EHMT1/2 Inhibition Promotes Regression of Therapy-Resistant Ovarian Cancer Tumors in a CD8 T-cell-Dependent Manner

Abstract

Poly ADP-ribose polymerase inhibitors (PARPi) are first-line maintenance therapy for ovarian cancer and an alternative therapy for several other cancer types. However, PARPi-resistance is rising, and there is currently an unmet need to combat PARPi-resistant tumors. Here, we created an immunocompetent, PARPi-resistant mouse model to test the efficacy of combinatory PARPi and euchromatic histone methyltransferase 1/2 inhibitor (EHMTi) in the treatment of PARPi-resistant ovarian cancer. We discovered that inhibition of EHMT1/2 resensitizes cells to PARPi. Markedly, we show that single EHMTi and combinatory EHMTi/PARPi significantly reduced PARPi-resistant tumor burden and that this reduction is partially dependent on CD8 T cells. Altogether, our results show a low-toxicity drug that effectively treats PARPi-resistant ovarian cancer in an immune-dependent manner, supporting its entry into clinical development and potential incorporation of immunotherapy. Implications: Targeting the epigenome of therapy-resistant ovarian cancer induces an antitumor response mediated in part through an antitumor immune response.

Figure 1:. An immunocompetent model of PARP inhibitor resistant ovarian cancer.

A) ID8 Trp53−/−, Brca2−/− sensitive (ID8) and Olaparib resistant (ID8-OR, red line) treated with increasing doses of Olaparib. B) Volcano plot of ID8-OR transcriptome compared to ID8 cells. C) In vivo evaluation of tumor burden 35 days after cell implantation (N=5). ID8 and ID8-OR tumors were fixed and used for evaluation via multispectral IHC. The density of different cell types is graphed. D) B cells via B220+ cells, E) macrophages via F4/80+ cells, F) non-T regulatory cells via CD3+, FOXP3−, G) T regulatory cells via CD3+, FOXP3+, H) dendritic cells via CD11c+. Error bars, SEM. Statistical test, unpaired t-test with Welch’s correction.

Figure 2:. EHMT inhibition resensitizes PARPi-resistant cells to Olaparib and induces interferon signaling and transposable element expression when combined with PARP inhibition.

A) Protein from ID8 and ID8-OR used for immunoblotting against EHMT1, EHMT2, H3K9me2, and H3K27me3. Loading controls, Actin and Histone H3. B) Quantification of immunoblots from A. C) ID8-OR cells treated with increasing doses of UNC0642 and protein blotted against H3K9me2. Loading control, Histone H3. D) ID8-OR cells treated with control (DMSO, black lines) or EHMT inhibitor (EZM8266, red lines) and increasing doses of Olaparib for 12 days. E) Volcano plot of differentially regulated genes in ID8-OR cells treated with Olaparib/UNC0642 compared to vehicle control (red dots, adj. p-value<0.05). F-H) Hallmark geneset enrichment plot for F) Interferon-α response, G) Interferon-γ response, and H) Inflammatory response in ID8-OR cells treated with combination compared to control. I) Heatmap of significantly differentially regulated genes in DMSO control (Ctrl) and Olaparib/UNC0642 (O+U) in the inflammatory pathway. J) Volcano plot of differentially regulated transposable elements in ID8-OR cells treated with Olaparib/UNC0642 compared to DMSO control (red dots, adj. p-value<0.05). K) Immunofluorescence imaging of dsRNA in ID8-OR cells treated with DMSO control, single, and combinatory Olaparib and/or UNC0642. L) Quantification of dsRNA immunofluorescence in panel K. Error bars, SEM. Statistical test, unpaired t-test (B). Error bars, SEM. Statistical test, unpaired t-test with Welch’s correction and Log-rank. *p<0.05 (D); Error bars, SEM. Statistical test, multicomparison ANOVA with Tukey correction. **p<0.01, ****p<.0001 (L).

Figure 3:. Targeting EHMT in an immunocompetent, PARPi-resistant, in vivo model.

A) Study design and timeline. Population size is indicated by treatment group (N = 9 or 10). B) Tumor progression monitored via total flux (photons/sec). C) Omental weight. D) Number of tumor disseminated tumor nodules detected. E) Weight of solid tumors. F) Histological score (H-score) of formalin fixed tumor used for IHC against cleaved caspase 3 (CC3). G) Histological score (H-score) of formalin fixed tumor used for IHC against Ki67. H) IHC against H3K9me2. I) Percent H3K9me2 positive cells in tumor compartment. J) Percent H3K9me2 positive cells in non-tumor compartment. Error bars, SEM. Statistical test, mixed model effect and multicomparison ANOVA with Tukey correction. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Figure 4:. Immune milieu of the tumor microenvironment is transformed by combinatory EHMTi/PARPi therapy.

A) Heatmap of multiplex ELISA for indicated cytokines in ascites fluid collected from treated tumor-bearing mice. B) Graphical data of TNF-α concentration. C) Graphical data of CCL5. Tumors from treated mice were used for mIHC and 5 ROI were selected from each tumor ((N=9 or 10, n=5-14). Tumor infiltration of D) NK cells, E) CD8+ cells, F) CD4+ cells, G) CD4+FOXP3+ cells, and H) F4/80+ cells. Error bars, SEM. Statistical test, multicomparison ANOVA with Tukey correction. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Figure 5:. EHMT inhibition induces a T cell dependent anti-tumor response.

A) Study design and population size indicated next to treatment (n=9 or 10/group). B) Tumor progression measured via total flux. C) Representative images of luminescence in tumor-bearing mice at the end of study for each treatment group. D) Percentage of splenic CD8a+ T cells at the end of study. E) Tumor burden measured as total flux on Day 28. F) Solid tumor weight. G) Number of disseminated tumor nodules. Error bars, SEM. Statistical test, multicomparison ANOVA with Tukey correction (E, I-L). *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.