The antitumor effects of entinostat in ovarian cancer require adaptive immunity

Abstract

Background: Ovarian cancer is poorly immunogenic; however, increased major histocompatibility complex class II (MHCII) expression correlates with improved immune response and prolonged survival in patients with ovarian cancer. The authors previously demonstrated that the histone deacetylase inhibitor entinostat increases MHCII expression on ovarian cancer cells. In the current study, they evaluated whether entinostat treatment and resultant MHCII expression would enhance beneficial immune responses and impair tumor growth in mice with ovarian cancer.

Methods: C57BL/6 mice bearing intraperitoneal ID8 tumors were randomized to receive entinostat 20 mg/kg daily versus control. Changes in messenger RNA (mRNA) expression of 46 genes important for antitumor immunity were evaluated using NanoString analysis, and multicolor flow cytometry was used to measure changes in protein expression and tumor-infiltrating immune cells.

Results: Entinostat treatment decreased the growth of both subcutaneously and omental ID8 tumors and prolonged survival in immunocompetent C57BL/6 mice. NanoString analysis revealed significant changes in mRNA expression in 21 of 46 genes, including increased expression of the MHCI pathway, the MHCII transactivator (CIITA), interferon γ, and granzyme B. C57BL/6 mice that received entinostat had increased MHCII expression on omental tumor cells and a higher frequency of tumor-infiltrating, CD8-positive T cells by flow cytometry. In immunocompromised mice, treatment with entinostat had no effect on tumor size and did not increase MHCII expression.

Conclusions: In the current murine ovarian cancer model, entinostat treatment enhances beneficial immune responses. Moreover, these antitumor effects of entinostat are dependent on an intact immune system. Future studies combining entinostat with checkpoint inhibitors or other immunomodulatory agents may achieve more durable antitumor responses in patients with ovarian cancer.

Keywords: adaptive immunity; entinostat; major histocompatibility complex class II (MHCII); ovarian cancer.

Figure 1:. Entinostat (ENT) inhibits tumor growth and prolongs survival in a syngeneic murine ovarian cancer model.

A) ENT treatment led to a significant decrease in tumor growth compared to vehicle in a subcutaneous ID8 model by day 11 of treatment (19.9 vs. 32.9 mm², p=0.046), which persisted at day 14 (19.0 vs. 35 mm², p=0.022). Treatment was started at 32 days post tumor implantation. Data shown is representative of 3 experiments. Error bars represent standard deviation. B) In an intraperitoneal ID8 model, ENT treatment led to a significant decrease in omental tumor weight compared to vehicle (44.7 vs. 100.4 mg, p<0.001), and fewer ENT-treated mice had ascites at time of tumor harvesting. Data shown is representative of 2 experiments. C) ENT treatment significantly prolonged overall survival (OS) compared to vehicle in an intraperitoneal ID8 model (median OS 74 vs 62 days, p=0.005, HR 0.26 (95% CI 0.06–1.21). Mice were treated from day 28 to 42 (indicated by shaded portion of graph). n=5 mice per group. *indicates p<0.05

Figure 2:. By NanoString^® analysis, entinostat (ENT) induces significant changes in mRNA expression of multiple genes involved in anti-tumor immunity.

C57Bl/6 mice were treated with either ENT or vehicle control for 14 days (n=4 per group). NanoString^® analysis was performed on omental tumors. A) In a selected panel of 50 genes specific to the adaptive immune response, unsupervised hierarchical clustering shows distinct gene expression patterns in omental tumors from C57Bl/6 mice treated with ENT compared to vehicle controls. Green indicates downregulated genes, while red indicates upregulated genes. B) ENT treatment led to a significant increase in expression of T cell effector molecules, including interferon gamma (8.5 vs. 2.9, p=0.039), granzyme B (125.4 vs. 9.5, p=0.008), and CXCL10 (1,636.4 vs. 417.0, p=0.023). C) ENT-treated mice showed significantly increased expression of the immune checkpoint molecule PDL1 (1,090.6 vs. 294.2, p=0.016) and significant decreases in immune checkpoints LAG3 (78.0 vs. 106.5, p<0.001) and BTLA (26.6 vs. 65.11, p=0.017). D) ENT treatment significantly increased expression of CIITA (267.4 vs. 132.1, p=0.011) and the MHCI pathway (20,728.3 vs. 10,272.8, p=0.013). There was a trend towards increased MHCII pathway expression, although this was not statistically significant (51,667.3 vs. 39,683.12, p=0.098). H2-Kb and I-Ab alpha chain are shown and are representative of the other genes in the MHCI and MHCII pathways, respectively. *indicates p<0.05

Figure 3:. Entinostat (ENT) increases MHCII and PDL1 expression on omental tumor cells.

Gating was on single, viable, CD45 negative cells. Expression was compared using mean fluorescent intensity (MFI). A) ENT treatment led to significant increase in MHCII expression (1,585.0 vs. 581.4 MFI p=0.002). B) ENT treatment led to a near-significant increase in PDL1 expression (2,832.0 vs. 1,593.0, p=0.052). n=5 mice per group. *indicates p<0.05.

Figure 4:. Entinostat (ENT) increases recruitment of CD8+ T cells into the tumor.

A) Gating strategy for identification of CD8+ and CD4+ T cells and CD25+FoxP3+ T cells (Tregs). Initial gating was on single, viable, CD45+ cells. B) Entinostat treatment increased the percentage of CD8+ T cells in omental tumors of C57bl/6 mice (21.0% vs. 18.0%, p=0.039). There was no difference in percentage of CD4 T cells (27.4% vs. 29.4%, p=0.255) or Tregs (4.7% vs. 3.4%, p=0.075). n=5 mice per group. *indicates p<0.05.

Figure 5:. Entinostat (ENT) does not significantly inhibit tumor growth or increase MHCII or PDL1 expression in Rag1 knockout mice.

A) In Rag1 knockout mice, there was no difference in omental tumor weight between mice treated with ENT versus vehicle control (54.3 vs. 79.8 mg, p=0.093) and no difference in ascites formation. B) There was no difference in either MHCII (1,250.0 vs. 1,483.0 MFI, p=0.152) or PDL1 expression (805.8 vs. 706.0 MFI, p=0.258) in Rag1 knockout mice treated with ENT versus vehicle control. n=6 mice per group.