The gene expression profile data (GSE36376, GSE39791 and GSE60502) were downloaded from the GEO (http://www.ncbi.nlm.nih.gov/geo/geo2r) (24-26). The inclusion criteria for genomics data were as follow: i) The samples were obtained from Homo sapiens; ii) all tissues were classified as HCC or normal tissues; and iii) the sample sizes were >10 per study. GSE36376 included 240 HCC tumor tissues and 193 adjacent non-tumor tissues; GSE39791 included 72 tumor tissues and 72 matched adjacent non-tumor tissues; GSE60502 included 18 tumor tissues and 18 adjacent non-tumor tissues. GEO2R (http://www.ncbi.nlm.nih.gov/geo/geo2r) was used to analyze differentially expressed genes (DEGs) between HCC and non-tumor samples. The cut-off was set as |log2 fold change (FC)|>1 and adjusted P<0.01. TPR expression and clinical data were obtained from TCGA (https://portal.gdc.cancer.gov/) and GTEx (https://gtexportal.org/home/). Given that the data were obtained from the online databases, additional approval from an ethics committee was not required.

Xiantao tool (https://www.xiantao.love/) is a platform using R software (3.6.3) for acquiring data from TCGA and GTEx. TPR expression in HCC tissues, tissues adjacent to carcinoma and normal tissues was compared using Xiantao tool and presented in box, scatter and violin plots. Diagnostic performance of TPR was assessed using receiver-operating characteristic curves performed on Xiantao tool. All the DEGs (|log2FC|>1.5 and adjusted P<0.05) gained from single gene differential analysis through Xiantao tool were presented in volcano plots.

A total of 14 pairs of HCC and matched normal fresh frozen tissues were obtained through hepatectomy at Sun Yat-sen University Cancer Center (Guangzhou, China), from 2019 to 2021 (12 males and 2 females, ages from 20 to 78 years with an average age of 56). The patients were diagnosed according to their clinicopathologic characteristics at the Sun Yat-sen University Cancer Center. No patients had received radiotherapy and/or chemotherapy prior to surgery. Informed consent was obtained from all patients and approved (approval no. B2022-238-02) by the Research Medical Ethics Committee of Sun Yat-sen University Cancer Center.

The liver cancer cell lines (Hep3B, SNU449, MHCC97H, Huh7, HepG2 and HCCLM3) were obtained from the American Type Culture Collection (ATCC) and cultured according to the instructions from the ATCC. All cells were grown in DMEM supplemented with 10% FBS (both from Invitrogen; Thermo Fisher Scientific, Inc.) at 37˚C and 5% CO₂. All cell lines in the present study were authenticated utilizing short-tandem repeat profiling.

The total RNA of the liver cancer cell lines (Hep3B, SNU449, MHCC97H, Huh7, HepG2 and HCCLM3) or tissue was isolated utilizing TRIzol^® reagent (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. First-strand cDNA was synthesized utilizing the Revert Aid^™ First Strand cDNA Synthesis Kit (Fermentas; Thermo Fisher Scientific, Inc.). The thermocycling conditions used were as follows: 37˚C for 15 min, 95˚C for 5 sec and were then cooled to 4˚C upon completion. Quantitative PCR assays were performed using a A28134 QuantStudio^® 5 Real-Time PCR Instrument (Thermo Fisher Scientific, Inc.) and iTaq Universal SYBR Green Supermix reagent (Bio-Rad Laboratories, Inc.). The thermocycling conditions used were as follows: 95˚C for 30 sec, 40 cycles at 95˚C for 3 sec and 60˚C for 30 sec, then 60˚C for 20 sec and 95˚C for 1 sec. The relative expression of each gene was calculated using the 2^-ΔΔCq method with GAPDH as the internal reference (27). The primers used to amplify the indicated genes are shown in Table SI.

The GeneMANIA database (http://www.genemania.org) was used to explore the genes that interacted with TPR. The Search Tool for Interaction Gene/Proteins (STITCH) website (http://stitch.embl.de/) was used to analyze the protein-protein interactions of TPR.

TIMER (https://cistrome.shinyapps.io/timer/) and Xiantao tool were used to analyze the infiltration levels of different immune cells. TIMER was also applied to explore the interrelation between TPR expression and different gene marker sets of immune cells. The correlations were evaluated by purity-correlated partial Spearman's correlation. An integrated repository portal for TISIDB (http://cis.hku.hk/TISIDB/) was used to investigate the relationship between TPR and immunoinhibitors, immunostimulators, chemokines and receptors.

Metascape (http://metascape.org/) and Xiantao tool was used to perform Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis to explore the biological processes (BPs), cellular components (CCs) and molecular functions (MFs) of TPR in HCC. Gene Set Enrichment Analysis (GSEA) was used to probe the potential mechanisms of TPR performed on Xiantao tool. Gene sets with false discovery rate (FDR) <0.25 and P. adjust <0.05 were considered as significant.

The oligonucleotide sequences targeting TPR mRNA were as follows: #1, GCAGCTTGTTGATTCCATA (5'-3') and #2, GGAGCGATCTGAAACAGAA (5'-3') synthesized from Guangzhou RiboBio Co., Ltd. The siRNA negative control (siN0000001-1-5) is designed and produced by Guangzhou RiboBio Co., Ltd. and its sequence is proprietary (28). Transfection was performed according to the manufacturer's instructions using Lipofectamine RNAi MAX transfection reagent (Invitrogen; Thermo Fisher Scientific, Inc.) and 50 nM siRNA. siRNA transfection was performed at 37˚C for 6 h and the culture medium was then replaced with fresh culture medium. Follow-up procedures were performed 24 h later.

Briefly, the liver cancer cell lines Hep3B and SNU449 were collected, lysed in RIPA buffer (150 mM NaCl, 0.5% EDTA, 50 mM Tris, 0.5% NP40, pH=8.0) and centrifuged for 20 min at 13,500 x g and 4˚C. The protein concentration was measured using a BCA Protein Assay Kit (Beijing Solarbio Science & Technology Co., Ltd.). The proteins (20 µg/lane) were separated using SDS-PAGE electrophoresis in 4 to 20% polyacrylamide gels and transferred to PVDF membranes (Invitrogen; Thermo Fisher Scientific, Inc.). The PVDF containing protein membranes were blocked in 5% skim milk at room temperature for 120 min. The membranes were then incubated overnight with primary antibodies against AKT (1:1,000), phosphorylated (p)-AKT (1:1,000) or GAPDH (1:1,000) at 4˚C. Membranes were then washed with TBST with 0.1% Tween-20 and incubated with goat anti-rabbit IgG (H&L) HRP secondary antibodies (1:10,000) at 25˚C for 1 h. The protein bands were then visualized using the ECL chemiluminescence system (Pierce; Thermo Fisher Scientific, Inc.). The antibodies used in the present study are shown in Table SII.

An MTT assay was conducted to measure cell viability. Briefly, SNU449 or Hep3B cells were seeded at a density of 2,000 cells per well in a 96-well microplate. The cells were incubated with MTT at 37˚C for 4 h, then the culture medium was removed, and 200 µl dimethylsulfoxide (DMSO) was added to dissolve the purple formazan. The optical density (OD) was detected at 490 nm with the microplate reader once per day for 5 days. The results are presented as the mean ± SD of three independent experiments.

Data were collected from three independent experiments. The GraphPad Prism 9 (GraphPad Software; Dotmatics) and the SPSS software (version 16.0, SPSS Inc.) were used for data analysis. Data were presented as the mean ± SD. Unpaired Student's t-test was used to analyze differences between groups and one-way analysis of variance with Tukey's post hoc test was used to analyze multiple groups. Survival analysis was performed using the KM method, and differences among the survival curves were analyzed with the log-rank test. The follow-up threshold was 30 months. Wilcoxon rank sum test was used to analyze α-fetoprotein (AFP). Dunn's test was used to analyze the pathological stage. Two-way ANOVA with Dunnett's multiple comparison test was used to analyze the MTT assays. P<0.05 was considered to indicate a statistically significant difference.

The flow diagram presented in Fig. 1 was constructed to reveal the process of the present study. To identify the key genes in the progression of HCC, GSE36376, GSE39791 and GSE60502 datasets were analyzed to explore the DEGs in HCC. As demonstrated by Venn diagram (Fig. 2A), 187 congruous DEGs were identified. Notably, TPR was the only upregulated gene and there were no downregulated genes included in the overlap of the three datasets (Figs. 2B and S1). An additional 186 genes were upregulated or downregulated in two datasets and expressed inversely in the other dataset.

The differential expression of TPR in HCC tissues and normal tissues was then explored. Among the expression of TPR in 33 types of tumors assessed from TCGA and GTEx, 24 types expressed significantly more TPR than corresponding normal tissues, 6 types demonstrated no significant difference and 3 types lacked sufficient samples (Fig. 2C). It was further identified that the expression of TPR was significantly higher in 374 tumor cases compared with 50 normal cases (P<0.001; Fig. 2D) and 50 tumor tissues compared with their matched non-tumor tissues from TCGA (P<0.001; Fig. 2E).

Furthermore, RT-qPCR was used to determine TPR expression in 14 pairs of HCC and peritumoral tissues. The findings demonstrated a higher TPR expression in most tumor tissues (Fig. 2F). Subsequently, TPR expression was confirmed at the mRNA level in liver cancer cell lines (Fig. 2G). Next, the correlation of TPR expression with several types of current HCC therapeutic targets, including programmed cell death protein 1 (PDCD1) (29), vascular endothelial growth factor A (VEGFA) (30), fibroblast growth factor receptor 4 (FGFR4) (31), platelet derived growth factor subunit B (PDGFB) (32) and KIT proto-oncogene (KIT) (33), were investigated (Fig. 2H and L). Notably, there were markedly positive correlations between the expression of TPR and these targets. Collectively, these results demonstrated that TPR was highly expressed in HCC and was associated with several existing therapeutic targets.

It was then determined whether the expression of TPR is associated with the clinicopathologic variables of patients with HCC. A total of 374 patients with HCC from TCGA, including 353 males and 121 females, were divided into two categories according to the median expression of TPR: The low expression group and the high expression group, and the clinical and gene expression data of these patients were subsequently analyzed (Table SIII). The results revealed that TPR expression was significantly associated with histologic grade (P=0.046), age (P=0.042), height (P=0.039) and weight (P=0.034). An association between TPR expression and other clinicopathologic features was not identified. Based on univariate analyses using logistic regression, it was determined that TPR upregulation in HCC was significantly associated with age (P=0.020), histologic grade (P=0.016) and AFP level (P=0.007) (Table I). As revealed in Fig. 3, TPR expression was significantly associated with AFP (P<0.001) and pathological stage (stage I vs. stage III, P=0.019), but not sex, tumor status, tumor (T) stage, adjacent hepatic tissue inflammation, vascular invasion, residual tumor or ethnicity. These results indicated that the expression of TPR was associated with several clinicopathologic variables of patients with HCC.

To evaluate the clinical significance of TPR expression, survival rates between high and low TPR expression levels were compared to determine the association between TPR expression and prognosis. The KM survival analysis demonstrated that there were significant differences in overall survival (OS) relapse-free survival (RFS) and progression-free survival (PFS), between the TPR high and low expression groups (P=0.08; P=0.033; P=0.02, respectively) (Fig. 4A-C). The univariate Cox proportional hazards model was then applied, and T stage, metastasis stage, pathological stage and tumor status were identified as potential prognostic factors with P<0.05 for OS. In addition, after multivariate analyses was applied, the expression of TPR was not identified as an independent risk prognostic factor for OS among patients with HCC (Table II). Based on the risk factors identified in the multivariate analysis, nomograms were developed to predict 1-, 3-5-year OS rates for HCC patients (Fig. 4D). Calibration plots were constructed to evaluate the agreement between the predicted and the actual OS using the prognosis model (Fig. 4E). In addition, the area under curve (AUC) of TPR in HCC was 0.885 (95% CI, 0.852-0.918), which suggested high diagnostic accuracy of TPR in HCC (Fig. 4F). These results indicated that TPR was a valuable marker to predict clinical outcomes, along with other clinical features in patients with HCC.

After dividing the patients with HCC into two groups according to their TPR expression, an analysis of the DEGs from the two groups was performed. Based on the analysis, 676 DEGs were identified, among them, 498 genes were upregulated and 169 genes were downregulated (Fig. 5A). The top 10 most downregulated genes were AC107396.1, SAA2, SAA1, HAMP, SAA2-SAA4, CLEC1B, CLEC4G, ACTBP12, AC091729.2 and OR52E8 (data not shown). In addition, the top 10 most upregulated genes (LGALS14, HMGA2, SLC6A14, SLC6A15, MYO3A, FER1L6, BPIFA1, CAPN6, MIR483 and COL2A1) are shown in the gene expression heatmap in Fig. 5B. Notably, HMGA2, which had been demonstrated to be associated with tumor progression (34-36), was the upregulated gene of significant expression according to the volcano plots (Fig. 5A).

Therefore, the potential biological functions of TPR in HCC were examined. GO and KEGG analyses were performed using the Metascape database, and the top 20 GO enriched pathways are listed in Fig. 5C. It was notable that the GO enriched items included cell differentiation, which is involved in tumor progression. Moreover, a total of 132 BPs, 34 CCs, 37 MFs and 7 KEGG annotations were associated with TPR co-expressed genes, the results of which are shown by bubbly plots (Fig. 5D-G). Next, GSEA between high and low TPR expression data sets were conducted to explore the signaling pathways differentially activated in HCC. Gene sets with false discovery rate (FDR) <0.25 and P.adjust <0.05 were considered as significant. The top five most significant signaling pathways enriched in the high TPR expression phenotype according to their normalized enrichment score (NES) are listed in Fig. 5H, including retinoblastoma gene in cancer (NES=2.801; FDR=0.002; P.adjust=0.003), resolution of sister chromatid cohesion (NES=2.703; FDR=0.002; P.adjust=0.003), mesodermal commitment pathway (NES=2.681; FDR=0.002; P.adjust=0.003), mitotic prometaphase (NES=2.658; FDR=0.002; P.adjust=0.003) and endoderm differentiation (NES=2.650; FDR=0.002; P.adjust=0.003).

Finally, the gene-gene interaction networks of TPR constructed using GeneMania indicated the top 20 most frequently altered genes (Fig. 5I), while the protein-protein interaction networks generated using STITCH presented the top 10 proteins that most interacted with TPR (Fig. 5J; Table III). These results prompted a further investigation into the potential biological functions of TPR in HCC.

GSEA between high and low TPR expression data sets also identified several immune and inflammation-related pathways enriched in the high TPR expression phenotype, including resistin as a regulator of inflammation, immunoregulatory interactions between a lymphoid and a non-lymphoid cell, the inflammatory response pathway, ADORA2B-mediated anti-inflammatory cytokine production, and intestinal immune network for IgA production (data not shown). Given that targeting the tumor microenvironment (TME) is a new treatment strategy for HCC and that immune infiltration is a core component of the TME, the correlation between TPR and tumor immune infiltration was determined to further investigate the effect of TPR on the TME. For this purpose, the associations between TPR expression and the number of different tumors infiltrating immune cells in HCC were analyzed utilizing the Xiantao tool. The results indicated that the expression of TPR was negatively correlated with StromalScore (ρ=-0.121; P=0.020), ESTIMATEScore (ρ=-0.192; P<0.001) and ImmuneScore (ρ=-0.218; P<0.001) (Fig. 6A). The correlations between TPR and infiltrating immune cells are shown in Fig. 6B. Elevated expression of TPR was positively correlated with the infiltration of T helper cells (ρ=0.436), Th2 cells (ρ=0.319) and central memory T cell (Tcm) (ρ=0.276), and negatively correlated with cytotoxic cells (ρ=0.421), dendritic cells (DCs; ρ=-0.382) and plasmacytoid dendritic cells (pDCs; ρ=-0.366) (all P<0.001) (Fig. 6C-H).

A validation study of TPR expression and the diverse immune signature was further conducted to understand the crosstalk between TPR and immune response by TIMER. The gene markers were used to characterize immune cells (Table IV) and functional T cells (Table V). The results revealed that the expression of TPR was significantly associated with most immune and T cell markers after adjusting for tumor purity. As shown in the aforementioned tables, markers of M1 macrophages (IRF5, PTGS2 and NOS2), M2 macrophages (CD163, VSIG4 and MS4A4A), Treg (FOXP3, CCR8 and TGFB1) and exhausted T cells (HAVCR2, CXCL13 and LAYN) were correlated with TPR in HCC (P<0.001). It was suggested TPR could regulate exhaustion and macrophage polarization in HCC. Next, the association between TPR with immunoinhibitors, such as CSF1R (ρ=-0.285; P<0.001), HAVCR2 (ρ=-0.275; P<0.001) and TGFBR1 (ρ=0.133; P=0.0104) (Fig. 7A), and immunostimulators, including CD40 (ρ=-0.296; P<0.001), IL6R (ρ=0.395; P<0.001) and TNFRSF14 (ρ=-0.388; P<0.001) (Fig. 7B), were investigated. In addition, the association between the expression of TPR and chemokines, containing CCL5 (ρ=0.377; P<0.001), CCL14 (ρ=-0.382; P<0.001) and CXCL2 (ρ=-0.324; P<0.001) (Fig. 7C), were explored. Consistently, TPR expression was identified to be significantly associated with chemokine receptors, such as CCR5 (ρ=-0.181; P<0.001), CXCR3 (ρ=-0.22; P<0.001) and CXCR6 (ρ=-0.238; P<0.001) (Fig. 7D). These results demonstrated that TPR was an immunoregulatory factor in HCC and played an integral role in regulating the immune response.

To confirm the bioinformatics results and to verify the effect of TPR on HCC cells, TPR-silenced SNU449 and Hep3B cell lines were constructed (Fig. 8A). It has been reported that regulation of TPR expression affects the phosphorylation activity of the AKT pathway (37). Consistently, the results of WB demonstrated that silencing TPR expression decreased the phosphorylation levels of AKT (Fig. 8B). Next, an MTT assay was conducted to explore the effect of siTPR on the proliferation ability of HCC cells (Fig. 8C). The results demonstrated that after silencing TPR in SNU449 and Hep3B cells, proliferation was significantly inhibited. Therefore, TPR may promote the proliferation of HCC cells through the AKT pathway.