Tumors driven by RAS signaling harbor a natural vulnerability to oncolytic virus M1

RAS/RAF/MAPK/p‐ERK signaling promotes the replication and oncolytic effect of M1 virus via inhibiting P21 and subsequent suppression of Cyclin‐dependent kinase 2 (CDK2). Inhibition of RAS/RAF/MAPK/p‐ERK signaling by sorafenib or U0126 suppresses the oncolytic effect of M1. Knockout of p21 enhances the effect of M1, while overexpression of p21 inhibits it. Inhibition of CDK2 by K03861 attenuates the efficacy of M1.

Oncolytic viruses are potent anticancer agents that replicate within and kill cancer cells rather than normal cells, and their selectivity is largely determined by oncogenic mutations. M1, a novel oncolytic virus strain, has been shown to target cancer cells, but the relationship between its cancer selectivity and oncogenic signaling pathways is poorly understood. Here, we report that RAS mutation promotes the replication and oncolytic effect of M1 in cancer, and we further provide evidence that the inhibition of the RAS/RAF/MEK signaling axis suppresses M1 infection and the subsequent cytopathic effects. Transcriptome analysis revealed that the inhibition of RAS signaling upregulates the type I interferon antiviral response, and further RNA interference screen identified CDKN1A as a key downstream factor that inhibits viral infection. Gain-and loss-of-function experiments confirmed that CDKN1A inhibited the replication and oncolytic effect of M1 virus. Subsequent TCGA data mining and tissue microarray (TMA) analysis revealed that CDKN1A is commonly deficient in human cancers, suggesting extensive clinical application prospects for M1. Our report indicates that virotherapy is feasible for treating undruggable RASdriven cancers and provides reliable biomarkers for personalized cancer therapy.

Introduction
Oncolytic viruses offer a new approach for cancer therapy, which exploits tumor mutations to specifically replicate within and kill tumor cells without causing harm to normal cells [1][2][3][4][5]. Oncolytic viruses also activate the antitumor immune response in addition to direct killing [6][7][8]. Since Talimogene laherparepvec became the first oncolytic virus approved by the US Food and Drug Administration (FDA) [9], numerous clinical trials have begun [10] and are expected to further boost the development of oncolytic therapies.
The tumor selectivity of oncolytic viruses is largely conferred by tumor-specific aberrations in signaling pathways that normally sense and block viral replication. It is now well established that cancer-specific aberrations in BCL-2, WNT, EGFR, RAS, TP53, RB1, PTEN, and other cancer-related genes predispose cancer cells to viral infection [7,8,11]. For example, Newcastle disease virus targets cancer cells overexpressing BCL-XL, which prevents apoptosis and thereby permits the virus to utilize the transcription and translation machinery for the synthesis of the viral nucleocapsid [12]. The activation of WNT signaling, a key pathway in embryonic development that directs cell proliferation, polarity, and developmental fate, has been found to attenuate the host antiviral response and facilitate the infection and replication of several kinds of viruses [13][14][15]. In addition, cancer cells with RAS mutations cannot activate the PKR pathway which functions to prevent the production and spread of virus, rendering cancer cells permissive to reovirus, herpesvirus, and vaccinia virus infection [16][17][18][19].
M1 virus is an enveloped alphavirus with an 11.7 kb positive single-stranded RNA genome [20], which contains four nonstructural proteins and five structural proteins. Our previous studies demonstrated that M1 is a potent oncolytic virus that selectively targets and induces irreversible endoplasmic reticulum stress-mediated apoptosis in different cancers in vitro and in vivo [21][22][23]. More interestingly, the oncolytic effect of M1 can be enhanced by small-molecule compounds, including BCL-XL inhibitors, Smac mimetics, and DNA-PK inhibitors [24][25][26][27][28][29]. Our previous study established that M1 virus is a promising oncolytic virus for clinical cancer therapy. Though we have identified that the deficiency of zinc-finger antiviral protein mediates the cancer selectivity of M1, however, the relationship between the cancer selectivity of M1 virus and oncogenic signals has not yet been illuminated.
In this study, we analyzed the relationship between viral infection and oncogenic mutations in 52 tumor cell lines and found that among the mutations in these cell lines, K-RAS aberration promotes viral infection. Further expression profiling identified CDKN1A as a key factor downstream of the RAS/RAF/MEK signaling pathway that inhibits the replication of M1 virus. The knockdown of CDKN1A enhances the oncolytic effect of M1 virus in nude mice bearing human tumor cells, which largely represents the characteristics of human cancer. This study identifies RAS mutation and deficiency of CDKN1A as candidate biomarkers for personalized anticancer virotherapy.

Cell culture and M1 viruses
All cell lines were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). Cells were cultured in Dulbecco's Modified Eagle's medium supplemented with 10% (vol/vol) FBS (Thermo Fisher Scientific, Waltham, MA, USA) and 1% penicillin/streptomycin (Thermo Fisher Scientific). All cell lines were cultured at 37°C in a 5% CO 2 environment. All cell lines were authenticated by the short tandem repeat assay and were mycoplasma free according to the MycoGuard Mycoplasma PCR Detection Kit (MPD-T-050; GeneCopoeia, Rockville, MD, USA).
The M1 virus was grown in the Vero cell line and collected for experiments. The M1-c6v1 strain of virus was provided by Guangzhou Virotech Pharmaceutical Co., Ltd, Guangzhou, Guangdong, China. M1-GFP is a recombinant M1 engineered to express jellyfish green fluorescent protein [28]. The viral titer was determined by the TCID50 method using the BHK-21 cell line and converted to plaque-forming unit (PFU).

Cell viability assay
Cells were seeded in 96-well plates at 3000 cells per well. After different treatments indicated in the figure legends were administered, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) was added (1 mgÁmL À1 ) and incubated at 37°C for 3 h. The supernatants were removed, and the MTT precipitate was dissolved in 100 lL of DMSO. The optical absorbance was determined at 570 nm by a microplate reader (Synergy H1; Gene Company, Hong Kong, China).

RT-qPCR
Total RNA was extracted using TRIzol (Thermo Fisher Scientific), and 2 lg of total RNA was reversetranscribed to cDNA with oligo (dT) (synthesized by Thermo Fisher Scientific) and RevertAid Reverse Transcriptase (Thermo Fisher Scientific). The expression levels of the specific genes were calculated by the comparative C t method using SuperReal PreMix

Western blot analysis
Cell samples were prepared using M-PER Mammalian Protein Extraction Reagent (Thermo Fisher Scientific) and then separated by SDS/PAGE. The membranes were visualized with a ChemiDoc XRS + System (Bio-Rad, Hercules, CA, USA) using Immobilon Western Chemiluminescent HRP Substrate (Millipore, Darmstadt, Germany).

Animal models
The mouse study was approved by the Animal Ethics and Welfare Committee of Sun Yat-sen University, and all experiments were conducted according to the US 'Public Health Service Policy on Humane Care and Use of Laboratory Animals'. HCT-15-negative control (NC) and HCT-15-shCDKN1A (5 9 10 6 cells/ mouse) tumor cells were implanted subcutaneously into the hind flanks of 5-week-old 16 g female BALB/ c-nu/nu mice (the mice were bought from Nanjing Biomedical Research Institute, China, and housed in an SPF facility with normal temperature and food). After 6 days, tumors were observed (~50 mm 3 ). M1 virus (3.48 9 10 8 TCID50 per mouse) was injected intravenously for 14 days. The lengths and widths of the tumors were measured every 3 days, and the tumor volume was calculated according to the formula (length 9 width 2 )/2. At the termination of the experiment, all mice were euthanized by overdose anesthesia, and the tumors were removed and fixed in 4% paraformaldehyde for subsequent immunohistochemistry (IHC) assays. The study was randomized and blind.
For detection of M1 viral copy number in tumor tissues, 3 days after the first medication, three mice in M1 virus-treated HCT-15-NC and HCT-15-shCDKN1A tumor groups were euthanized by overdose anesthesia, and tumors were stripped out. Total RNA in tumors was extracted by EastepÒ Super Total RNA Extraction Kit (Promega, Madison, Wisconsin, USA). Viral copy numbers were detected by TaqMan qRT-PCR with FastKing One Step RT-qPCR Kit (TIANGEN) according to the manufacturer's instructions. The sequence of primers and probe was listed as below: Forward primer: 5 0 -GGGATTCACTACACCTGCT-TAGAC-3 0 Reverse primer: 5 0 -GCTGACTCTGTCTGCGTAA CC-3 0 Prober: 5 0 -CTCTCATCAGCAGCGAGCCTCCT-3 0

Immunohistochemistry (IHC) assay
The expression of Ki-67 and cleaved caspase 3 in the tumors was assessed by IHC. Briefly, tumor sections were dewaxed in xylene, hydrated in decreasing concentrations of ethanol, immersed in 0.3% H 2 O 2 -methanol for 30 min, washed with phosphate-buffered saline, and probed with monoclonal antibodies or isotype controls at 4°C overnight. After being washed, the sections were incubated with biotinylated goat anti-rabbit or anti-mouse IgG at room temperature for 2 h. Immunostaining was visualized with streptavidin/ peroxidase complex and diaminobenzidine, and sections were counterstained with hematoxylin.

Statistics
All statistical analyses were performed using GRAPHPAD PRISM software 8.0 (RRID: SCR_002798) (GraphPad Software Inc., San Diego, California, USA) and SPSS 18.0 software (RRID: SCR_002865) (IBM SPSS Statistics Inc., Chicago, Illinois, China). Most of the data were analyzed by a two-tailed Student's t-test or oneway ANOVA with Dunnett's test for pairwise comparisons. Tumor volumes were analyzed by a two-tailed paired Student's t-test. Correlations were analyzed by the Pearson's test. The expression of CDKN1A in paired cancer and adjacent non-neoplastic tissue was analyzed by a two-tailed paired Student's t-test. Phasecontrast and fluorescence pictures were taken with a Nikon Eclipse A1 microscope (Tokyo, Japan). The IHC staining intensity was analyzed by IMAGESCOPE software (ImageScope, RRID: SCR_014311). The Wilcoxon signed-rank test was used to compare paired non-normally distributed data. Bars show the mean AE SD or SEM of three independent repeated experiments. Significant differences were accepted if the P-value was < 0.05.

Results
3.1. Tumor cell lines harboring K-RAS mutation are more sensitive to M1 virus than those without K-RAS mutation To investigate the relationship between the selectivity of M1 virus and oncogenic signals in tumors, we analyzed the relationship between the oncolytic effect (represented by values of EC50) of M1 virus (Table S1) and cellular oncogenic mutations in 52 tumor cell lines originating from various types of tissue. The oncogenic mutation data were retrieved from the Cancer Cell Line Encyclopedia (CCLE) database [30], and all of the mutations in the 52 cell lines are listed in Table S2. By analyzing the EC50 values of M1 virus in these cell lines with or without oncogenic mutations, we found that the K-RAS, MDN1, RYR3, and PIEZO2 genes are frequently mutated in the cell lines with higher sensitivity to M1 virus (Fig. 1A). Of the genes listed, K-RAS is the most notable and undruggable target, and further statistical analysis confirmed that M1 virus showed lower EC50 values which represent better antitumor effects in the cell lines with K-RAS mutation than those with wild-type K-RAS (Fig. 1B and  Table S1). To test whether mutant K-RAS indeed affects the sensitivity of cancer cells to M1 virus, specific siRNAs to K-RAS were used to knock down the expression of K-RAS in HCT-15 and MIA PaCa-2 cell lines (Fig. S1), which harbor the mutant K-RAS (indicated in Table S1). In these cell lines, knockdown of K-RAS inhibited the cell killing by M1 virus (Fig. 1C,D) and suppressed the infection rate of M1 virus as shown by flow cytometry and fluorescence imaging (Fig. 1E-G). Moreover, knockdown of K-RAS inhibited the replication of M1 virus as shown by viral titer detected by TCID50 method (Fig. 1H-I). Taken together, the results suggest that K-RAS mutation promotes the replication and subsequent oncolytic effect of M1 virus.

RAS/RAF/MEK signaling inhibitors suppress the oncolytic efficiency and gene expression of M1 virus
RAS mutation results in the activation of the RAS/ RAF/MEK pathway, so we investigated whether inhibitors of the RAS/RAF/MEK pathway affect the oncolytic efficiency and gene expression of M1 virus in tumor cells. Sorafenib is an approved drug for the treatment of different types of cancer that has been reported to preferably inhibit the activity of RAF [31], while U0126 selectively inhibits MEK1/2 [32]. In the pancreatic carcinoma cell line MIA PaCa-2 and colorectal carcinoma cell line HCT-15, both sorafenib and U0126 modestly but significantly inhibited the oncolytic effect of M1 ( Fig. 2A-D). Cobimetinib and trametinib, two other MEK inhibitors approved by the FDA, also modestly but significantly inhibited the oncolytic effect of M1 virus (Fig. S2A-D). We previously reported that the cancer targeting and killing properties of M1 virus depend on the replication of the virus in cancer cells [23], so we used M1 virus engineered to express the reporter protein GFP (M1-GFP) to trace the gene expression of the virus in tumor cells. Phase-contrast and fluorescence imaging showed that U0126 suppressed the reporter gene expression and M1-induced cytopathic effects (Fig. 2E,F). Cytometry analysis consistently proved that the M1 virus infection rate was modestly but significantly inhibited by U0126 treatment (Fig. 2G,H). Furthermore, the expression of viral proteins E1 and NS3 decreased significantly following U0126 treatment, which effectively inhibited the phosphorylation of ERK (Fig. 2I,J). In conclusion, RAS/RAF/MEK signaling promotes the oncolytic effect of M1 virus by upregulating viral gene expression, which suggests the acceleration of viral replication.

Inhibition of the RAS/RAF/MEK pathway upregulates the expression of antiviral signaling pathway members
To identify the mechanism by which RAS/RAF/MEK signaling promotes the replication of M1, gene expression profiling was performed in the HCT-15 cell line under vehicle, U0126, M1, or M1 plus U0126 treatment conditions. Gene set enrichment analysis (GSEA) proved that U0126 effectively downregulated the RAS signaling pathway (Fig. 3A and Table S3). It is well known that the promotion of viral replication by oncogenic signals is due to crosstalk of them with antiviral pathways, which consist mainly of interferon  (IFN) signaling [7]. Among the three classes of IFNs, type I IFNs are known to be essential for mounting a robust host response against viral infection [33]. We hypothesized that U0126 might upregulate antiviral IFN pathway activity to inhibit the replication and oncolytic effect of M1 virus.
Gene set enrichment analysis revealed that U0126 strongly upregulated the IFN-a and IFN-b response pathways after M1 virus infection (Fig. 3B,C and Tables S4-S5). To identify the key factors upregulated by U0126 to inhibit the replication and oncolytic effect of M1 virus, we focused on the expression of 317 IFN-regulated genes (IRGs), which were identified to be crucial antiviral effectors for alphavirus M1 [23]. The heatmap of the top 20 IRGs upregulated by U0126 plus M1 compared with M1 (MU/M) is shown in Fig. 3D. In addition, qPCR was used to verify the expression of the top 10 IRGs in HCT-15 cell line. We found that the expression of IFIT3, CDKN1A, MX2, IFI27, and ID1 was significantly increased in the M1 plus U0126 group compared with the M1 treatment group (Fig. 3E,F). When we ranked the fold change between the M1 plus U0126 group and the M1 group (MU/M), these five genes comprised the top five (Fig. 3G). Taken together, these results indicate that the inhibition of RAS signaling by U0126 upregulates the IFN-mediated innate immune response of cancer cells, which may result in the attenuation of the replication and oncolytic effect of M1 virus.

CDKN1A is a key antiviral factor downstream of RAS signaling that inhibits the replication of M1 virus
To further identify the specific IRG that inhibits M1 viral infection, we used siRNAs to knock down the     five genes identified above in the HCT-15 cell line, including CDKN1A, IFI27, IFIT3, MX2, and ID1. The siRNA effectively knocked down the expression of these genes (Fig. S3), and only the knockdown of CDKN1A significantly increased the infection of M1 virus (Fig. 4A), indicating that CDKN1A might be a key factor. CDKN1A, also known as p21, is a universal cell cycle inhibitor directly controlled by p53 and p53-independent pathways [34]. To further elucidate the antiviral function of CDKN1A, we knocked down the expression of CDKN1A in two more pancreatic carcinoma cell lines, MIA PaCa-2 and PANC-1 (Fig. 4B). Knocking down CDKN1A increased both the M1 infection and viral titer in these cell lines (Fig. 4C-F). Further knockout of CDKN1A in HCT-15 and PANC-1 cell lines by shRNA also increased the infection and titer of M1 virus (Fig. 4G-I). Further, the lentivirus-mediated overexpression of CDKN1A notably reduced the M1 virus infection and titer (Fig. 4J,K). Moreover, the overexpression of CDKN1A in SW620 colorectal carcinoma cell line which is CDKN1A-defective (Fig. S4) decreased the infection rate of M1 virus and attenuated the subsequent cell killing of M1 (Fig. 4L-N). In summary, these loss-and gain-of-function experiments demonstrated that CDKN1A is the key IRG that is suppressed by RAS/RAF/MEK signaling to promote the replication of M1 virus.
CDKN1A is a tumor suppressor that mainly functions as a cell cycle checkpoint by binding to CDK1, Cyclin-dependent kinase 2 (CDK2), or the CDK4/6 complex [35]. We investigated whether the cell cycle plays an important role in the replication of M1 virus by inhibiting the activity of CDK2, a primary factor mediating the cell cycle inhibition activity of CDKN1A. K03861, an inhibitor of CDK2, modestly but significantly decreased the infection rate of M1 virus in a concentration-dependent manner in HCT-15 and MIA PaCa-2 cell lines (Fig. 5A,B), and titer of M1 virus was subsequently significantly inhibited by K03861 in these cell lines (Fig. 5C,D), phase-contrast and fluorescence imaging also showed the inhibition of  A and B) was detected by flow cytometry, the titer was detected by TCID50 (C and D), and phase-contrast and fluorescence pictures were taken (E). Statistical analysis was performed by two-tailed Student's t-test or one-way ANOVA with Dunnett's test for pairwise comparisons. Error bars represent the mean AE SD obtained from three independent experiments. n.s., not significant; *P < 0.05, **P < 0.01, ***P < 0.001. For TCID50 assay, the starting cell numbers of compared group are the same.  (Fig. 5E). The results indicate that CDKN1A inhibits the replication of M1 virus by suppressing the cell cycle.

Knockdown of CDKN1A enhances the oncolytic effect of M1 virus in vivo
To further validate whether the suppression of CDKN1A enhances the oncolytic effect of M1 virus in vivo, we established an HCT-15 subcutaneous xenograft model in nude mice. HCT-15-NC cells and HCT-15-shCDKN1A cells were implanted in the left and right hind flanks of mice, and M1 virus was injected through the caudal vein (Fig. 6A). Compared with that of HCT-15-NC tumors, the growth of HCT-15-shCDKN1A tumors treated with M1 virus was modestly but significantly inhibited (Fig. 6B), and the tumor size in the HCT-15-shCDKN1A group treated with M1 virus was smaller than that in the other three groups (Fig. 6C). Consistent with the cellular results, M1 viral copy number in HCT-15-shCDKN1A tumors was increased compared with HCT-15-NC tumors (Fig. 6D). Furthermore, immunohistochemical (IHC) staining was performed on tumor tissues to detect the expression of Ki-67 and Cl-casp3 to measure the proliferation and apoptosis of tumor cells. We observed that Ki-67 expression was significantly downregulated and Cl-casp3 expression was coordinately upregulated in the M1 virus-treated HCT-15-shCDKN1A tumors (Fig. 6E,F). These results further support the hypothesis that CDKN1A acts as an antiviral factor to inhibit the oncolytic effect of M1 virus in vivo.

CDKN1A deficiency is a biomarker for M1 therapy
The above CDKN1A loss-and gain-of-function experiments indicated that M1 may specifically target cancer cells deficient in CDKN1A. Therefore, we detected the protein expression of CDKN1A in pancreatic carcinoma and colorectal carcinoma cell lines, which have been reported to frequently harbor K-RAS mutations, and analyzed the correlation between the cell-killing effect of M1 virus and CDKN1A expression. In both pancreatic and colorectal cancer cell lines, there is a significant negative correlation between CDKN1A expression and the killing capacity of M1 virus, which means that the lower the protein expression of CDKN1A is, the greater the cell-killing effect of M1 virus is (Fig. 7A,B, the protein expression and cell-killing effect of M1 are shown in Fig. S4 and Table S6). Furthermore, the relationship between the oncolytic effect of M1 and the expression of CDKN1A was also analyzed in 44 tumor cell lines, and CDKN1A expression was measured via reversed-phase protein array (RPPA) in the CCLE database. The expression of CDKN1A negatively correlated with the oncolytic effect of M1 in 44 tumor cell lines (Fig. 7C, Table S7). These results suggest that the protein expression of CDKN1A might predict the killing effect of M1 virus.
To elucidate the potential of M1 virus personalized therapy by detecting CDKN1A deficiency, we compared the expression of CDKN1A in both tumor and adjacent non-neoplastic tissue specimens from 38 colon cancer patients in the TCGA database [36]. The expression of CDKN1A in tumor tissues was lower than that in adjacent non-neoplastic tissues in over 90% (35/38) of patients (Fig. 7D). Furthermore, we performed IHC on two tissue microarrays (TMAs) containing paired tumor and adjacent non-neoplastic clinical specimens from 143 colon cancer patients. CDKN1A expression was represented by the mean staining intensity, calculated by IMAGESCOPE software (Leica, Weztlar, Germany). The expression of CDKN1A in tumor tissues was lower than that in adjacent non-neoplastic tissues in 58% (83/143) of patients (Fig. 7E,F). Both the database and our experiments strongly suggest that the expression of CDKN1A was frequently deficient in colorectal cancers, which implies the application of M1 virus.

Discussion
In this paper, we report that the oncogenic RAS/RAF/ MEK pathway promotes the replication and oncolytic effect of M1 virus by inhibiting the expression of the key antiviral factor CDKN1A. More importantly, cancer cell lines with lower expression levels of CDKN1A showed greater sensitivity to M1 virus than those with higher levels, and the deficiency of CDKN1A is a ubiquitous event in colorectal cancer patients. Our data suggest that CDKN1A is a suitable biomarker for M1 virus therapy. Our report provides a candidate research program and working model to screen and identify oncogenic pathways as well as biomarkers that other oncolytic viruses utilize to facilitate their replication.
Despite more than three decades of intensive effort, no effective pharmacologic inhibitors of the RAS oncoproteins have reached the clinic, promoting the widely held perception that RAS proteins are 'undruggable' [37]. According to our results, M1 virus might offer a renewed hope for the treatment of cancer with abnormal RAS activation, making RAS-driven tumors 'druggable'. Among all cancer-driven genes, RAS is the most frequently mutated oncogene family in human cancers. The top four cancer types harbor RAS mutation including pancreatic ductal adenocarcinoma (97.7%), colorectal adenocarcinoma (52.2%), multiple myeloma (42.6%), and lung adenocarcinoma (32.2%) [37]. The high percentage of RAS mutation in tumors indicates that M1 virus might be effective in a large proportion of cancer patients.
The hypothesis that oncolytic viruses exploit the deficiencies of antiviral pathways and oncogenic signaling between tumor and normal cells to selectively target and kill cancer cells has been strongly proven. The RAS pathway is a typical oncogenic pathway reported to facilitate the replication of various oncolytic viruses. However, to date, no clinical trials have validated RAS mutations as a biomarker for any oncolytic virus, which may be due to the upregulation of other compensatory pathways downstream of RAS that control the expression of antiviral genes in patients; thus, identifying the antiviral genes regulated by the RAS pathway as biomarkers for oncolytic viruses is urgently required. The expression of components of the IFN pathway, such as PKR and IRF1, has been reported to be downregulated by the RAS pathway. However, the expression of these genes has not been proven to be substantially decreased in tumors. Here, we found a new antiviral gene, CDKN1A, downstream of the RAS pathway to suppress the replication and oncolytic effect of M1 virus, and it has been reported to be a tumor suppressor. Our data clearly support the hypothesis that deficiency in both tumor suppressor and antiviral pathways supports the tumor selection mechanism of oncolytic viruses and provides a new biomarker for accurate oncolytic virotherapy.
CDKN1A is a universal cell cycle inhibitor directly controlled by p53 and p53-independent pathways, and the loss of CDKN1A causes carcinogenesis by inducing growth arrest, regulating the expression of genes associated with senescence, and protecting cells from apoptosis [34,35,38]. Consistently, our data prove that CDKN1A expression is consistently deficient in colon cancer, further supporting the reports that CDKN1A is a tumor suppressor. In this study, we demonstrate that CDKN1A inhibits the replication of M1, whereas the inhibition of CDKN1A promotes the replication and oncolytic effect of M1 virus. In tumors with low  (Fig. S4), and the oncolytic effect of M1 virus is shown in Table S6. tissue than that in adjacent non-neoplastic tissue was calculated. Blue dots represent tumor tissues, red dots represent paired adjacent non-neoplastic tissues, and black lines connect paired tumor and adjacent nonneoplastic tissue from the same patient. n = 38. Statistical analysis was performed by two-tailed paired Student's t-test. (E) IHC staining intensity of CDKN1A expression in TMAs containing paired tumor and adjacent non-neoplastic clinical specimens from 143 colon cancer patients. The IHC level is presented as the mean staining intensity calculated using IMAGESCOPE software (Leica); the percentage of patients showing lower expression of CDKN1A in tumor tissue than that in adjacent non-neoplastic tissue was calculated. Statistical analysis was performed by two-tailed paired Student's t-test. (F) Representative pictures of the IHC in (E). Scale bars, 50 lm. *P < 0.05, **P < 0.01, ***P < 0.001. See also Fig. S4 and Tables S6 and S7. CKDN1A expression, the oncolytic effect of M1 virus is stronger than that in tumors with high CDKN1A expression. We hypothesize that CDKN1A controls the cell cycle, but viral replication needs increased amounts of material and energy, so in cancer cells without CDKN1A to control proliferation, the proliferation rate increases and provides sufficient cellular resources for the synthesis and assembly of new viral particles.
Genetic heterogeneity represents one of the most significant hallmarks of cancer, indicating that universal treatment for all patients is problematic and that personalized medicine is required for cancer therapy. The discovery of biomarkers in tumors provides a potential strategy to solve this problem [39]. The detection of biomarkers before treatment with an anticancer drug decreases the chance of a patient receiving an ineffective medication and largely increases the efficiency of the drug. In 2017, the PD-1 antibody pembrolizumab was approved by the FDA for patients with microsatellite instability-high and different mismatch repair [40]. It is the first anticancer strategy that is distinguished by a biomarker but not the organ origin of the tumor. In this study, we found that the oncolytic effect of M1 virus negatively correlated with the expression of CDKN1A in tumor cells. CDKN1A serves as a tumor suppressor and is deficient in various cancer types in addition to colorectal carcinoma [41], which indicates that CDKN1A expression may serve as an M1 virus biomarker in pan-cancer types.
Sorafenib is the first oral multikinase inhibitor that targets RAF. It was first approved by the FDA for the treatment of advanced renal cell carcinoma in 2005 [31]. Subsequently, it was approved for other indications, including hepatocellular carcinoma and radioactive iodine-refractory differentiated thyroid cancer [42]. We have previously reported that M1 virus has a natural tropism for hepatocellular carcinoma and multiple kinds of cancers [23]. Moreover, we have identified various anticancer chemicals in clinical use or in the clinical trial stage, such as VCP inhibitors [27], DNK-PK inhibitors [28], Smac mimetics [24], and Bcl-XL inhibitors [26], can sensitize tumors to M1 virus. Here, we report that sorafenib inhibits the replication and oncolytic effect of M1 virus. These results suggest that in the future, the combination of M1 virus and sorafenib should be avoided in the clinic.

Conclusions
In summary, our research shows that tumors involving RAS signaling harbor a natural vulnerability to oncolytic M1 virus and illustrate that CDKN1A is the key downstream antiviral factor to predict the oncolytic efficacy of M1 virus.

Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article. Table S1. Oncolytic effect of M1 virus and status in k-Ras in 52 tumor cells. Table S2. The mutation status of whole genome in 52 tumor cells. Table S3. Genes regulated by k-Ras activation in expression profile database. Table S4. Interferon alpha response genes in expression profile database. Table S5. Interferon beta response genes in expression profile database. Table S6. Oncolytic effect of M1 virus in colon cancer cell lines and pancreatic cancer cell lines. Table S7. Oncolytic effect of M1 virus and the protein expression of CDKN1A in 44 tumor cells. Fig. S1. The efficiency of siRNAs to K-RAS.