Remodeling of the tumor microenvironment using an engineered oncolytic vaccinia virus improves PD-L1 inhibition outcomes

Abstract Immune checkpoint inhibitor (ICI) immunotherapies have vastly improved therapeutic outcomes for patients with certain cancer types, but these responses only manifest in a small percentage of all cancer patients. The goal of the present study was to improve checkpoint therapy efficacy by utilizing an engineered vaccinia virus to improve the trafficking of lymphocytes to the tumor, given that such lymphocyte trafficking is positively correlated with patient checkpoint inhibitor response rates. We developed an oncolytic vaccinia virus (OVV) platform expressing manganese superoxide dismutase (MnSOD) for use as both a monotherapy and together with anti-PD-L1. Intratumoral OVV-MnSOD injection in immunocompetent mice resulted in inflammation within poorly immunogenic tumors, thereby facilitating marked tumor regression. OVV-MnSOD administration together with anti-PD-L1 further improved antitumor therapy outcomes in models in which these monotherapy approaches were ineffective. Overall, our results emphasize the value of further studying these therapeutic approaches in patients with minimally or non-inflammatory tumors.


Introduction
The development of immune checkpoint inhibitor (ICI) immunotherapies including anti-programmed cell death-1 (PD-1)/PD-L1 and anti-cytotoxic T lymphocyte antigen 4 (CTLA4) have significantly improved survival and clinical outcomes in a subset of patients with certain tumor types [1][2][3]. In a study of lymphoma patients, ICI treatment exhibited synergistic activity with other antitumor therapies relative to monotherapeutic interventions, but these treatments were associated with systemic toxicities associated with the robust activation of the immune system [4]. Further research is thus required to optimize the clinical efficacy of these therapeutic approaches while minimizing toxicity via utilizing different ICIs, agonistic drugs, cytokines, and combinations thereof [5,6].
The local administration of immunotherapeutic agents represents an ideal approach to minimizing off-target systemic toxicity. The use of tumor-specific oncolytic viruses (OVs) capable of replicating in the tumor microenvironment (TME) and killing tumor cells to release tumor-specific antigens has the potential to improve cancer treatment outcomes [7,8]. Such virus-induced tumor cell lysis can also activate innate immune receptors owing to the concomitant release of damage-associated molecular patterns, thereby bolstering antitumor immune responses [9]. Tumor-selective OVs can also be leveraged as vectors to deliver therapeutic genes to the TME, and several such viruses have been tested in this context. For example, talimogene laherparepvec (T-VEC) [10,11], which is a herpes simplex virus strain that was engineered to encode the immune adjuvant granulocyte-macrophage colony-stimulating factor (GM-CSF), was approved as a first-in-class OV in U.S.A. in 2015 to treat metastatic melanoma. Phase III trials have also been conducted to assess the efficacy of the GM-CSF-encoding Pexastimogene devacirepvec (Pexa-Vec, JX-594) in advanced hepatocellular carcinoma patients [12,13]. Whether the encoding of these immunomodulatory genes improves overall patient clinical outcomes, however, remains uncertain, as these genes can promote local inflammation that suppresses the oncolytic activity of the viral vectors and further suppresses the expression of these transgenes [14]. In addition to directly lysing tumor cells, OVs have been shown to effectively activate immune cells and promote their infiltration into tumors, circumventing immunosuppression in the microenvironment. Considering the immunomodulatory properties of OVs, strategies combining virotherapy with other immunotherapies, such as ICIs, have been proposed [15][16][17][18]. OV-based immunotherapy combined strategy can elicit tumor killing effect via multiple targets and mechanisms, which may be expected to improve the situation [19].
We have previously demonstrated the ability of oncolytic vaccinia virus (OVV), which is a promising immune-oncolytic therapy strategy [20], to induce T cell and NK cell infiltration and to reduce myeloid-derived suppressor cell (MDSC) numbers in a model of subcutaneous lymphoma [21]. We also generated the novel tumor-homing E1B55K gene-deleted ZD55-MnSOD oncolytic adenovirus encoding the manganese superoxide dismutase (MnSOD) gene, and we found that this virus was able to promote robust tumor cell death in vitro and in vivo in models of ovarian and colorectal cancer [22]. Several studies have shown that MnSOD −/+ mice are prone to spontaneous lymphoma development and MnSOD is crucial for proper thymocyte differentiation, homeostatic survival of peripheral T cells as well as for T cell-mediated immune responses [23][24][25]. So we presumed that overexpression of MnSOD not only suppresses tumor cell growth, but also promotes stronger antitumor immune response. Herein, we sought to explore the intratumoral changes in immune status when utilizing an MnSOD-expressing OVV platform either alone or in combination with anti-PD-L1. Through these experiments, we found that OVV-MnSOD markedly enhanced intratumoral inflammation, improved systemic antitumor efficacy, and increased lymphoma sensitivity to PD-L1 blockade. Together, our results highlight novel approaches to overcoming ICI resistance in tumors, providing a foundation for efforts to expand these results to human clinical trials.

OVV-MnSOD synthesis
OVV and OVV-MnSOD vectors were prepared via gene recombination [26,27]. Briefly, the full MnSOD gene sequence was then inserted into the pCB plasmid. The resultant sequenced pCB-MnSOD or control pCB vectors were then transfected into HEK293A cells that had already undergone WT vaccinia virus infection. In brief, HEK293 cells were infected with WT vaccinia virus at a multiplicity of infection (MOI) of 1 for 2 h and then transfected with the corresponding shuttle plasmid. The cell extraction solution was used to infect the HEK293 cells in the presence of 25 mg/ml mycophenolic acid (MPA; Cat# A600640, Sangon Biotech, Shanghai, China), 250 mg/ml xanthine (Cat# A601197, Sangon Biotech), and 15 mg/ml hypoxanthine (Cat# A500336, Sangon Biotech). After three cycles of screening, EGFP-positive plaques were isolated, resuspended, and further HEK293 cells were infected for two cycles of plaque purification. After completing the first and second rounds of plaque purification, the following primers were used to amplify the target gene and the viral thymidine kinase (TK) gene to identify whether the recombinant virus was adulterated with the parental vaccinia virus. Primer of target gene: 5 -CTCCCCGACCTGCCCTACGACT-3 , 5 -TGCAAGCCATGTATCTTTCAGTTAC-3 ; primer of TK: 5 -tgtgaagacgataaattaatgatc-3 , 5 -gtttgccatacgctcacag-3 . Recombinant vaccinia virus successfully screened by plaque purification was further expanded by Hela-S3 cells in six-well plates, cell culture dishes, and cell culture spinner flasks.

Quantitative RT-PCR
TRIzol (#15596-026; Invitrogen) was utilized to extract total RNA from infected cells, after which PrimeScript RT Master Mix (#DRR036A, TaKaRa, Shiga, Japan) was used to prepare cDNA. Next, FastStart Universal SYBR Green Master Mix (#04913914001; Roche) was employed to conduct quantitative RT-PCR (qPCR) with a Real-Time PCR System (Applied Biosystems, CA, U.S.A.). The comparative C t method was used to evaluate relative gene expression, with GAPDH for normalization.

Western blot
Cells were lysed using RIPA buffer and BCA assay was then conducted to measure protein levels in these samples based on the provided directions. Equal protein amounts were separated via 10-15% SDS/PAGE and transferred on to PDF membranes that were probed with anti-MnSOD (Abcam, ab68155), anti-Caspase-3 (Abcam, ab3251), or anti-GAPDH (Abcam, ab9485). Blots were then probed with HRP-linked secondary antibodies for 1 h (1:4000; HuaAn Biotechnology Co. Ltd). The ImageJ software was then employed to measure protein band density.

CCK8 assay
Murine lymphoma A20 and EL4 cells were added to 960-well-plates (10000 cells/well) and were infected with OVV or OVV-MnSOD for 48 h, after which 10 μl of CCK8 reagent was added per well for 4 h at 37 • C. Absorbance at 450 nm was then assessed with a microplate reader to assess the cytolytic activity of these viral preparations. PBS was used as a negative control.

Animal experiments
The Animal Ethics Committee of Zhejiang Provincial People's Hospital approved all animal studies (A20190029). C57BL/6 mice were purchased from Zhejiang Chinese Medical University, Hangzhou, China) and housed in a specific pathogen-free facility with free food and water access in Animal Laboratory of Zhejiang Provincial People's Hospital. Digital calipers were used to measure subcutaneous tumor diameter, with tumor volume being defined as: volume = length × width 2 × 0.5. The body weight and tumor sizes of all mice were regularly monitored, and mice were killed by carbon dioxide suffocation if they exhibited acute weight loss or tumors ≥ 3000 mm 3 in size. When tumors were no longer palpable, mice were considered to have achieved a complete response (CR).
A20 and EL4 cells (5 × 10 6 or 2 × 10 6 cells, respectively) were subcutaneously injected into the right flank of model mice. After tumors were ∼50 mm 3 in size, 50 μl of PBS, OVV, or OVV-MnSOD were injected into the tumor every other day (three doses in total). Combination therapy efficacy was assessed by also intraperitoneally injecting these mice with 100 μg of anti-PD-L1 (clone RMP1-14, Bio X Cell) beginning on the second day of initial viral treatment every 3 days for three times.

OVV-MnSOD characterization
We employed a homologous recombination approach to prepare OVV-MnSOD as detailed previously [28], and as shown in Figure 1A. Successful exogenous expression of MnSOD in these viral particles was confirmed via qPCR. OVV and OVV-MnSOD were then used to infect A20 and EL4 lymphoma cells (MOI = 2) for 24 h, after which significant MnSOD expression was detectable in cells infected with OVV-MnSOD but not in cells infected with OVV or treated with PBS ( Figure 1B). Western blotting yielded comparable results regarding MnSOD protein levels in these cells, confirming that we had successfully prepared an OVV strain capable of overexpressing MnSOD in target cells. Subsequent Western blotting also revealed that OVV-MnSOD infection significantly induced caspase-3 cleavage in both the cell lines at 48 h post-infection, consistent with the apoptotic death of these target cells ( Figure 1C).
As an additional control, we evaluated the impact of MnSOD expression on OVV viral replication in these two lymphoma cell lines. This analysis revealed no significant difference in viral yield when comparing the OVV and OVV-MnSOD viruses, suggesting that this transgene had no adverse impact on such replication ( Figure 1D). As such, the ability of these OVV preparations to selectively replicate within tumor cells is unaffected by the deletion of TK or the insertion of the MnSOD transgene.

In vitro antitumor activity of OVV-MnSOD
A 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay was next conducted using A20 and EL4 cells following a 48-h infection with OVV-MnSOD in order to gauge the cytotoxic activity of this virus. OVV-MnSOD exhibited superior inhibition of lymphoma cell proliferation relative to OVV in a dose-dependent manner in this assay (Figure 2A,B), whereas neither virus was able to suppress peripheral blood mononuclear cell (PBMC) replication ( Figure 2C). Overall, these findings indicated that OVV-MnSOD can selectively and specifically suppress the in vitro growth of tumor cells.

OVV-MnSOD promotes lymphocyte infiltration into tumors
In order to establish the ability of our virotherapy approach to promote remodeling within the TME, we next intratumorally administered OVV or OVV-MnSOD into immunocompetent mice bearing A20 or EL4 tumors (Supplementary Figure S1). In the former model, OVV-MnSOD resulted in an 84.9% inhibition of tumor growth relative to 53.8% for OVV, while in the latter tumor model these percentages were 77.8 and 52.5%, respectively ( Figure 3A). Animals were killed when tumors were over 2000 mm 3 in size or mice appeared moribund. All mice administered an intratumoral PBS injection were killed within 20 days of treatment, while 75% of A20 model mice and 62.5% of EL4 model mice treated with OVV-MnSOD survived ( Figure 3B). To establish the impact of OVV-MnSOD on tumor-infiltrating lymphocyte (TIL) populations, we conducted flow cytometry on tumor samples collected on day 7 following the third treatment of these mice. Tumor-derived single-cell suspensions were stained with antibodies specific for CD45, CD3, CD4, CD8, FoxP3, NK1.1, and CD103. This analysis revealed that OVV-MnSOD administration was associated with a significant increase in the frequency of CD8 + T cells, CD4 + T cells, Treg, NK cells, and dendritic cells (DCs) within EL4 and A20 tumors relative to those treated with PBS, and to increase the frequency of CD8 + T cells, CD4 + T cells, and DCs in both of these tumor types relative to OVV treatment ( Figure 3C,D).

OVV-MnSOD enhances tumor sensitivity to anti-PD-L1 treatment
We next speculated that the increases in TIL levels within tumors following OVV-MnSOD injection would enhance the susceptibility of these tumors to ICI treatment. To test this, we treated mice bearing these lymphoma model tumors with anti-PD-L1 with or without prior intratumoral OVV-MnSOD treatment (Supplementary Figure S2). While anti-PD-L1 monotherapeutic treatment exhibited limited efficacy and did not facilitate CR in any of the treated mice, combination OVV-MnSOD + anti-PD-L1 treatment was associated with CR in the majority of treated mice (7/10 CR in the A20 model, 8/10 in the EL4 model) ( Figure 4A). Importantly, 100% of mice in both of these combination treatment groups survived the study period ( Figure 4B). Intratumoral injection of OVV-MnSOD prior to PD-L1 blockade increased CD8 + T cells, activated CD8 + T cells, and Tregs in A20 ( Figure 4C) and EL4 ( Figure 4D) tumors. This suggests that locally injecting OVV-MnSOD can increase tumor sensitivity to ICI treatment. No murine weight loss was detected over the course of the study period (data not shown).

Discussion
A range of immunotherapies has been tested to treat lymphoma patients. ICI-based treatments including anti-PD-1/PD-L1, anti-CTLA4, and combinations thereof are increasingly common standards of care for certain forms of cancer [29]. However, these immune checkpoint blockage approaches are only efficacious in a limited number of patients owing to the potential for systemic toxicity and the role of multiple different immune signaling pathways within the TME [30]. It is thus essential that novel approaches to safely and effectively improving tumor patient clinical response rates be developed. In the present study, we found that intratumoral OVV-MnSOD administration both altered the local and the systemic immune status in mice bearing lymphoma model tumors refractory to anti-PD-L1 treatment. Importantly, this virotherapy approach was sufficient to sensitize mice to ICI treatment without inducing significant toxicity. Increased mitochondrial reactive oxygen species (ROS) production is a common hallmark of tumor cells. After being produced, ROS can suppress T-cell activation and proliferation within the TME such that high levels of ROS can impair the development of antitumor immune responses in cancer patients, whereas low levels of ROS can enhance T cell receptor-induced signal transduction [31,32]. MnSOD is a mitochondrial antioxidant protein found in Escherichia coli and encoded by nuclear genes on chromosome 6q21. Reductions in MnSOD activity are frequently observed in tumor tissues and cell lines from humans, and it has been shown to function as a tumor suppressor protein such that MnSOD overexpression can suppress the growth of a range of tumor types [33,34]. There are multiple limitations to the present study that should be considered when evaluating this therapeutic approach. For one, while we were able to demonstrate OVV-MnSOD-induced changed in the local tumor microenvironmental immune status based on analyses of particular cell subsets; further work is necessary to fully understand the functional roles of cells such as antigen-presenting cells, stromal cells, and Tregs within tumors in the context of this virotherapy approach. In particular, more work is needed to understand the link between OVV therapy and adaptive antitumor immunity through gene expression and T cell repertoire analyses [14]. In addition, more research is required to establish the tumor types that are most likely to be amenable to virotherapy treatment by identifying key therapeutic biomarkers. While we initiated anti-PD-L1 treatment on the first day of OVV-MnSOD administration in light of prior reports suggesting that ICIs can suppress vaccinia virus replication, additional dosing optimization is required to improve the clinical outcomes associated with this therapeutic strategy. Clinical trials evaluating the combination of ICIs with OVs, such as T-VEC followed by pembrolizumab, will provide key insights regarding the promise of such combination virotherapy.
Together, our results suggest that the intratumoral administration of OVV-MnSOD can enhance antitumor immunity and increase tumor immunogenicity and amenability to checkpoint blockade therapy. This OVV-MnSOD preparation may offer value as both a monotherapy and together with different immunotherapies when used to treat diverse tumor types.

Data Availability
Some or all data generated or used during the study are available from the corresponding authors on request.