HTRA3 Is a Prognostic Biomarker and Associated With Immune Infiltrates in Gastric Cancer

HtrA serine peptidase 3 (HTRA3) participates in multiple signal pathways and plays an important regulatory role in various malignancies; however, its role on prognosis and immune infiltrates in gastric cancer (GC) remains unclear. The study investigated HTRA3 expression in tumor tissues and its association with immune infiltrates, and determined its prognostic roles in GC patients. Patients with GC were collected from the cancer genome atlas (TCGA). We compared the expression of HTRA3 in GC and normal gastric mucosa tissues with Wilcoxon rank sum test. And logistic regression was used to evaluate the relationship between HTRA3 and clinicopathological characters. Gene ontology (GO) term analysis, Gene set enrichment analysis (GSEA), and single-sample Gene Set Enrichment Analysis (ssGSEA) was conducted to explain the enrichmental pathways and functions and quantify the extent of immune cells infiltration for HTRA3. Kaplan-Meier analysis and Cox regression were performed to evaluate the correlation between HTRA3 and survival rates. A nomogram, based on Cox multivariate analysis, was used to predict the impact of HTRA3 on prognosis. High HTRA3 expression was significantly correlated with tumor histological type, histological grade, clinical stage, T stage, and TP53 status (P < 0.05). HTRA3-high GC patients had a lower 10-year progression-free interval [PFI; hazard ratio (HR): 1.46; 95% confidence interval (CI): 1.02–2.08; P = 0.038], disease-specific survival (DSS; HR: 1.65; CI: 1.08–2.52; P = 0.021) and overall survival (OS; HR: 1.59; CI: 1.14–2.22; P = 0.006). Multivariate survival analysis showed that HTRA3 was an independent prognostic marker for PFI (HR: 1.456; CI: 1.021–2.078; P = 0.038), DSS (HR: 1.650; CI: 1.079–2.522; P = 0.021) and OS [hazard ratio (HR): 1.590; 95% confidence interval (CI):1.140–2.219; P = 0.006]. The C-indexes and calibration plots of the nomogram based on multivariate analysis indicated an effective predictive performance for GC patients. GSEA showed that High HTRA3 expression may activate NF-κB pathway, YAP1/WWTR1/TAZ pathway, and TGFβ pathway. There was a negative correlation between the HTRA3 expression and the abundances of adaptive immunocytes (T helper cell 17 cells) and a positive correlation with abundances of innate immunocytes (natural killer cells, macrophages etc.). HTRA3 plays a vital role in GC progression and prognosis and could be a moderate biomarker for prediction for survival after gastrectomy.

HtrA serine peptidase 3 (HTRA3) participates in multiple signal pathways and plays an important regulatory role in various malignancies; however, its role on prognosis and immune infiltrates in gastric cancer (GC) remains unclear. The study investigated HTRA3 expression in tumor tissues and its association with immune infiltrates, and determined its prognostic roles in GC patients. Patients with GC were collected from the cancer genome atlas (TCGA). We compared the expression of HTRA3 in GC and normal gastric mucosa tissues with Wilcoxon rank sum test. And logistic regression was used to evaluate the relationship between HTRA3 and clinicopathological characters. Gene ontology (GO) term analysis, Gene set enrichment analysis (GSEA), and single-sample Gene Set Enrichment Analysis (ssGSEA) was conducted to explain the enrichmental pathways and functions and quantify the extent of immune cells infiltration for HTRA3. Kaplan-Meier analysis and Cox regression were performed to evaluate the correlation between HTRA3 and survival rates. A nomogram, based on Cox multivariate analysis, was used to predict the impact of HTRA3 on prognosis. High HTRA3 expression was significantly correlated with tumor histological type, histological grade, clinical stage, T stage, and TP53 status (P < 0.05). HTRA3-high GC patients had a lower 10-year progression-free interval [PFI; hazard ratio (HR): 1.46; 95% confidence interval (CI): 1.02-2.08; P = 0.038], disease-specific survival (DSS; HR: 1.65; CI: 1.08-2.52; P = 0.021) and overall survival (OS; HR: 1.59; CI: 1.14-2.22; P = 0.006). Multivariate survival analysis showed that HTRA3 was an independent prognostic marker for PFI (HR: 1.456; CI: 1.021-2.078; P = 0.038), DSS (HR: 1.650; CI: 1.079-2.522; P = 0.021) and OS [hazard ratio (HR): 1.590; 95% confidence interval (CI):1.140-2.219; P = 0.006]. The C-indexes and calibration plots of the nomogram based on multivariate analysis indicated an effective predictive performance for GC patients. GSEA showed that High HTRA3 expression may activate NF-kB pathway, YAP1/ WWTR1/TAZ pathway, and TGFb pathway. There was a negative correlation between the HTRA3 expression and the abundances of adaptive immunocytes (T helper cell 17 cells) and a positive correlation with abundances of innate immunocytes (natural killer cells,

INTRODUCTION
Gastric cancer (GC) is one of the most lethal carcinomas worldwide, and almost half of all GC cases are diagnosed in East Asia (1). Despite the developments in multimodal therapy strategies, such as surgery, chemotherapy, radiotherapy and immunotherapy, the five-year survival rates for the patients with stage II and stage IIIA GC were about 34 and 20%, respectively (2). GC is a heterogeneous malignancy (2,3) and the existing microsatellite instability, HER2 mutation, and amplification cannot completely explain the different prognosis or therapeutic response of GC (4,5). And the current biomarkers for the prognosis of GC are not suitable for clinical needs. Numerous studies currently on molecular targeted therapy and associated molecular pathways referred to the gastric tumorigenesis have illuminated the pathogenesis of GC and helped ameliorate the prognosis of patients with GC (6). For example, Trastuzumab can extend the survival time of the patients with HER2-positive GC (7). Therefore, there is an urgent necessity for the elucidation of the identification of novel biomarkers for carcinoma diagnosis and therapeutic targets in GC.
HtrA Serine Peptidase 3 (HTRA3), which was first reported in 2003, is a protein encoding gene located on chromosome 4p16.1 (8). HTRA3 protein (expressed in many cell types and organs) has two isomers, which both represent active serine proteases which cleave beta-casein/CSN2 as well as several extracellular matrix (ECM) proteoglycans (9). Previous studies indicated that HTRA3 inhibits signaling mediated by TGF-beta family proteins possibly indirectly by degradation of these ECM proteoglycans (10).
HTRA3 was first shown to be associated with cancer in a report demonstrating that it may act as a tumor suppressor, which is a proapoptotic protease that promotes drug-induced cytotoxic effects in lung cancer cells (11). Overexpressed HTRA3 inhibits the carcinogenic role of TGFb1 and thus inhibits metastasis in the early stages of non-small cell lung cancer (10). HTRA3 expression was negatively correlated with lymph node metastasis in breast cancer, but not with positive or negative expression of ER and PR (12). Recent studies illuminated it represents as a tumor promoter. The expression of HTRA3 in the peritumor stroma of patients with stage II colorectal cancer is associated with high-grade tumor budding, which may be a new marker of poor prognosis (13). HTRA3 may be related to the acquisition of invasive phenotype of and may be a potential prognostic indicator of oral cancer (14). These findings suggest that HTRA3 has multifaceted functional roles in various malignancies. Moreover, the underlying functions and mechanisms of HTRA3 in tumor progression and tumor immunology are still unclear.
In this study, we aimed to systematically analyze the significance of HTRA3 in GC using RNA sequencing data retrieved from TCGA database, along with bioinformatics and statistical methods including differentially expressed genes (DEG) analysis, gene ontology (GO) term analysis, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis, gene set enrichment analysis (GSEA), single-sample gene set enrichment analysis (ssGSEA), Kaplan-Meier survival analysis, and logistic & Cox regression analysis. Moreover, we further developed a nomogram to predict the patients' prognosis.

Data Source and Preprocessing
Gene expression data with clinical information from STAD projects (included 32 normal and 375 tumor tissues, Workflow Type: HTSeq-FPKM) were collected from TCGA. The exclusion criteria were normal STAD samples and an overall survival less than 30 days. Next, level 3 HTSeq-FPKM data were transformed into TPM (transcripts per million reads), and the TPM data of 375 GC patients were used for further analyses. Unavailable or unknown clinical features were regarded as missing values. The data are summarized in Supplemental Table 1.

HTRA3 Differential Expression in GC Tissues in the TCGA Database
Boxplots and scatter plots, using disease state (tumor or normal) as the variable, were generated to calculate differential expression of HTRA3. The diagnostic performance of HTRA3 was estimated using receiver operating characteristic (ROC) curves. Statistical ranking for HTRA3 expression above or below the median value was defined as HTRA3-high or HTRA3low, respectively.
Each sample (60 mg) was electrophoresed in 10% polyacrylamide gel and transferred to a polyvinylidene difluoride membrane (Millipore, Bedford, MA, USA) using a BG-blotMiMi transfer machine (Baygene, Beijing, China). After blocking with 5% nonfat milk for 2 h at room temperature (about 20 to 25°C.), overnight incubation with primary antibodies for HTRA3 (1:500 dilution; Abcom) or GAPDH (1:2,000 dilution; Santa Cruz) was performed at 4°C. The next day, after incubation with secondary antibodies (1:5000; Santa Cruz) at 37°C for 2 h, and after washing, the immunoreactive protein bands were visualized using an electrochemiluminescence (ECL) detection kit (Thermol Biotech, Rockford, IL, USA). The ratio between the optical density of the protein and GAPDH was calculated as the relative content of the protein detected. Each experiment was repeated three times.

Analysis of DEGs Between HTRA3-High and -Low Expression GC Groups
DEGs between HTRA3-high and HTRA3-low patients from TCGA datasets were identified by the unpaired Student's t-test, within the DESeq2 (3.8) package (15). Genes with the adjusted P value <0.05 and the absolute FC larger than 1.5 were considered to be statistically significant. All the DEGs were presented in a heat map and volcano plots.

Functional Enrichment and Analysis of Immune Cell Infiltration
In this study, Metascape (http://metasape.org) (16) was used as a tool to analyze the enrichment of HTRA3 related DEGs by process and pathway. The threshold conditions included: P <0.01, a minimum count of 3, and the enrichment factor >1.5 to obtain significant statistical differences.
GSEA starts with the HTRA3 differentially expressed matrix and analyzes the differences in signal pathways between the HTRA3-high and -low groups to predict the HTRA3-related phenotypes and signal pathways. A permutation test with 1,000 times was used to identify the significantly changed pathways. Adjusted P <0.01 and FDR <0.25 were identified as significant related genes. Statistical analysis and graphical plotting were conducted using R package clusterProfiler (3.8.0) (17). To construct the protein-protein interaction (PPI) network, the DEGs were input into STRING database (18). And PPI pairs with an interaction score >0.95 were chosen to build the PPI network.
The relative tumor infiltration levels of 24 immune cell types were quantified by ssGSEA to interrogate expression levels of genes in published signature gene lists (19). The signatures we used included a diverse set of adaptive and innate immune cell types and comprised 509 genes in total. To explore the correlation between HTRA3 and the infiltration levels of immune cells and the association of infiltration of immune cells with the different expression groups of HTRA3, Wilcoxon rank sum test, and Spearman correlation were adopted.

Clinical Statistical Analysis on Prognosis, Model Construction, and Evaluation
All statistical analyses were performed in R package (V3.6.2). The relationship between clinical pathologic features and HTRA3 were analyzed with the Wilcoxon signed-rank sum test and logistic regression. Clinicopathological characteristics associated with the 10-year overall survival (OS), progressionfree interval (PFI), and disease-specific survival (DSS) in TCGA patients using Cox regression and the Kaplan-Meier method. Multivariate Cox analysis was used to compare the influence of HTRA3 expression on survival along with other clinical characteristics (stage, myometrial invasion, lymph node status, distant metastasis status, histological grade and subtype). The cut-off value of HTRA3 expression was determined by its median value. P-values less than 0.05 were considered significant in all tests. The difference of 10-year OS, PFI, and DSS between HTRA3-high and -low group was calculated by the Kaplan-Meier method with a two-sided log-rank test.
Based on Cox regression models, the independent prognostic factors obtained from multivariate analysis were used to establish nomograms, individualizing the predicted survival probability for 1-, 3-, and 5-year. The RMS package (Version: 5.1-4; https:// cran.r-project.org/web/packages/rms/index.html) was employed to generate nomograms that included significant clinical characteristics and calibration plots. The calibration curves were graphically assessed by mapping the nomogram-predicted probabilities against the observed occurrences, and the 45°line represented the best predictive values. A concordance index (Cindex) was used to determine the discrimination of the nomogram, and it was calculated by a bootstrap approach with 1,000 resamples. The predictive accuracies of the nomogram and separate prognostic factors were compared using the C-index. All statistical tests were two tailed with a statistical significance level set at 0.05 in this study.  Table 1). However, there was no significant difference in the expression of HTRA3 in 27 GC samples and matched paracancerous samples (p = 0.117) ( Figure 1C), that may be due to the small number of paired samples in TCGA database. Therefore, we expanded the number of paired samples to verify mRNA and protein differential expression of HTRA3 in gastric cancer tissues and cell lines. We verified that HTRA3 is highly expressed in GC tissues by QPCR (P = 0.0035) and WB analysis (P = 0.0013) (Supplementary Figure 1A) on 44 paired GC samples and paracancerous samples. Among the GC cells, HTRA3 mRNA and protein expression was found at a comparatively higher level in MGC803, HGC27, and MKN45 cells and was lowest in SGC7901 (Supplementary Figure 1B).

Abnormally High Expression of HTRA3 in GC
ROC was used to analyze the distinguishing efficacy of HTRA3 between GC tissues and normal gastric mucosa tissue. The area under the curve (AUC) of HTRA3 is 0.710, suggesting  that HTRA3 may be a potentially moderate identification molecule for GC tissues ( Figure 1D).

Identification of DEGs in GC
We compared 187 GC HTRA3-high samples with 188 HTRA3low controls. A total of 607 DEGs, covering 299 upregulated GENEs and 308 downregulated GENEs, were identified to be statistically significant between the two cohorts (adjusted p-value < 0.05, |Log2-fold change| > 1.5) ( Figure 1E; Supplementary Table  2). Then, DEGs in HTSeq-Counts were further analyzed by DESeq2 package. Relative expression values of the top 10 DEGs between the two cohorts were showed in Figure 1F.

Functional Enrichment and Analyses of HTRA3 Related Genes in GC
In order to predict the functional enrichment information of HTRA3 interactive genes, we used Metascape for GO enrichment analysis, which showed that HTRA3-related genes were involved in many biological processes (BPs), cellular compositions (CCs), and molecular functions (MFs), including extracellular matrix organization, extracellular structure organization, transmembrane receptor protein serine/threonine kinase signaling pathway. Moreover, regulation of transmembrane receptor protein serine/ threonine kinase signaling pathway, extracellular matrix disassembly, regulation of extrinsic apoptotic signaling pathway via death domain receptors, and positive regulation of cell-substrate adhesion were also involved in the regulation of HTRA3 interactive genes ( Figure 2A; Supplementary Table 3).

Protein-Protein Interaction (PPI) Network Analysis
To obtain the interactions between the 607 DEGs in the GC group, a PPI network was constructed using the STRING database and the interaction threshold is set at 0.90. A total of 319 proteins and 1,425 edges were selected and 4 hub gene clusters were selected from PPI network with scores ≥6,600 ( Supplementary Figures 2A-E). Additionally, top 10 hub genes included IVL, TGM1, SPRR3, SPRR2A, SPRR2E, SPRR2D, LCE3E, SPRR2B, LCE3D, and SPRR2G.

GSEA Identifies HTRA3-Related Signaling Pathways
To identify HTRA3-related signaling pathways in GC, GSEA between HTRA3-high and -low expression data sets was conducted to reveal significant differences (adjusted P < 0.05, FDR q value < 0. 25 Table 2). Increased HTTA3 expression in GC is positively associated with T stage (OR = 2.39 for T3 and T4 vs. T1 and T2), histological grade (OR = 2.14 for G3 vs. G1 and G2); meanwhile negatively associated with TP53 status (OR = 0.59 for mutation vs. wild type) significantly (all P < 0.05). These results suggested that GCs with high HTRA3 expression were prone to progress to a more advanced stage and less susceptible to TP53 mutations than those with low HTRA3 expression.

High HTRA3 Expression Was Closely Associated With Poor Prognosis of Patients With GC
The 10-year OS rates were significantly higher among patients with low HTRA3 expression than those with high HTRA3 expression (37.6 vs. 24.7%; P = 0.006; Figure 5A). Similarly, the 10-year PFI rates and DSS in the HTRA3-low group were significantly higher than those in the HTRA3-high group (46.9 vs. 36.7%; P = 0.038; 57.3 vs. 38.2%; P = 0.021; Figures 5B, C). Next, we conducted subgroup survival analyses of OS, DSS, and PFI, which showed that the prognosis of patients with HTRA3-high was poor in stage III-IV, stage III, T4, N2&N3, and M0 subgroups of OS and M0 subgroup of DSS ( Figures 5D-I) Table 6) and short PFI (HR: 1.116; CI: 0.656-1.898; P = 0.686) (Supplemental Table 7) in patients with GC.
Based on the independent adverse prognostic factors chosen by multivariate Cox analysis, we studied the effect of   Table 8). However, the PFI prognostic analyses did not find that the difference in HTRA3 expression had a significant effect on the prognosis in any subgroup (Supplemental Table 9).

Construction and Validation of a Nomogram Based on the HTRA3
To provide a quantitative approach predicting the prognosis of GC patients, HTRA3 and independent clinical risk factors were used to construct a nomogram ( Figure 6A). In the nomogram based on multivariate Cox analysis, a point scale was used to assign points to these variables. The sum of points assigned to each variable was readjusted to a range from 1 to 100. The points of the variables were accumulated and recorded as the total scores. The probability of survival in GC patients at 1, 3, and 5 years was determined by drawing a vertical line directly down from the total point axis to the outcome axis. For instance, a GC patient with high HTRA3 risk (45 points), lymph node metastasis (42.5 points), and poor therapy outcome (37 points) received a total point score of 124.5. The probabilities of 1-, 3-, 5year survival were about 87.5, 57, and 38% ( Figure 6A). We also analyzed the prediction efficiency of the nomogram, and the result illuminated that the C-index of the model was 0.743(CI: 0.720-0.766), which suggested that the prediction efficiency of this model is moderately accurate. The biascorrected line in the calibration plot was utilized to be close to the ideal curve (the 45-degree line), which showed a fine

DISCUSSION
To our knowledge, the expression of HTRA3 and its potential prognostic impact on GC has not been explored. Hence, the potential role of HTRA3 in GC was the focus of the present study. In the present study, bioinformatics analysis using highthroughput RNA-sequence data from TCGA database demonstrated that there were evident individual variation and heterogeneity in these RNA transcripts, and HTRA3 might be a potential moderate marker for GC tissues. Increased expression of HTRA3 in GC was associated with fewer TP53 mutations, advanced clinical pathologic characteristics, shorter survival time, and poor prognosis. Human HTRA3 is a trimer multitasking serine protease related to cell function and pathogenicity. It may serve as a potential therapeutic target (9). HTRA3 dissolves XIAP and BAX, therefore promoting apoptosis. Additionally, it is associated with epithelial-mesenchymal transformation (EMT) (11). HTRA3 may inhibit the signal transduction of TGF-b family members and prohibit cell migration through MEK/ ERK pathway (20). HTRA3 may promote the instability of actin and vimentin cytoskeleton and affect the dynamics of cytoskeleton (21).
In some malignancies, HTRA3 has anti-tumor effect. Previous studies have elucidated the low expression of HTRA3 in endometrial carcinoma and ovarian cancer had anti-tumor effect (22,23). The ectopic expression of HTRA3 leaded to the decrease of cell proliferation and the increase of the expression of apoptotic protein Bax, suggesting that HTRA3 has anti-tumor effect on pancreatic cancer cells (24). The expression of HTRA3 in lung cancer tissue was significantly down-regulated. HTRA3 promoted drug-induced apoptosis of lung cancer cells (11). The expression of HTRA3 was negatively correlated with lymph node metastasis in breast cancer, but not with positive or negative expression of ER and PR (12). In non-small cell lung cancer cell lines, exogenous transforming growth factor b-1 significantly reduced HTRA3 expression during EMT (10). However, HTRA3 can promote tumor progression in other malignancies. recent studies have also shown that HTRA3 was highly expressed in thyroid carcinoma and related to the occurrence of thyroid cancer (25). HTRA3 was related to the   acquisition of invasive phenotype of oral squamous cell carcinoma and may be a potential prognostic indicator of oral cancer (14). HTRA3 is highly expressed in the stroma of the invasive front of colorectal cancer. The high expression of HTRA3 in tumor core was significantly correlated with the decrease of 5-year overall survival rate. In addition, the expression of HTRA3 in the peritumor stroma of patients with stage II colorectal cancer is related to high grade tumorous budding, which may be a new marker of poor prognosis (13).
To further investigate the functions of HTRA3 in GC, we performed GO, GESA, and ssGSEA analyses using TCGA data. The results revealed that NF-kB signaling pathway, YAP1/ WWTR1/TAZ pathway, TGFb signaling pathway, RAS/CA2P/ NMDA pathway, RHO/CIT pathway, and RAC1 signaling pathway were differentially enriched in the HTRA3-high phenotype and fewer abundance of adaptive immune cells were also observed. These data suggested that HTRA3 might serve as a potential prognostic marker and therapeutic target in GC.
HTRA3 might promote the instability of cytoskeleton, thereby regulating the EMT process of various tumor cells. And HTRA3 is a potential therapeutic target involved in various cancers (21). Researchers revealed that the modulation of HTRA3 in tumorigenesis might be dual, either inhibitory or promoting, depending on the specific tissues, stages of cancer progression etc. On one hand, HTRA3 expression is low in endometrial cancer, ovarian cancer, and lung cancer (23,26). HTRA3 promotes druginduced apoptosis through XIAP cleavage in lung cancer cells (11).
In non-small cell lung cancer (NSCLS) cell lines, overexpression of HTRA3 inhibits the carcinogenesis of TGF-b1, thus inhibiting tumor metastasis in the early stage of cancer. The role of HTRA3 is weakened, and transforming growth factor b1 effectively promoted EMT without HTRA3 brakes in the later stage of cancer (10). The ectopic expression of HTRA3 in pancreatic cancer leads to the decrease of cell proliferation and the increase of the expression of apoptotic protein Bax, suggesting that HTRA3 has anti-tumor effect on pancreatic cancer cells (24). On the other hand, HTRA3 is highly expressed in oral squamous cell carcinoma, thyroid cancer and colorectal cancer, and the acquisition of invasive phenotype in oral squamous cell carcinoma and colorectal tumor is closely related to poor prognosis (13,14,25). Here, we found that HTRA3 is highly expressed in GC. The overexpression of HTRA3 was significantly correlated with histological type, histological grade, clinical stage, T stage, and TP53 status of gastric cancer. HTRA3-high patients with GC had worse histological types (diffuse type), lower tumor differentiation, later clinicopathological stages (especially greater primary tumor growth), and were less prone to have TP53  Given the limited data on HTRA3 function, we performed functional annotation based on GO and GESA. We demonstrated that HTRA3-high phenotype was associated with NF-kB signaling pathway, YAP1/WWTR1/TAZ signaling pathway, TGFb pathway, RAS/CA2P/NMDA pathway, RHO/ CIT pathway, and RAC1 signaling pathway. Recent studies have elucidated that Enhanced activity of NF-kB signaling pathway can promote proliferation, metastasis, and angiogenesis of GC cells (27). YAP1/TAZ activity controls GC stem cells tumorigenic properties (28), indicating that HTRA3 may promote GC cell growth, metastasis, and poor survival via the NF-kB and YAP1/WWTR1/TAZ pathway. Additionally, the PPI network of HTRA3-associated genes was constructed in this study, and these genes were involved in various signaling pathways and biological processes. In future studies, we will further determine the correlation between HTRA3-associated genes and the prognosis of GC.
Stromal cells in tumor microenvironment can change the carcinogenic characteristics of tumor cells. Among them, tumorinfiltrating lymphocytes (TILs) play an important role in the occurrence and development of tumors (29,30). TILs establish a complex intercellular interaction network, which helps improve and maintain the immunosuppressive microenvironment, promote immune escape, and thus promote tumor progression (31). Our results demonstrated there were more NK cells, TAMs, and fewer Th17 cells in the HTRA3-high GC group than those in the HTRA3-low GC group, suggesting the improvement of innate immunity was accompanied by the decrease of adaptive immunity. Furthermore, infiltrating NK cells and TAMs in the tumor microenvironment have strong immunosuppressive activity, which reduces the secretion of IFN-g and induces T cell dysfunction (32,33). Additionally, the subgroup with more Th17 cell infiltration in breast and ovarian cancer is less likely to have lymph node metastasis and more likely to have a better prognosis, which suggests that Th17 cells have anti-tumor effect (34)(35)(36). Therefore, our data elucidated the immunosuppression induced by fewer Th17 cells in the primary tumor microenvironment might result in a lower 10-year survival rate in HTRA3-high patients with GC.
High level of HTRA3 expression was correlated with poor prognosis of GC in stage III -IV, T4, N2-3, and M0 subgroups, with the highest HR for poor OS, DSS, and PFI when HTRA3 was highly expressed in GC. We found that the expression of HTRA3 remained a powerful predictor of prognosis within these subsets, suggesting that HTRA3 was independent of these important clinicopathological parameters. Subsequently, a nomogram with comprehensive evaluation combining HTRA3 with other important clinical patterns (HTRA3 status, N stage, primary therapy outcome, age status, and tumor status) was performed. Based on the calibration plot, there was a favorable consistency between the actual and predicted values for 1-, 3-, 5-year OS. Our model was constructed based on the complementary perspective for respective tumors and provided a personalized score for individual patients. Consequently, our nomogram could be a valuable new prognostic method for clinicians in the future.
Although these results improved our understanding of the relationship between HTRA3 and GC, there were some limitations. First, to clarify the specific role of HTRA3 in the development of GC comprehensively, several clinical factors and parameters should be considered, such as the details on treatments received by patients involved. However, this information was lacked or inconsistent in public databases because the experiments were performed in different centers. Second, the number of healthy subjects used as controls was considerably different from that of patients with cancer in the current study, hence additional studies were required to maintain a balance of sample size. Third, although multicenter study in public databases intends to complement the drawbacks of single center study, retrospective studies have their limitations, especially non-uniform intervening measures, and lacking of some information. Therefore, a prospective study should be performed in the future to avoid analysis bias arising due to the retrospective nature of the current study. Lastly, since the current study was performed based on RNA sequencing from TCGA database only, it is necessary to further study the direct mechanism of HTRA3 in GC.
In this study, we firstly reported that the high expression of HTRA3 was significantly associated with the progression, poor survival, and immune infiltration of GC, which might promote tumorigenesis through abnormal inflammation and immune response. HTRA3 has the potential to predict treatment outcomes and may become a new biomarker of GC. The mechanism of HTRA3 promoting the progression and metastasis of GC will be verified in further studies. This study provided a new and promising insight for further elucidating the clinicopathological significance and molecular pathogenesis of GC.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.

ETHICS STATEMENT
Ethical review and approval were not required for the study on human participants in accordance with the local legislation and institutional requirements. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
B-CJ contributed to concept and design of this article. CJ participated in manuscript writing. L-SS, FX, and NN participated in data collection and data analysis. H-LG, J-WD, and NZ analyzed and interpreted data. All authors contributed to the article and approved the submitted version.