ACE2 Expression and its Prognostic Signicance in Head and Neck Squamous Cell Carcinoma

Background and objective: Angiotensin-converting enzyme 2 (ACE2), a membrane structural glycoprotein that acts as a key receptor in the process of SARS-CoV-2 infection, has been identied as an oncogene in some tumor types. However, few studies have explored the role of the ACE2 gene in head and neck squamous cell carcinoma (HNSCC). The purpose of this study was to investigate the potential relationship between ACE2 and HNSCC and explore early markers and molecular targets for the treatment of HNSCC. Methods: Integrative bioinformatics analyses were applied to uncover the potential role of ACE2 in HNSCC development and tumor-associated immunology. Results: The results showed that ACE2 was highly expressed in HNSCC and signicantly correlated with clinical features such as sex. In addition, ACE2 may be a potential prognostic marker for HNSCC, as it was correlated with shorter recurrence-free survival (RFS) according to the Kaplan-Meier method. The PPI network revealed that STAT1 is the gene most closely related to ACE2 and that the NOD-like receptor signaling pathway was the most relevant pathway. Moreover, ACE2 expression was closely associated with the immune-inltrating levels of CD8 + T cells, myeloid dendritic cells, and neutrophils. Conclusions: The viral entry molecule ACE2 plays an important role in the tumorigenesis and cancer-immune interactions of HNSCC, suggesting that it is a novel molecular target and a new immune checkpoint in the diagnosis and treatment of HNSCC.


Introduction
As a homolog of the angiotensin-converting enzyme (ACE), ACE2 can negatively regulate the activated renin angiotensin system by degrading Ang II to heptapeptide angiotensin 1-7 (1). Some studies have shown that the RAS system is involved in the important pathological process of head and neck squamous cell carcinoma. For example, Rho/Ras CO activation, PLC epsilon-ca2 + signal transduction, and Raf/ERK are necessary for the development of head and neck squamous cell carcinoma (2). In terms of invasion and metastasis, Ang can promote the invasion and migration of head and neck squamous cell carcinoma (HNSCC) cells (3). In drug resistance, activation of RAS signaling leads to continuous extracellular signal-regulated kinase 1/2 signaling, resulting in a resistance to cetuximab (4). In immune in ltration, KRAS mutations in the RAS family increase TGF-1 levels in head and neck squamous cell carcinoma patients, while anti-tumor drugs used for its treatment need to have anticancer effects that overcome the anti-tumor immunosuppression induced by TGF-1 (5). However, the role of the novel coronavirus pneumonia receptor ACE2 is still undetermined in HNSCC.
In general, squamous cell carcinoma of the head and neck is a malignant tumor covered by squamous epithelium that occurs in the mouth, pharynx, nose, and throat (6). In terms of human malignant tumors, HNSCC has the sixth highest incidence, and its mortality rate has remained high (7). Despite the rapid development of medical technology in recent decades, the ve-year survival rate of patients with HNSCC has not been greatly improved (8). Since head and neck squamous cell carcinoma is highly invasive and complex to treat, it is urgent to nd early markers and molecular targets for the treatment of HNSCC (9).
Previous studies have shown that NOTCH1, TP53, TP63, and CDKN2A are the malignant tumor genes of HNSCC, and that dysregulation of these genes is the driving factor of HNSCC cancer (10)(11)(12)(13). However, few studies have explored the role of the ACE2 gene in HNSCC. Based on the corresponding bioinformatics websites, we found that ACE2 has the potential to be a molecular biomarker in HNSCC based on its expression level, methylation level, survival analysis, and corresponding immune invasion results. To nd out whether ACE2 can play a crucial role in the occurrence, development, and evolution of HNSCC, as has been demonstrated previously by various HNSCC malignant genes, experiments that are more adequate are still necessary.

Materials And Methods
Use of different public biology websites for data collection and reanalysis The relevant ACE2 bioinformatics data obtained from different bioinformatics network resources are summarized in Table S1(Table S1).
As a database for tumor-related gene research, the Oncomine database integrates RNA and DNA SEQ data from GEO, TCGA, and published literature, and has become the largest cancer gene chip database and integrated data mining platform in the world (14). UALCAN is an effective online cancer data analysis and mining website, which is mainly based on the relevant cancer data in the TCGA database, such as biomarker identi cation, expression pro le analysis, survival analysis, and so on (15). CancerRNA-Seq Nexus, also known as the CRN database, is a website for the direct analysis of tumor transcriptome data.
Its main data sources are the geo database and TCGA database, including 40 types of tumors, 89 tumor sequencing datasets, 325 phenotypic sets and 12167 samples (16). Through the above public bioinformatics platform, we can understand the expression level of ACE2 in human HNSCC. The Kaplan Meier plotter is a tool used to evaluate the effect of genes on the survival rates of cancer samples. It provides tools such as overall survival rates and post-progression survival rates, which can help to evaluate the prognosis of the disease (17). MethSurv is a methylation biomarker that is primarily used to explore the survival rates of cancer patients. It includes 7,358 methylated data points from 25 different human cancers as well as their corresponding survival analyses (18). cBioPortal is a multi-dimensional data network resource platform for cancer genomes, which can visually analyze gene changes in cancer research samples and explore the relationship between gene changes and clinical practice (19).
Therefore, we used it to screen the ACE2 co-expression genes in human HNSCC. These co-expressed genes were put into the string database to obtain the protein-protein interaction (PPI) network of these coexpressed genes (20). Then, a detailed visual analysis was carried out using Cytoscape software (version 3.7.2) and the Kobas website (21,22). Next, we used a web-based gene set analysis kit (Webgestalt) to perform a gene ontology (GO) enrichment analysis and a KEGG pathway analysis (23). At the same time, a pathview web was used to organize the analyzed data into graphs (24).
Tumor IMmune Estimation Resource (TIMER)is a component analysis software for tumor-in ltrating immune cells. It can associate tumor-in ltrating immune cells with their relevant gene expression, gene mutation, somatic copy number variation, and other data, and can predict the immune cell composition of each tumor sample (25).

Statistical analysis
The difference of mRNA expression between cancer and non cancer tissues was analyzed by Student t test. Chi square test and a generalized linear model analysis were used to analyze the relationship between the expression of ACE2 and clinicopathological characteristics of head and neck squamous cell carcinoma. If P < 0.05, the results are considered to be statistically signi cant. All the above methods were calculated by SPSS (SPSS 23.0, IBM Analytics).

ACE2 was highly expressed in HNSCC
We obtained and analyzed the expression of ACE2 in human HNSCC from the data from the different databases. First, two HNSCC groups named Toruner Head-Neck and Ye Head-Neck were downloaded from the Oncomine public bioinformatics website. After data processing, we found that the expression of ACE2 was upregulated in two HNSCC groups compared to normal tissues, and there was a signi cant difference found (Fig. 1A, B). In addition, from the TCGA database, we found that the expression of ACE2 in the HPV-negative group GSE40774 was signi cantly increased. These data indicate that ACE2 and HPV infection may opposite effects in HNSCC, and their speci c effects and mechanisms still require further experimental study (Fig. 1C). Finally, we further veri ed the expression level of ACE2 in HNSCC from the UALCAN database, which proved that the expression of ACE2 was upregulated in HNSCC ( Fig. 1D). When examining the methylation level of ACE2 in the tissues of patients with HNSCC compared to normal tissues, we can clearly see from the UALCAN that ACE2 methylation levels were downregulated in grade 1-3 HNSCC ( Fig. 2A). However, in the N stage, the methylation levels of ACE2 in patients with stage N0-2 HNSCC were signi cantly different from the levels seen in healthy patients, and the differences were statistically signi cant (P < 0.05) (Fig. 2B). In addition, the expression of ACE2 mRNA in HNSCC was further determined using the cancer RNA-seq Nexus (CRN) database (Table 1).

Correlation Between Ace2 Expression And Clinical Features Of Hnscc
In order to explore the correlation between ACE2 expression and the clinicopathological characteristics of HNSCC patients, we downloaded the relevant clinical data of HNSCC patients from the TCGA database.
Following the data analysis and statistical analysis, we found that high expression of ACE2 in HNSCC was signi cantly correlated with gender, N stage, and HNSCC grade (P = 0.008, P = 0.014, P = 0.034, respectively) ( Table 2). Based on the univariate analysis, the clinical features with P < 0.24 were included in the multivariate analysis. The results showed that high expression level of ACE2 was only related to gender, but not to N stage, M stage, and HNSCC grade (Table 3). In summary, we have reason to believe that ACE2 may play a role in the occurrence and development of HNSCC.  (Fig. 3A, 3B, and 3C). Finally, relevant data were downloaded from the Methsurv database, and it was found that the high expression of ACE2 was correlated with the short survival time of patients with HNSCC (P < 0.05) (Fig. 3D and 3E). Based on the above ndings, we concluded that ACE2 can be used as a potential biomarker for the prognosis of patients with HNSCC.
Analysis of the co-expression network and enrichment pathway of the ACE2 gene We used the cBioportal database to analyze the biological function of ACE2. First, all co-expressed genes were downloaded from the database, and then these genes were screened with an absolute value of correlation 0.33, and P < 0.05; 217 differentially expressed genes (Table 3S) were obtained. The PPI network was constructed using the string public information website, and then the corresponding coexpression network map was constructed using Cytoscape software. In the network diagram, we can see that STAT1 is the gene most closely related to ACE2 (Fig. 4A). At the same time, the KEGG enrichment pathway of these co-expressed genes was studied using the KOBAS biological website (Table 4S). We found that the NOD-like receptor signaling pathway was the most relevant pathway (Fig. 4B). Finally, we used the WebGestalt website to carry out a GO biological analysis on the 217 different co-expressed genes and identi ed the molecular functions, cellular components and the biological processes of ACE2 biology. The main processes were found to be protein binding, membranes, and biological regulation, respectively (Fig. 4C).
Expression of the ACE2 gene is associated with immune in ltration in HNSCC TIMER provides high-throughput data on immune cell in ltration for the study of ACE2 in relation to immune responses in the tumor microenvironment (TME). In the data on immune invasion in HNSCC, we can clearly obtain the following results: in B cells, CD8 + T cells, myeloid dendritic cells, and neutrophils, the expression of ACE2 was positively correlated with the level of immune cell in ltration (P < 0.05) (Fig. 5). We found that among the invasive immune cell signal markers of HNSCC, M1 macrophage cell immune gene signal markers such as PTGS2 (r = 0.101, P = 2.51e-02) and NOS2 (r = 0.092, P = 4.11e-02) were correlated with the expression of ACE2 in HNSCC. Meanwhile, other immune cell signal markers were not correlated with the expression level of ACE2 in HNSCC (Fig. 6). These results suggest that ACE2 expression plays a potentially important role in the immune in ltration of HNSCC.

Discussion
In this study, we explored the potential relationship between angiotensin converting enzyme 2 (ACE2) and the development of head and neck squamous cell carcinoma (HNSCC) in humans. On the one hand, we used public datasets to analyze the expression level of ACE2 in HNSCC. On the other hand, we constructed a co-expression network of ACE2 and screened several important co-expressed genes and signaling pathways that may signi cant for the tumor's progression. Then, the GEO and TCGA databases were analyzed. According to the results, ACE2 expression was signi cantly upregulated in HNSCC. In addition, ACE2 signi cantly affects the HNSCC survival time and tumor-associated immune response.
Angiotensin converting enzyme (ACE2) is a structural membrane glycoprotein that participates in the renin-angiotensin reaction; it also an important member of the renin-angiotensin system (RAS). It is known during SARS-CoV-2 infection, the v2irus invades a key receptor in the body. After entering the cell, the virus can induce heterogeneous protein synthesis and the release of in ammatory factors, promoting vasoconstriction, an in ammatory response, hypertension, oxidation, and brosis, resulting in multiple organ dysfunction (26). The entry of the coronavirus into the cells depends on the binding of viral spike proteins to the cell receptor and the initiation of S proteins by host cell proteases. Hoffmann et al. (27)demonstrated that SARS-CoV-2 uses ACE2 as a SARS-CoV receptor to enter cells, and the virus also uses the serine protease TMPRSS2 for S protein priming. Sacconi et al. (28) found that the expression of TMPRSS2 in HNSCC cells was signi cantly lower than in normal tissues, while ACE2 was slightly upregulated in female patients; however, the overall expression levels of ACE2 were comparable to normal tissues. ACE2 hydrolyzes angiotensin I (Ang I) to produce angiotensin II (Ang II). It has been proven that Ang II can promote the invasion and migration of HNSCC cells in an autocrine manner or by triggering stromal tumor-paracrine interactions (3). Narayan et al. (29) performed RT-PCR and a protein analysis on thyroid cancer and normal tissues, respectively, using ACE-and ACE2 speci c primers or antibodies. The results showed that the expression of ACE2 in thyroid carcinoma was signi cantly increased, and the higher the degree of differentiation, the higher the ratio of ACE2/ACE. In addition, Carlos et al. (30)found that ACE2 is signi cantly associated with epithelial malignancies and can be used as a therapeutic target for malignant epithelial tumors, especially oral squamous cell carcinoma (OSCC).
Through a previous bioinformatics analysis, it was found that ACE2 mainly plays a role in regulating STAT1, and also has a role in bioregulatory and metabolic pathways. STAT1 is an important component of the IFNγ/ STAT1 signaling pathway, which is involved in many cell life activities such as cell growth inhibition and apoptosis promotion. Jiang et al. (31) showed that myotubularin-related protein 2 (MTMR2) can promote the invasion and metastasis of gastric cancer cells by inhibiting the IFNγ/ STAT1 pathway. It has also been shown that the IFNγ/STAT1 pathway promotes tumor cell survival and induces an adaptive immune resistance via CD4 + T cell loss and PD-L1 upregulation (32). Ryan et al. (33)found that in mouse HNSCC, STAT1 could promote a T cell immune response and inhibit myeloid-derived suppressor cell aggregation, thus mediating an anti-tumor immune response. Aldo keto reductase family 1 member C1 (AKR1C1) was positively correlated with cisplatin resistance and was a poor prognostic factor for HNSCC. A transcriptomic analysis by Chang et al. (34)showed that STAT1 and STAT3 could activate an AKR1C1-induced cisplatin resistance, which could be overcome by treatment with ruxolitinib. Metabolic pathways play an important role in regulating tumorigenesis and the development of many tumors. Tumor cells possess new metabolic pathways that enable them to increase the uptake e ciency of nutrients through metabolic reprogramming, in order to meet their requirements for growth and invasion (35). Sur et al. (36) found that oral cancer cells induced the generation of mitochondrial reactive oxygen species (ROS) and inhibited cell apoptosis by altering glycolysis and lipid metabolism pathways. Common reprogramming metabolic pathways include the lKB1-AMP kinase (AMPK) signaling pathways(37). Chen et al. (38)showed that nuclear AMPK recruits PMK2 and β-Catenin by interacting with them, which plays an important role in promoting the cell migration of thyroid cancer. In addition, metabolic pathways are also involved in maintaining tumor stemness. Liu et al. (39)demonstrated that activation of the HSP27/hK2 pathway caused cancer stem cells (CSCs) to exhibit reprogrammed metabolic features and enhanced stem cell phenotypes, such as increased ALDH activity, chemoresistance, and tumor formation. However, the speci c relationship between ACE2 and metabolic pathways in tumorigenesis, as well as its role in the pathogenesis, development, and prognosis of HNSCC remains to be further explored.
The tumor microenvironment (TME) and tumor-related immune responses have always been the focus of tumor research, and they have also directed new therapy regimens (40). HNSCC cells can evade the host's immune system by manipulating their own immunogenicity, producing immunosuppressive mediators, and promoting the creation of immunoregulatory cell types (41). Mandal et al. (42) comprehensively described the immune landscape of HNSCC using the transcriptome data of 280 tumor pro les depicted by The Cancer Genome Atlas (TCGA), and found that both HPV + and HPV-HNSCC tumors were among the most highly immune-in ltrated cancer types. HNSCC has a high level of Treg/CD8 + T cell and NK cell in ltration, which is statistically correlated with its prognosis. ACE2 is an important molecule that participates in TME regulation. Zhang et al. (43) used TCGA to explore the relationship between pancancer ACE2 expression and several factors including anti-tumor immunity, immunotherapy responses, the carcinogenic pathway, tumor progression phenotypes, and clinical outcomes. It was found that ACE2 upregulation was associated with increased antitumor immune signatures and increased PD-L1 expression, as well as favorable anti-PD-1/PD-L1/CTLA-4 immunotherapy responses. ACE2 may provide a reference for prognosis following tumor immunotherapy. Previous studies have shown that the expression of ACE2 is negatively correlated with immune cell in ltrates, such as neutrophils and macrophages (44,45). Cheng et al. (46) showed that overexpression of ACE2 could inhibit the synthesis of vascular endothelial growth factor in TME and it inhibited tumor invasion and in ammation. However, by using the GSE30589 database, Yang et al. (47)found that the expression level of ACE2 was positively correlated with the level of immune in ltration of macrophages, B cells, CD4 + T cells, neutrophils, and dendritic cells in uterine corpus endometrial carcinoma (UCEC). We also found that the expression of ACE2 was positively correlated with the level of immune in ltration of B cells, CD8 + T cells, neutrophils, and dendritic cells. Therefore, the mechanism by which ACE2 participates in the HNSCC immune response warrants further study.

Conclusions
We can conclude that ACE2 expression plays an important role in the tumor immune response in HNSCC, suggesting that ACE2 can be a novel molecular target and a new immune checkpoint in tumor immune escape and tolerance.
ACE2 is a potential biomarker of HNSCC and is involved in the formation of the TME. However, whether ACE2 promotes or inhibits the immune escape of HNSCC tumor cells remains unknown. Further analysis  The methylation level of ACE2 in HNSCC in different grades and stages A: The association between the methylation level of ACE2 and the grades of HNSCC was analyzed in UALCAN database. (Among them, * represents signi cant difference, i.e, P < 0.05). B :The association between the methylation level of ACE2 and the N stages of HNSCC was analyzed in UALCAN database. (Among them, * represents signi cant difference, i.e, P < 0.05). Please refer to Table S2(Table S2) for detailed data. Analysis of ACE2 co-expression network A: Using cBioPortal database, string public information website website and Cytoscape software, the network of co-expression gene with ACE2 was constructed. B: The KEGG enrichment pathway of co-expressed genes was analyzed by KOBAS biological website, and bubble diagram was drawn by R language package. C: The analysis of GO biological processes, molecular functions and cell components is from WebGestalt website.