Comprehensive Bioinformatics Analysis Reveals PTPN1 (PTP1B) Is a Promising Immunotherapy Target Associated with T Cell Function for Liver Cancer

Recently, PTP1B was identified as a novel immune checkpoint whose removal can unleash T cell responses. However, research on the influence of PTP1B as an immune regulator on liver cancer is limited. This study aimed to investigate the immunological correlation and function of PTP1B in liver cancer. The expression profiles and corresponding clinical information of liver cancer patients were obtained from the TCGA and ICGC databases. GSE146115 and GSE98638 retrieved from the GEO database were used for the single-cell RNA-seq analysis. The mRNA expression of PTP1B (PTPN1) was increased in patients with most malignancies (all p < 0.05), including liver cancer (p < 0.001). Furthermore, up-regulated PTPN1 was connected to advanced tumor stage (p < 0.05) and worse prognosis (p < 0.01) in liver cancer. Through Cox regression analysis, PTPN1 was considered as an independent prognosis factor of overall survival (p < 0.05) and acted as a high-risk factor (hazard ratio > 1). Gene function and pathway analysis suggested PTPN1 was involved in T cell-related immune responses. Moreover, a close relationship was also found between PTPN1 expression and immune checkpoints as well as immune cells, especially with T cell-related checkpoints (all p < 0.001) and T cells (all p < 0.001). Single-cell RNA-seq analysis further illustrated that the enrichment of PTPN1 in the T cell population may be linked to its exhaustion in the liver cancer microenvironment. Overall, PTPN1 (PTP1B) closely related to T cell may function as an immunotherapy target for liver cancer.


Introduction
According to the Global Cancer Statistics of 2020, liver cancer became the seventh most common type of cancer worldwide and the third leading cause of cancer-related death [1]. Clinically, most patients with liver cancer are diagnosed at an advanced stage because the symptoms at an early stage are absent or concealed [2]. In the past decades, although we have made remarkable achievements in improving treatment for liver cancer, the therapy for patients with advanced liver cancer is still limited and inefective [3]. For patients with advanced liver cancer, systemic chemotherapy is currently the only available and efective treatment [3,4]. Unfortunately, a large proportion of patients with advanced liver cancer sufer from severe liver dysfunction, often showing poor treatment with systemic chemotherapeutics due to drug resistance [3].
Numerous research studies have demonstrated the complex immunosuppressive network and tumor microenvironment in liver cancer, indicating a promising prospect of immunotherapy, including immune checkpoint therapy and chimeric antigen receptor (CAR) T cell therapy [2,5]. As an emerging therapy, immunotherapy has been proven to signifcantly prolong the survival time of patients with liver cancer and improve their prognosis [6]. Notably, the combination treatment of tyrosine kinase inhibitors (TKIs) and anti-PD-1 antibodies proved to be a feasible conversion therapy for patients with advanced or unresectable liver cancer [7]. Nevertheless, there still exist some limitations to immune checkpoint therapy for liver cancer, mainly because of the resistance to immune checkpoint blockades, which is partly related to Tcell infltration defects [2,8]. Furthermore, CAR-T therapy for the treatment of liver cancer remains inadequate [9]. Tus, searching for new immune checkpoints closely associated with T cell function may be helpful in improving the situation.
In humans, protein-tyrosine phosphatase 1B (PTP1B), encoded by protein tyrosine phosphatase nonreceptor type 1 (PTPN1), is known to act as a distinct regulatory role in various diseases, such as neurodegenerative diseases, liver disease, cancer, and diabetes [10]. Recently, PTP1B was identifed as a novel intracellular immune checkpoint that limits the anti-tumor function of T cells [11]. Tis leading research also revealed that targeting PTP1B is signifcant for promoting CAR T cell-mediated antitumor immunity [11]. Accordingly, PTP1B is expected to be a promising target for cancer immunotherapy. However, little is known about the infuence of PTP1B on patients with liver cancer [12]. Terefore, elucidating the immunological relevance and function of PTP1B in liver cancer are urgently needed.
In this study, we conducted a comprehensive bioinformatics analysis on PTPN1 (mRNA expression level of PTP1B) in liver cancer based on multiple public databases. A brief fow chart of our research is shown in Supplement Figure S1. Firstly, we investigated the expression profle of PTPN1 at pancancer level via Te Cancer Genome Atlas (TCGA) database and Clinical Proteomic Tumor Analysis Consortium (CPTAC) database, as well as the diferential expression of PTPN1 in liver cancer from the International Cancer Genome Consortium (ICGC) database and the Human Protein Atlas (HPA) database. Ten, we explored the association between the expression of PTPN1 and the clinical characteristics and survival of patients with liver cancer in the TCGA and ICGC databases. To explore whether the biological function and signal pathway of PTPN1 is related to immune responses, three functional and pathway enrichment methods were conducted [13][14][15]. Tese methods were the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis on diferentially expressed genes (DEGs) between low-and high-PTPN1 expression groups, the Gene Set Enrichment Analysis (GSEA), and the GO and KEGG analysis on PTPN1-correlated genes. Herein, various immune-related enriched GO terms and KEGG pathways of PTPN1 were observed in the TCGA and ICGC databases. Tus, we furthermore analyzed the correlation of PTPN1 expression with immune checkpoints and tumor infltrated immune cells, respectively. Using GEO database, single-cell RNA-seq analysis was then performed to identify the PTPN1expressed immune infltration cell in liver cancer. Overall, our study shed light on the upregulation of PTPN1 with advanced stage and poor prognosis in liver cancer and revealed PTPN1 (PTP1B) as a potential immunotherapy target due to its close association with immune-related pathways, immune checkpoints, and infltrating T cells.

Data Resource and Preprocess.
Te expression profles (TPM) and corresponding clinical data of liver cancer were downloaded from the TCGA database using the R package "TCGA biolinks," including 50 normal samples and 374 liver cancer samples. We also downloaded the expression profles (normalized counts) and clinical data of liver cancer from the LIRI-JP project of the ICGC database (https://dcc.icgc. org/releases/current/Projects/LIRI-JP), containing 202 normal samples and 243 liver cancer samples, as the validation dataset for our study. Samples with missing or uncertain information in these datasets were excluded. Tese samples were then divided into low-and high-PTPN1 expression groups according to the median expression of PTPN1. Moreover, GSE146115 and GSE98638 were obtained from the GEO database for the single-cell RNA-seq analysis [16,17].

Analysis of PTPN1 Expression in Pancancer and Liver
Cancer. Te TIMER (https://cistrome.shinyapps.io/timer/) and UALCAN (https://ualcan.path.uab.edu) online tools were used to analyze the diferential expression of PTPN1 in distinct cancer types and liver cancer from the TCGA database and the CPTAC database, respectively [18,19]. Te expression of PTPN1 in liver cancer was then analyzed in the ICGC database by the R package "ggpubr." Te HPA database (https://www.proteinatlas.org/) was further utilized to study the protein expression levels of PTPN1 in liver cancer and normal liver. In our study, the expression of PTPN1 in the TCGA database was normalized as log 2 (TPM), which is consistent with the method that TIMER used. Furthermore, Z-value and log 2 (normalized counts) were separately applied to measure the expression of PTPN1 in the CPTAC database and the ICGC database.

Correlation Analysis of Clinical Characteristics.
To explore the correlation between PTPN1 expression and distinct clinical characteristics of liver cancer, the normal samples were excluded, and clinical characteristics (gender, tumor stage, and Tumor Node Metastasis classifcation (TNM classifcation)) were converted into numerical variables for further analysis. Te R package "ggplot2" and "ComplexHeatmap" were used to describe the distribution of survival status and clinical characteristics of patients in the TCGA database and the ICGC database according to the increased expression of PTPN1. Ten, the expression of PTPN1 among diferent groups based on age, gender, tumor grade, and TNM classifcation was investigated using the R package "ggpubr."

Survival Analysis.
To identify the prognostic value of PTPN1 in liver cancer, tumor samples were classifed into low-and high-PTPN1 expression groups according to the best cutof value of PTPN1 using the function "surv_cutpoint" inserted in the R package "survminer." Next, K-M curves were plotted to illustrate the relationship of PTPN1 expression and overall survival (OS) using the R package "survminer" and "survival." To test whether PTPN1 was an independent prognostic factor of OS, univariate and multivariate Cox regression analyses were performed using the R package "survival." 2.5. Gene Function and Pathway Analysis by GO, KEGG, and GSEA. To investigate the potential biological function and pathway, we frst screened the DEGs between the low-and high-PTPN1 expression groups using the R package "limma." Te genes with log 2 |fold change (FC)| larger than one and adjusted p value (q value) lower than 0.05 were considered signifcant DEGs [14]. Next, these DEGs were used to perform GO and KEGG pathway enrichment analysis. GO terms and KEGG pathways with q value lower than 0.05 were signifcantly enriched in DEGs [14]. Furthermore, we also utilized GSEA pathway analysis on PTPN1 by the KEGG gene sets. GO, KEGG, and GSEA analysis were conducted by the R package "clusterProfler."

PTPN1-Related Genes Enrichment Analysis by GO and KEGG.
Considering the limitations of GO and KEGG analysis based on the DEGs, the method using correlated genes was utilized to annotate the function and pathways of PTPN1 for verifcation. Firstly, the GeneMANI database (https://genemania.org/) and STRING database (https://cn. string-db.org/) were used to display PTPN1-dominated gene interaction network and protein interaction network, respectively. Next, the interactive genes with PTPN1 from GeneMANI database and STRING database were analyzed by Pearson's correlation to obtain the signifcant part, combined with 60 genes encoding immune regulators (Supplement Table 1) [20,21]. Moreover, we used Pearson's correlation method to screen out the top 100 PTPN1related genes in the TCGA database and ICGC database, respectively, and the top 10 signifcant correlated genes in the TCGA and ICGG databases were then visualized as a chordal graph by the R package "circlize." Accordingly, the important genes positively related with PTPN1 (p < 0.001 and Pearson R > 0) were obtained (Supplement Table 2), which were used for GO and KEGG pathway analysis in the TCGA and IGCG databases. GO terms and KEGG pathways with q value lower than 0.05 were considered the signifcantly enrichment of PTPN1-correlated genes. GO and KEGG analyses were performed by the R package "clusterProfler." 2.7. Correlation Analysis of Immune Checkpoints. To explore whether the expression of PTPN1 was correlated with immune checkpoints in liver cancer, 47 members of immune checkpoints were obtained for our study [20]. Using the R package "ggpubr," the expression profles of these immune checkpoints among low-and high-PTPN1 expression groups in the TCGA database and ICGC database were revealed. Next, we assessed and visualized the signifcant correlations between the immune checkpoints and PTPN1 expression by the R package "corrplot." Te top 10 important immune checkpoints correlated with PTPN1 in the TCGA database and ICGC database were displayed in chordal graph by the R package "circlize," respectively.

Correlation Analysis of Immune Infltration Cells.
To illustrate the relationship between PTPN1 and immune infltration cells in liver cancer, the TIMER online tool was adopted to analyze. Herein, the relationship between PTPN1 and immune infltration cells was investigated from the infltration abundances and immune gene expressions, respectively. TIMER applied the constrained least squares method to estimate the infltration abundance of six immune cells in liver cancer from the TCGA database [18]. Firstly, we evaluated the relationship between PTPN1 expression and infltration levels of six immune cells (B cell, CD8+ T cell, CD4+ T cell, macrophage, neutrophil, and dendritic cell) using Spearman's correlation method. To further explore more specifc types of immune cells, the gene markers of

Single-Cell RNA-Seq Analysis.
To precisely identify the PTPN1-expressed immune cells infltrated in liver cancer, we conducted single-cell RNA-seq analysis on GSE146115 obtained from the GEO database. Te R package "Seurat" was used for clustering analysis based on PCA and UMAP algorithms. Ten, we utilized the R package "SingleR" to identify the cell types of diferent clusters. In order to annotate cells more accurately, the marker genes of liver cancer cells and immune cells from the Cell Marker 2.0 database were obtained to correct the wrongly identifed cell types. Next, we further identifed the T cell subtypes in GSE146115 and then utilized the R package "Monocle" for trajectory analysis using the "DDRTree" method. Te additional processes of single-cell RNA-seq analysis on GSE146115 are shown in Supplement Figures S2 and S3. Furthermore, we analyzed GSE98638 dataset using TISCH online tool (https://tisch.comp-genomics.org/) to verify the results of GSE146115 [23].

Statistical Analysis.
Statistical analysis was performed using R programming language (version 4.2.0). Wilcoxon test or unpaired t-test was applied to evaluate the expression of PTPN1 in tumor tissues versus normal tissues. Unpaired t-test was also utilized to investigate the relationship between PTPN1 expression and various clinical characteristics (all converted to numerical variables) of liver cancer. Log rank test was used to analyze the diference of OS between lowand high-PTPN1 expression groups. Pearson's correlation method was conducted to screen the correlated genes of PTPN1, and the correlation between PTPN1 expression and immune checkpoints was evaluated. Spearman's correlation analysis was performed to test the correlation of PTPN1 expression between infltration levels and gene markers of immune cells. Te statistical analyses were all two-tailed tests, and p < 0.05 was considered statistically signifcant.

PTPN1 Is Up-regulated in Liver
Cancer. Te increased expression of PTPN1 in liver cancer was confrmed by the ICGC database. Te result of the ICGC database (p < 0.001, Figure 2(c)) was consistent with the TCGA database (p < 0.001, Figure 2(a)) and CPTAC database (p < 0.001, Figure 2(b)). We further assessed the protein expression level of PTPN1 in liver cancer using the HPA database. Immunohistochemical staining showed that PTPN1 was signifcantly upregulated in liver cancer tissue compared to normal liver tissue (Figures 2(d) and 2(e)).

Association between PTPN1 Expression and Clinical
Characteristics of Liver Cancer. To explore the relationship between PTPN1 and clinical characteristics of liver cancer, scatter plots and heatmaps were frst utilized to show the distribution of survival status and clinical features of liver cancer patients in the TCGA database (Figure 3(a)) and the ICGC database (Figure 3(b)). Te results indicated that the higher expression of PTPN1 was linked with the accumulation of dead patients and patients with advanced stage (stage III-IV). We then compared the PTPN1 expression in distinct patients based on clinical characteristics. Te TCGA and ICGC databases demonstrated that the expression of PTPN1 did not correlate with patients' age (p > 0.05, Figures 3(c) and 3(i)) or gender (p > 0.05, Figures 3(d) and 3(j)). Patients with stage III-IV showed higher PTPN1 expression in the TCGA database (p < 0.05, Figure 3(e)) and the ICGC database (p < 0.05, Figure 3(k)). Moreover, PTPN1 expression was correlated with the T stage (p < 0.05, Figure 3(f )) and M stage (p < 0.001, Figure 3(h)) but not with the N stage (p > 0.05, Figure 3(g)). Tese results suggested that PTPN1 was upregulated in patients with advanced stage, indicating its important role in the development of liver cancer.

Prognostic Value of PTPN1 in Liver Cancer.
To evaluate the prognostic value of PTPN1 in liver cancer, Kaplan-Meier analysis and Cox regression analysis were conducted. Te Kaplan-Meier curve for OS showed that patients with higher expression of PTPN1 had a worse prognosis in the TCGA database (p � 0.002, Figure 4(a)) and the ICGC database (p � 0.003, Figure 4(b)). Univariate and multivariate Cox regression analysis was then used to confrm whether PTPN1 was an independent prognostic factor of liver cancer. Te results indicated that PTPN1 was independent of clinical features (including age, gender, and stage) and acted as a high-risk factor (hazard ratio > 1) for liver cancer patients in the TCGA database (Figures 4(c) and 4(e)) and the ICGC database (Figures 4(d) and 4(f )).

PTPN1 Is Involved in Immune-Related and Cell Adhesion-Related Signaling Pathways.
To illustrate the potential biological functions and pathways of PTPN1 in liver cancer, we frst screened DEGs between the high-expression group and the low-expression group of PTPN1. Ten, GO and KEGG analyses were performed based on these DEGs. In the TCGA database, the enriched GO terms (q < 0.05, Figure 5(a)) and KEGG pathways (q < 0.05, Figure 5(b)) were closely associated with immune response receptors and cell adhesion, including immune receptor activity, chemokine binding, cytokine-cytokine receptor interaction, T cell receptor signaling pathway, ECM-receptor interaction, and proteoglycans in cancer. Ten, GO and KEGG analyses were utilized in the ICGC database to validate the above fndings. Te results of the ICGC database also showed that the enriched GO terms (q < 0.05, Figure 5(c)) and KEGG pathways (q < 0.05, Figure 5(d)) were related to immune response and cell adhesion. To further identify the specifc molecular mechanism afected by PTPN1 in liver cancer, we also conducted a GSEA pathway analysis using KEGG gene sets. Interestingly, cytokine-cytokine receptor interaction was also found and signifcantly enriched in the up-regulated groups of PTPN1 in the TCGA database (p < 0.05, Figure 5(e)) and the ICGC database (p < 0.05, Figure 5(f )). Terefore, PTPN1 may be involved in immune responses of liver cancer, in which the cytokine-cytokine receptor interaction pathway may be the underlying mechanism.

PTPN1 Regulates Immune Response Processes with Its
Interactive Genes. To ensure the accuracy of the above results, GO and KEGG analyses based on PTPN1-related genes were performed in the TCGA database and the ICGC database as another strategy to annotate biological functions and pathways. As shown in Figures 6(a) and 6(b), the GeneMANI database and STRING database provided us with the gene-gene and protein-protein interaction network of PTPN1, respectively. Based on these networks, 27 genes interacting with PTPN1 were obtained for further analysis (duplicate genes are excluded, see Supplement Table 1). Furthermore, the top 100 PTPN1-correlated genes and 60 genes encoding immune regulators are also included in our analysis (Supplement Table 1). Pearson correlation analysis was then used to screen out the signifcantly correlated genes with PTPN1 (p < 0.001). Herein, we ultimately obtained a set of 165 genes and a set of 138 genes in the TCGA database and the ICGC database, respectively (Supplement Table 2). Te top 10 genes positively correlated with PTPN1 in the TCGA database and the ICGC database are shown in Figures 6(c) and 6(f ), respectively. In the TCGA database, PTPN1 interacting genes are enriched in many immune response-related functions and pathways (q < 0.05, Figures 6(d) and 6(e)), such as regulation of T cell activation, cytokine-cytokine receptor interaction, T cell receptor signaling pathway, and T17 cell diferentiation. Tese results are then confrmed in the ICGC database (q < 0.05, Figures 6(g) and 6(h)). Notably, most of the results pointed to T cell-related immune responses and pathways, indicating that the function of T cells might be afected by PTPN1 and its related genes in liver cancer.

Correlation Analysis between PTPN1 Expression and
Immune Checkpoint Members in Liver Cancer. PTPN1 was proven to show synergistic efects with some immune checkpoints, such as PD-1 [11]. Herein, we analyzed the

Correlation Analysis between PTPN1 Expression and
Immune Cell Infltration in Liver Cancer. To elucidate whether PTPN1 was related to immune infltration cells in liver cancer, the TIMER database was used to analyze the correlation between the expression of PTPN1 and six types of immune infltration cells, including B cell, CD8+ T cell, CD4+ T cell, macrophage, neutrophil, and dendritic cell. Figure 8(a) shows that PTPN1 expression was positively correlated with the infltration levels of the above six immune cells (all Spearman R > 0, all p < 0.001). Next, we assessed the expression correlation between PTPN1 and gene markers of nine immune cells. Te expression of PTPN1 was closely positive (most Spearman R > 0.2) associated with B cell (Figure 8(b)), T cell (Figure 8(c)), CD4+ T cell (Figure 8(d)), CD8+ T cell (Figure 8(e)), Treg cell (Figure 8(f )), macrophage (Figure 8(h)), and dendritic cell (Figure 8(i)) but weakly positive (all Spearman R < 0.2) associated with T cell (Figure 8(g)) and NK cell (Figure 8(j)). Tese results indicated a close linkage between PTPN1 and most immune cells in liver cancer, in which T cell might be an essential cell type afected by PTPN1.

Distribution of PTPN1 in Diferent Subtypes of T Cells.
To further explore the PTPN1-expressed immune cells in liver cancer, two datasets (GSE146115 and GSE98638) were obtained for the single-cell RNA-seq analysis. Te entire cell profle of the GSE146115 dataset was projected onto the UMAP plot (Figure 9(a)). Using the R package "singleR" and specifc cell markers, fve main types of cells were annotated in the GSE146115 dataset, including tumor cells, B cells, T cells, macrophages, and NK cells. We then explore the expression of PTPN1 and gene markers of two classical T cell subtypes (CD4+ and CD8+ T cell) in these cells. As can be seen from Figure 9(b), CD4, CD8A, and CD8B were all expressed in the T cell population, and PTPN1 was also enriched. Considering that the PTPN1 expression may be infuenced by the heterogeneity of the T cell population, we further identifed the specifc subtype of T cells in GSE146115. As shown in Figures 9(c) and 9(d), the T cells were divided into CD4+ and CD8+ subtypes according to the gene markers. Figure 9(d) further demonstrated that CD4+ and CD8+ T cells were all PTPN1-expressed cell types while exhibiting the maker genes of T cells infltrating liver cancer. Moreover, Monocle 2 trajectory analysis was performed to distinguish the T cell population based on the pseudotime state. Te T cell population is classifed into three distinct states as shown in Figure 9(e). Figure 9(f ) further demonstrated the expression linkage pattern of PTPN1 and T cell exhaustion-related genes (CD27, CTLA4, and LAG3) in CD4+ and CD8+ T cells, and had minor relationships with the cell states. GSE98638 dataset was then analyzed by TISCH online tool to verify the above results. Te profle of T cells infltrated in liver cancer from the GSE98638 dataset is releveled in Figure 9(g), including conventional CD4+ T cells (CD4Tconv), CD8+ T cells (CD8T), regulatory T cells (Treg), exhausted CD8+ T cells (CD8Tex), proliferating T cells (Tprolif), and other cells. Signifcantly, the overexpression distribution pattern of PTPN1 was observed in the fve distinct types of Tcells from GSE98638 (Figure 9(h)). Overall, the data suggested that PTPN1 was enriched in diferent types of T cells in GSE146115 and GSE98638, some of which were exhausted cells or displayed exhaustion characteristics; thus, indicating the potential relationship between PTPN1 expression and exhausted T cells infltrated in liver cancer.

Discussion
As the founding member of the protein tyrosine phosphatase (PTP) superfamily, PTP1B (encoded by PTPN1) has been reported to exert critical roles in many physiological and pathological processes, such as insulin signal transduction and cancer development [10,11]. Accordingly, PTP1B has long been regarded as a very attractive therapeutic target against metabolic syndrome and cancer [10,11]. Recently, a leading study uncovered that PTP1B is an immune checkpoint upregulated in tumor T cells and the inactivation of PTP1B can enhance T cell-mediated antitumor immunity [11]. Te research suggested that PTP1B is a promising immunotherapy target, especially for providing an alternative for CAR-T cell immunotherapy in solid tumors [11].
Nevertheless, our understanding of the infuence and function of PTP1B in liver cancer remains limited. Terefore, more attention should be paid to exploring the role of PTP1B in liver cancer. In this study, the TIMER and UALCAN online tools were frst used to investigate the diferential mRNA expression of PTP1B (PTPN1) at a pancancer level. We found that PTPN1 was overexpressed in various cancers, including liver cancer. Meanwhile, the upregulation of PTPN1 in liver cancer was validated in the ICGC database and the HPA database, respectively. Te results indicated the oncogenic role of PTPN1 in the development of liver cancer. Previous research has shown that PTPN1 increased in liver cancer tissues or liver cancer cell lines [24,25], which was consistent with our analysis from public databases. CD44  TNFRSF9  LAG3  CD200  CD40  NRP1  CD40LG  CD276  CD86  HHLA2  CD48  CD160  TNFSF4  CD274  TNFSF18  TNFRSF8  CD80  CD244  TNFSF9  CD70  ADORA2A  IDO1  VTCN1  HAVCR2  CD27  TNFRSF14  ICOSLG  CTLA4  ICOS  CD200R1  LAIR1  LGALS9  CD28  TNFSF15  TIGIT  BTLA  TNFRSF4  TNFRSF18  PDCD1  PDCD1LG2  BTNL2  TNFRSF25   TCGA   Gene CD160  CD200  CD200R1  CD27  CD274  CD28  CD40  CD40LG  CD44  CD48  CD80  CD86  CTLA4  HAVCR2  ICOS  IDO1  LAIR1  LGALS9  NRP1  PDCD1  PDCD1LG2  TIGIT  TMIGD2  TNFRSF4  TNFRSF8  TNFRSF9  TNFSF15  TNFSF18   ICGC   Gene  Next, we explore the relationship between PTPN1 expression and clinical characteristics of liver cancer. Te results demonstrated that PTPN1 expression was enriched in patients with advanced stage in the TCGA and the ICGC databases, which further illustrated the oncogenic role of PTPN1 in liver cancer. Since PTPN1 was considered a potential oncogenic gene related to patients with advanced liver cancer, we also wonder whether PTPN1 has a prognostic signifcance for liver cancer. Kaplan-Meier analysis showed that the higher expression of PTPN1 was connected to a shorter prognosis time (OS) of liver cancer patients. Univariate and multivariate Cox regression analysis then revealed PTPN1 as an independent prognostic factor with a high risk of liver cancer. Overall, these results suggested that the up-regulated expression of PTPN1 was associated with advanced tumor stage and poor prognosis of patients with liver cancer.
Inspired by the reported role of PTPN1 as an immune checkpoint [11], gene function and pathway analysis were performed to explore whether PTPN1 was involved in immune response processes in liver cancer. Herein, three strategies were adopted for more credible results. First, the GO and KEGG analyses based on DEGs were conducted to annotate the PTPN1-related biological functions and pathways. Te majority of enriched GO terms and KEGG signaling pathways closely related to immune response and cell adhesion were observed in the TCGA database and the ICGC database, which mainly included immune receptor activity, chemokine binding, cytokine-cytokine receptor interaction, T cell receptor signaling pathway, ECM-receptor interaction, and proteoglycans in cancer. However, this traditional analysis using overlap statistics merely focuses on the DEGs' enriched ontology terms or pathways [26].  Compared to traditional GO/KEGG analyses, GSEA comprehensively considered all the genes rather than DEGs determined by certain threshold values [26]. Tus, GSEA as an additional method was performed for functional and pathway analysis. Te GSEA results signifcantly showed that upregulated PTPN1 was involved in the cytokinecytokine receptor interaction in the TCGA and ICGC databases. Subsequently, we may speculate that the cytokine-cytokine receptor interaction acts as a critical mechanism in PTPN1-regulated immune responses in liver cancer.
To deepen our understanding of the mechanism of PTPN1 in liver cancer and verify the above results, we also carried out GO and KEGG analyses based on PTPN1correlated genes. Te correlation genes from four sources (Supplement Table 1) were adopted for analysis. In the TCGA and ICGC databases, PTPN1 interactive genes were observed participating in many T cell-related immune functions and pathways, such as regulation of T cell activation, cytokine-cytokine receptor interaction, and T cell receptor signaling pathway. Interestingly, the enrichment results also included the cytokine-cytokine receptor interaction, consistent with our present analysis. Together, we found strong evidence that PTPN1 may function as an immune regulator that afects the functions of T cells infltrated in liver cancer. Te closely relevant pathway was the cytokine-cytokine receptor interaction. Numerous research have underscored the linkage between cytokine and T cell function [27,28]. Impressively, Wiede et al. illustrated that the expression of PTPN1 would dampen cytokine signaling through attenuating JAK, Tyk2, and STAT5 phosphorylation, thereby inhibiting the expansion and activation of T cells [11]. A similar PTPN1-regulated mechanism was also observed in lymphoid malignancies [29] and Hodgkin lymphoma [30]. It is worth noting that our analysis was interpretable, consistent with the results of molecular experiments to a certain extent. Besides, the results also indicated the feasibility and accuracy of integrating various methods to annotate the enrichment pathway for a single gene. Liver cancer is known to be a multistep development where various signal pathways are altered, such as VEGF receptor signaling, Ras-MAPK signaling and Wnt/β-Catenin pathway [31]. Herein, we defne an additional pathway in liver cancer: the cytokine-cytokine receptor interaction closely associated with PTPN1.
To further investigate the immune-related role of PTPN1 in liver cancer, we frst analyzed the correlation between PTPN1 expression and 47 types of immune checkpoints in the TCGA and ICGC databases. Te results showed that most immune checkpoints were increased in the PTPN1 upregulated groups, and most immune checkpoints were positively linked with PTPN1 expression. Among the top 10 PTPN1-correlated immune checkpoints in the TCGA and ICGC databases, the insertion checkpoints including CD28, CD80, and CD86 functioned as T cell regulators [32]. Te results suggested that PTPN1 might have a similar expression pattern and immune response efect on T cells as these checkpoints. Next, the TIMER online tools explored six types of immune infltration cells (B cell, CD4+ T cell, CD8+ T cell, macrophage, neutrophil, and dendritic cell) in liver cancer. Te correlation between gene markers of nine immune cells (B cell, T cell, CD4+ T cell, CD8+ T cell, Treg cell, T cell, macrophage, dendritic cell, and NK cell) and PTPN1 was also analyzed. Te results showed a close positive association between PTPN1 and B cell, T cell, CD4+ T cell, CD8+ T cell, Treg cell, macrophage, and dendritic cell. As expected, the infltration levels or gene markers of distinct T cell subtypes were correlated with PTPN1 expression, further indicating the relationship between PTPN1 and T cell function. Besides, previous research has illustrated the essential role of PTPN1 for T cell activation in mice [33], which undoubtedly supported our conclusion.
Single-cell RNA-seq analysis was further implemented to identify the PTPN1-expressed cell type in the complex tumor microenvironment of liver cancer, which also served as a verifcation for the above immune analysis. Te characteristic profle of liver cancer was obtained through feature reduction, cell clustering, and cell type annotation on GSE146115. Tumor cells, B cells, T cells, macrophages, and NK cells were identifed in GSE146115, and PTPN1 was enriched in T cells. Subsequently, we analyzed the T cell population and classifed it into CD4+ and CD8+ T cells, and they were PTPN1-expressed cells. Considering that the expression of PTPN1 and T cell markers may be afected by the diferent cell states, pseudotime analysis on these T cells was then conducted, and it was found that three distinct diferentiation states existed. Te coexpression pattern of PTPN1 and T cell exhaustion-related genes (CD27, CTLA4, and LAG3) was observed, independent of cell states. Te enrichment of PTPN1 in T cells was then validated with the GSE98638 dataset. Using the TISCH online tool, we obtained a comprehensive landscape of T cells infltrating liver cancer. It was also found that PTPN1 was highly expressed in T cells, including conventional CD4+ T cells, CD8+ T cells, regulatory T cells, exhausted CD8+ T cells, and proliferating Tcells. Tus, it might be speculated that PTPN1 was elevated in liver cancer infltrating T cells and appeared to be associated with T cell exhaustion.
Te present fndings provided insights into understanding the role of PTP1B in liver cancer, but some limitations still existed. First, we explored mRNA expression levels of PTP1B instead of protein levels because there is a lack of protein omics data matched with adequate clinical information for integrated analysis, just like many bioinformatics analyses on a single gene [13-15, 34, 35]. In this research, the immunohistochemical staining results from the HPA database were used to verify the protein expression level of PTPN1. Second, we speculated that PTPN1 might regulate T cell responses mainly through cytokine-cytokine receptor interaction by pathway enrichment analysis. However, the specifc mechanism of how PTPN1 regulates and afects the function of T cell needs more profound molecular research. Tird, our analysis was performed using public databases and bioinformatics methods. Terefore, further in vitro and in vivo research on PTPN1 are needed to verify these computational results, which will make the results more convincing and advance this work.

Conclusions
Our study showed that the expression of PTPN1 was increased in advanced liver cancer, leading to a worse prognosis for patients. Moreover, PTPN1 was closely associated with some immune response pathways, immune checkpoints, and immune infltration cells. Notably, T cells were enriched in PTPN1, some of which showed exhaustion characteristics, indicating the inhibition efect of PTPN1 on T cell function. Taken together, these results indicated that PTPN1 (PTP1B) is a promising immunotherapy target associated with T cell function for liver cancer.

Data Availability
Te mRNA expression profles of liver cancer and corresponding clinical data were obtained from the TCGA database (https://portal.gdc.cancer.gov/) and ICGC database