Oncogenic Ras Inhibits IRF1 to Promote Viral Oncolysis

ORIGINAL ARTICLE Oncogenic Ras inhibits IRF1 to promote viral oncolysis

Y Komatsu1, SL Christian2,NHo1, T Pongnopparat1, M Licursi1 and K Hirasawa1

Oncolytic viruses exploit common molecular changes in cancer cells, which are not present in normal cells, to target and kill cancer cells. Ras transformation and defects in type I interferon (IFN)-mediated antiviral responses are known to be the major mechanisms underlying viral oncolysis. Previously, we demonstrated that oncogenic RAS/Mitogen-activated protein kinase kinase (Ras/MEK) activation suppresses the transcription of many IFN-inducible genes in human cancer cells, suggesting that Ras transformation underlies type I IFN defects in cancer cells. Here, we investigated how Ras/MEK downregulates IFN-induced transcription. By conducting promoter deletion analysis of IFN-inducible genes, namely guanylate-binding protein 2 and IFN gamma inducible protein 47 (Ifi47), we identified the IFN regulatory factor 1 (IRF1) binding site as the promoter region responsible for the regulation of transcription by MEK. MEK inhibition promoted transcription of the IFN-inducible genes in wild type mouse embryonic fibroblasts (MEFs), but not in IRF1− / − MEFs, showing that IRF1 is involved in MEK-mediated downregulation of IFN-inducible genes. Furthermore, IRF1 protein expression was lower in RasV12 cells compared with vector control NIH3T3 cells, but was restored to equivalent levels by inhibition of MEK. Similarly, the restoration of IRF1 expression by MEK inhibition was observed in human cancer cells. IRF1 re-expression in human cancer cells caused cells to become resistant to infection by the oncolytic vesicular stomatitis virus strain. Together, this work demonstrates that Ras/MEK activation in cancer cells downregulates transcription of IFN-inducible genes by targeting IRF1 expression, resulting in increased susceptibility to viral oncolysis.

Oncogene (2015) 34, 3985–3993; doi:10.1038/onc.2014.331; published online 27 October 2014

INTRODUCTION panel of human cancer cells, Stojdl et al.4 demonstrated that Oncolytic viruses preferentially replicate within cancer cells, cancer cells generally have lower sensitivity to IFN than normal leading to their destruction, while normal cells remain unharmed. cells. A common strategy in designing oncolytic viruses is to Although the concept of oncolytic viral therapy has been studied disarm their anti-IFN viral proteins in wild type viruses such that for a century, recent advances in molecular virology and cell they exploit IFN deficiency to selectively replicate in cancer cells. biology have brought oncolytic viral therapy to the forefront of We and others previously reported that activation of the Ras/ Mitogen-activated protein kinase kinase (MEK) pathway sup- cancer therapies. As the outcomes of animal studies and early 15–17 stage clinical trials are very promising, a few viruses are currently, presses the host antiviral response induced by IFN, clearly or will be under evaluation, in phase III clinical trials.1 Oncolytic demonstrating that the two major mechanisms of viral oncolysis fi viruses exploit tumor-specific molecular changes in cancer cells for (Ras-dependency and IFN-de ciency) are indeed connected. By fi 2 3 conducting microarray analysis using human cancer cells, we their replication such as p53 de ciency, oncogenic Ras activation fi and defects in the type I interferon (IFN)-induced antiviral further identi ed the presence of a group of IFN-inducible genes 4,5 (MEK-downregulated IFN-inducible (MDII) genes), which are response. Nevertheless, despite its promise in clinical applica- 18 tion, the precise molecular mechanisms supporting viral oncolysis induced commonly by IFN and by the MEK inhibitor U0126. remain incompletely understood. These results suggest that transcriptional dysregulation of MDII Ras-dependent oncolysis was first reported as the mechanism genes in human cancer cells is one of the underlying mechanisms allowing oncolytic reovirus infection.3 Constitutive Ras activation of Ras-dependent viral oncolysis. Here, we sought to clarify the in transformed cells was shown to be a pre-requisite for the precise mechanism of how Ras/MEK suppresses transcription of the MDII genes. replication of oncolytic reovirus. Since then, the efficacy of reovirus in cancer treatment has been extensively studied both in animal models and in clinical settings, making reovirus a RESULTS promising anti-cancer agent.6 Following the discovery of Ras- dependent oncolysis by reovirus, other viruses were found to Ras/MEK downregulates expression of MDII genes in RasV12 cells similarly exploit the activated Ras signaling pathway for their We previously reported a group of IFN-inducible genes which can – oncolysis.7 12 Multiple cellular mechanisms have been identified be downregulated by Ras/MEK (MDII genes) in human cancer that underlie the Ras dependent viral oncolysis.3,11,13 cells.18 To clarify the molecular mechanisms underlying the Another novel concept of viral oncolysis is the exploitation of downregulation of IFN-inducible genes by Ras/MEK, we analyzed IFN defects in cancer cells by IFN-sensitive viruses.4 Insensitivity of global gene expression in RasV12 transformed NIH3T3 (RasV12) cancer cells to IFN has been known as one of the major obstacles cells, which do not have the confounding mutations that occur in of IFN therapy in cancer patients.14 By systematically testing a human cancer cells. To identify the MDII genes in the mouse

1Division of BioMedical Sciences, Faculty of Medicine, St John’s, Newfoundland, Canada and 2Department of Biochemistry, Faculty of Science, Memorial University of Newfoundland, St John’s, Newfoundland, Canada. Correspondence: Dr K Hirasawa, Division of BioMedical Sciences, Faculty of Medicine, Memorial University of Newfoundland, 300 Prince Philip Dr, St John’s, Newfoundland, Canada A1B3V6. E-mail: [email protected] Received 12 February 2014; revised 5 July 2014; accepted 11 September 2014; published online 27 October 2014 Ras/MEK regulation of IRF1 Y Komatsu et al 3986 fibroblast system, the gene expression profiles of RasV12 cells The expression of Gbp2, Ifi47, Il15, Rig-I, Stat2 and Xaf1 genes treated with U0126, IFN or left untreated were determined using were significantly induced by U0126 only and IFN only treatment, DNA microarrays (Figure 1a). We identified 1264 genes and 1258 and identified as MDII genes. Interestingly, combined treatment of genes with ⩾ 2.5-fold increased expression by either U0126 or IFN IFN and U0126 significantly increased expression of Gbp2, Il15 and treatment, respectively (listed in Supplementary Dataset S1). Stat2 compared with those in cells treated with IFN only or U0126 fi Furthermore, we identi ed 619 genes that were upregulated by only, suggesting synergistic activation of their transcription by IFN both MEK inhibition and IFN treatment as the responsive MDII and MEK inhibition. Iigp2 and Ifit1, which were induced only by IFN fi genes in mouse broblast cells. treatment but not by U0126, represent non-MDII IFN-inducible Based on the microarray data, the expression changes of a genes. In contrast, Ptx3, which was induced by U0126, but not by subset of genes (guanylate-binding protein 2 (Gbp2), IFN gamma IFN, represents a non-MDII MEK-downregulated gene. inducible protein 47 (Ifi47), IFN-induced protein with tetratrico- peptide repeats 1 (Ifit1), immunity-related GTPase family M member 2 (Iigp2), interleukin 15 (Il15), pentraxin related gene Ras/MEK suppresses the transcription of MDII genes (Ptx3), retinoic acid-inducible gene 1 (Rig-I), signal transducer and To further determine whether Ras/MEK regulates the MDII gene activator of transcription 2 (Stat2) and XIAP associated factor 1 transcription, we conducted promoter gene analysis for a subset (Xaf1)) were selected for validation by reverse transcription of MDII genes. RasV12 cells were transfected with the pGL3-Basic polymerase chain reaction (RT–qPCR) analysis (Figure 1b). vector containing the promoter region of either the Gbp2, Ifi47 or

Figure 1. Identification of mouse MEK-downregulated IFN-inducible (MDII) genes. (a) Venn diagrams from DNA microarray analysis showing gene upregulation (⩾2.5-fold expression) in RasV12 cells treated with U0126 (20 μM) or IFN-α (500 U/ml). MDII genes represent genes upregulated by both MEK inhibition and IFN treatment. (b) RasV12 cells were treated with U0126 (20 μM), IFN-α (500 U/ml) or U0126/IFN-α or left untreated for 6 h. RT–qPCR analysis for Gbp2, Ifi47, Il15, Rig-I, Stat2, Xaf1, Iigp2, Ifit1 and Ptx3. The relative expression levels were normalized to Gapdh and reported as compared with the untreated controls. (n = 3, upregulation compared with the untreated control **Po0.01, *Po0.05, upregulation by U0126/IFN-α treatment compared with that by U0126 only or with IFN only treatment ##Po0.01, #Po0.05).

Oncogene (2015) 3985 – 3993 © 2015 Macmillan Publishers Limited Ras/MEK regulation of IRF1 Y Komatsu et al 3987 Rig-I genes, or the IFN-stimulated response element (ISRE) and − 80 and − 35 was also found to be responsive. In contrast, the then treated with U0126 or IFN for 24 h (Figure 2a). We found that elements responsible for regulation of Gbp2 by IFN are contained IFN treatment significantly activated transcription of the ISRE in regions from − 512 to − 83 and from − 83 to − 68 of the Gbp2 reporter construct and all the reporter constructs containing the promoter, distinct from the U0126 responsive sites. These results MDII gene promoter region. MEK inhibition significantly increased suggest that independent promoter elements are required for the promoter activity of Gbp2, Ifi47 and ISRE in RasV12 cells. MDII gene transcription by MEK inhibition compared with IFN Although Rig-I promoter activity increased two-fold in RasV12 cells stimulation. treated with U0126, the induction was not statistically significant. We examined the Gbp2 and Ifi47 promoter regions essential for These results demonstrate that Ras/MEK suppresses the promoter regulation by U0126 (Gbp2: − 68 to − 47 and Ifi47: − 148 and − 80) activities of these MDII genes. for transcription factor binding sites. Using the JASPAR database, To identify the promoter region responsive to Ras/MEK, we we identified five putative transcription factor-binding sites (for conducted promoter deletion analysis of the Gbp2 and Ifi47 genes, paired box gene 2 (PAX2), IFN regulatory factor (IRF)-binding which showed the most robust up-regulation by U0126 treatment element (IRFE), AT rich interactive domain 3A (ARID3A), SRY-box (Figure 2b). Deletion of 21-bp region between − 68 and − 47 of the containing gene 17 (SOX17) and SRY-box containing gene 10 Gbp2 promoter significantly reduced the U0126-induced transcrip- (SOX10)) that were located in both of the U0126-responsive regions. Among the five candidate binding sites, we focused on tional response, indicating that this region contains elements 19 essential for transcriptional regulation of Gbp2. Similarly, we the IRFE because of its relevance to the IFN system. observed a significant reduction in U0126-induced activation of the Ifi47 promoter when the region from − 762 to − 148 or from IRF1 regulates the transcription of MDII genes − 148 to − 80 was deleted. Importantly, the deletion between Chromatin immunoprecipitation (ChIP) assays were performed to − 148 and − 80 of the Ifi47 promoter completely abolished determine whether IRF1 binds to the Gbp2 and Ifi47 gene activation of its promoter activity by U0126. Regulation of Ifi47 promoters in response to MEK inhibition. We observed a relatively transcription by IFN was primarily regulated by the same region small amount of IRF1 present at the putative IRFE sites of Gbp2 (−762 to − 80) as regulated by U0126, although a region between and Ifi47 promoter in untreated RasV12 cells (Figure 3a).

Figure 2. Identification of MDII promoter regions responsible for transcriptional activation by U0126 or IFN. (a) Promoter activity of control pGL3-Basic plasmid, pISRE-Luc plasmid or pGL3-Basic plasmid containing promoters of Gbp2, Ifi47 or Rig-I in RasV12 cells treated with U0126 (20 μM) or IFN-α (500 U/ml) for 24 h. Relative luciferase activities (RLU) were reported as compared with the untreated controls. (b) Promoter activity of control pGL3-Basic plasmid, or pGL3-Basic plasmid containing various deletion constructs of the Gbp2 and the Ifi47 promoters in RasV12 cells treated as above (n = 3, *Po0.05, **Po0.01).

Figure 3. Effects of Ras/MEK inhibition on IRF1 binding to the Gbp2 and Ifi47 promoter. ChIP assay was performed on chromatin isolated from RasV12 cells treated with U0126 (20 μM) or IFN-α (500 U/ml), or left untreated. The ChIP DNA and input DNA were analyzed by (a) end-point PCR and (b) qPCR using primers designed to amplify the U0126 responsive and distal promoter regions as shown. Vertical bars in the promoter diagram indicate the location of predicted IRFEs (Gbp2: − 61 to − 44. Ifi47: − 156 to − 145, − 128 to − 116, − 74 to − 56 and − 46 to − 35). Arrows and the numbers below the arrows indicate binding sites of the primer sets (thick arrows: amplifies IRFE regions, thin arrows: amplifies distal control regions). IRF1 binding to the promoter of Gbp2 or Ifi47 are presented as percentage of input (% Input) (n = 3, *Po0.05; **Po0.01).

Treatment with either U0126 or IFN substantially enhanced IRF1 IRF1 expression is suppressed by the Ras/MEK pathway recruitment at the Gbp2 and Ifi47 promoter. This interaction was To determine how the Ras/MEK pathway regulates IRF1 function, specific since the IRFE sites of Gbp2 and Ifi47 promoter regions we examined IRF1 expression in vector control NIH3T3 cells and were not detected in the pull-down with normal IgG, and binding RasV12 cells treated with or without U0126 (Figure 5a). The of IRF1 to the distal control sites in the two promoters was not expression level of IRF1 was lower in RasV12 cells than in vector observed. These results demonstrate specific binding of IRF1 to control NIH3T3 cells and was restored by 6 h of MEK inhibition. the IRFE sites of Gbp2 and Ifi47 upon Ras/MEK inhibition in RasV12 Double bands of IRF1 were observed in vector control NIH3T3 cells. To validate and quantify these results, quantitative ChIP cells, whereas only one band was apparent in RasV12 cells. We analysis was performed, which revealed that U0126 treatment believe that the upper band is one of the IRF1 isoforms; however, fi signi cantly increased IRF1 binding at the IRFE sites of both Gbp2 it remains to be determined why it is present only in vector fi and I 47 (Figure 3b). control cells. We also examined the mRNA and protein expression To determine if IRF1 regulates MDII gene expression in response levels of IRF1 in RasV12 cells after treatment with U0126 to inhibiting the Ras/MEK pathway, we established primary and (Figure 5b). RT–qPCR analysis revealed that the IRF1 mRNA level fi immortalized mouse embryonic broblasts (MEFs), which repre- was significantly increased by MEK inhibition after 4 h of U0126 sent different stages of transformation, from wild type, or IRF1- treatment, indicating that IRF1 mRNA levels are downregulated by fi de cient, C57BL/6 mice. Expression of the MDII genes Gbp2 and the active Ras/MEK pathway. Similarly, we observed increased IRF1 fi Xaf1 was signi cantly induced by U0126 in primary wild type MEFs protein levels as early as 2 h after U0126 treatment. These results (Figure 4a). In support of the role of IRF1 in regulating MDII genes, demonstrated that the Ras/MEK pathway downregulates IRF1 U0126 did not increase transcription of Gbp2 or Xaf1 in primary fi fi expression at both the mRNA and protein levels. IRF1-de cient MEFs. In contrast, there was no upregulation of I 47, To determine if regulation of IRF1 by Ras/MEK also occurs in Il15 and Rig-I in wild type primary MEFs treated with U0126. human cancer cells, we analyzed the effect of MEK inhibition on Therefore, we examined the transcription of MDII genes in IRF1 expression in the human breast carcinoma cells MDA-MB-468 response to U0126 and IFN in immortalized MEFs, which are at and the human colon cancer cells DLD-1 (Figures 5c and d). We a more advanced stage of transformation. Treatment with U0126 fi found that IRF1 protein is upregulated in both MDA-MB-468 and increased expression of all the MDII genes tested (Gbp2, I 47, Il15, DLD-1 cells after 24 h of treatment, even at sub-optimal Rig-I and Xaf1) in immortalized wild-type MEFs (Figure 4b). This concentrations of U0126, as determined by ERK phosphorylation suggests that the degree of MDII gene induction by U0126 levels. Furthermore, we found that IRF1 protein increased in correlates with the level of cellular transformation and the level of response to three different MEK inhibitors (U0126, PD98059 and IRF1 downregulation, which are in turn correlated with the degree SL327), demonstrating that the regulation of IRF1 was not because of constitutive Ras activation. In contrast, induction of these genes of non-specific effects of U0126 (Figure 5d). Knockdown of ERK 1/2 was not observed in immortalized IRF1-deficient MEFs, confirming in RasV12 cells also increased the IRF1 expression levels, further that IRF1 expression is necessary for this regulation. The absence supporting the interaction between Ras/MEK and IRF1 (Figure 5e). of IRF1 also significantly reduced the induction of Ifi47, Il15 and Rig-I by IFN in both primary and immortalized MEFs and that of Gbp2 in primary MEFs and that of Xaf1 in immortalized MEFs, Ras/MEK-mediated downregulation of IRF1 impairs the IFN supporting the critical role of IRF1 in IFN-mediated transcription.19 antiviral response Taken together, these results demonstrated that IRF1 is the To determine whether IRF1 downregulation is the key factor in the primary transcriptional regulator of these MDII genes. impairment of antiviral response in cells with activated Ras/MEK,

Figure 4. IRF1 involvement in the modulation of MDII gene transcription by Ras/MEK. RT-qPCR analysis for the MDII genes (Gbp2, Xaf1, Ifi47, Il15 and Rig-I). Primary (a) and immortalized (b) MEFs derived from wild type or IRF1-deficient mice were treated with U0126 (20 μM) or IFN-α (500 U/ml), or left untreated for 6 and 12 h, respectively. Relative expression levels of the MDII genes were normalized to Gapdh and the data from wild type MEFs were compared with that of IRF1-deficient MEFs of the same treatment (n = 3, *Po0.05; **Po0.01).

Figure 5. Restoration of IRF1 expression by Ras/MEK inhibition. (a) Vector control NIH3T3 or RasV12 cells were treated with the indicated concentration of U0126 for 6 h, and IRF1 and ERK phosphorylation levels were detected by western blot. (b)RT–qPCR analysis for IRF1 mRNA (top panel) and western blot analysis for IRF1 protein (bottom panel) in RasV12 cells treated with U0126 (20 μM) for the indicated period of time. IRF1 expression levels were normalized against the 0.5 h controls. (c) MDA-MB-468 cells were treated with the indicated concentration of U0126 for 6 h. (d) DLD-1 cells were treated with indicated concentration of the different MEK inhibitors U0126, PD98059 or SL327. IRF1 and ERK phosphorylation levels were analyzed by western blot. (e) RasV12 cells transfected with random nucleotide sequences (NG) or two independent ERK 1/2 oligonucleotides (#1 and #2) for indicated time points were analyzed by western blot (n = 3, *Po0.05).

Figure 6. IRF1 regulates viral oncolysis. (a) RasV12 cells transfected with control or mouse IRF1 pCMV-SPORT6 (0.5 μg) were treated with indicated concentration of IFN-α for 16 h and then challenged with VSV (MOI = 1). (b) HT1080 and HT29 cells transfected with control (1 μg) or human IRF1 pCMV-SPORT6 (1, 0.5, 0.25, 0.12 and 0.06 μg) were infected with VSVM51R (MOI = 5). (c) DLD-1 cells transfected with control or human IRF1 pCMV-SPORT6 (1 μg) were infected with VSVM51R (MOI = 0, 5, 1.25, 0.3, 0.08 or 0.02) for 24 h. Protein levels of VSV-G, IRF1 or GAPDH were determined by western blot analysis.

we restored IRF1 expression in RasV12 cells and in human cancer DISCUSSION cells. First, to determine if IRF1 overexpression is sufficient to In our previous study, we reported that the transcription of a restore IFN-induced antiviral activity in RasV12 cells, the cells were group of IFN-inducible genes was suppressed by Ras/MEK in transfected with either mouse IRF1 or control vector (pCMV- human cancer cells, which we defined as MEK downregulated IFN- SPORT6), followed by treatment with different concentrations of inducible (MDII) genes.18 In this study, we aimed to clarify the IFN for 16 h and then challenged with vesicular stomatitis virus molecular mechanism underlying Ras/MEK-mediated downregula- (VSV). IRF1 introduction restored IFN’s ability to protect RasV12 tion of the MDII gene transcription. This study demonstrates that cells from VSV infection, particularly with 120 and 60 U/ml of IFN (1) IRF1 is the transcriptional regulator of MDII genes, (2) activated (Figure 6a). Ras/MEK reduces IRF1 expression in mouse fibroblast cells and Next, we determined whether IFN-sensitive oncolytic viruses human cancer cells and (3) the expression level of IRF1 modulates exploit Ras-mediated IRF1 downregulation in human cancer cells cellular susceptibility to viral oncolysis. by overexpressing IRF1 followed by infection with the oncolytic IRF1, which was initially identified as a transcriptional activator version of VSV (VSVM51R; AV1).5 We found that the human cancer of the IFN-β gene, has been found to be broadly involved in cell lines, HT1080 and HT29, were susceptible to the oncolytic regulating gene transcription important for innate immune virus when transfected with control vector (pCMV-SPORT6), but responses.19–21 IRF1 is a potent antiviral effector, and deficient that viral oncolysis was severely restricted in a concentration levels of this protein result in dramatically increased cellular dependent manner with overexpression of IRF1 (Figure 6b). Next, susceptibility to different types of viruses.22–24 In addition, IRF1 has human cancer DLD-1 cells were infected with different MOIs of been characterized as a tumor suppressor, which simultaneously VSVM51R after transfection with control or IRF1 vector (Figure 6c). upregulates the transcription of tumor suppressor genes, while Similar to the results obtained above, the oncolytic virus replicated downregulating the transcription of oncogenes to induce efficiently in cells transfected with control vector, whereas IRF1 apoptosis and growth inhibition.25 Therefore, it is likely that overexpression substantially restricted viral oncolysis. Together tumor cells acquire oncogenic cellular characteristics to dysregu- these results demonstrated the critical role of IRF1 in defining the late anti-tumor functions of IRF1 during their development. In fact, susceptibility of cancer cells to certain oncolytic viruses. the IRF1 gene is often lost, mutated or downregulated in several

Oncogene (2015) 3985 – 3993 © 2015 Macmillan Publishers Limited Ras/MEK regulation of IRF1 Y Komatsu et al 3991 types of cancer26 and is negatively correlated with the tumor There are other possible ways that translation of IRF1 may be grade, risk of recurrence and death.27,28 Activating mutations of directly regulated by the Ras/MEK pathway. MicroRNAs can the Ras protein have been found in ~ 30% of all human cancers.29 regulate protein synthesis by repressing translation or, in some Moreover, in cancers where direct activating mutations of Ras cases, by promoting translation.34,35 In particular, miR-23a and protein are absent, the upstream or downstream signaling miR-383 bind to the 3′-UTR of IRF1 mRNA to regulate both its components of the Ras pathway are often found to be over-active mRNA and protein level.36,37 Ras/MEK may regulate the activity of instead,30 suggesting that Ras/MEK activation may underlie IRF1 an IRF1-binding microRNAs to suppress translation rate of IRF1 impairment in a broad range of human cancer cells. Such mRNA. Alternatively, Ras/MEK could regulate translation of IRF1 observations suggest that the IRF1 downregulation may be a mRNA through the regulation of eIF4E, which is phosphorylated common aberrant characteristic in human cancer cells that can be by the MAP-kinase signal-integrating kinases 1 and 2, both of exploited by Ras-dependent or IFN-sensitive oncolytic viruses. which are downstream elements of Ras/MEK.38 The effect of As we show here, IRF1 is the key transcriptional regulator of the phosphorylation on eIF4E activity is controversial and may either MDII genes that are responsible for the IFN-mediated antiviral reduce.39 or increase40 the binding affinity for the cap structure of response, IFN production or Jak/STAT activation (Figure 7). There- mRNA. eIF4E phosphorylation also plays an important role in the fore, the low-expression IRF1/MDII genes in cancer cells results in a translocation of certain mRNAs.41,42 IRF1 mRNA may be sensitive delay in establishing antiviral responses, and allows replication of to eIF4E-regulated translation or translocation via the Mnk1/2- oncolytic viruses. In contrast, normal cells expressing higher levels eIF4E. The precise mechanism of IRF1 regulation by Ras/MEK, of IRF1, and therefore higher levels of IRF1-regulated MDII genes, however, remains to be elucidated. can induce the antiviral response rapidly and efficiently upon Both Type I and Type II IFNs are known as a potent inducer of oncolytic virus infection. In contrast to the initial stage of viral IRF1.19 IRF1 promotion by Ras/MEK may be a secondary effect of oncolysis that is promoted by IRF1/MDII genes-downregulation by IFN induction and activation of the IFN pathway. However, this is MEK, other tumor-specific molecular changes may play critical unlikely the case because MEK inhibition did not activate the Jak- roles in maintaining cancer cell susceptibility to oncolytic viruses STAT pathways16 and because transcriptional activators of IFNs, at the later phase of infection when IFNs and cytokines are such as IRF3 and IRF7, were not induced by U0126 treatment in actively produced. As recently shown, the steady state levels of the microarray analysis. Although further study is required, we antiviral genes correlate with sensitivity of cancer cells to viral 31,32 believe that Ras/MEK directly regulates IRF1 expression at its oncolysis. transcriptional or translational level. IRF1 stimulates its own transcription as a positive regulatory 33 Within cancer cells, Ras activation and defects in the innate factor. As IRF1 expression was restored earlier at the protein immune response are the two major cellular characteristics level than at the mRNA level in RasV12 cells treated with U0126 targeted by oncolytic viruses. Ras activation promotes replication (Figure 5b), we hypothesized that Ras/MEK inhibition restores IRF1 of oncolytic viruses in multiple ways, such as inhibition of antiviral protein stability, which leads to the induction of IRF1 mRNA function of PKR, or enhancement of virus entry and uncoating, expression. However, this was not the case since U0126 treatment viral particle release and translation of viral mRNA.3,11,14 Deficien- did not change IRF1 protein stability in cycloheximide-treated cies in IFN signaling components43 and epigenetic modifications RasV12 cells and human cancer cells (Supplementary Figure S1). of chromatin44 have also been reported to contribute to the generation of a defective innate immune response. Here, we demonstrate that IRF1 downregulation is responsible for transcriptional suppression of IFN-inducible genes in cells with activated Ras/MEK. Although it remains to be determined whether the Ras/MEK-mediated downregulation of IRF1 expression underlies other oncolytic mechanisms previously reported, we believe that this may represent a common mechanism for a broad range of oncolytic viruses. Although IRF1 is a well-known tumor suppressor and antiviral protein, its involvement in viral oncolysis has never been investigated. Our findings will contribute toward the design of novel oncolytic viruses and their therapeutic uses for cancer treatment with improved efficacy and safety. Moreover, while IFN therapy has shown significant therapeutic effects on cancer patients with certain types of cancer,45 its efficacy is often halted by the presence of cancer cells resistant to IFN therapy.43,46–48 Since IRF1 downregulation by Ras/MEK may underlie development of resistance to IFN stimulation, combined treatment with Ras/MEK inhibitors would improve the efficacy of IFN therapy in cancer patients.

MATERIALS AND METHODS

Figure 7. Schematic diagram illustrating the suppression of MDII Cells and reagents gene expression by Ras/MEK via inhibition of IRF1. The expression of Murine fibroblast cells (NIH3T3 and L929) and human cancer cells (HT29, the MDII genes is regulated by the level of IRF1 expression, which is HT1080, DLD-1 and MDA-MB-468) were obtained from the American Type in turn regulated by Ras/MEK activity. The MDII genes are involved in Culture Collection (Manassas, VA, USA) and immortalized wild type MEFs promoting host defense against oncolytic viruses at the different from Dr Nahum Sonenberg (McGill University, Montreal, Canada). All the stages, such as antiviral response (that is, Gbp2, Xaf1), IFN production human cancer cell lines used in this study were authenticated by DDC (that is, Rig-I, Il15) and IFN sensitivity (that is, Stat2). The decreased Medical (Fairfield, OH, USA) by STR DNA analysis. IRF1-deficient and wild- expression of the MDII genes by Ras/MEK support replication of type MEFs were established from day 14 embryos of C57BL/6-Irf1tm1Mak49 oncolytic virus in cancer cells, but not in normal cells. The relative and C57BL/6J mice from the Jackson Laboratory (Bar Harbor, ME, USA). expression levels of the proteins are symbolized by the size of the Immortalized IRF1-deficient MEFs were established by serial passages of protein. primary MEFs over the course of seven passages over 21 days.

© 2015 Macmillan Publishers Limited Oncogene (2015) 3985 – 3993 Ras/MEK regulation of IRF1 Y Komatsu et al 3992 Recombinant mouse IFN-α was purchased from PBL Interferon Source incubated with 4 μg of anti-IRF1 antibody or normal Rabbit IgG (Santa (Piscataway, NJ, USA), U0126 from Cell Signaling Technology (Danvers, Cruz Biotechnology) overnight. The antibody-chromatin complexes were MA, USA) and PD98059 and SL327 from Calbiochem (La Jolla, CA, USA). incubated with 50 μl of preblocked protein A agarose beads and washed Antibodies to VSV-G were purchased from Alpha Diagnostic (San Antonio, twice each with low-salt wash buffer (0.1% SDS, 1.0% Triton X-100, 2 mM TX, USA), phospho-ERK 1/2 from Calbiochem, GAPDH (6C5) from Abcam EDTA, 20 mM Tris–HCl pH 8.1, 150 mM NaCl), high-salt wash buffer (0.1% (Toronto, ON, Canada), mouse IRF1 (M-20) and total ERK (sc-94) from Santa SDS, 1.0% Triton X-100, 2 mM EDTA, 20 mM Tris–HCl pH 8.1, 500 mM NaCl), Cruz Biotechnology (Santa Cruz, CA, USA) and human IRF1 from BD LiCl wash buffer (250 mM LiCl, 1.0% IGEPAL-CA630, 1% Deoxycholic Acid, Transduction Laboratories (Mississauga, ON, Canada). Mouse IRF1 and 1.0 mM EDTA, 10.0 mM Tris pH 8.1), and TE buffer (10.0 mM Tris pH 8.1, human IRF1 cDNAs were purchased from Thermo Fisher Scientific 1.0 mM EDTA). The chromatin complex was then eluted from beads using (Waltham, MA, USA), pGL3-Basic vectors from Promega (Madison, WI, elution buffer (1.0% SDS, 100.0 mM NaHCO ), and treated with 5 M NaCl USA) and pISRE-Luc from Stratagene (La Jolla, CA, USA). ERK 1/2 siRNA and 3 control siRNA were obtained from Santa Cruz Biotechnology and used as overnight. Samples were then treated with RNase A (QIAGEN) and previously described.16 Proteinase K (QIAGEN). ChIP end-point PCR and ChIP qPCR were performed using primers listed in Supplementary Table S2 and S3, respectively. DNA microarray analysis and quantitative RT–PCR Western blot analysis RNA was isolated from RasV12 cells treated with 20 μM U0126 or 500 U/ml IFN-α, or left untreated, for 6 h. Total RNA treated with DNase using TURBO Protein samples were prepared and western blot was conducted and DNA-freeTM kit (Ambion, ON, Canada) were sent to the Centre for Applied evaluated as previously described15 using antibodies listed above. Genomics (TCAG, Toronto, ON, Canada) for analysis using Affymetrix 430 2.0 mouse DNA microarrays (Affymetrix, Santa Clara, CA, USA). RNA integrity number was determined to be greater than 8.9 for all samples Statistical analysis (Agilent 2100 Bioanalyzer, Agilent, Santa Clara, CA, USA). Data from three One-way analysis of variance with Tukey’s post-hoc test were performed biological replicates were analyzed using GeneSpring (v7.3, Agilent). Genes using GraphPad Prism 4.0c software (GraphPad Software, La Jolla, CA, USA). with greater than 2.5-fold induction compared with the untreated control were compared using Venn diagrams. A fold-change cut-off strategy was used to reduce the Type II (false negative) error rate in order to identify all CONFLICT OF INTEREST possible candidate genes. Data are deposited in the Gene expression The authors declare no conflict of interest. omnibus50 (GEO accession number GSE49469). Quantitative RT-PCR (RT- qPCR) was performed in triplicate using the previously described validation 18 strategies. The primer sequences are shown in Supplementary Table S1. ACKNOWLEDGEMENTS This work was supported by grants (to KH) from the Canadian Institutes for Health Promoter reporter assays Research (CIHR), the Regional Development Corporation of Newfoundland and fi Promoter constructs of Gbp2, I 47 and Rig-I were obtained by PCR Labrador (RDC) and Canadian Breast Cancer Foundation (CBCF). SLC was supported fi ampli cation of mouse genomic DNA and ligation into the XhoI and HindIII by a post-doctoral award from CIHR and RDC and a separate trainee award from the sites of pGL3-Basic vector. The promoter deletion constructs of Gbp2 and Beatrice Hunter Cancer Research Institute with funds provided by the Cancer Ifi47 were made using the Erase-a-Base System (Promega) according to the Research Training Program as part of The Terry Fox Foundation Strategic Health manufacturer’s instructions followed by sequencing to determine the remaining promoter region. RasV12 cells were transiently transfected with Research Training Program in Cancer Research at CIHR. Special thanks to Dr John Bell 1 μg of the reporter plasmids using SuperFect Transfection Reagent (Centre for Innovative Cancer Therapeutics, Ottawa Hospital Research Institute, (QIAGEN, Mississauga, ON, Canada). Twenty-four hours after transfection, Ottawa, Canada) for VSV and VSVM51R, and Nebiur E A for graphic design. We thank cells were treated with U0126 or IFN-α for additional 24 h. Luciferase Susan Wakeham and Rahul-dev Kalsi for technical support. activity was measured by the Luciferase Assay System (Promega) and luminescence measured using Fluoroskan Ascent FL (Thermo Labsystems, Waltham, MA, USA). Transcription factor binding elements were identified REFERENCES 51 using the JASPAR database. 1 Bell J. Oncolytic viruses: an approved product on the horizon? Mol Ther 2010; 18: 233–234. Virus infection 2 Bischoff JR, Kirn DH, Williams A, Heise C, Horn S, Muna M et al. An adenovirus mutant that replicates selectively in p53-deficient human tumor cells. Science Wild-type VSV (Indiana strain) and attenuated VSV (VSVM51R), provided by 274 – Dr John C Bell (Centre for Innovative Cancer Therapeutics, Ottawa Hospital 1996; :373 376. Research Institute, Ottawa, Canada) were amplified and titers determined 3 Strong JE, Coffey MC, Tang D, Sabinin P, Lee PW. The molecular basis of viral 17 as described previously.4,15 To generate IRF1-overexpressing cells, RasV12 oncolysis: usurpation of the Ras signaling pathway by reovirus. EMBO J 1998; : – cells were transfected with mouse IRF1 pCMV-SPORT6 or control pCMV- 3351 3362. SPORT6 using SuperFect Transfection Reagent (QIAGEN). At 24 h after the 4 Stojdl DF, Lichty B, Knowles S, Marius R, Atkins H, Sonenberg N et al. Exploiting fi transfection, the cells were treated with IFN-α for 16 h and then infected tumor-speci c defects in the interferon pathway with a previously unknown 6 – with VSV. IRF1-overexpressing human cancer cells were generated using oncolytic virus. Nat Med 2000; : 821 825. human IRF1 pCMV-SPORT6 or control pCMV-SPORT6. 5 Stojdl DF, Lichty BD, tenOever BR, Paterson JM, Power AT, Knowles S et al. VSV strains with defects in their ability to shutdown innate immunity are potent systemic anti-cancer agents. Cancer Cell 2003; 4:263–275. Chromatin immunoprecipitation assay 6 Norman KL, Lee PW. Not all viruses are bad guys: the case for reovirus in cancer RasV12 cells treated with U0126 or IFN-α, or left untreated for 4 h were therapy. Drug Discov Today 2005; 10: 847–855. incubated with 37% formaldehyde for 10 min to cross-link chromatin and 7 Farassati F, Yang AD, Lee PW. Oncogenes in Ras signalling pathway dictate host- then treated with 0.125 M glycine to terminate the cross-linking. Cells were cell permissiveness to herpes simplex virus 1. Nat Cell Biol 2001; 3: 745–750. then washed twice with PBS, and the cell pellet was resuspended in ChIP 8 Balachandran S, Porosnicu M, Barber GN. Oncolytic activity of vesicular stomatitis lysis buffer (Santa Cruz Biotechnology) (1.0% SDS, 10.0 mM EDTA, 50 mM virus is effective against tumors exhibiting aberrant p53, Ras, or myc function and Tris-Hcl, pH 8.1). Chromatin was sonicated into 200–500-bp length involves the induction of apoptosis. J Virol 2001; 75: 3474–3479. fragments and incubated with 50 μl of Pierce Protein A Agarose beads 9 Bergmann M, Romirer I, Sachet M, Fleischhacker R, Garcia-Sastre A, Palese P et al. (Thermo Scientific, Rockford, IL, USA) pre-blocked with Herring Sperm DNA A genetically engineered influenza A virus with ras-dependent oncolytic prop- (Life Technologies, Burlington, ON, Canada) and BSA (Promega). One erties. Cancer Res 2001; 61:8188–8193. milligram of the cell lysate was diluted to a final volume of 1 ml with ChIP 10 Cascallo M, Capella G, Mazo A, Alemany R. Ras-dependent oncolysis with an dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.20 mM EDTA, 16.7 mM Tris- adenovirus VAI mutant. Cancer Res 2003; 63: 5544–5550. HCl pH 8.1, 167 mM NaCl) supplemented with protease and phosphatase 11 Goetz C, Everson RG, Zhang LC, Gromeier M. MAPK signal-integrating kinase inhibitors (Aprotinin, Sigma-Aldrich, Oakville, ON, Canada), (PMSF, Sigma- controls cap-independent translation and cell type-specific cytotoxicity of an Aldrich), (Halt Phosphatase Inhibitor Cocktail, Thermo Scientific) and was oncolytic poliovirus. Mol Ther 2010; 18: 1937–1946.

Oncogene (2015) 3985 – 3993 © 2015 Macmillan Publishers Limited Ras/MEK regulation of IRF1 Y Komatsu et al 3993 12 Puhlmann J, Puehler F, Mumberg D, Boukamp P, Beier R. Rac1 is required for 32 Kurokawa C, Iankov I, Aderca I, Petell C, Middha S, Giannini C et al. Pre-existing oncolytic NDV replication in human cancer cells and establishes a link between antiviral state can impact oncolytic infection in glioblastoma patients treated with tumorigenesis and sensitivity to oncolytic virus. Oncogene 2010; 29: 2205–2216. oncolytic measles virus. Mol Ther 2014; 22(Suppl 1): S248. 13 Marcato P, Shmulevitz M, Pan D, Stoltz D, Lee PW. Ras transformation mediates 33 Reis LF, Harada H, Wolchok JD, Taniguchi T, Vilcek J. Critical role of a common reovirus oncolysis by enhancing virus uncoating, particle infectivity, and transcription factor, IRF-1, in the regulation of IFN-beta and IFN-inducible genes. apoptosis-dependent release. Mol Ther 2007; 15:1522–1530. EMBO J 1992; 11:185–193. 14 Marcato P, Shmulevitz M, Pan D, Stoltz D, Lee PW. Ras transformation mediates 34 Fabian MR, Sonenberg N, Filipowicz W. Regulation of mRNA translation and reovirus oncolysis by enhancing virus uncoating, particle infectivity, and stability by microRNAs. Annu Rev Biochem 2010; 79: 351–379. apoptosis-dependent release. Mol Ther 2007; 15:1522–1530. 35 Vasudevan S, Tong Y, Steitz JA. Switching from repression to activation: micro- 15 Battcock SM, Collier TW, Zu D, Hirasawa K. Negative regulation of the alpha RNAs can up-regulate translation. Science 2007; 318: 1931–1934. interferon-induced antiviral response by the Ras/Raf/MEK pathway. J Virol 2006; 36 Liu X, Ru J, Zhang J, Zhu LH, Liu M, Li X et al. miR-23a targets interferon regulatory 80: 4422–4430. factor 1 and modulates cellular proliferation and paclitaxel-induced apoptosis in 16 Christian SL, Collier TW, Zu D, Licursi M, Hough CM, Hirasawa K. Activated Ras/ gastric adenocarcinoma cells. PLoS One 2013; 8: e64707. MEK inhibits the antiviral response of alpha interferon by reducing STAT2 levels. 37 Lian J, Tian H, Liu L, Zhang XS, Li WQ, Deng YM et al. Downregulation of J Virol 2009; 83: 6717–6726. microRNA-383 is associated with male infertility and promotes testicular embry- 1 17 Noser JA, Mael AA, Sakuma R, Ohmine S, Marcato P, Lee PW et al. The RAS/Raf1/ onal carcinoma cell proliferation by targeting IRF1. Cell Death Dis 2010; : e94. MEK/ERK signaling pathway facilitates VSV-mediated oncolysis: implication 38 Fukunaga R, Hunter T. MNK1, a new MAP kinase-activated protein kinase, isolated for the defective interferon response in cancer cells. Mol Ther 2007; 15: by a novel expression screening method for identifying protein kinase substrates. 16 – 1531–1536. EMBO J 1997; : 1921 1933. 18 Christian SL, Zu D, Licursi M, Komatsu Y, Pongnopparat T, Codner DA et al. 39 Scheper GC, van Kollenburg B, Hu J, Luo Y, Goss DJ, Proud CG. Phosphorylation of fi Suppression of IFN-induced transcription underlies IFN defects generated by eukaryotic initiation factor 4E markedly reduces its af nity for capped mRNA. 277 – activated Ras/MEK in human cancer cells. PLoS One 2012; 7: e44267. J Biol Chem 2002; : 3303 3309. 19 Taniguchi T, Ogasawara K, Takaoka A, Tanaka N. IRF family of transcription factors 40 Minich WB, Balasta ML, Goss DJ, Rhoads RE. Chromatographic resolution of in vivo as regulators of host defense. Annu Rev Immunol 2001; 19:623–655. phosphorylated and nonphosphorylated eukaryotic translation initiation factor fi 20 Kimura T, Nakayama K, Penninger J, Kitagawa M, Harada H, Matsuyama T et al. eIF-4E: increased cap af nity of the phosphorylated form. Proc Natl Acad Sci USA 91 – Involvement of the IRF-1 transcription factor in antiviral responses to interferons. 1994; :7668 7672. 41 Topisirovic I, Ruiz-Gutierrez M, Borden KL. Phosphorylation of the eukaryotic Science 1994; 264: 1921–1924. translation initiation factor eIF4E contributes to its transformation and mRNA 21 Miyamoto M, Fujita T, Kimura Y, Maruyama M, Harada H, Sudo Y et al. Regulated transport activities. Cancer Res 2004; 64: 8639–8642. expression of a gene encoding a nuclear factor, IRF-1, that specifically binds to 42 Phillips A, Blaydes JP. MNK1 and EIF4E are downstream effectors of MEKs in the IFN-beta gene regulatory elements. Cell 1988; 54:903–913. regulation of the nuclear export of HDM2 mRNA. Oncogene 2008; 27: 1645–1649. 22 Pine R. Constitutive expression of an ISGF2/IRF1 transgene leads to interferon- 43 Sun WH, Pabon C, Alsayed Y, Huang PP, Jandeska S, Uddin S et al. Interferon-alpha independent activation of interferon-inducible genes and resistance to virus resistance in a cutaneous T-cell lymphoma cell line is associated with lack of infection. J Virol 1992; 66: 4470–4478. STAT1 expression. Blood 1998; 91:570–576. 23 Kanazawa N, Kurosaki M, Sakamoto N, Enomoto N, Itsui Y, Yamashiro T et al. 44 Nguyen TL, Abdelbary H, Arguello M, Breitbach C, Leveille S, Diallo JS et al. Regulation of hepatitis C virus replication by interferon regulatory factor 1. J Virol Chemical targeting of the innate antiviral response by histone deacetylase inhi- 2004; 78: 9713–9720. bitors renders refractory cancers sensitive to viral oncolysis. Proc Natl Acad Sci USA 24 Brien JD, Daffis S, Lazear HM, Cho H, Suthar MS, Gale M Jr et al. Interferon 2008; 105: 14981–14986. regulatory factor-1 (IRF-1) shapes both innate and CD8(+) T cell immune 45 Goldstein D, Laszlo J. Interferon therapy in cancer: from imaginon to interferon. responses against West Nile virus infection. PLoS Pathog 2011; 7: e1002230. Cancer Res 1986; 46: 4315–4329. 25 Tamura T, Ishihara M, Lamphier MS, Tanaka N, Oishi I, Aizawa S et al. An 46 Landolfo S, Guarini A, Riera L, Gariglio M, Gribaudo G, Cignetti A et al. Chronic IRF-1-dependent pathway of DNA damage-induced apoptosis in mito- myeloid leukemia cells resistant to interferon-alpha lack STAT1 expression. 376 – gen-activated T lymphocytes. Nature 1995; : 596 599. Hematol J 2000; 1:7–14. 26 Willman CL, Sever CE, Pallavicini MG, Harada H, Tanaka N, Slovak ML et al. 47 Sakai I, Takeuchi K, Yamauchi H, Narumi H, Fujita S. Constitutive expression of Deletion of IRF-1, mapping to chromosome 5q31.1, in human leukemia and SOCS3 confers resistance to IFN-alpha in chronic myelogenous leukemia cells. 259 – preleukemic myelodysplasia. Science 1993; :968 971. Blood 2002; 100: 2926–2931. 27 Moriyama Y, Nishiguchi S, Tamori A, Koh N, Yano Y, Kubo S et al. Tumor- 48 Shang D, Liu Y, Ito N, Kamoto T, Ogawa O. Defective Jak-Stat activation in renal suppressor effect of interferon regulatory factor-1 in human hepatocellular car- cell carcinoma is associated with interferon-alpha resistance. Cancer Sci 2007; 98: 7 – cinoma. Clin Cancer Res 2001; : 1293 1298. 1259–1264. 28 Doherty GM, Boucher L, Sorenson K, Lowney J. Interferon regulatory factor 49 Tanaka N, Ishihara M, Kitagawa M, Harada H, Kimura T, Matsuyama T et al. Cellular 233 – expression in human breast cancer. Ann Surg 2001; :623 629. commitment to oncogene-induced transformation or apoptosis is dependent on 29 Adjei AA. Blocking oncogenic Ras signaling for cancer therapy. J Natl Cancer Inst the transcription factor IRF-1. Cell 1994; 77:829–839. 93 – 2001; : 1062 1074. 50 Barrett T, Wilhite SE, Ledoux P, Evangelista C, Kim IF, Tomashevsky M et al. NCBI 30 Downward J. Targeting RAS signalling pathways in cancer therapy. Nat Rev Cancer GEO: archive for functional genomics data sets—update. Nucleic Acids Res 2013; 2003; 3:11–22. 41: D991–D995. 31 Iankov I, Kurokawa C, Middha S, Aderca I, Allen C, Keeney G et al. Gene expression 51 Portales-Casamar E, Thongjuea S, Kwon AT, Arenillas D, Zhao X, Valen E et al. profile of innate immune anti-viral factors in ovarian cancer patients treated with JASPAR 2010: the greatly expanded open-access database of transcription factor oncolytic measles virus. Mol Ther 2014; 22:S12. binding profiles. Nucleic Acids Res 2010 38: D105–D110.

Supplementary Information accompanies this paper on the Oncogene website (http://www.nature.com/onc)