Blocking of EGFR Signaling is a Latent Strategy for the Improvement of Prognosis of HPV Induced Cancer

Human papillomavirus (HPV) is a dsDNA virus and its high-risk subtypes increase cancer risks. Yet, the mechanism of HPV infection and pathogenesis still remain unclear. Therefore, understanding the molecular mechanisms, and the pathogenesis of HPV are crucial in the prevention of HPV related cancers. In this study, we analyzed cervix squamous cell carcinoma (CESC) and head and neck carcinoma (HNSC) combined data to investigate various HPV induced cancer common feature. We showed that epidermal growth factor receptor (EGFR) was downregulated in HPV positive (HPV+) cancer, and that HPV+ cancer patients exhibited better prognosis than HPV negative (HPV−) cancer patients. Our study also showed that TP53 mutation rate is lower in HPV+ cancer than in HPV− cancer and that TP53 can be modulated by HPV E7 protein. However, there was no significant difference in the expression of wildtype TP53 in both groups. Subsequently, we constructed HPV-human interaction network and found that EGFR is a critical factor. From the network, we also noticed that EGFR is regulated by HPV E7 protein and hsa-miR-944. Moreover, while phosphorylated EGFR is associated with a worse prognosis, EGFR total express level is not significantly correlated with prognosis. This indicates that EGFR activation will induce a worse outcome in HPV+ cancer patients. Further enrichment analysis showed that EGFR downstream pathway and cancer relative pathway are diversely activated in HPV+ cancer and HPV− cancer. In summary, HPV E7 protein downregulates EGFR that downregulates phosphorylated EGFR and inhibit EGFR related pathways which in turn and consequently induce better prognosis. Importance Although HPV infection has been studied in various cancer types, there are only limited studies that have focused on the common effect of HPV related cancer. Consequently, this study focused on CESC and HNSC, two cancer types with high HPV infection proportion in cohort, thereby, intending to dig out the common effects and mechanisms of HPV+ cancers. Unlike some virus-human interaction prediction studies, the P-HIPSter database provides virus-human protein interaction based on protein structure prediction. Through this data, our interaction network was able to uncover previously unnoticed protein interactions. Our finding revealed that HPV infection caused various gene expression differences, and a great amount of which interact with EGFR, a cancer related gene. Therefore, since EGFR is associated with HPV+ cancer patients’ survival, some FDA proved EGFR inhibitors would be potential anti-HPV drugs.


70
Tumor can be caused by several factors (1). A virus is a small pathogen that often cause 71 pathological changes or diseases in the target host (2). Some viral infections have been 72 linked to be essential factors that induce numerous forms of cancer such as liver cancer 73 and nasopharyngeal carcinoma (3,4). Virus lifecycle requires intracellular environment 74 owing to its simple structure (5, 6). It hijacks the cell's complex protein and nucleic acid 75 synthesis system for self-proliferation and also controls cells' functional protein to 76 modulate the normal cell signaling pathway (7). consequences of HPV infection (8,9). HPV preferably infect the mucosal layer and no 81 evidence shows that HPV has the ability to infect other cells except basal cells of the 82 epithelia. Basal cell is the high differential ability cell of the epithelia. Hence, host cell 83 development and differentiation ability are probably required for HPV infection (10). The 84 carcinogenesis of squamous cell carcinoma is often accompanied by changes in 85 development-related functions (11). Therefore, epithelia development regulated protein 86 may be the key target of HPV infection and oncogenesis. HPV genome encodes seven 87 early phase proteins (E1 to E7) and two late phase proteins (L1 and L2) for its 88 proliferation. E6 and E7 proteins can modulate p53 and Rb through downregulation or 89 inhibition, which is the basic mechanism of HPV+ cancer genesis (9). Therefore, E6 and 90 E7 can be regarded as the most essential HPV oncogenic proteins (12). Due to its small 91 genome and limited virus encoded protein, virus proteins require high efficiency and 92 5 multifunctionality for complicated manipulation. For example, evidences showed that E6 93 and E7 proteins can interact with many human proteins and participate in a lot of 94 biological processes (13). Likewise, HPV capsid protein L1 and L2 have been reported to 95 interact with human proteins (14).    The miRNA-Seq read counts data were also analyzed using PCA and differential 155 expression analysis. PCA distribution showed no obvious difference between HPV+ and 156 HPV-samples (Fig 1E), which indicates that there is no clear difference in miRNA 157 expression level between HPV+ cancer and HPV-cancer. Differentially express analysis 158 further showed that only 5 miRNAs were significantly differentially expressed. They 159 were hsa-miR-944, hsa-miR-196, hsa-miR-206, hsa-miR-10a and hsa-miR-548k ( Fig 1F-  number of TP53 mutation is significantly higher in HPV-cancer than in HPV+ cancer.

184
Whereas most of the genes with high mutation rate showed no significant difference in 185 both groups ( Fig 2E). This result suggests that TP53 mutation is an important mechanism 186 for the occurrence of HPV-. However, HPV+ cancer shows no relative involvement with 187 TP53 mutation, cancer occurrence may be involved in other mechanisms. subnetwork, we showed that degrees of tumor suppressor gene TP53 (regulated by HPV 205 E7 protein) is the highest of all genes ( Fig 3C). This result indicates that the regulation of 206 TP53 by HPV is a crucial mechanism for HPV+ tumor to occur. 207 208 EGFR is the crucial gene that regulate HPV+ tumor differentially expressed genes. 209 We used degree = 55 to screen hub nodes of miRNA-mRNA-protein network, and 22 tumor related genes ( Fig 4A). Remarkably, EGFR is the only hub node that belongs to 213 both DEGs and directly interacts with HPV ( Fig 4B). We further extracted a subnetwork 214 constructed with EGFR and its first neighbor. The result showed that EGFR interacts with 215 a considerable amount of DEGs and it is also regulated by hsa-miR-944 and HPV protein 216 E7 ( Fig 4C). This indicates that EGFR is an essential gene that regulates the differentially 217 expressed genes.

219
Furthermore, we predicted that EGFR is regulated by hsa-miR-944; and that the 220 upregulation of has-miR-944 caused the downregulation of EGFR (Fig 4B and 1J). In 221 order to confirm whether hsa-miR-944 combine stably with EGFR, RIsearch2 software 222 was used for RNA combined analysis. The result shows that EGFR and hsa-miR-944 223 have low energy binding site at 3' end of hsa-miR-944 ( Fig 4D). Our findings revealed 224 that EGFR is regulated by both E7 protein, and hsa-miR-944. This shows that E7 protein 225 does not only induces carcinogenesis in HPV+ tissues, but also causes the difference in 226 appearance in HPV+ and HPV-tumor. A module with EGFR was identified using module analysis of global network. Since gene 229 in the same module interact closely, there is possibility that they can participate in the 230 same biological process. HPV protein E2, a key protein that plays a pivotal role in HPV 231 infection from the early stage to the late stage, was also identified (Fig 4E). This finding 232 suggests that EGFR participates in all HPV infection stages and could probably influence Activation of EGFR-related pathway is an important factor that decrease survival. 240 In order to figure out how EGFR influence prognosis, we merged RPPA data of CESC 241 and HNSC and a total of 133 protein was obtained. Student-t test was used to test for the  representing EGFR was activated to form a dimer that binds with its ligand; thus, further 251 12 activating downstream pathways like PI3K/Akt, MAPK, WNT pathway. We also 252 highlighted that AR (amphiregulin), an EGFR ligand, and some significant proteins 253 belongs to PI3K/Akt or MAPK pathway. Those downstream proteins also showed similar 254 properties of phosphorylated form of EGFR that are significantly negatively correlated 255 with survival (Fig 5). shown and it could be correlated with prognosis. Remarkably, EGFR related pathways, 265 "PI3K/Akt pathway" and "ECM-receptor interaction", were also included in the top list 266 (Fig 6A). For DEmiR target enrichments, numerous of cancer related pathway were 267 shown and EGFR related pathway, "PI3K/Akt pathway", was also enriched (Fig 6B). the representative of data and limited data volume, we combined both CESC and HNSC 287 data into our study. Also, since our study focus on HPV-induced tumorigenesis rather 288 than ontogenesis, we merged the two cancer data for analysis instead of separate analysis.

298
The tumors that are not driven by virus usually shows mutations at oncogenes and/or 299 tumor suppressor genes. In the total of 84 HPV-regulated genes, only 4 genes (EGFR, 300 SNF, UBD and VCAM1) were differentially expressed when compared with HPV-tumor.

301
Among the 4 genes, SNF and VCAM1 directly interact with EGFR through string 302 estimation, this suggests that HPV-regulated DEG tends to interact, and that they 303 participate in similar biological processes that affects patient's survival. to that of HPV-tumors. Second, these samples were probably infected by low-risk HPV 315 that rarely induce cancer, which implies that their oncogenesis ability is relatively lower.

316
In spite of diversity in mRNA expression, miRNA expression between HPV+ and HPV-317 shows little differences. However, based on the enrichment analysis of DEmiRs target 318 genes, we showed that there are several differences in tumor-related pathways of HPV+ 319 and HPV-cancers. One possible reason for this is that only a small percentage of 320 15 miRNAs are differentially expressed in HPV+ and HPV-tumors. And these differentially 321 expressed miRNA are highly correlated with tumor-related biological processes.

323
Since certain mutations occur across different cancer types (25, 26), we therefore focus 324 on differences in mutation types between HPV+ and HPV-tumor. Our study showed that 325 TP53 mutation rate in HPV+ tumor is dramatically lower than in HPV-tumor. TP53 is an 326 important tumor suppressor gene. The mutation of tumor suppressor gene is considered 327 more serious than its dysfunction. HPV+ cancer patients showed better outcome than 328 HPV-cancer patients and that can probably be attribute to low TP53 mutation rate (9, 329 27). We also showed that genes that belong to the same family or participate in the same  Since EGFR is a potent oncogene, EGFR dysregulation will cause several forms of 353 cancer. High proportion of non-small cell lung carcinomas express EGFR and the EGFR 354 mutant as its signature (35,36). Likewise, EGFR has become a biomarker of HNSC (37).

355
In our study, EGFR shows different express pattern for HPV+ and HPV-cancers. EGFR      String database (https://string-db.org/). MCODE plugin of Cytoscape (version 3.7.0) was 444 used for network module identification. The parameters for MCODE were set as default.

445
The degree calculation was determined by NetworkAnalyzer, and genes with degree 446 higher than the threshold were defined as hub genes.