Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Identification of the novel prognostic biomarker SERPINH1 reveals its relationship with immunology in gastric cancer

Identification of the novel prognostic biomarker SERPINH1 reveals its relationship with... IntroductionIn 2020, there were 1,089,000 new cases of and 768,000 deaths from gastric cancer worldwide, ranking fifth in incidence and fourth in mortality globally [1]. GC is a multifactorial disease that, based on molecular and genomic studies, can be classified into Epstein-Barr virus-infected, genomically stable tumors, tumors with chromosomal instability, and MSI tumors [2]. Gastric cancer risk factors subsume smoking, alcohol consumption, chronic Helicobacter pylori infection, and eating preserved foods [3], [4], [5], [6]. In addition, a low intake of fruits and a high intake of processed and smoked meats may increase this risk [7], [8], [9].Despite recent developments in modern medical technologies, including, chemotherapy targeted therapy, and radiotherapy, the risk of death from GC remains high (13.6 %). Most gastric cancers are discovered at an advanced stage, making treatment extremely challenging. Therefore, regular diagnoses and early treatments may reduce the mortality rate of cancer. Recent studies have shown that lncRNAs affect the diagnosis, treatment, and prognosis of gastric cancer [10], [11], [12]. Targeted therapy, mRNA vaccines, and immunotherapy are potential feasible therapies for advanced gastric cancer [13].However, owing to regional, cultural, genetic, and other factors, achieving a complete cure of gastric cancer is challenging, with short-term curative effects causing financial burdens and failing to relieve patient pain. Therefore, effective biomarkers that can detect the disease at an early stage and lead to highly effective treatments are necessary. Many studies have been conducted on biomarkers related to gastric cancer, with microRNAs being used as non-invasive biomarkers [14]. CircRNA molecules rich in miRNA binding sites are also considered as promising biomarkers for gastric cancer owing to their various advantages [15]. Proteins such as TFF3, a secretory protein stably expressed in the gastrointestinal mucosa, have also been identified as effective biomarkers of gastric cancer [16]. In addition, abnormal DNA methylation affects pathways involved in cancer therapy and may act as a therapeutic biomarker for gastric cancer [17].SERPINH1 encodes a superfamily of serine protease inhibitors [18], [19], [20]. This protein is localized to the endoplasmic reticulum and functions as a collagen-specific molecular chaperone during collagen biosynthesis. In this study, we used GEO, TCGA, GTEx, and other databases to identify SERPINH1 as an effective biomarker for gastric cancer through screening, comparison, and experimental verification (Figure 1).Figure 1:Overview of the study.Overall, the development of highly effective biomarkers is crucial in the early identification and therapy of gastric cancer, and further research on promising candidates, such as SERPINH1, may lead to improved patient outcomes.Materials and methodsData from microarraysThree gene expression profiles (GSE54129, GSE79973, and GSE65801) were obtained from the Gene Expression Omnibus dataset. Among the 132 samples in the GSE54129 array, 111 gastric cancer and 21 normal samples were obtained. Twenty patient samples were collected from GSE79973, including ten malignant and ten normal gastric tissues. GSE65801 comprises 32 GC samples and 32 matched noncancerous tissues.Identification of genes with differential expressionGEO2R (https://www.ncbi.nlm.nih.gov/geo/geo2r/) was utilized to identify the differentially expressed genes (DEGs) between GC and noncancerous samples. DEGs were selected based on an adjusted |log FC| > 1 and p-value <0.05. After creating a Wayne diagram using Bioinformatics & Evolutionary Genomics (http://bioinformatics.psb.ugent.be/webtools/Venn/), the DEGs shared by all three datasets were selected for further analysis.Analysis of DEGs’ gene ontology and pathway enrichmentThis study utilized the Database for Annotation, Visualization, and Integrated Discovery (DAVID, https://david.ncifcrf.gov/summary.jsp) v6.8 owing to its wealth of biological information. Researchers may analyze the biological significance of several genes using the analytical tools and biological data this database provides, as well as complete functional annotation tools. Using p-value <0.05 as the cutoff criterion, gene ontology analysis and pathway enrichment of DEGs were performed using the DAVID online tool.PPI network construction and module selectionWe further analyzed the contacts among DEGs using the Search Tool for the Retrieval of Interacting Genes (STRING, http://string.embl.de/). Cytoscape software was used to visualize the integration results of the Protein-Protein Interaction networks. An important built-in plug-in in Cytoscape, Molecular Complex Detection (MCODE), can further analyzed to identify modules and hub genes. The criteria for identification were degree ≥12 and an MCODE score ≥11.Expression of the hub genes in mRNAGEPIA (http://gepia.cancer-pku.cn/detail.php?clicktag=degenes) is a multifunctional online analysis tool that can be directly matched to TCGA database. To increase the reliability of the analysis, TCGA was wielded to value the mRNA expression of hub genes in GC tissues and noncancerous tissues.Survival analysisAs an online analysis tool. The TCGA database’s sources include EGA, GEO, and TCGA, as well as 1440 GC samples. Based on the expression levels of specific genes, patients with GC were separated into two groups. Using the Kaplan-Meier (KM) plotter (http://kmplot.com/analysis/index.php?p=backgro und), overall survival (OS) was analyzed to produce a KM plot.RT-qPCRThe tissue samples were milled, and a suitable amount of Trizol (Carlsbad, California, USA) solution was added. After 10 min, trichloromethane (Xilong Chemical Co., Ltd., Guangdong, China) was added at a 5:1 ratio and mixed thoroughly. After standing for 2 min, the clear supernatant liquid was extracted via centrifuge, and an equal amount of isopropyl alcohol was added and cooled to −80 °C overnight. The precipitate was discarded, centrifuged, washed twice, and nuclease-free water (TIANGEN BIOTECH, Beijing, China) was added. A RevertAid First Strand cDNA Synthesis Kit (Waltham, Massachusetts, USA) was used to collect cDNA, and mRNA expression levels were detected by LightCycler® 96 Instrument (F. Hoffmann-La Roche AG, Basel, Switzerland).Immunoregulatory analysisThe datasets TCGA TARGET GTEx (PANCAN, n=19,131, G=60,499) and TCGA Pan-cancer (PANCAN, n=10,535, G=60,499) were obtained from the University of California Santa Cruz (UCSC, https://xenabrowser.net/) using Sangerbox (http://www.sangerbox.com/tool). The TCGA datasets were unified and standardized. The conditions were set in a Sanger box to screen 405 gastric cancer expression data points from the TCGA dataset. The differently expression of SERPINH1 in GC samples from various clinical stages were obtained. The TCGA TARGET GTEx dataset was filtered using a specific source in Sangerbox and processed as in the TCGA Pan-Cancer dataset, and 388 sample data points were obtained. Additionally, the gastric cancer gene expression profile was mapped to GeneSymbol, and the immune cell infiltration score of each patient with gastric cancer was reassessed based on their SERPINH1 gene expression. We extracted 60 genes (inhibitory (24) and stimulatory (36)) of the immune checkpoint pathway related to SERPINH1 from TCGA TARGET GTEx. Pearson’s correlations between SERPINH1 and five immune pathway marker genes were screened and calculated.The TIMER: Tumor Immune Estimation Resource gene module (https://cistrome.shinyapps.io/timer/) was applied to create a scatter plot showing the correlation between the expression of SERPINH1 and the degree of immune infiltration in GC. The SCNA module compares the extent of tumor infiltration according to various somatic copy number alterations in SERPINH1 Box plots show the distribution of individual immune subsets according to the copy number status within a given cancer type.Cell transfectionA liposome-mediated transfection of hSERPINH1-824 into AZ521 gastric cancer cells was performed. The experiment was divided into a control group, an NC control group (Sense: UUCUCCGAACGUGUCACGUTT; Antisense: ACGUGACACGUUCGGAGAATT), and a siSERPINH1 group (Sense: CCUCUACAACUACUACGACGATT; Antisense: UCGUCGUAGUAGUUGUAGAGGTT). Taking siSERPINH1 transfection as an example, the gastric cancer cell line AZ521 at a density of 1 × 106/mL was inoculated into a 6-well plate. When the cell confluency reached 50–70 %, the original medium was thrown away. After cleaning, the well was replenished with serum-free 1,640 medium (Shanghai VivaCell Biosciences Ltd, Shanghai, China) for 2 h. 125 μL OPTI-MEM (Thermo Fisher Scientific, Waltham, MA) containing at room temperature, LipofectmineTM3000 (5 μL) (Thermo Fisher Scientific, Waltham, MA) and 125 μL OPTI-MEM containing hSERPINH1-824 (5 μL, 20 μM) were combined, and incubated for 15 min. The mixture was transferred to a 6-well plate for 6 h. Replaced with the RPMI 1640 medium, the culture medium had been continuing for 24 h, after that it was used for subsequent detection.EdU assayAZ521 cells were inoculated in 12-well plates (5 × 106/mL). Each group was represented by three replicate wells. After 24 h of incubation, the culture medium was removed and EdU medium was added and then incubated for an additional 2 h. Next, the bottom of the hole was covered with 4 % paraformaldehyde (Lanjike Technology Co. LTD, Beijing, China) for 15 mins to fixed. After washing, 0.3 % Triton X-100 (Lanjike Technology Co. LTD, Beijing, China) was filled, and the mixture was incubated for 15 min and replaced with the reaction solution. The well plates were covered with tin foil and the mixture was incubated for 30 min. Finally, the nuclei were labeled with Hoechst 33,342 (Beyotime Institute of Biotechnology, Shanghai, China) staining solution for 10 min. An inverted fluorescence microscope was used to capture images, and EdU-positive cells were counted.Transwell assayCells in three group were set with three replicate wells and were starved for 12 h. We then added 500 μL 1,640 medium (10 % FBS) to the lower chamber of Transwell, while the upper chamber was replenished with 200 μL corresponding cell suspension (5 × 105/mL). The cells were cultured for 24 h before being fixed with 4 % paraformaldehyde. After 30 min, the cells were dyed with crystal violet (Beyotime Institute of Biotechnology, Shanghai, China) for 15 min. The remaining cells in the chamber were cleaned with a cotton swab. Each chamber was then placed under a microscope to capture images.ResultsIdentification of DEGsAfter normalization and differential expression analysis, according to the standards |log FC| > 1 and p-value <0.05, the GSE79973, GSE65801, and GSE54129 expression profiles showed 1,406, 3,894, and 2,722 differentially expressed genes, respectively. Among them, 360 genes were differentially expressed genes that overlapped in the three expression profiles (Figure 2A). GO terms and KEGG pathway enrichment analysis were performed using the online program DAVID, and we filtered the enrichment analysis findings with the standard p-value of <0.01 to further examine differentially expressed genes. According to the GO analysis, the changes in the molecular function (MF) of the differentially expressed genes were mainly focused on calcium ion binding, heparin binding, and extracellular matrix binding (Figure 2B). The changes of differentially expressed genes in cell components (CC) were mainly concentrated on the extracellular exosome, extracellular space, and proteinaceous extracellular matrix (Figure 2C). The changes in the differentially expressed genes in the biological process (BP) were mainly concentrated in cell adhesion, angiogenesis, and collagen fibril organization (Figure 2D). KEGG analysis results had shown the main enrichment in the PI3K-Akt signaling pathway, protein digestion, and ECM-receptor interaction (Figure 2E).Figure 2:Identification of DEGs and enrichment analysis of GO and KEGG pathways. (A) GSE54129, GSE79973, and GSE65801 obtained from the GEO database including 216 samples, among which 153 are of gastric cancer (GC) and 63 are of normal tissue. Using the adjusted |log FC| > 1 and p-value <0.05 as the standard screening, the expression profiles of the three genes were screened, and the overlapping genes of the three selected gene expression profiles were defined as differentially expressed genes. According to the identification criterion with p-value <0.01, significantly enriched GO terms were screened out among molecular functions (MF) (B), cell components (CC) (C), biological processes (BP) (D), and KEGG signaling pathways (E).PPI network construction and module analysisThe PPI network (Figure 3A) was comprised of 119 nodes and 267 edges, including 119 differentially expressed genes. With degrees ≥12 as the cutoff criterion, the top 19 genes were selected (Figure 3B). Then we selected a significant module with an MCODE score ≥11 and 14 genes were obtained (Figure 3C). Finally, 14 differentially expressed genes satisfying degrees ≥12 and MCODE score ≥11 were selected as hub genes for further analysis. We performed GO (Figure 3D) and KEGG (Figure 3E) pathway enrichment analyses for the 14 differentially expressed genes.Figure 3:PPI network and two modules. (A) PPI network of DEGs. (B–C) Genes identified by degree and MCODE score in the PPI network. (D) GO analysis. (E) KEGG pathway enrichment.Hub gene mRNA expression and survival analysisEight mRNA transcription group data sets, including GSE2685 (22 GC tissues and 8 para-carcinoma tissues), GSE13861 (65 primary gastric adenocarcinomas and 19 para-carcinoma tissues), GSE13911 (38 GC tumor and 31 para-carcinoma tissues), GSE26899 (96 GC tissues and 12 para-carcinoma tissues), GSE29998 (50 GC tumor and 49 para-carcinoma tissues), GSE33651 (40 gastric tumor and 12 para-carcinoma tissue), GSE51575 (26 pairs of neighboring non-tumor tissues and GC tumors) and GSE118916 (15 pairs of neighboring non-tumor tissues and GC tumors), were acquired from the GEO database. The above analysis results were verified by screening differentially expressed genes after GEO2R processing of the original data. The verification results showed that SERPINH1, COL1A1, COL4A1, and COL6A3 exhibited differential expression rates of up to 100 % (Table 1). Except for P4HA3, the remaining 13 differentially expressed genes had obviously higher expression levels in normal samples, according to the sample analysis results from TCGA and GTEx databases (Figure 4) Table 2.Table 1:Screening hub genes for mRNA expression, survival analysis, and validation results in eight data sets.Hub genesmRNA expressionSurvival analysis p-ValueConfirm rateSERPINH1a3.60E-11100 %COL4A1a5.50E-07100 %COL1A1a8.90E-05100 %COL6A3a0.0015100 %COL8A1a6.80E-0887.50 %COL1A2a0.001587.50 %COL12A1a0.00287.50 %COL10A1a0.01187.50 %COL5A1a4.30E-1475 %COL4A2a1.80E-1375 %COL3A1a9.30E-0575 %COL18A1a2.90E-1362.50 %arepresents a significant difference in mRNA expression between tumor and non-tumor samples.Table 2:Detailed data of gene validation results of 12 hubs in eight data sets.GSE2685GSE13861GSE13911GSE26899GSE29998GSE33651GSE51575GSE118916adj.p-ValueLog FCadj.p-ValueLog FCadj.p-ValueLog FCadj.p-ValueLog FCadj.p-ValueLog FCadj.p-ValueLog FCadj.p-ValueLog FCadj.p-ValueLog FCCOL1A11.95E-042.22.79E-072.23.22E-112.31.12E-042.21.38E-172.52.92E-062.21.06E-092.71.54E-032.2COL1A23.02E-052.38.63E-082.34.24E-142.62.55E-052.01.38E-172.48.56E-092.4––8.42E-041.9COL3A11.33E-031.61.90E-031.29.85E-071.1––1.41E-111.53.64E-082.2––2.12E-031.2COL4A13.18E-031.21.34E-061.45.43E-121.42.27E-051.52.34E-121.42.30E-133.21.53E-091.83.94E-072.0COL4A26.27E-051.53.52E-051.34.56E-101.13.35E-031.1––5.18E-132.2––2.66E-051.4COL5A1––––5.58E-071.24.85E-031.29.52E-091.06.25E-052.01.60E-031.13.65E-031.4COL6A33.27E-041.96.86E-071.66.56E-121.81.36E-031.61.03E-101.53.51E-092.87.07E-041.41.43E-052.5COL8A1––1.30E-082.31.41E-113.17.85E-042.11.49E-112.64.45E-072.54.67E-041.53.41E-062.6COL10A14.93E-021.85.36E-052.02.97E-083.65.12E-031.91.77E-104.9––2.05E-092.91.62E-021.6COL12A1––9.77E-061.41.35E-112.03.72E-041.53.72E-121.93.20E-031.65.65E-071.81.41E-042.2COL18A14.79E-041.2––––1.32E-031.21.19E-081.21.13E-061.3––1.29E-051.5SERPINH18.95E-041.28.04E-081.42.36E-183.61.33E-051.33.08E-151.45.10E-061.51.62E-111.72.00E-072.0Figure 4:Expression landscape of 14 hub genes across TCGA normal and GTEx data. The mRNA expressions of COL1A1 (A), COL1A2 (B), COL3A1 (C), COL4A1 (D), COL4A2 (E), COL5A1 (F), COL5A2 (G), COL6A3 (H), COL8A1 (I), COL10A1 (J), COL12A1 (K), COL18A1 (L), SERPINH1 (M) and P4HA3 (N) were all derived from the GEPIA online database.The survival analysis results obtained from the Kaplan-Meier Plotter suggested that, except for COL5A2 (p=0.18), the high expressions of the other 13 differentially expressed genes were closely associated with poor overall survival (Figure 5). Therefore, through comprehensive boxplot and survival analyses, we found 12 hub genes (Table 1) that were significant in the development and progression of GC.Figure 5:Prognostic values of 14 hub genes.RT-qPCR validationFour genes (SERPINH1, COL1A1, COL4A1, and COL6A3) were screened by eight profile comparisons and deemed differentially expressed in tumor and non-tumor samples. To validate this finding, we performed RT-qPCR to analyze the mRNA expression levels of these four genes in GC tissues. The expression levels of SERPINH1, COL1A1, COL4A1, and COL6A3 were upregulated (Figure 6).Figure 6:Five tumor and non-tumor samples selected for RT-qPCR verification. The expression levels of COL1A1 (A), COL4A1 (B), COL6A3 (C), and SERPINH1 (D) in tumor tissues increased compared with noncancerous tissues.Clinical staging and immunological evaluation of SERPINH1SERPINH1 is a precursor of COL1A1, COL4A1, and COL6A3. Therefore, we studied its relationship with gastric cancer. Analysis of TCGA data showed a significant connection between high SERPINH1 expression and the clinical stage of patients with GC (Figure 7A). TIMER database analysis revealed that SERPINH1 expression levels were negatively correlated with macrophages and B cells. Owing to the numerous tumor samples in the TCGA data, there was a large gap between the number of samples and that of the control group (Figure 7B). Therefore, we collected samples from the GTEx database and conducted relevant analyses of immune cells, immune checkpoints, and immune regulatory genes. The SERPINH1 expression was strongly linked to CD4+ T cells, B cells, and CD8+ T cells in patients with gastric cancer (Figure 7C). SCNA analysis demonstrated that the infiltration levels of CD4+ T cells, B cells, CD8+ T cells, dendritic cells, macrophages, and neutrophils were substantially linked to SERPINH1 expression (Figure 7D). The analysis of immune checkpoints suggested that the expression of SERPINH1 was linked to inhibitory immune checkpoints such as CD276, LAG3, TIGIT, IDO1, KIR2DL3, VEGFA, HAVCR2, ADORA2A, VEGFB, TNFRSF14, CD27, CD40LG, ICOS, PRF1, and GZMA. Significant correlations were observed between TNFRSF4, TNFRSF18, ITGB2, CD40, BTN3A1, CD70, and ICOSLG (Figure 7E). In the figures, we visualize the immunoregulation-related genes grouped according to SERPINH1 expression levels.Figure 7:Detailing the relationship between SERPINH1 and the development of gastric cancer. (A) SERPINH1 linked to a poor prognosis of gastric cancer. (B) Connection between SERPINH1 expression and the level of immune infiltration in gastric cancer. The X-axis shows the infiltration level, whereas the Y-axis shows the gene expression levels against tumor purity. (C) Eight types of immune cell infiltration scores in gastric cancer samples. The color represents the correlation coefficient. The asterisks denote statistically significant p-values. (D) The relationship between tumor immune cell infiltration and somatic copy-number. The samples were separated into four categories based on the copy number of SERPINH1, and the allocation of infiltrated immune cells in each group was compared. (E) The pearson connection between SERPINH1 and immune pathway marker genes.Effect of SERPINH1 on GC cellsTo demonstrate the effect of SERPINH1 on GC cells, a SERPINH1 knockdown transient cell line was constructed and the transient effect was illustrated using RT-qPCR (Figure 8A). The EdU experiment showed that the number of SERPINH1 knockdown cells that proliferated was significantly less than those of the control and NC groups, thereby demonstrating that the low expression of SERPINH1 inhibited the proliferation of gastric cancer cells (Figure 8B). Transwell assay results also revealed that the migration rate of normal GC cells, the SERPINH1 knockdown group’s migration rate was lower. This indicates that low expression of SERPINH1inhibited the migration capacity of GC cells (Figure 8C).Figure 8:SERPINH1 knockdown cell model used to verify the effect of SERPINH1 on GC. (A) RT-qPCR model. (B–C) EdU experiments proving that the knockdown of SERPINH1 inhibited the proliferation of gastric cancer cells. (D–E) Gastric cancer cell migration decreased by the SERPINH1 knockdown in a transwell assay.DiscussionGastric cancer therapy requires a comprehensive collaboration. Targeted therapy, immunotherapy, surgery, systemic chemotherapy, and radiotherapy are crucial in treating gastric cancer. Additionally, identifying ideal molecular markers may improve the direction of treatment. In recent years, with the continuous updating and iteration of omics technology, big data analysis along with experimental verification has offered a fresh perspective on the search for new and effective tumor-screening markers.The expression profiles of GSE54129, GSE79973, and GSE65801 in the GEO database were analyzed to identify 360 overlapping differentially expressed genes in this study. Constructed using STRING data and Cytoscape software, the PPI network was analyzed. Analysis of the other eight expression profiles in the GEO datasets revealed that SERPINH1, COL1A1, COL4A1, and COL6A3 were significantly differentially expressed in all samples. Additionally, RT-qPCR analysis of clinical tissue samples revealed that gastric cancer tissues had significantly higher transcription levels of SERPINH1, COL1A1, COL4A1, and COL6A3.COL1A1, COL4A1, and COL6A3 are members of the collagen superfamily, which consists of 28 members and is the most prevalent protein in mammals. One of the two forms of collagen I is COL1A1, a heterotrimer made of two α1 chains and one α2 chain [21]. The primary structural component of ECM is collagen Type I, and an aberrant ECM expression frequently coexists with the development of epithelial tumors [22, 23]. COL1A1 is essential for GC cell migration, invasion, and proliferation [22]. The migration, proliferation, and invasion abilities of GC cells with low COL1A1 expression were attenuated, thereby verifying the pro-tumor role of COL1A1 [24].COL4A1 is a type-IV collagen protein found in the basement membrane. COL4A1 abnormalities have been observed in familial nephropathy [25], cerebral palsy, chronic kidney disease, and myocardial infarction [26]. A bioinformatics analysis of GSE2685 showed that overexpression of COL4A1 led to a decreased overall survival rate, which was significantly correlated with T staging and GC recurrence [27]. COL6A3 is overexpressed regularly in GC cells, and most of the 62 genes in the list of potential partner genes have influenced cancer-related biological processes [28]. In previous studies, silencing of COL1A2, THBS2, and COL6A3 caused delayed migration and limited the proliferation of GC cells, whereas transfection promoted apoptosis [29].Previous studies have indicated that COL1A1, COL4A1, and COL6A3 are closely associated with tumor migration, invasion, and apoptosis. The collagen chaperone encoded by SERPINH1 can assist the collagen precursor in forming the correct spatial structure for further function. Because SERPINH1 acts as a precursor of COL1A1, COL4A1, and COL6A3, we selected SERPINH1 to further investigate its relationship with the development and prognosis of gastric cancer.SERPINH1 encodes the protein HSP-47 localized in the endoplasmic reticulum (ER) and is required for proper folding of the precursor collagen (procollagen) in the ER [30]. Therefore, SERPINH1 abnormalities are often present in various collagen-related disorders, including brittle bone disease [31], keloids [32], and pulmonary fibrosis [33], [34], [35]. SERPINH1 has also been implicated in the development of various tumors. Additionally, an overexpression of SERPINH1 has been observed in ccRCC [36], glioma [37], and nasopharyngeal carcinoma [19]. In nasopharyngeal carcinoma, SERPINH1 regulates the expression of many proteins, including BCAT1, BMI1, and BRD7, by inhibiting c-Myc ubiquitination-degradation, thus affecting the development of nasopharyngeal carcinoma [19]. In clear cell renal cell carcinoma (ccRCC), SERPINH1 may mediate the TGF-β-induced EMT process and affect the prognosis of ccRCC patients [36].In this study, using databases and related software, we found that SERPINH1 is closely related to gastric cancer. Sangerbox, TIMER, and other tools were used to analyze SERPINH1, and we observed that SERPINH1 promoted the progression of gastric cancer and has a greater impact on many cells related to immunity. In addition, we constructed a SERPINH1 knockdown cell model and corroborated its ability of SERPINH1 to inhibit the proliferation and migration of GC cells using EdU and Transwell assays; this indicated that SERPINH1 may be a potential target for the treatment of GC.In conclusion, SERPINH1 is a promising biomarker of GC. Further in-depth studies of SERPINH1 may provide new ideas for treating gastric cancer and new hope for patients with GC.ConclusionsIn conclusion, based on a comparative analysis of multiple databases and detection of clinical samples, we determined that SERPINH1 is highly expressed in gastric cancer. Combined with clinical stage information, patient prognoses, and protein functions, a SERPINH1 overexpression significantly promotes the pathological process and negatively impacts prognoses of gastric cancer. Therefore, SERPINH1 is a promising marker of gastric cancer. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png ONCOLOGIE de Gruyter

Identification of the novel prognostic biomarker SERPINH1 reveals its relationship with immunology in gastric cancer

ONCOLOGIE , Volume 25 (4): 13 – Jul 1, 2023

Loading next page...
 
/lp/de-gruyter/identification-of-the-novel-prognostic-biomarker-serpinh1-reveals-its-TlkPCKzZ0S

References (33)

Publisher
de Gruyter
Copyright
© 2023 the author(s), published by De Gruyter, Berlin/Boston
eISSN
1765-2839
DOI
10.1515/oncologie-2023-0017
Publisher site
See Article on Publisher Site

Abstract

IntroductionIn 2020, there were 1,089,000 new cases of and 768,000 deaths from gastric cancer worldwide, ranking fifth in incidence and fourth in mortality globally [1]. GC is a multifactorial disease that, based on molecular and genomic studies, can be classified into Epstein-Barr virus-infected, genomically stable tumors, tumors with chromosomal instability, and MSI tumors [2]. Gastric cancer risk factors subsume smoking, alcohol consumption, chronic Helicobacter pylori infection, and eating preserved foods [3], [4], [5], [6]. In addition, a low intake of fruits and a high intake of processed and smoked meats may increase this risk [7], [8], [9].Despite recent developments in modern medical technologies, including, chemotherapy targeted therapy, and radiotherapy, the risk of death from GC remains high (13.6 %). Most gastric cancers are discovered at an advanced stage, making treatment extremely challenging. Therefore, regular diagnoses and early treatments may reduce the mortality rate of cancer. Recent studies have shown that lncRNAs affect the diagnosis, treatment, and prognosis of gastric cancer [10], [11], [12]. Targeted therapy, mRNA vaccines, and immunotherapy are potential feasible therapies for advanced gastric cancer [13].However, owing to regional, cultural, genetic, and other factors, achieving a complete cure of gastric cancer is challenging, with short-term curative effects causing financial burdens and failing to relieve patient pain. Therefore, effective biomarkers that can detect the disease at an early stage and lead to highly effective treatments are necessary. Many studies have been conducted on biomarkers related to gastric cancer, with microRNAs being used as non-invasive biomarkers [14]. CircRNA molecules rich in miRNA binding sites are also considered as promising biomarkers for gastric cancer owing to their various advantages [15]. Proteins such as TFF3, a secretory protein stably expressed in the gastrointestinal mucosa, have also been identified as effective biomarkers of gastric cancer [16]. In addition, abnormal DNA methylation affects pathways involved in cancer therapy and may act as a therapeutic biomarker for gastric cancer [17].SERPINH1 encodes a superfamily of serine protease inhibitors [18], [19], [20]. This protein is localized to the endoplasmic reticulum and functions as a collagen-specific molecular chaperone during collagen biosynthesis. In this study, we used GEO, TCGA, GTEx, and other databases to identify SERPINH1 as an effective biomarker for gastric cancer through screening, comparison, and experimental verification (Figure 1).Figure 1:Overview of the study.Overall, the development of highly effective biomarkers is crucial in the early identification and therapy of gastric cancer, and further research on promising candidates, such as SERPINH1, may lead to improved patient outcomes.Materials and methodsData from microarraysThree gene expression profiles (GSE54129, GSE79973, and GSE65801) were obtained from the Gene Expression Omnibus dataset. Among the 132 samples in the GSE54129 array, 111 gastric cancer and 21 normal samples were obtained. Twenty patient samples were collected from GSE79973, including ten malignant and ten normal gastric tissues. GSE65801 comprises 32 GC samples and 32 matched noncancerous tissues.Identification of genes with differential expressionGEO2R (https://www.ncbi.nlm.nih.gov/geo/geo2r/) was utilized to identify the differentially expressed genes (DEGs) between GC and noncancerous samples. DEGs were selected based on an adjusted |log FC| > 1 and p-value <0.05. After creating a Wayne diagram using Bioinformatics & Evolutionary Genomics (http://bioinformatics.psb.ugent.be/webtools/Venn/), the DEGs shared by all three datasets were selected for further analysis.Analysis of DEGs’ gene ontology and pathway enrichmentThis study utilized the Database for Annotation, Visualization, and Integrated Discovery (DAVID, https://david.ncifcrf.gov/summary.jsp) v6.8 owing to its wealth of biological information. Researchers may analyze the biological significance of several genes using the analytical tools and biological data this database provides, as well as complete functional annotation tools. Using p-value <0.05 as the cutoff criterion, gene ontology analysis and pathway enrichment of DEGs were performed using the DAVID online tool.PPI network construction and module selectionWe further analyzed the contacts among DEGs using the Search Tool for the Retrieval of Interacting Genes (STRING, http://string.embl.de/). Cytoscape software was used to visualize the integration results of the Protein-Protein Interaction networks. An important built-in plug-in in Cytoscape, Molecular Complex Detection (MCODE), can further analyzed to identify modules and hub genes. The criteria for identification were degree ≥12 and an MCODE score ≥11.Expression of the hub genes in mRNAGEPIA (http://gepia.cancer-pku.cn/detail.php?clicktag=degenes) is a multifunctional online analysis tool that can be directly matched to TCGA database. To increase the reliability of the analysis, TCGA was wielded to value the mRNA expression of hub genes in GC tissues and noncancerous tissues.Survival analysisAs an online analysis tool. The TCGA database’s sources include EGA, GEO, and TCGA, as well as 1440 GC samples. Based on the expression levels of specific genes, patients with GC were separated into two groups. Using the Kaplan-Meier (KM) plotter (http://kmplot.com/analysis/index.php?p=backgro und), overall survival (OS) was analyzed to produce a KM plot.RT-qPCRThe tissue samples were milled, and a suitable amount of Trizol (Carlsbad, California, USA) solution was added. After 10 min, trichloromethane (Xilong Chemical Co., Ltd., Guangdong, China) was added at a 5:1 ratio and mixed thoroughly. After standing for 2 min, the clear supernatant liquid was extracted via centrifuge, and an equal amount of isopropyl alcohol was added and cooled to −80 °C overnight. The precipitate was discarded, centrifuged, washed twice, and nuclease-free water (TIANGEN BIOTECH, Beijing, China) was added. A RevertAid First Strand cDNA Synthesis Kit (Waltham, Massachusetts, USA) was used to collect cDNA, and mRNA expression levels were detected by LightCycler® 96 Instrument (F. Hoffmann-La Roche AG, Basel, Switzerland).Immunoregulatory analysisThe datasets TCGA TARGET GTEx (PANCAN, n=19,131, G=60,499) and TCGA Pan-cancer (PANCAN, n=10,535, G=60,499) were obtained from the University of California Santa Cruz (UCSC, https://xenabrowser.net/) using Sangerbox (http://www.sangerbox.com/tool). The TCGA datasets were unified and standardized. The conditions were set in a Sanger box to screen 405 gastric cancer expression data points from the TCGA dataset. The differently expression of SERPINH1 in GC samples from various clinical stages were obtained. The TCGA TARGET GTEx dataset was filtered using a specific source in Sangerbox and processed as in the TCGA Pan-Cancer dataset, and 388 sample data points were obtained. Additionally, the gastric cancer gene expression profile was mapped to GeneSymbol, and the immune cell infiltration score of each patient with gastric cancer was reassessed based on their SERPINH1 gene expression. We extracted 60 genes (inhibitory (24) and stimulatory (36)) of the immune checkpoint pathway related to SERPINH1 from TCGA TARGET GTEx. Pearson’s correlations between SERPINH1 and five immune pathway marker genes were screened and calculated.The TIMER: Tumor Immune Estimation Resource gene module (https://cistrome.shinyapps.io/timer/) was applied to create a scatter plot showing the correlation between the expression of SERPINH1 and the degree of immune infiltration in GC. The SCNA module compares the extent of tumor infiltration according to various somatic copy number alterations in SERPINH1 Box plots show the distribution of individual immune subsets according to the copy number status within a given cancer type.Cell transfectionA liposome-mediated transfection of hSERPINH1-824 into AZ521 gastric cancer cells was performed. The experiment was divided into a control group, an NC control group (Sense: UUCUCCGAACGUGUCACGUTT; Antisense: ACGUGACACGUUCGGAGAATT), and a siSERPINH1 group (Sense: CCUCUACAACUACUACGACGATT; Antisense: UCGUCGUAGUAGUUGUAGAGGTT). Taking siSERPINH1 transfection as an example, the gastric cancer cell line AZ521 at a density of 1 × 106/mL was inoculated into a 6-well plate. When the cell confluency reached 50–70 %, the original medium was thrown away. After cleaning, the well was replenished with serum-free 1,640 medium (Shanghai VivaCell Biosciences Ltd, Shanghai, China) for 2 h. 125 μL OPTI-MEM (Thermo Fisher Scientific, Waltham, MA) containing at room temperature, LipofectmineTM3000 (5 μL) (Thermo Fisher Scientific, Waltham, MA) and 125 μL OPTI-MEM containing hSERPINH1-824 (5 μL, 20 μM) were combined, and incubated for 15 min. The mixture was transferred to a 6-well plate for 6 h. Replaced with the RPMI 1640 medium, the culture medium had been continuing for 24 h, after that it was used for subsequent detection.EdU assayAZ521 cells were inoculated in 12-well plates (5 × 106/mL). Each group was represented by three replicate wells. After 24 h of incubation, the culture medium was removed and EdU medium was added and then incubated for an additional 2 h. Next, the bottom of the hole was covered with 4 % paraformaldehyde (Lanjike Technology Co. LTD, Beijing, China) for 15 mins to fixed. After washing, 0.3 % Triton X-100 (Lanjike Technology Co. LTD, Beijing, China) was filled, and the mixture was incubated for 15 min and replaced with the reaction solution. The well plates were covered with tin foil and the mixture was incubated for 30 min. Finally, the nuclei were labeled with Hoechst 33,342 (Beyotime Institute of Biotechnology, Shanghai, China) staining solution for 10 min. An inverted fluorescence microscope was used to capture images, and EdU-positive cells were counted.Transwell assayCells in three group were set with three replicate wells and were starved for 12 h. We then added 500 μL 1,640 medium (10 % FBS) to the lower chamber of Transwell, while the upper chamber was replenished with 200 μL corresponding cell suspension (5 × 105/mL). The cells were cultured for 24 h before being fixed with 4 % paraformaldehyde. After 30 min, the cells were dyed with crystal violet (Beyotime Institute of Biotechnology, Shanghai, China) for 15 min. The remaining cells in the chamber were cleaned with a cotton swab. Each chamber was then placed under a microscope to capture images.ResultsIdentification of DEGsAfter normalization and differential expression analysis, according to the standards |log FC| > 1 and p-value <0.05, the GSE79973, GSE65801, and GSE54129 expression profiles showed 1,406, 3,894, and 2,722 differentially expressed genes, respectively. Among them, 360 genes were differentially expressed genes that overlapped in the three expression profiles (Figure 2A). GO terms and KEGG pathway enrichment analysis were performed using the online program DAVID, and we filtered the enrichment analysis findings with the standard p-value of <0.01 to further examine differentially expressed genes. According to the GO analysis, the changes in the molecular function (MF) of the differentially expressed genes were mainly focused on calcium ion binding, heparin binding, and extracellular matrix binding (Figure 2B). The changes of differentially expressed genes in cell components (CC) were mainly concentrated on the extracellular exosome, extracellular space, and proteinaceous extracellular matrix (Figure 2C). The changes in the differentially expressed genes in the biological process (BP) were mainly concentrated in cell adhesion, angiogenesis, and collagen fibril organization (Figure 2D). KEGG analysis results had shown the main enrichment in the PI3K-Akt signaling pathway, protein digestion, and ECM-receptor interaction (Figure 2E).Figure 2:Identification of DEGs and enrichment analysis of GO and KEGG pathways. (A) GSE54129, GSE79973, and GSE65801 obtained from the GEO database including 216 samples, among which 153 are of gastric cancer (GC) and 63 are of normal tissue. Using the adjusted |log FC| > 1 and p-value <0.05 as the standard screening, the expression profiles of the three genes were screened, and the overlapping genes of the three selected gene expression profiles were defined as differentially expressed genes. According to the identification criterion with p-value <0.01, significantly enriched GO terms were screened out among molecular functions (MF) (B), cell components (CC) (C), biological processes (BP) (D), and KEGG signaling pathways (E).PPI network construction and module analysisThe PPI network (Figure 3A) was comprised of 119 nodes and 267 edges, including 119 differentially expressed genes. With degrees ≥12 as the cutoff criterion, the top 19 genes were selected (Figure 3B). Then we selected a significant module with an MCODE score ≥11 and 14 genes were obtained (Figure 3C). Finally, 14 differentially expressed genes satisfying degrees ≥12 and MCODE score ≥11 were selected as hub genes for further analysis. We performed GO (Figure 3D) and KEGG (Figure 3E) pathway enrichment analyses for the 14 differentially expressed genes.Figure 3:PPI network and two modules. (A) PPI network of DEGs. (B–C) Genes identified by degree and MCODE score in the PPI network. (D) GO analysis. (E) KEGG pathway enrichment.Hub gene mRNA expression and survival analysisEight mRNA transcription group data sets, including GSE2685 (22 GC tissues and 8 para-carcinoma tissues), GSE13861 (65 primary gastric adenocarcinomas and 19 para-carcinoma tissues), GSE13911 (38 GC tumor and 31 para-carcinoma tissues), GSE26899 (96 GC tissues and 12 para-carcinoma tissues), GSE29998 (50 GC tumor and 49 para-carcinoma tissues), GSE33651 (40 gastric tumor and 12 para-carcinoma tissue), GSE51575 (26 pairs of neighboring non-tumor tissues and GC tumors) and GSE118916 (15 pairs of neighboring non-tumor tissues and GC tumors), were acquired from the GEO database. The above analysis results were verified by screening differentially expressed genes after GEO2R processing of the original data. The verification results showed that SERPINH1, COL1A1, COL4A1, and COL6A3 exhibited differential expression rates of up to 100 % (Table 1). Except for P4HA3, the remaining 13 differentially expressed genes had obviously higher expression levels in normal samples, according to the sample analysis results from TCGA and GTEx databases (Figure 4) Table 2.Table 1:Screening hub genes for mRNA expression, survival analysis, and validation results in eight data sets.Hub genesmRNA expressionSurvival analysis p-ValueConfirm rateSERPINH1a3.60E-11100 %COL4A1a5.50E-07100 %COL1A1a8.90E-05100 %COL6A3a0.0015100 %COL8A1a6.80E-0887.50 %COL1A2a0.001587.50 %COL12A1a0.00287.50 %COL10A1a0.01187.50 %COL5A1a4.30E-1475 %COL4A2a1.80E-1375 %COL3A1a9.30E-0575 %COL18A1a2.90E-1362.50 %arepresents a significant difference in mRNA expression between tumor and non-tumor samples.Table 2:Detailed data of gene validation results of 12 hubs in eight data sets.GSE2685GSE13861GSE13911GSE26899GSE29998GSE33651GSE51575GSE118916adj.p-ValueLog FCadj.p-ValueLog FCadj.p-ValueLog FCadj.p-ValueLog FCadj.p-ValueLog FCadj.p-ValueLog FCadj.p-ValueLog FCadj.p-ValueLog FCCOL1A11.95E-042.22.79E-072.23.22E-112.31.12E-042.21.38E-172.52.92E-062.21.06E-092.71.54E-032.2COL1A23.02E-052.38.63E-082.34.24E-142.62.55E-052.01.38E-172.48.56E-092.4––8.42E-041.9COL3A11.33E-031.61.90E-031.29.85E-071.1––1.41E-111.53.64E-082.2––2.12E-031.2COL4A13.18E-031.21.34E-061.45.43E-121.42.27E-051.52.34E-121.42.30E-133.21.53E-091.83.94E-072.0COL4A26.27E-051.53.52E-051.34.56E-101.13.35E-031.1––5.18E-132.2––2.66E-051.4COL5A1––––5.58E-071.24.85E-031.29.52E-091.06.25E-052.01.60E-031.13.65E-031.4COL6A33.27E-041.96.86E-071.66.56E-121.81.36E-031.61.03E-101.53.51E-092.87.07E-041.41.43E-052.5COL8A1––1.30E-082.31.41E-113.17.85E-042.11.49E-112.64.45E-072.54.67E-041.53.41E-062.6COL10A14.93E-021.85.36E-052.02.97E-083.65.12E-031.91.77E-104.9––2.05E-092.91.62E-021.6COL12A1––9.77E-061.41.35E-112.03.72E-041.53.72E-121.93.20E-031.65.65E-071.81.41E-042.2COL18A14.79E-041.2––––1.32E-031.21.19E-081.21.13E-061.3––1.29E-051.5SERPINH18.95E-041.28.04E-081.42.36E-183.61.33E-051.33.08E-151.45.10E-061.51.62E-111.72.00E-072.0Figure 4:Expression landscape of 14 hub genes across TCGA normal and GTEx data. The mRNA expressions of COL1A1 (A), COL1A2 (B), COL3A1 (C), COL4A1 (D), COL4A2 (E), COL5A1 (F), COL5A2 (G), COL6A3 (H), COL8A1 (I), COL10A1 (J), COL12A1 (K), COL18A1 (L), SERPINH1 (M) and P4HA3 (N) were all derived from the GEPIA online database.The survival analysis results obtained from the Kaplan-Meier Plotter suggested that, except for COL5A2 (p=0.18), the high expressions of the other 13 differentially expressed genes were closely associated with poor overall survival (Figure 5). Therefore, through comprehensive boxplot and survival analyses, we found 12 hub genes (Table 1) that were significant in the development and progression of GC.Figure 5:Prognostic values of 14 hub genes.RT-qPCR validationFour genes (SERPINH1, COL1A1, COL4A1, and COL6A3) were screened by eight profile comparisons and deemed differentially expressed in tumor and non-tumor samples. To validate this finding, we performed RT-qPCR to analyze the mRNA expression levels of these four genes in GC tissues. The expression levels of SERPINH1, COL1A1, COL4A1, and COL6A3 were upregulated (Figure 6).Figure 6:Five tumor and non-tumor samples selected for RT-qPCR verification. The expression levels of COL1A1 (A), COL4A1 (B), COL6A3 (C), and SERPINH1 (D) in tumor tissues increased compared with noncancerous tissues.Clinical staging and immunological evaluation of SERPINH1SERPINH1 is a precursor of COL1A1, COL4A1, and COL6A3. Therefore, we studied its relationship with gastric cancer. Analysis of TCGA data showed a significant connection between high SERPINH1 expression and the clinical stage of patients with GC (Figure 7A). TIMER database analysis revealed that SERPINH1 expression levels were negatively correlated with macrophages and B cells. Owing to the numerous tumor samples in the TCGA data, there was a large gap between the number of samples and that of the control group (Figure 7B). Therefore, we collected samples from the GTEx database and conducted relevant analyses of immune cells, immune checkpoints, and immune regulatory genes. The SERPINH1 expression was strongly linked to CD4+ T cells, B cells, and CD8+ T cells in patients with gastric cancer (Figure 7C). SCNA analysis demonstrated that the infiltration levels of CD4+ T cells, B cells, CD8+ T cells, dendritic cells, macrophages, and neutrophils were substantially linked to SERPINH1 expression (Figure 7D). The analysis of immune checkpoints suggested that the expression of SERPINH1 was linked to inhibitory immune checkpoints such as CD276, LAG3, TIGIT, IDO1, KIR2DL3, VEGFA, HAVCR2, ADORA2A, VEGFB, TNFRSF14, CD27, CD40LG, ICOS, PRF1, and GZMA. Significant correlations were observed between TNFRSF4, TNFRSF18, ITGB2, CD40, BTN3A1, CD70, and ICOSLG (Figure 7E). In the figures, we visualize the immunoregulation-related genes grouped according to SERPINH1 expression levels.Figure 7:Detailing the relationship between SERPINH1 and the development of gastric cancer. (A) SERPINH1 linked to a poor prognosis of gastric cancer. (B) Connection between SERPINH1 expression and the level of immune infiltration in gastric cancer. The X-axis shows the infiltration level, whereas the Y-axis shows the gene expression levels against tumor purity. (C) Eight types of immune cell infiltration scores in gastric cancer samples. The color represents the correlation coefficient. The asterisks denote statistically significant p-values. (D) The relationship between tumor immune cell infiltration and somatic copy-number. The samples were separated into four categories based on the copy number of SERPINH1, and the allocation of infiltrated immune cells in each group was compared. (E) The pearson connection between SERPINH1 and immune pathway marker genes.Effect of SERPINH1 on GC cellsTo demonstrate the effect of SERPINH1 on GC cells, a SERPINH1 knockdown transient cell line was constructed and the transient effect was illustrated using RT-qPCR (Figure 8A). The EdU experiment showed that the number of SERPINH1 knockdown cells that proliferated was significantly less than those of the control and NC groups, thereby demonstrating that the low expression of SERPINH1 inhibited the proliferation of gastric cancer cells (Figure 8B). Transwell assay results also revealed that the migration rate of normal GC cells, the SERPINH1 knockdown group’s migration rate was lower. This indicates that low expression of SERPINH1inhibited the migration capacity of GC cells (Figure 8C).Figure 8:SERPINH1 knockdown cell model used to verify the effect of SERPINH1 on GC. (A) RT-qPCR model. (B–C) EdU experiments proving that the knockdown of SERPINH1 inhibited the proliferation of gastric cancer cells. (D–E) Gastric cancer cell migration decreased by the SERPINH1 knockdown in a transwell assay.DiscussionGastric cancer therapy requires a comprehensive collaboration. Targeted therapy, immunotherapy, surgery, systemic chemotherapy, and radiotherapy are crucial in treating gastric cancer. Additionally, identifying ideal molecular markers may improve the direction of treatment. In recent years, with the continuous updating and iteration of omics technology, big data analysis along with experimental verification has offered a fresh perspective on the search for new and effective tumor-screening markers.The expression profiles of GSE54129, GSE79973, and GSE65801 in the GEO database were analyzed to identify 360 overlapping differentially expressed genes in this study. Constructed using STRING data and Cytoscape software, the PPI network was analyzed. Analysis of the other eight expression profiles in the GEO datasets revealed that SERPINH1, COL1A1, COL4A1, and COL6A3 were significantly differentially expressed in all samples. Additionally, RT-qPCR analysis of clinical tissue samples revealed that gastric cancer tissues had significantly higher transcription levels of SERPINH1, COL1A1, COL4A1, and COL6A3.COL1A1, COL4A1, and COL6A3 are members of the collagen superfamily, which consists of 28 members and is the most prevalent protein in mammals. One of the two forms of collagen I is COL1A1, a heterotrimer made of two α1 chains and one α2 chain [21]. The primary structural component of ECM is collagen Type I, and an aberrant ECM expression frequently coexists with the development of epithelial tumors [22, 23]. COL1A1 is essential for GC cell migration, invasion, and proliferation [22]. The migration, proliferation, and invasion abilities of GC cells with low COL1A1 expression were attenuated, thereby verifying the pro-tumor role of COL1A1 [24].COL4A1 is a type-IV collagen protein found in the basement membrane. COL4A1 abnormalities have been observed in familial nephropathy [25], cerebral palsy, chronic kidney disease, and myocardial infarction [26]. A bioinformatics analysis of GSE2685 showed that overexpression of COL4A1 led to a decreased overall survival rate, which was significantly correlated with T staging and GC recurrence [27]. COL6A3 is overexpressed regularly in GC cells, and most of the 62 genes in the list of potential partner genes have influenced cancer-related biological processes [28]. In previous studies, silencing of COL1A2, THBS2, and COL6A3 caused delayed migration and limited the proliferation of GC cells, whereas transfection promoted apoptosis [29].Previous studies have indicated that COL1A1, COL4A1, and COL6A3 are closely associated with tumor migration, invasion, and apoptosis. The collagen chaperone encoded by SERPINH1 can assist the collagen precursor in forming the correct spatial structure for further function. Because SERPINH1 acts as a precursor of COL1A1, COL4A1, and COL6A3, we selected SERPINH1 to further investigate its relationship with the development and prognosis of gastric cancer.SERPINH1 encodes the protein HSP-47 localized in the endoplasmic reticulum (ER) and is required for proper folding of the precursor collagen (procollagen) in the ER [30]. Therefore, SERPINH1 abnormalities are often present in various collagen-related disorders, including brittle bone disease [31], keloids [32], and pulmonary fibrosis [33], [34], [35]. SERPINH1 has also been implicated in the development of various tumors. Additionally, an overexpression of SERPINH1 has been observed in ccRCC [36], glioma [37], and nasopharyngeal carcinoma [19]. In nasopharyngeal carcinoma, SERPINH1 regulates the expression of many proteins, including BCAT1, BMI1, and BRD7, by inhibiting c-Myc ubiquitination-degradation, thus affecting the development of nasopharyngeal carcinoma [19]. In clear cell renal cell carcinoma (ccRCC), SERPINH1 may mediate the TGF-β-induced EMT process and affect the prognosis of ccRCC patients [36].In this study, using databases and related software, we found that SERPINH1 is closely related to gastric cancer. Sangerbox, TIMER, and other tools were used to analyze SERPINH1, and we observed that SERPINH1 promoted the progression of gastric cancer and has a greater impact on many cells related to immunity. In addition, we constructed a SERPINH1 knockdown cell model and corroborated its ability of SERPINH1 to inhibit the proliferation and migration of GC cells using EdU and Transwell assays; this indicated that SERPINH1 may be a potential target for the treatment of GC.In conclusion, SERPINH1 is a promising biomarker of GC. Further in-depth studies of SERPINH1 may provide new ideas for treating gastric cancer and new hope for patients with GC.ConclusionsIn conclusion, based on a comparative analysis of multiple databases and detection of clinical samples, we determined that SERPINH1 is highly expressed in gastric cancer. Combined with clinical stage information, patient prognoses, and protein functions, a SERPINH1 overexpression significantly promotes the pathological process and negatively impacts prognoses of gastric cancer. Therefore, SERPINH1 is a promising marker of gastric cancer.

Journal

ONCOLOGIEde Gruyter

Published: Jul 1, 2023

Keywords: bioinformatics; gastric cancer; prognostic; survival analysis

There are no references for this article.