Targeting BCL2 with Venetoclax Enhances the Efficacy of the KRAS^G12D Inhibitor MRTX1133 in Pancreatic Cancer

Abstract

MRTX1133 is currently being evaluated in patients with pancreatic ductal adenocarcinoma (PDAC) tumors harboring a KRAS^G12D mutation. Combination strategies have the potential to enhance the efficacy of MRTX1133 to further promote cell death and tumor regression. In this study, we demonstrated that MRTX1133 increased the levels of the pro-apoptotic protein BIM in PDAC cells and conferred sensitivity to the FDA-approved BCL2 inhibitor venetoclax. Combined treatment with MRTX1133 and venetoclax resulted in cell death and growth suppression in 3D cultures. BIM was required for apoptosis induced by the combination treatment. Consistently, BIM was induced in tumors treated with MRTX1133, and venetoclax enhanced the efficacy of MRTX1133 in vivo. Venetoclax could also re-sensitize MRTX1133-resistant PDAC cells to MRTX1133 in 3D cultures, and tumors established from resistant cells responded to the combination of MRTX1133 and venetoclax. These results provide a rationale for the clinical testing of MRTX1133 and venetoclax in PDAC patients.

INTRODUCTION

Pancreatic ductal adenocarcinoma (PDAC) is currently the third leading cause of cancer-related death in the US. Despite advances in surgery, radiation, chemotherapy, and immunotherapy, the 5-year survival rate for patients with PDAC is ~13% (1–3). Mutational activation of the KRAS gene, critical for PDAC initiation and maintenance in mouse models (4, 5), is present in over 90% of human PDAC tumors, with KRAS^G12D mutation accounting for ~45% of the KRAS mutants (6, 7). Recently, inhibitors targeting KRAS^G12C, present in ~1–2% of human PDAC tumors, have been developed and tested in PDAC patients (8, 9). However, single-agent treatment with KRAS^G12C inhibitors mainly demonstrates brief disease stabilization, with cancer cells activating bypass pathways and developing resistance to KRAS inhibitors (8, 9). As inhibitors of the more common KRAS mutations in PDAC are now emerging and with the KRAS^G12D inhibitor MRTX1133 in the early phase of clinical testing (10), there is increasing interest in identifying novel combination strategies to further enhance the efficacy of KRAS inhibitors (11, 12).

Several combination therapies with MRTX1133 have been tested in pre-clinical models. For example, in xenograft mouse studies, the efficacy of MRTX1133 in human tumors is enhanced by combining it with an EGFR antibody (13). In immunocompetent mice, as the T cells contribute to the effects of MRTX1133 in PDAC models (14, 15), the efficacy can be enhanced by combining with immune checkpoint inhibitors (15). Notably, while MRTX1133 effectively suppresses cancer cell proliferation in tumors in immunocompetent mice, CD8+ T cells are required to induce cell death and tumor regression (15). The findings that MRTX1133 primarily suppresses cancer cell proliferation without directly inducing apoptosis in immunocompetent mouse model suggest a need to identify additional combination regimens with KRAS^G12D inhibitors that promote cell death and tumor regression.

The intrinsic (mitochondrial) apoptosis pathway is tightly regulated by the BCL-2 protein family, which includes the BH3-only pro-apoptotic proteins (e.g., BIM), the multi-domain effector proteins BAK and BAX, and the anti-apoptotic proteins (e.g., BCL2) (16–18). The BCL-2 family of anti-apoptotic proteins restrain cell death mediators to maintain cellular viability (16–18). In contrast, the BH3-only pro-apoptotic proteins inhibit the pro-survival BCL-2 proteins to induce apoptosis (16–18). Significantly, these proteins can be regulated through post-translational mechanisms. For example, ERK1/2 signaling can promote phosphorylation and subsequent proteasome-dependent degradation of BIM (19–21).

In this study, we demonstrate that while MRTX1133 increases BIM protein levels, MRTX1133 fails to suppress the growth or induce apoptosis of tumor cells grown in confluent 3D collagen cultures. However, adding the FDA-approved BCL2 inhibitor venetoclax to MRTX1133 induces cell death and suppresses growth in confluent 3D collagen cultures. We also show that BIM is induced in tumors treated with MRTX1133 and demonstrate that venetoclax enhances the efficacy of MRTX1133 in vivo. We have also found that venetoclax can re-sensitize the resistant cells to MRTX1133 in collagen cultures, and tumors established from the resistant cells respond to the combination of MRTX1133 and venetoclax. These results provide a rationale for the clinical testing of MRTX1133 and venetoclax in PDAC patients.

MATERIALS AND METHODS

Chemicals –

MRTX1133 (HY-134813), Trametinib (HY-10999), and venetoclax (HY-1553) were purchased from MedChemExpress. z-VAD-FMK (S7073) and additional venetoclax (S8048) were purchased from Selleck Chemicals. The membrane-permeable thiazolyl blue tetrazolium bromide (MTT) dye (ab146345) and the thymidine analog BrdU (5-bromo-2’-deoxyuridine) (ab142567) were obtained from Abcam.

Tumor cell lines –

Mouse PDAC cell lines KPC-2138 (2138) and KPC-3213 (3213) were derived from PDAC tumors developing in the KPC (LSL-Kras^G12D/+/LSL-Trp53^R172H/+/Pdx-1-Cre) mouse model in the C57BL/6 background (22). KPC-1245 (1245) cell line was provided by David Tuveson (Cold Spring Harbor Laboratories, Cold Spring Harbor, New York, US) (23). KPC-K cell lines were generated by culturing KPC cells over six weeks in increasing concentrations of MRTX1133 to ≥2μM, with cells maintained in 2μM MRTX1133. The human PDAC cell line PANC1 was obtained from ATCC (CRL-1469). PANC1-K cells were generated similarly to the KPC-K cell lines and maintained in 2μM MRTX1133. The mouse and human PDAC-K cell lines were cryopreserved after being generated. All cell lines were used within 15 passages after revival from frozen stocks. These cell lines did not undergo authentication and were not tested for mycoplasma. All cell lines were cultured in DMEM supplemented with 10% Fetal Bovine Serum (FBS) and antibiotics (100 U/mL Penicillin and 100 μg/mL Streptomycin).

Floating collagen gel cultures –

Acidified rat tail collagen I (Corning) was neutralized with 0.34N NaOH and diluted to a final concentration of 1.2 mg/ml (24). Single-cell suspensions were mixed with diluted collagen to obtain 10⁵ cells/ml, and 1 mL of collagen/cell mixture was plated into Ultra-Low Attachment 6-well plates (Corning) (25, 26). After 1 hour of polymerization at 37°C, 2 mL of the media were added, and the gels were detached from the well using a fine pipette tip. For ‘low’ confluent collagen cultures, the cells in the floating collagen gels were treated after 48 hours, while for the ‘high’ confluent collagen cultures, the cells were treated after 96–120 hours. The cells were imaged using the EVOS XL Core Imaging System (ThermoFisher), and the relative growth was quantified.

BIM knockdown –

Silence-select pre-designed siRNA against mouse Bim (s201095) and human BIM (195011) were obtained from ThermoFisher. The siRNA transfections were carried out using RNAiMAX (Invitrogen) according to the manufacturer’s instructions (27).

MTT assay –

PDAC cells were seeded in low density (2×10³ cells/well) or high density (10×10³ cells/well) in each well of a 96-well plate in serum-free DMEM. After 16 hours, media and drugs were added, and cells cultured for 72 hours. At this time, 20μL of a 5mg/ml MTT solution was added to each well (1:10). After two hours, media was aspirated, crystals dissolved in DMSO, and 570 nm absorbance determined by plate reader (28).

BrdU incorporation assay –

PDAC cells growing in ‘low’ and ‘high’ confluent 3D collagen cultures were treated with 2μM BrdU for 1 hour. Cells were extracted from collagen gels using collagenase I (Worthington Biochemical), fixed with 4% paraformaldehyde, and then collected for immunocytochemistry. Antigen retrieval was carried out at 37°C in 2N HCl for 1 hour. The slides were incubated with 1% BSA/PBS for 1 hour at room temperature. The cells were stained overnight at 4°C with BrdU antibody (MoBU-1, B35139, ThermoFisher, RRID:AB_2536439) at 1:100 dilution followed by Secondary Alexa Fluor 488 antibody (A32723, ThermoFisher, RRID:AB_2633275). The nuclei were co-stained with DAPI, the slides mounted and visualized with EVOS M7000 microscope (ThermoFisher), and quantified by ImageJ.

Western blot analysis –

Cells were extracted from collagen gels using collagenase I (Worthington Biochemical) and lysed in RIPA buffer supplemented with phosphatase and protease inhibitors (Calbiochem) (29). Equal amounts of protein were run in reducing conditions as described previously (22, 29). The following antibodies were used at the dilution recommended by the manufacturers: p-ERK1/2 (#9101, Cell Signaling, RRID:AB_331646), total ERK1/2 (#9102, Cell Signaling, RRID:AB_330744), mouse BIM (#2933, Cell Signaling, RRID:AB_1030947), mouse BCL2 (554218, BD Pharmingen, RRID:AB_395311), cleaved caspase-3 (#9664, Cell Signaling, RRID:AB_2070042), human BCL2 (#3498, Cell Signaling, RRID:AB_1903907), PARP (#9542, Cell Signaling, RRID:AB_2160739), HSP90 (#4877 Cell Signaling, RRID:AB_2233307), and GAPDH (MAB374, Millipore Sigma, RRID:AB_2107445). Secondary anti-mouse IgG were purchased from Sigma-Aldrich (A4416, RRID:AB_258167) or Cell Signaling (#7076, RRID:AB_330924). Anti-rabbit IgG antibodies were purchased from Sigma-Aldrich (A6667, RRID:AB_258307) or Cell Signaling (#7074, RRID: AB_2099233).

Animal studies –

For subcutaneous tumor studies, 2138, 3213, 1245, 2138-K, 3213-K, and 1245-K cells (25 × 10³ cells/100 μL of Matrigel) were injected under the skin into the flanks of 6–8-week-old C57BL/6 mice. When tumors achieved a volume of ~250–300 mm³ for 2138, 3213, and 1245 cells or ~80–100 mm³ for 2138-K, 3213-K, and 1245-K cells, mice were randomized and treated with vehicle control, MRTX1133 (30mg/kg 2x daily (13–15)), venetoclax (15 mg/kg daily (30)), or the combination of MRTX1133 and venetoclax. In additional experiments, 2138, 3213, and 1245 tumors measuring ~80–100 mm³ were treated with vehicle control or MRTX1133 (30mg/kg 2x daily). Tumor volume was calculated using the formula V = (W² × L)/2 where V is tumor volume, W is tumor width, and L is tumor length by caliper measurement. At the study endpoint (e.g., tumor volume exceeding 1200–1500 mm³, tumor ulceration, weight loss of >20%, or weakness and inactivity), tumor-bearing mice were euthanized by CO₂ inhalation and cervical dislocation.

IHC staining and analysis –

Mouse tumors were stained for pERK (#9101, Cell Signaling, RRID:AB_331646, 1:100), Bim (#2933, Cell Signaling, RRID:AB_1030947, 1:1200), and cleaved caspase-3 (#9664, 1:1000, Cell Signaling, RRID:AB_2070042). Antigen retrieval was carried out, as previously described (31), with either pH 6.0 or pH 9.0 buffer according to the manufacturer’s instructions. After antigen retrieval, tumor sections were incubated with BLOXALL (Vector Laboratories) for 20 minutes at room temperature. Sections were incubated with primary antibodies in 1% BSA/PBS overnight at 4°C. ImmPRESS^® Secondary HRP anti-rabbit IgG (Peroxidase) Polymer Detection Kit was purchased from Vector Laboratories. Photographs were taken on the FeinOptic microscope using the Jenoptik ProgRes C5 camera, and pERK, BIM, and cleaved caspase-3 signals were quantified. At least 5 different fields of view were examined for final quantification.

Immunofluorescence staining –

For tumor specimens, antigen retrieval was carried out as previously described using pH 6.0 citrate buffer (28). After antigen retrieval, sections were incubated with 10% normal goat serum (Agilent) for 20 minutes at room temperature, followed by incubation with 1% BSA/PBS buffer for 1 hour at room temperature. The sections were incubated with Ki67 (#12202, 1:1000, Cell Signaling, RRID:AB_2620142) and cytokeratin 19 (CK19, 1:100, DHSB, RRID:AB_2133570) in 1% BSA/PBS overnight at 4°C. Secondary Alexa Fluor 488 and Alexa 647 antibodies were purchased from ThermoFisher and used at 1:400–1:1000 dilution factor. DAPI was used to counterstain the nuclei. Final pictures, taken on EVOS^™ M7000 microscope, were quantified using ImageJ. At least 5 different fields of view were examined for final quantification.

Statistical analysis –

Error bars represent the standard error of mean (SEM) or standard deviation (SD), as specified in the figure legends. The values for n, p, and the specific statistical test performed for each experiment are specified in the figure legends. All statistical analyses were done using GraphPad. A p<0.05 value was considered significant.

Study approval –

All animal work and procedures were approved by the Northwestern University Institutional Animal Care and Use Committee. The animal experiments were performed in accordance with relevant guidelines and regulations.

Data availability –

All raw data generated in this study are available upon request from the corresponding author.

RESULTS

MRTX1133 inhibits the growth of PDAC cells in ‘low’ but not ‘high’ confluent collagen cultures.

PDAC is hallmarked by a desmoplastic collagen-rich stroma that is not represented in traditional 2D cell culture systems (32, 33). As cells grown in 3D floating collagen gels better mirror tumor morphogenesis than those grown on plastic (25, 26), we utilized the floating collagen gel system to test the efficacy of MRTX1133 against PDAC cell lines. When tested in ‘low’ confluent collagen cultures (Fig. 1A), MRTX1133 suppressed ERK1/2 phosphorylation and effectively blocked the growth of mouse and human PDAC cell lines (Figs. 1B, 1C). However, when tested in ‘high’ confluent collagen cultures (Fig. 1D), MRTX1133, even though it blocked ERK1/2 phosphorylation (Fig. 1E), failed to suppress the growth of these cell lines (Fig. 1F).

Figure 1: MRTX1133 inhibits the growth of PDAC cells in ‘low’ but not ‘high’ confluent collagen cultures.

A. PDAC cells (1×10⁵) are grown in floating collagen gels (1.2 mg/ml) for 48 hours to generate ‘low’ confluent cultures. Depiction of ‘low’ confluent floating collagen cultures. B. PDAC cell lines (2138, 3213, 1245, PANC1) grown in ‘low’ confluent collagen cultures were treated with MRTX1133 (0.5μM) for 8 hours, and the effect on ERK1/2 phosphorylation was analyzed by Western blotting. Blots are representative of at least three biological replicates. C. PDAC cell lines grown in ‘low’ confluent collagen cultures were treated with DMSO or 0.5μM MRTX1133 for 72 hours. The cells were imaged at baseline and after 1 day, 2 days, and 3 days of treatment, and the relative growth was quantified. Error bars ± SD, n=3, unpaired t-test. **, p<0.01; ***, p<0.001. Scale bar = 100μm. D. PDAC cells (1×10⁵) are grown in floating collagen gels (1.2 mg/ml) for 96–120 hours to generate ‘high confluent’ cultures. Depiction of ‘high confluent’ floating collagen cultures. E. PDAC cell lines grown in ‘high confluent’ collagen cultures were treated with MRTX1133 (0.5μM) for 8 hours, and the effect on ERK1/2 phosphorylation was analyzed by Western blotting. Blots are representative of at least three biological replicates. F. PDAC cell lines grown in ‘high confluent’ collagen cultures were treated with DMSO or 0.5μM MRTX1133 for 72 hours. The cells were imaged at baseline and after 1 day, 2 days, and 3 days of treatment and the relative growth was quantified. Error bars ± SD, n=3, unpaired t-test. ns, not significant. Scale bar = 100μm.

To understand the differential response to MRTX1133 in ‘low’ and ‘high’ confluent collagen cultures, we evaluated the effect of confluency on PDAC cell proliferation. PDAC cells in ‘low’ and ‘high’ confluent collagen cultures were treated with MRTX1133 for 8 hours, and BrdU was added an hour before collection. The samples were processed for BrdU incorporation by immunocytochemistry. Cells in ‘high’ confluent collagen cultures exhibit reduced BrdU incorporation compared to cells in ‘low’ confluent collagen cultures, indicating that the cells in ‘high’ confluent cultures proliferate significantly slower (Fig. S1). While MRTX1133 decreases BrdU incorporation in both ‘low’ and ‘high’ confluent collagen cultures, the effect was more pronounced in the ‘low’ confluent collagen cultures (Fig. S1).

PDAC cells grown on tissue culture plastic exhibit increased sensitivity to MRTX1133 in ‘low’ confluent than in ‘high’ confluent 2D cultures.

Given the lack of efficacy of MRTX1133 in ‘high’ confluent collagen cultures, we evaluated the efficacy of MRTX1133 in PDAC cells plated in ‘low’ and ‘high’ density on tissue culture plastic. We show that while MRTX1133 suppresses ERK1/2 phosphorylation in both ‘low’ and ‘high’ confluent 2D cultures (Figs. S2A, S2B), MRTX1133 has minimal effect on suppressing the cell viability in ‘high’ confluent 2D cultures (Fig. S2C). Moreover, consistent with a recent report (34), the efficacy of MRTX1133 in ‘low’ confluent cultures is lower when cells are grown on 2D tissue culture plastic compared to when cells are grown in 3D collagen gels (Figs. 1C, S2C).

MRTX1133 increases the levels of the pro-apoptotic protein BIM.

The KRAS/MEK/ERK pathway is a central regulator of cell survival and can suppress apoptosis (5, 35). For example, the MEK/ERK signaling can decrease the protein levels of the pro-apoptotic BIM by promoting its degradation (19–21). Thus, to identify potential combination strategies to improve the efficacy of MRTX1133 in ‘high’ confluent 3D collagen cultures, we evaluated the effects of MRTX1133 on the pro-apoptotic BIM. We found that MRTX1133, like the MEK inhibitor Trametinib, increased the levels of the BIM protein in PDAC cells (Fig. 2A).

Figure 2: BCL2 inhibitor venetoclax enhances the efficacy of MRTX1133 in ‘high’ confluent collagen cultures.

A. PDAC cells (2138, 3213, 1245, PANC1) in ‘high’ confluent collagen cultures were treated with MRTX1133 (0.5μM) or the MEK1/2 inhibitor Trametinib (0.1μM) for 8 hours. The effect on BIM and BCL2 was analyzed by Western blotting. Blots are representative of at least three biological replicates. B. PDAC cells in ‘high’ confluent collagen cultures were co-treated with venetoclax (2.5μM) and MRTX1133 (0.5μM) for 8 hours, and the effect on cell death (cleaved caspase 3, cleaved PARP) was analyzed. Blots are representative of at least three biological replicates. C. ‘High’ confluent collagen cultures of PDAC cells were co-treated with venetoclax (V, 2.5μM) and MRTX1133 (M, 0.5μM) for 48 hours, and the effect on cell growth was analyzed. Error bars ± SD, n=3–4. One-way ANOVA, followed by Tukey’s multiple comparison test. D. PDAC cells were transfected with control siRNA or siRNA against BIM for 72 hours, plated in collagen, and treated with MRTX1133 (0.5μM) for 8 hours. The effect on BIM was analyzed using Western blotting. The transfected cells were also co-treated with venetoclax (2.5μM) and MRTX1133 (0.5μM) for 8 hours, and the effect on cell death (cleaved caspase 3) was analyzed. Blots are representative of three biological replicates. E. ‘High’ confluent collagen cultures of PDAC cells were co-treated with venetoclax (V, 2.5μM) and MRTX1133 (M, 0.5μM) for 48 hours with or without the pan-caspase inhibitor zVAD-FMK (zVAD, 10μM) and the effect on cell growth was analyzed. Error bars ± SD, n=3. One-way ANOVA, followed by Tukey’s multiple comparison test. **, p<0.01; ***, p<0.001; ****, p<0.0001.

Venetoclax enhances the efficacy of MRTX1133 in ‘high’ confluent collagen cultures.

Despite the increase in BIM protein levels, MRTX1133 failed to suppress the growth of tumor cells in ‘high’ confluent collagen cultures. Thus, to alter the balance between pro- and anti-apoptotic proteins towards apoptosis, we treated the cells with the FDA-approved BCL2 inhibitor venetoclax (36, 37). MRTX1133 had minimal to no effects on the anti-apoptotic protein BCL2 in the PDAC cells (Fig. 2A). While venetoclax did not induce apoptosis (cleaved caspase-3 (c-C3) or cleaved-PARP) or suppress the growth of collagen cultures (Figs. 2B, 2C), adding venetoclax to MRTX1133 significantly induced cell death and suppressed growth of ‘high’ confluent 3D collagen cultures (Figs. 2B, 2C).

We also evaluated the combination of venetoclax and MRTX1133 in PDAC cells grown on 2D tissue culture plastic. In contrast to the effect of venetoclax in 3D collagen culture, venetoclax had minimal to no effect on the ability of MRTX1133 to decrease the cell viability of PDAC cells grown on 2D tissue culture plastic (Fig. S3).

BIM is required for apoptosis induced by the combination therapy in ‘high’ confluent collagen cultures.

Since we found that the induction of BIM was not sufficient for MRTX1133 to induce cell death, we evaluated whether BIM was required for the response of the combination treatment with MRTX1133 and venetoclax in cells grown in ‘high’ confluent 3D collagen cultures. Using siRNA to suppress MRTX1133-induced BIM expression, we show that siRNA against BIM attenuates apoptosis induced by the combination treatment (Fig. 2D). We also show that the pan-caspase inhibitor zVAD-FMK limits the growth suppression seen with the combination treatment (Fig. 2E).

Venetoclax enhances the efficacy of MRTX1133 in vivo.

We next evaluated the combination of venetoclax and MRTX1133 in vivo. KPC cell lines 2138, 3213, and 1245 were implanted into the flank of syngeneic C57BL/6 mice. Once the tumors reached ~250–300 mm³, the mice were randomized to four treatment groups. Mice received intraperitoneal (IP) injections of DMSO (vehicle control), MRTX1133 (30 mg/kg 2x/day), venetoclax (15mg/kg daily), or the combination of MRTX1133 and venetoclax. As with our findings in the collagen gels, single-agent treatment with venetoclax did not affect the growth of 2138, 3213, or 1245 tumors (Fig. 3A). MRTX1133 had varying effects on the growth of these three KPC cell lines. MRTX1133 prevented the growth of 2138 tumors transiently (Fig. 3A). Comparing the tumor volumes at the study endpoint, the majority of the 2138 tumors treated with MRTX1133 showed an increase in tumor size compared to the start of treatment (Fig. 3B). While the 3213 tumors are overall more responsive to MRTX1133 compared to the 2138 tumors, the 3213 tumors showed a similar trend in tumor growth following MRTX1133 treatment as the 2138 tumors (Figs. 3A, 3B). In contrast, the single-agent treatment of 1245 tumors with MRTX1133 led to prolonged disease stability (Figs. 3A, 3B).

Figure 3: Venetoclax enhances the efficacy of MRTX1133 in vivo.

2138, 3213, and 1245 KPC cells (2.5×10⁴) were implanted via subcutaneous injection into the flanks of C57BL/6 mice. Once animals developed 250–300mm³ tumors, they were treated with vehicle control (C, DMSO), venetoclax (V, 15mg/kg daily), MRTX1133 (M, 30mg/kg 2x/day), or the combination of venetoclax and MRTX1133. A. Tumor size was measured 2–3 times/week by caliper, and tumor volume was calculated using the formula V = (W2 × L)/2 and normalized to tumor volume at the start of treatment. Error bars ± SEM, n=8–9 per treatment group. Significance between the MRTX1133-treated group and the combination treatment group was determined by two-way ANOVA with the Sidak multiple comparisons test. **, p<0.01; ***, p<0.001; ****, p<0.0001. Changes in tumor volume between the dotted lines (−30%, +20%) were considered as stable disease. B. Relative tumor volumes at study endpoints were compared to tumors at the start of treatment and change in tumor volume was calculated. Unpaired t-test. **, p<0.01; ***, p<0.001. C. 3213 tumors at the study endpoints were H&E stained and stained for pERK, BIM, the proliferation marker Ki-67, the epithelial marker CK19, and the apoptosis marker cleaved caspase 3 (c-C3). Error bars ± SD. One-way ANOVA, followed by Tukey’s multiple comparison test. *, p<0.05; **, p<0.01; ***, p<0.001; ****, p<0.0001; ns, not significant. Scale bars = 100μm.

Venetoclax enhanced the efficacy of MRTX1133 in all three KPC tumors. The combination of MRTX1133 and venetoclax suppressed the growth rate of 2138 tumors, resulting in prolonged disease stability (Figs. 3A, 3B). In contrast, compared to single-agent treatment with MRTX1133, the combination treatment of 3213 tumors resulted in tumor regression (Figs. 3A, 3B). The co-treatment of 1245 tumors with venetoclax and MRTX1133 also resulted in tumor regression (Figs. 3A, 3B). Importantly, tumor-bearing mice treated with the combination treatment maintained their body weights (Fig. S4), suggesting that the combination treatment was well tolerated.

Histologic analysis of the KPC tumors at the study endpoint showed that MRTX1133 decreased ERK1/2 phosphorylation and induced BIM expression (Figs. 3C, S5). While treatment with MRTX1133, either as a single-agent or as part of the combination treatment, decreased Ki67 staining, only the combination treatment significantly increased staining for c-C3 (Figs. 3C, S5).

Venetoclax re-sensitizes the resistant cells to MRTX1133 in collagen cultures.

In additional studies, we established an in vitro model of MRTX1133 resistance by culturing PDAC (KPC, PANC1) cell lines over six weeks in increasing concentrations of MRTX1133 to ≥2μM to generate PDAC-K (KPC-K, PANC1-K) cell lines. These mouse and human PDAC-K cell lines are poorly responsive to MRTX1133-induced suppression of ERK1/2 activation in 2D cultures and growth inhibition in 3D collagen cultures (Figs. 4A, 4B). Next, we evaluated whether the PDAC-K cells upregulated BIM expression when treated with MRTX1133 in collagen cultures. MRTX1133 treatment decreased pERK signaling and upregulates BIM expression in 3D collagen cultures (Fig. 4C). Given the BIM upregulation in collagen cultures, we treated the PDAC-K cells growing in collagen cultures with venetoclax. Co-treatment with venetoclax and MRTX1133 resulted in increased apoptosis (c-C3 and cleaved-PARP) and growth suppression in collagen cultures (Figs. 4D, 4E).

Figure 4: MRTX1133-resistant cells respond to the combination of MRTX1133 and venetoclax in collagen cultures.

PDAC (2138, 3213, 1245, PANC1) cells were treated over 6–8 weeks with increasing concentrations of MRTX1133 to 2μM or greater to generate PDAC-K cells. A. PDAC and PDAC-K cells growing on tissue culture plastic were treated with increasing concentrations of MRTX1133 for 8 hours, and ERK activation was evaluated by Western blotting. Blots are representative of three biological replicates. B PDAC and PDAC-K cells grown in ‘low’ confluent collagen cultures were treated with DMSO or 2μM MRTX1133 for 72 hours. The cells were imaged at baseline and after 1 day, 2 days, and 3 days of treatment, and the relative growth was quantified. Error bars ± SD, n=3. Unpaired t-test. *, p<0.05; **, p<0.01. C. PDAC-K cells growing in ‘high’ confluent collagen cultures were treated with MRTX1133 (2μM) and the effect on p-ERK and BIM protein levels was analyzed by Western blotting. Blots are representative of three biological replicates. D. PDAC-K cells in ‘high’ confluent collagen cultures were treated with venetoclax (2.5μM) and MRTX1133 (2.0μM) for 8 hours, and the effect on cell death (c-C3, cleaved PARP) was analyzed. Blots are representative of at least three biological replicates. E. PDAC-K cells growing in ‘high’ confluent collagen cultures were co-treated with venetoclax (V, 2.5μM) and MRTX1133 (M, 2.0μM) for 48 hours, and the effect on cell growth was analyzed. Error bars ± SD, n=3. One-way ANOVA, followed by Tukey’s multiple comparison test. **, p<0.01; ****, p<0.0001.

Tumors established from KPC-K cells respond to the combination of MRTX1133 and venetoclax.

Finally, we evaluated the efficacy of the combination of MRTX1133 and venetoclax in tumors established from 2138-K, 3213-K, and 1245-K cells. Due to the rapid growth of some of the KPC-K cells in vivo (Fig. S6), we treated mice bearing small (~80–100 mm³) KPC-K tumors (Fig. 5A). As with the parental KPC tumors, single-agent treatment with venetoclax did not affect their growth. While single-agent treatment of 2138-K tumors with MRTX1133 did not slow their growth, single-agent treatment with MRTX1133 of the small 3213-K and 1245-K tumors resulted in disease stability (Fig. 5A). To demonstrate that the KPC-K tumors, in particular the 3213-K and 1245-K tumors, exhibit relative resistance to MRTX1133, we also treated similarly sized (~80–100 mm³) tumors established from 2138, 3213, and 1245 cell lines with MRTX1133. In contrast to the small 3213-K and 1245-K tumors, the corresponding small parental KPC tumors treated with MRTX1133 demonstrated tumor regression (Fig. S7), indicating that the tumors established from the KPC-K cell lines exhibit relative resistance to MRTX1133 in vivo.

Figure 5: Tumors established from KPC-K cells respond to the combination of MRTX1133 and venetoclax.

2138-K, 3213-K, and 1245-K KPC cells (2.5×10⁴) were implanted via subcutaneous injection into the flanks of C57BL/6 mice. Once animals developed ~80–100 mm³ tumors, they were treated with vehicle control (C, DMSO), venetoclax (V, 15mg/kg daily), MRTX1133 (M, 30mg/kg 2x/day), or the combination of venetoclax and MRTX1133. A. Tumor size was measured 2–3 times/week by caliper, and tumor volume was calculated using the formula V = (W2 × L)/2 and normalized to tumor volume at the start of treatment. Error bars ± SEM, n= 7–9 per treatment group. Significance between the MRTX1133-treated group and the combination treatment group was determined by two-way ANOVA with the Sidak multiple comparisons test. ***, p<0.001; ****, p<0.0001. Changes in tumor volume between the dotted lines (−30%, +20%) were considered as stable disease. B. Relative tumor volumes at study endpoints were compared to tumors at the start of treatment and change in tumor volume was calculated. Unpaired t-test. C. 3213-K tumors at the study endpoints were H&E stained and stained for pERK, BIM, the proliferation marker Ki-67, the epithelial marker CK19, and the apoptosis marker cleaved caspase 3 (c-C3). Error bars ± SD. One-way ANOVA, followed by Tukey’s multiple comparison test. *, p<0.05; **, p<0.01; ***, p<0.001; ****, p<0.0001; ns, not significant. Scale bars = 100μm.

As with the parental tumors, venetoclax enhanced the response of MRTX1133 in all three KPC-K tumors (Figs. 5A, 5B), with mice maintaining their weights with the combination treatment (Fig. S8). While the 2138-K tumors continued to grow, the combination treatment significantly slowed their growth. In contrast, the combination treatment caused regression of 3213-K and 1245-K tumors (Figs. 5A, 5B). The histologic analysis showed that MRTX1133 decreased pERK and induced BIM expression (Figs. 5C, S9). Treatment with MRTX1133, either alone or in combination with venetoclax, decreased proliferation in all three KPC-K tumors (Figs. 5C, S9). While single-agent treatment with venetoclax induced apoptosis in the 1245-K tumors (Fig. S9), the combination of MRTX1133 and venetoclax induced apoptosis in all three KPC-K tumors (Figs. 5C, S9).

DISCUSSION

There is increasing excitement with inhibitors targeting the RAS protein, which was previously considered undruggable (11, 12). The KRAS^G12C inhibitors have received accelerated approval for lung cancer and are being considered for additional approvals as part of combination regimens (38). For example, the combination of the KRAS^G12C inhibitor Adagrasib and the EGFR antibody cetuximab is under FDA consideration as a subsequent-line treatment for colon cancers with the KRASG12C mutation (39, 40). The KRAS^G12C inhibitors have also been evaluated in pancreatic cancer, where the KRASG12C mutation is seen in only 1–2% of pancreatic cancer patients (8). Clinical trials of inhibitors targeting KRAS^G12D, present in 40–45% of human PDAC tumors (6, 7), have recently been initiated (10–12). Given that the KRAS^G12C inhibitors demonstrate a response rate of ~20% and a progression-free survival of ~4–5 months in pancreatic cancer (8, 9), it is likely that MRTX1133 and other KRAS^G12D inhibitors will need to be combined with other drugs for enhanced anti-tumor response.

In this study, we show that venetoclax can enhance the efficacy of MRTX1133 in collagen cultures and in animal studies. Notably, we show that while MRTX1133 is effective in low-confluent collagen cultures, MRTX1133 has minimal to no effect when tested in high-confluent collagen cultures. Similarly, we show while MRTX1133 has some activity in cells grown in low density on 2D tissue culture plastic, MRTX1133 has minimal activity when the cells are plated at high density. By BrdU staining, we show reduced proliferation under high-confluent collagen culture conditions compared to when cells are grown in low-confluent culture conditions, thus explaining the lack of response to single-agent treatment in high-confluent collagen culture conditions. Others have demonstrated that targeting KRAS is more effective in 3D than 2D cultures (34, 41, 42). While our data agree with these previous observations when the cells are plated in low density in 2D and low-confluent collagen cultures in 3D, we show that the confluency of the cells in the 3D culture conditions can significantly limit the efficacy of MRTX1133. Nonetheless, our confluent collagen cultures enabled us to identify the combination regimen of venetoclax and MRTX1133 that shows efficacy in vivo.

We show that the efficacy of the combination therapy is dependent on increased MRTX1133-mediated BIM expression. Others have demonstrated that BIM is critical in mediating the response of inhibitors targeting the EGFR-RAS-RAF-MEK signaling pathway. Apoptosis induced by MEK inhibition in lung cancer cells is mediated by BIM (43). Similarly, EGFR inhibitor-induced apoptosis of lung cancer cells requires BIM, which is induced due to the suppression of MEK/ERK signaling (44). Moreover, the anti-tumor response of B-RAF mutant human tumor cells to MEK inhibition requires BIM (45). Similarly, induction of BIM expression contributes to the response of the combination of Sorafenib and a MEK inhibitor in hepatocellular carcinoma cells (46).

While MRTX1133 increases BIM protein levels, we show that the increased BIM protein expression is not sufficient to induce apoptosis following single-agent treatment with MRTX1133 in high-confluent collagen cultures. This is consistent with reports in lung and pancreatic cancer models where the efficacy of MEK inhibitors can be increased by co-targeting the BCL-2 family proteins (47). Similarly, the response of MEK inhibitors in B-RAF mutant human tumor cells and RAS-MAPK-mutated neuroblastoma cells can be enhanced by co-targeting the BCL-2 family of proteins (45, 48). The combination of a MEK inhibitor and an inhibitor targeting the BCL-2 family protein BCL-xL showed in vivo tumor regression in KRAS mutant lung cancer models (49). Also, EGFR inhibitor-induced killing of lung cancer cells can be enhanced by co-targeting the BCL-2 family proteins (44). We show that co-treatment with the BCL2 inhibitor venetoclax is required to induce apoptosis in high-confluent collagen cultures and in vivo.

We show that tumors established from parental KPC cell lines exhibit further regression when co-treated with MRTX1133 and venetoclax. Notably, a recent report showed that cancer cell killing and tumor regression in an immunocompetent model requires CD8+ T cells to induce apoptosis of cancer cells (15). Interestingly, MRTX1133 induces expression of FAS on cancer cells, leading to engagement of FAS ligand (FASL) on CD8+ T cells and subsequent killing of cancer cells through the extrinsic apoptotic pathway (15). Using our collagen culture model, we show that we need both BIM induction and BCL2 targeting to enhance apoptosis through the intrinsic apoptotic pathway with the combination regimen of venetoclax and MRTX1133.

We show that the combination therapy is also effective against cells that have become resistant to MRTX1133. While these cells exhibit a reduced response to MRTX1133 suppression of pERK in 2D, we show that in 3D collagen cultures, MRTX1133 decreases pERK and increases BIM protein levels. The combination therapy suppresses growth and induces apoptosis in 3D collagen cultures. The combination therapy also causes tumor regression in two of the three resistant cell lines (3213-K and 1245-K). However, in the 2138-K tumors that grow rapidly in vivo, we show that while single-agent treatment with MRTX1133 does not affect their growth, the combination treatment significantly slows their growth. In future studies, we will evaluate whether replacing venetoclax with navitoclax, which targets BCL2, BCL-xL, and BCL-W (17, 50), in the combination treatment with MRTX1133 causes regression of 2138-K tumors in vivo.

In conclusion, our study underscores the potential of the combination of MRTX1133 and the FDA-approved drug venetoclax in promoting cell death and tumor regression, even in pancreatic tumors established from cells resistant to MRTX1133. Notably, there is increasing realization that venetoclax and other inhibitors targeting the BCL-2 family proteins can overcome adaptive resistance to various therapies (18). Our results suggest testing venetoclax with MRTX1133 and other KRAS^G12D inhibitors in patients with pancreatic cancer.

SIGNIFICANCE.

The combination of MRTX1133 and the FDA-approved drug venetoclax promotes cancer cell death and tumor regression in pancreatic ductal adenocarcinoma, providing rationale for testing venetoclax with KRAS^G12D inhibitors in pancreatic cancer patients.