Combination Therapy Using Ruxolitinib and Oncolytic HSV Renders Resistant MPNSTs Susceptible to Virotherapy

Malignant peripheral nerve sheath tumors (MPNSTs) are highly aggressive and clinically challenging soft-tissue sarcomas characterized by high local recurrence rates, metastatic potential, and resistance to chemotherapy. MPNSTs can develop from a peripheral nerve, a preexisting peripheral nerve sheath tumor, or from a precursor plexiform neurofibroma in the setting of cancer genetic predisposition syndrome, neurofibromatosis type 1 (NF1; ref. 1). MPNST is the most common malignancy in patients with NF1, with an incidence rate of 2% and a lifetime risk of 8% to 13% (1–3). Complete surgical resection with negative margins is the treatment of choice in the case of localized disease, but in metastatic MPNST, outcomes are generally poor. The 5-year survival range is between 15% and 50% (4). Poor clinical outcomes with conventional agents have urged the need for new therapeutic agents.

Oncolytic viruses are biological anticancer agents that preferentially target tumor cells while sparing normal cells and may be engineered to improve their efficacy and/or safety profile. Viral replication leading to direct tumor cell lysis was previously considered the principal mechanism contributing to oncolytic viruses’ antitumor activity: once the virus enters tumor cells, the virus propagates and lyses the tumor cells. However, the oncolytic virotherapy paradigm has evolved from a focus primarily on viral replication and direct oncolysis to an emphasis on the immune-mediated antitumor mechanism (5, 6). In addition to its beneficial activity, the immune response could also negatively affect oncolytic virus antitumor activity by prematurely restricting viral infection, leading to ineffective viral replication (7, 8). One of the main challenges facing oncolytic virotherapy is the optimization of efficient viral replication and the generation of an anticancer immune response.

Oncolytic herpes simplex viruses (oHSVs) used in clinical trials to date are derived from recombinant HSVs, from which the principal neurovirulence gene (γ₁34.5) has been deleted (9–11). These attenuated Δγ₁34.5 oHSVs are useful for the development of a therapeutic platform for cancer treatment and have been safely administered to patients with peripheral and brain tumors in phase I clinical trials (10). Oncolytic HSV-1 has been shown to be safe and effective in phase I to III clinical trials (12), and in 2015, the FDA approved the first oncolytic HSV called talimogene laherparepvec (T-VEC) for the treatment of advanced, inoperable malignant melanoma by intralesional injection (13).

However, limited replication and spread in tumors has impaired the clinical benefit of first-generation oHSV-1. The need for efficient replication and improved immune activity led to next-generation oHSVs, such as T-VEC, and the virus we used here, C134. C134 is an oHSV that expresses the human cytomegalovirus (HCMV) IRS1 gene, which is essential for HCMV replication (14), restoring the virus's ability to replicate and synthesize late viral proteins but not its neurovirulence, allowing it to remain as safe as the parent Δγ₁34.5 (15–17). However, C134 remains sensitive to IFN-mediated restriction and stimulates early IRF3-dependent signaling in nonmalignant cells and some tumors with intact signaling pathways (16). This IRF3-mediated acute inflammatory response recruits the immune cells and improves the immune-mediated antitumor response (16).

Our studies demonstrated that 50% of human MPNST tumor lines restrict both first-generation oHSV and C134. We investigated the molecular mechanism underlying this oHSV resistance and showed that HSV entry was not the primary obstacle (18). Instead, basal NF-kB and IFN signaling activation primed an antiviral environment hostile to viral replication in these resistant tumor cells (19). Constitutive expression of antiviral interferon-stimulated genes (ISG), such as retinoic acid-inducible gene I (RIG-I), melanoma differentiation-associated protein 5 (MDA-5), interferon-induced protein with tetratricopeptide repeats 3 (IFIT3), myxovirus resistance 1 (MX1), and 2′,5′-oligoadenylate synthetase-1 (OAS1), is driven by JAK/STAT-mediated signaling in these cells and contributes to oHSVs’ inability to replicate efficiently, resulting in a negligible antitumor effect (19). A similar mechanism has been described in carcinomas and inflammation-driven tumors, where chronic low cytokine signaling promotes myeloid and stromal cell infiltration, as well as NF-kB and STAT-mediated chemokine, cytokine, and growth factor production in tumors. This chronic cytokine and chemokine production creates a paradoxically inflamed (IL10, PgE2, TGFβ) and immunosuppressive tumor environment, that, when coupled with chronic tumor antigen stimulation, can lead to T-cell exhaustion (20, 21). We sought to identify if murine MPNST tumor lines utilize a similar oHSV-resistance mechanism, which would allow us to dissect the complex relationship between viral replication, the intrinsic IFN-mediated antiviral response, and the resultant cellular immune-mediated antitumor response.

In the current study, we showed that murine MPNST tumor lines recapitulated our findings in human MPNSTs. Resistant murine MPNSTs rapidly expressed and accumulated higher ISG expression and activated signal transducers and activators of the transcription-1 (STAT1) signaling pathway in response to viral infection, thus reducing oHSV replication. To test how JAK inhibition influenced oHSV therapy, we chose one of the more resistant MPNST lines (67C-4), which constitutively activates STAT1 pathway signaling and restricts viral replication. Pretreatment of the resistant MPNSTs with the FDA-approved JAK1/2 inhibitor ruxolitinib (RUX, Jakafi, Incyte Corporation) increased MPNST susceptibility and improved viral spread and gene expression in vitro. On the basis of these results, we hypothesized that RUX pretreatment would improve C134 replication and antitumor effects in vivo. Here, we present evidence that combination therapy using RUX and C134 improves oHSV antitumor activity in resistant MPNSTs in a syngeneic model and that this combinatorial therapy involves both enhanced viral replication and an essential, adaptive CD8⁺ T-cell response.

B76 and B96 cells were generously provided by Dr. Steven Carroll at Medical University of South Carolina and were propagated in Dulbecco's modified eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS). 67C-4 and 5NPCIS cells were kindly provided by Dr. Tim Cripe and were developed and provided to him by his collaborator Dr. Nancy Ratner and maintained in DMEM supplemented with 10% FBS. Tumor lines were tested negative for Mycoplasma contamination using the ATCC universal Mycoplasma detection kit. Tumor cells with relative low passage numbers (<12 passages) were used in the study before returning for a “low” passage form of the cell line to minimize genetic drift in our studies. Viruses have been previously described (22), but in brief; HSV-1 (F) strain and R3616, the Δγ₁34.5 recombinant, were kindly provided by Dr. Bernard Roizman (University of Chicago, Chicago, IL; ref. 23). C134 has been described previously (15, 17). Briefly, C134 is a Δγ₁34.5 virus that contains the HCMV IRS1 gene under control of the CMV IE promoter in the U_L3/ U_L4 intergenic region (ref. 15; Supplementary Table S1). C154 is an EGFP-expressing version of C134.

B76, B96, 67C-4, and 5NPCIS cells were plated into clear, 48-well flat-bottom polystyrene tissue culture–treated microplates (Corning) and allowed to adhere overnight at 37°C. Cells were infected the following day with an EGFP-expressing second-generation oHSV-1 (C154) at the indicated multiplicity of infection (MOI), and the plates were monitored using the IncuCyte Zoom platform, which was housed inside a cell incubator at 37°C with 5% CO₂ until the end of the assay. Nine images per well from three replicates were taken every 3 hours for 3 days using a 10× objective lens and then analyzed using the IncuCyte Basic Software. Green channel acquisition time was 400 ms in addition to phase contrast.

Animal studies were approved by the Nationwide Children's Hospital Institutional Animal Care and Use Committee (IACUC; protocol number AR16-00088) and performed in accordance with guidelines established by NIH Guide for the Care and Use of Laboratory Animals. To establish tumors, 2 × 10⁶ 67C-4 MPNST cells were injected subcutaneously into the flanks of 6- to 8-week-old C57BL/6 mice (Envigo). Tumor sizes were measured biweekly by caliper after implantation, and tumor volume was calculated by length × width × depth. When tumors reached 25 to 150 mm³ in size, animals were pooled and randomly divided into the specified groups, discussed below, with comparable average tumor size. Mice were administered 3 doses of RUX (INCB018424, AbexBio; 60 mg/kg) every day intraperitoneally (i.p.) as described previously (24). Studies were repeated 3 times to ensure biological validity. Mice were treated with saline or C134 (3.5 × 10⁷ in 100 μL 10% glycerol in PBS) intratumorally (i.t.) on day 4 (1 day after the last RUX dose) and again a week later.

For survival studies, animals were monitored for tumor volumes three times per week after the initial treatment, until total tumor volume/mouse exceeded 2,000 mm³ or an individual tumor was >1,500 mm³. Once overall tumor size exceeded these criteria, mice were sacrificed based upon IACUC requirements. For cell recruitment analysis and in vivo gene expression, tumors were harvested, as described below, 1, 3, 5, and 7 days after the initial C134 injection. Tumors were washed in PBS and finely minced into small pieces. Then tissues were digested in RPMI 1640 containing collagenase D (2 mg/mL; Roche) and DNase I (0.01 mg/mL; Roche) for 30 minutes at 37°C on a shaking platform. After collagenase digestion, the medium containing the mononuclear cells was strained and centrifuged at 400 × g for 10 minutes at 4°C, and the resulting cells were resuspended in RPMI 1640 supplemented with 1% FBS and penicillin/streptomycin, and then used for flow cytometry analysis and RNA extraction. For the CD8 depletion studies, mice were treated with RUX similar to that described above, but upon initiation of the RUX therapy, mice were randomized into anti-CD8 depletion or isotype treatment cohorts. Mice were treated with 100 μg of anti-CD8 (Clone 2.43, Bio X Cell) or the isotype control (Clone LTF-2, rat IgG2a. Bio X Cell) i.p. twice weekly throughout the experiment. Mice were then treated with IT C134 as described above. To quantify CD8 depletion, mice underwent a tail-vein bleed (1 week after initiating CD8 depletion), and the CD8⁺ T-cell populations were analyzed using FITC-conjugated anti-CD8b (clone H35-17.2; eBioscience).

For the acyclovir (ACV, Carlsbad Technology Inc.) and UV-inactivated virus treatment studies, a similar design was used, except UV-inactivated C134 (300 mj) was administered. For the ACV-treated cohort, the RUX/C134 treatment was performed as described above, except mice were provided with 1 mg/mL ACV in their drinking water beginning the day prior to virus administration and for 4 days after treatment (25).

67C-4 tumors were established in 6- to 8-week-old female C57BL/6 mice as described in the previous section. When tumors reached 25 to 150 mm³ in size, mice were randomized and administered either 3 doses of RUX (60 mg/kg) or 3 doses of vehicle i.p. daily for 3 days. On the fourth day, all mice were treated with C134 i.t. (3.5 × 10⁷ pfu in 100 μL 10% glycerol and PBS). On days 1, 3, and 5 after virus treatment, tumor samples were harvested and homogenized as mentioned previously. DNA was extracted by DNeasy blood and tissue kit (Qiagen) per the manufacturer's instructions. Virus recovery was measured by TaqMan quantitative PCR (22). Briefly, extracted DNA samples were incubated with the following HSV-specific primers and probes for HSV polymerase (sequences kindly provided by Dr. Fred Lakeman University of Alabama at Birmingham, Birmingham, Alabama): PolF (forward), 5′-ACC GCC GAA CTG AGC AGA C-3′; and PolR (reverse), 5′ -TGA GCT TGT AAT ACA CCG TCA GGT-3′. The fluorescent-labeled probe sequence was 5′-6FAM-CGC GTA CAC CAA CAA GCG CCT G-TAMRA-3′. HSV genome equivalents of the amplified product were measured from triplicate samples using a StepOne Plus real-time PCR system (Applied Biosystems) and compared against logarithmic dilutions of a positive control DNA sequence (10⁶–10¹ copies). Descriptive statistical analyses (mean and SD) were used to compare differences in DNA copy numbers between samples using Prism 7.0 statistical software (GraphPad).

Total RNA was isolated from tumor samples using the Direct-zol RNA Miniprep Plus kit (Zymo Research) according to the manufacturer's instructions. RNA quantity and purity was determined using a NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific). Two micrograms of total RNA was used to synthesize cDNA using SuperScript III Reverse Transcriptase (Life Technologies) according to the manufacturer's instructions. Quantitative real-time PCR was performed using the StepOne Plus real-time PCR system (Applied Biosystems). The primer pairs used were as follows: RIG-1: sense, TGTGGGCAATGTCATCAAAA, anti-sense, GAAGCACTTGCTACCTCTTGC; MDA5: sense, GGCACCATGGGAAGTGATT, anti-sense, ATTTGGTAAGGCCTGAGCTG; IFN-a: sense, AAAGAA ATGTAA GAAAGC TTTTGATGA, antisense, TACACTTTG GCTCAGGACTCATTT GAPDH: sense, GACAACTTTGGTATCGTGGAA, anti-sense, CCAGGAAATGAGCTTGACA (26, 27). PowerUp SYBR Green Master Mix (Thermo Fisher Scientific) was used to quantify the gene transcripts in 20 μL reactions according to the manufacturer's instructions. Cycling conditions included the initial step of 2 minutes at 50°C and 2 minutes denaturation at 95°C, followed by 40 thermal cycles of denaturation at 95°C for 15 seconds, annealing at 58°C for 15 seconds, and elongation at 72°C for 30 seconds. Results were expressed in relative copy numbers (RCN) as described elsewhere (28). Briefly, RCN = 2 − ΔCt × 100, with ΔCt calculated by subtracting the average Ct of housekeeping control (GAPDH) from the experimental sample Ct.

Single-cell suspensions from tumors were obtained as described previously (29). Briefly, tumors were washed in PBS and finely minced into small pieces. Then tissues were digested in RPMI 1640 containing collagenase D (2 mg/mL; Roche) and DNase I (0.01 mg/mL; Roche) for 30 minutes at 37°C on a shaking platform. After collagenase digestion, the medium containing the mononuclear cells was filtered and centrifuged at 400 × g for 10 minutes at 4°C, and the resulting cells were resuspended in PBS supplemented with 1% FBS and then used for flow cytometry analysis. Single-cell suspensions from tumors were lysed with RBC lysis buffer (Sigma) and blocked with 5% mouse Fc blocking reagent (2.4G2, BD Biosciences) in FACS buffer (1% FBS and 1 mmol/L EDTA in PBS). Cells were labeled with the following antibody staining panels for analysis of the adaptive immune cells: CD11b-Violet 421 (M1/70), CD4-BV785 (GK1.5), CD25-PE (7D4), CD8a-BV510 (53-6.7), CD3ϵ-BV 711 (145-2C11), CD44-APC, CD45-BV605, NKp46–PE-Cy7, and B220-AF488 (RA3-6B2) from BioLegend. Dead cells were excluded by staining with Live/Dead Near/IR staining (APC-Cy7; Thermo Fisher Scientific). Single samples were stained with the above staining panels for 30 minutes on ice and washed one time with FACS buffer. After labeling, cells were fixed in 1% paraformaldehyde and analyzed on a BD FACS LSR II (BD Biosciences). Analysis was carried out using the FlowJo software, version 10.0.3 (TreeStar Inc.).

For T-cell proliferation assays, splenocytes were prepared from mechanically dissociated and filtered spleens using a 70-μm cell strainer and a sterile 5 mL syringe plunger. Then, splenocytes were labeled with Cell Trace Violet (CTV) in accordance with the manufacturer's guidelines (Life Technologies). Dye levels were titrated to achieve the brightest, most uniform staining distribution and the best viability in culture for the cells that were being used. Briefly, cells are counted, washed once in warm PBS, and resuspended in prewarmed PBS containing 0.2% FBS labeling solution at a density of 20 × 10⁶/mL and incubated with an equal volume of CTV staining solution in PBS at 37°C for 12 minutes. RPMI supplemented with 1% FBS was added at 4 times of the original volume to absorb remaining free dye. Cells were centrifuged and resuspended in fresh media.

For proliferation and activation assays, splenocytes (5 × 10⁵) from the treated tumor-bearing mice were plated in round-bottom 96-well plates and stimulated with 10 μmol/L EphA2 peptide (671-FSHHNIIRL-679) or control peptide (498-SSIEFARL-505) for 6 hours. In parallel studies, CFSE-labeled splenocytes (5 × 10⁵) were incubated in 96-well plates alone (no cells) or with mitomycin C (50 μg/mL for 45 minutes) treated and washed 67C-4 cells at 37°C for 3 days. Samples were incubated with protein transport inhibitor containing 1 μL/mL brefeldin A (GolgiPlug, BD Biosciences) for 6 hours prior to flow cytometry staining (as described above) and CD8⁺ T lymphocytes were analyzed by flow cytometry for IFNγ intracellular staining, proliferation (CFSE), and activation (CD25).

Cellular lysates from tumor samples were collected on ice in disruption buffer (10 mmol/L Tris-Cl pH 8.0, 1 mmol/L EDTA, 1% Triton X100, 0.1% sodium deoxycholate, 0.1% SDS, 140 mmol/L NaCl, 20% β-mercaptoethanol, 0.04% bromophenol blue) with complete, mini protease inhibitor cocktail (Roche). The protein concentrations were determined using Pierce BCA Protein Assay Kit (Thermo Scientific). Samples were denatured at 98°C for 5 minutes, chilled on ice, separated by PAGE, transferred to a nitrocellulose membrane (Thermo Scientific) and blocked for 1 hour at room temperature with 5% dry milk (S.T. Jerrell Co.) or bovine serum albumin (Fisher Scientific). Membranes were incubated overnight at 4°C with primary antibody diluted in Tris-buffered saline with 0.1% Tween-20 (TBST). Primary antibodies against RIG-I (clone D14G6), MDA-5 (clone D74E4), and p-STAT-1 (clone 58D6) were purchased from Cell Signaling Technology and against actin (clone C4) from Chemicon. Membranes were repeatedly washed with TBST, incubated for 1 hour with HRP-conjugated goat anti-rabbit (Pierce) for RIG-I, MDA-5, and p-STAT-1 or goat anti-mouse (Pierce) for actin diluted in TBST (1:20,000 dilution) at room temperature, and subsequently washed with TBST. Membranes were developed using SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific) and exposed to X-ray film (Research Products International).

Statistical analysis was performed using Prism 7 (GraphPad Software). One-way ANOVA with correction for multiple comparisons (Holm–Sidhak or Kruskal–Wallis as specified) was used for analysis involving multiple cell lines or otherwise specified. For comparing tumor growth over time between two treatment groups, two-way ANOVA with Sidak multiple comparisons test was used. Survival was assessed using log-rank assay, and data were shown using Kaplan–Meier curves. For all analyses, the cutoff for statistical significance was set at P < 0.05. The following notation was used: (ns) P > 0.05; *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001.

We previously showed that basal ISG accumulation and rapid IFN signaling in ∼50% of human MPNST lines restricts oHSV infection and was a potential impediment to oncolytic HSV therapy (19). We were interested in how viral replication and the ensuing immune-mediated antitumor response were linked and their role in the oHSV antitumor effect. We, therefore, sought to develop a syngeneic MPNST model to evaluate how JAK/STAT signaling modulation influences the oHSV-1 antitumor effect in resistant MPNSTs. We screened four different murine MPNST lines for STAT1 activation and ISG expression to determine if murine MPNSTs share a similar resistance mechanism to human MPNSTs. Figure 1A shows that, consistent with our past studies (19), B76 cells were susceptible to C134 infection and permitted cell-to-cell spread (quantified by the EGFP expression over time), whereas the other MPNST lines tested (B96, 5NPCIS, and 67C-4) restricted C134 infection and spread. Immunoblotting analysis showed that IFN signaling (phosphorylated STAT1: p-STAT1) did not occur prior to infection or within 6 hours of C134 infection in the sensitive MPNST B76 cell line. In contrast, the more resistant lines demonstrated varied IFN responsiveness to C134 infection, with two of the lines (5NPCIS and 67C-4) exhibiting basal p-STAT1 activity prior to oHSV infection. Consistent with the p-STAT1 results, we also saw increased accumulation of RIG-I, our surrogate protein for ISG expression. The cell lines that had the highest basal p-STAT1 and ISG accumulation (67C-4 and 5NPCIS) were the most resistant to oHSV infection and spread, consistent with our prior human MPNST results (Fig. 1B).

To examine whether RUX pretreatment improved C134 infection and spread in resistant murine MPNSTs, similar to our previous findings in human MPNSTs, we exposed 67C-4 to increasing doses of RUX for two days and evaluated STAT1 activation (p-STAT1) and ISG protein expression (RIG-I and MDA-5). We also performed studies that showed that RUX (125–1,000 nmol/L) did not have a direct cytotoxic or antiproliferative effect on the tumor cells (Supplementary Fig. S1). RUX pretreatment decreased basal Stat activation and ISG accumulation (RIG-I and MDA-5) in a dose- and time-dependent manner, with maximal inhibition attained with pretreatment for 2 days with 250 nmol/L (Fig. 2A and B). Next, we examined how RUX pretreatment (250 nm) for 2 days with or without cotreatment affected STAT signaling and ISG accumulation, as well as subsequent oHSV gene expression, infection, and spread. Figure 2C shows that both regimens (pretreatment with or without cotreatment) diminished ISG accumulation, demonstrated by decreased RIG-I and MDA-5 expression. STAT1 activation was blocked completely by the combined regimen (pretreatment and cotreatment). Pretreatment and removal at the time of infection permitted STAT1 phosphorylation in both mock and infected cells, while reducing ISG expression prior to infection. We then examined the effect of RUX pretreatment on virus infection and spread. The results showed that in the untreated cells, C154 (GFP-expressing C134) did not readily infect or generate a cytopathic effect, whereas in the RUX-pretreated cells, C154 gene expression (GFP), infection, and cytopathic effect was enhanced (Fig. 2D). Kinetic analysis of the varying drug doses and their effect on C154 replication and spread over time is provided (Fig. 2E).

After showing that murine MPNSTs recapitulated a similar oHSV-resistance mechanism as human MPNSTs, we next sought to determine whether RUX pretreatment reduced ISG accumulation and STAT response in vivo. As shown in the schematic (Fig. 3A), C57BL/6J immunocompetent mice with established 67C-4 tumors were treated with RUX (60 mg/kg/d) for 3 consecutive days. One day later (immediately prior to C134 treatment), mice were sacrificed, and ISG expression and STAT activation were measured in their tumors. Our results showed that, consistent with in vitro studies, RUX pretreatment significantly reduced IFNα and ISG (RIG-I, MDA-5) gene expression (Fig. 3B). Immunoblotting analysis from the untreated and RUX-pretreated tumors also showed reduced ISG expression in vivo (Fig. 3C). A day after their final RUX dose, mice were treated with C134 i.t., and viral replication was measured by HSV qPCR. As shown in Fig. 3D, in the untreated mice, C134 did not replicate effectively and virus recovery declined after injection (2.7 × 10⁵ copies), but RUX pretreatment improved C134 replication (2.61 × 10⁷ copies), leading to a 96-fold increase in viral recovery by day 5 after C134 treatment. In conclusion, these results showed that RUX pretreatment reduced IFN signaling and ISG expression and improved viral replication in syngeneic 67C-4 tumors, similar to our earlier in vitro 67C-4 studies (Fig. 2A–C).

Next, we sought to determine if RUX pretreatment improved C134 antitumor activity in the oHSV-resistant model. Similar to the discussion above, 67C-4 tumors were established in C57BL/6J mice and RUX-treated for 3 consecutive days, followed by 2 i.t. doses of C134, 1 week apart (Fig. 4A). When used alone, neither RUX nor C134 injection had any antitumor activity in these resistant tumors. However, when the two therapies were combined, they significantly reduced tumor growth (Fig. 4B and C) and improve animal survival (Fig. 4D). To determine if the observed C134 replication advantage was integral to this antitumor efficacy, we repeated the RUX/C134 combination therapy but included matched cohorts where virus replication was suppressed using either C134 or UV-inactivated C134 in the presence of the HSV antiviral drug acyclovir (ACV). RUX pretreated tumor-bearing mice that were administered UV-inactivated C134 or were treated with C134 in conjunction with antiviral therapy did not exhibit any antitumor effect, suggesting that viral replication is necessary for the RUX/C134 antitumor effect (Supplementary Fig. S2).

Combination therapy produced an initial antitumor effect (tumor growth reduction) that could be explained by the improved C134 replication and direct oncolytic effect. However, the degree and timing of the antitumor effect appeared out of proportion to direct oncolytic activity (Fig. 3D). The antitumor activity and reduced tumor growth weeks after C134 treatment suggested that an immune-mediated mechanism also contributed to combination therapy's efficacy. We, therefore, repeated the studies and analyzed the immune cell infiltrates. Although no proportional differences between the treatment cohorts with regard to their overall T-cell, CD4⁺, or CD8⁺ T-cell populations were seen (Supplementary Fig. S3A–S3C), a significant increase in the activation status (CD44⁺ and CD25⁺) of the CD8⁺ T cells in the combination therapy group occurred (Fig. 5A–C; Supplementary Fig. S3D). This activation was unique to CD8⁺ T cells and did not occur in the CD4⁺ T-cell population (Supplementary Fig. S3E and S3F). By day 5 after virotherapy, the CD8⁺ T-cell population significantly increased in the RUX/C134-treated tumors (Fig. 5D). These data indicated that only Rux/C134 therapy was capable of increasing this effector population in the tumors. When we examined the response in the periphery, we found that both C134 and Rux/C134 combination therapy also induced CD8⁺ T cells in the spleen (Supplementary Fig. S4A and S4B).

To elucidate the mechanisms underlying T-cell restriction within the tumors, we next assessed expression of inhibitory receptor PD-1 on CD8⁺ T cells. Single-cell suspensions of enzymatically digested tumors were analyzed by flow cytometry. As demonstrated in Fig. 5E, RUX/C134 combination therapy significantly increased the PD-1–negative CD8⁺ T cells, a cytotoxic T-lymphocyte (CTL) population not subject to costimulatory PD-L1 restriction in the tumor. Our results showed that RUX/C134 combination therapy not only increased CTLs but also decreased regulatory T cells (Treg) in the tumors (Fig. 5F), including the described CD8⁺FoxP3⁺ T-cell population (Supplementary Fig. S4C; refs. 30, 31). These data showed that RUX/C134 therapy significantly shifted the CTL response in the tumors, increasing the effector T-cell-to-Treg ratio (Fig. 5G), and suggest that CD8⁺ CTLs contribute to the RUX/C134 antitumor efficacy (32).

Next, to determine if the CD8⁺ T-cell population was required for C134 antitumor activity, we repeated the in vivo studies and included a CD8⁺ T-cell–depleted cohort. The depletion conditions were validated by flow cytometry of peripheral blood and included an isotype-treated control cohort (Fig. 6A and B). In contrast to the isotype-treated cohort, CD8⁺ T-cell depletion eliminated the antitumor benefit of RUX/C134 therapy, indicating that the CD8⁺ T cells were integral to the antitumor activity (Fig. 6C and D).

Having shown that the combination therapy improved antitumor activity and confirmed that CD8⁺ T cells were integral to tumor regression, we next evaluated if combination therapy induced a tumor antigen–specific response. To assess this, we measured CTL activation and proliferation after tumor antigen exposure using day 5 after treatment splenocytes. As shown in Fig. 7A, no basal difference in T-cell activity between the 4 treatment cohorts was seen. However, when the splenocytes were incubated with mitomycin C–treated 67C-4 cells, CD8⁺ T cells from the RUX + C134 cohort produced a significantly higher IFNγ response than the CD8⁺ T cells from the other treatment cohorts. Exposure of cells to the specific 67C-4 tumor antigen EphA2, but not control peptide (Supplementary Fig. S5A and S5B), induced CD8⁺ T-cell proliferation and activation, represented by the CD25 expression (Fig. 7B). In summary, our data suggest that combination RUX/C134 therapy improves viral replication and antitumor immune activity, both of which are integral to the antitumor effect.

IFN and STAT1 activation are generally considered to be antitumor effectors due to their proapoptotic, antiangiogenic, genotoxic potentiating effects. However, high STAT1 activation also confers cancer therapy resistance (33). STAT1-mediated radioresistance has been described in many tumors, is associated with the expression of a subset of ISGs termed IRDS (IFN-related DNA damage resistance signature), and occurs in squamous cell carcinoma, breast cancer, head and neck cancer, colon cancer, prostate carcinoma, and gliosarcoma (34–36). A sustained STAT1 activation was also described in cisplatin-resistant ovarian cancer cells and human lung cancer cells resistant to the topoisomerase inhibitor etoposide (37, 38).

Our previous studies in human MPNSTs showed that basal IFN and ISG expression predicted oncolytic virotherapy resistance and that by pretreating the resistant cells with a JAK1/2 inhibitor, RUX, we could reset ISG expression in the cells and make them oHSV susceptible again. To test our hypothesis that ISG downregulation improves viral replication and enhances the oHSV antitumor effect, we sought a preclinical, syngeneic, resistant MPNST model. In this study, we showed that RUX pretreatment followed by oHSV treatment reduced tumor growth and improved survival in vivo. Our results showed that murine MPNSTs had a similar oHSV-resistance mechanism as human MPNSTs, composed of basal IFN activity, rapid STAT1 phosphorylation, and ISG overexpression that restricted oHSV infection. Antiviral ISGs defend the host cell from viral infection and replication. Viruses have evolved genes to evade these host antiviral proteins (39–42). HSV encodes numerous genetic mechanisms (e.g., γ₁34.5, Us11, and Us12) to subvert host mediated antiviral or immune-mediated responses. HSV also encodes genes (VHS and α27) that give it a selective transcriptional advantage over the host cell (43, 44). Our data showed that host RIG-I protein expression decreases in HSV-infected cells. However, we do not know if this was a targeted viral response or a function of viral transcriptional control of the cell (39–42).

RUX pretreatment reduced ISG expression and increased murine MPNST oHSV susceptibility both in vitro and in vivo. To test if RUX pretreatment enhanced oncolytic HSV antitumor activity, we used 67C-4 MPNST tumor cells because of their high ISG expression and STAT1 activation. We anticipated that this constitutive STAT1 pathway activation produced an environment unconducive to virus replication in vivo and would restrict oHSV antitumor activity. We found that, alone, neither RUX nor C134 treatment produced detectable antitumor effects. However, when the two were administered sequentially (RUX followed by C134), the combination reduced tumor growth and improved animal survival.

Although our viral recovery studies showed improved viral replication, the timing and persistence (2–3 weeks after virus administration) of the tumor growth suppression suggested an immune-mediated mechanism was also involved. We, therefore, examined the immune cell infiltrates in the tumors, with a focus on the adaptive immune response. Combination treatment with RUX followed by C134 increased T-cell activation and antitumor activity. The T-cell distributions (CD4⁺, CD8⁺, or total) in the tumor-infiltrating lymphocytes (TIL) were similar in all the treatment cohorts early. However, only mice pretreated with RUX and then treated with C134 exhibited CD8⁺ T-cell activation (CD25/CD44) and generated an antitumor response. To examine whether similar CTL activity was detectable in the periphery, we examined the lymphocyte populations in the spleen. Our results showed that CD8⁺ T-cell activity was increased in both oHSV-treated cohorts (C134 and RUX + C134) on days 3 to 5 after treatment. By day 7 after virus treatment, the infiltrating CD8⁺ T-lymphocyte population was significantly increased in the RUX + C134–treated cohort, and these CTLs did not express PD-1 and, therefore, appeared capable of increased activity. Based upon these findings, we sought to determine if CTLs were integral to RUX/C134 therapy. We depleted CD8⁺ T cells and confirmed their essential role in the oHSV-mediated antitumor effect. When the CD8⁺ T cells were exposed to tumor antigens, cells from animals treated with the combination therapy showed a more robust response (activation and proliferation), indicating that OV therapy can induce a specific antitumor antigen response and not simply an antiviral response.

We conclude that CTLs are integral to C134 therapy but that basal STAT signaling in MPNSTs impairs TIL function and prevents an effective antitumor immune response after oHSV therapy. We postulate that this is likely related to an IFN-induced chronic inflammatory state and T-cell immune exhaustion. Pathologic type I IFN production is well described in chronic viral infections and induces a dysfunctional immune state where T cells become exhausted by the chronic inflammatory state, thus unable to clear the virus (45). Paradoxically, chronic expression of an antitumor cytokine (type I IFN) creates both an inflamed, yet immunosuppressive, environment that prevents an effective antitumor immune response after oHSV treatment (46). Chronic inflammation promotes tumor initiation, progression, and metastasis by providing a tumor-supporting microenvironment. Evidence indicates that chronic inflammation also establishes immunosuppression that promotes immune exhaustion and limits CTL function, which has been well described in many tumors, such as gliomas, melanoma, gastrointestinal malignancies, and hepatic and breast carcinoma (47–50). Bellucci and colleagues and Liu and colleagues previously reported that activation of JAK1, JAK2, and STAT1 in tumor cells results in the upregulation of PD-L1 and inhibition of T-cell functions (51, 52). We showed that MPNSTs should be included in this list of inflammatory tumors.

In the case of MPNSTs, constitutive JAK/STAT signaling pathway activation not only acts as a resistance mechanism to protect tumor cells against oncolytic viruses and other antitumor modalities such as chemoradiotherapeutics (19, 53), but also protects tumor cells against the cytotoxic CD8⁺ T lymphocytes. We, therefore, postulated that modalities that enhance CD8⁺ T-cell cytotoxicity could further enhance the antitumor activity of our combination therapy. Inhibiting the interaction between the PD-L1–expressing cells and T lymphocytes by checkpoint inhibitors, such as anti–PD-1 or anti–CTLA-4, and the role of PD-L1 expression on tumor and myeloid immune cells, as well as on regulating the cytotoxic T-lymphocyte function, are key questions to be answered in ensuing studies. Although we have shown that CTLs are an endpoint effector of the oHSV-induced antitumor immune response, a complex set of immune regulatory events lead up to this CTL response. Their role in oHSV combination therapy has not been fully investigated. How RUX “resets” this chronic inflammatory environment, the role of both myeloid and CD4⁺ T-cell regulation of the CTL response, and whether the chronic inflammatory environment reestablishes or diminishes the antitumor CTL response have not been addressed. Our results showed that basal IFN/STAT1 activity in some MPNSTs produced a chronically inflamed antiviral, yet immunosuppressive, environment that limited oHSV efficacy. RUX pretreatment not only enhanced virus infection and replication but also boosted the immune-mediated antitumor immune response, improving survival and reducing tumor growth.

No potential conflicts of interest were disclosed.

Conception and design: M.G. Ghonime, K.A. Cassady

Development of methodology: M.G. Ghonime, K.A. Cassady

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M.G. Ghonime, K.A. Cassady

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.G. Ghonime, K.A. Cassady

Writing, review, and/or revision of the manuscript: M.G. Ghonime, K.A. Cassady

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M.G. Ghonime, K.A. Cassady

Study supervision: M.G. Ghonime

This study was supported by Hyundai Hope on Wheels, Alex's Lemonade Stand Foundation, CancerFreeKids, and Department of Defense (NFRP-IIRA).

The authors wish to thank Hannah Mohd Amir for her assistance with these studies and manuscript editing.

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.