A Vaccinia Virus Armed with Interleukin-10 Is a Promising Therapeutic Agent for Treatment of Murine Pancreatic Cancer

Pancreatic cancer is the fourth leading cause of cancer-related death worldwide (1) and remains consistently lethal with a 5-year survival rate of less than 5%. This situation signifies a need for radically new therapeutic strategies that are not subject to cross-resistance with conventional therapies.

Oncolytic viruses have emerged as attractive therapeutic candidates for cancer treatment due to their inherent ability to specifically target and lyse tumor cells and induce antitumor effects. An engineered replication-competent Adenovirus, dl1520 (ONYX-015), was the first of these viruses to be tested for human pancreatic cancer treatment. The treatments were well tolerated, but no objective responses with virus therapy alone were seen in any of the patients (2).

Vaccina virus has strong potential for exploitation as both an oncolytic agent and vector for therapeutic gene delivery to tumors. Extremely promising clinical trial data have recently emerged in which GM-CSF–armed vaccinia virus induced objective responses in patients with liver, colon, kidney, and lung cancer and melanoma (3, 4). Vaccinia virus has several inherent features that make it particularly suitable for use as an oncolytic agent, including fast and efficient replication with rapid cell-to-cell spread, natural tropism for tumors, a well-documented safety record, and an ability to replicate in many different cell types, a feature not shared by adenoviruses. We recently demonstrated that hypoxia, which contributes to the aggressive and treatment-resistant phenotype of pancreatic ductal adenocarcinoma (PDAC; ref. 5), does not inhibit and may even enhance the potency of oncolytic vaccinia virus (6). In addition, vaccinia virus has recently been shown to be effective at human tumor targeting after intravenous delivery (4).

Interleukin-10 (IL10), first described as a factor produced by Th2 clones capable of inhibiting Th1 cytokine production (7), is a potent inhibitor of T cell–mediated antiviral responses by prevention of dendritic cell (DC) activation of the CD4⁺Th1 inflammatory pathway (8, 9). IL10 is a key player in the establishment and perpetuation of viral persistence in vivo (10, 11). Therefore, arming vaccinia virus with IL10 may prolong viral persistence and enhance the antitumor efficacy. IL10 has historically been regarded as an immunosuppressive cytokine that has extensively been described in association with cancer, including pancreatic cancer (12, 13), as a mechanism of tumor escape from immunosurveillance (14, 15). However, accumulating evidence demonstrates that IL10 also has immunostimulatory and antitumor properties (16). Functional mechanisms investigated include activation of natural killer (NK) cells (17) that have been associated with tumor clearance in murine models of breast and colorectal cancer (18); inhibition of angiogenesis; enhancement of macrophage infiltration into tumors (19); and prevention of metastasis by inhibition of matrix metalloproteinase-2 (20). A number of preclinical (21) and clinical trials have consistently demonstrated safety of IL10 administration in treatment of diseases, including psoriasis (22), Crohn disease (23), and chronic hepatitis C infection (24), which make a strong case for its use as a therapeutic modality in cancer. IL10 has been reported to enhance the therapeutic effectiveness of a vaccinia virus–based vaccine against murine cancer cells (25), which may be connected to its ability to enhance the growth and proliferation of T cells (26) or its role as a chemotactic agent for CD8⁺ T cells. Unfortunately, the half-life of IL10 is only approximately 20 minutes and it is difficult to maintain a high concentration after administration of recombinant protein (27). Nonreplicating adenovirus-mediated delivery has shown promise in retaining therapeutically effective levels of IL10 in vivo (28). Given its pleiotropic effects, IL10 may be an effective agent with which to improve the antitumor potential of vaccinia virus.

In this study, we have tested a Lister strain, TK-deleted replicating vaccinia virus armed with murine IL10 (VVLΔTK-IL10) in subcutaneous and transgenic murine models of pancreatic cancer and demonstrated that VVLΔTK-IL10 has far superior antitumor activity compared with unarmed vaccinia virus (VVLΔTK), resulting in almost complete tumor clearance, significantly increased survival times, and the production of long-term tumor immunity in the host. Our results suggest that VVLΔTK-IL10 has strong potential as an effective treatment for pancreatic cancer and lays the foundation for translation of this therapeutic into a clinical setting.

The murine PDAC cell line DT6606 and the preinvasive pancreatic cancer (PanIN) cell line DT4994 were cultured from LSL-Kras^G12D/+;Pdx-1-Cre mice that had developed PDAC (29). These were kindly provided by David Tuveson (Cancer Research UK Cambridge Research Institute, Cambridge, United Kingdom; now at Cold Spring Harbor Laboratory). The DT6606-ovalbumin (OVA) stable cell line was created by transfection of DT6606 cells at with pCI-neo-cOVA (Addgene) using Effectene transfection reagent (Qiagen) according to the manufacturers' protocol. CV1 (African monkey kidney) cells and PT45 (human pancreatic carcinoma) cells were obtained from the American Type Culture Collection (ATCC).

Construction and production of recombinant vaccinia virus Lister strains VVLΔTK-IL10 (rVV-IL10, armed with murine IL10) and VVLΔTK (rVV-L15) were previously described (30, 31).

Appropriate cell lines were seeded in triplicate and infected 16 hours later with VVLΔTK or VVLΔTK-IL10 at a multiplicity of infection (MOI) of 1 plaque-forming unit (PFU) per cell. Cells and supernatant were collected at 24, 48, and 72 hours after infection and titers were determined by measuring the median tissue culture infective dose (TCID₅₀) on indicator CV1 cells. The Reed–Muench mathematical method was used to calculate the TCID₅₀ value for each sample (32). Viral burst titers were converted to PFU per cell based on the number of cells present at viral infection. One-way ANOVA followed by Bonferroni post-test was used to assess significance.

The cytotoxicity of the viruses in each cell line was assessed 6 days after infection with virus using an MTS nonradioactive cell proliferation assay kit (Promega) according to the manufacturer's instructions, which allowed determination of an EC₅₀ value (dose required to kill 50% of cells).

Subcutaneous tumors collected from treated mice were homogenized before DNA was extracted using the QIAamp DNA Blood Mini Kit (Qiagen Ltd.) according to the manufacturer's instructions. TaqMan system primers and probes (Supplementary Table S1) were designed using Primer Express v3.0 software (Applied Biosystems) and constructed by Sigma-Aldrich and Applied Biosystems, respectively. Samples, controls, and standards were tested in triplicate by quantitative polymerase chain reaction (qPCR) using 7500 Real-time PCR System. Results were normalized to NanoDrop readings and expressed as genome copy number/0.01 g tumor tissue. One-way ANOVA followed by Bonferroni post-test was used to assess significance.

IL10 or interferon-γ (IFNγ) protein levels were quantified using an IL10-specific or IFNγ-specific ELISA (R&D Systems) according to the manufacturer's instructions. Where appropriate, data were normalized to cell number present at time of infection.

Spleens were extracted from mice, combined with complete T-cell medium (RPMI medium, 10% BCS, 1% penicillin–streptomycin, and 1% sodium pyruvate), and cells were separated using a 70-μm cell strainer. Cells were resuspended in red blood cell lysis buffer (Sigma-Aldrich), washed in PBS, and the pellet was resuspended in T-cell medium.

Cells (2 × 10⁶) were aliquotted into each well of a 96-well plate in duplicate. Cells were restimulated with either a vaccinia virus–specific B8R peptide (TSYKFESV; Proimmune) at a final concentration of 20 μg/mL or 5 × 10⁵ mitomycin C–treated DT6606-OVA cells. Restimulated splenocytes were incubated at 37°C/5% CO₂ for 72 hours and the supernatant was collected.

Tumor cell suspensions were prepared by incubation with 1× collagenase/hyaluronidase (STEMCELL TECHNOLOGIES) for 30 minutes at 37°C. Cells were separated using a 70-μm cell strainer and resuspended in complete T-cell medium.

All fluorochrome-conjugated antibodies were supplied by eBiosciences and used at a 1:200 dilution. The B8R and OVA H-2Kb-restricted, MHC class I pentamers were synthesized by Proimmune and used at a 1:20 dilution.

Splenocytes and tumors were prepared and aliquotted into 96-well U-bottom plates. Pentamer staining was carried out by resuspending cells in FACS buffer (FB; PBS+1% heat-inactivated BCS+0.1%NaN₃) plus pentamer and incubating at room temperature for 10 minutes. Cells were washed before being incubated in FB plus appropriate fluorescent marker-conjugated anti-immune cell marker antibodies for 30 minutes on ice. Cells were washed and fixed in 2% formalin before analysis using a BD LSR Fortessa flow cytometer. Data were analyzed using FlowJo software (TreeStar Inc).

All animal studies were carried out under the terms of the Home Office Project Licence PPL 70/6030 and subject to Queen Mary University of London ethical review, according to the guidelines for the welfare and use of animals in cancer research (33).

The C57/BL6 mouse is H-2 haplotype-identical to the injected DT6606 cells thus DT6606 allografts could be established in the right flank of 3- to 4-week male C57/BL6 mice by injecting 3 × 10⁶ DT6606 cells. When tumors reached around 0.6 cm in diameter, mice were stratified by tumor size into groups of 8 and received 100-μL intratumoral (i.t.) injections of 1 × 10⁸ PFU of VVLΔTK, VVLΔTK-IL10, or PBS daily for 5 days. Tumor size was measured twice weekly until the death of the first animal in each group and volume was estimated [volume = (length × width² × π)/6]. Survival analysis was carried out using the Kaplan–Meier survival curves with log-rank (Mantel–Cox) tests used to assess significance. Mice that had cleared tumor after treatment were rechallenged 4 weeks after clearance in the opposite flank with 4 × 10⁶ DT6606 cells and tumor volume was estimated as previously. For immune depletion studies, DT6606 subcutaneous tumors were established as described and 1 day before commencement of viral treatment 200 μg of anti-CD4 IgG (antibody clone GK1.5), anti-CD8 IgG (antibody clone TIB210), anti-NK IgG (antibody clone PK136), or control rat IgG was injected intraperitoneally (i.p.) in 200-μL PBS. Injections were continued twice weekly for the duration of the experiment and FACS analysis was used to verify depletion for the duration of the experiment. Five mice per group were treated and the experiment was carried out twice.

LSL-Kras^G12D/+;LSL-Trp53^R172H/+;Pdx-1-Cre (KPC) mice were kindly provided by David Tuveson (Cancer Research UK Cambridge Research Institute) and have been described previously (29). Mice were treated when they reached 2.5 months, previously demonstrated to be the mean age at which PanINhas progressed to PDAC (29). Mice were treated i.p. with 2 × 10⁸ PFU/injection VVLΔTK or VVLΔTK-IL10 on days 1, 3, and 5. Mice were examined daily for signs of disease progression and culled when they showed symptoms of sickness. Survival data were compared using Prism (GraphPad Software) and a log-rank (Mantel Cox) test was used to determine significance of survival differences.

Seven days after treatment of KPC or KP mice, the biodistribution of VVLΔTK was determined in anesthetized animals (2% isofluorane inhalation) after i.p. injection of d-luciferin (150 mg/kg; Xenogen) and fluorescence measured with the IVIS camera (Xenogen Corp.).

KPC or control KP mice (2.5-month old) were treated i.p. with 2 × 10⁸ PFU/injection VVLΔTK or VVLΔTK-IL10 on days 1, 3, and 5. On relevant days, animals were sacrificed, the pancreas removed, snap-frozen, and stored at −80°C. Frozen tissue was processed for immunohistochemistry (IHC) analysis of vaccinia virus coat protein [1:50 rabbit anti-vaccinia virus coat protein polyclonal antibody (MorphoSys UK Ltd.)], macrophage [1:2,000 anti-F4/80 antibody (Serotech)], CD3⁺ T cell [1:200 anti-CD3 antibody (BioLegend)], or CD8⁺ T cell [1:300 anti-CD8 antibody (BioLegend)] as described previously (31).

To determine whether inclusion of IL10 affected on characteristics of VVLΔTK in vitro, replication and cytotoxicity in three cell lines were examined; DT6606, representing late-stage invasive PDAC and DT4994, representing PanIN, were both derived from the K-ras transgenic mouse model of pancreatic cancer (29). DT6606-OVA, in which the OVA antigen is overexpressed in DT6606 cells, was also examined. All cell lines supported production of infectious virions of VVLΔTK and VVLΔTK-IL10 (Fig. 1A–C) and IL10 did not act to inhibit nor promote viral replication. The EC₅₀ was comparable between VVLΔTK and VVLΔTK-IL10 (Fig. 1D). Furthermore, IL10 was expressed in all three cell lines over 72 hours after infection (Fig. 1E). Thus, arming VVLΔTK with IL10 does not adversely affect the in vitro oncolytic effect desired for our virotherapy strategy.

VVLΔTK-IL10 infection was also assessed in the human pancreatic cancer cell line PT45 to demonstrate potential translation of this therapy into human cells. VVΔTK-IL10 showed efficient replication, cytotoxicity, and IL10 expression in this cell line (Supplementary Fig. S1).

In vivo efficacy of VVLΔTK-IL10 was examined using a subcutaneously established pancreatic cancer model. DT6606 subcutaneous tumors were established in male C57/BL6 mice and the animals received i.t. injections of 1 × 10⁸ PFU of VVLΔTK, VVLΔTK-IL10, or PBS daily for 5 days. The selected viral dose was 10 times lower than the most commonly reported 1 × 10⁹ PFU/dose in the literature (34). Both VVLΔTK and VVLΔTK-IL10 demonstrated antitumor efficacy (Fig. 2A). However, treatment with VVLΔTK-IL10 resulted in a superior antitumor efficacy by day 44, with 87.5% of mice showing tumor clearance and significantly improved overall survival rates compared with both VVLΔTK- and PBS-treated animals (Fig. 2B). The C57/Black6 mouse is H-2 haplotype-identical to the injected DT6606 cells. Growth of tumors in PBS-treated animals confirmed that there was no immunologic rejection of the DT6606 cell line due to MHC or minor antigen mismatches.

To determine whether VVLΔTK-IL10 remained efficacious in a more pathologically relevant model of pancreatic cancer, KPC transgenic mice were used. In these mice, pancreas-specific expression of mutant Kras^G12D and Trp53^R172H results in progressive development of PDAC (35). Three doses of virus (2 × 10⁸ PFU/day) were given i.p. to 2.5-month old, PDAC-bearing mice. To confirm the specificity of virus for pancreatic tumors after i.p. injection, VVLΔTK, which expresses a luciferase transgene in the viral TK region, was injected into either experimental KPC mice or control KP mice. Two days later, mice were imaged for luciferase expression (Fig. 2C). Strong luciferase signals were obtained specifically in the pancreatic area of KPC transgenic mice (Fig. 2C, left), while no signal was obtained from control mice (Fig. 2C, right). The vaccinia virus proteins were expressed in cancer cells and proliferative acinar cells in KPC mice (Fig. 2C, left, bottom), whereas no viral protein expression was observed in the ductal epithelial cells and acinar cells in KP mice (Fig. 2C, right, bottom), confirming specificity of replication of TK-deleted vaccinia virus for pancreatic tumor cells. Efficacy of viral treatment in this model was assessed by survival (Fig. 2D). Treatment with VVLΔTK-IL10 resulted in significantly improved survival rates compared with treatment with VVLΔTK. Mean survival time for VVLΔTK-IL10–treated animals after commencement of treatment was 138.5 days compared with 69.7 days for VVLΔTK–treated animals, suggesting VVLΔTK-IL10 as an extremely effective treatment for PDACs even in the most complex murine models of the disease.

Successful OV strategies aim not only to eradicate the primary tumor, but also to induce long-term antitumor immunity to prevent disease recurrence. Thus, animals were rechallenged with 4 × 10⁶ DT6606 cells 4 weeks after complete regression of the primary tumor (Fig. 3A). Treatment with both viruses resulted in long-term immunity to DT6606 tumor cells as evidenced by rapid clearance of these cells that necessitated no further viral treatments. Interestingly, VVLΔTK-IL10–treated animals were able to clear the secondary tumor more quickly and more consistently than VVLΔTK-treated animals.

Long-term immunity suggests an activation of specific antitumor immune responses after treatment. To assess the contribution of different immune cells to treatment efficacy, CD8⁺, CD4⁺, or NK immune subsets were depleted from mice before treatment of subcutaneous DT6606 tumors with VVLΔTK-IL10 (Fig. 3B). Depletion of both CD4⁺ and CD8⁺ cells had a significantly detrimental effect on the efficacy of treatment, suggesting VVLΔTK-IL10 is acting via these immune subsets to eliminate the tumor. Surprisingly, given previous reports that IL10 can activate NK cells to mediate tumor clearance (17), depletion of NK cells in our experiment had no effect on treatment efficacy.

Given the involvement of T cells in VVLΔTK-IL10 treatment efficacy, tumor T-cell populations were analyzed in more detail. Pancreatic tumors of KPC transgenic mice treated as previously were harvested posttreatment and T-cell populations analyzed by IHC. We noted a significant increase in CD3⁺CD8⁺ infiltrate after treatment with both viruses compared with PBS (Supplementary Fig. S2) and a significant increase in CD3⁺CD8⁺ cells in VVLΔTK-IL10–treated animals at day 22 after infection compared with VVLΔTK–treated animals.

DT6606-subcutaneous tumors were also harvested for analysis of T-cell populations by FACS. In accordance with data obtained from KPC mice, we found a significant increase in tumor T-cell infiltrate after treatment with both viruses, with a significant increase in CD8⁺ infiltrate into tumors of VVLΔTK-IL10–treated animals (Fig. 4A). However, most interesting was that in CD4⁺ (data not shown) and, more significantly, CD8⁺ populations (Fig. 4B), the proportion of activated (CD45RB^lo/CD44^hi) T cells in tumors treated with VVLΔTK was higher than those treated with VVLΔTK-IL10. IFNγ expression within VVLΔTK-IL10–treated tumors was also significantly reduced compared with VVLΔTK–treated tumors (Fig. 4C).

Tumor-associated macrophage populations were also assessed in KPC (Supplementary Fig. S3) and DT6606 tumor-bearing mice (Fig. 4D) after infection. We found that treatment with either virus increased macrophage infiltrate into tumors of KPC mice compared with PBS, but that treatment with VVLΔTK-IL10 resulted in a reduced macrophage tumor infiltrate compared with treatment with VVLΔTK. This result was mirrored in DT6606 tumor-bearing mice. Further assessment of macrophage activation status in the DT6606 subcutaneous model revealed that VVLΔTK-IL10 induces a downregulation of MHCII expression compared with VVLΔTK (Fig. 4E).

To assess the impact of these phenomena on viral persistence, viral DNA load in the tumors (6 mice/group/time point) was analyzed after i.t. treatment at days 8, 16, and 24 after infection using qPCR (Fig. 4F, i) and TCID₅₀ (Fig. 4F, ii). We found that by day 24, both viruses had been cleared from the tumor to the same extent, but at days 12 and 16, significantly more VVLΔTK-IL10 was recovered from tumors than VVLΔTK, indicating a delay in clearance of VVLΔTK-IL10 compared with VVLΔTK. These results were confirmed by IHC analysis of viral load in pancreatic tumors of KPC mice (Supplementary Fig. S4).

It is clear that VVLΔTK-IL10 treatment efficacy involves modulation of the immune system, thus splenic immune cell population dynamics in response to treatment were assessed in greater detail. DT6606 tumor-bearing mice were treated as described and their spleens collected and assessed for presence of various immune cell subsets. No differences were found in splenic B cell (B220⁺ cells), Treg (CD4⁺, CD25^hi cells), NK (CD3⁻, CD49b⁺ cells), or NKT populations (CD3⁺, CD49b⁺ cells) after treatment with either virus compared with PBS-treated animals (Supplementary Fig. S5).

Analysis of CD4⁺ and CD8⁺ populations revealed that frequencies of these populations were altered at early time points (Fig. 5A and D; Supplementary Fig. S6). At days 8 and 16, a significant increase in the frequency of total CD8⁺ cells was seen after treatment with either virus; however, VVLΔTK-IL10 treatment resulted in fewer total CD8⁺ cells than treatment with VVLΔTK (Fig. 5D). This phenomenon was also observed in the CD4⁺ populations at day 16 after treatment (Fig. 5A).

Further examination revealed that after treatment with VVLΔTK-IL10 or VVLΔTK, T-cell populations shifted toward an effector/memory phenotype (Fig. 5B–F) at days 8 and 16. However, VVLΔTK-IL10 induced statistically fewer activated CD4⁺ and CD8⁺ T cells than VVLΔTK at days 8 and 16 after infection (Fig. 5B–F), as noted previously within the tumor.

To clarify the proportions of virus-specific and tumor-specific splenic effector CD8⁺ cells elicited after treatment with VVLΔTK and VVLΔTK-IL10, splenocytes from DT6606-OVA tumor-bearing animals were analyzed. For virus-specific T cells, an MHCI-specific pentamer against an immunogenic vaccinia virus antigen, B8R, was used (Fig. 6A and Supplementary Fig. S7A). As expected, viral treatment resulted in detection of B8R-specific CD8⁺ cells in both treatment groups. However, VVLΔTK-treated animals had a significantly higher proportion of B8R-specific T cells than VVLΔTK-IL10–treated animals at all time points, suggesting a decreased virus-specific immune response after treatment with VVLΔTK-IL10, which could account for the fewer effector CD8⁺ cells noted after VVLΔTK-IL10 treatment. We confirmed the decreased frequency of antivirus-specific T cells using an in vitro restimulation assay, in which IFNγ production from splenocytes in response to B8R peptide restimulation was measured (Fig. 6B). At all time points, significantly less IFNγ was detected from VVLΔTK-IL10 treatment groups compared with VVLΔTK treatment groups.

To assess T-cell reaction to tumor antigens, an MHCI OVA-specific pentamer was used in FACS staining (Fig. 6C and Supplementary Fig. S7B). At day 8, no differences in OVA-specific CD8⁺ T cells was observed after treatment with either virus when compared with PBS; however, by day 16, VVLΔTK-IL10–treated animals showed an increase in production of OVA-specific antigens compared with VVLΔTK–treated animals. This result was reflected in restimulation assays (Fig. 6D).

Taken together, these results indicate that although VVLΔTK-IL10 treatment resulted in a reduction in antiviral T-cell production, the frequency of antitumor-specific CD8⁺ T cells was comparable or even increased compared with VVLΔTK–treated mice.

Efficacy of oncolytic virotherapy is dependent on both the oncolytic action of the virus itself and the effective stimulation of a local immune response to viral infection (36, 37). Oncolytic viruses may represent a method of achieving vaccination in situ, enabling the adaptive arm of the immune system to clear residual disease and provide long-term surveillance against relapse. To date, however, the use of oncolytic viruses alone has proved unsuccessful in clinical trials and this is likely due to their early clearance preventing their oncolytic effects and an effective immune-stimulating release of tumor-associated antigens (TAAs). Many viruses encode homologs of the cytokine IL10, generally considered immunosuppressive, to dampen the antiviral immune response and circumvent early viral clearance (11, 38). We aimed to adopt this natural strategy of viruses by arming vaccinia virus with IL10, which has been reported to be effective at prevention of vaccinia virus clearance (39). We hypothesized that prolonging viral persistence in the host would improve the antitumor efficacy by enhancing both the direct oncolytic effect and release of TAAs.

The pancreatic cancer subcutaneous tumor model we developed was based on the use of a DT6606 cell line, which was originally derived from the transgenic KPC spontaneous model of pancreatic cancer (29), and therefore accurately reflect the PDAC populations of cells within these mice. Previous study has demonstrated that these cancer cells resemble human PDAC in many respects, including their expression of oncogenic KrasG12D and the TAA mesothelin, and both spontaneous and subcutaneous tumors show similar histopathologic features such as the presence of FAP⁺ stromal cells (40).

The long-held paradigm of IL10 function suggests it as an immunosuppressive cytokine, commonly investigated therapeutically in the context of treatment for inflammatory autoimmune conditions and allograft survival (41, 42). However, using these two different murine models of pancreatic cancer, we observed significantly enhanced therapeutic responses after treatment with our IL10-armed vaccinia virus compared with unarmed virus. In both models, low doses of the virus were sufficient to induce objective responses and in agreement with previous reports, no IL10-related toxicity was observed (43). Treatment also resulted in rejection of tumors after rechallenge, confirming the development of effective long-term immunity against tumor antigens. These results are consistent with those of others investigating the antitumor properties of IL10 in which systemic administration of recombinant protein or tumor cells transfected with IL10 induced tumor clearance and long-term memory responses in mice bearing sarcoma (16), melanoma (16, 44), colorectal cancers (16), breast cancers (45), and prostate cancers (20).

In vitro studies indicated that IL10 did not alter vaccinia virus replication or cytotoxicity and no effect on cell proliferation was observed. To determine other possible mechanisms for the superior efficacy associated with this virus, viral persistence within tumors was assessed. Although both IL10-armed and unarmed viruses were effectively cleared from animals, greater titers of VVLΔTK-IL10 were recovered at days 12 and 16 compared with VVLΔTK in both the transgenic and subcutaneous models of pancreatic cancer, suggesting that IL10 could significantly delay viral clearance.

Given previous reports of the ability of IL10 to stimulate NK cells (17) and as a cytotoxic T-cell differentiation factor (46), we examined reliance of our treatment on these immune subsets. Depletion of NK cells had no effect on treatment efficacy in vivo and we found no evidence of altered splenic or tumor (data not shown) NK populations after treatment with VVLΔTK-IL10. In contrast, depletion of CD4⁺ and CD8⁺ T-cell populations had a negative impact on treatment efficacy. It has previously been reported that progression from PanIN to PDAC is accompanied by progressive infiltration of T cells into the tumor in KPC transgenice mice (47, 48); however, no antitumor response is induced by this infiltrate. Our analysis of T-cell populations in spleens and tumors revealed that treatment with both unarmed and IL10-armed viruses induced a high level of adaptive immunity in mice compared with untreated mice. However, an interesting finding was that the magnitude of the activated splenic CD4⁺ and CD8⁺ population response in VVLΔTK-IL10 treated mice was lower compared with the unarmed virus. This difference correlated with a reduction in virus-specific CD8⁺ T cells and IFNγ recovery from tumors after VVLΔTK-IL10 treatment, which accounted for the delayed viral clearance from tumors. Interestingly, although VVLΔTK-IL10 treatment reduced antiviral CD8⁺ populations, IL10 had no inhibitory effect on production of antitumor CD8⁺ cells. Indeed, at day 16 after injection, an increase in anti-OVA CD8⁺ cells was observed, which we postulate is a result of the increased oncolysis occurring with VVLΔTK-IL10 treatment, which improves TAA release.

These results suggest that IL10 improves the efficacy of OV by modulation of the early immune response to infection, resulting in dampening of antiviral, but not antitumor immunity. However, the mechanism by which IL10 elicits this alteration remains unclear. Our investigations revealed that local IL10 expression results in modification of the tumor macrophage population, which is highly sensitive to IL10 exposure (49). Numerous investigators have reported that IL10 can negatively regulate macrophages by (i) inhibiting their infiltration into tumors, and (ii) downregulating MHCII expression and suppressing production of proinflammatory cytokines and reactive nitrogen oxides (50). Although we found that VVLΔTK-IL10 treatment increased macrophage infiltrate into tumors in both the spontaneous and subcutaneous models of pancreatic cancer, we found that in accordance with previous data, VVLΔTK-IL10 treatment results in a significant downregulation of MHCII expression. Thus, it is feasible that in our model, tumor macrophages are responsible for viral antigen presentation to T cells and a reduction in macrophage activation by IL10 leads to reduced cross-priming of the antiviral immune response. A further consideration is that this model suggests distinct pathways of viral and tumor antigen presentation, which are the subject of ongoing investigation in our laboratory.

These findings demonstrate that IL10 armed vaccinia virus shows great promise as a novel therapeutic for pancreatic cancer, and that IL10 in combination with oncolytic virotherapy is clearly able to enhance tumor rejection through modulation of the innate and adaptive immune responses.

No potential conflicts of interest were disclosed.

Conception and design: L.S. Chard, I. Fodor, N.R. Lemoine, Y. Wang

Development of methodology: L.S. Chard, E. Maniati, P. Wang, F. Cao

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): L.S. Chard, E. Maniati, Z. Zhang, D. Gao, J. Wang, J. Ahmed, M. El Khouri, J. Hughes, B. Denes

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): L.S. Chard, E. Maniati, J. Hughes, B. Denes, I. Fodor, Y. Wang

Writing, review, and/or revision of the manuscript: L.S. Chard, E. Maniati, J. Ahmed, J. Hughes, B. Denes, I. Fodor, N.R. Lemoine, Y. Wang

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): L.S. Chard, J. Wang, S. Wang, X. Li, T. Hagemann, Y. Wang

Study supervision: I. Fodor, Y. Wang

This study was supported by The UK Charity Pancreatic Cancer Research Fund, the National Natural Science Foundation of China (81101608 and 81201792), Ministry of Sciences and Technology, China (2013DFG32080), the Henan Provincial Department of Science and Technology as well as the Department of Health, Henan Province, China (124200510018 and 104300510008).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.