STAT1 and NF-κB Inhibitors Diminish Basal Interferon-Stimulated Gene Expression and Improve the Productive Infection of Oncolytic HSV in MPNST Cells

Malignant peripheral nerve sheath tumors (MPNST) are a highly aggressive cancer of the peripheral nervous system derived from cells of the Schwann cell lineage. Patients have a median survival of 26 months following diagnosis (1). Beyond advances in surgical resection, further treatment with chemotherapy and radiation has not demonstrated an overall benefit to survival (2, 3). A promising alternative approach is the use of conditional replicating oncolytic herpes simplex viruses type 1 (oHSV) to treat MPNSTs. These oHSVs have been safely used in clinical trials in a number of cancer types, however, the tumor response has varied and we anticipate this would be true for patients with MPNSTs. We have previously demonstrated that MPNST cell lines exhibit variable oHSV susceptibility (4), and we have sought to understand the potential mechanisms of resistance that are detrimental to oHSV therapy.

Tumor resistance to oHSV has been attributed to oncogenic signaling cascade changes, namely, Ras-mediated suppression of protein kinase R (PKR) in the infected cells (5). PKR is a host antiviral kinase that limits γ₁34.5 deleted oHSV late-gene expression and replication (6). In brief, HSV produces double-stranded RNA during viral gene expression which induces PKR dimerization, autophosphorylation, and activation. Activated PKR phosphorylates eukaryotic initiation factor α (eIF2α), leading to translational arrest in the infected cell. Wild-type HSV counters translational arrest by expressing the neurovirulence gene γ₁34.5 which encodes a multifunctional viral protein, infected cell protein 34.5 (ICP34.5). ICP34.5 recruits a host phosphatase, protein phosphatase 1 alpha (PP1α), to dephosphorylate eIF2α thus restoring protein translation in the infected cell. Though wild-type HSV causes lethal encephalitis in the central nervous system (CNS), deletion of one or both copies of the diploid γ₁34.5 gene attenuates neurogenic virulence. It is believed that malignant cells partially complement the loss of γ₁34.5 through oncogenic processes such as Ras-induced suppression of PKR.

In addition to PKR, cells express a diverse set of pattern recognition receptors (PRR) that detect pathogen-associated molecular patterns such as viral nucleic acids. Stimulation of certain PRRs including DEAD (Asp-Glu-Ala-Asp) box polypeptide 58 (DDX58; commonly known as retinoic acid-inducible gene I; RIG-I), interferon induced with helicase C domain 1 (IFIH1; commonly known as melanoma differentiation antigen 5; MDA5), transmembrane protein 173 (TMEM173; commonly known as stimulator of interferon genes; STING), and members of the Toll-like receptor family ultimately lead to the activation of transcription factors which promote expression of the type I IFNs IFNα and IFNβ which are potent antiviral cytokines. Secreted type I IFN interacts with transmembrane IFNα receptors (IFNAR) in an autocrine and paracrine manner leading to the activation of the intracellular Janus kinases JAK1 and TYK2. Activation of JAK1 and TYK2 leads to phosphorylation of signal transducer and activator of transcription 1 (STAT1) and STAT2, which together with interferon regulatory factor 9 (IRF9) form the heterotrimeric transcription factor interferon-stimulated gene factor 3 (ISGF3). The ISGF3 complex localizes to the nucleus where it promotes transcription of several hundred interferon-stimulated genes (ISG) that contain interferon-stimulated response elements (ISRE) in their promoter regions. ISGs have diverse functions, modulating viral and cellular functions to promote intrinsic antiviral resistance. ISGs can directly inhibit mechanisms unique to certain viruses (e.g., myxovirus resistance 1, MX1), inhibit cellular processes involving transcription and translation (e.g., PKR; and 2′-5′-oligoadenylate synthetase 1, OAS1), or increased expression of PRRs and IFN/STAT1 signaling modulators to promote amplification of the antiviral response (e.g., RIG-I; MDA5; interferon-induced protein with tetratricopeptide repeats 3, IFIT3).

The research presented here tests the hypothesis that ISG upregulation contributes to oHSV resistance in a subset of MPNST tumor lines. Our results show that: (i) PKR activation is not specifically associated with oHSV-resistant phenotypes; (ii) STAT1 activation is observed in 10 of 21 MPNST cell lines and is statistically associated with diminished productivity of both a first generation Δγ₁34.5 oHSV and a second generation Δγ₁34.5 oHSV, C134, capable of evading PKR-mediated translational arrest; (iii) resistant MPNST cell lines exhibit greater ISG expression than oHSV-sensitive lines prior to oHSV infection suggesting they are primed toward an antiviral state; (iv) pretreatment of resistant MPNST tumor lines with a small-molecule JAK inhibitor reduces basal ISG expression and improves viral replication and spread; (v) conversely, either IFN stimulation or stable ISGF3 overexpression in MPNST cell lines increased ISG expression leading to decreased oHSV productivity; and finally (vi) we have provided evidence that basal expression of ISGs may be dependent on the NF-κB signaling network.

Human MPNST cell lines or their firefly luciferase–expressing derivatives (-luc; ref. 4) as well as MPNSTs lines derived from a genetically engineered mouse model (7) have been previously described. All MPNST cell lines were maintained in growth media containing DMEM with 10% FBS and 10 mmol/L l-glutamine (Sigma-Aldrich). The African green monkey kidney cell line Vero was obtained from the ATCC and maintained in minimum essential medium with 5% bovine growth serum. All viruses except M201 have been described previously (4, 8) and are summarized in Supplementary Table S1. M201 is an oHSV with CMVp–GFP inserted between UL3 and UL4, and murine IL12 inserted in γ₁34.5 locus and was constructed using homologous recombination with M002 (mIL-12 virus) viral DNA (8) and pCK1029 plasmid DNA using a method described previously (6). Sequential plaque purification was performed using Vero cells using eGFP expression and candidate viruses were verified by southern blot analysis (data not shown).

The small-molecule inhibitors ruxolitinib and TPCA-1 (Selleckchem) were stored as 10 mmol/L aliquots in DMSO (Sigma-Aldrich). Primary antibodies were obtained as follows: p-T446-PKR (3076), p-S51-eIF2α (3398), eIF2α (2103), p-Y701-STAT1 (9167), STAT1 (9172), MDA5 (5321), RIG-I (3743), p-S536–NF-κB/P65 (3033), NF-κB/P65 (3987) IκB (4814), p-MEK1/2 (2338), MEK1/2 (4694), p-ERK1/2 (4376), and ERK1/2 (9102) from Cell Signaling Technology; MX1 (13750), OAS1 (14955), IFIT3 (15201), STAT2 (16674), and IRF9 (14167) from Proteintech; HSV ICP4 (H1A021, Virusys Corporation), PKR (sc707, Santa Cruz Biotechnology), and β-actin (A2228, Sigma-Aldrich). Horseradish peroxidase–conjugated goat anti-mouse (Immunopore) and goat anti-rabbit antibodies were used as secondary antibodies.

Viral recovery (replication) assays have been described previously (4) with the following modifications: virus inoculation was performed in 150 μL regular growth media (10% FBS) for 1 hour followed by addition of growth media up to 1 mL per well in 24-well plates. Viral spread assays have been described previously (4). In brief, cells were seeded 1.5 × 10⁵ per well in 24-well plates and infected at an multiplicity of infection (MOI) of 0.1 with GFP-expressing viruses or mock infected (media alone). After 48 hours, we performed flow cytometry to assess the percentage of cells positive for viral GFP. Flow Cytometry Absolute Count Standard fluorescent beads (Bangs Laboratories) were added to the final suspension of cells to calculate the ratio of cells surviving infection compared with mock-infected samples. All samples were collected in triplicate and the means reported.

Cloning methods and construction of lentivector plasmids are provided as supplemental information. Lentivirus production has been described previously (4). Stable cell lines were produced via lentiviral transduction of the target cell lines. Hygromycin (Sigma-Aldrich) selection (300 μg/mL) was applied not earlier than 48 hours after transduction.

Cellular lysates were collected on ice in RIPA 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) with protease inhibitor cocktail (Roche) and diluted in 4x sample buffer (240 mmol/L Tris-Cl pH 6.8, 40% glycerol, 4% SDS, 20% β-mercaptoethanol, 0.04% bromophenol blue). Samples were denatured at 98°C for 5 minutes, chilled on ice, separated by PAGE, and transferred to a nitrocellulose membrane (Thermo Fisher Scientific) and blocked for 1 hour at room temperature with 5% dry milk (S.T. Jerrell Company) or BSA (Thermo Fisher Scientific). Membranes were incubated overnight at 4°C with primary antibody diluted in TBS with 0.1% Tween-20 (TBST). Membranes were repeatedly washed with TBST, incubated for 1 hour with secondary antibody diluted in TBST (1:20,000) at room temperature, and subsequently washed with TBST. Membranes were wetted with SuperSignal West Pico Chemiluminescent Substrate (Thermo Fisher Scientific) and exposed to X-ray film (Research Products International).

The STS26T-luc and ST8814-luc cell lines used in reporter assays have been previously transduced with firefly luciferase under a constitutive cytomegalovirus (CMV) promoter. These cell lines were transduced via lentivirus with a nanoluciferase reporter (with a C-terminal PEST sequence; Promega) under the control of either ISRE or NF-κB promoters. Luminescence assays were performed in opaque 96-well plates with the Nano-Glo Dual Luciferase Reporter Assay (Promega) according to the manufacturer's instructions. Luminescence was measured using a FLUOStar Optima (BMG LABTECH) plate reader. Nanoluciferase activity was normalized to that of firefly luciferase and reported as arbitrary units. Data were collected in triplicate or quadruplicate.

Statistical analysis was performed using Prism 5 (GraphPad Software). One-tailed Mann–Whitney U test was used for analysis involving multiple cell lines. Student t test was used for inhibitor and transduction experiments within individual cell lines. 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.

To assess the contribution of antiviral signaling pathways to oHSV resistance in MPNSTs, we assessed PKR activation and eIF2α phosphorylation in response to a Δγ₁34.5 oHSV (R3616, kindly provided by Dr. Bernard Roizman, University of Chicago, Chicago, IL). The relevant characteristics of R3616 and other viruses used in the following experiments are provided in Supplementary Table S1. We first determined the susceptibility of 8 human and 13 mouse MPNST cell lines by viral recovery assay 24 hours after cells were infected at a MOI of 1. Titers of recovered virus ranged from 7.9 × 10³ to 4.1 × 10⁵ plaque-forming units (PFU) for human cell lines and 1.5 × 10³ to 2.0 × 10⁵ PFU for mouse lines (Fig. 1A and B). Although mouse lines yielded 3-fold lower average titers of virus than human-derived lines (3.2 × 10⁴ and 9.5 × 10⁵ PFU, respectively), the distributions of human and mouse lines were statistically indistinguishable (Supplementary Fig. S1). Immunoblots against phosphorylated PKR (p-PKR) and p-eIF2α in human cell lines, or p-eIF2α in mouse cell lines, revealed PKR activation and eIF2α phosphorylation following R3616 infection (Fig. 1C and D) at 12 hours postinfection in nearly all cell lines tested. There was no apparent difference in p-PKR/p-eIF2α between cell lines with high or low viral recovery. We conclude that activation of PKR is not sufficient to exclusively define the resistant phenotypes observed in MPNST cell lines.

Because deletion of the HSV γ₁34.5 gene increases HSV-1 sensitivity to type I IFNs (9) which activate STAT1, we hypothesized that oHSV-induced STAT1 activation was associated with decreased viral productivity in MPNST cells. We determined that 6 hours postinfection was the optimal time to observe STAT1 Y701 phosphorylation (Supplementary Fig. S2). R3616 infection induced STAT1 activation in 3 of 8 (38%) human (Fig. 2A) and in 7 of 13 (54%) mouse cell lines (Fig. 2B). When exposed to exogenous IFNβ (200 IU/mL) STAT1 Y701 phosphorylation was evident in all human MPNST cell lines indicating that mechanisms for signal transduction were functional (Supplementary Fig. S3). When R3616 titers from all MPNST cell lines were sorted into STAT1-unresponsive (pSTAT1−) and STAT1-responsive (pSTAT1+) groups, cell lines which were STAT1 responsive were associated with significantly lower viral recovery (Fig. 2C). To further test the association of the STAT1 response of each cell line with viral productivity, we assessed viral spread within an in vitro monolayer. In this assay, the percentage of cells infected with an eGFP-expressing Δγ₁34.5 virus (C101) in a multistep infection (MOI = 0.1, 48 hours postinfection) was measured by flow cytometry. In general, MPNST cell lines tended to be resistant to the spread of C101 in the multistep assay, however, permissive cell lines which supported spread were associated with an unresponsive STAT1 phenotype (Fig. 2D). To determine whether differences in STAT1 activation was cytoprotective following oHSV infection, we measured the number of gated cells by flow cytometry at 48 hours postinfection following multistep infection with C101 and compared the counts to mock-infected cells. The results showed a trend toward higher cell counts (lower cytotoxicity) after C101 infection in STAT1-responsive cell lines, however, similar to the previous assessment, the majority of cell lines were resistant to the cytotoxic effects of C101 (Fig. 2E).

To identify whether the STAT1 response was associated with diminished oHSV productivity in the setting of a Δγ₁34.5 oHSV capable of PKR evasion, we repeated the spread and cytotoxicity studies using C134, a chimeric Δγ₁34.5 oHSV expressing the human CMV IRS1 gene product which inhibits PKR-mediated translational arrest. Immunoblots against PKR and eIF2α verified that C134 inhibited PKR-induced eIF2α phosphorylation in MPNST tumor cells (Fig. 3A) similar to what has been observed in other cell lines (10). In a viral spread assay with C154 (an eGFP-expressing variant of C134), we found that similar to the results with C101, the spread and cytotoxic effect of C154 was significantly diminished in STAT1-responsive cell lines (Fig. 3B and C). Next, to identify whether the productivity of wild-type HSV-1 was associated with STAT1 response, we repeated the spread and cytotoxicity assays with a wild-type HSV-1(F) that expresses eGFP (M2001). The results show that unlike Δγ₁34.5 oHSVs, the STAT1 response was not significantly associated with wild-type HSV-1 spread or cytotoxicity in the MPNST cell lines (Fig. 3D and E). Both viral spread and cytotoxicity induced by M2001 was generally high in all cell lines compared with the Δγ₁34.5 attenuated viruses.

These results demonstrate that the restricted productivity of Δγ₁34.5 oHSVs is significantly associated with the capacity of a cell to activate the STAT1 signaling cascade in response to Δγ₁34.5 oHSV infection. This association exists irrespective of PKR-mediated translational arrest inasmuch as the spread and cytotoxicity of C134, a Δγ₁34.5 HSV capable of PKR evasion, is also restricted in these STAT1-responsive cell lines. In contrast, the productivity of wild-type HSV-1 was not associated with the capacity for a STAT1 response in similar assays.

Next, to determine whether the observed STAT1 activation was directly responsible for diminished Δγ₁34.5 viral productivity in resistant MPNST cell lines, we inhibited STAT1 Y701 phosphorylation using the JAK1/2 inhibitor ruxolitinib. We hypothesized that by inhibiting this response, Δγ₁34.5 oHSV productivity could be improved. To test this hypothesis, we treated the oHSV-resistant MPNST cell lines STS26T-luc and ST8814-luc with 250 nmol/L ruxolitinib (Rux) or vehicle (DMSO) following Δγ₁34.5 oHSV infection. Cotreatment with ruxolitinib did not improve Δγ₁34.5 oHSV infection and spread compared with DMSO in multistep (low MOI) assay with C101 (Fig. 4A) despite the fact that ruxolitinib cotreatment prevented STAT1 Y701 phosphorylation (Fig. 4G). In viral recovery assays, ruxolitinib cotreatment did not improve R3616 replication in a single-step (MOI = 1) assays (Fig. 4B), but did slightly improve R3616 replication in multistep (MOI = 0.1) replication assays in both cell lines (Fig. 4C).

Multistep replication assays, using a low virus/cell ratio and longer time points (24 vs. 48 hours postinfection), challenge the virus to undergo multiple rounds of replication and spread beyond the initially infected cells. The modest increases in viral titer observed in the multistep assay (Fig. 4C), but not the single-step assay, led us to hypothesize that uninfected cells in the multistep assay benefitted to some degree from “pre-treatment” with ruxolitinib prior to becoming infected. To further test the effects of pretreatment, the above experiments were repeated by exposing cells to ruxolitinib for 48 hours followed by removal of the inhibitor prior to infection and addition of DMSO following infection. The results show that pretreatment alone with ruxolitinib (Rux + DMSO) significantly improved the spread of C101 (Fig. 4D) as well as titers of R3616 in both single-step and multistep assays in both cell lines (Fig. 4E and F) compared with DMSO pretreatment (DMSO + DMSO). Interestingly, pretreatment alone, whereby cells were washed of ruxolitinib prior to infection did not inhibit subsequent oHSV-induced STAT1 phosphorylation (Fig. 4G). Sequential pretreatment and cotreatment with ruxolitinib (Rux +Rux) further improved viral spread beyond pretreatment alone (Fig. 4D), however, it did not significantly improve viral replication in either the single-step or multistep viral recovery assays (Fig. 4E and F). Similar improvements were obtained with a M002 series oHSV (8; Supplementary Fig. S4). To explain the unique effects of pretreatment, we hypothesized that ruxolitinib inhibited low-level stimulation and activation of the JAK/STAT pathway in the STS26T-luc and ST8814-luc cell lines. To test this hypothesis, we stably transduced these cell lines with a nanoluciferase reporter under the control of a series of ISRE promoter elements and showed that ruxolitinib diminished the basal (uninfected) ISRE reporter activity in a concentration-dependent manner relative to DMSO (Fig. 4H). An assessment of protein expression by Western blot analysis revealed that the expression of five representative ISGs (MDA5, RIG-I, MX1, IFIT3, and OAS1) decreased following 48-hour treatment with ruxolitinib as compared with DMSO (Fig. 4I). We conclude that basal STAT1/ISRE promoter activity leads to ISG expression which negatively impacts the productive infection of oHSVs.

To determine the extent to which basal ISGs expression occurs among other MPNST cell lines, we evaluated the relative levels of five ISGs (MDA5, RIG-I, MX1, IFIT3, and OAS1) by immunoblot in all human MPNST cell lines. The results show that for the five probed ISGs, greater protein expression was detected in the cell lines ST8814-luc, T265-luc, 2XSB, STS26T-luc, and 90-8-luc whereas YST-1, NMS2-PC, and S462-luc had low or undetectable ISG expression (Fig. 5A). The low ISG expression in S462-luc is notable as this cell line was the most permissive to oHSV infection producing the highest viral titers of MPNST cell lines tested. We hypothesized that higher basal ISG expression would increase oHSV resistance in S462-luc. To test this, S462-luc cells were exposed to different concentrations of IFNβ for 24 hours and probed for ISGs by Western blot analysis. Figure 5B shows the associated increase in ISG expression with increasing concentrations of IFNβ, whereas IFNβ-treated cells which were infected with C101 in a multistep assay were less supportive of cell to cell spread in a manner dependent on the concentration of IFNβ. Because ISGs are only transiently expressed in S462-luc in response to IFNβ (data not shown) we stably coexpressed the ISGF3 component transcription factors (STAT1, STAT2, and IRF9), a strategy which has previously been reported to elevate basal ISG expression (11). Overexpression of ISGF3 in S462-luc similarly resulted in increased ISG expression (Fig. 5C) and was associated with greater than 10-fold reduction of R3616 titers in both single-step and multi-step replication assays (Fig. 5D). We conclude that elevated ISG expression in MPNST cell lines diminishes oHSV productivity.

Recent reports have implicated the transcription factor NF-κB, specifically the p65/RelA subunit, in promoting low-level expression of IFNβ resulting in higher basal ISG expression (12). We hypothesized that MPNST cell lines with constitutively active NF-κB increased ISG expression in oHSV-resistant cell lines. To assess this, immunoblots were performed and showed relatively high p65 phosphorylation in the resistant cell lines STS26T-luc and ST8814-luc (Fig. 6A) suggesting elevated NF-κB activity. To identify whether basal ISRE transcriptional activity and ISG expression was related to NF-κB transcriptional activity, we stably expressed nuclear factor of kappa light polypeptide gene enhancer in B-cell inhibitor, alpha (NFKBIA; commonly known as inhibitor of κB; IκB), a native inhibitor of NF-κB, with S32A/S36A mutations that prevent proteasomal degradation and act as an NF-κB “super repressor.” The results show that IκB super repressor (IκB-SR) expression decreases both NF-κB and ISRE reporter activity (Fig. 6B). In contrast, control transduction had no effect upon NF-κB or ISRE activity. IκB-SR expression also led to decreased expression of ISGs (Fig. 6C). Together these results suggest that constitutive stimulation of the NF-κB pathway can be related to the basal stimulation of ISRE elements and ISG expression.

Previous reports have shown that the small-molecule inhibitor TPCA-1, a dual inhibitor of IkB kinase (IKK; a positive regulator of NF-κB) and JAK1, can benefit the productivity of the oncolytic viruses vesicular stomatitis virus (VSV) and encephalomyocarditis virus (EMCV; refs. 13, 14). We hypothesized that TPCA-1 could also benefit oHSV by downregulating expression of ISGs. Indeed, TPCA-1 inhibited the activity of both NF-κB (Fig. 6D) and ISRE (Fig. 6E) nanoluciferase reporters in a concentration-dependent manner and reduced ISG expression in the treated cells (Fig. 6F). To test the effect of TPCA-1 on oHSV productivity, we used combinations of pretreatment and cotreatment with TPCA-1 or DMSO, similar to the ruxolitinib studies described above. As with ruxolitinib, cotreatment (DMSO + TPCA1) with TPCA-1 did not significantly improve Δγ₁34.5 infection and spread, however, pretreatment (TPCA1 + DMSO) was sufficient to significantly improve the spread of both of the Δγ₁34.5 recombinants C101 and M201 in the ST8814-luc and STS26T-luc cell lines (Fig. 6G and H). Interestingly, in contrast to the C101 and M201, the Δγ₁34.5 PKR evasion virus C154 benefited from either TPCA-1 cotreatment or from pretreatment alone (Fig. 6I). Pretreatment followed by cotreatment (TPCA1 + TPCA1) further improved spread in the majority of conditions tested.

The oHSVs engineered within our laboratory are derived from HSV-1 mutants with dual deletions of the γ₁34.5 neurovirulence gene rendering them safe to administer in the CNS (15). Although normal cells restrict Δγ₁34.5 oHSV replication, a partial explanation for the oncolytic selectivity of Δγ₁34.5 HSVs for cancerous tissue is that malignant transformation of these cells results in an aberrant antiviral response that complements viral loss of the γ₁34.5 gene. However, despite the wide variety of cancer types that are susceptible to oHSVs, resistance is commonly observed. In an in vivo setting, resistance can be attributed to inefficient delivery methods, low interstitial penetration, tumor heterogeneity, and activation of the innate and adaptive immune response. However, resistance is observed in vitro. This implies that the cancer cells themselves can independently restrict or resist oHSV infection and replication through intrinsic mechanisms.

Our previous work showed that the abundance of viral entry receptors was not responsible for restricted oHSV replication and growth in MPNST tumor cells (4). This suggests that an intracellular response contributes to intrinsic oHSV restriction. There are conflicting reports regarding the role of Ras signaling and its benefit to oHSV in the context of MPNSTs (16, 17). Our determination of the MEK/ERK status (Supplementary Fig. S5) shows that this pathway can be activated in both permissive and resistant MPNST cell lines similar to what was reported by Mahller and colleagues (16). Together with our observations regarding PKR, this suggests that Ras/MEK/ERK and PKR status are not sufficient to explain oHSV resistance in MPNST cells. Furthermore, an oHSV capable of evading the PKR response, C134, was also restricted in resistant MPNST cell lines despite inhibiting PKR activation. Although PKR-mediated translational arrest is a critical obstacle, it appears not to be the sole impediment to Δγ₁34.5 oHSV replication in MPNST cells.

We demonstrated that viral productivity in MPNST tumor lines was inversely associated with an oHSV-induced STAT1 response; however, STAT1 inhibition using ruxolitinib during HSV infection did not improve oHSV replication and spread. Instead, ruxolitinib pretreatment was necessary and sufficient to improve oHSV productivity. One potential explanation for the observable STAT1 response in oHSV-resistant cells is that these cells have a greater basal upregulation of ISGs. A subset of ISGs are pattern recognition receptors (e.g., MDA5 and RIG-I) which lead to the expression of type I IFN and subsequent phosphorylation of STAT1 in response to infection. This capacity to respond to oHSV infection with STAT1 phosphorylation may therefore reflect (or act as a surrogate marker for) the coexpression of other ISGs which are directly antagonistic to oHSV replication and spread. Therefore, suppressing the STAT1 response at the time of infection with ruxolitinib cotreatment is insufficient to improve oHSV replication because basal ISG accumulation is sufficient to impart this viral resistance. The downregulation of basal ISRE activity and ISG expression following ruxolitinib pretreatment supports the hypothesis that reduced accumulation of antiviral effectors is sufficient to improve oHSV infection and spread. The presence of faint p-Y701 STAT1 staining in some of the uninfected resistant MPNST cell lines (e.g., STS26T-luc and A382 in Fig. 2A) supports the possibility that basal STAT1 activity contributes to the expression of ISGs in these cell lines. Finally, consistent with our observation that viral productivity of both Δγ₁34.5 (9, 18) and C134 (Δγ₁34.5 IRS1) are associated with the capacity for a STAT1 response, these viruses have been previously been reported to be sensitive to type I IFN pretreatment (9, 18) whereas wild-type HSV-1 remains insensitive to those effects (9, 18).Although basal expression of ISGs has been previously documented in a number of tumor-derived cell lines (19–23), this has not been shown for MPNSTs. However, in the context of oHSV infection, Mahller and colleagues showed through gene expression analysis that the “JAK/STAT pathway” and “Tyrosine Phosphorylation of STAT protein” were significantly upregulated in response to G207 (a lacZ-expressing variant of R3616) in a panel of MPNST cell lines (24). Interestingly, knockdown of suppressor of cytokine signaling 1 (SOCS1), a negative regulator of STAT1, decreased the viral titers of the G207 by more than 10-fold in the highly permissive cell line S462.

Activation of the JAK/STAT pathway and ISG expression limits other oncolytic viruses including vesicular stomatitis virus (13, 19, 22, 23,25), measles virus (20), Newcastle disease virus (21), respiratory syncytial virus (26), Semliki Forest virus (27), and adenovirus (28, 29). In several of these studies, inhibitors of JAK/STAT1 signaling enhanced viral productivity (19, 22, 23, 25,27). With respect to HSV, a head and neck squamous cell carcinoma, which became radioresistant through the upregulation of STAT1 and other ISGs, suppressed the replication of oHSV R3616 by 40-fold compared with the original tumor (30). Haseley and colleagues found that overexpression of the extracellular matrix protein cysteine-rich 61 (CYR61) induced upregulation of type I IFNs and ISGs which suppressed oHSV productivity in glioma cells (31). Although Δγ₁34.5 HSVs are known to be sensitive to exogenous IFN (9, 18), presumably through the upregulation of ISGs by STAT1, these appear to be the only reports in the literature between a cancer-associated STAT1/ISG signature and resistance to oHSV.

The STAT1/ISG expression signature has been detected in a number of patient tumor specimens including glioblastoma (32, 33), squamous cell carcinoma (34), melanoma (35), leukemia (36), breast (36–38), ovarian (36), and pancreatic (28) cancers suggesting that cancer-associated STAT1/ISG expression is not an in vitro artifact. In these reports, it is not possible to determine whether the STAT1 signature is a phenotype driven by the intrinsic nature of the tumor cells (autocrine activation) or whether expression of this signature occurs in response to inflammatory stimuli from the surrounding microenvironment (paracrine activation). Regardless of the mechanism, STAT1 and ISG expression signatures have been implicated in the resistance of cancer to radiation (30, 39, 40) and chemotherapy (39, 41, 42). ISG expression in MPNSTs may offer a possible explanation for the lack of efficacy by these treatment modalities in patients with MPNSTs (2, 3).

It has been shown that innate immune effectors including microglia, macrophages, and natural killer cells actively restrict replication of oncolytic HSV in vivo (43, 44). However, oncolytic viruses may be beneficial as an immunotherapy by provoking an immune response, specifically involving natural killer cells and cytotoxic T cells, leading to recognition and clearance of the tumor, despite the detriment to viral replication and oncolysis (45, 46). Indeed, paradoxical responses to oHSV have been reported whereby tumor cells which were characterized as oHSV resistant in vitro were more susceptible to the antitumor effect of oHSV in vivo (46). Although we have demonstrated associated STAT1 activation to be detrimental to oHSV productivity in vitro, it will be necessary to identify how tumor cells that are STAT1-responsive interact with the peripheral immune elements in this context because the efficacy of certain immunotherapies and the recruitment of CTLs is dependent upon the type I IFN response (47). In addition, the use of powerful immunosuppressants such as JAK inhibitors may counteract the effects of immunotherapy or diminish the safety profiles which have been previously established for oHSV.

Finally, the driver for basal ISG expression in tumor cells remains incompletely understood. Virally induced NF-κB activation is known to promote IFNβ transcription (48), but the extent to which the constitutive NF-κB activation, which is commonly overserved in certain cancers, affects IFNβ expression is unclear. There is some evidence that upstream signaling components of the NF-κB pathway, independent of NF-κB–driven gene expression, may be involved in STAT1 cross-talk (49, 50). Further work is needed to determine the prevalence of NF-κB–related ISG expression and the complete mechanism in MPNSTs.

In conclusion, our current research has identified a previously unexplored component of MPNST resistance to Δγ₁34.5 oHSVs. We believe the novel finding of basal ISG expression in MPNST cells may have further implications for MPNST biology and their treatment with conventional antitumor therapies.

J.M. Markert has ownership interest (including patents) in oHSV, K.A. Cassady has patent interests in oHSV. No potential conflicts of interest were disclosed by the other authors.

Conception and design: J.D. Jackson, J.M. Markert, S.L. Carroll, K.A. Cassady

Development of methodology: J.D. Jackson, S.L. Carroll, K.A. Cassady

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J.D. Jackson, J.M. Markert, L. Li, S.L. Carroll, K.A. Cassady

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.D. Jackson, K.A. Cassady

Writing, review, and/or revision of the manuscript: J.D. Jackson, J.M. Markert, S.L. Carroll, K.A. Cassady

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

Study supervision: J.M. Markert, K.A. Cassady

The authors thank the UAB Comprehensive Flow Cytometry Core, UAB Heflin Genetics Center for Genomic Science, and Dr. Bernard Roizman (University of Chicago, Chicago, IL) for providing R3616.

Funding for this research was provided through DOD-W81XWH-10-NFRP-IIRA-NF1001157, P20 CA151129-03, P01 CA71933-15, and P50 CA97247-05.

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.