Upregulation of PTGFRN in Hepatocellular Carcinoma Predicts Poor Prognosis: A Study Based on the TCGA Database

Background. The importance of prostaglandin F2 receptor inhibitors (PTGFRN) in the progression of a variety of malignant tumors has been recognized in recent years. So far, no role of PTGFRN in hepatocellular carcinoma (HCC) has been reported. In this study, we focused on the possible mechanisms of PTGFRN in HCC based on the Cancer Genome Atlas (TCGA) data. Methods. The mRNA gene expression data of PTGFRN were downloaded from TCGA database to analyze the expression level of PTGFRN in HCC. According to the human protein atlas database, the expression difference of PTGFRN protein between HCC and adjacent tissues was veried. Wilcoxon signed-rank test and logistic regression were used to analyze the relationship between PTGFRN and clinicopathological characteristics. Kaplan Meier method and Cox proportional hazards model were used to explore the prognostic role of PTGFRN in HCC. The ROC curve was used to evaluate the diagnostic value of PTGFRN in HCC. Gene Set Enrichment Analysis (GSEA) was used to investigate the function of PTGFRN related Gene sets. Finally, obtain the co-expressed genes of PTGFRN through the cBioPortal database, and use the Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) function enrichment analysis to further explore the role of PTGFRN in HCC regulated related pathways. up-regulated grade involved in extracellular matrix tissue, epithelial-mesenchymal transition, cell adhesion and cell cycle, and PI3K-Akt/NF-kB signaling pathways.


Introduction
Hepatocellular carcinoma (HCC) is one of the most lethal malignant tumors in the world and the second leading cause of cancer-related death [1,2]. And Guangxi Province in China is one of the high-incidence areas of HCC [3]. Due to the high heterogeneity and complex etiology of HCC, it is di cult to predict the prognosis of tumors [4].
In recent years, progress has been made in the early diagnosis and treatment of HCC, which has effectively improved the prognosis of patients [5]. However, due to the high recurrence rate and high metastasis rate, the survival of HCC patients is still facing challenges [6,7]. Traditional prognostic models, such as tumor lymph node metastasis (TNM) stage and vascular in ltration, are not ideal for prediction due to the extreme heterogeneity of HCC [8,9]. In addition, biomarkers and histological classi cation, such as CD34 and AFP, etc., are widely used clinically to evaluate the prognosis of HCC patients. But the prognostic effect is limited by sample size, validation bias and other factors [1,10]. Thus, new speci c molecular markers are urgently needed to improve the prognosis and treatment of patients.
With the rapid development of genome sequencing technology, more and more evidence show that molecular markers have a great contribution to the diagnosis and prognosis prediction of HCC [8,11]. Prostaglandin F2 Receptor Inhibitor (PTGFRN, also known as CD9P-1, FPRP, EWIF) is a cell surface immune globulin (Ig) superfamily protein that has an effect on angiogenesis and tumor growth in vivo [12]. Its Ig domains include CD101, EWI-2 and IgSF3 [13]. Moreover, PTGFRN can act as a scaffold protein to bind to other proteins in TEM and regulate downstream signaling pathways [14]. Up to now, more and more evidence has shown that PTGFRN is signi cantly associated with a variety of cancer types and has different expressions and functions in different tumors. For example, Colin et al. [12] found that the expression level of PTGFRN is positively correlated with lung cancer metastasis. Guilmain et al. [15] found that the up-regulation of PTGFRN expression is closely related to the poor prognosis of patients with renal cell carcinoma. Brittany Aguila et al. [14,16] showed that inhibiting the expression of PTGFRN gene can signi cantly down-regulate the PI3-K/AKT signaling pathway, which affects the survival of solid tumor cells. S Colin et al. [12] found that the high expression of PTGFRN could negatively affect the growth and metastasis of solid tumors. In addition, inhibition of PTGFRN can reduce DNA damage perception and repair, increasing the sensitivity of glioblastoma multiforme (GBM) cells to radiotherapy and chemotherapy [14]. In conclusion, we can see that PTGFRN is related to tumor metastasis, important signaling pathways and so on. However, there is no report about PTGFRN in HCC, and its diagnostic effect and molecular mechanism remain unclear.
In addition, multiple family members of PTGFRN have also played an important role in the development of HCC. Ricardo et al. found that the increased expression and activity of prostaglandin reductase 1 (PTGR1) was often closely related to the progression of HCC [17]. Min Zhang et al. found that the expression level of prostaglandin E2 (PGE2) in HCC cells was signi cantly higher than that in normal liver cells. PGE2 can activate the EGFR/Akt/NF-κB pathway, promote Snail expression, and thus accelerate the migration and invasion of HCC [18]. Ricard et al. found that high level of 8-EPI-prostaglandin F2α (8-EPI-PGF2α) was positively associated with the risk of HCC and could be used as a biomarker to diagnose people at high risk of HCC [19]. However, the potential role of PTGFRN in HCC remains to be clari ed.
In this study, we conducted the rst bioinformatics analysis of PTGFRN expression in HCC, and discussed the correlation between PTGFRN expression and clinical features and prognosis. We also evaluated the effect of PTGFRN on overall and relapse-free survival. GSEA was used to further understand the regulatory network of PTGFRN involved in biological pathways of disease in HCC. PTGFRN co-expressed genes were screened by cBioPortal database, and they were analyzed by KEGG and GO for annotation, visualization and integrated discovery (DAVID) to analyze the potential function of PTGFRN in HCC. Our results demonstrate the important role of PTGFRN in the prognosis of HCC, which may provide a potential target and strategy for the diagnosis and prognosis of HCC.

Patient characteristics
Based on the data of TCGA hepatocellular carcinoma, the clinicopathological features and gene expression data of 377 primary tumors in August 2020 were retrospectively analyzed, the details are shown in Table 1.

High expression of PTGFRN in HCC
The TCGA data set provides support for our research on the expression level of PTGFRN in LIHC. Through the differential expression scatter plot and paired differential analysis, RNA-seq data from the TCGA-LIHC cohort indicated that PTGFRN expression was signi cantly increased in HCC tissues (n = 374; p = 7.147e-14; compared with normal liver tissues (n = 50), as shown in Fig. 1A. The expression of PTGFRN in paired cancer tissues is also highly statistically signi cant (P = 1.653e-09), as shown in Fig

Correlation of PTGFRN gene expression on survival and multivariate analysis
We performed relevant analyses to evaluate the effect of PTGFRN on the prognosis of patients with HCC. As shown in Fig Table 2 showed that PTGFRN-high was signi cantly associated with poor OS (hazard ratio [HR]: 1.343; 95% con dence interval [CI]: 1.056-1.707; P = 0.016). Other clinicopathological variables related to poor survival include clinical stage and pathology M. The signi cant risk factors identi ed in univariate analysis were used in multivariate Cox analysis and nally found that the expression of PTGFRN (HR = 1.277, P = 0.045) may be a potential independent risk factor for HCC (Table 3). Forest plot analysis (Fig. 3C) showed that clinical stage (P<0.001) and PTGFRN expression (P = 0.045) were statistically signi cant with the prognosis of HCC patients. and the expression of PTGFRN (P = 0.045) and the outcome of LIHC patients. In general, PTGFRN can be used as a reliable and effective independent prognostic factor.   Table 4). The Fig. 4 and Fig. 5A-F shows that "endocytosis", "oocyte meiosis", "cancer pathways", "purine metabolism", "renal cell carcinoma", "apoptosis", "glioma" and "ERBB signaling pathway", "ubiquitin-mediated proteolysis", "focal adhesion", "jak stat", "axon guidance", "apoptosis" and "phosphatidylinositol" were signi cantly enriched in PTGFRN high expression phenotype.

GO analysis and KEGG analysis of PTGFRN co-expressed genes
Genes with similar expression patterns may be linked with some functions. In order to explore the function and signaling pathways of PTGFRN in HCC, GO and KEGG enrichment analyses of PTGFRN co-expressed genes were performed. A total of 215 co-expressed genes of PTGFRN were obtained by screening. As shown in Fig. 6A, PPI analysis using STRING database revealed that 10 genes, including PTGFR, CD81, CD9, TSPAN15, TSPAN2, CD63, CD82, GDPD5, CD151, and ENSG00000250349 have interacted with PTGFRN.

Discussion
Due to the cellular heterogeneity and complex pathogenesis of HCC, its fatality rate remains relatively high worldwide [20,21]. In recent years, more and more evidence have shown that prognostic gene markers have great potential in the prognosis of HCC [8]. In this study, the PTGFRN gene was used as the research object, and 377 HCC patients from TCGA were evaluated and found factors related to high expression of PTGFRN, namely clinical stage, histological grade, and survival time. The potential functions and related pathways of PTGFRN in HCC were obtained by enrichment analysis. It further con rmed the correlation between PTGFRN and the malignant phenotype and prognosis of HCC, and clari ed the importance of PTGFRN as a potential biomarker of HCC.
In this study, we used the TCGA database for bioinformatics assessment to explore the expression level of PTGFRN in HCC. Previous studies have shown that PTGFRN is overexpressed in renal cell carcinoma and glioblastoma multiforme and is closely associated with poor prognosis [14,22]. In addition, PTGFRN has also been shown to be increased in lung tumors and is closely associated with lung tumor growth restriction and lung transplantation metastatic status [15]. We can see that the gene PTGFRN is highly expressed in many cancers, and we found that PTGFRN is also highly expressed in HCC patients. In addition, we analyzed the relevant expression data and found that the expression of PTGFRN in HCC was signi cantly higher than that of adjacent non-tumor tissues. Therefore, we speculated that the high expression of PTGFRN might affect the related progression of HCC, and carried out further analysis.
Next, we performed a series of correlation analyses in order to explore whether PTGFRN could be used as an independent target. In our study, we discovered increased expression PTGFRN in HCC associated with latestage clinical and pathological features and predict a poor prognosis. The elevated expression level of PTGFRN at the high stage suggests a negative impact on the progression of HCC. Importantly, we found that the high expression of PTGFRN in tumor tissues is closely related to shorter OS and DFS. Global gene expression pro le analysis of HCC tissue samples showed that PTGFRN mRNA expression was signi cantly up-regulated compared with normal liver tissue. The ROC curve of PTGFRN showed that its mRNA expression level had a high diagnostic value for HCC. Previous studies have not clari ed the relationship between PTGFRN expression level and HCC tissue stage. However, the present study found that PTGFRN expression has a good predictive ability in low, high and histological grades. Through the analysis of HCC stage M and N, although the results were not statistically signi cant, it was undeniable that the expression level of PTGFRN in M1 and N1 stages was higher than that in M0 and N0 stages. In addition, multivariate Cox analysis showed that overexpression of PTGFRN was an independent risk factor affecting the survival of patients with HCC. Therefore, according to the data we analyzed above, PTGFRN can serve as a potential new therapeutic target and an independent prognostic factor for HCC. Moreover, we urgently need to understand how the high expression of PTGFRN affects the occurrence and development of HCC.
To further explore the potential pathways regulated by PTGFRN in HCC, we performed GSEA analysis of PTGFRN. We have observed that the high expression phenotype of PTGFRN is mainly enriched to "endocytosis", "oocyte meiosis", "cancer pathways", "purine metabolism", "renal cell carcinoma", "regulation of actin cytoskeleton", "glioma" and "ERBB signaling pathway". Studies have pointed out that changes in the endocytosis of cancer cells will affect the changes in related cancer signaling pathways, thereby promoting the occurrence and development of tumors [23]. Zhou et al. [24] found that overexpression of SNX5 can promote the migration and invasion of HCC cells by mediating endocytosis. Qi et al. [25] found that CD147 was up-regulated in HCC, and its endocytosis played a key role in the occurrence, invasion, and distant metastasis of HCC. We suspect that the overexpression of PTGFRN may affect the progression of HCC by mediating endocytosis. In addition, related study has shown that purine nucleotides are the basic and necessary factors of tumor cell proliferation, abnormal purine metabolism can promote the rapid proliferation and growth of tumor cells [26]. For cancer pathways, overexpression of miR-296-5p has been reported to inhibit the oncogenic role of the ERBB signaling pathway in HCC with a favorable prognosis [27]. In addition to the ERBB pathway, oocyte meiosis and jak/stat pathways have also been shown to play an important role in the carcinogenesis or progression of HCC [28,29]. Perhaps this prompted that HCC activates the various signaling pathway by increasing the expression of PTGFRN, which in turn leads to a poor prognosis for patients. GSEA results also showed that PTGFRN was involved in the regulation of the actin cytoskeleton.
Actin regulatory proteins are considered to be a basic mechanism of cell migration. Studies have shown that actin regulatory proteins are considered to be the basic mechanism of cell migration. It is closely related to the poor prognosis of breast cancer and liver cancer, and is involved in the development of many cancers [30]. Based on GSEA, we also identi ed pathways associated with the cell cycle. However, the effect of transcriptional activation of PTGFRN on cell cycle progression and HCC progression still needs to be further studied. It can be seen that PTGFRN is related to many important pathways, and its up-regulation changes may affect the changes of a series of important pathways.
Finally, GO and KEGG enrichment analysis was performed on the co-expression genes of PTGFRN, and some important ndings were obtained. The results showed that PTGFRN co-expressed genes were mainly involved in biological processes such as "PI3K-Akt signaling pathway", "ECM-receptor interaction", "extracellular matrix organization", "cell development", "angiogenesis", "adhesion", "TNF-α/NF-kB signaling pathway" and "epithelial-mesenchymal transformation". As we all know, the growth and migration of solid tumor cells are closely related to angiogenesis. S Colin et al. found that the domain GS-168AT2 of PTGFRN can inhibit the proliferation and migration of human endothelial cells (hEC) in vitro and in vivo, and its anti-angiogenic activity is likely to be the main reason for the anti-tumor growth effect [15]. It indicated that PTGFRN might be involved in the malignant biological behavior of HCC. Moreover, our study also demonstrated that overexpression of PTGFRN was signi cantly associated with histological grade of HCC and predicted poor prognosis. In addition, the co-expression genes of PTGFRN are also importantly involved in the extracellular matrix organization (EMO) process. Previous studies have shown that EMO can provide a complex microenvironment for HCC cells and actively participate in the tumor development of HCC, such as angiogenesis, epithelial-mesenchymal transformation (EMT), invasion and metastasis [31,32]. Besides, EMT is believed to play an important role in the anti-apoptosis, invasion, chemical drug resistance and other pathological processes of HCC cells [33,34]. In addition, we found that PI3K-Akt and NF-kB signaling pathways were also signi cantly enriched. Studies have shown that NF-kB can act as downstream effectors of the PI3K-Akt pathway, and their abnormal activation is positively correlated with the malignant progression of HCC and leads to poor prognosis of HCC patients [35]. These enriched pathways suggest that PTGFRN may be involved in the progression of HCC at different stages. In addition, GO term enrichment analysis showed that PTGFRN may affect the cycle progression of tumor cells through cell adhesion and cell proliferation. Studies have con rmed that cell adhesion molecule overexpression can interact with ECM receptor, which is the key factor to accelerate the metastasis of HCC cells in blood vessels [36]. In addition, recent studies have shown that inhibition of PTGFRN can lead to a decrease in cell proliferation levels [14].
Furthermore, we reported for the rst time that the expression of PTGFRN may also be related to chemotaxis, lateral plasma membrane, and in ammatory response pathways. But its regulatory mechanism needs to be further elucidated. In summary, the abundant GO and KEGG pathways explain the important role of PTGFRN in the progression of HCC to a certain extent.

Conclusions
In summary, our study preliminarily con rmed that PTGFRN is a promising biomarker for the diagnosis and prognosis of HCC. We also further explored its potential mechanisms in HCC. To our knowledge, this is the rst study to determine the correlation between PTGFRN expression and clinical features of HCC based on database mining. However, this study was limited by its retrospective nature, and further studies are needed to con rm these results. Current studies on the functional role of PTGFRN are super cial, and future studies should focus on the molecular mechanisms by which PTGFRN promotes the occurrence and development of HCC.

Data Mining and Collection
The corresponding mRNA sequencing data and clinical data of HCC patients were downloaded from TCGA. Samples with missing expression data were excluded from the study. Finally, the transcriptome sequencing were drawn and hazard ratios (HRs) were calculated for each selected gene individually. The P-value was calculated by log-rank test. The Cox proportional hazard ratio and the 95% con dence interval information can also be included in the survival plot [39]. The P-ROC package was used to draw the receiver operating characteristic curve (ROC) to evaluate the diagnostic value [40].

Statistical analysis
All acquired data were statistically analyzed by R (V.3.6.3). The relationship between clinicopathological characteristics and PTGFRN were analyzed with the Wilcoxon signed-rank test and logistic regression.
Univariate Cox was used to analyzing and evaluate potential prognostic factors, and multivariate Cox analysis was used to compare the effect of PTGFRN expression on survival and other clinical characteristics (age, gender, stage, lymph node status, and distant metastasis). P<0.05 was considered statistically signi cant [41]. Concentration graph of multiple GSEA enrichment pathways Pathways associated with GSEA enrichment. A-F: The results showed that "ubiquitin-mediated proteolysis", "adhesion plaque", "axon-directed", "apoptosis", "jak stat" and "phosphatidylinositol" signaling pathways were signi cantly enriched in the high expression phenotype of PTGFRN Figure 6 PPI interaction analysis and functional enrichment analysis. A: PTGFRN protein interaction network diagram in String; B: Biological process enrichment of KEGG co-expressed genes; C: Biological process enrichment of GO co-expressed gene