The Bcl-2/Bcl-X_L/Bcl-w Inhibitor, Navitoclax, Enhances the Activity of Chemotherapeutic Agents In Vitro and In Vivo

The Bcl-2 family comprises 2 broad categories of prosurvival (Bcl-2, Bcl-X_L, Bcl-w, Mcl-1, and A1) and proapoptotic (Bax, Bak, Bim, Bid, Puma, Bad, Noxa, Bik, Bmf, and Hrk) proteins (1–4). In general, the balance between these proteins determines whether a cell lives or dies. The overexpression of prosurvival Bcl-2 family members provides a means by which cancer cells can overcome the continual assault of cellular stresses that would cause normal cells to undergo apoptosis. Their overexpression has also been shown to correlate with poor prognosis and resistance to therapy (5). For example, overexpression of Bcl-2 has been shown to confer resistance to cytotoxic agents in lung cancer (6, 7), whereas both Bcl-2 and Bcl-X_L are associated with the emergence of androgen independence and chemoresistance in prostate cancer (8–11). Thus, directly targeting the apoptotic machinery represents an attractive strategy for enhancing the efficacy of certain chemotherapeutics (12).

Recently, we described the discovery of ABT-263 (navitoclax), an orally bioavailable Bcl-2 family protein inhibitor with high affinity (<1 nmol/L) to Bcl-2, Bcl-X_L, and Bcl-w but not to Mcl-1 or A1 (13). Navitoclax exhibits potent single-agent antitumor activity in multiple murine xenograft models of small cell lung cancer (SCLC), leukemia, and lymphoma (13, 14) and, in addition, potentiates the activity of clinically relevant chemotherapeutic regimens, for example, rituximab, rapamycin, rituximab-cyclophosphamide-Adriamycin-vincristine-prednisone (CHOP), and bortezomib, in several models of hematologic malignancies (14, 15). However, the interrogation of the ability of navitoclax to enhance the activity of chemotherapeutic agents in solid tumors has been limited (16).

To address this question, we have examined navitoclax in combination with 19 chemotherapeutic agents across a panel of 46 cell lines representing 16 tumor types. These cell lines are commercially available and often used for drug development. Navitoclax enhanced the activity of several chemotherapeutic agents, representing diverse mechanisms of action, across a broad spectrum of tumor types. To extend these observations, potent combinations were tested in murine xenograft models. Navitoclax showed activity in combination with erlotinib or docetaxel in lung or ovarian cancer models in vitro and in vivo. Subsequent mechanism of action and short interfering RNA (siRNA)-mediated knockdown experiments indicated that this efficacy was associated with the neutralization of Mcl-1. Taken together, our studies provide a mechanistic rationale for the use of navitoclax in combination with standard-of-care agents that neutralize Mcl-1 or increase proapoptotic BH3 proteins such as Bim and Noxa.

Chemotherapeutic agents were purchased from Sigma or Calbiochem or synthesized internally at Abbott Laboratories. The SCC25, SCC9, FADU, HN, 786-O, RENCA, Hep3B, HepG2, A549, Calu-6, DMS-53, H196, H460, H1299, D54MG, DoHH2, SuDHL4, SKNFI, HEYA8, OVCAR-3, OVCAR-4, OVCAR-5, OVCAR-8, SKOV3, MiaPaCa, PANC-1, BxPC3, PC3, 22RV1, LNCAP, HT-1080, SK-LMS-1, A375, A431, AGS (CRL-1739), N87, HCT116, DLD1, SW620, Colo-205, EJ-1, MDA-MB-231, MDA-MB-435, MCF-7, Mx-1, BT474, K562, and H1650 cell lines were purchased from American Type Culture Collection or Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (DSMZ). These cell lines were not reauthenticated before using in this study. They are grown in medium conditions recommended by the suppliers or in RPMI medium (Invitrogen) supplemented with 10% FBS (Invitrogen).

Five thousand cells were plated per well in a 96-well plate and incubated with each therapeutic agent with or without 1 μmol/L navitoclax for 72 hours in a total volume of 100 μL culture medium supplemented with 10% FBS. Cytotoxicity was assessed using the CellTiter-Glo Assay that measures the ATP content of the cells according to manufacturer's specifications (Promega). Agents that induced 20% or more reduction in combination with navitoclax below what was observed with agent alone across 2 doses were considered synergistic hits from the screen. To quantitatively assess the combination effects of navitoclax with other chemotherapeutic agents, experiments were carried out to determine the combination index (CI), which is a value derived from the multiple drug–effect equation of Chou–Talalay (17). CalcuSyn software from Biosoft was used for this analysis. Briefly, this method of analysis distinguishes between synergistic, additive, and antagonistic drug–drug interactions. A CI value of 1.0 indicates an additive effect. CI values less than 1.0 reflect a synergistic effect, whereas CI values greater than 1.0 reflect an antagonistic effect.

Twenty to thirty micrograms of whole-cell or fractionated lysates were separated by SDS-PAGE (4%–20% gels) and transferred to Immobilon-P membranes (Millipore). The membranes were incubated with primary antibodies against Bcl-2 (Epitomics), Bcl-X_L (Epitomics), Mcl-1 (Santa Cruz Biotechnology), Bim (Epitomics), Bid (BD Biosciences), Bax (Epitomics), Puma (Sigma), Noxa (Abcam), pEGFR (Invitrogen), EGFR (Invitrogen), pERK (Cell Signaling Technology), ERK (Cell Signaling Technology), or cytochrome c (BD Biosciences) overnight at 4°C. Primary antibodies were detected using either a goat antimouse or goat antirabbit Alexa Fluor 680–conjugated or IRDye 800CW–conjugated secondary antibody (Molecular Probes) diluted 1:10,000 for 30 minutes. Bands were detected using the Odyssey Infrared Imaging System (LI-COR Biosciences).

SKOV3 cells were pretreated with 1.5 nmol/L docetaxel or 2.5 μmol/L gemcitabine for 48 hours, washed, and resuspended in cell permeability buffer (20 mmol/L HEPES, pH 7.2, 100 mmol/L KCl, 5 mmol/L MgCl₂, 1 mmol/L EDTA, 1 mmol/L EGTA, 250 mmol/L sucrose, 0.015% digitonin, Complete Protease Inhibitor; Roche), and incubated on ice for 3 to 5 minutes. Permeabilized cells were then incubated with 10 μmol/L navitoclax, enantiomer, Bim-BH3 peptide, or vehicle at 30°C for 75 minutes. Mitochondrial and cytosolic fractions were separated by centrifugation at 16,000 × g at 4°C for 10 minutes, resolved on a 10% SDS PAGE gel, and analyzed by immunoblotting using a mouse anti-cytochrome c antibody (BD Biosciences).

A standard immunofluorescence protocol was used to prepare samples for high content analysis. SKOV3 cells were propagated in RPMI-1640 supplemented with 10% FBS. Cells were plated at a density of 2,000 cells per well (100 μL) in 96-well Collagen I–coated microplates (BD Biosciences) and incubated overnight under standard cell culture conditions. On the next day, cells were treated with docetaxel or vehicle by addition of a 50 μL solution containing a 3× compound stock and incubated for an additional 48 hours. Cells were then challenged with navitoclax for the indicated times. Cells were subsequently fixed by the addition of 100 μL of prewarmed 10% formaldehyde (Sigma) made up in PBS, pH 7.4 lacking Ca²⁺/Mg²⁺ (Invitrogen) and incubated at room temperature for 10 minutes. All subsequent reagents were made up in PBS, pH 7.4 lacking Ca²⁺/Mg²⁺. Wells were then washed twice with 200 μL PBS and permeabilized with 0.1% Triton X-100 for 10 minutes, washed with PBS, blocked with 1% bovine serum albumin for 20 minutes and then incubated with 50 μL of a 1:100 dilution of a rabbit antibody raised against activated caspase-3 (Cell Signaling Technology). Primary antibody was incubated overnight at 4°C, washed with 200 μL PBS, incubated for an additional hour with a 50 μL mixture of a 1:200 dilution of Alexa Fluor 488–conjugated goat anti-rabbit antibody and 1 mg/mL Hoechst 33342 (Invitrogen). After a 1-hour incubation in the dark, cells were washed 3× with 200 μL PBS and imaged on an ArrayScan VTI HCS Reader (Thermo Scientific) with a ×10 (0.3 numerical aperture) objective. The Cell Health Profiling BioApplication was used to quantify the percentage of cells exhibiting caspase-3 activation above baseline levels determined from cells that received vehicle only.

SKOV3 cells were plated at 1 × 10⁴ cells/100 μL in 96-well or 3 × 10⁵ cells/2 mL in 6-well tissue culture plates and transfected with siRNA duplexes targeting Mcl-1 (M-004501-02) or scrambled siRNA (D-001210-01) from Dharmacon. Briefly, siRNAs were prepared by incubating 1.5 μL of a 20 μmol/L siRNA stock with 100 μL Opti-MEM (Invitrogen) for 10 minutes and 3 μL Lipofectamine 2000 (Invitrogen) with 100 μL Opti-MEM. Each solution was then mixed and incubated for an additional 20 minutes at room temperature. siRNA mixtures were then added to the cells. siRNA knockdown was confirmed by immunoblotting using whole-cell lysates of SKOV3 treated with Mcl-1 siRNA or scrambled control siRNA. Cell viability was assessed after 72 hours using the CellTiter-Glo Assay as described above.

All animal studies were conducted according to guidelines established by the Abbott and Genentech Institutional Animal Care and Use Committee. Xenograft studies were conducted by implantation of cells or tumor fragments in the flanks of immunocompromised mice. Tumor-bearing mice were size matched (∼235 mm³; day 0) into treatment and control groups, and tumor size was monitored serially in individual animals using electronic calipers. Tumor volume was calculated 2 to 3 times weekly using the following equation: volume = length × width²/2. Percentages of tumor growth inhibition (TGI), tumor growth delay (TGD), partial regression (PR), complete regression (CR), and overall response rates (ORR) are defined as previously described (14). All studies used 6 to 10 mice per group. Statistical comparisons of tumor growth rate and TGD used the Wilcoxon rank-sum test and the Mantel–Cox log-rank test, respectively. Navitoclax was formulated in 10% ethanol, 30% polyethylene glycol 400, and 60% Phosal 50 PG (a dispersion of 50% phosphatidylcholine in a propylene glycol/ethanol carrier) and administered orally by gavage. Docetaxel was purchased from Sanofi Aventis, and gemcitabine was purchased from Eli Lilly. These agents were dosed intravenously and intraperitoneally, respectively, at the schedule and dose described in the table descriptions. Erlotinib was formulated in 7.5% Captisol (sulfobutylether-β-cyclodextran) and dosed daily by oral gavage.

The single-agent cellular activity of navitoclax has only been shown in SCLC and hematologic malignancies (13, 18). To evaluate the breadth of cellular activity in combination with other chemotherapeutic agents, 19 therapeutic agents were tested in the presence or absence of 1 μmol/L navitoclax against 46 human cancer cell lines representing 16 tumor types (Fig. 1A). Using a 6-point dose–response curve, positive combination activity was defined as 20% or greater reduction in survival across 2 or more doses in the presence versus absence of navitoclax. Notably, navitoclax enhanced the activity of multiple therapeutic agents broadly across multiple tumor types, with 196 combination hits identified of 774 examined (∼25% hit rate). Combination synergy, defined as a response that is greater than the expected additive effect, was evaluated using CI analysis according to the multiple drug–effect equation of Chou–Talalay (17). To select the best positive combination activities for further investigation in vivo, CI analyses were conducted across different cell line models to determine the CI_50–90 for these combinations. Navitoclax displayed synergy with DNA-damaging agents (e.g., gemcitabine, etoposide, camptothecin, and doxorubicin) and antimitotic agents (e.g., docetaxel and vincristine; Fig. 1B) in a variety of tumor types in vitro, as confirmed by CI analyses (Table 1).

As shown in Fig. 1B, docetaxel exhibited positive combination activities with navitoclax in a high percentage of cancer cell lines (78%). Docetaxel represents a clinically relevant anti-microtubule agent approved for use in a variety of tumors (19). To extend these observations to in vivo, navitoclax was tested in combination with docetaxel in the SKOV3 ovarian cancer xenograft model using a variety of different schedules (Fig. 2). As a monotherapy, navitoclax dosed orally once a day at 100 mg/kg for 2, 14, or 21 days was not efficacious in the SKOV3 xenograft model (Fig. 2). Docetaxel administered once weekly for 3 cycles at 10 mg/kg/d produced a significant decrease in tumor burden (TGI) of 82% with a delay in regrowth (TGD) of 114% (Table 2). In contrast, bolus dosing of docetaxel dosed 30 mg/kg once intravenously inhibited tumor growth by only 48% with no significant effect on time to progression. ORRs were slightly higher when cyclical dosing was administered, rather than bolus dosing (30% vs. 0%).

Navitoclax enhances the activity of docetaxel under most schedules. Chronic (14 or 21 days) daily administration of navitoclax concurrently with either docetaxel regimen yielded greater than additive inhibition of tumor growth, although this only reached statistical significance in the bolus dosing group (Fig. 2A and B; Table 2). Effects on delay of tumor regrowth and overall response with concurrent, chronically dosed navitoclax were significantly improved in both docetaxel regimens compared with docetaxel alone. TGD, a measurement of durability of antitumor effect, was also significantly improved from 12% to 135% and from 114% to more than 175% in the bolus dosing and cyclical dosing, respectively. ORRs were also significantly improved from 0% to 89% and 30% to 100% in the bolus dosing and cyclical dosing, respectively. Concurrent, acute dosing of navitoclax (2 days) with bolus administration of docetaxel also yielded greater than additive results, with TGI from 57% to 86% and TGD from 12% to 54% (Fig. 2C; Table 2). Sequential dosing of docetaxel and navitoclax was also investigated. Administration of docetaxel on day 1, followed by 14 days of navitoclax therapy, induced a significant inhibition in tumor growth and a greater than additive delay in regrowth, similar to concurrent, acute administration of navitoclax and docetaxel (Fig. 2D). In contrast, acute navitoclax treatment following a bolus dose of docetaxel yielded only modest increases in TGI and TGD, as did administration of a bolus dose of docetaxel following 14 days of navitoclax (Fig. 2E and F).

To gain a better mechanistic understanding of how docetaxel synergizes with navitoclax, we examined the effect of single-agent and combination treatments on apoptotic signaling in SKOV3 in vitro. Neither docetaxel nor navitoclax elicited cytochrome c release in permeabilized SKOV3 cells (Fig. 3A). In contrast, 48 hours of pretreatment with 1.5 nmol/L docetaxel followed by 75 minutes of navitoclax treatment led to cytochrome c release similar to that observed by treatment with Bim BH3 peptide. The enantiomer of navitoclax with much lower target affinity is unable to elicit this effect. To examine effects on apoptosis in intact cells, high content analysis was conducted. As shown in Fig. 3B, no increase in activated caspase-3 is detected in cells treated with navitoclax alone. However, pretreatment of SKOV3 cells with 1.5 nmol/L docetaxel and subsequent incubation with navitoclax induced a concentration-dependent increase in activated caspase-3 within 2 hours. Taken together, these data indicate that docetaxel acts synergistically with navitoclax by increasing mitochondrial sensitivity to Bcl-2 and/or Bcl-X_L inhibition.

To determine which events might be responsible for the combination synergy observed with navitoclax, the expression levels of several key Bcl-2 family proteins were examined in response to docetaxel treatment. Treatment of SKOV3 cells with docetaxel resulted in a decrease in Mcl-1 levels (Fig. 3C). No other significant changes were observed with the exception of a modest increase in BH3-only protein Noxa levels at 12 nmol/L docetaxel. Because Mcl-1 is a well-known resistance factor for ABT-737 (20, 21), we next evaluated whether targeted knockdown of Mcl-1 would enhance navitoclax activity in these cells. While knockdown of Mcl-1 on its own had no effect on cell viability (Fig. 3D), it enhanced the cytotoxicity of 1 and 5 μmol/L navitoclax by 35% and 25%, respectively. The P values are less than 0.0002 and 0.002 for 1 and 5 μmol/L navitoclax treatment, respectively. These data suggest that the combination synergy observed between docetaxel with navitoclax is due, at least in part, to the neutralization of Mcl-1.

In addition to the broad synergy observed with cytotoxic agents, navitoclax also enhanced the activity of erlotinib, an epidermal growth factor receptor (EGFR)-specific kinase inhibitor. Positive combination activity between navitoclax and erlotinib was observed in 5 of the 46 lines examined (5 of 774 combinations examined; Fig. 4A). Interestingly, in all 5 of these lines (A431, BxPC3, N87, OVCAR-3, and SKOV3), EGFR has been reported either amplified (A431) or overexpressed (22–25). Clinically, tumors with activation mutations, such as the EGFR deletion 746 to 750, have been most responsive to erlotinib therapy (26). We next investigated the combination in a non-SCLC (NSCLC) line (NCI-H1650) with this genetic alteration to assess the ability of navitoclax to enhance apoptotic potential and antitumor efficacy of erlotinib. CI analysis of the navitoclax–erlotinib combination in NCI-H1650 cells confirmed synergistic activity with an average CI of 0.44 across the EC₅₀ to EC₉₀ (CI_50–90; Fig. 4B). Examination of Bcl-2 family proteins indicated the coordinated downmodulation of Mcl-1 and upregulation of Bim_EL after 6 hours of treatment with erlotinib in vitro. These data are consistent with previous findings showing synergy between erlotinib (or gefitinib) and ABT-737, an earlier generation Bcl-2 family protein inhibitor (27, 28).

To confirm the synergistic effects of navitoclax and erlotinib on tumor cell proliferation observed in vitro in the NCI-H1650 model, we tested the combination of these 2 targeted agents in vivo. As a single agent, 100 mg/kg navitoclax alone dosed daily had no significant antitumor activity, whereas daily dosing of erlotinib at 50 mg/kg resulted in significant tumor stasis (%TGI = 52) during a 21-day treatment period (Fig. 4C). Notably, the combination of navitoclax and erlotinib dosed daily for 21 consecutive days resulted in 98% TGI and durable tumor regressions in 100% of treated tumor-bearing mice (Fig. 4C). The latter was sustained for 25 days even after the cessation of dosing (Fig. 4C). In addition, the combination of navitoclax and erlotinib was well tolerated, resulting in less than 5% body weight loss during the treatment period.

To investigate potential mechanisms of action underlying the enhanced combination activity of navitoclax and erlotinib in the NCI-H1650 xenograft model in vivo, tumors were harvested posttreatment and analyzed for EGFR and Bcl-2 pathway markers. Treatment with erlotinib resulted in decreased phosphorylation levels of extracellular signal–regulated kinase (pERK) and induced significant downmodulation of Mcl-1 and upregulation of Bim (Fig. 4D), which is consistent with previous in vitro findings (Fig. 4B). Thus, the enhanced antitumor activity of erlotinib in combination with navitoclax in vivo is because of simultaneous inhibition of the EGFR and Bcl-2 prosurvival pathways.

Defects in the apoptotic cascade are a hallmark of cancer and are often associated with chemoresistance (29). Thus, agents that restore the ability of cancer cells to undergo apoptosis may enhance the activity of chemotherapies when used in combination. We recently reported the discovery of navitoclax, an orally bioavailable small-molecule Bcl-2 family protein inhibitor currently in phase I and II clinical development. Preclinically, navitoclax exhibits single-agent activity in SCLC and leukemia and lymphoma cell lines but has limited activity in the majority of solid tumor cell lines examined (13–15, 30). In combination, navitoclax has been shown to enhance the efficacy of rapamycin, rituximab, rituximab-CHOP, and bortezomib in various leukemia and lymphoma models (13, 15). In NSCLC xenograft tumor models, navitoclax enhances the efficacy of taxanes (16). However, investigation of the ability of navitoclax to enhance the efficacy of clinically relevant therapeutics in other solid tumor models has been limited. To address this, we conducted a screen of navitoclax in combination with 19 therapeutically relevant agents across 46 solid tumor cell lines to identify potential combinations for clinical interrogation.

In vitro, navitoclax enhanced the activity of several chemotherapeutic agents across multiple cell lines. Notably, the greatest synergy was observed in combination with therapeutics that act as antimitotics (docetaxel and vincristine) or induce DNA damage (camptothecin, gemcitabine, etoposide, and doxorubicin). These in vitro findings were confirmed in vivo. Navitoclax treatment in the SKOV3 xenograft model significantly enhanced the antitumor activity of the antimitotic agent docetaxel with significantly improved ORRs compared with docetaxel treatment alone. In addition, combination treatment resulted in more durable antitumor responses, as evidenced by the significant increase of %TGD. In that it has been noted that the scheduling of docetaxel dosing can have a significant effect on antitumor activity (16, 31), we investigated a variety of doses and schedules in our combination studies. Addition of navitoclax to docetaxel treatment resulted in enhanced efficacy under a variety of dose schedules. We find that navitoclax is effective when administered concurrent with or immediately after dosing of docetaxel, whereas predosing with navitoclax before docetaxel administration did not significantly enhance efficacy. The combinations of navitoclax–docetaxel and navitoclax–gemcitabine were evaluated once in N87 (gastric) and OVCAR-5 (ovarian) xenograft models (Supplementary Fig. S1). Consistent with our in vitro findings, these combinations showed enhanced antitumor activity in vivo compared with that of the chemotherapeutic treatment alone, as evidenced by a significant increase in %ORR (from 50% to 100%) for navitoclax–docetaxel in N87 and significant improvement in %TDG (from 39% to 76%) for navitoclax–gemcitabine in OVCAR-5 (Supplementary Table S1).

Mechanistically, we observed that navitoclax induces cytochrome c release and caspase-3 activation upon pretreatment with docetaxel. As such, we hypothesized that docetaxel may alter the sensitivity of the mitochondria (and hence the cell viability), in part, by neutralization of Mcl-1 via a downregulation of Mcl-1 protein levels. The role of Mcl-1 as a resistance factor to navitoclax and ABT-737 has been heavily investigated in multiple cellular systems (13, 20, 21, 32, 33). Indeed, siRNA knockdown of Mcl-1 in SKOV3 cells sensitized these cells to navitoclax, further indicating that neutralization of Mcl-1 may, in part, underlie the observed combination synergy. Our data indicate that chemotherapies that indirectly modulate Mcl-1 levels or function may exhibit in vivo combination synergy with navitoclax. If so, this criterion may be used as a rational approach in the selection of combination agents.

Applying this criterion, we observed that erlotinib treatment was capable of downmodulating Mcl-1. As such, we investigated whether navitoclax enhanced the antitumor activity of erlotinib in vivo. Consistent with our hypothesis, the combination of erlotinib with navitoclax in an NSCLC xenograft model showed durable tumor regressions and 100% ORRs in mice, whereas mice treated with erlotinib alone showed only a modest TGI during the treatment period. In addition to downmodulation of Mcl-1 in tumors isolated from mice treated with erlotinib, we also noted an upregulation of the proapoptotic BH3 member Bim.

In summary, we report that navitoclax significantly enhances the activity of chemotherapeutic and targeted agents against a variety of solid tumor types in vitro which may, in part, be due to restoration of the apoptotic potential of the cell. While we used CI values from these studies as a guide to select the best combinations to be tested in vivo, they were not always predictive of antitumor activity. For example, navitoclax did not consistently enhance the antitumor activity of docetaxel in BxPC3 and of gemcitabine in SKOV3 xenograft models, even though their CI values were less than 0.1 in vitro (data not shown). The latter shows the inherent challenges in using in vitro studies to predict in vivo efficacy when differences in drug exposures and microenvironment are not accounted for. Using a mechanism-based approach, we identified multiple efficacious combinations that were confirmed in vivo. Continued exploration of mechanism-based studies with navitoclax and other therapeutic agents should ultimately aid clinical trial design and success in the clinic.

All the authors are employees of Abbott Laboratories or Genentech whose potential product was studied in the present work.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.