BH3 mimetics targeting BCL-XL have efficacy in solid tumors with RB1 loss and replication stress

Abstract

BH3 mimetic drugs that inhibit BCL-2, BCL-XL, or MCL-1 have limited activity in solid tumors. Through assessment of xenograft-derived 3D prostate cancer models and cell lines we find that tumors with RB1 loss are sensitive to BCL-XL inhibition. In parallel, drug screening demonstrates that disruption of nucleotide pools by agents including thymidylate synthase inhibitors sensitizes to BCL-XL inhibition, together indicating that replication stress increases dependence on BCL-XL. Mechanistically we establish that replication stress sensitizes to BCL-XL inhibition through TP53/CDKN1A-dependent suppression of BIRC5 expression. Therapy with a BCL-2/BCL-XL inhibitor (navitoclax) in combination with thymidylate synthase inhibitors (raltitrexed or capecitabine) causes marked and prolonged tumor regression in prostate and breast cancer xenograft models. These findings indicate that BCL-XL inhibitors may be effective as single agents in a subset of solid tumors with RB1 loss, and that pharmacological induction of replication stress may be a broadly applicable approach for sensitizing to BCL-XL inhibitors.

Fig. 1. Subset of PCa tumors is responsive to single-agent BCL-XL inhibition.

A BH3 mimetic screening in PCa 2D and 3D models. Growth inhibition of PCa models treated for 4 days with single agent Navitoclax (BCL-XL inhibitor), Venetoclax (BCL-2 inhibitor), or S63845 (MCL-1 inhibitor) over a dose range was assessed compared to control. The fraction of models that could be inhibited by >50% for each treatment at the highest concentration (1 μM) is shown. Left panel created in Biorender, Yuan, X. (2025) https://BioRender.com/r54j854B Heat map based on ranking of IC50 values showing distribution of PCa model sensitivity to BH3 mimetics. C BIDPC5 spheroids were treated with navitoclax and S63845 for 6 h. Apoptosis was assessed with luminescence-based caspase 3 and 7 activity assay. Mean and SEM for 5 biological replicates are shown. Data were analyzed by one-way ANOVA ***p < 0.001. D BIDPC5 (navitoclax sensitive) and BIDPC4 (navitoclax resistant) spheroids were treated with navitoclax (1 μM) for 6 h. Apoptosis markers cleaved PARP and cleaved caspase 3 (CC3) were assessed with immunoblotting. E BIDPC5 spheroids were treated with navitoclax (1 μM) for 6 h. Apoptosis was assessed with fluorescence-based caspase 3 and 7 activity assay. Scale bar is 100 μM (F) BIDPC1 and BIDPC5 patient derived xenografts were treated with intraperitoneal DMSO or navitoclax (50 mg/kg every other day) for 14 days. Fold change in tumor volume was assessed after completion to therapy. Asterisk (*) represents mice that developed toxicity requiring euthanasia during the 14 day treatment period. Data were analyzed by unpaired t-test ***p < 0.001 (for PC1 p = 0.0001 and for PC5 p = 0.0008). Upper panel was created in Biorender. Yuan, X. (2025) https://BioRender.com/f87k274G Kaplan-Meyer curves show overall survival. Data were analyzed by logrank test. Data are presented as mean values +/- SEM. Source data are provided as a Source Data file.

Fig. 2. RB1 loss increases dependency on BCL-XL.

A Genomic characterization of PCa models based on PCa-commonly altered genes. B Validation of RB1-loss induced sensitivity to BCL-XL inhibitor using an independent PCa PDX-derived spheroid cohort. Upper panels: RB1 expression by IHC and H-scores. Lower panels: organoid cultures from each PDX were treated with navitoclax for 6 h and caspase activity was assessed (left), or treated for 4 days and cells viability was assessed (right). Mean and SEM for 5 biological replicates are shown. Data were analyzed by one-way ANOVA *p < 0.05 (left panel: p = 0.003 for model CP267C and p = 0.002 for model CP336C, right panel: p = 0.002 for model CP267C and p = 0.001 for model CP336C). Scale bar is 100 μM. C Comprehensive analysis of solid tumor cell lines sensitivity to BCL-XL inhibitors (Navitoclax, WEHI and ABT737) based on RB1 alteration (combined mutation or copy number loss). Data were analyzed by unpaired t-test. D Volcano plot with effect size (x axis) and significance (y axis) of large-effect cancer-specific pharmacogenomic interactions based on RB1 alteration. Each circle represents an association between RB1 status and drug sensitivity analyzed using ANOVA (Genomics of Drug Sensitivity in Cancer- Sanger Institute/Mass General Cancer Center database). E Effect of short term RB1 silencing on BCL-XL sensitivity in LNCaP (RB1 proficient PCa cell line). Cells were treated with siRNA targeting RB1 (siRB1) or nontarget control (siNT) for 48 h. Navitoclax was then added to media for 6 h. Apoptotic effect was assessed by luminescence assay over a range of navitoclax concentrations (left panel) and apoptosis markers at 500 nM navitoclax by immunoblotting (right panel). Mean and SEM for 3 biological replicates are shown. Effects of RB1 siRNA at each navitoclax concentration were analyzed by unpaired t-test, *p < 0.05. Two-way ANOVA then showed that the effect of the shRNA on response to navitoclax was significant (p = 0.002). F Effect of long-term RB1 silencing on BCL-XL sensitivity in LNCaP cells. Cells were infected with shRNA targeting RB1 (shRB1) or nontarget control (shNT) constructs and treated with increasing doses of enzalutamide until development of resistance to 5 μM. Left panel: RB1 expression in the enzalutamide adapted cells. Middle panel: apoptosis activity of cells treated with navitoclax for 6 h based on luminescence assay. Right panel: viability assay of cells treated with navitoclax for 4 days based on luminescence assay. Mean and SEM for 3 biological replicates are shown. Effects of RB1 shRNA at each Navitoclax concentration were analyzed by unpaired t-test, *p < 0.05. Two-way ANOVA then showed that the effect of the shRB1 on response to Navitoclax was significant (p = 0.0001 for both apoptosis and viability analysis). Source data are provided as a Source Data file.

Fig. 3. Replication stress increases dependency on BCL-XL.

A Mechanism of action drug screening to identify agents that synergize with navitoclax. LNCaP cells were cultured in navitoclax (500 nM) or DMSO containing media. Compounds from the ICCB-Longwood Mechanism of Action Library (ICCB-L MoA) were then added in duplicate at 4 concentrations were then added, and viability assay was performed after 48 h. This figure was created in Biorender. Yuan, X. (2025) https://BioRender.com/p38v199B Left panel: average luminescence (viability) for cells cultured with library drug alone (Y axis) versus with navitoclax (X-axis). Results for all 4 drug concentrations are plotted. Circles above blue lane represent potential positive hits. Right panel: drug class of positive hits. C Single sample gene set enrichment analysis of TCGA primary prostate cancer database was carried out. Activation of ATR Response to Replication Stress gene set was calculated for individual samples and plotted relative to RB1 copy number. Data were analyzed by one-way ANOVA *p < 0.05 (p = 0.002). D LNCaP cells were treated with nolatrexed or vehicle for 48 h followed by navitoclax or vehicle for 4 days for cell recovery or 6 h for apoptosis. Left panel: cell cycle analysis of nolatrexed-treated LNCaP cells. Middle panel: cell recovery assessed by CTGlo assay. Right panel: caspase activation assessed by Caspase Glo 3/7 assay. Data are mean and SEM for biological replicates. Data at each nolatrexed concentration were analyzed by unpaired t-test, *p < 0.05. Two-way ANOVA then showed the effect of nolatrexed on response to navitoclax was significant (p < 0.0001 for both apoptosis and viability analysis). Upper panel was created in Biorender. Yuan, X. (2025) https://BioRender.com/h91a470E Analysis of apoptosis markers by immunoblotting of LNCaP cells treated with nolatrexed or vehicle for 2 days followed by navitoclax or vehicle for 6 h. F LNCaP cells were treated with nolatrexed alone or combined with thymidine for 2 days, followed by 6 h with navitoclax or vehicle, and apoptosis was assessed by CaspaseGlo 3/7 assay. Mean and SEM for 5 biological replicates are shown. Data at each nolatrexed concentration were analyzed by unpaired t-test, *p < 0.05. Two-way ANOVA then showed that the effect of adding thymidine was significant, p < 0.001. Source data are provided as a Source Data file.

Fig. 4. Pharmacologic induction of replication stress sensitizes solid tumor cells to BCL-XL inhibitors.

A Immunoblotting analysis of replication stress and apoptosis markers in LNCaP cells treated with nolatrexed (48 h) followed by navitoclax for 6 h. B Single cell electrophoresis (Comet assay) of LNCaP cells treated for 48 h with DMSO, nolatrexed (1 μM), raltitrexed (1 μM), or doxorubicin (250 nM). Data were analyzed by one-way ANOVA ***p < 0.001 (p = 0.0007). Scale bar is 10 μM. C Immunoblotting analysis of cells treated with nolatrexed or raltitrexed, alone or combined with thymidine rescue for 48 h. D Caspase 3/7 activity in LNCaP cells treated with raltitrexed and navitoclax. Mean and SEM for 3 biological replicates are shown. Data at each raltitrexed concentration were analyzed by unpaired t-test, *p < 0.05. Two-way ANOVA showed effect of raltitrexed was significant, p < 0.001. E, F Caspase 3/7 activity in LNCaP cells treated with BML-277 plus navitoclax (E) or AZD4320 (F). Both were significant by two-way ANOVA, p = 0.0001. G Caspase 3/7 activity in LNCaP cells treated with nolatrexed and AZD4320. Mean and SEM for 3 biological replicates are shown. p < 0.0001 by two-way ANOVA. H Caspase 3/7 activity in indicated cell lines (A375P - melanoma, RKO - poorly differentiated colorectal carcinoma, ZR75 - breast ductal carcinoma, A549 - lung carcinoma) treated with nolatrexed or raltitrexed for 48 h followed by navitoclax for 6 h. Mean and SEM for 3 biological replicates are shown. Two-way ANOVA showed effects of nolatrexed and raltitrexed were significant, p < 0.001. I Immunoblotting of BCL-2 family proteins in LNCaP cells treated with PZ18753B (BCL-XL/BCL2 degrader) for 24 h. J Caspase 3/7 activity in LNCaP cells treated with nolatrexed (48 h) and PZ18753B (16 h). Data are presented as mean values +/− SEM. Data analyzed by two-way ANOVA showed significant effect of nolatrexed, p < 0.001. Source data are provided as a Source Data file.

Fig. 5. Replication stress sensitization to navitoclax is mediated by decreased BIRC5/Survivin.

A LNCaP cells were treated with siRNA targeting BAX or BAK for 24 h followed by 48 h treatment with nolatrexed (4 μM), and addition of navitoclax for last 6 h. Whole cell lysates were then immunoblotted as indicated. B LNCaP cells were treated as in A with nolatrexed (4 or 8 μM). Apoptosis was quantified by CaspaseGlo 3/7 assay and displayed as a heat map. C Analysis of anti-apoptotic proteins in LNCaP cells treated with nolatrexed and navitoclax. Fold change in BCL-2 is quantified. D Analysis of pro-apoptotic proteins in LNCaP cells treated with nolatrexed and navitoclax. E Analysis of further pro- and anti-apoptotic genes in LNCaP cells treated with nolatrexed and navitoclax. Fold change in Survivin is quantified. F Survivin/BIRC5 was silenced with siRNA (versus nontarget siRNA) for 3 days in LNCaP cells and apoptosis markers after nolatrexed and navitoclax treatment were assessed with immunoblotting. G Survivin/BIRC5 was silenced with siRNA and Caspase 3/7 activity in response to navitoclax and S63845 was assessed. Mean and SEM for 5 biological replicates are shown. Data were analyzed by unpaired t-test, ***p < 0.001. H–J LNCaP cells treated with YM-155 were immunoblotted (H) and assessed for Caspase 3/7 activity in combination with navitoclax (I) or AZD4320 (J). Mean and SEM for 5 biological replicates are shown. Data at each YM-155 concentration were analyzed by unpaired t-test, *p < .05. Analysis by two-way ANOVA showed significant enhancement of apoptosis by YM-155 in I and J, p < 0.001 in both. Source data are provided as a Source Data file.

Fig. 6. Activation of p53 mediates survivin depletion and sensitivity to BCL-XL inhibitors.

A LNCAP cells treated with nolatrexed (2 μM) for 48 h were immunoblotting for DNA damage response and p53 activation. Fold changes in P-RPA32 and p53 are quantified. B Time-course for DNA damage response and p53 induction in nolatrexed treated LNCaP cells. Fold changes in P-RPA32 and p53 are quantified. C Immunoblotting for DNA damage response and p53 targets in cells treated with raltitrexed and/or navitoclax. D Immunoblotting for p21 in LNCaP cells treated with thymidylate synthase inhibitors with and without thymidine rescue. E LNCaP cells were treated with 5-FU for 2 days followed by navitoclax for 6 hand caspase activation was assessed. Mean and SEM for 6 biological replicates are shown. Data at each 5-FU concentration were analyzed by unpaired t-test, *p < .05. Analysis by two-way ANOVA showed significant effect of 5-FU, p < 0.001. F Immunoblotting for MCL-1, p21, and survivin in LNCaP cells treated with 5-FU for 1–2 days. G LNCaP cells were treated with Nutlin-3a (MDM2 inhibitor) for 1–2 days. Left panel: LNCaP cells were treated with Nutlin-3a for 1–2 days followed by immunoblotting. Middle and left panels: LNCaP cells were treated for 24 h with Nutlin-3a followed by 6 h with navitoclax (middle panel) or AZ4320 (right panel) and assessed for apoptosis by Caspase Glo 3/7 assay. Mean and SEM for 3 biological replicates are shown. Data at each Nutlin-3 concentration were analyzed by unpaired t-test, *p < .05. Analysis by two-way ANOVA showed effects were significant (p = 0.006 for left panel and p = 0.0013 for right panel). H LNCaP and A549 cells were treated with TP53 or nontarget control siRNA. The effect of 5-FU on the p53 pathway and integrated stress response pathway was assessed with immunoblotting. I LNCaP cells were treated with p21 or nontarget control siRNA. The effect of 5-FU on survivin expression was then assessed with immunoblotting. Source data are provided as a Source Data file.

Fig. 7. Combined inhibition of thymidylate synthase and BCL-XL is effective in vivo.

A LNCaP xenografts were developed in immunodeficient male mice. Mice were randomized to control, raltitrexed (50 mg/kg q4 days), navitoclax (50 mg/kg q2 days), or the combination as indicated in schema. Actual tumor size measurements are shown on left. Individual tumor size changes at 14 days relative to baseline (tumor size at randomization) are shown on right panel. Five (5) biological replicates are shown. Data were analyzed by Wilcoxon Rank Sum Test ***p < 0.001. Upper panel was created in Biorender. Yuan, X. (2025) https://BioRender.com/n41n675B Tumor-bearing mice were treated as in (A) for 14 days were monitored for tumor growth and euthanized when tumors reached 2.0 cm³. Graph shows Kaplan-Meier curve for mice survival from randomization. Data were analyzed by logrank test. C Tumors in another cohort were collected 6 h after the second raltitrexed dose (combination treated mice received 2 navitoclax doses). Apoptosis markers (cleaved PARP and cleaved caspase 3, CC3) were assessed with immunoblotting. D Tumors were collected 6 h after second raltitrexed dose and survivin protein expression was assessed with immunoblotting. Fold change in survivin is quantified. E Immunoblotting for MCL-1 and survivin in A549 (lung adenocarcinoma) and ZR75 (breast cancer) cells treated with 5-FU. Fold change in survivin is quantified. F ZR75 cells were treated with 5-FU for 48 h followed by navitoclax for 6 h, and apoptosis was assessed by CaspaseGlo 3/7 assay. Mean and SEM for 3 biological replicates are shown. Data at each 5-FU concentration were analyzed by unpaired t-test, *p < .05. Analysis by two-way ANOVA showed raltitrexed significantly enhanced apoptosis, p < .001. G ZR75 xenografts were developed in immunodeficient mice. Mice were randomized to control, capecitabine (200 mg/kg daily), navitoclax (50 mg/kg q2 days), or combination therapy as indicated in schema. Actual tumor size measurements are shown in the left panel. Individual tumor size changes at 28 days relative to baseline are shown in right panel. Five biological replicates are shown. Data were analyzed by Wilcoxon Rank Sum Test, ***p < 0.001. Upper panel was created in Biorender. Yuan, X. (2025) https://BioRender.com/obtzw73H Tumor-bearing mice were treated for 28 days. After that, mice were monitored for tumor growth as described above. Graph shows Kaplan-Meier curve for mice survival from randomization. Data were analyzed by logrank test. I Tumors were collected 6 h after the fourth capecitabine dose, and survivin protein expression was assessed with immunoblotting. Fold change in survivin is quantified. Source data are provided as a Source Data file.