The addition of recombinant vaccinia HER2/neu to oncolytic vaccinia-GMCSF given into the tumor microenvironment overcomes MDSC-mediated immune escape and systemic anergy

Abstract

Effective immunotherapeutic strategies require the ability to generate a systemic antigen-specific response capable of impacting both primary and metastatic disease. We have built on our oncolytic vaccinia a granulocyte-macrophage colony-stimulating factor (GM-CSF) strategy by adding recombinant tumor antigen to increase the response in the tumor microenvironment and systemically. In the present study, orthotopic growth of a syngeneic HER2/neu-overexpressing mammary carcinoma in FVB/N mice (NBT1) was associated with increased Gr1(+)CD11b(+) myeloid-derived suppressor cells (MDSCs) both systemically and in the tumor microenvironment. This MDSC population had inhibitory effects on the HER2/neu-specific Th1 immune response. VVneu and VVGMCSF are recombinant oncolytic vaccinia viruses that encode HER2/neu and GM-CSF, respectively. Naive FVB mice vaccinated with combined VVneu and VVGMCSF given systemically developed systemic HER2/neu-specific immunity. NBT1-bearing mice became anergic to systemic immunization with combined VVneu and VVGMCSF. Intratumoral VVGMCSF failed to result in systemic antitumor immunity until combined with intratumoral VVneu. Infection/transfection of the tumor microenvironment with combined VVGMCSF and VVneu resulted in development of systemic tumor-specific immunity, reduction in splenic and tumor MDSC and therapeutic efficacy against tumors. These studies demonstrate the enhanced efficacy of oncolytic vaccinia virus recombinants encoding combined tumor antigen and GM-CSF in modulating the microenvironment of MDSC-rich tumors.

Fig. 1. Vaccination of naïve FVB/N mice with VVneu induces a systemic HER2-specific Th1 response

(a) Female FVB/N mice were injected twice, two weeks apart, with VVneu + VVGMCSF (either s.c. or i.m.f.), VVBGal + VVGMCSF (i.m.f.), or vehicle (i.m.f.). Two weeks after the second and final vaccination, spleens were restimulated with irradiated splenocytes from naïve female FVB/N mice that had been pulsed with immunodominant RNEU_420-429 peptide. (b) NBT1 tumor-specific systemic (spleen) CTL activity against NBT1 target cells after restimulation. Differences of both s.c. and i.m.f. VVneu + VVGMCSF groups compared to controls were significant (** P < 0.01, * P < 0.05). Results are representative of 3 independent experiments.

Fig. 2. NBT1 growth in female FVB/N mice leads to increased levels of Ly6G⁺ gMDSC, both systemically and in the tumor microenvironment

(a) Flow cytometry was performed on spleen, TDN, and tumor samples stained for Gr-1, CD11b, Ly-6C, and Ly-6G. Levels of CD11b⁺Gr1⁺ cells in spleen, TDN, and tumor were compared to levels in spleen and axillary LN of naïve female FVB/N mice. Cumulative results are presented from 5 independent experiments (*P < 0.05). (b) Cells from spleen and tumor were gated to include monocytic and granulocytic populations for evaluation of Gr-1 and CD11b. This population was then further gated for CD11b⁺ cells, which were evaluated for Ly-6G and Ly-6C. Results are representative of 3 independent experiments.

Fig. 3. MDSCs suppress a HER2/neu-specific Th1 response in NBT1 tumor-bearing mice

MDSCs from spleen of NBT1 tumor-bearing mice were depleted using a MDSC (Ly6G⁺) magnetic bead isolation kit. (a) Flow cytometry was performed on pre and post-depletion samples stained for Gr-1, CD11b, Ly-6C, and Ly-6G. (b) The MDSC-depleted negative fraction was restimulated with irradiated naïve splenocytes pulsed with control NP_118-126 peptide, RNEU_420-429 peptide, lysate from Sf21 insect cells that had been infected with baculovirus expressing HER2/neu, and control Sf21 cell lysate. IFN-γ ELISA was performed in triplicate on supernatant from MDSC-depleted effectors. Results representative of 3 independent experiments. (†<0.001 compared to all other stimulation conditions.)

Fig. 4. Vaccination with VVneu into the tumor microenvironment generates a systemic antitumor CTL response

(a) Female FVB/N mice were injected with 2×10⁶ NBT1 tumor cells into the right second mammary fat pad. Mice were treated twice (days 14 and 28) with VVneu + VVGMCSF + KLH (s.c. or i.t.), VVBGal + VVGMCSF + KLH (i.t.), or vehicle control (i.t.). On day 42, spleens and VDN were restimulated with irradiated naïve splenocytes that had been pulsed with RNEU_420-429 peptide. (b) Ability of splenocyte effectors to lyse target NBT1cells was measured by ⁵¹Cr release assay. Difference between i.t. VVneu + VVGMCSF + KLH and all other conditions was significant (** P < 0.01). (c) Tumor-specific CTL activity of VDN effector cells was measured by ⁵¹Cr lysis. Differences between i.t. VVneu + VVGMCSF + KLH and all other groups were statistically significant (** P < 0.01, * P < 0.05). (d) Restimulated effectors were evaluated for RNEU tetramer-positive CD8+ cells using flow cytometry. The percent of CD8+ cells that are RNEU tetramer positive was measured, with representative dot plots and cumulative results of 3 independent experiments presented (**P < 0.01).

Fig. 5. Vaccination with VVneu into the tumor microenvironment results in decreased levels of systemic and tumor microenvironment MDSCs

Splenocytes and tumor were stained ex vivo for Gr-1 and CD11b using flow cytometry on day 28. Representative dot plots and combined mean %CD11b⁺Gr-1⁺ cells from 3 independent experiments are shown (*P < 0.05 compared to i.t. Vehicle tumor, **P < 0.01 compared to i.t. Vehicle spleen, ¥P < 0.05 compared to i.t. VVBGal + VVGMCSF + KLH tumor; all other comparisons not significant).

Fig. 6. Vaccination with VVneu into the tumor microenvironment results in regression of NBT1 tumor

(a) Mean tumor size and standard deviation from 5 combined independent experiments are shown. (b) Day 42 tumor size from 5 combined experiments, each point representing one mouse from indicated treatment condition († P < 0.001). (c) Kaplan-meier survival curve from 5 combined experiments († P < 0.001).