Co-targeting of epigenetic regulators and BCL-XL improves efficacy of immune checkpoint blockade therapy in multiple solid tumors

Abstract

Epigenetic modulators in combination with proapoptotic drugs have become the standard of care treatment in hematological malignancies. Conversely, these combinations have failed to demonstrate clinical efficacy in solid tumors. To address this discrepancy, we conducted a comprehensive analysis of the anti-tumor activity of epigenetic inhibitors in combination with BH3 mimetics that block anti-apoptotic proteins BCL-XL, BCL2 or MCL1 in a large set of solid tumor cell lines derived from patients and mouse models. Treatment with epigenetic drugs targeting DNA methyltransferase, histone methyltransferase, and histone deacetylase enzymes in combination with a BCL-XL inhibitor resulted in marked synergistic in vitro responses both in human and mouse solid tumor cell lines. This unique BCL-XL dependency was in clear contrast to hematological malignancies, which are largely dependent on BCL2 or MCL1 inhibition under epigenetic drug treatment. Mechanistically, co-targeting of epigenetic regulators and BCL-XL induced expression of endogenous retroelements that led to immunogenic cell death. We thus hypothesized that this response may sensitize tumor cells to immune checkpoint blockade (ICB). Accordingly, treatment with a triple combination of epigenetic and BCL-XL inhibitors with an anti-PD-1 monoclonal antibody in vivo reduced tumor growth and prolonged overall survival in a panel of murine syngeneic and orthotopic models of lung, colorectal and breast carcinomas, melanoma, and glioblastoma, as well as in an immunocompetent human colon cancer model. Using flow cytometry and single-cell RNA sequencing of the tumor microenvironment, we found that the broad activity of the triple therapy relied on the expansion of T and NK cells with cytotoxic potential, an increase in the M1/M2 macrophage ratio, and a reduction of immunosuppressive Treg cells, dendritic cells, and B lymphocytes. In conclusion, we report a novel regimen combining epigenetic and BCL-XL inhibitors with ICB that produces potent anti-tumor responses in multiple preclinical models of solid tumors.

Keywords: A1331852; Anti-PD-1; Anti-apoptotic proteins; BCL-XL; CM272; Epigenetic modulators; Immune checkpoint blockade; Solid tumors.

Fig. 1

Epigenetic drug-mediated tumor cell death is enhanced by concomitant BCL-XL inhibition. A. Heatmap of the induced expression of endogenous retroviruses (ERVs), helicases and interferon-stimulated genes (ISGs) after treatment of human lung cancer cell lines with the epigenetic drug CM272. Expression was measured by real-time PCR. B. Heatmap of the induced expression of ERVs, helicases and ISGs after treatment of mouse lung cancer cell lines with the epigenetic drugs azacitidine, vorinostat or CM272. Expression was measured by real-time PCR. C. ADP/ATP ratio measured in human (A549, H460) and mouse (LLC, 393P) lung cancer cell lines treated with CM272 at IC₅₀ and IC₇₀. D. Fold change in oxygen consumption rate (OCR) measured in human (A549, H460) and mouse (LLC, 393P) lung cancer cell lines treated with CM272 at IC₅₀ and IC₇₀. E. Schematic representation of the synergistic anti-tumor effect of epigenetic and proapoptotic drugs. F. Combination index (CI) values and cell viability curves obtained after treatment of human (H226) and mouse (LLC) lung cancer cells with the epigenetic drug CM272 (at its IC₅₀) in combination with the proapoptotic drugs against BCL-XL (A1331852), BCL2 (venetoclax) or MCL-1 (S63845) at increasing concentrations. A synergistic effect between the two drugs was considered when at least four drug-drug interactions had CI values < 1. G. Relative IC₅₀ values obtained when proapoptotic drugs were combined with the epigenetic drug CM272 (at its IC₅₀) in human and mouse lung cancer cell lines. Data are expressed as mean ± SEM. Student’s t-test was used for statistical analysis in C and D

Fig. 2

A. Left panel: Representative Annexin V/PI staining of human (H460) and mouse (393P) lung cancer cells following treatment with the epigenetic drug CM272 and the BCL-XL inhibitor A1331852. Right panel: Quantification of late apoptotic cells (high Annexin V and PI staining) from three independent experiments. B. Representative activated caspase 3/7 and 7-ADD staining of human (H460) and mouse (393P) lung cancer cells following treatment with the epigenetic drug CM272 and the BCL-XL inhibitor A1331852. Right panel: Quantification of high active caspase 3/7 and 7-ADD staining from three independent experiments. C. Western blot analysis of caspase-3 cleavage in human and mouse lung cancer cell lines treated with CM272 and A1331852 alone or in combination. D. Calreticulin staining of human (H460 and A549) and mouse (393P and LLC) lung cancer cells following treatment with the epigenetic drug CM272 and the BCL-XL inhibitor A1331852. The graphs are representative of the results of two independent experiments. Treatment conditions were: H460: 5 µM CM272/2.5 µM A1331852; A549: 15 µM CM272, 5 µM A1331852; 393P and LLC: 30 µM CM272/2.5 µM A1331852. Cells were treated for 24 h. Data are presented as mean ± SEM. Statistical comparisons between treatment groups were conducted using one-way ANOVA followed by Tukey’s post hoc test

Fig. 3

The epigenetic/proapoptotic drug combination CM272/A1331852 sensitizes lung tumors to anti-PD-1 therapy. A. Schematic representation of the hypothetical synergy between epigenetic drugs, BCL-XL inhibition and immunotherapy in vivo. B. Schematic representation of the therapeutic regimens followed in the cancer model based on subcutaneous inoculation of 393P lung cancer cells into syngeneic mice. The therapeutic regimens of the other models presented in this figure, Lacun.3 and LLC, are shown as supplementary material (Supplementary Fig. 9). C. Mice bearing 393P, Lacun.3 or LLC tumors were treated intraperitoneally with vehicle (control), CM272, A1331852 and/or anti-PD-1 as shown in B. Eight mice were used per group. Left panels: tumor size monitoring. Middle panels: tumor volumes on the last day of follow-up. Right panels: overall survival. D. Orthotopic LLC tumors treated with vehicle (control) or the triple combination (CM272/A1331852/anti-PD-1). Tumor growth was measured on day 14 after cell inoculation by bioluminescence (left) and presented as bio-layer interferometry (BLI) values (middle). Overall survival is shown in on the right. Eight mice were used in each group. Data are expressed as mean ± SEM. Comparisons between treatment strategies were performed by one-way ANOVA with Tukey’s post hoc test, except for D, where Student’s t-test was used. Survival curves were compared using the log rank test

Fig. 4

The epigenetic drug CM272 synergizes with proapoptotic drugs in vitro and in vivo in a variety of solid tumors. A. Reduction of IC₅₀ of proapoptotic drugs against BCL-XL (A1331852), BCL2 (venetoclax) or MCL-1 (S63845) in combination with the epigenetic drug CM272 in human cancer cell lines of melanoma, glioblastoma, colorectal cancer, breast cancer and pancreatic cancer. B. Reduction of IC₅₀ of proapoptotic drugs in combination with the epigenetic drug CM272 in mouse cancer cell lines of melanoma, glioblastoma, colorectal cancer, breast cancer and pancreatic cancer. C. Subcutaneous growth monitoring of KPC pancreatic tumors treated intraperitoneally with vehicle (control) or CM272 (five days per week), A1331852 (three days per week), and anti-PD-1 (two days per week). Eight mice were used in each group. D. Subcutaneous growth monitoring of MC38 colorectal tumors treated as above. E. Subcutaneous growth monitoring of B16.F10 melanoma tumors treated as above. F. Orthotopic growth monitoring of ANV5 breast tumors treated as above. G. Survival of mice bearing orthotopic CT-2 A glioblastomas treated as above. H. Left: humanized mouse model of subcutaneous HT-29 tumors. Right: percentage volume change from baseline at the end of the experiment is shown for each humanized mouse. Mice were treated with vehicle, CM272 (5 days per week starting on day 11), pembrolizumab (days 12, 15, 18), and/or A1331852. Five mice per group were used. Data are expressed as mean ± SEM. Comparisons between treatment strategies were carried out by Student's t-test analysis, except for H, where one-way ANOVA with Tukey’s post hoc test was used. Survival curves were compared using the log rank test

Fig. 5

The anti-tumor activity of the triple combination against lung tumors is associated with a reduction of immunosuppressive populations in the TME. A. LLC tumor-bearing mice were treated intraperitoneally with vehicle (control) or CM272 (five days per week)/A1331852 (three days per week)/anti-PD-1 (two days per week) in the presence or absence of the indicated depleting antibodies (days 6, 10, 14 and 18). Six mice were used per experimental group. Upper panel: follow-up of tumor size. Bottom panel: tumor volumes at the end of the experiment on day 18 post inoculation. B. Flow cytometry analysis of tumor-infiltrating lymphoid immune cells performed on day 15 after implantation of LLC cells in mice treated as above with vehicle (control), CM272, A1331852 and/or anti-PD-1. Eight mice were used per experimental group. Values are expressed as percentage of CD45⁺ cells. Data are expressed as mean ± SEM. C. As in B, for cells in the myeloid compartment. D. Multiplex immunophenotyping in the orthotopic LLC tumors treated as above. Tumors were harvested at day 14 post-inoculation. Left: Representative merged immunofluorescence images for the quantification of CD4 (green), CD8 (yellow), FoxP3 (pink) and nucleus (DAPI) in panel 1; and CD86 (green), F4/80 (red) and nucleus (DAPI) in panel 2. Right: Quantification of Treg cells (CD4⁺ FoxP3⁺ cells) as a percentage of total CD4⁺ cells and of M1-like macrophages (F4/80⁺ CD86⁺ cells). Comparisons were made by one-way ANOVA with Tukey's post hoc test in A, B and C and by Student's t-test in D

Fig. 6

A scRNA-seq analysis reveals that the anti-tumor activity of the triple combination therapy is associated with a decrease in immunosuppressive subpopulations and an increase in cytotoxic subpopulations in the lymphoid and myeloid compartments of the TME. LLC tumor-bearing mice were treated with the triple combination CM272/A1331852/anti-PD-1. Three mice were used per experimental group. A. Two-dimensional t-distributed stochastic neighbor embedding (t-SNE) plot showing cell clusters of immune cells in the TME of LLC tumors. B. Prevalence of each cell type estimated by Ro/e score. C. t-SNE plot showing the sub-clusters of T cells. D. Prevalence of each T-cell subtype estimated by Ro/e score. E. Dot plot showing the expression of genes associated with effector or thymic-derived lung resident Treg cells. F. t-SNE plot showing the sub-clusters of NK cells. G. Prevalence of each NK subtype estimated by Ro/e score. H. Dot plot showing the expression of cytotoxic-associated genes in the sub-clusters of NK cells. I. t-SNE plot showing the sub-clusters of monocyte/macrophages (Mo/MØs). J. Prevalence of each subtype of Mo/MØs estimated by the Ro/e score. K. t-SNE plot showing the expression of genes associated with an M2 phenotype. L. t-SNE plot showing the sub-clusters of DCs. M. Prevalence of each subtype of DCs estimated by the Ro/e score. N. Dot plots showing the expression of genes associated with tumor promotion and immunosuppression (top panel), or the expression of genes associated with antigen presentation (bottom panel) in the sub-clusters of DCs