The Role of SRMS in Colorectal Cancer by Bioinformatics Analysis

Background:Src-related kinase lacking C-terminal regulatory tyrosine and N-terminal myristoylation sites (SRMS) is a non-receptor tyrosine kinase that has been found to be overexpressed in various tumors. Therefore, it may be an important carcinogenic factor whose role in colorectal cancer (CRC) has not been established. Methods: We evaluated the expression patterns of SRMS in CRC using GEPIA, Oncomine and HPA datasets while the association between SRMS and clinicopathological features was analyzed using a UALCAN dataset. LinkedOmics was used to determine co-expression and functional networks associated with SRMS. Besides, we used TISIDB to assess the correlation between SRMS and immune signatures, including tumor-inltrating immune cells and immunomodulators. Lastly, protein-protein interaction network establishment and pathway enrichment analysis of the SRMS-associated 33 immunomodulators and 191 immune cell marker genes were performed using the STRING portal. Results: Compared to normal colorectal tissues, SRMS was found to be upregulated in CRC tissues, and was correlated with higher pathological stages and nodal metastasis status. Functional network analysis suggested that SRMS regulates intermediate lament-based processes, protein autophosphorylation, translational initiation and elongation signaling through pathways involving ribosomes, proteasomes, oxidative phosphorylation, and DNA replication. In addition, SRMS expression was correlated with inltrating levels of CD4+ T cells, CD56dim, MEM B, Neutrophils, Th2, Th17, and Act DC. Pathway enrichment analysis of SRMS-associated 33 immunomodulators and 191 immune cell marker genes indicated that they are involved multiple cancer-related pathways. Conclusions: SRMS is a promising prognostic biomarker and a potential therapeutic target for CRC patients. In particular, SRMS regulates CRC progression by modulating chemokines, IL-17, intestinal immune networks for IgA production, and cytokine-cytokine receptor interaction signaling pathways among others. However, more studies are needed to validate these ndings.


Introduction
Globally, colorectal cancer (CRC) is one of the most prevalent malignancies and the fourth leading cause of cancer-related mortalities, resulting in almost 900,000 annual mortalities [1]. The high incidence of colorectal cancer is correlated with age, diet, race, lifestyle, genetic alteration and other factors. Since the early symptoms of colorectal cancer are not typical, about 25% of all colorectal cancers present distant metastasis at the time of initial diagnosis [2]. Currently, the major therapeutic options for CRC include surgery, chemotherapy, radiotherapy, biotherapy and immunotherapy. However, the 5-year survival outcomes for advanced CRC patients are approximately 10% [3,4]. Identi cation of genes associated with tumor formation and metastasis will inform the development of new therapeutic options for CRC.
Src-related kinase lacking C-terminal regulatory tyrosine and N-terminal myristoylation sites (SRMS) encode a 53 kDa non-receptor tyrosine kinase protein that was rst cloned in mouse neural precursor cells [5]. SRMS is composed of 488 amino acids, and is similar to Src family kinases. Its functional regions are a Src homology 3 (SH3) domain, a Src homology 2 (SH2) domain and a protein kinase domain [6]. These domains are important in mediating a series of intra-molecular or inter-molecular interactions as well as downstream signal transductions. The non-receptor tyrosine kinase plays a role in the regulation of cell growth, proliferation and invasion by activating the downstream substrate to initiate tyrosine phosphorylation [7]. Compared to other members of the non-receptor tyrosine kinase family, studies aimed at elucidating the roles and functions of SRMS are still in the early stages. A recent review article noted that SRMS was overexpressed in six breast cancer cell lines and its levels were elevated in breast tumors compared to adjacent normal tissues [8]. A proteomic study of gastric cancer patients showed that SRMS was the only differentially expressed kinase [9]. However, expression levels and biological functions of SRMS in CRC have not been clearly elucidated.
In this study, we evaluated and validated the expression levels of SRMS in CRC patients in multiple independent cohorts. Then, we systemically determined the relationship between SRMS expression and clinic-pathological features such as histological grade, metastasis and prognosis. In addition, we performed co-expression analysis and assessed the gene sets associated with SRMS in CRC through gene set enrichment analysis (GSEA). Lastly, we assessed the roles of SRMS in tumor immunity.

Materials And Methods
SRMS differential expression analysis. GEPIA (http://gepia.cancer-pku.cn/) is a database used to perform comprehensive and customizable functions using TCGA and GTEx data [10]. It includes 9736 tumor and 8587 normal samples. We used GEPIA to analyze the mRNA expression levels of SRMS in tumor and normal CRC tissue samples. The Y axis represents the average log2 abundance in transcripts-per-million (TPM). DNA copy number variations (CNV) of SRMS in CRC and normal tissues were examined using the Oncomine 4.5 database.
Oncomine (https://www.oncomine.org/) is the world's largest cancer microarray database and datamining platform [11]. It contains 715 datasets and 86,733 samples. Lastly, we compared SRMS protein expression levels between CRC tissues and paired normal colorectal tissues in the Human Protein Atlas (HPA) database. The HPA database (https://www.proteinatlas.org/) is aimed at mapping immunohistochemistry (IHC)-based protein expression pro les in cancerous and normal tissues as well as in cell lines [12]. UALCAN (http://ualcan.path.uab.edu/) is a comprehensive and interactive web-portal that provides an easy access to publicly available cancer OMICS data [13]. Moreover, it has gene expression pro les and patient survival information. In this study, SRMS expression levels in various sub-groups of clinical characteristics (age, gender, race, stages, histological and nodal metastasis status) were examined in UALCAN. p≤0.05 was considered statistically signi cant. LinkedOmics analysis.
LinkedOmics (http://www.linkedomics.org/ login. php) is a web-based platform that includes multiple omics data and proteomics data based on mass spectrometry (MS) generated by the CPTAC [14]. The "LinkFinder" module of LinkedOmics was used to evaluate gene co-expressions that are related to SRMS in the TCGA CRC cohort (n=379). Statistical analyses were done using Pearson's correlation coe cients and results were presented in form of heatmaps, scatter plots, and volcano plots. In the "LinkInterpreter" module of LinkedOmics, we also performed functional enrichment analysis of co-expressed genes. Data from LinkFinder were used to perform gene ontology (GO) and KEGG pathway enrichment analysis through GSEA. Rank criterion was an FDR<0.05 and 500 simulations. TISIDB analysis. TISIDB (http://cis.hku.hk/TISIDB/) is a web portal for tumor and immune system interactions, including genomics and transcriptomics of 30 cancer types from TCGA, RNA sequencing data set of patient cohorts treated with immunotherapy [15]. This online tool has "Function", "Literature", "Screening" "Immunotherapy", "Lymphocyte", "Immunomodulator", "Chemokine", "Subtype", "Clinical", and "Drug" modules. We used the TISIDB database to determine the correlations between SRMS expression and 28 tumor-in ltrating lymphocytes (TILs) and immunomodulators. STRING analysis. STRING (https://string-db.org/) is a database of known and predicted protein-protein interactions [16]. Currently, it has 24,584,628 proteins from 5,090 organisms. In this study, STRING was used to visualize protein-protein interaction networks and to predict the functions of the SRMS-associated 33 immunomodulators and 191 immune cell marker genes. GO and KEGG enrichment analysis was performed using Cluster-Pro ler package 3.14.0. Proteins with a minimum required interaction score greater than or equal to 0.400.

Result SRMS expression in CRC.
Initially, we compared mRNA expression levels of SRMS in GEPIA, which matched TCGA normal and GTEx data. mRNA expression levels of SRMS were signi cantly elevated in CRC tissues than in the paired adjacent normal tissues (p 0.05, Figure 1A). Moreover, from the HPA database, protein expression levels of SRMS were moderate in normal tissues and high elevation in tumor tissues ( Figure 1B). Data in the Oncomine database showed that CNVs of SRMS were signi cantly elevated in colon adenocarcinoma (COAD) tissues than in paired normal tissues (p 0.01, Figure 1C). Therefore, SRMS expression is a potential diagnostic indicator for CRC. SRMS expression in clinical characteristic sub-groups.
The association between SRMS and clinicopathological features was further evaluated in the online cancer OMICS database, UALCAN. In subgroup analyses based on age, gender, race, clinical stage, histological and nodal metastasis status, the transcription level of SRMS was signi cantly elevated in COAD patients than in healthy individuals (p<0.05, Figure 2A-F). Moreover, the higher the expression of SRMS, the higher the clinical stages and nodal metastasis. The expression of SRMS in stage 4 was higher than in stage 1 and 2 (p=0.0028, 0.015, respectively). The expression of SRMS in N2 was high than in N0 (p=0.048).These ndings implythat SRMS playsanimportantroleintumorprogression.
SRMS co-expression networks in CRC.
To elucidate on the functional properties of SRMS in CRC, the "LinkFinder" module in LinkedOmics was used to analyze the co-expression networks of SRMS. As shown in the volcano plot ( Figure 3A), a total of 2176 genes (red dots) were signi cantly positively correlated with SRMS while 1368 genes (green dots) were signi cantly negatively correlated (p<0.05). The heatmaps show the top 50 genes that were positively and negatively correlated with SRMS ( Figure 3B-C). SRMS expression was positively correlated with the expression of C20orf195 (positive rank #1, r = 0.389, p = 3.90E-15), HES2 (r = 0.301, p = 2.20E-9), PTK6(r = 0.300, p = 3.04E-9) and SYNGR3 (r = 0.300, p = 2.43E-9) among others. This result suggests that SRMS has various effects on the transcriptome.
Gene Ontology term and KEGG pathway by GSEA showed that co-expressed genes were mainly involved in the intermediate lament-based processes, epidermis development and protein autophosphorylation ( Figure 3D). In contrast, translational initiation and elongation, mitochondrial gene expression, and cytoplasmic translation were inhibited. KEGG pathway analysis showed enrichment in inositol phosphate metabolism, ribosomes, proteasomes, oxidative phosphorylation and DNA replication among others.
Immune signatures associated with the SRMS gene.
The SRMS-associated 33 immunomodulators and 191 immune cell marker genes were analyzed in the STRING database to validate functional connectivity. From the STRING database, we obtained the SRMSassociated immune gene PPI network (enrichment p-value<1.0e-16) of a total of 214 nodes and 829 edges, which represented proteins and functional interactions ( Figure 5B). Next, we performed an enrichment analysis of these genes used GO and KEGG ( gure5C). Consequently, we probed the signaling pathways through which SRMS regulates immune responses in CRC. Figure 5D shows that these genes were mainly enriched in chemokine, cancer, IL-17, in ammatory bowel disease (IBD), intestinal immune network for IgA production, and in cytokine-cytokine receptor interaction signaling pathways. These signaling pathways may related to SRMS-mediated immune events.

Discussion
Compared to paired normal tissues, SRMS was highly expressed in CRC tissues. A higher SRMS expression level was associated with late TNM stages, more lymph node metastasis, and poor prognosis.
Moreover, SRMS expression was correlated with many immune cells and immunostimulators, and these immune signatures were focused on in ammation and cancer signaling pathways.
Tumor occurrence is closely associated with cell proliferation, differentiation, and apoptosis. Non-receptor tyrosine kinases, including ptk6 (protein tyrosine kinase 6), fak (focal adhesion kinase), and Jak (janus kinase) are involved in cell differentiation, apoptosis, and proliferation through their interactions with transmembrane receptors [17]. PTK6 is the most studied member of the non-receptor tyrosine kinase family. Important roles of PTK6 in various cancers, including breast, prostate and colon cancers have been reviewed [18]. SRMS and PTK6 genes are closely linked on the chromosome. Additionally, SRMS has been shown to biochemically interact with PTK6, which phosphorylates the C-terminus of PTK6 [19]. It has been reported that PTK6 expression is highest in normal colon epithelial tissues and decreases during colon tumor progression [20]. Contrastingly, we found that proteins, CNVs, and mRNA expression levels of SRMS were signi cantly elevated in CRC than in normal colorectal tissues.
Copy number variants play a signi cant role in genetic variations and evolution, and can also cause genetic diseases and cancer [21]. We also found that different histologic stages and nodal metastasis status were associated with SRMS expression. Elevated SRMS expression was associated with advanced tumor stages, indicating that it may be involved in CRC progression. From this hypothesis, we evaluated the association between SRMS expression levels and survival using the UALCAN database.
Elevated SRMS expression levels were correlated with poor prognostic outcomes. Therefore, SRMS in CRC should be clinically validated as a possible diagnostic or prognostic marker.
Within the gene co-expression network of SRMS, we identi ed genes that were signi cantly positively correlated, such as C20orf195, HES2, PTK6 and SYNGR3 among others. Hes genes are Notch downstream target genes, and could re ect expression levels of Notch signals [22]. Aberrant activation of Notch signaling has been associated with various tumors, such as breast and colorectal cancers [23,24]. As mentioned earlier, PTK6 and SRMS have a close relationship. PTK6 is mainly expressed in epithelial tissues, with highest levels in gastrointestinal linings [25]. In colon cancer cell line models, PTK6 was shown to suppress tumorigenesis by inhibiting EMT [20]. Co-expression functional networks were found to be mainly involved in protein autophosphorylation, translational initiation and elongation, mitochondrial gene expression, as well as cytoplasmic translation. Thus, the SRMS co-expression network is involved in post-transcriptional regulation, which is closely associated with protein translation and phosphorylation. KEGG pathway analysis showed enrichment in the ribosomal, proteasomal, oxidative phosphorylation and DNA replication pathways among others. Ribosomes are comprised of ribonucleoprotein and non-coding ribosomal RNAs in eukaryotes and are conserved molecular structures required for protein synthesis [26]. Cancer development and progression are associated with ribosomal dysregulation, which affects the expression of key factors involved in tumorigenesis [27]. It has been reported that a single ribosomal assembly factor promotes LUAD progression through the Notch signaling pathway[28]. One of the major mechanisms of protein degradation is the proteasome pathway with 26S proteasomes being the most active isoforms that are involved in cell cycle progression, apoptosis and transcription [29]. Therefore, they are potential targets for anticancer therapy. Suppressed oxidative phosphorylation is a basic feature of tumor cells and tumors [30]. Huang et al. showed that LYRM2 directly interacts with complex I and enhances its activity, thereby promoting oxidative phosphorylation to induce colorectal cancer cell growth [31]. Since coexpressed genes share functions and affect each other, SRMS may play a role in CRC occurrence and progression through the above factors.
In the TISIDB database, we found that the Reactome pathway of SRMS was mainly involved in the immune system. Moreover, SRMS expression levels were positively correlated with in ltration levels of CD56dim, neutrophils, Th17 and negatively correlated with Act CD4, MEM B, Tem CD4, as well as Th2 in CRC. Tumor in ltrating immune cell patterns are associated with CRC initiation and prognosis [32][33][34]. In addition we identi ed the immunomodulators that were signi cantly associated with SRMS in CRC.
Tumor cells utilize various immune escape mechanisms, creating a microenvironment that is favorable for tumor growth and metastasis [35]. Immunomodulators have the potential for cancer treatment. KEGG pathway analysis of SRMS-associated immunomodulators and immune cell marker genes revealed that the chemokine pathway, IL17 and intestinal immune networks for IgA production may be involved in SRMS-mediated immune responses. It has been reported that the CXCL5 chemokine enhances the migratory and invasive properties of colorectal cancer cells by inducing epithelial-mesenchymal transition [36]. Chemokine signaling systems play critical roles in either the promotion or inhibition of tumor growth, proliferation, angiogenesis, or metastasis [37]. Aberrated activation of intestinal immune networks for IgA production signaling pathway promotes tumorigenesis [38]. IL-17 promotes tumor development through chronic tissue in ammation signals. For many cancers, Th17-cell signatures (RORC, IL17, IL23, STAT3) are correlated worse clinical outcomes [39]. It has been reported that IL-17, produced in response to microbial colonization in the intestines of mice, promotes CRC initiation and progression [40]. Therefore, it is biologically plausible that SRMS promotes tumor immunity by regulating multiple signaling pathways. This study provides the rst evidence for the link between SRMS and tumor immunity, opening up more avenues for research. This study has several limitations. First, the data we analyzed were obtained from several public datasets, which lack validation data. Second, the mechanisms through which SRMS-mediates tumor immunity were not fully evaluated. More clari cation and basic data are required to better assess the potential relationships between SRMS and CRC.

Conclusions
In summary, SRMS is over-expressed in CRC and is associated with adverse clinical outcomes. Moreover, SRMS expression levels were signi cantly correlated with various immune signatures. These ndings indicate that SRMS might play a role in the control of tumor immune microenvironments. However, more in vivo and in vitro experiments are required to verify these ndings.