Ritonavir reverses resistance to docetaxel and cabazitaxel in prostate cancer cells with acquired resistance to docetaxel

Materials and reagents

Docetaxel, cabazitaxel, ritonavir, and elacridar were obtained from Selleckchem (Houston, TX, USA), dissolved in DMSO, and stored at -20 °C. From stock solution, target concentrations were made in DMSO. These were then diluted 1:22 in cell culture medium, and of these, 10 µL were taken into 140 µL final cell culture volume. The final DMSO concentration was 0.3% v/v in all experiments. Bacto Agar was from BD Biosciences, CellTiter-Blue® cell viability assay from Promega. Heat-inactivated fetal calf serum (FCS) was from Sigma-Aldrich, gentamicin from Life Technologies, Iscove’s Modified Dulbecco’s Medium (IMDM) and phosphate buffered saline (PBS) Dulbecco’s w/o calcium, magnesium, w/o sodium bicarbonate from Life Technologies, and RPMI 1640-Medium with 25 mM HEPES buffer, with L-glutamine from Biochrom, all from local subsidiaries. Human prostate cancer cell lines were obtained from standard sources: 22Rv1 was from American Type Culture Collection (ATCC, Rockville, MD, USA), whereas DU-145 and PC-3 were obtained from National Cancer Institute (NCI, Bethesda, MD, USA). Protease inhibitors for Western blotting Na₃VO₄, NaF, PMSF, and benzonase were obtained from Sigma-Aldrich (Taufkirchen, Germany). The 22Rv1 prostate cancer cell line was obtained from the CWR22R xenograft model. It is an androgen receptor (AR)-positive prostate cancer model. The PC3 cell line, of a low differentiated cell type, was isolated from human prostate cancer bone metastases. It is an androgen-independent prostate cancer cell known to have moderate metastatic potential. The DU-145 cell line was the first prostate cancer cell line and was derived from a brain metastatic prostate tumor. The DU-145 cells do not show AR protein expression.

Cell culture and generation of docetaxel-resistant prostate cancer sublines

Cells were grown at 37 °C in RPMI 1640 medium supplemented with 10 % (v/v) fetal calf serum and 50 µg/mL gentamicin and passaged twice weekly. The percentage of viable cells was determined by the CASY TT cell counter (OMNI Life Science). DU-145 and 22Rv1 docetaxel-resistant clones were generated by exposure to 2 nM docetaxel and stepwise increase to 4, 6, 8, and finally 10 nM docetaxel. PC-3 cells were exposed to 0.5, 1, 2, and finally 3 nM docetaxel. In brief, after 4 weeks of incubation with docetaxel, cytotoxicity to docetaxel was determined again when sufficient cells were available. Concentrations of docetaxel were increased when the growth of treated cells in the presence of docetaxel was comparable to parental cells. Then, after another 4 weeks, the procedure was repeated. After about 3 months, no further increase in resistance to docetaxel was measured and the cells were stored as aliquots in liquid N₂. Cells of one aliquot were cultured to have enough sufficient cells (approx. 1-2 weeks). Cells were seeded in the assay plates 24 h before treatment.

Determination of cytotoxicity

Cytotoxicity was determined in two models: a 2D cell proliferation assay with cells adhering to the tissue plate and a 3D clonogenic assay with cells growing non-attached in soft agar. The viability of cells in the 2D assay was analyzed by using the CellTiter-Blue® Cell Viability Assay. The assay is based on the ability of living cells to convert a redox dye (resazurin) into a fluorescent end-product (resorufin)^[34]. Briefly, cells were harvested from exponential phase cultures, counted, and plated in 96-well flat-bottom microtiter plates at a cell density of 4,000-20,000 cells/well in a volume of 140 µL per well. After a 24-hour recovery period, 10 μL cell culture medium with or without test compound was added. After 72 h of total incubation, 20 µL/well CellTiter-Blue® reagent was added. After up to four hours, fluorescence (FU) was measured by using an Enspire Multimode Plate Reader. The 3D-assy (clonogenic assay) was carried out in a 96-well plate format using ultralow attachment plates. For each test, cells were prepared as described above and assay plates were prepared as follows: each test well contained a layer of semi-solid medium with tumor cells (50 µL), and 24 h later, a second layer of medium supernatant with or without test compound (100 µL) was added. The cell layer consisted of 2,000 to 7,500 tumor cells per well, which were seeded in 50 µL/well cell culture medium [IMDM, supplemented with 20% (v/v) FCS, 50 µg/mL gentamicin, and 0.4% (w/v) agar]. After 24 h, the soft-agar layer was covered with 100 µL medium without agar, consisting of 90 µL of the same culture and 10 µL of control medium or test compound pre-diluted from DMSO in cell culture medium (see above), and left on the cells for the duration of the experiment (continuous exposure, 100 µL drug overlay). Every plate included six untreated control wells and drug-treated groups. Duplicate values were obtained by testing single values per condition on one plate. Cultures were incubated at 37 °C and 7.5% CO₂ in a humidified atmosphere for 13 days and monitored closely for colony growth using an inverted microscope. Within this period, tumor growth led to the formation of colonies with a diameter of > 50 µm (area > 2,000 µm²). At the time of maximum colony formation, vital colonies were stained for 48 h with a sterile aqueous solution of 2-(4-iodophenyl)-3-(4-nitrophenyl)-5-phenyltetrazolium chloride (INT, 1 mg/mL, 25 µL/well), and colony counts were performed with an automatic image analysis system (CellInsight NXT, Thermo Scientific or Bioreader 5000 V-alpha, BIO-SYS GmbH). In general, results of the 3D clonogenic assay are found to be highly predictive for in vivo efficacy of test compounds, including tubulin targeting agents^[35].

Assessment of drug response

Sigmoidal concentration-response curves were fitted through the data points obtained for each tumor model using a 4-parameter non-linear curve fit (Charles River Discovery Research Services Germany Data-Warehouse Software). IC₅₀ values are reported as relative IC₅₀ values. The relative IC₅₀ value is the concentration of the test compound that gives a response (inhibition of colony formation) halfway between the top and bottom plateau of the sigmoidal concentration-response curve (inflection point of the curve). Drug combinations were evaluated using Bliss independence analysis. Measured treated vs. control (T/C) values of the combination are shown for each pair of concentrations of the compounds (modeled T/C). The expected effect (= Bliss neutral value) for simple additivity (E12) of a combination was computed by multiplication of the measured effect of compound 1 alone (E1) and effect of compound 2 (E2) alone, with E1 and E2 being fraction unaffected like percentage of viable cells (T/C values) in an inhibition assay as it was used in the present study:
Bliss neutral value: E12 = E1 * E2
E12 = T/C [compound 1] * T/C [compound 2]

The difference between the Bliss neutral value and the measured T/C value (= modeled T/C) for the combination was taken:
Bliss Index: E12 (Bliss neutral) - T/C [combination]

The results of drug combination tests were presented in heat maps. In the Modeled T/C heat maps, the efficacy of the different combination conditions is shown as the modeled T/C value, which is the mean of the measured T/C values of the individual conditions on a scale ranging from 0 to 1.0, where 1.0 corresponds to a T/C value of 100%. In the Bliss- Modeled T/C, the Bliss index is shown as the difference between the expected T/C value (Bliss neutral) and the measured T/C value (modeled T/C) on a scale ranging from -1.0 to 1.0. Positive values (Bliss Index ≥ 0.15, blue) indicate synergy, negative values (Bliss Index ≤ -0.15, red) indicate antagonism, and zero is the neutral value (white). If the observed effect is consistent across several test concentrations, an additive/synergistic/ antagonistic effect of the combination can be assumed.

RNA-sequencing

DU-145, DU-145DOC10, 22Rv1, and 22Rv1DOC8 were grown in triplicate and RNA was isolated. Paired-end RNA-sequencing data were generated using the Illumina NovaSeq 6000 with between 60 and 75 k reads per sample. 58,243 targets were tested, of which 16,387 features with a read count equal to zero across all samples were excluded from analysis. Raw expression data were evaluated using four automated outlier tests i.e., sum of Euclidean distance, Hoeffding’s D-statistic, mean Pearson correlation, and the Kolmogorov-Smirnov test statistic. Normalization was carried out using trimmed mean normalization and data were transformed with voom^[36]. The data were also manually inspected with the density plots, principal component plots, correlation heat map, and principal component analysis association plots. No samples were identified as outliers based on the results of these tests. Two statistical contrasts were performed to determine significantly differentially expressed genes (DEGs) between the docetaxel-resistant and respective parental samples. The difference in cell viability was calculated from independent duplicate/triplicate samples using a Students’ t-test. Values P < 0.05 were considered statistically significant. Significant differentially expressed genes (DEGs) were identified as those that surpassed a statistical threshold corrected for multiple testing using the false discovery rate (FDR) adjustment < 0.05 and a ≥ 2-fold change in expression value.

Western blotting

Cells were grown in cell culture flasks, washed with PBS twice, and then detached by incubation in PBS with 5% EDTA at 37 °C for 5 min. Cells were washed with PBS, centrifuged down at 300 g for 5 min, transferred into 1.5 mL Eppendorf tubes, and centrifuged again. The resulting pellet was snap-frozen in liquid nitrogen. On the day of lysis, the pellet was re-suspended in 300 µL of pre-cooled OT-lysis buffer (20 mM Tris pH 7.4 with 100 mM NaCl, 1 mm EDTA, 1 mM EGTA, 1% NP-40, 0.5% deoxycholate, 10 mM NaP₂O₇ and with the protease inhibitors Na₃VO₄, NaF, PMSF, and benzonase, all from Sigma-Aldrich, Taufkirchen, Germany). The cells were lysed by incubation on ice for 15 min. After 5, 10, and 15 min, the tube was pulse-vortexed for 3 s. Samples were then centrifuged at 16,000 g, 4 °C for 20 min. The supernatants of about 200 µL were transferred into new tubes and an aliquot was taken for protein concentration determination. Samples were snap frozen until further use. Protein concentrations were measured using the DC Protein Assay Kit II (BioRad, #500-0112). The samples were thawed on the day of western blotting and their concentration was adjusted to 2.5 mg/mL. Lysates were denatured using 2× Laemmli Sample Buffer (Biorad, #1610737) with TCEP (Sigma, #646547) as the reducing agent. The mixture was incubated for 5 min at 95 °C, while shaking at 600 rpm in a Thermomixer comfort (Eppendorf, #5360 000.011). The denatured lysates were spun down in a tabletop centrifuge and 15 µL lysate was loaded into a Criterion TGX precast gradient gel (Biorad, #5671095). Electrophoresis was performed for 50 min at 180 V, using a peqPower300 power supply (PeqLab, #55-E300-230V) in a Criterion Cell chamber (Biorad, #165-6001). The transfer was done using a Trans-Blot Turbo Transfer System (Biorad, #170-4155) and a ready-to-use PVDF membrane (Biorad, #170-4157) for 10 min at 2.5 A. The membrane was blocked in 5% (w/v) milk (Roth, #T145.3) in TBS/T (Sigma, #T5912 diluted 10-fold and Sigma, #P1379, diluted 1000-fold). The primary rabbit monoclonal IgG antibodies were diluted in TBS/T with 5% (w/v) BSA (Sigma, #A2153) and incubated for MDR1 shaking at 4 °C overnight and for β-actin 1hr at RT. The secondary goat anti-rabbit IgG antibody was diluted in TBS/T with 5% (w/v) milk and shaken with the membrane for 1 h at RT. Detection was performed using the Amersham ECL Prime reagent (GE Healthcare, #28980926) in an Image Quant 4000 (Cytiva, #28955810).

Caco-2 substrate assessment of docetaxel and bidirectional permeability assays

Bidirectional transport of docetaxel was determined through Caco-2 cell monolayers. Assay buffer containing docetaxel at 5 and 10 μM with and without reference inhibitor (10 μM valspodar or 50 μM ritonavir) was added to the appropriate apical or basolateral chamber. Samples were taken from the helper plate before the incubation and from the donor chambers after the incubation to determine the initial concentration (C0) and recovery (R) of docetaxel and the controls. After incubation, the receiver chambers were sampled to determine the amount of translocated docetaxel and controls. Samples were analyzed using LC-MS/MS. E3S (0.2 μM) for BCRP and digoxin (5 μM) for MDR1 were used as positive control substrates, while Ko143 (5 μM) for BCRP and valspodar (10 μM) for MDR1 were applied as reference inhibitors.

Bidirectional permeability assays were performed using Caco-2 (C2Bbe1 clone, ATCC-2102) cell monolayers. The cells were cultured in Dulbecco’s Modified Eagle’s Medium with 4.5 g/L glucose (DMEM) supplemented with 10% (v/v) fetal bovine serum (FBS) at 37 °C in a humidified atmosphere containing 5% CO₂ in cell culture flasks prior to seeding into collagen-coated 24-transwell inserts. Cells were cultured on the inserts with 400 μL medium per well on the apical side and 25 mL in a single-well receiver tray for all 24 wells on the basolateral side for 20-23 days. The medium was changed 24 h before the experiment. In all experiments, cells were pre-incubated in assay buffer for 15 min to allow cells to adjust to the assay conditions. Permeability incubations were carried out in Hank’s Balanced Salt Solution (HBSS) at 37 °C. The apical-to-basolateral permeability of Lucifer Yellow (LY) was assessed as a low permeability control. LY was also incubated in the presence of 5 and 10 μM docetaxel to assess the effect of the compound on the monolayer integrity. LY samples were analyzed by measuring fluorescence with the following wavelengths: excitation - 430 nm, emission - 520 nm. Estrone-3-sulfate (E3S) and digoxin efflux ratios (ERs) were determined as positive controls for BCRP and MDR1 function, respectively. The transepithelial electric resistance (TEER) of each well was measured to confirm the confluency of the monolayer after the experiments. Bidirectional transport of docetaxel (5 and 10 μM) was determined in Caco-2 cell monolayers in the absence and presence of the MDR1 reference inhibitor valspodar (10 μM) and ritonavir (50 μM). Cells were pre-incubated in HBSS for 15 min to allow cells to adjust to the assay buffer. HBSS containing docetaxel at 5 and 10 μM in the absence and presence of valspodar (10 μM) or ritonavir (50 μM) was then added to the appropriate apical (400 μL) or basolateral chambers (800 μL). After a 60-minute incubation at 37 °C, aliquots (100 μL) were taken from the receiver chambers to determine the amount of translocated docetaxel and controls. Samples were taken from the donor chambers before and after a 120-minute incubation to determine the initial concentrations (C0) and recovery (R) of the test compounds and the controls. The cellular compartment was also sampled for a more complete recovery of docetaxel. Bidirectional transport of docetaxel was determined by Liquid chromatography-Mass Spectrometry (LC-MS). The bioanalysis of docetaxel was performed with the . A minimum of six calibration concentration levels was used for fitting the calibration curve.

The following equation was used to calculate apparent permeability coefficient (Papp):

dQ: amount of transporter test drug in pmol;
dT: incubation time in seconds;
A: surface of porous membrane in cm² (standard: 0.7);
C₀: initial concentration of test agent in the donor chamber in pmol/cm³.

The efflux ratio (ER) is given as:

P_app(A-B): apparent permeability in the apical to basolateral direction;
P_app(B-A): apparent permeability in the basolateral to apical direction.

Recovery (R) was calculated according to the following formula to allow for estimation of metabolism and/or non-specific binding:

Q_apical: quantity of TA detected on apical side in pmol;
Q_basolateral: quantity of TA detected on basolateral side in pmol;
Q_cells: quantity of TA in the cellular compartment in pmol;
Q₀: quantity of TA detected at t₀ in pmol.

Assay parameters and treatment groups are summarized in Table 2.

Activity of docetaxel and ritonavir in a cell line panel

Before starting experiments in resistant cell lines, the activity of docetaxel, ritonavir, and the combination of both drugs was established in a panel of 15 tumor cell lines, including four prostate cancer cell lines (22Rv1, DU-145, PC-3, and PC-3M) [Table 3].

Docetaxel was tested as a single agent at 10 concentrations in the 2D proliferation assay to identify an appropriate test range for subsequent combination experiments. Ritonavir tested as a single agent showed no considerable or only low activity at the concentration ranges tested. Docetaxel inhibited tumor cell growth in a concentration-dependent manner with mean relative IC₅₀ values of 2.6 nM. For drug combination tests, the docetaxel concentrations were selected between 0.1 and 100 nM, whereas the ritonavir concentrations used for all 15 cell lines were 0.3, 0.949, 3, 9.49 and 30 µM. Docetaxel and ritonavir combinations were tested using a 5 × 5 matrix layout 2D proliferation assay and Bliss independence analysis was performed. The combination of docetaxel and ritonavir was synergistic in ovarian cancer cell lines 899 and NCI-ADR-RES as well as for the triple-negative breast cancer cell line MCF 10A. Ritonavir induced an increase in sensitivity to docetaxel in these cell lines. In all other cell lines, except the prostate cancer cell line DU-145 where a tendency to slight antagonism was observed, docetaxel and ritonavir had limited additive effects. An overview of all drug combination test results is provided in Table 4.

Generation of docetaxel-resistant clones

The DU-145 and 22Rv1 resistant sublines were generated by exposure to 2 nM docetaxel and stepwise increase to 4, 6, 8, and finally 10 nM docetaxel. PC-3 cells were exposed to 0.5, 1, 2, and finally 3 nM docetaxel. This resulted in increased resistance to docetaxel in DU-145 and 22Rv1 [Figure 1], but only imperceptibly for the PC-3 assay. Due to the lack of increased resistance for docetaxel in the PC-3 cells, these cells were not used for further experiments. Inhibitory concentrations of docetaxel in parental and docetaxel-resistant DU-145, 22Rv1, and PC-3 cells have been listed in Table 5.

Figure 1. Docetaxel concentration-effect curves in cell lines in 2D culture. The prostate cancer cell lines 22Rv1 and DU-145 were exposed to increasing concentrations of docetaxel over a period of 3 months. Clones of the parental cell lines DU-145 became docetaxel-resistant for 8 and 10 nM (DU-145DOC10). The cell line 22Rv1 became docetaxel-resistant for 4, 6, and 8 nM (22Rv1DOC8).

Characterization of parental and resistant clones

To determine the mechanism of resistance to docetaxel in the different cell lines, genome-wide screening using RNA-seq was applied. RNA expression levels of docetaxel-resistant cell lines DU-145DOC10 and 22Rv1DOC8, as well as their parent cell lines DU-145 and 22Rv1, were matched. Of each cell line, RNA was isolated from three different cultures (triplicates). 58,243 features were generated, of which 16,387 features with reads equal to zero were eliminated. After rigorous quality control (see methods), no outliers were detected, and the data were evaluated. 2,360 DEGs were identified between DU-145 DOC10 and DU-145 samples, and 1,058 DEGs were identified between 22Rv1 DOC8 and 22Rv1 [Figure 2].

Figure 2. Double volcano plot of differential gene expression between the parental human prostate cancer cell lines DU-145 and 22Rv1 and their docetaxel resistance generated clones DU-145DOC10 and 22Rv1DOC8. Red dots are differentially expressed genes between DU-145 vs. DU-145DOC10, blue for 22Rv1 vs. its resistant clone 22Rv1DOC8. Green dots are differentially expressed genes in both setups. The grey area represents not significantly regulated genes. Cut-offs of 2-fold and a FDR P < 0.05 were used. Note the extreme differentially and significantly expressed green dot for MDR1 (arrow). FDR: False detection rate; MDR1: multidrug resistance protein 1.

The genes that were most highly up-regulated in DU-145DOC10 samples compared to DU-145 were those associated with metabolism and homeostasis, including ABCB1 (3287-fold), Dermatan Sulfate Epimerase Like (DSEL, 530-fold), Glypican 4 (GPC4, 260-fold), Matrix Metall-Protease 1 (MMP1, 60-fold), Intern Subunit alpha 4 (ITGA4, 45-fold), and Glypican 6 (GPC6, 44-fold). The most down-regulated genes in DU-145DOC10 samples compared to DU-145 were genes associated with signaling pathways, such as Butyrylcholinesterase (BCHE, 6770-fold), Secreted Frizzled Related Protein (SFRP1, 3325-fold), PBX Homeboy 1 (521-fold), R-Spondin (RSOP3, 497-fold), and Glucagon Like Peptide 2 Receptor (GLP2R, 163-fold), among others. Some CYP4F-family members were down-regulated, but CYP3A4 was 1.7-fold up-regulated. In the 22Rv1DOC8 clone vs. the 22Rv1 cell line, genes associated with metabolic or homeostasis pathways were most up-regulated, including ABCB1 (2772-fold), Glutathione S Transpherase Pi1 (GSTP1, 20-fold), GATA5 (19-fold), SERPINA4 (14-fold). The most down-regulated genes in 22Rv1 DOC8 samples compared to 22RV1 were genes associated with neuronal system pathways, such as KCNJ8 (208-fold), SLITRK2 (150-fold), GABA type A receptor (GABRA2, 56-fold), and ABCC9 (30-fold), among others. Some CYP4F-family members were also down-regulated as with DU-145DOC10, but here CYP3A4 was 4-fold down-regulated. It was evident that, in both docetaxel-resistant cell lines, the gene that was up-regulated the most and with the highest significance was ABCB1, 3287-fold in DU-145 DOC10 and 2773-fold in 22Rv1DOC8.

The gene expression data for ABCB1 were then confirmed by western blotting. Both parental cell lines DU-145 and 22Rv1 expressed no detectable amount of the P-gp protein. In both docetaxel- resistant cell clones DU-145DOC10 and 22Rv1DOC8, P-gp expression increased, but was especially high in DU-145DOC10 [Figure 3].

Figure 3. Western blot for the P-gp protein in the parental human prostate cancer cell lines (DU-145 and 22Rv1) and their docetaxel resistance clones (DU-145DOC10 and 22Rv1DOC8). The size of the P-gp protein is indicated by an arrow (130-180 kDa). The lower panel shows equal load by β-actin staining (45 kDa, arrow). P-gp: P-glycoprotein.

Reversal of resistance to docetaxel

DU-145 and docetaxel-resistant DU-145DOC10 cell lines were treated with docetaxel in combination with ritonavir in an 8 × 2 matrix combination setup using the 3D tumor clonogenic assay. The baseline IC₅₀ value for docetaxel (i.e., without ritonavir) in the DU-145 cell line was established at 2.9 nM and in the resistant DU-145DOC10 at 9.5 nM; a relative resistance of 3.3. The addition of 10 µM ritonavir to docetaxel in the DU-145DOC10 cell line increased the sensitivity to docetaxel, resulting in a 3.4-fold lower IC₅₀ value of 2.8 nM, which is almost identical to the IC₅₀ value of the parental cell line. Bliss independence analysis of the DU-145 DOC10 assay showed synergistic effects at 10 μM ritonavir when combined with medium concentrations of docetaxel (3.16 and 10 nM). This confirmed the back-shift of the concentration-effect curves and indicates that the addition of 10 μM ritonavir reverses the resistance to docetaxel. The concentration-effect curve of the docetaxel/ritonavir combination, as well as the Bliss- Model T/C heat map of the DU-145DOC10 assay [Figure 4].

Figure 4. Effect of the docetaxel ritonavir combination in the docetaxel-resistant DU-145DOC10 cell line in the 3D clonogenic assay. (A) Concentration-effect curves showing a decrease of the docetaxel IC₅₀ from 9.5 to 2.8 nM in the presence of 10 μM ritonavir; (B) Bliss Modeled T/C analysis shows synergistic effects at 10 μM ritonavir when combined with medium concentrations of docetaxel (3.16 and 10 nM).

Similarly, parental 22Rv1 and docetaxel-resistant 22Rv1DOC8 cells were treated with docetaxel in combination with ritonavir in an 8 × 2 matrix combination setup using a 3D tumor clonogenic assay. Bliss independence analysis was used to assess synergistic effects. The baseline IC₅₀ value for docetaxel in the 22Rv1 and 22Rv1DOC8 cell lines was established at 1.3 and 54.6 nM, respectively, i.e., a relative resistance of 42.0. The addition of 10 µM ritonavir to docetaxel in the 22Rv1DOC8 cell line reduced the IC₅₀ value to 46.2, which is still a 36-fold higher relative resistance as compared to the IC₅₀ value of 1.3 in the parental cell line. At a concentration of 32 μM, ritonavir decreased the relative IC₅₀ value of docetaxel from 219 to 26 nM (84-fold) in the 22Rv1DOC8 cell line and strong synergistic effects were observed at mid concentrations of docetaxel (3.16 to 316 nM) as shown in shown in Figure 5.

Figure 5. Effect of the docetaxel ritonavir combination in the docetaxel-resistant 22Rv1DOC8 cell line in the 3D clonogenic assay. (A) Concentration-effect curve showing a decrease of the docetaxel IC₅₀ from 219 to 26 nM in the presence of 32 μM ritonavir; (B) Bliss Modeled T/C analysis shows synergistic effects at 32 μM ritonavir when combined with medium concentrations of docetaxel (3.16 and 10 nM).

To further investigate if ABCB1 up-regulation affects docetaxel-induced resistance, we used elacridar, a specific P-gp inhibitor, and compared the IC₅₀ values of docetaxel in parental and docetaxel-resistant cell lines. In this confirmatory experiment, the objective was to investigate the ability of elacridar to revert resistance towards docetaxel in 22Rv1DOC8 and DU-145DOC10. For this purpose, resistant and parental cell lines were treated with docetaxel in combination with elacridar using the 3D clonogenic experimental setup. Bliss independence analysis was used to assess synergistic effects. The IC₅₀ value for docetaxel in the DU-145 cell lines was established at 1.1 nM (DU-145) and 11.7 nM (DU-145-DOC10) in this experiment. For the 22Rv1 cell lines, the docetaxel IC₅₀ value was 1.3 (22Rv1) and 134 (22Rv1DOC8). In the resistant cell lines, the relative IC₅₀ value of docetaxel decreased 6- and 7.1-fold in DU-145DOC10 and 91- and 116-fold in 22Rv1DOC8 in the presence of 1 or 10 µM elacridar. The elacridar concentrations of 0.316 and 1 µM with docetaxel showed synergistic effects in the DU-145DOC10 assay in the mid concentrations of docetaxel (3.16 and 10 nM). The concentration-effect curve of the docetaxel/elacridar combination, as well as the Bliss Model T/C heat map of the DU-145DOC10 assay [Figure 6].

Figure 6. Effect of the docetaxel-elacridar combination in the docetaxel-resistant DU-145DOC10 cell line in the 3D clonogenic assay. (A) Concentration-effect curve showing a decrease of the docetaxel IC₅₀ from 11.4 to 1.9 nM combined with 1 μM elacridar and from 11.4 to 1.6 nM combined with 1 μM elacridar; (B) Bliss Modeled T/C analysis shows synergistic effects at both test concentrations of elacridar (0.316 and 1 µM) when combined with medium concentrations of docetaxel (3.16 and 10 nM).

The 0.316 and 1 µM elacridar concentrations showed strong synergistic effects with docetaxel in the 22Rv1DOC8 assay almost throughout the whole concentration range of docetaxel (1 to 316 nM). Thus, the resistance-reversing effect of ritonavir that was observed in the previous experiment combining docetaxel with ritonavir could be reproduced by elacridar in both DU-145DOC10 and 22Rv1DOC8 cell lines, strongly suggesting that inhibition of P-gp is responsible for this effect. Elacridar alone had no effect on the parental cells (data not shown). Ritonavir itself was also not toxic in the used concentrations (data not shown). Results on the reversal of resistance by ritonavir and elacridar in both cell lines have been summarized in Table 6.

Reversal of resistance to cabazitaxel

The docetaxel-resistant prostate cancer sublines were tested for sensitivity to cabazitaxel, an analog compound of docetaxel, to investigate whether ritonavir could also reverse resistance to this second taxane. DU-145 DOC10 and 22Rv1 DOC8 and their respective parental cell lines were treated with cabazitaxel in combination with ritonavir in an 8 × 2 matrix combination setup using a 3D tumor clonogenic assay. Docetaxel-resistant DU-145DOC10 and 22Rv1DOC8 sublines were also resistant to cabazitaxel. DU-145DOC10 cells were 2.5 times less sensitive to cabazitaxel than the parent DU-145 cells (IC₅₀ of 1.5 nM in DU-145DOC10 vs. an IC₅₀ of 0.6 nM in DU-145). 22Rv1DOC8 cells were 13.6 times less sensitive to cabazitaxel than the parent 22Rv1 cells (IC₅₀ of 9.5 nM in 22Rv1DOC8 vs. an IC₅₀ of 0.7 nM in 22Rv1). Similar to the observations with the docetaxel ritonavir combination, cytotoxicity to cabazitaxel was restored with 10 µM ritonavir in DU-145DOC10 (IC₅₀ of 0.6 nM and a relative resistance of 1.0) and with 32 µM ritonavir in 22Rv1DOC8 (IC₅₀ of 0.6 nM and a relative resistance of 1.0). These results have been summarized in Table 7.

In the Bliss analysis of the DU-145DOC10 assay, synergistic effects were observed of 10 µM ritonavir combined with 1 or 3.16 nM cabazitaxel. The concentration-effect curve of the cabazitaxel/ritonavir combination, as well as the Bliss Model T/C heat map of the DU-145DOC10 assay [Figure 7]. In the 22Rv1DOC8 assay, synergistic effects were observed in the combination of 32 µM ritonavir and 1 to 10 nM cabazitaxel.

Figure 7. Effect of the cabazitaxel ritonavir combination in the docetaxel-resistant DU-145DOC10 cell line in the 3D clonogenic assay. (A) Concentration-effect curve showing a decrease in the cabazitaxel IC₅₀ from 1.5 to 0.6 nM in the presence of 10 μM ritonavir; (B) Bliss Modeled T/C analysis shows synergistic effects at 10 μM ritonavir when combined with cabazitaxel concentrations of 1 and 3.16 nM.

Interaction of docetaxel with the P-gp transporter

The human Caco-2 gastrointestinal cell line has been used frequently to establish drug efflux driven by P-gp. The expression of P-gp on the apical surface of this cell line makes this cell line particularly useful^[37]. Moreover, the CYP3A4 enzyme is missing in Caco-2 cells. To further elucidate the mechanism of the reversal of docetaxel-mediated drug resistance by ritonavir, the bidirectional permeability of docetaxel was tested in Caco-2 monolayer assays at 5 and 10 µM in the absence and in the presence of the P-gp specific reference inhibitor valspodar and ritonavir applying 60- and 120-minute incubation periods confirming that docetaxel is an in vitro substrate of the P-gp transporter [Supplementary Table 1]. Specifically, the bidirectional permeability from Apical to Basolateral (A-B), from Basolateral to Apical (B-A), and ER values of docetaxel in the presence of the inhibitors valspodar and ritonavir were determined. The effect of docetaxel on the monolayer integrity was also investigated by applying low permeability control Lucifer LY in the absence and presence of the docetaxel (data not shown).

In the bidirectional Caco-2 monolayer assay, docetaxel showed active transport as the obtained Papp values in the B-A direction were higher than in the A-B direction and the ER values were > 2 both at 5 and 10 μM. The ER values were 69.5 and 43.9 at 5 μM and 31.8 and 32.7 at 10 μM of docetaxel after 60- and 120-minute incubation time, respectively [Table 8]. The P-gp reference inhibitor valspodar and ritonavir inhibited the active transport of docetaxel, resulting in ERs < 2. These show that docetaxel is transported via P-gp in the B-A direction through Caco-2 monolayers and this drug transport can be inhibited by ritonavir.