Oncolytic Virotherapy Blockade by Microglia and Macrophages Requires STAT1/3

Oncolytic viruses (OV) have been used as a therapeutic arsenal for specifically destroying cancer cells through oncolysis, a killing mechanism characterized by cancer cell lysis through the course of lytic virus replication (1). In addition to direct cell killing by the virus, it has been demonstrated that a virally induced immune response plays a pivotal role in OV therapy (2). As OVs can kill cancer cells via a mechanism distinct from the killing effects of conventional chemotherapy and radiotherapy, these viruses may be ideal for treating cancers that are nonresponsive to conventional treatment. Among the various OVs, those that are herpes simplex virus type I based are the furthest advanced. A herpes virus–based OV (T-Vec) was recently approved by the FDA for the treatment of melanoma after a successful completion of clinical trials in North America (3, 4).

Glioblastoma multiforme (GBM) is a treatment-refractory brain tumor with a poor prognosis (5). The most intensely investigated oncolytic HSV-1 (oHSV-1) viruses for treating GBM are mutant HSV-1s with deletions in viral gene ICP6 (6–8) or neurovirulent γ34.5 or both (9–12). Despite their excellent safety profile, the clinical efficacy of many oHSV-1s has been disappointing (13, 14).

Among many possible factors, one important reason for this lack of efficacy can be attributed to heterogeneity among tumors. Clinically, the tumor bed has been observed to be highly heterogenetic at the cellular level, and a significant portion of cells in tumor mass is nontumorous (15), which, by OV standards, makes them resistant to OV replication. We termed these cells tumor-associated nonpermissive (TANP) cells.

Microglia/macrophages are important innate immune cells in GBM and many other non-CNS tumors. They are probably the most common nontumor cells among all types of cellular infiltrates (15–18), comprising around 40% (range, 5%–78%) of the content of the total tumor mass (19, 20) for glioma. Badie and Schartner have further verified the source of microglia/macrophages in a rodent GBM model. They showed that microglia/macrophages contribute up to 46% of the cellular mass of a rodent GBM tumor, with up to 34% originating from microglia (CD45⁻/CD11b⁺ cells) present in the brain since embryonic development and the remaining 12% originating from blood-derived macrophages (identified as CD45⁺/CD11b⁺ cells; ref. 21).

Tumor-associated microglia/macrophages play an important role in tumor progression (15–17, 22). There is a positive correlation between the number of microglia/macrophages and the malignancy of the brain tumor. Malignant gliomas, such as glioblastomas and anaplastic gliomas, showed the largest number of mixed cell populations containing microglia/macrophages, and low malignancy glial tumors contain fewer microglial cells (18). Regardless of the origin of macrophages in the tumor microenvironment, a large body of work has suggested that tumor cells communicate with macrophages (15). Importantly, the communication has been shown to occur in both directions, in that glioma cells attract macrophage infiltration, while macrophages promote glioma growth and metastasis (15, 17).

The abundant microglia/macrophages in human GBM compromise the major component of TANP cells and contribute significantly to the poor efficacy of OVs. A previous study found that chemical ablation of microglia/macrophages subsequently enhances the antitumor effect of oHSV-1 in a glioma-bearing rodent model (23). Even for some so-called “cold” tumors that do not have large number of macrophages in the tumor mass, oncolytic virotherapy often turns the tumor to “hot” with significant macrophage infiltration as fast reacting innate immune responses. Although those macrophages are probably M1-type that are important in killing tumor cells as shown by a recent study (24), they are equally or more effective in clearing out the OVs to hinder the antitumor effectiveness of OV therapy.

As the most predominant TANP cells in GBM, a better understanding of this interaction is essential for the development of novel strategies for oncolytic virotherapy against GBM and other solid tumors. This report addresses the mechanism and effect of microglia/macrophages in response to oHSV-1–infected human glioblastoma and prostate cancer cells. We hypothesized that, by modifying the cellular activity of tumor-associated microglia/macrophages, we might enhance the efficacy of oHSV-1 in GBM. Indeed, we found that the presence of microglia/macrophages is sufficient to hinder oHSV-1 replication in glioma cells. Our findings suggest that microglia impede the dissemination of oHSV-1 among glioma cells mainly by uptaking oHSV-1 through phagocytosis. Furthermore, although oHSV-1 fails to replicate in microglia/macrophages, some transgenes carried by the virus can still be expressed. We further found that phosphorylation of STAT-1 and 3 are critical for inhibition of viral replication in microglia/macrophages. We then discovered that C16, an oxindole/imidazole derivative (25), selectively inhibits the phosphorylation of STAT-1 and 3 to allow viral replication in tumor-associated macrophages, thereby increasing viral dissemination and tumor destruction in a subcutaneous mouse glioma model.

E18 Sprague-Dawley rats were obtained from Charles River Laboratories. Rat primary microglia isolation and culture was conducted according to the standard protocol (26). In brief, cortices were isolated from day 18 embryonic E18 Sprague-Dawley rat brains. After a 30-minute incubation in trypsin/EDTA (Invitrogen), harvested tissue was washed with culture medium and minced in the presence of DNase I (Invitrogen). The cell suspension was then centrifuged, resuspended in fresh culture medium, and plated on 10-cm culture dishes to high confluence. Microglia cells were grown in DMEM (Sigma) supplemented with 10% FBS (Invitrogen) and 1% antibiotics (penicillin and streptomycin) and maintained at 37°C in 5% CO₂. The culture medium was changed every 3 to 4 days. After 7 to 10 days, microglia were harvested by gently rocking the plate a few times. Finally, microglia floating in the supernatant were plated on a poly-l-lysine–coated plate. Cell purity was routinely tested by immunocytochemistry staining for ITGAM (1:200; ProSci Inc.), which is a microglia-specific integrin protein.

The NF-κB inhibitor, Bay-11, the imidazolo-oxindole derivative C16, and the iNOS inhibitor, aminoguanidine hydrochloride were obtained from Santa Cruz Biotechnology, Millipore, and Tocris Bioscience, respectively. U87 (human GBM) cells and Vero (African green monkey kidney) cells were obtained from ATCC. BV2 (mouse microglia) cells were kindly provided by Dr. Stephanie Booth, Department of Medical Microbiology and Infectious Diseases, University of Manitoba (Manitoba, Canada). All cells were maintained in DMEM supplemented with 10% FBS and 1% antibiotics (penicillin and streptomycin). Total passages of the cells were less than 30 times, and they were routinely tested for mycoplasma. However, cell lines were not authenticated in our hand. All cells were maintained at 37°C in 5% CO₂.

G207 virus was obtained from NeuroVir Therapeutics Inc. U87 (5 × 10⁴) cells alone, or with the indicated number of microglia cells in coculture, were incubated overnight with culture medium (DMEM with 10% FBS and 1% antibiotics). The next day, cells were infected with G207 virus at a multiplicity of infection (MOI) of 1. Viruses were harvested after 2 to 4 days postinfection. After three freeze-thaw cycles, viruses were titrated in triplicate on Vero cells by a standard plaque assay on 12-well plates.

The phagocytosis assay was performed according to the manufacturer's protocol (Phagocytosis Assay Kit, Cayman Chemical). Briefly, microglia cells in wells of a 24-well plate were infected with G207 or vehicle at an MOI of 1, and a 10% latex bead-rabbit IgG-FITC solution was added to each well. At 1 hour postinfection, cells were washed three times with PBS. After 24 hours, cells were incubated with Trypan blue solution for 2 minutes to quench nonspecific staining. After washing twice, cells were lysed using 1× lysis reagent (Promega). Fluorescence intensity was measured in a plate reader (Envision 2103 Multilabel reader, PerkinElmer) using an excitation of 485 nm and an emission of 535 nm.

Virus and IgG latex bead interactions were detected by coincubating cells with either vehicle or 1%, 10%, or 20% IgG latex beads and G207 (MOI:1). After 24 hours of coincubation, cells were incubated overnight with 1 mg/mL X-gal solution (Sigma) for LacZ staining. Cells were counterstained with DAPI to quantify the total number of cells in each microscopic field. Stained cells were then visualized and imaged by using a light microscope (magnification, ×20). LacZ-positive cells and DAPI-stained cells were manually quantified. Data represented the ratio of the number of LacZ-positive cells to the total number of cells in each image (n = 5).

For collecting the cell-conditioned medium, U87 (5 × 10⁴), microglia (5 × 10⁴), or a U87 + microglia coculture (5 × 10⁴ + 5 × 10⁴) were incubated in cell culture medium for 24 hours in a 24-well plate. To examine the effect of cell-conditioned medium on G207 growth, previously collected conditioned medium (50%) was added to U87 (5 × 10⁴) cells (seeded on a 24-well plate) and then infected with G207. Viruses were harvested 3 days postinfection and titrated by a plaque assay on Vero cells.

Total protein was harvested with sample buffer (125 mmol/L Tris-HCL, 50% glycerol, 4% bromophenol blue, and 5% 2-mercaptoethanol) and boiled for 5 minutes. Protein samples were subjected to SDS-PAGE (8% gel), transferred to nitrocellulose membranes, and blocked with 5% nonfat milk (Bio-Rad) in TBS-Tween 20 (TBS-T) for 1 hour at room temperature. The membranes were then incubated with primary antibody (anti–β-actin 1:1,000; Cell Signaling Technology or anti-STAT1 1:1,000; Cell Signaling Technology, anti-STAT3 1:1,000; Cell Signaling Technology, anti–phospho-STAT1 (Tyr701) 1:1,000; Cell Signaling Technology, anti–phospho-STAT3 (Tyr705) 1:1,500; Cell Signaling Technology, anti–phosphor-eIF2α (ser51) 1:1,000; Cell Signaling Technology, anti-ICP27 1:1,000; Abcam, or anti-ICP4 1:750; Abcam) overnight at 4°C. The following day, membranes were washed three times with TBS-T and incubated with the corresponding secondary antibody (1:3,000; PerkinElmer) for one hour at room temperature. Membranes were washed three times with TBS-T before visualization using ECL reagent (PerkinElmer) and a VersaDoc imaging system (Bio-Rad). Band densities were measured using ImageJ software (NIH, Bethesda, MD).

U87 and BV2 microglia cells were infected with oHSV-1 at an MOI of 1. Total RNA was isolated 24 hours postinfection from BV2 or U87 cells using TRIzol reagent (Invitrogen). RT-PCR was performed by a one-step real time PCR using KAPA SYBR FAST One-Step qRT-PCR Universal (D-MARK Biosciences) following the manufacturer's protocol. cDNA was amplified with the following primers. ICP4: 5′-GGCCTGCTTCCGGATCTC-3′ (forward) and 5′-GGTGATGAAGGAGCTGCTGTT-3′ (reverse); ICP27: 5′-GTCTGGCGGACATTAAGGACA-3′ (forward) and 5′-TGGCCAGAATGACAAACACG-3′ (reverse); β-actin: 5′-ACGAGGCCCAGAGCAAGAG-3′ (forward) and 5′-TCTCCATGTCGTCCCAGTTG-3′ (reverse); ICP8: 5′-GCGCCCCATGGTCGTGTT-3′ (forward) and 5′-CTCCGCCGCCGAGGTTC-3′ (reverse); GC: 5′-GCCGCCGCCTACTACCC-3′ (forward) and 5′-GCTGCCGCGACTGTGATG-3′ (reverse); VP5: 5′ - TGAACCCCAGCCCCAGAAACC 3′ (forward) and 5′ - CGAGTAAACCATGTTAAGGACC 3′ (reverse). Results were expressed as 2^−ΔΔC_t.

Cells plated onto 8-well chamber slides were infected with G207 virus, and mock-infected cells were used as a control. Twenty-four hours postinfection, cells were fixed with 0.5% glutaraldehyde solution. Fixed cells were washed twice with PBS and then incubated with 1 mg/mL X-gal solution (Sigma) diluted with X-gal staining solution (5 mmol/L K₃Fe, 5 mmol/L K₄Fe and 2 mmol/L MgCl₂) at 37°C for 1 hour. Stained cells were then visualized and imaged by a light microscope.

Cells were seeded in a 96-well plate at a density of 1 × 10⁴ (U87). After an overnight incubation, cells were treated with vehicle only, viruses with a specified MOI. After 3 days of treatment, cell viability was measured by means of an MTT assay (Sigma) following the manufacturer's instructions. In brief, cells were incubated with the MTT solution for 3 hours at 37°C and were then incubated with lysis buffer for overnight at 37°C. Cell viability was measured at 595 nm using a plate reader (Envision 2103 Multilabel reader, PerkinElmer).

Female athymic nude mice 5 to 6 weeks old were obtained from Harlon Laboratories. Human glioma U87 cells were implanted subcutaneously into the lower flank. When the tumor size reached approximately 75 to 100 mm³, vehicle or C16 (5 mg/kg) was administered intraperitoneally. At 2 to 4 days after the initial C16 administration, the vehicle or oHSV-1 was injected intratumorally. Tumor volumes were then measured using a caliper (height × length × width/2). At the end of the experiment, mice were euthanized by CO₂ asphyxiation. All in vivo procedures were approved by the UBC Animal Care Committee and performed according to the guidelines of the Canadian Council on Animal Care.

DNA was extracted from 4% paraformaldehyde-fixed tumor tissues using an EZNA Tissue DNA Kit (Omega Bio-tek). Extracted DNAs were subjected to qPCR analysis using Sybr Green Master Mix (Invitrogen) supplemented with ICP27 primers: 5′-GTCTGGCGGACATTAAGGACA-3′ (forward) and 5′-TGGCCAGAATGACAAACACG-3′ (reverse); and β-actin primers: 5′-ACGAGGCCCAGAGCAAGAG-3′ (forward) and 5′-TCTCCATGTCGTCCCAGTTG-3′ (reverse). Amplification was carried out using a Quant Studio 6 Flex qPCR system (Applied Biosystems).

Harvested tumor tissues were fixed for 24 hours with 4% paraformaldehyde, followed by 72-hour incubation with 30% sucrose. Tissues were then embedded in OCT (Sakura Tissue Tek), sectioned (20 μm) using a cryostat (Leica CM 3050 S), and placed on Fisherbrand Superfrost Plus microscope slides (Thermo Fisher Scientific). Slides were then washed with PBS, and cells were permeabilized with 0.125% Triton X-100 for 5 minutes, and incubated with 5% goat serum (Santa Cruz Biotechnology) for an hour to block nonspecific binding. Cells were then incubated overnight at 4°C with either anti–HSV-1 antibody (1:50; Abcam) or anti F4/80 antibodies (1:50; Abcam). On the following day, sections were washed three times and then incubated with either goat anti-rabbit IgG Alexa Fluor 488 or goat anti-rat IgG Alexa Fluor 568 secondary antibody (1:500; Invitrogen) for an hour at room temperature. After washing three times, sections were then mounted with DAPI Fluoromount G (Electron Microscopy Sciences) and visualized and imaged by using a confocal microscope (Olympus).

Statistical analysis was performed with SPSS 18 or Microsoft Excel, and significance (P < 0.05) was determined using an independent sample t test, or significance at P < 0.001, P < 0.01, or P < 0.05 was determined using a two-tailed Student t test, respectively. Data were expressed as means ± SD or ± SE.

The efficiency of G207 replication in U87 cells was determined by a one-step viral growth assay (Fig. 1A). The antiproliferative effect of G207 was evaluated by means of an MTT assay. A dose-dependent antiproliferative effect was observed, and the IC₅₀ for G207 at MOI = 1 was determined after 72 hours of infection (Fig. 1B). Our results confirmed that G207 can effectively replicate and lyse U87 cells.

We then measured the growth of G207 in U87 cells in the presence of different numbers of microglia cells. The addition of microglial to U87 culture inhibited G207 replication in a “dose”-dependent manner. G207 replication was reduced by 50% and almost 100% with the addition of 6.25 × 10³ and 1 × 10⁵ microglial cells to U87 (5 × 10⁴) cultures, respectively (Fig. 1C). To further confirm that microglia-mediated oHSV-1 growth suppression is not strain specific, we tested different HSV-1 strains, including HrR3 (ICP6 mutated), b17-TK (a TK mutant), and KOS (wild type) and observed similar inhibition of viral replication among all HSV-1 strains tested (Fig. 1D). Moreover, we observed that oHSV-1 trigger classical activation (M1 polarization) of microglial cells (Supplementary Fig. S1).

To understand the mechanism of microglia-mediated suppression of oHSV-1 oncolysis of glioma cells, we first asked whether the microglia-mediated inhibition of viral production in U87 cells could be caused by products secreted by microglia. To answer this question, we measured G207 production in U87 cells cultured with conditioned medium from microglia cultures. To our surprise, neither microglia nor microglia–glioma coculture conditioned medium showed any inhibition of viral production (Fig. 2A). Moreover, conditioned medium from oHSV-1–infected microglia (virus-free) had no effect on virus replication in glioma cells (Supplementary Fig. S2). In addition, a noncytotoxic dose of nitric oxide also did not have any anti–HSV-1 effect (Supplementary Fig. S3).

We next asked whether oHSV-1 can infect and replicate in microglia cells. Both rodent primary cultured microglial cells and BV2 microglia infected with G207 showed LacZ staining, indicating that the virus can enter the cell and express the reporter gene carried by the virus (27) (Fig. 2B). Concentration-dependent HSV-1 infection in microglial cells was also observed with GFP-expressing UL5354 virus, an ICP27 nonfunctional HSV-1 that is replication deficient but expresses the GFP reporter gene (Supplementary Fig. S4).

As previous studies reported that HSV-1 gains entry into corneal fibroblasts (28) and ocular cells (29) through a phagocytosis-like uptake, we then examined whether G207 can also gain entry into microglia through this form of uptake as microglia/macrophages are well known for their phagocytic capability. Accordingly, we measured phagocytic activity of the microglia in the presence of G207. Interestingly, the results of the phagocytosis assay using IgG latex beads (Phagocytosis Assay Kit, Cayman Chemical) demonstrated that the capacity of microglia to engulf the beads by phagocytosis was significantly reduced in the presence of G207. The addition of G207 at a MOI of 0.1 and 1 hindered the phagocytosis of IgG latex beads by 20% and 60%, respectively (Fig. 2C, I). Similar dose-dependent phagocytosis hindrance was observed in BV2 microglia cells as well (Fig. 2C, II). As one possible explanation is that G207 and IgG latex beads may compete with each other in phagocytosis, we measured virally infected microglia in the presence of an excessive amount of IgG latex beads. The latex beads caused a reduction in the number of G207 “infected” microglial cells in a concentration-dependent fashion (Fig. 2D). Thus, it appears that oHSV-1 entered the microglia via the same phagocytotic mechanism as the IgG beads. It is worth noting that when BV2 cells were pretreated with G207 and an hour later, were then treated with IgG latex, the uptake of IgG latex beads was significantly increased, suggesting that the virus may stimulate the phagocytic activity of these cells (Fig. 2E).

Interestingly, the result of the viral growth assay in primary cultured rat microglial cells (Fig. 2F, I) and BV2 cells (Fig. 2F, II) demonstrated that G207 failed to produce its progeny in the microglia. These results suggest that microglia internalized the viruses through a phagocytosis-like uptake but did not allow viral replication.

To further understand the molecular processes by which HSV-1 replication is prevented in microglia, we measured transcript levels of a panel of viral genes in glioma (U87) and microglial (BV2) cells. The viral genes included ICP4, ICP27, ICP8, VP5, and glycoprotein C (gC), representing immediate early, early, and late genes, respectively. Quantitative RT-PCR results showed that transcription of ICP27, ICP8, VP5, and gC but not ICP4 was significantly suppressed in BV2 cells compared with U87 cells (Fig. 3A).

To understand the events in intracellular signaling that inhibit viral gene expression in G207-stimulated microglia, we tested the effects of a PKR inhibitor (C16), an NF-κB inhibitor (Bay11) and an iNOS inhibitor (AG) on viral immediate early ICP27 and ICP4 gene expression in BV2 cells. C16 but neither Bay11 nor aminoguanidine hydrochloride upregulated expression of ICP4 and ICP27 by 1.8- and 25-fold, respectively (Fig. 3B). To verify that C16-mediated viral gene transcriptional augmentation is not due to the double deletion of ICP34.5 and ICP6 in G207, we then assessed the effect of C16 on wild-type (KOS), and ICP6-mutated (HrR3) HSV-1–infected BV2 cells. C16 treatment also upregulated ICP4 and ICP27 expression in KOS and HrR3-infected BV2 cells (Fig. 3C). In agreement with the result, treatment with 1 and 10 μmol/L C16 significantly enhanced the replication by 9 and 8 times, respectively (Fig. 3D).

Finally, we asked whether C16 can overcome the microglia-mediated suppression of viral replication in glioma cells. As shown in Fig. 3E, oHSV-1 viral replication increased by 33% in the glioma–microglia cocultures treated with C16. However, C16 did not enhance oHSV-1 replication in U87 glioma cultures alone (Supplementary Fig. S5).

As part of an effort to understand the mechanism of C16 in facilitating viral replication in microglia/macrophages, we first looked at the activity level of phosphorylated eIF2α in oHSV-1–infected BV2 cells, as C16 was reported to be a PKR inhibitor (25, 30). To our surprise, there was no change in the phosphorylated eIF2α status in C16-treated microglia at the concentrations used (Fig. 4A). We then examined activities of STAT1 and STAT3 in BV2 microglia with and without G207 infection. G207 upregulated phosphorylation of STAT1 (Tyr701) and STAT3 (Tyr705) in microglia cells (Fig. 4A), which was significantly suppressed by C16. The overall expression level of STAT1 but not STAT3 was upregulated in G207-infected microglial cells, and that level was also reduced after C16 treatment. The C16-induced inhibition of STAT1/3 phosphorylation was also confirmed in LPS-stimulated microglia (Fig. 4B).

To demonstrate that the effect of C16 described above can be translated into the enhanced efficacy of intratumoral replication of oHSV in vivo, C16 was intraperitoneally injected into animals bearing subcutaneously implanted U87 tumors that received G207 intratumorally. Administration of C16 significantly enhanced the oHSV-1 titer in the tumor mass (Fig. 5A). IHC demonstrated an increased number of cells harboring replicating HSV-1 in animals cotreated with oHSV-1 and 5 mg/kg C16 compared with those treated with oHSV-1 alone. Furthermore, numerous cells with colocalization of HSV-1 and macrophage markers (F4/80) were seen in animals cotreated with oHSV-1 and C16 but not oHSV-1 alone, indicating increased viral replication in the macrophages after C16 treatment (Fig. 5B).

Finally, we asked whether C16-mediated facilitation of oHSV-1 replication in a tumor is capable of augmenting antitumor oncolysis. Indeed, tumors in animals treated with C16 (5 mg/kg) and the oHSV-1 combination were 9.3-, 8.2-, and 6-fold smaller than those treated with the vehicle, C16 alone, or oHSV-1 alone, respectively (Fig. 6A). Furthermore, qPCR virus copy numbers in the liver, brain, and gastrointestinal tract were virtually undetectable, demonstrating that viral replication was restricted to the tumors only without spreading to normal organs (Fig. 6B).

In this study, we demonstrated that the presence of microglia/macrophages impedes oHSV-1 oncolytic replication in GBM cells. This finding is consistent with previous reports that inhibition of the innate immune response enhances the efficacy of oncolytic viruses (31–34).

As microglia/macrophages release a wide array of proinflammatory mediators, including nitrogen intermediates, cytokines, and chemokines, in response to pathogen stimulation (35), one may expect that these cells would inhibit OV replication by releasing these factors (36). Surprisingly, we did not find that microglia-conditioned medium had any effect on oHSV-1 replication, albeit the same number of microglia is sufficient to suppress viral replication when cocultured with glioma cells. As we do not know the concentrations and stability of the factors released by microglia in the tumor mass, we cannot rule out the possible antiviral effect of these soluble proinflammatory mediators in vivo as suggested by Meisen and colleagues (36).

Our data suggest that the microglia phagocytosis pathway may play a major role in the clearance of oHSV-1. More importantly, oHSV-1 entering macrophages/microglia via this pathway were unable to replicate. Therefore, macrophages/microglia form a physical barrier that can effectively block dissemination of OVs within the tumor mass. Furthermore, we also found that upon viral infection, the phagocytotic activity of microglia was enhanced.

Our gene expression profiling of microglial cells infected by oHSV-1 demonstrated that expression of the immediate early ICP27 gene, the early ICP8 gene, and the late VP5 and gC genes carried by HSV-1 are significantly decreased in microglia. Downregulation of early and late genes, such as ICP8 (37) and gC (38), might be the result of the silence of the immediate early gene ICP27 that is required for viral DNA synthesis and late gene expression (37, 39). However, although the silence of ICP27 is necessary, it is not sufficient for blocking HSV replication in microglia as restoring ICP27 expression by transfecting an exogenous copy of ICP27 could not rescue viral replication (Supplementary Fig. S6).

Interestingly, some viral genes such as ICP4 and virally carried exogenous genes such as lacZ or GFP reporter can still be expressed after oHSV is taken up by microglia/macrophages. These findings suggest that HSV-1 can potentially target microglia/macrophages to express therapeutic genes that may modify the tumor microenvironment.

It is well known that the IFN-inducible, dsRNA-dependent protein kinase, PKR, plays a key role in the innate immunity response to viral infection (40). After infection with HSV-1, normal host cells activate PKR, thereby shutting down protein synthesis through inactivating eukaryotic initiation factor-2 (eIF-2α) via increased phosphorylation. It is also known that γ34.5 of HSV-1 counteracts PKR and IFN-mediated antiviral activity (41) by recruiting the cellular protein phosphatase 1α, which reverses PKR-mediated phosphorylation of eIF2α (41, 42). Originally, we thought that inhibition of viral protein synthesis might be attributable to eIF2α-mediated blockage of protein synthesis, which led us to the use of the oxindole/imidazole derivative C16, a known PKR inhibitor (25, 30). However, subsequent results indicated that activation of PKR is unlikely to be the mechanism behind the blockage of viral replication in microglia/macrophages for the following reasons: (i) the inhibition of viral gene expression occurred at the transcriptional level; (ii) C16 enhanced viral replication in all viral strains tested regardless of their γ34.5 status (43, 44); (iii) in agreement with previous studies, we also observed inhibition of eIF-2α phosphorylation by C16 in primary astrocytes (25, 30); however, no change was seen in microglia or glioma cells (Supplementary Fig. S7), suggesting that C16 has a possible cell-type specificity in terms of eIF-2α phosphorylation inhibition. Instead, we found that C16 significantly inhibited phosphorylation of STAT1/3 in oHSV-1–infected microglia/macrophages and thereby rescued the replication of oHSV-1. More interestingly, inhibition of STAT1/3 phosphorylation by C16 was only evident in oHSV-1–infected cells and not in mock-infected cells. Previously, C16 was reported to inhibit three dormant host fucosyltransferase genes (FUT3, FUT5, and FUT6) expression in wild-type HSV-1–infected human diploid embryonic lung fibroblasts, which thereby stimulated HSV-1–induced expression of sialyl Lewis X independently to PKR-stimulated translation inhibition (45). Interestingly, they also observed that C16 treatment in HSV-1–infected cell reduced IL6 production, which is one of the main cytokine to activate STAT signaling. STAT blockade effect of C16 observed in our study might be the outcome of C16-mediated inhibition of IL6 (45).

STATs are well known for their antimycobacterial and antiviral effects (46–48). Among the different members of the STAT family, STAT1 is the most known for regulating the antiviral IFN signaling cascade (49), and IFN has in turn been reported to inhibit HSV-1 (50). As expected, elevated STAT1 phosphorylation was observed in oHSV-1–infected microglia/macrophages, which might have been the consequence of the cellular defense against the viral infection.

We found that the phosphorylation of STAT3 was also upregulated in oHSV-1–infected microglia/macrophages. STAT3 is another member of the STAT family that is known for its oncogenic function, tumor cell proliferation, survival, and invasion (51). Inhibition of STAT-3 activity induces cell death mediated by apoptosis (52). The role of STAT3 in HSV-1 infection is not clearly understood. In fact, there have been no studies on the role of STAT3 in the antiviral function of microglia/macrophages. Our data demonstrated enhanced oHSV-1 replication in microglia/macrophages upon C16-mediated inhibition of STAT1 and STAT3. However, C16-mediated virus replication enhancement was only evident in microglia/macrophages and not in glioma cells. In agreement with our findings, functions of both STAT1 and STAT3 were reported to be cell-type specific (53, 54). Shin and colleagues have demonstrated that HSV-1–mediated countermeasures against IFN are ineffective in monocytes (54). Apparently, the STAT1 pathway is more functional in monocytes and macrophages. As ICP27 plays a vital role in countering the cellular IFN response (41, 55), silencing ICP27 by STAT1/3 may render HSV-1 ineffective in evading the intracellular antiviral mechanism within microglia/macrophages.

Our in vivo data using a U87 human glioma xenograft model also confirmed that C16 rescued oHSV-1 gene expression in tumor-infiltrating macrophages and augmented the oHSV-1 viral load in the tumor mass. As the permeability of blood–brain barrier to C16 is not known, we did not use intracranial glioma model in the current study. However, both brain-resident microglia and peripheral macrophages are abundant in intracerebral tumor (19, 21), and their function is almost identical (15). To further demonstrate that the effect of C16 is mainly on macrophage/microglia but not tumor type specific, we used a human prostate cancer LNCaP xenograft model in the same animals. C16 also significantly enhanced oHSV-1 antitumor efficacy in that model (Supplementary Fig. S8). As infiltrated macrophages are evident in almost all tumor types (56), our finding that macrophages prevent oncolytic herpes virus antitumor efficacy by a STAT-dependent mechanism might be applicable to all types of tumors. However, C16 effect on other immune cells, specially activated natural killer (NK) cells remained to be evaluated, as Alvarez-Breckenridge and colleagues 2012 (57) observed that activated NK cells are recruited into infection site and reduce the antitumor efficacy of oHSV-1.

Importantly, we observed that the effect of C16 on enhanced viral replication in macrophages was restricted to tumors; no viral replication was detectable outside of tumor tissue using a sensitive PCR method. It remains to be confirmed whether the selective facilitation of oHSV-1 oncolysis by C16 was due to locally delivered virus or to the difference between normal microglia/macrophages and tumor-associated macrophages/microglia.

In any case, our results suggest that activation of STAT1/3 is sufficient for inhibiting viral replication of oHSV-1 in microglia/macrophages even when the PKR–eIF2α axis remains active. Furthermore, as STAT3 has been a target for cancer treatment and as our data demonstrated that C16 inhibits the proliferation of various types of cancer cells (Supplementary Fig. S9), inhibition of STAT3 activity by C16 may significantly facilitate oHSV-1–induced tumor regression and enhance viral dissemination in a tumor mass.

In summary, our current study suggests that the antitumor efficacy of oHSV-1 is partially determined by the number of TANP cells in the tumor mass, especially the amount of microglia/macrophages that uptake the viruses but do not permit viral replication. This is mainly due to upregulated STAT1/3 phosphorylation. However, we found that some viral and exogenous genes carried by oHSV-1 can still be expressed after uptake of the virus by microglia/macrophages. This suggests that tumor-associated microglia/macrophages may be utilized for expressing virally carried therapeutic genes. More importantly, we discovered that C16 is a strong phosphorylation inhibitor of STAT1/3 and selectively restore viral replication in microglia/macrophages by reversing the character of TANP cells from nonpermissive to permissive, thereby allowing enhanced intratumoral viral production and dissemination (summarized in Fig. 7). Inhibition of STAT3 phosphorylation by C16 may further suppress tumor growth. The combined effects of the above dramatically increase the antitumor efficacy of the oncolytic virus. Our research has provided strong evidence that C16 is an excellent candidate for clinical application in combination with oncolytic virotherapy because of its capacity to significantly augment the antitumor efficacy of oHSV-1.

Z.M. Delwar is a research scientist at Valeo Solutions Inc. and reports receiving a research grant, which is partially sponsored by Virogin Biotech Ltd. through Mitacs. W. Jia is the chief scientific officer at Virogin Biotech Ltd. No potential conflicts of interest were disclosed by the other authors.

Conception and design: Z.M. Delwar, P.S. Rennie, W. Jia

Development of methodology: Z.M. Delwar, Y.H. Wen, P.S. Rennie, W. Jia

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Z.M. Delwar, Y. Kuo

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Z.M. Delwar, P.S. Rennie, W. Jia

Writing, review, and/or revision of the manuscript: Z.M. Delwar, P.S. Rennie, W. Jia

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): P.S. Rennie

Study supervision: P.S. Rennie, W. Jia

We thank Mary Bowden, Hanadi Qashqari, and Erin Kwa for helping with in vivo experiments, Western blot analyses, and cytotoxicity assays, respectively. W. Jia had been awarded grants from Canadian Cancer Society. P.S. Rennie and Z. Delwar had been awarded a grant from Mitacs. Z. Delwar had been awarded a graduate training award (4-year doctoral fellowship) from University of British Columbia.

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.