Construction and Comprehensive Analysis of a Stratification System Based on AGTRAP in Patients with Hepatocellular Carcinoma

Background With the development of sequencing technology, several signatures have been reported for the prediction of prognosis in patients with hepatocellular carcinoma (HCC). However, the above signatures are characterized by cumbersome application. Therefore, the study is aimed at screening out a robust stratification system based on only one gene to guide treatment. Methods Firstly, we used the limma package for performing differential expression analysis on 374 HCC samples, followed by Cox regression analysis on overall survival (OS) and disease-free interval (PFI). Subsequently, hub prognostic genes were found at the intersection of the above three groups. In addition, the topological degree inside the PPI network was used to screen for a unique hub gene. The rms package was used to construct two visual stratification systems for OS and PFI, and Kaplan-Meier analysis was utilized to investigate survival differences in clinical subgroups. The ssGSEA algorithm was then used to reveal the relationship between the hub gene and immune cells, immunological function, and checkpoints. In addition, we also used function annotation to explore into putative biological functions. Finally, for preliminary validation, the hub gene was knocked down in the HCC cell line. Results We discovered 6 prognostic genes (SKA1, CDC20, AGTRAP, BIRC5, NEIL3, and CDC25C) for constructing a PPI network after investigating survival and differential expression genes. According to the topological degree, AGTRAP was chosen as the basis for the stratification system, and it was revealed to be a risk factor with an independent prognostic value in Kaplan-Meier analysis and Cox regression analysis (P < 0.05). In addition, we constructed two visualized nomograms based on AGTRAP. The novel stratification system had a robust predictive value for PFI and OS in ROC analysis and calibration curve (P < 0.05). Meanwhile, AGTRAP upregulation was associated with T staging, N staging, M staging, pathological stage, grade, and vascular invasion (P < 0.05). Notably, AGTRAP was overexpressed in tumor tissues in all pancancers with paired samples (P < 0.05). Furthermore, AGTRAP was associated with immune response and may change immune microenvironment in HCC (P < 0.05). Next, gene enrichment analysis suggested that AGTRAP may be involved in the biological process, such as cotranslational protein targeting to the membrane. Finally, we identified the oncogenic effect of AGTRAP by qRT-PCR, colony formation, western blot, and CCK-8 assay (P < 0.05). Conclusion We provided robust evidences that a stratification system based on AGTRAP can guide survival prediction for HCC patients.


Introduction
Globally, primary liver cancer is one of the most common cancers, with a high mortality. Hepatocellular carcinoma (HCC) accounts for the great majority of liver cancer, which is genetically defined as a malignancy with unique molecular events and is anticipated to be with high heterogeneity [1]. Despite the widespread application of pathological staging, it still has limits in predicting survival [2]. As a result, a novel stratification system must be developed that provides for precise clinical intervention.
AT1R-associated protein (ATRAP) is a molecule specifically interacting with the carboxyl-terminal domain of the angiotensin II (Ang II) type 1 receptor (AT1R) [3]. Although AGTRAP (ATRAP) is found in a variety of human tissues, little is known about it in tumor tissue [4]. Because AT1R is involved in the pathophysiology of hypertension, fundamental research has focused on the function of AT1R and ATRAP in hypertension development [5,6]. A recent study on illnesses like hypertension and nephrotic syndrome has focused on AGTRAP. ATRAP expression was positively associated with AT1R gene expression in 22 kidney biopsy specimens of IgA nephropathy [7]. Furthermore, ATRAP expression of hypertension patients was significantly lower than that of normotensive patients in 36 visceral adipose tissues from abdominal surgery [4]. In particular, ATRAP expression was strongly associated with inflammatory indicators such as granulocyte and monocyte counts in outpatients with noncommunicable illnesses [8]. Hence, therefore, in order to fill the gap of AGTRAP in tumorigenesis, the role of AGTRAP in HCC was explored in depth using bioinformatics analysis and assays.
Several signatures have been reported in the previous references for the prediction of prognosis in patients with HCC [9][10][11]. However, the above signatures are characterized by excessive factors and cumbersome application in clinical settings. Therefore, the study is aimed at screening out a robust stratification system based on only one gene to facilitate prognosis prediction of HCC patients.

Materials and Methods
2.1. Differential Expression Analysis. We download RNAsequence data (HTseq-FPKM) from the Pancancer Project in The Cancer Genome Atlas (TCGA) database. The expression of AGTRAP was compared between the normal and tumor tissues after log2 transformation. In addition, the RNA-sequence data of TCGA-LIHC in the same format was downloaded to perform AGTRAP differential analysis of paired (50 pairs) and unpaired (374 HCC tissues and 50 normal adjacent tissue). The thresholds were set to log 2ð FCÞ ≥ 4, P < 0:05, in the limma package in R software, and the gene set of differential expression was identified.

Construction of PPI Network and Screening of Prognostic
Risk Factors. Prognostic risk factors in HCC patients were identified by univariate Cox regression analysis, and thresholds were set to HR > 1, P < 0:001. According to Cox regression analysis, gene sets related to overall survival (OS) and progress-free interval (PFI) were identified. Then, we intersected the above three gene sets and screened the hub genes involved in the construction of a protein-protein interaction (PPI) network. The PPI network was constructed by using the STRING tool and Cytoscape software. We retained only a gene with top topological degree.

Clinical Prognosis Analysis and Construction of
Stratification System. The clinical data was downloaded from the TCGA database, including age, gender, pathological stage, T staging, N staging, M staging, grade, residual size, AFP, albumin, and vascular invasion. We calculated the median expression of AGTRAP of HCC patients, which is used to select "high-risk" and "low-risk" groups. Kaplan-Meier survival analysis and log-rank test were used to suggest the survival differences in the two groups. In addition, independent prognostic factors were identified by univariate and multivariate Cox regression analysis. We used the "rms" package in R software to plot a nomogram for visualizing the prognosis value of AGTRAP. The distinction and calibration were evaluated by the ROC curve and calibration curve.

Enrichment Analysis of Differentially Expressed Genes.
We divided all HCC samples into two groups (AGTRAPhigh and AGTRAP-low) based on the median expression of AGTRAP. The differentially expressed genes (DEGs) in AGTRAP-high samples and AGTRAP-low samples were screened using the limma package in R software. The thresholds were set to |log 2ðFCÞ | >2 and P:adj < 0:05. Moreover, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses were performed using related packages.
2.5. Immune-Infiltration Analysis. The immune-infiltration algorithm used in this study was ssGSEA, which was implemented through the GSVA package in R software. The correlation between AGTRAP and immune function, immune cells, and immune checkpoints was analyzed. Pearson correlation analysis was used to verify the correlation between the risk group and immune cell infiltration.
2.6. In Vitro Assays. In this study, we used cell culture, transfection, CCK-8, and qRT-PCR as in vitro assays. The Shanghai Cell Institute Country Cell Bank provided the normal and HCC cell lines. GenePharma generated and annealed small-interfering RNA (si-RNA-1/2/3) oligos for AGTRAP and a general negative control. Following the manufacturer's procedure, each siRNA duplex was transfected into the cells using Lipofectamine® 2000 (Invitrogen, Carlsbad, CA, USA). The antibodies against AGTRAP and GAPDH were all obtained from Abcam. GAPDH served as the internal control. The sequences of the primers used for qRT-PCR are as follows in reference [12]. The details of the methods are provided in reference [13]. In addition, transwell, clone

Disease Markers
formation, western blot, and other detailed experimental processes are discussed in our previous study [14].

Statistical Analysis.
All statistical analyses were performed using the R software (v.4.0.1). Detailed statistical methods about transcriptome data are covered in the bioinformatics method section. P < 0:05 was considered statistically significant.

Screening of a Hub Prognosis Gene in Patients with
Hepatocellular Carcinoma. Differential expression analysis was performed on the transcriptome data of all samples in the TCGA-LIHC cohort (50 adjacent normal samples and 374 HCC tissues samples). In order to identify potential protooncogenes, we screened only the upregulated genes in HCC tissues, and finally, 474 hub protein-coding genes were identified. Meanwhile, univariate Cox regression analysis was performed on 374 patients corresponding to transcriptome data. We selected risk factors with HR > 1 for screening and finally identified 721 risk factors related to OS and 838 risk genes related to PFI for HCC. Finally, we intersected the above genes and further screened the 6 hub genes involved in the construction of the PPI network, as shown in Figure 1(a). In the PPI network, we calculated the topological degree for the 6 hub genes and determined AGTRAP as the final factor, as shown in Figure 1(b). In addition, our survival analysis for 374 HCC patients also showed significant predictive performance in AGTRAP (P < 0:001), as shown in Figures 1(c) and 1(d).

Independent Prognostic Role of AGTRAP in Progress-
Free Interval and Overall Survival. In order to further explore the independent prognostic ability of AGTRAP with clinicopathologic factors, we again conducted univariate and multivariate Cox regression analysis in 374 HCC patients. In OS, univariate Cox analysis revealed that AGTRAP with the same pathological staging and T staging were risk factors (P < 0:05), as shown in Figure 2(a). Moreover, further multivariate Cox analysis showed that only AGTRAP was independently associated with OS (P < 0:05), as shown in Figure 2(b). In PFI, similar to the results of Cox analysis in OS, AGTRAP was also a high-risk factor for HCC recurrence (P < 0:05), as shown in Figures 2(c) and 2(d). Taken together, the results may imply that AGTRAP may be an independent prognostic predictor for HCC patients.

Construction and Validation of Visual Prognostic
Stratification System. Considering the clinical value of the pathological stage, we combined the pathological stage and the significance of AGTRAP in multivariate Cox analysis to construct two visual stratification systems for OS and PFI, as shown in Figures 3(a) and 3(d). It is worth mentioning that ROC analysis and calibration curve also showed that the stratification system has good predictive value, as shown  Disease Markers  The expression of AGTRAP

Expression Landscape and Clinical Correlation Analysis of AGTRAP.
To further explore the expression of the stratification system in pancancers, we downloaded expression profile raw data of pancancers from the TCGA database. The results showed that AGTRAP was upregulated in most tumors in the unpaired differential analysis, as shown in Figure 4(a). In particular, AGTRAP was overexpressed in all tumor tissues in the paired differential analysis, as shown in Figure 4(d). In detail, compared with normal liver tissues, AGTRAP was overexpressed in HCC tissues (Figures 4(b) and 4(c)). As illustrated in Figures 4(e)-4(j), for the significant differential clinicopathologic factors, AGTRAP was overexpressed in T staging, N staging, M staging, pathological stage, vascular infiltration, and G3-G4 (P < 0:05).

Potential Prognostic Significance of AGTRAP in Clinical
Subgroups.       14 Disease Markers overall survival time of HCC patients with high expression of AGTRAP was significantly shorter than that with low expression of AGTRAP (P < 0:05).
3.6. Analysis of Potential Biological Mechanisms Involving AGTRAP. To further explore the biological mechanisms of AGTRAP, we calculated the median expression of AGTRAP in HCC patients, which is used to select the "high-risk" and "low-risk" groups. We used the limma package in R software to explore the differential expression of genes in the two groups. According to the threshold of the method section, we screened a total of 2467 upregulated genes and 711 downregulated genes, which may be involved in the regulation with AGTRAP. KEGG and GO enrichment analyses were performed for the above genes, as shown in Figures 6(a) and 6(b). GO enrichment analysis showed that

Disease Markers
AGTRAP and its coexpressed genes may be involved in cotranslational protein targeting to the membrane, RNA catabolic process, mRNA catabolic process, etc. KEGG analysis revealed that the above genes may be involved in pathways of neurodegeneration, Alzheimer disease, amyotrophic lateral sclerosis, and other autoimmune pathways. On the basis of differential genes, we analyzed the coexpression gene network of AGTRAP and conducted gene enrichment analysis again. Only 20 genes are coexpressed with AGTRAP, including NECAP2, GP6C3, CAPZB, ALDOA, MIIP, ENO1, ATAD3B, TMEM234, UBE2M, ATAD3B, 2BTB17, 4GRN, ARPC, UBE2J2, SZED1, MAD2L2, ADPRHL2, GIT1, TRNAU1AP, and LYPLA2, as shown in Figure 6(c). Interestingly, these genes may be involved in the HIF-1 signaling pathway, glycolysis, and other biological processes, as shown in Figure 6(d).

A Comprehensive Analysis of Immune Function
Based on the Stratification System. Metabolic reprogramming and immune escape are independent predictors of patient survival, and changes in the immune microenvironment regulate tumor progression [15]. Therefore, we used the ssGSEA algorithm to comprehensively analyze the immune cell content, immune cell correlation, immune function, and immune checkpoint in liver cancer tissue. HCC patients were divided into a high-risk group and a low-risk group. It showed that Th2 cells, Th17 cells, Th1 cells, TFH, Tcm, T cells, NK CD56 bright cells, macrophages, iDC, and aDC differed significantly between the high-risk and low-risk groups (Figure 7(a)). Moreover, the results showed that AGTRAP was negatively correlated with only 3 immune cells, including eosinophils, Tcm, and Th17 cells, and 10 immune cells were positively correlated (Figure 7(b)). Interestingly, analysis of immunologic function confirmed significant differences between the low-and high-risk groups for other immunological functions except MHC class I (P > 0:05), as shown in Figure 7(c). Finally, it is worth noting that given the importance of checkpoint immunotherapy, all have significant differences in the expression of immune checkpoints between different risk groups, as shown in Figure 7(d).
3.8. In Vitro Assays for Validation. To further validate the above bioinformatics results, we detected the expression level of AGTRAP mRNA in HCC cell lines. The results showed that expression of AGTRAP was upregulated in HCC cell lines (PLC, HEp3B, and HEpG2) compared to THLE-3, as shown in Figure 8(a). In addition, si-AGTRAP and si-NC were transfected in PLC and HEp3B cells, respectively, and qRT-PCR and western blot were used to detect the protein expression of AGTRAP. It was found that AGTRAP expression was downregulated in HCC cell lines with transfection, as shown in Figure 8(b). Similarly, CCK-

16
Disease Markers 8 assays showed that HCC cell proliferation was inhibited after transfection with AGTRAP, as shown in Figure 8(d).
We calculated the IOD value of IHC for 13 clinical samples in the HPA database about the AGTRAP protein expression, as shown in Figure 8(c). We found that the AGTRAP protein level is also overexpressed in tumor samples. In the cloning formation assay, we also found that knocking down AGTRAP could affect proliferation of HCC cells, as shown in Figure 8(f). Unfortunately, the downexpression of the AGTRAP may have no effect on the invasion and migration of HCC cells, as shown in Figure 8(e).

Discussion
Currently, although the TNM system is used to roughly determine the prognosis of HCC patients [1], different stratification systems are not always effective in predicting prognosis depending on the different genetic characteristics. The risk signatures in the previous references are too complicated to be used by clinicians and are expensive to use. Furthermore, as the largest immune organ, the liver plays an important role in the immune response [16]. Therefore, the study focused on the stratified prognostic value of AGTRAP and its impact on the immune microenvironment. Targeted therapy has been intensively studied in a variety of tumors and has now been shown to be a possible new therapeutic approach [17]. However, compared to other targets, there has been relatively little research on AGTRAP and cancer, particularly with regard to its specific mechanisms in HCC. In this study, we performed a comprehensive analysis of AGTRAPA in HCC. We compared the genetic landscape of AGTRAP across TCGA databases; next, we found a robust correlation between AGTRAP and clinicopathologic factors. At the same time, the remarkably prognostic predictive value of AGTRAP was through a series of studies. In addition, we further explored the potential mechanism of AGTRAP and the impact on immune function. Finally, we further explored the role of AGTRAP in the HCC cell line by qRT-PCR and western blot.
However, there are a number of limitations to our study that need to be considered with caution. Our study is based on the TCGA database alone, with no validation of an external dataset and clinical samples. Finally, more functional assays are needed to confirm our findings and to better understand the role of AGTRAP in HCC.

Conclusions
We provide strong evidence that a stratification system based on AGTRAP can guide survival prediction in HCC patients and may have an impact on the immune microenvironment in HCC tissues.

Data Availability
The following information was supplied regarding data availability: data is available at the TCGA (https://portal .gdc.cancer.gov/).

Disclosure
The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.