Oncolytic VSV Primes Differential Responses to Immuno-oncology Therapy

Abstract

Vesicular stomatitis virus encoding the IFNβ transgene (VSV-IFNβ) is a mediator of potent oncolytic activity and is undergoing clinical evaluation for the treatment of solid tumors. Emerging preclinical and clinical data suggest treatment of tumors with oncolytic viruses may sensitize tumors to checkpoint inhibitors and increase the anti-tumor immune response. New generations of immuno-oncology molecules including T cell agonists are entering clinical development and could be hypothesized to enhance the activity of oncolytic viruses, including VSV-IFNβ. Here, we show that VSV-IFNβ exhibits multiple mechanisms of action, including direct cell killing, stimulation of an innate immune response, recruitment of CD8 T cells, and depletion of T regulatory cells. Moreover, VSV-IFNβ promotes the establishment of a CD8 T cell response to endogenous tumor antigens. Our data demonstrate a significant enhancement of anti-tumor function for VSV-IFNβ when combined with checkpoint inhibitors, but not OX40 agonists. While the addition of checkpoint inhibitors to VSV-IFNβ generated robust tumor growth inhibition, it resulted in no increase in viral replication, transgene expression, or immunophenotypic changes beyond treatment with VSV-IFNβ alone. We hypothesize that tumor-specific T cells generated by VSV-IFNβ retain activity due to a lack of immune exhaustion when checkpoint inhibitors were used.

Keywords: VSV; agonist; checkpoint inhibitor; immune-oncology; oncolytic virus.

Figure 1

VSV Encoding IFNβ Transgene Has Oncolytic Properties (A) 29 human tumor cell lines derived from multiple histological origins were infected with VSV-hIFNβ, herpes simplex virus 1 (HSV-1), or adenovirus type 5 (Ad5) at MOIs from 10 to 0.00001 and assayed 72 hrs post infection for cell viability. EC₅₀ values were calculated using a variable slope log (agonist) versus response algorithm. (B) Murine cell lines infected VSV-mIFNβ, HSV-1, or Ad5 as in (A). (C and D) Single-step (MOI = 3.0, n = 3) or (D) multi-step (MOI = 0.003, n = 3) growth kinetics of VSV-mIFNβ on BHK, B16-F10, and CT26 cell lines reported as the TCID₅₀/mL over time (hr post infection).

Figure 2

Single-Agent Activity of VSV-mIFNβ in Syngeneic Mouse Tumor Models (A) Spider plots of B16-F10 tumor volume for individual mice treated with sham, HI-VSV-IFNβ, or VSV-IFNβ. Mice were randomized based on a mean tumor volume of 200 mm³ on day 0. Mice were dosed intratumorally (IT) in a volume of 30 μL twice weekly for a total of four doses of sham, HI VSV-IFNβ, or 1 × 10⁹ TCID₅₀ of VSV-IFNβ. (B) Kaplan-Meier survival plots of animals in (A). The significance was determined by log rank (Mantel-Cox) test. (C) Spider plots of CT26 tumor volume for individual mice treated with sham, HI-VSV-IFNβ, or VSV-IFNβ. Mice were randomized based on a mean tumor volume of 200 mm³ on day 0. Mice were dosed IT in a volume of 30 μL twice weekly for a total of four doses of sham, HI VSV-IFNβ, or at 1 × 10⁹ TCID₅₀ of VSV-IFNβ. (D) Kaplan-Meier survival plots of animals in (C). ***p < 0.001 and *p < 0.05; NS, not significant.

Figure 3

VSV-mIFNβ Affects Immune Cell Populations within the Tumor Microenvironment and Induces Tumor- and Viral-Specific T Cell Responses (A–C) Flow cytometry analysis of the percentage of (A) intratumoral CD45⁺ cells in B16-F10 and CT26 untreated tumors (B) intratumoral CD45⁺CD8⁺ T cells (C) and intratumoral regulatory T cells (CD45⁺CD4⁺FoxP3⁺) in B16-F10 and CT26 tumors tested 48 hr after the first or fourth dose of test article administered to the mice. (D) Tumor and spleen were collected 48 hr after the fourth dose of test articles in B16-F10 tumor model, processed into single-cell suspensions, stimulated with tumor (GP100), virus (NP52), or irrelevant (Ova) peptides and analyzed by flow cytometry for cytokine production. The datasets were concatenated from five individual mice per group due to a low number of available events. *p < 0.05 and **p < 0.01.

Figure 4

VSV-mIFNβ Combines with Checkpoint Inhibitors to Elicit Diverse Anti-tumor Responses against B16-F10 Tumors (A) Spider plots are shown of the tumor volume for individual mice bearing B16-F10 tumors after VSV-mIFNβ monotherapy, checkpoint inhibitor monotherapy, or combination treatment. Mice were randomized based on a mean tumor volume of 200 mm³ on day 0. Next, mice were concurrently administered VSV-mIFNβ IT (1 × 10⁹ TCID₅₀) in a volume of 30 μL and/or i.p. with checkpoint inhibitors (anti-CTLA-4 or anti-PD-L1 mAbs) at doses of 20 mg/kg in a volume of 200 μL. The antibodies and/or virus were dosed to mice two times a week for a total of four doses. (B) Kaplan-Meier survival plots of animals in (A). The significance was determined by Mantel-Cox test. ***p < 0.001 and **p < 0.01; NS, not significant.

Figure 5

α-CTLA-4 Antagonist mAb Combines with VSV-mIFNβ Treatment to Elicit Increased Therapeutic Responses to CT26 Tumors (A) Spider plots are shown of the tumor volume for individual mice bearing CT26 tumors after VSV-mIFNβ monotherapy, checkpoint inhibitor monotherapy, or combination treatment. Mice were randomized based on a mean tumor volume of 200 mm³ on day 0. Next, mice were concurrently administered VSV-mIFNβ IT (1 × 10⁹ TCID₅₀) in a volume of 30 μL and/or i.p. with checkpoint inhibitors (anti-CTLA-4 at 20 mg/kg or anti-PD-L1 mAb at a dose of 10 mg/kg) in a volume of 200 μL. The antibodies were dosed two times a week for a total of four doses. (B) Kaplan-Meier survival plots of animals in (A). The significance was determined by log rank (Mantel-Cox) test. ***p < 0.001 and *p < 0.05; NS, not significant.

Figure 6

Immunophenotypic, Viral, and T Cell Response Changes Observed in the Tumors of Mice following Treatment with Checkpoint Blockade Combined with VSV-mIFNβ (A) Percentage of intratumoral CD8⁺ T cells in B16-F10 and CT26 tumors 24 hr after the final administration of VSV-mIFNβ, anti-PDL1 mAb, anti-CTLA-4 mAb alone, or in combination. (B) Percentage of intratumoral T-regulatory cells as a subset of CD45⁺CD4⁺ cells in B16-F10 or CT26 tumors 24 hr after the final administration of VSV-mIFNβ, anti-PDL1 mAb, anti-CTLA-4 mAb alone, or in combination. The percentage of ex vivo peptide stimulated CD8⁺ cells producing IFNγ, TNF-α, or both cytokines is shown. (C) Tumor cells from mice bearing B16-F10 tumors were collected 24 hr after fourth dose of sham, VSV-IFNβ, and mAbs specific for PD-L1 and CTLA4 alone, or in combination were cultured ex vivo with GP100 or NP52 peptides. The percentage of ex vivo peptide stimulated CD8⁺ cells producing IFNγ, TNF-α, or both cytokines is shown (cytokine analysis from n = 4 mice/group). (D) Tumor cells from mice bearing CT26 tumors were collected 24 hr after fourth dose of sham, VSV-mIFNβ, and mAbs specific for PD-L1 and CTLA4 alone, or in combination were cultured ex vivo with AH-1 or P622 peptides (cytokine analysis from n = 5 mice/group). *p < 0.05, **p < 0.01, and***p < 0.001; NS, not significant; ITC, isotype control.

Figure 7

OX40 Agonist mAb Combined with VSV-IFNβ Skews Early T Cell Responses to Viral, Not Tumor, Antigens in B16-F10 Bearing Mice (A) Spider plots are shown of the tumor volume for individual mice bearing B16-F10 after isotype control, VSV-mIFNβ monotherapy, α-OX40 agonist mAb monotherapy, or combination treatment. Mice were randomized based on a mean tumor volume of 200 mm³ on day 0. Next, mice were concurrently administered VSV-mIFNβ IT (1 × 10⁹ TCID₅₀) in a volume of 30 μL and/or i.p. with α-OX40 agonist mAb at a dose of 12.5 mg/kg in a volume of 200 μL. The antibodies were administered two times a week for a total of four doses. (B and C) Percentage of intratumoral CD45⁺CD8⁺ T cells (B) and intratumoral regulatory T cells (CD45⁺CD4⁺FoxP3⁺) in B16-F10 tumors as determined by flow cytometry 24 hr after the fourth dose of the indicated treatment regimen (C). (D and E) Percentage of ex vivo stimulated CD8⁺ cells producing IFNγ, TNF-α, or both cytokines. Single-cell suspensions of tumors were produced as described above in Figure 6, except isolation occurred 24 hr after both the second and the last dose administered to the mice. Suspensions were cultured with GP100 tumor antigen peptide (D) or NP52 VSV antigen peptides (E) and then CD8 T cells in the cultures were measured for cytokine production using a flow cytometry based system (antigen-specific analysis conducted on n = 5 mice/group). *p < 0.05 and ***p < 0.001; NS, not significant.

Figure 8

VSV-mIFNβ Induces Type I and II IFN Responses and an Increase in PD-L1 Expression on Tumor Cells (A and B) Concentrations of IFNβ (A) or IFNγ in sera at the indicated times (B) after the first (1) and fourth (4) dose of sham, HI VSV-mIFNβ, or 1 × 10⁹ TCID₅₀ VSV-mIFNβ in mice bearing B16-F10 and CT26 tumors. (C) PD-L1 expression measured on CD45⁻ cells from tumor homogenates after first or fourth dose of sham, HI VSV-mIFNβ, or 1 × 10⁹ TCID₅₀ VSV-mIFNβ in mice bearing B16-F10 (n = 4 mmice/group) and CT26 (n = 4 mice/group) tumors. (D) Analysis of major histocompatibility complex class I (MHC-I) expression levels on the cell surface of B16-F10 and CT26 cell lines 24 hr after infection with sham (media), or MOI = 0.03 of heat-inactivated (HI), or live VSV-mIFNβ. (E) Bar representation of MFI from (D). MFI, mean fluorescence intensity; *p < 0.05.