 to 35.17) and 5.67-fold > CDDP (95% CI: 2.43 to 13.35). One-way ANOVA revealed no major variations in the log EC₅₀ of DXR and EXR (p = 0.711), DXR and CDDP (p = 0.971) as well as EXR and CDDP (p = 0.684), but significant differences between the log EC₅₀ of DXR and VCR (p = 0.041), EXR and VCR (p = 0.016), and CDDP and VCR (p = 0.044). TM potentiated the cytotoxicity of all the drugs by effectively decreasing their EC₅₀ and, correspondingly, increasing the potency ratio for each drug in the following order of magnitude: EXR (102-fold) > DXR (88-fold) > CDDP (17-fold) > VCR (5-fold) (Table 1). Unpaired t-tests for the differences in log EC₅₀ for drug alone and drug in combination with TM yielded p values less than 0.001 for DXR, EXR and CDDP, but not for VCR (p = 0.335). Thus, although TM increased the potency ratio for VCR in UWOV2 cells, the enhanced toxicity was not statistically significant.

The influence of TM on the efficacy of the anticancer drugs under study was further appraised in respect to the differences of the measured responses and the expected values generated from the CombiTool programme for the Loewe additivity and Bliss independence models for drug interaction (Figure 4). Loewe and Bliss antagonism were observed at the lower range (lower effect levels) of drug concentrations (0.0001 to 0.003 μg/ml) for DXR and CDDP and at 0.0001 for VCR (not shown), whereas all the drugs in combination with TM displayed Loewe and Bliss synergism at higher effect levels (Figures 4A and 4B, respectively). The synergistic action between TM and the different anticancer drugs was confirmed by the generalized isobolar median-effect method, using the CombiTool computer programme. The results, presented in Figures 4C], show that synergism occurs at effect levels (fraction affected, F_a) of 0.2 and above. The combination index (I_x) for each drug at multiple effect levels was also compared with a null hypothesized value of 1 and yielded p values < 0.001 in all cases, thus indicating a high degree of synergism, according to the Loewe additivity model, for all the TM-drug combinations (Table 1).

Based on the relatively high degree of resistance of UWOV2 cells to VCR and the synergistic action of TM on its potency, an investigation into the direct effects which TM might exert on the transport of VCR in these cells was undertaken. This, along with the assumption that VCR transport in these cells is coupled to the expression of Pgp, therefore provided an approach that could reflect a general association between inhibition of protein and glycoprotein synthesis and the drug transport mechanism. The effects of TM on the transport of [G-³H]VCR into (uptake) and out of (efflux) UWOV2 cells are shown in Figure 5. TM had no appreciable effect on the saturable uptake of VCR by these cells Figure 5A], but VCR efflux was significantly and consistently inhibited at post-incubation times 60 min (p < 0.001), 120 min (p = 0.004) and 180 min (p = 0.004) (Figure 5B). Data obtained for separate experiments which emulated efflux studies, on the retention of VCR in response to TM treatment, are summarized in Figure 6. TM induced a consistent 20% increase in the fractional retention of VCR in UWOV2 cells at post-incubation times 60 min (p = 0.004), 80 min (p = 0.029) and 100 min (p = 0.027). The results show that by lowering the efflux rate of VCR from the cells, probably via inhibition of Pgp synthesis and function, TM concomitantly raises the amount of drug retained in the cells. Such increased retention of VCR brought about by TM may well explain the observed increased efficacy or potency of VCR and the other drugs when combined with TM, especially in view of the molecular components targeted in the experimental setup. Therefore, overall inhibition of both protein and glycoprotein synthesis as well as drug efflux which possibly involve Pgp in the interconnected system may account for the observed enhancing effects of TM on drug cytotoxicity in UWOV2 cells.

The MDR status of UWOV2 ovarian carcinoma cells was demonstrated, in part, by measuring the specific binding of [³H]AZD, a photoactive dihydropyridine calcium channel blocker known to bind to Pgp, and [¹⁴C]DXR to intact cells in culture at 4°C in the absence (control) or presence of 5 μg/ml TM (TM-treated) following an initial 16h pre-incubation at 37°C with or without the antibiotic. The binding of DXR to the cells was saturable in the linear concentration range of 0–80 nM (Figure 7A). The measured binding affinity for DXR in control cultures was K_d = 28.48 ± 14.94 nM (95% CI: 0–59.5) and the receptor/cell surface specific activity was B_max = 0.61 ± 0.14 pmol/mg (95 % CI: 0.33 to 0.89). In the presence of TM, the B_max for DXR was increased 2.87-fold (95 % CI: 1.05 to 7.33, p = 0.045) as determined by Fieller's ratio of means test. The unpaired t-test for the difference in K_d for control and that for TM-treated cells (K_d = 60.82 ± 26.66 nM (95 % CI: 5.52 to 116) indicated no significant change in this parameter for DXR (p = 0.331). Taken together, the results show that the increased binding of DXR to UWOV2 cells in the presence of TM is probably caused by a mechanism which facilitates DXR attachment to the cells. This supposition was corroborated by an analysis of the results in terms of fractional occupancy of binding sites as predicted by the law of mass action at equilibrium (saturation), i.e., a plot of fractional occupancy ([DXR]/([DXR] + K_d) vs [DXR]/K_d yielded a rectangular hyperbola (not shown) similar to Figure 7A. The plot revealed that maximal occupancy of DXR binding sites was attained at much lower concentrations of DXR when cells had been treated with TM. Although the precise molecular mechanism has yet to be established, it is apparent that TM pre-treatment facilitates DXR binding on UWOV2 cell surfaces. The binding of the photoaffinity label of Pgp, [³H]AZD, to intact cells (control) and cells treated with TM was studied as described for DXR. Binding of AZD remained linear (i.e., non-saturation binding) between 0 and 80 nM (Figure 7B). TM treatment also significantly increased the binding of AZD to cells at all the concentrations studied (p = 0.008). In the presence of equimolar concentrations (range 30–80 nM) of unlabelled DXR, the binding of labelled AZD to UWOV2 cells was significantly reduced (Figure 7B, p < 0.005), suggesting interaction of DXR with Pgp. This reduction of AZD binding to UWOV2 cells in the presence of DXR was attenuated by treatment of cells with TM (Figure 7B, p < 0.005). Non-linear regression analysis of AZD binding did not converge and could not be performed as the incubation was in the linear region, i.e., saturation (equilibrium) was not reached. This could well have been accomplished in a study design with log increments of AZD concentration, but was not considered crucial for the purpose of this investigation.