Oncolytic adenovirus MEM-288 encoding membrane-stable CD40L and IFNβ induces an anti-tumor immune response in high grade serous ovarian cancer

Abstract

Single agent immune checkpoint inhibitors have been ineffective for patients with advanced stage and recurrent high grade serous ovarian cancer (HGSOC). Using pre-clinical models of HGSOC, we evaluated the anti-tumor and immune stimulatory effects of an oncolytic adenovirus, MEM-288. This conditionally replicative virus encodes a modified membrane stable CD40L and IFNβ. We demonstrated this virus successfully infects HGSOC cell lines and primary human ascites samples in vitro. We evaluated the anti-tumor and immunostimulatory activity in vivo in immune competent mouse models. Intraperitoneal delivery of MEM-288 decreased ascites and solid tumor burden compared to controls, and treatment generated a systemic anti-tumor immune response. The tumor microenvironment had a higher proportion of anti-tumor macrophages and decreased markers of angiogenesis. MEM-288 is a promising immunotherapy agent in HGSOC, with further pre-clinical studies required to understand the mechanism of action in the peritoneal microenvironment and clinical activity in combination with other therapies.

Keywords: Ascites; ELISA; ELISPOT; Immunotherapy; Intraperitoneal; Microenvironment; Oncolytic virus; Ovarian cancer.

Fig. 1

MEM-288 infection causes transgene expression in human ovarian cancer cell lines and human ascites samples. (A-B) TykNu cells demonstrated elevated concentrations of (A) IFNβ (p = 0.00001) and (B) expression of CD40L (p = 0.0001) after exposure to MEM-288, compared to Adv-GFP and saline exposure. (C-D) Jhos2 cells demonstrated elevated concentrations of (A) IFNβ (p = 0.00001) and (B) expression of CD40L (p = 0.0001) after exposure to MEM-288, compared to Adv-GFP and saline exposure. (E-F) Primary human ascites samples demonstrated elevated concentrations of (E) IFNβ (p < 0.0001) and (F) Expression of CD40L (p < 0.0001) after exposure to MEM-288, compared to Adv-GFP and saline exposure. (G) Human ascites samples containing both tumor and immune cells demonstrated increased concentrations of IFNγ, a marker of T-cell activation when exposed to MEM-288, compared to Adv-GFP or saline (p = 0.0002). All error bars represent standard error of the mean (SEM). All experiments were performed in triplicate or quadruplicate. Statistical significance was defined as p ≤ 0.05 (*), p ≤ 0.01 (**), p ≤ 0.001 (***), p ≤ 0.0001 (***).

Fig. 2

Anti-tumor and immune stimulatory activity of MEM-288 in vivo. (A-C) Two weeks after intraperitoneal treatment with saline, control oncolytic virus (Adv-GFP), or MEM-288, mice were assessed for tumor burden as measured by (A) ascites volume, (B) tumor weight, and (C) number of metastatic sites. (D) Generation of a systemic anti-tumor immune response was evaluated by IFNγ ELISA, measuring number of tumor-reactive splenocytes in mice treated with saline, Adv-GFP, or MEM-288. (E) Representative wells from IFNγ ELISPOT with each column representing a biologic replicate. Splenocytes were placed in duplicate (not shown) with irradiated STOSE-luc target cells in sample wells. Positive control wells contained splenocytes with a non-specific T-cell stimulator (PMA/I, phorbol myristate acetate/ ionomycin). Negative control wells contained splenocytes with media. Number of spots per well was quantified, with each spot representing a tumor-reactive T-cell. (F) Representative images are shown from imaging experiments. Mice were monitored with in vivo imaging at least weekly from time of cell injection (day 0) to experimental endpoint (day 27). The experimental endpoint was selected to be 2 weeks after treatments were administered as the control mice (saline, Adv-GFP) approached humane endpoints, with large abdominal ascites and tumor burden on necropsy. Error bars represent standard error of the mean (SEM) with each point representing a biologic replicate. Statistical significance was defined as p ≤ 0.05 (*), p ≤ 0.01 (**), p ≤ 0.001 (***), p ≤ 0.0001 (***).

Fig. 3

MEM-288 and Adv-GFP altered the tumor microenvironment in vivo. (A-I) Proportions of cell populations were assessed with flow cytometry. There was no significant difference in (A) immune cells (CD45+) between treatment groups. There was a significant increase in proportion of (B) CD4+ T-cells with MEM-288 vs saline (p = 0.004) and Adv-GFP vs saline (p = 0.02). (C) CD8+ T-cells were increased in Adv-GFP compared to saline (p < 0.0001) and MEM-288 (p = 0.003). Proportions of (D) CTLA4+ and PD1+ CD8 T-cells were decreased with MEM-288 versus saline (p < 0.0001) and Adv-GFP versus saline (p = 0.02) but not significantly different between Adv-GFP and MEM-288. MEM-288 treatment did not affect proportions of (E) dendritic cells or (F) neutrophils compared to saline. Proportions of (G) resident monocytes were lower with MEM-288 treatment compared to Adv-GFP (p = 0.01) and saline (p < 0.0001) with no difference in (H) inflammatory monocytes. The (I) ratio of M1 “anti-tumor” macrophage phenotype: M2 “pro-tumor” macrophage phenotype was increased with MEM-288 treatment versus Adv-GFP (p = 0.05) and saline (p < 0.0001). (J-M) Immunohistochemistry (IHC) was used to quantify (J) CD3+ and (K) CD8+ cells as an average number of positive staining cells per high power field. There was no difference in CD3 or CD8 staining between treatment groups. (L) CD31 staining, defined as the number of vessels per high power field staining positive for CD31, was decreased with MEM-288 treatment versus Adv-GFP (p = 0.02). (M) Ki-67 index was not different between treatment groups. Error bars represent standard error of the mean (SEM) with each point representing a biologic replicate. Statistical significance was defined as p ≤ 0.05 (*), p ≤ 0.01 (**), p ≤ 0.001 (***), p ≤ 0.0001 (***).

Fig. 4

Efficacy of MEM-288 in vivo in epithelial ovarian cancer mouse models (A) STOSE-luc cells (1 × 10⁶) were injected intraperitoneal (i.p.) and tumor formation confirmed via in vivo imaging on day 10. Mice were randomized to MEM-288 or Adv-GFP i.p. treatment was administered on days 15, 18, 46, and 49. Median survival was 51 days for Adv-GFP and 51 days for MEM-288 (p = 0.24). (B) IG10-luc cells (2 × 10⁶) were injected intraperitoneal (i.p.) and tumor formation confirmed via in vivo imaging. Mice were randomized to MEM-288 or Adv-GFP i.p. treatment was administered on days 46 and 49. Median survival was 106 days for Adv-GFP and 110 days for MEM-288 (p = 0.018). Significance was assessed using the log-rank (Mantel-Cox) tests. Statistical significance was defined as p ≤ 0.05 (*), p ≤ 0.01 (**), p ≤ 0.001 (***), p ≤ 0.0001 (***).