Δγ₁34.5 Herpes Simplex Viruses Encoding Human Cytomegalovirus IRS1 or TRS1 Induce Interferon Regulatory Factor 3 Phosphorylation and an Interferon-Stimulated Gene Response

TEXT

Recombinant herpes simplex viruses (HSV) deleted for the diploid γ₁34.5 neurovirulence gene (Δγ₁34.5) are in development for use in antitumor therapeutics and have been safely administered to patients with brain tumors in independent phase I clinical trials (12, 20, 21, 33). HSV triggers an innate interferon (IFN)-mediated antiviral response, inducing effector proteins that limit viral gene expression, replication, and spread in the host (24, 34, 42). This innate response is an important determinant in the safety and efficacy of the oncolytic HSV (oHSV) vectors (4, 9, 34). IFN-regulated gene expression also alters the chemokine expression and surface receptor expression responsible for adaptive immune cell recruitment and immune targeting of infected cells (1, 4, 5, 41). In this way, the innate and adaptive immune response elicited during oHSV infection can suppress viral replication while contributing to the immune-mediated antitumor effect and is an important determinant of oHSV antitumor efficacy (2, 4, 5, 25, 26, 30, 42).

Host pattern recognition proteins bind unique viral products (double-stranded RNA [dsRNA] and cytosolic DNA) produced during infection and activate IFN signaling pathways in the cell (14). Ultimately, these signaling changes lead to transcriptional regulation of host genes (chemokines and IFN-stimulated genes) that mediate an antipathogen response and occur in three phases (14, 29). The initial phase occurs after viral nucleic acids bind host pattern recognition receptors and activate signaling cascades, ultimately leading to phosphorylation of IFN regulatory factor 3 (IRF3) by tank binding kinase 1 (TBK1) (14, 40). Phosphorylated IRF3 (p-IRF3) dimerizes and translocates to the nucleus, where it upregulates beta IFN 1 (IFN-β1), chemokine, and IFN-stimulated gene (ISG) expression (14, 29). The inductive phase follows next and is characterized by IFN-β1 activation of the Janus kinase (JAK)-signal transducers and activators of transcription (STAT) pathway, resulting in IRF3- and IRF7-mediated ISG and IFN-α gene production. The third or amplification phase consists of increased ISG production, including production of effector antiviral proteins (e.g., protein kinase R [PKR], RNase L, and myxovirus resistance 1 [Mx1]) that limit viral replication and spread (14, 23, 28, 35). PKR can trigger further IRF3 activation, thus amplifying ISG and chemokine signaling and IFN induction (43). IRF3 is an important target for viral host evasion, because it is integral to all 3 phases (sensitization, induction, and amplification) of the type I IFN response.

HSV type 1 (HSV-1) encodes γ₁34.5, a multifunctional gene that suppresses both the early sensitization and the late amplification phase of the type I IFN response (13, 40). Early in infection, the γ₁34.5 gene product, infected-cell protein 34.5 (ICP34.5), binds TBK1, limiting IRF3-mediated antiviral gene expression (40). Late in infection, ICP34.5 enables HSV evasion of dsRNA-activated protein kinase R (PKR), a host antiviral kinase that orchestrates translational arrest (7, 8). We previously showed that replacement of the γ₁34.5 gene with either of the human cytomegalovirus (HCMV) PKR-evasion genes, TRS1 or IRS1, produced attenuated oncolytic HSV (oHSV) capable of late viral protein synthesis, improving replication in the tumor and antitumor activity (36). The prior studies showed that, while the HCMV genes replaced one γ₁34.5 function (late viral protein synthesis), they did not restore γ₁34.5 neurovirulence.

Other viral (vaccinia virus, parainfluenza virus, influenza virus) PKR-evasion genes target other dsRNA-activated pathways, including IRF3 signaling. The HCMV TRS1 and IRS1 genes contain a dsRNA binding domain that has been hypothesized to bind and sequester dsRNA, precluding PKR activation in a manner similar to that seen with vaccinia virus E3L (11). To determine whether the HCMV PKR evasion genes TRS1 and IRS1 suppress IRF3 signaling, immunostaining studies were performed using rabbit anti-phospho-IRF3 (Ser396) monoclonal antibody (MAb) (4D4G; Cell Signaling Technology, Beverly, MA), using viruses summarized in Table 1. As seen in previous studies (40), a higher level of phosphorylated IRF3 (p-IRF3) is detectable at 4 h postinfection (hpi) in the Δγ₁34.5-infected cells (C101) than in wild-type HSV-infected cells (Fig. 1A, lanes 7 and 8). Increased p-IRF3 staining was detected in the chimeric HSV-infected cells at 2 and 4 hpi compared with the level exhibited by the other samples, and the highest level of p-IRF3 staining occurred in the C134-infected cells. Immunostaining studies using a rabbit polyclonal Ig against total IRF3 (Santa Cruz Biosystems, Santa Cruz, CA) demonstrated equivalent IRF3 levels in the lysates (Fig. 1A, lower panel).

In unstimulated cells, latent IRF3 remains unphosphorylated and cytoplasmic. After induction of the IFN response by dsRNA, IRF3 is phosphorylated, dimerizes, and translocates to the nucleus, where it becomes transcriptionally active. Nuclear and cytoplasmic fractions were prepared using infected Neuro2A cells and hypotonic and hypertonic solutions as described previously (3), and IRF3 immunostaining was repeated. The results showed that p-IRF3 was detectable in the nuclear fraction of the C101-, C130-, and C134-infected cells (Fig. 1B, lanes 8 to 10), with the highest level of pIRF3 staining seen in the C134-infected nuclear sample (Fig. 1B, lane 10). While most of the pIRF3 staining occurred in the nuclear fractions, trace p-IRF3 was also detected in the C134 cytoplasmic sample (Fig. 1B, lanes 5 and 10).

The p-IRF3 staining results suggested that the IRF3 was transcriptionally active. To test whether this was the case, Neuro2A cells were mock infected or infected with a high multiplicity of infection (MOI) of virus (10 PFU/cell) and RNA was collected 4 hpi using an RNeasy total RNA isolation kit (Qiagen, Valencia, CA) per the manufacturer's instructions, and IRF3-regulated chemokine (CXCL10) and ISG (IFIT1) expression was examined by TaqMan quantitative PCR (qPCR) using commercial primers (ABI Mm99999072_m1, Mm00439552_s1, Mm00515153_m1, and 4352930E) specific for mRNA transcripts (2). CXCL10 and IFIT1 gene expression levels were increased in all virus-infected samples relative to uninfected samples (Fig. 2A and B; mock-infected cells [relative quantity {RQ} = 1]). The chimeric HSVs increased both CXCL10 and IFIT1 gene expression compared to Δγ₁34.5-infected cell results, but C134 provided the highest level of gene expression. In contrast to the differences in p-IRF3 levels detected in the wild-type and Δγ₁34.5-infected Neuro2A cells, no differences were detected in CXCL10 and IFIT1 expression. IRF3 phosphorylation is sufficient for CXCL10 and IFIT1 transcriptional upregulation; however, additional transcription factors (NF-κB and AKT) are required for IFN-β1 gene upregulation (1). To determine whether the chimeric HSV induced IFN-β1 gene expression changes, TaqMan qPCR studies were performed. The results showed that all of the Δγ₁34.5-based recombinants (C101, C130, and C134) significantly increased IFN-β1 transcription compared to wild-type or mock-infected sample results.

Next, to identify differences between recombinant and wild-type infected cells in IFN sensitivity, single-step replication studies were performed with IFN-treated cells. Neuro2A cells (1 × 10⁵/well) were treated with IFN (400 IU of murine IFN-β1/ml) 14 h prior to infection and then infected in parallel at a high MOI (10 PFU/cell) with wild-type, C101, C130, or C134 virus in the presence or absence of IFN-β (ProSci, Poway, CA). Neuro2A cells provide a unique environment for examining the effect of IFN on viral replication. Unlike primary cells and glioma tumor cells, in which translational arrest limits Δγ₁34.5 replication, Δγ₁34.5 recombinants synthesize abundant late viral proteins in Neuro2A cells (reference 38 and data not shown). In Neuro2A cells, the chimeric HSV (C130 and C134) and the Δγ₁34.5 virus replicated like wild-type HSV in single-step replication assays (Fig. 3). However, the recombinants exhibited differing levels of IFN-β1 sensitivity in IFN-treated cells. Wild-type HSV replication was relatively insensitive to IFN-β1, producing >10⁸ PFU, a significantly higher level than that seen with the Δγ₁34.5-based recombinants. IFN-β1 treatment suppressed C130 viral replication by 100-fold to 10⁶ PFU and had the greatest effect on C134 and C101 replication, suppressing replication by 1,000-fold.

To determine whether the IRF3 phosphorylation and gene expression changes were unique to Neuro2A cells, IRF3 phosphorylation and ISG expression studies were performed using primary human foreskin fibroblast (HFF) cells. As seen with the Neuro2A results, more p-IRF3 was detected in the C101-, C130-, and C134-infected samples than in the mock-infected and wild-type HSV-infected HFFs, and the highest levels of p-IRF3 staining and CXCL10 gene expression were seen in the C134-infected samples (Fig. 4).

We previously showed that replacement of the γ₁34.5 gene with either of the HCMV PKR-evasion genes (TRS1 or IRS1) restored late viral protein synthesis and improved viral replication in tumor cells (36). The chimeric HSV, however, did not suppress IRF3 signaling or early IFN signaling or induce IRF3 phosphorylation, nuclear translocation, and transcriptional activity in Neuro 2A or primary HFFs. Of the chimeric HSV, C134 (Δγ₁34.5 IRS1) produced the greatest IRF3 phosphorylation and ISG and chemokine gene expression and was most sensitive to IFN-β1 effects. As shown with published results of studies using HeLa cells, we detected suppression of IRF3 phosphorylation in Neuro2A and HFF cells. Despite the apparent differences in IRF3 phosphorylation and nuclear translocation data, significant differences in chemokine or ISG expression in wild-type HSV- and Δγ₁34.5 virus-infected cells were not seen. This suggests that there are redundant pathways that elicit these transcriptional changes despite IRF3 suppression by the γ₁34.5 gene or that additional host factors are required that were absent in the cell lines we examined.

Suppression of PKR activity does not restore wild-type HSV virulence, and C134 is aneurovirulent (27, 36). It is possible that induction of IRF3-mediated antiviral gene expression compensates for PKR evasion in chimeric HSV-infected cells. IRF3-mediated type I IFN signaling induction is an important determinant for CNS replication of neurotropic viruses such as rabies virus and HSV-1 (6, 22, 24). Deletion of the IFN receptor or IRF3 signaling molecule in transgenic animal models increases HSV-1 replication and neurovirulence (9, 24, 32).

The IRF3 response could also affect oHSV antitumor activity (10, 15, 25, 42). C134 elicits a robust ISG and chemokine response that could potentially impede viral replication and spread in the tumor. However, increased ISG and chemokine expression could enhance both direct and indirect antitumor activity. IRF3 induces proapoptotic genes (encoding the zinc finger regulator of apoptosis and cell cycle arrest [Zac1], death domain-associated protein [Daxx], and early gene response 2 [EGR2]) that could enhance oncolysis and alter tumor growth (1). IFN and dsRNA induce antiangiogenic effects as well (16, 37). Investigations examining corneal vascularization have shown that dsRNA activation of TLR3 inhibits lymphatic and blood vessel growth in the cornea (17, 19, 39). Increased chemokine and ISG responses could enhance immune cell recruitment and improve antitumor activity. Investigators have shown differences in cytokine responses to wild-type and Δγ₁34.5 recombinant virus infection in preclinical models (18, 31). Increased chemokine expression may contribute to the efficacy of C134 as an antitumor oHSV. In addition to chemokine-mediated immune cell recruitment, IRF3 also transcriptionally upregulates genes that improve antigen recognition and immune cell function such as CD166 (an adhesion and T cell proliferative-response molecule) and the nonclassical major histocompatibility class I genes Q-I, T-10, and T-22, which provide ligands for natural killer (NK) and γ/δ T cells.

C134 shares a beneficial characteristic of a Δγ₁34.5 virus in that it is aneurovirulent. It selectively inhibits one component of the innate antiviral response (PKR-mediated translational arrest) but induces the type I IFN response, increasing IRF3-regulated chemokine, ISG, and IFN-β1 gene expression. Future studies would need to examine this response in vivo in the context of the cellular heterogeneity of nonmalignant and malignant tissue with an intact cellular immune response.