Cancer Upregulated Gene 2, a Novel Oncogene, Confers Resistance to Oncolytic Vesicular Stomatitis Virus Through STAT1-OASL2 Signaling

ORIGINAL ARTICLE Cancer upregulated gene 2, a novel oncogene, confers resistance to oncolytic vesicular stomatitis virus through STAT1-OASL2 signaling

W Malilas1, SS Koh2, R Srisuttee1, W Boonying1, I-R Cho1, C-S Jeong3, RN Johnston4 and Y-H Chung1

We have recently found a novel oncogene, named cancer upregulated gene 2 (CUG2), which activates Ras and mitogen-activated protein kinases (MAPKs), including ERK, JNK and p38 MAPK. Because activation of these signaling pathways has previously been shown to enhance cancer cell susceptibility to oncolysis by certain viruses, we examined whether vesicular stomatitis virus (VSV) could function as a potential therapeutic agent by efﬁciently inducing cytolysis in cells transformed by CUG2. Unexpectedly, NIH3T3 cells stably expressing CUG2 (NIH-CUG2) were resistant to VSV because of the activation of signal transducers and activators of transcription 1 (STAT1). The result was supported by evidence showing that suppression of STAT1 with short interference RNA (siRNA) renders cells susceptible to VSV. Furthermore, 20–50 oligoadenylate synthetase-like (OASL) 2 was the most affected by STAT1 expression level among anti-viral proteins and furthermore suppression of OASL2 mRNA level caused NIH-CUG2 cells to succumb to VSV as seen in NIH-CUG2 cells treated with STAT1 siRNA. In addition, Colon26L5 carcinoma cells stably expressing CUG2 (Colon26L5-CUG2) exhibited resistance to VSV, whereas Colon26L5 stably expressing a control vector yielded to VSV infection. Moreover, Colon26L5-CUG2 cells stably suppressing STAT1 succumbed to VSV infection, resulting in apoptosis. Taken together, we propose that VSV treatment combined with the selective regulation of genes such as STAT1 and OASL2 will improve therapeutic outcomes for CUG2-overexpressing tumors.

Cancer Gene Therapy (2013) 20, 125–132; doi:10.1038/cgt.2012.96; published online 11 January 2013 Keywords: VSV; STAT1; OASL2; CUG2; virotherapy

INTRODUCTION peripheral matrix protein.4 VSV causes vesicular lesions in the To discover new genes that have a crucial role in common mucous membranes of mouse and in the nasal tissues of horses, tumorigenesis regardless of tissue origin, we analyzed commonly cattle and pigs. However, VSV does not induce significant upregulated unknown genes using 242 normal and 300 tumor symptoms in humans. Moreover, evidence of human exposure to 5 samples originating from 11 different tissues with an Affymetrix VSV is generally uncommon in seroprevalence studies, which is gene chip system. An uncharacterized gene, later named cancer beneficial for the potential use of VSV as an oncolytic agent in upregulated gene 2 (CUG2), was identified as a potential human trials. Furthermore, VSV can replicate in malignant cells 6–9 candidate that was commonly upregulated in various tumor more efficiently than in normal cells as do other oncolytic viruses. tissues, including ovary, liver, lung, pancreas and colon. CUG2 was Successful treatments with VSV have been reported in laboratory 10–13 mapped to chromosome 6q22.32, extends about 8.5 kb in length models of breast, lung, brain and colon cancers. However, VSV with a 3-exon structure, encoding a 88-amino-acid polypeptide.1 is very sensitive to the anti-viral activity of interferon (IFN) induction, Further study has revealed that CUG2 protein is a novel and therefore improved strategies for the application of VSV have component of centromeres required for proper kinetochore been sought for preferential replication in tumor cells and 10,14 function during cell division.2 Of interest, CUG2 has been shown resistance to the anti-viral action of host cells. to possess an oncogenic effect in a transplanted model using IFN-induced anti-viral proteins have been extensively studied. NIH3T3 cells expressing CUG2, behaving similarly to Ras.1 Our 20–50 Oligoadenylate synthetase (OAS) stimulated by double- previous study has shown that CUG2 activates Ras and mitogen- stranded RNA polymerizes ATP to 20- and 50-linked oligomers of activated protein kinases (MAPKs) including p38 MAPK, which adenosine and activates RNaseL by inhibiting initiation of protein eventually facilitate reoviral replication and the death of virus- synthesis or degrading RNA.15 The OAS gene family contains four infected cancer cells.3 OAS members: OAS1, OAS2, OAS3 and OASL.16 IFN-stimulated gene Vesicular stomatitis virus (VSV), also known as a potential 15 (ISG15), a 15-kDa ubiquitin-like protein, has recently emerged therapeutic anticancer agent, is a non-enveloped, negative- as an important tool in the defense against viral infections. Like stranded RNA virus and belongs to the Rhabdoviridae family. ubiquitin and other ubiquitin-like proteins, ISG15 is attached to The RNA genome of VSV encodes five proteins, including target proteins through a C-terminal Gly–Gly motif, which is nucleocapsid, polymerase proteins, surface glycoprotein and referred to as ISGylation and is implicated in anti-viral activity.17

1WCU, Department of Cogno-Mechatronics Engineering, Pusan National University, Busan, Republic of Korea; 2Immunotherapy Research Center, Korea Research Institute of Bioscience and Biotechnology, Department of Functional Genomics, University of Science and Technology, Daejeon, Republic of Korea; 3School of Biological Sciences, Ulsan University, Ulsan, Republic of Korea and 4Department of Biochemistry and Molecular Biology, University of Calgary, Calgary, Alberta, Canada. Correspondence: Professor Y-H Chung, WCU, Department of Cogno-Mechatronics Engineering, Pusan National University, Busan 609-735, Republic of Korea. E-mail: [email protected] Received 2 October 2012; revised 13 December 2012; accepted 13 December 2012; published online 11 January 2013 CUG2 confers resistance to VSV through STAT1 W Malilas et al 126 Guanylate-binding proteins (GBPs) are a family of large IFN- Active ras capture assays 18 induced GTPases. Although it is currently unclear how GBP To capture GTP-bound Ras from cell lysates, a GST-Raf-RBD fusion protein functions in anti-viral activity, ectopic expression of GBP1 or GBP2 was used as described elsewhere.24 The cells were lysed with lysis buffer protects cells against VSV and encephalomyelitis virus infections.19 and glutathione-agarose beads (Clontech, Mountain View, CA, USA) were IFN-inducible protein-10 (IP-10), a non-ELR CXC chemokine, is a added to the supernatant, and the solutions were rocked on ice for 30 min chemoattractant for natural killer and T cells and is believed to be to exclude nonspecific binding. After centrifugation, the supernatant was an important regulator of the T-helper type 1, interleukin-12- harvested, followed by the addition of GST-Raf-RBD (Cytoskeleton, Denver, CO, USA). The solution was rocked on ice for 1 h, followed by the addition driven inflammatory responses as an inducer of cellular 20 of glutathione-agarose beads to capture active Ras, which binds to GST- infiltration. IFN-induced protein with tetratricopeptide repeat 3 Raf-RBD. After washing three times, the pellet was mixed with 40 mlof2Â (IFIT3) is significantly induced upon RNA virus infection. Recent SDS loading buffer. To detect the captured Ras, immunoblotting was study has shown that IFN-induced tetratricopeptide repeat protein performed with anti-Ras antibody (Santa Cruz Biotechnology). IFIT3 potentiates anti-viral signaling by bridging mitochondrial anti-viral signaling complex and tumor necrosis factor receptor- Qualitative reverse transcription-polymerase chain reaction associated factor family member-associated nuclear factor-kB activator-binding kinase 1.21 Total RNA was extracted from cells using the RNeasy mini kit (Qiagen, Valencia, CA, USA) in accordance with the manufacturer’s instructions. In Several lines of evidence have shown that Ras/Raf/MEK all, 2 mg of total RNA were converted to cDNA using Superscript II reverse signaling reduces the threshold for VSV infection by attenuation transcriptase (Invitrogen, Carlsbad, CA, USA), and polymerase chain 22,23 of IFN signaling. Our observation that CUG2 upregulates Ras reaction (PCR) was performed using the specific primers designed by the and MAPK activity3 led us to predict that NIH3T3 cells transformed Vector NTI software (Invitrogen). The cDNA products of each reaction were with CUG2 should become susceptible to VSV infection. In this diluted, and PCR was run at the optimized cycle number. Glyceraldehyde study, we tested this hypothesis by asking whether VSV efficiently 3-phosphate dehydrogenase mRNA was used as an internal standard. induces cytolysis in NIH3T3 cells transformed with CUG2. We unexpectedly found that CUG2 activates STAT1, leading to the Real-time RT-PCR upregulation of the anti-viral gene OASL2, which actually confers In addition, for a quantitative real-time reverse transcription (RT)-PCR, resistance to VSV infection. We also confirmed that activation of several sets of primers were designed by Bioneer Corporation (Daejeon, STAT1 mediated by CUG2 provides resistance to VSV in colon South Korea). Gene amplification was carried out using cDNA samples with carcinoma cells. We herein suggest that VSV treatment combined a SYBR Green-based real-time PCR. All real-time PCR amplification and with selective downregulation of genes such as STAT1 and OASL2 analysis was conducted by Accupower qPCR array using Exicycler 96 Real- will enhance therapeutic efficiency for CUG2-overexpressing Time Quantitative Thermal Block (Bioneer, Daejeon, South Korea). The tumors. entire system and all experiments involved therein were performed according to the MIQE (Minimum Information for Publication of Quantitative Real-Time PCR Experiments) guidelines. Amplification para- meters were as follows: one cycle of 95 1C for 10 min and 65 1C for 5 min MATERIALS AND METHODS followed by 40 cycles of 95 1C for 5 s, 58 1C for 25 s and 72 1C 30 s. The last Cell cultures and virus amplification step was carried out at 65–95 1C(11CsÀ 1) for melting curve analysis. Data NIH3T3 cells stably transfected with empty vector (NIH-Vec), oncogenic analysis is based on relative quantitation method and DDCT value was Ras(V12) expression plasmid (NIH-Ras) or CUG2 expression plasmid (NIH- used to determine the relative fold changes.25 All the data were CUG2) were used. In addition, Colon26-L5 cells stably transfected with the normalized with the housekeeping glyceraldehyde 3-phosphate empty vector (Colon26L5-Vec) or the CUG2 expression plasmid dehydrogenase expression level. (Colon26L5-CUG2), and Colon26L5-CUG2 cells stably transfected with a short hairpin RNA control vector (Colon26L5-CUG2-ShCon) or short hairpin Transfection with short interference RNA RNA STAT1 vector (OriGene, Rockville, MD, USA) (Colon26L5-CUG2- ShSTAT1) were used. Cells were cultured in Dulbecco’s modified Eagle’s Cells were trypsinized and incubated overnight to achieve 60–70% medium supplemented with 10% fetal bovine serum and 1% penicillin, confluence before short interference RNA (siRNA) transfection. STAT1 0 and streptomycin at 5% CO and 37 1C. The stable cells were selected siRNA (Bioneer Corporation; 200 nM; sense, 5 -ACUUUGCUGUAACCCU- 2 0 0 0 under puromycin (1 mgmlÀ 1) or zeocin (200 mgmlÀ 1) based on trans- GUA(dTdT)-3 ; antisense, 5 -UACAGGGUUACAGCAAAGU(dTdT)-3 ), OASL2 fected vectors. VSV (Indiana strain) purchased from the ATCC (Manassas, siRNA (200 nM; commercially pre-made at Qiagen) or negative control VA, USA) was propagated in L929 cells and the viral titer was measured in siRNA (Bioneer Corporation) were mixed with Lipofectamine 2000 plaque-forming units. (Invitrogen). The cells were incubated with the transfection mixture for 6 h and then rinsed with Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum. The cells were incubated for 48 h before harvest. Reagents and antibodies For immunoblotting, antibodies against phospho-STAT1, STAT1, phospho- STAT2, STAT2, caspase 8 and poly-ADP-ribose polymerase (PARP) were Microarrays acquired from Cell Signaling Biotechnology (Danvers, MA, USA). Anti-Ras Total RNA was amplified and purified using the Ambion Illumina RNA and b-actin antibodies were obtained from Santa Cruz Biotechnology amplification kit (Ambion, Austin, TX, USA) to yield biotinylated cRNA (Santa Cruz, CA, USA). Anti-VSV glycoprotein (G) antibody was obtained according to the manufacturer’s instructions. Labeled cRNA samples from Abcam (Cambridge, MA, USA). Glutathione S-transferase (GST)-Raf- (750 ng each) were hybridized to each mouse-8 expression bead array for RBD fusion protein was obtained from Cytoskeleton (Denver, CO, USA). 16–18 h at 58 1C, according to the manufacturer’s instructions (Illumina, IFN-b was also purchased from R&D systems (Minneapolis, MN, USA). San Diego, CA, USA). Detection of array signal was carried out using Amersham fluorolink streptavidin-Cy3 (GE Healthcare Bio-Sciences, Little Chalfont, UK) following the bead array manual. Arrays were scanned with Western blotting assays an Illumina BeadArray Reader confocal scanner according to the Cells were harvested and lysed with lysis buffer (150 mM NaCl, 1% NP-40, manufacturer’s instructions. The quality of hybridization and overall chip 50 mM Tris-HCl (pH 7.5)) containing 0.1 mM Na2VO3,1mM NaF and protease performance were monitored by visual inspection of both internal quality inhibitors (Sigma, St Louis, MO, USA). For immunoblotting, proteins from control and the raw scanned data. Raw data were extracted using the whole-cell lysates were resolved by 10 or 12% sodium dodecyl sulfate- software provided by Illumina GenomeStudio v.2009.2 (Gene Expression polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred to Module v.1.5.4). Array data were filtered by detection P-value o0.05 nitrocellulose membranes. Primary antibodies were used at 1 : 1000 or (similar to signal to noise) in at least 50% of samples (we applied a filtering 1 : 2000 dilutions, and secondary antibodies conjugated with horseradish criterion for data analysis; higher signal value was required to obtain a peroxidase were used at 1 : 2000 dilutions in 5% non-fat dry milk. After the detection P-value o0.05). Selected gene signal values were transformed as final washing, nitrocellulose membranes were exposed for an enhanced logarithms and normalized by quantiles. Comparative analysis between chemiluminescence assay using the LAS 4000 mini (Fuji, Tokyo, Japan). test and control samples was carried out.

Cancer Gene Therapy (2013), 125 – 132 & 2013 Nature America, Inc. CUG2 confers resistance to VSV through STAT1 W Malilas et al 127 Statistical analysis we examined Ras activity and levels in NIH-Vec, NIH-Ras and NIH- Data are presented as a means±s.d. The Student’s t-test was used for CUG2 cells. Ras protein was more abundant in NIH-CUG2 and NIH- statistical analysis, with P-value o0.05 defined as significance. Ras cells than in NIH-Vec cells (Figure 2a). In addition, the level of active GTP-bound Ras proteins, as detected by GST-Raf-RBD fusion protein binding, was much higher in NIH-Ras cells and relatively RESULTS higher in NIH-CUG2 cells than in NIH-Vec cells (Figure 2a), CUG2 expression confers resistance to VSV infection in NIH3T3 indicating that activation of Ras is not the only factor that cells determines susceptibility to VSV infection. When we examined the After elevated levels of CUG2 with oncogenic activity were phosphorylation of STAT1 and STAT2 (well known as signaling detected in various tumor tissues, such as ovary, liver, lung and molecules of type I IFN), as possible explanations for the enhanced colon,1 we wondered whether oncolytic VSV treatment might resistance to VSV infection in NIH-CUG2 cells, we found that NIH- effectively exert cytolysis in CUG2-expressing cells as a potential CUG2 cells exhibit much higher levels of phosphorylated STAT1 therapeutic anticancer strategy. To address this question, we used than did NIH-Vec and NIH-Ras cells (Figure 2b). However, NIH3T3 cell lines expressing vector only (NIH-Vec), active Ras (NIH- phosphorylated STAT2 was not seen in any of the cell lines Ras) or the CUG2 gene (NIH-CUG2) because we showed oncogenic (Figure 2b). We further compared the levels of phosphorylated activity of CUG2 in non-transformed NIH3T3 cells in our previous STAT1 and STAT2 during type I IFN treatment and VSV infection in study.1 NIH-Vec, NIH-Ras or NIH-CUG2 cells were infected with VSV each cell type. During IFN-b treatment, NIH-Vec cells reached peak (multiplicity of infection (MOI) ¼ 0.1) and harvested at 12, 24 and phosphorylation of STAT1 after 15 min and reduced the levels at 48 h post-infection. We found that NIH-Ras cells were more 30 and 60 min post-treatment (Figure 2c). NIH-Ras cells displayed susceptible to VSV infection compared with NIH-Vec cells (Figures no change in phosphorylated STAT1 levels during IFN treatment, 1a and b) as described elsewhere.22,23 However, NIH-CUG2 cells indicating that these cells may not respond to IFN-b (Figure 2c). that also showed higher levels and activity of Ras were NIH-CUG2 cells showed highly elevated phosphorylation of STAT1 unexpectedly resistant to VSV until 48 h and even at 72 h (Figure 2c). Of interest, phosphorylation of STAT2, which could not (Figures 1a and b: data not shown). To examine whether cell be seen in unstimulated conditions, was drastically enhanced in death observed in NIH-Vec and NIH-Ras cells are attributed to VSV NIH-CUG2 cells, compared with that seen in NIH-Vec and NIH-Ras replication, leading to apoptosis, VSV-G capsid protein and cells (Figure 2c). In addition, phosphorylation of STAT1 was cleaved caspase 8 were detected using immunoblotting. increased in NIH-Vec and NIH-Ras cells at 12 and 24 h following Increased VSV-G-protein levels in NIH-Vec cells over the time VSV infection, whereas elevated levels of phosphorylated STAT1 course of VSV infection and saturation of VSV-G-protein levels in were observed in NIH-CUG2 cells during VSV infection (Figure 2d). NIH-Ras cells at 24 h post-infection was observed (Figure 1c). Of However, VSV infection did not induce STAT2 phosphorylation in interest, we could not detect any VSV-G protein in NIH-CUG2 cells, any of these cells (Figure 2d). These results suggest that CUG2 suggesting that CUG2 inhibits VSV replication, contrary to our overexpression causes an aberrant response to type I IFN and VSV initiation prediction. Cleaved caspase 8, an indicator of apoptotic infection. cell death, was detected at 48 h post-infection in NIH-Vec and NIH- Ras cells but not in NIH-CUG2 cells, which is consistent with the Suppression of STAT1 renders NIH-CUG2 cells susceptible to VSV extent of VSV replication (Figure 1c). infection To confirm that activation of STAT1 due to CUG2 overexpression CUG2 induces aberrant STAT1 activation in NIH3T3 cells causes resistance to VSV, we suppressed STAT1 expression with As it has been reported that upregulation of Ras contributes to siRNA and then infected NIH-CUG2 cells with VSV. We first enhanced VSV replication due to suppression of IFN signaling,23,26 optimized STAT1 siRNA concentration to reduce STAT1 levels at

Figure 1. NIH-cancer upregulated gene 2 (CUG2) cells exhibit resistance to vesicular stomatitis virus (VSV) infection. (a, b) NIH-Vec, NIH-Ras and NIH-CUG2 cells were infected with VSV (multiplicity of infection (MOI) ¼ 0.1) and observed at 12, 24 and 48 h post-infection by light microscope. The live cells were evaluated by trypan blue exclusion at 48 h post-infection. Results were averaged from triplicate wells and the error bars indicate standard deviation. (c) Cell lysates from NIH-Vec, NIH-Ras and NIH-CUG2 cells were prepared and separated on a 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel. The expression of cleaved caspase 8 and VSV G protein were observed by immunoblotting with the corresponding antibodies.

& 2013 Nature America, Inc. Cancer Gene Therapy (2013), 125 – 132 CUG2 confers resistance to VSV through STAT1 W Malilas et al 128 200 nM of STAT1 siRNA (Figure 3a). VSV infection alone or STAT1 combined treatment with VSV and STAT1 siRNA induced siRNA treatment alone as a control did not induce cytolysis in NIH- signiﬁcant susceptibility to VSV infection at 60 h post-infection CUG2 cells at 60 h post-infection or post-treatment. However, (Figure 3b). Consistent with these results, combined treatment

Figure 2. NIH-cancer upregulated gene 2 (CUG2) cells aberrantly respond to interferon (IFN)-b treatment and vesicular stomatitis virus (VSV) infection. (a) Glutathione S-transferase (GST) or GST-Raf-RBD protein was added to cell lysates from NIH-Vec, NIH-Ras and NIH-CUG2 cells and active Ras was pulled down using glutathione agarose. The Ras protein was detected using anti-Ras antibody in the precipitation mixture after gel electrophoresis. (b) Cell lysates from NIH-Vec, NIH-Ras and NIH-CUG2 cells were separated on 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and phospho-signal transducers and activators of transcription 1 (STAT1), STAT1, phospho-STAT2 and STAT2 were detected using the corresponding antibodies. (c, d) NIH-Vec, NIH-Ras and NIH-CUG2 cells were treated with IFN-b (100 U) or VSV (multiplicity of infection (MOI) ¼ 0.1) and then harvested at the different time intervals after treatment. The cell lysates were separated on 10% SDS-PAGE and phospho-STAT1, STAT1, phospho-STAT2 and STAT2 were detected with the corresponding antibodies.

Figure 3. Suppression of signal transducers and activators of transcription 1 (STAT1) converts vesicular stomatitis virus (VSV) resistance to susceptibility in NIH-cancer upregulated gene 2 (CUG2) cells. (a) NIH-CUG2 cells were treated with control or STAT1 short interference RNA (siRNA) at the different concentrations for 48 h and the expression of STAT1 was examined by immunoblotting using its antibody. (b, c)NIH-CUG2cells were treated with control or STAT1 siRNA at 24 h before VSV infection and subsequently infected with VSV (multiplicity of infection (MOI) ¼ 0.1). The cells were observed for 60 h and live cells were measured at 60 h post-infection with trypan blue exclusion. Results were averaged from triplicate wells and the error bars indicate standard deviation (*Po0.01; control siRNA plus VSV vs STAT1 siRNA plus VSV). The cell lysates were prepared 60 h post-infection and the VSV G protein and caspase 8 precursor were detected using the corresponding antibodies.

Cancer Gene Therapy (2013), 125 – 132 & 2013 Nature America, Inc. CUG2 confers resistance to VSV through STAT1 W Malilas et al 129 with VSV and STAT1 siRNA reduced the number of surviving cells increased compared with that in NIH-Vec cells (Figure 4b). To compared with STAT1 siRNA treatment or control siRNA plus VSV confirm how much transcripts of these anti-viral genes were infection (Figure 3b; Po0.01). Furthermore, to investigate whether upregulated, quantitative real-time RT-PCR was performed. As cell death in NIH-CUG2 cells after the combined treatment may be seen in the results of microarrays and conventional RT-PCR attributable to apoptosis evoked by the activation of caspase 8, (Figures 4a and b), OASL2 transcripts were upregulated more than we examined the level of cleaved caspase 8 cleavage in cell lysates 40-fold in NIH-CUG2 cells compared with those in NIH-Vec cells following combined treatment. STAT1 siRNA treatment or VSV (Figure 4c). Transcripts of the other genes, IFIT3, GBP, IP-10 and infection alone did not induce activation of caspase 8, whereas the ISG-15, were increased more than 15-fold compared with those in combined treatment did (Figure 3c). In addition, observation of NIH-Vec cells (Figure 4c). Moreover, to identify which protein the VSV capsid G protein at 60 h confirmed that apoptosis is among the anti-viral proteins is most affected by STAT1 activation, directly related to VSV replication following combined treatment we examined mRNA levels of the proteins through RT-PCR and (Figure 3c). Accordingly, we propose that STAT1 activation due to real-time RT-PCR after STAT1 siRNA treatment. Transcripts of CUG introduction stimulates the formation of an anti-viral state, OASL2 were most significantly decreased among them after leading to resistance to VSV infection. STAT1 siRNA treatment, whereas transcripts of IP-10 were least decreased (Figures 4d and e). Based on these results, we suggest that upregulation of CUG2-mediated OASL2 transcripts is depen- Upregulation of OASL2 transcripts mediated by CUG2 is dent on STAT1 activation. dependent on STAT1 expression Based on our observation that STAT1 activation renders cells resistant to VSV, we investigated which downstream molecule of Suppression of OASL2 renders NIH-CUG2 cells susceptible to VSV STAT1 is directly involved in its anti-viral activity. To answer this infection question, we compared transcript levels of anti-viral proteins in To confirm that upregulation of OASL2 transcripts due to CUG2 NIH-Vec and NIH-CUG2 cells using a microarray assay. We then confers resistance to VSV, we suppressed OASL2 expression using found that transcripts of IP-10, GBP3, IFIT3 and OASL2 were its siRNA and then infected NIH-CUG2 cells with VSV. We first upregulated more than fourfold in NIH-CUG2 cells compared with optimized conditions for the suppression of OASL2 with its siRNA those in NIH-Vec cells (Figure 4a). In particular, OASL2 transcripts (200 nM; Figure 5a) and confirmed the reduced expression of were increased more than 10-fold in NIH-CUG2 cells compared OASL2. VSV infection alone or OASL2 siRNA treatment alone as a with those in NIH-Vec cells (Figure 4a). Furthermore, we confirmed control did not induce cytolysis in NIH-CUG2 cells at 72 h post- that the elevated transcripts observed in microarray were actually infection or post-treatment. However, combined treatment with detected as upregulated levels through RT-PCR assays (Figure 4b). VSV and OASL2 siRNA induced significant susceptibility to VSV When the transcript levels of ISG-15, a well-known anti-viral infection at 48 and 72 h post-infection (Figure 5b). Consistent with protein, were also examined using RT-PCR, the level of ISG-15 was these results, combined treatment with VSV and OASL2 siRNA

Figure 4. Signal transducers and activators of transcription 1 (STAT1) activation mediated by cancer upregulated gene 2 (CUG2) mostly affects upregulation of 20–50 oligoadenylate synthetase-like (OASL)2 transcripts. (a) Total RNAs isolated from NIH-Vec and NIH-CUG2 cells were subjected to a microarray assay. Genes exhibiting more than fourfold increase among interferon (IFN)-dependent genes were shown. (b, c) Total RNAs isolated from NIH-Vec and NIH-CUG2 cells were subjected to qualitative reverse transcription-polymerase chain reaction (RT-PCR) and real-time PCR. Lane 1 indicates transcripts from NIH-Vec cells and lane 2 indicates transcripts from NIH-CUG2 cells. All reactions were performed in triplicate and data are expressed as mean of relative mRNA levels in real-time PCR. (d, e) Total RNAs from NIH-CUG2 cells transfected with STAT1 short interference (siRNA) and control siRNA were also subjected to qualitative RT-PCR and real-time PCR. Lane 1 indicates transcripts from NIH-CUG2 cells transfected with control siRNA and lane 2 indicates transcripts from NIH-CUG2 cells transfected with STAT1 siRNA. All reactions were performed in triplicate and data are expressed as the mean of relative mRNA levels in real-time PCR.

& 2013 Nature America, Inc. Cancer Gene Therapy (2013), 125 – 132 CUG2 confers resistance to VSV through STAT1 W Malilas et al 130

Figure 5. Suppression of 20–50 oligoadenylate synthetase-like (OASL)2 mRNA levels sensitizes NIH-cancer upregulated gene 2 (CUG2) cells to vesicular stomatitis virus (VSV)-induced apoptosis. (a) NIH-CUG2 cells were treated with control or OASL2 short interference (siRNA) at different concentrations for 48 h and the OASL2 mRNA level was examined by reverse transcription-polymerase chain reaction (RT-PCR). (b, c) NIH-CUG2 cells were treated with control or OASL2 siRNA and subsequently infected with VSV (multiplicity of infection (MOI) ¼ 0.1) after 24 h. The cells were observed for 72 h and live cells were counted by trypan blue exclusion. Results were averaged from triplicate wells and the error bars indicate standard deviation (*Po0.01; OASL2 siRNA plus VSV vs control siRNA plus VSV, OASL2 siRNA alone or control siRNA alone). Cell lysates were prepared at 72 h post-infection and the VSV G protein and cleaved poly-ADP-ribose polymerase (PARP) were detected using the corresponding antibodies.

resulted in a reduction in surviving cell numbers compared with suppression renders Colon26L5-CUG2 cells susceptible to VSV. OASL2 siRNA treatment or VSV infection alone (Po0.01) To address this question, we introduced Colon26L5-CUG2 cells (Figure 5b). Furthermore, to test whether cell death in NIH-CUG2 stably suppressing STAT1 (Colon26L5-CUG2-ShSTAT1), which was cells after the combined treatment is attributable to apoptosis, we conﬁrmed by immunoblotting (in preparation) and its control examined the cleavage of PARP, which indicates activation of Colon26L5-CUG2-ShCon cells. Both of these cells were infected caspase 3 or 7, in cells subjected to combined treatment. The with VSV (MOI ¼ 0.1) and cell death was then examined for 36 h. combined treatment did induce the cleavage of PARP during VSV- We found that Colon26L5-CUG2-ShSTAT1 cells were susceptible to induced apoptosis (Figure 5c). Furthermore, observation of the VSV, whereas Colon26L5-CUG2-ShCon cells were resistant to VSV VSV capsid G protein at 48 h conﬁrmed that apoptosis is directly (Figure 6b). Moreover, we observed that cell death of Colon26L5- related to VSV replication following combined treatment CUG2-ShSTAT1 cells was mediated by VSV replication and (Figure 5c). Accordingly, we propose that STAT1 activation due attributed to apoptosis (Figure 6c). These results indicate that to CUG2 introduction stimulates the formation of an anti-viral CUG2 expression induces activation of STAT1, which contributes state, leading to resistance to VSV infection. resistance to VSV infection.

CU2G2-mediated STAT1 activation confers resistance to VSV in DISCUSSION Colon26-L5 carcinoma cells In general, it has been known that malignant cells are usually Furthermore, we wondered whether CUG2 expression induces more susceptible to oncolytic VSV infection, whereas normal cells activation of STAT1, and resistance to VSV in any other cells. We are relatively resistant to the viral infection. However, it is often introduced CUG2 into colon carcinoma, Colon26-L5 cells. We difficult to understand precisely why some malignant cell confirmed that STAT1 activation was detected in Colon26-L5 cells populations remain resistant to viral oncolysis. Several studies stably expressing CUG2 (Colon26L5-CUG2) (in preparation). have revealed that activation of the Ras/Raf/MEK signaling Colon26L5-CUG2 and its control, Colon26-L5 cells stably expres- pathway attenuates IFN signaling-mediated anti-viral action, sing a control vector (Colon26L5-Vec) were infected with VSV thereby preventing the normal inhibition of VSV RNA synthesis (MOI ¼ 0.1). Cell death was then monitored for 36 h. We found that or reduction of STAT2 levels.22,23,26 These lines of evidence would Colon26L5-Vec cells were susceptible to VSV, whereas Colon26L5- explain the preferential replication of VSV that is commonly CUG2 cells were resistant to VSV, as seen in NIH-Vec and NIH- observed in tumor cells rather than non-transformed cells. On the CUG2 cells (Figure 6a). We identified that cell death of Colon26L5- basis of these reports, we expected that NIH-CUG2 cells should be Vec cells was mediated by VSV propagation, leading to apoptosis susceptible to VSV infection as we observed earlier that CUG2 (Figure 6c). In addition, we next examined whether STAT1 upregulates the activity and levels of Ras and MAPKs including

Cancer Gene Therapy (2013), 125 – 132 & 2013 Nature America, Inc. CUG2 confers resistance to VSV through STAT1 W Malilas et al 131

Figure 6. Cancer upregulated gene 2 (CUG2)-mediated signal transducers and activators of transcription 1 (STAT1) activation determines resistance to vesicular stomatitis virus (VSV) infection in Colon26L5 cells. (a, b) Colon26L5-Vec, Colon26L5-CUG2, Colon26L5-CUG2-ShCon and Colon26L5-CUG2-ShSTAT1 cells were infected with VSV (multiplicity of infection (MOI) ¼ 0.1) and observed at 0, 12, 24 and 36 h post-infection under light microscope. (c) These cells were harvested at 24 and 36 h post-infection and cell lysates from the cells were prepared and separated on a 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel. The expression of cleaved poly-ADP-ribose polymerase (PARP) and VSV G protein were observed by immunoblotting with the corresponding antibodies.

ERK. However, we were surprised to find that NIH-CUG2- situation is different from theirs. To overcome the enervated transformed cells are actually resistant to VSV, whereas NIH-Ras- oncolytic action of VSV in cells responding to IFN signaling, one transformed cells are more susceptible to VSV compared with group has attempted to downregulate the IFN receptor in bladder non-transformed NIH-Vec cells. We attribute this unexpected cancer cells, which eventually enables VSV replication in the observation to CUG2-mediated STAT1 activation, leading to the cells.29 Similarly, in our experiments we have genetically upregulation of the OASL2 anti-viral protein. We furthermore downregulated the cellular anti-viral response by suppression of demonstrated that CUG2 introduction activates STAT1 in relevant the protein levels using siRNA against STAT1 or OASL2. Although colon cells, leading to resistance to VSV (Figure 6). Further our suppression by siRNA treatment of STAT1 or OASL2 did not investigation has also revealed that NIH-CUG2 cells also exhibit a completely eliminate the protein, it appears that we nevertheless constitutive STAT1 activation and more sensitive STAT2 activation surpassed a threshold of viral restriction that was enough to during IFN-b treatment compared with NIH-Ras and NIH-Vec cells. permit VSV replication in NIH-CUG2 cells. As recent studies have The result indicates that NIH-CUG2 cells are still responsive to IFN reported that the existence of new cellular anti-viral signaling signaling, which is contrary to other malignant cells with defective pathways such as Akt-EMSY(a transcriptional repressor)-ISG30 and IFN signaling. In contrast, when we consider that NIH-CUG2 cells protein kinase C-a-histone deacetylase6-b-catenin-IRF3,31 it will be are susceptible to reovirus,3 VSV seems to be more sensitive to interesting to examine whether blockage of these signaling IFN-related anti-viral signals when compared with reovirus. We pathways can render NIH-CUG2 cells susceptible to VSV. also assume that not only Ras but also p38 MAPK activation Herein, we found that the activation of STAT–OASL2 axis contributes to enhance susceptibility to reoviral infection to NIH- signaling mediated by CUG2 exerts unexpected anti-viral activity. CUG2 cells. Thus, we were curious about the mechanism by which CUG2 Mutants of VSV have been generated for use as cytolytic agents might activate STAT1. Interestingly, a recent study has shown that in various cancer cells. Among them, a mutant of matrix (M) CUG2 interacts with nucleophosmin (NPM1/B23) protein.32 NPM1/ protein of VSV displayed less efficient cytolysis of tumor cells but B23 protein has multiple functions, including cell growth, attenuated cytopathic effect on the tumor-bearing animal due to proliferation and resistance to stress.33 We additionally found lower capacity to induce IFN production during the viral infection that NPM1/B23 protein is closely related to cancer development.34 compared with wild-type VSV.27 Recent study has shown that a Based on these works, we speculate that CUG2, a component of mutant of glycoprotein in VSV exhibited efficient cytolysis and less kinetochores, has a role in the development of cancer at least in cytopathogenicity compared with a mutant of the matrix protein part through its interaction with NPM1/B23 protein. Cancerous in VSV.28 Cytopathogenicity of VSV was considered as an cells mediated by the CUG2-NPM1/B23 signaling axis produce important factor for the application of VSV to human patients as growth factors and cytokines, including epidermal growth factor a cancer therapeutic drug. Further these studies have reported and interleukin-6, which can then deliver signals to the secretory that normal cells are resistant to VSV infection due to the or neighboring cells in an autocrine or a paracrine manner. production of type I IFN, whereas tumor cells are sensitive to VSV Moreover, it is known that not only IFN signaling but also growth infection due to less production of type I IFN.27,28 However, our factor signaling (such as epidermal growth factor) can activate

& 2013 Nature America, Inc. Cancer Gene Therapy (2013), 125 – 132 CUG2 confers resistance to VSV through STAT1 W Malilas et al 132 35 STAT1. Taken together, we propose that the interaction of CUG2 12 Li Q, Tainsky MA. Epigenetic silencing of IRF7 and/or IRF5 in lung cancer cells with NPM1/B23 protein can activate epidermal growth factor leads to increased sensitivity to oncolytic viruses. PLoS One 2011; 6: e28683. release for cell proliferation, leading thereby to STAT1 activation. 13 Stephenson KB, Barra NG, Davies E, Ashkar AA, Lichty BD. Expressing human To answer more precisely how CUG2 overexpression might interleukin-15 from oncolytic vesicular stomatitis virus improves survival in a activate STAT1 through its interaction with NPM1/B23 protein, murine metastatic colon adenocarcinoma model through the enhancement of we are continuing our studies. We observed elevated transcripts anti-tumor immunity. Cancer Gene Ther 2012; 19: 238–246. 14 Porosnicu M, Mian A, Barber GN. The oncolytic effect of recombinant levels of OASL2, ISG15, IP-10, IFIT3 and GBP3 in NIH-CUG2 cells vesicular stomatitis virus is enhanced by expression of the fusion cytosine and discovered by the analysis of a public database (www. deaminase/uracil phosphoribosyltransferase suicide gene. Cancer Res 2003; 63: genecards.org) that all of them harbor the binding sequence for 8366–8376. STAT1 at their promoter sites, except for GBP3. This evidence 15 Hovanessian AG. On the discovery of interferon-inducible, double-stranded RNA supports our finding that the expression of OASL2 transcripts is activated enzymes: the 20–50 oligoadenylate synthetases and the protein kinase dependent on the STAT1 expression level (Figures 4d and e). PKR. Cytokine Growth Factor Rev 2007; 18: 351–361. Although it is known that OAS activated by double-stranded RNA 16 Justesen J, Hartmann R, Kjeldgaard NO. Gene structure and function of the 2’-5’- stimulates RNase L, leading to degradation of double stranded- oligoadenylate synthetase family. Cell Mol Life Sci 2000; 57: 1593–1612. viral RNA or decreased initiation of viral protein synthesis, we are 17 Skaug B, Chen ZJ. Emerging role of ISG15 in antiviral immunity. Cell 2010; 143: 187–190. interested in understanding the mechanism by which CUG2 18 MacMicking JD. IFN-inducible GTPases and immunity to intracellular pathogens. suppresses viral propagation by inhibition of RNA replication or Trends Immunol 2004; 25: 601–609. through blockade of viral protein synthesis. The molecular mecha- 19 Carter CC, Gorbacheva VY, Vestal DJ. Inhibition of VSV and EMCV replication by nism by which CUG2 confers resistance against VSV infection will the interferon-induced GTPase, mGBP-2: differential requirement for wild-type provide new information that will be useful in developing drugs to GTP binding domain. Arch Virol 2005; 150: 1213–1220. treat cancer cells overexpressing oncogenic CUG2. 20 Liu M, Guo S, Hibbert JM, Jain V, Singh N, Wilson NO et al. CXCL10/IP-10 in infectious diseases pathogenesis and potential therapeutic implications. Cytokine Growth Factor Rev 2011; 22: 121–130. CONFLICT OF INTEREST 21 Liu XY, Chen W, Wei B, Shan YF, Wang C. IFN-induced TPR protein IFIT3 The authors declare no conflict of interest. potentiates antiviral signaling by bridging MAVS and TBK1. J Immunol 2011; 187: 2559–2568. 22 Noser JA, Mael AA, Sakuma R, Ohmine S, Marcato P, Lee PW et al. The RAS/Raf1/ ACKNOWLEDGEMENTS MEK/ERK signaling pathway facilitates VSV-mediated oncolysis: implication for the defective interferon response in cancer cells. Mol Ther 2007; 15: This study was supported by a grant from the National R&D Program for Cancer 1531–1536. Control, Ministry of Health and Welfare, Republic of Korea (1120140) and the World 23 Battcock SM, Collier TW, Zu D, Hirasawa K. Negative regulation of the alpha Class University Program (R31-2008-000-20004-0) through the National Research interferon-induced antiviral response by the Ras/Raf/MEK pathway. J Virol 2006; Foundation funded by the Korean government. 80: 4422–4430. 24 Ayroldi E, Zollo O, Macchiarulo A, Di Marco B, Marchetti C, Riccardi C. Gluco- corticoid-induced leucine zipper inhibits the Raf-extracellular signal-regulated REFERENCES kinase pathway by binding to Raf-1. Mol Cell Biol 2002; 22: 7929–7941. 1 Lee S, Gang J, Jeon SB, Choo SH, Lee B, Kim YG et al. Molecular cloning and 25 Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time functional analysis of a novel oncogene, cancer-upregulated gene 2 (CUG2). quantitative PCR and the 2( À Delta Delta C(T)) Method. Methods 2001; 25: Biochem Biophys Res Commun 2007; 360: 633–639. 402–408. 2 Kim H, Lee M, Lee S, Park B, Koh W, Lee DJ et al. Cancer-upregulated gene 2 26 Christian SL, Collier TW, Zu D, Licursi M, Hough CM, Hirasawa K. Activated Ras/ (CUG2), a new component of centromere complex, is required for kinetochore MEK inhibits the antiviral response of alpha interferon by reducing STAT2 levels. function. Mol Cells 2009; 27: 697–701. J Virol 2009; 83: 6717–6726. 3 Park EH, Park EH, Cho IR, Srisuttee R, Min HJ, Oh MJ et al. CUG2, a novel oncogene 27 Ahmed M, Cramer SD, Lyles DS. Sensitivity of prostate tumors to wild type and M confers reoviral replication through Ras and p38 signaling pathway. Cancer Gene protein mutant vesicular stomatitis viruses. Virology 2004; 330: 34–49. Ther 2010; 17: 307–314. 28 Janelle V, Brassard F, Lapierre P, Lamarre A, Poliquin L. Mutations in the glyco- 4 Rose JK, Whitt MA. Rhabdoviridae: The Viruses and their Replication. In Knipe DM, protein of vesicular stomatitis virus affect cytopathogenicity: potential for onco- Howley PM (eds). Field Virology. 4th edn, Lippincott Williams and Wilkins: New lytic virotherapy. J Virol 2011; 85: 6513–6520. York, 2001, 1221–1244. 29 Zhang KX, Matsui Y, Hadaschik BA, Lee C, Jia W, Bell JC et al. Down-regulation of 5 de Mattos C, de Mattos CC, Rupprecht CE. Rhabdoviridae: The Viruses and their type I interferon receptor sensitizes bladder cancer cells to vesicular stomatitis Replication. In: Knipe DM, Howley PM (eds). Field Virology. 4th edn, Lippincott virus-induced cell death. Int J Cancer 2010; 127: 830–838. Williams and Wilkins: New York, 2001, 1245–1277. 30 Ezell SA, Polytarchou C, Hatziapostolou M, Guo A, Sanidas I, Bihani T et al. 6 Bischoff JR, Kirn DH, Williams A, Heise C, Horn S, Muna M et al. An adenovirus The protein kinase Akt1 regulates the interferon response through phosphor- mutant that replicates selectively in p53-deficient human tumor cells. Science ylation of the transcriptional repressor EMSY. Proc Natl Acad Sci USA 2012; 109: 1996; 274: 373–376. E613–E621. 7 Coffey MC, Strong JE, Forsyth PA, Lee PW. Reovirus therapy of tumors with 31 Zhu J, Coyne CB, Sarkar SN. PKC alpha regulates Sendai virus-mediated activated Ras pathway. Science 1998; 282: 1332–1334. interferon induction through HDAC6 and beta-catenin. EMBO J 2011; 30: 8 Varghese S, Rabkin SD. Oncolytic herpes simplex virus vectors for cancer vir- 4838–4849. otherapy. Cancer Gene Ther 2002; 9: 967–978. 32 Chun Y, Park B, Koh W, Lee S, Cheon Y, Kim R et al. New centromeric component 9 Stojdl DF, Lichty B, Knowles S, Marius R, Atkins H, Sonenberg N et al. Exploiting CENP-W is an RNA-associated nuclear matrix protein that interacts with nucleo- tumor-specific defects in the interferon pathway with a previously unknown phosmin/B23 protein. J Biol Chem 2011; 286: 42758–42769. oncolytic virus. Nat Med 2000; 6: 821–825. 33 Okuwaki M. The structure and functions of NPM1/Nucleophsmin/B23, a multi- 10 Fernandez M, Porosnicu M, Markovic D, Barber GN. Genetically engineered vesi- functional nucleolar acidic protein. J Biochem 2008; 143: 441–448. cular stomatitis virus in gene therapy: application for treatment of malignant 34 Grisendi S, Mecucci C, Falini B, Pandolfi PP. Nucleophosmin and cancer. Nat Rev disease. J Virol 2002; 76: 895–904. Cancer 2006; 6: 493–505. 11 Leveille S, Samuel S, Goulet ML, Hiscott J. Enhancing VSV oncolytic activity with an 35 Khodarev NN, Roizman B, Weichselbaum RR. Molecular pathways: interferon/stat1 improved cytosine deaminase suicide gene strategy. Cancer Gene Ther 2011; 18: pathway: role in the tumor resistance to genotoxic stress and aggressive growth. 435–443. Clin Cancer Res 2012; 18: 3015–3021.

Cancer Gene Therapy (2013), 125 – 132 & 2013 Nature America, Inc.