Introduction

Liver cancer is a significant public health issue and a leading cause of cancer-related deaths. According to global cancer statistics, there were approximately 905,677 reported cases of liver cancer and 830,180 related deaths in 2020, making it the third most common cause of cancer deaths.1 Hepatocellular carcinoma (HCC), as the main type of liver cancers, accounted for about 80% of all the cases.2 Chronic active hepatitis caused by hepatitis B virus (HBV) infection is a major risk factor for HCC,3 and 87.7% of HCC patients have contracted with HBV in Southern China.4 Despite advances in treatment strategies such as surgical resection, liver transplantation, percutaneous ablation and radiation as well as transarterial and systemic therapies, the 5-year survival rate for HCC patients remains discouraging at around 12.1%.5 Additionally, individual patients with the similar clinical characteristics and histopathological types may experience different survival outcomes, when received the same treatment methods. Even patients with early-stage HCC who received hepatectomy still have a variable survival,6 suggesting genetic susceptibility to risk of death (or survival) after surgery, indicating that genetic factors, such as single nucleotide polymorphisms (SNPs), may play a role.7 As a result, there is an urgent need to identify prognostic biomarkers for survival of HBV-HCC in the key genes and pathways to guide personalized management and treatment for HBV-HCC.

Previous studies have shown that Reticulons 3 (RTN3) is downregulated in HCC compared with normal hepatocytes, restraining HCC growth and inducing apoptosis by activating p53.8 Further research suggests that hepatitis B antigen (HBxAg) can bind to p53, leading to the inactivation of p53 function and “gain of function” mutations may occur during tumor progression. This combination is important for the occurrence and development of HCC.9–11 These findings indicate that abnormal activation of the p53 pathway plays a crucial role in the carcinogenesis and progression of HCC. The tumor suppressor protein, p53, is closely related to carcinogenesis and tumorigenesis. It consists of five main structural regions, including a proline-enriched region, a DNA binding region, tetramer structural region, a C-terminal regulatory structural region and a transcriptional activation structural region.12 The p53 could be activated upon different cell stresses, initiating the transcriptional program that plays an essential role in cell cycle arrest, cell apoptosis, cellular senescence, DNA recombination and repair, inhibition of angiogenesis, regulation of metabolism or other cellular responses.13,14 As a transcription factor, p53 is strictly regulated by a rich network of post-translational modifications, such as phosphorylation, acetylation, and ubiquitination. These modifications affect the stability, DNA binding ability, and transcriptional activity of p53, thereby exerting its role in tumor suppression. Mutations in the p53 gene are one of the most common events in cancer, and mutant p53 may completely lose its tumor suppressive function, and may even acquire new oncogenic functions. The roles of mutant p53 in tumor cells include promoting cell proliferation, invasion, metastasis, and treatment resistance in cancer.15–17 The MDM2 protein is a key negative regulator of the p53 pathway. One study found that SNP309 in the MDM2 promoter could increase the affinity of transcription activator Sp1 and promote the expression of MDM2 RNA and protein, leading to subsequent inactivation of the p53 pathway, which was associated with accelerating cancers formation.13 Other key genes in the p53 signaling pathway, such as tumor protein p63 (TP63) and tumor protein p73 (TP73), have also been shown to play important roles in the development of cancers.18

Genome-wide association studies (GWAS) were considered a powerful tool for genetic analysis of complex traits and diseases, including cancers.19 In the post-GWAS era, the method based on the biological pathway was a promising hypothesis-driven approach, which has been used to identify functionally reasonable SNPs related to patient survival.20 A previous study found that the TP73 rs747828 C allele in the p53 signaling pathway was significantly associated with reduced progression‐free survival in colorectal cancer and the functional genetic variants of TP73 mediated the response to chemotherapy in colorectal cancer.21 However, the specific genetic variants in candidate genes that affect the function of the p53 pathway in the development and progression of HCC are still unknown. Considering that the P53 gene plays an important role in many cancers, it is reasonable to assume that naturally occurring genetic variants in the p53 pathway may also influence the development and progression of HCC. However, there are currently no reports on the associations between genetic variants in the p53 pathway genes and survival of HBV-HCC patients based on the method of biological pathway. It is crucial to identify potentially functional SNPs in the p53 pathway genes as predictive biomarkers for survival of HBV-HCC. Therefore, the aim of the current study is to analyze the associations between SNPs in p53 pathway-related genes and overall survival (OS) of patients with HBV-HCC by utilizing a two-stage study design.

Materials and Methods Study Populations

In the current study, patients were eligible if they were pathologically diagnosed with HBV-HCC and had undergone hepatectomy at the Guangxi Medical University Cancer Hospital from July 2007 to December 2017. The specific inclusion and exclusion criteria and basic characteristics of patients have been described in the previous study.22,23 Briefly, peripheral whole blood samples were collected for genotyping, and clinical data such as age, sex, smoking and drinking status, alpha-fetoprotein (AFP) level, cirrhosis, cancer embolus, and Barcelona clinic liver cancer (BCLC) stage were also collected. According to the Chinese Guideline of surgical treatment for HCC patients,24 the operable patients included some patients with BCLC B or C stage. According to the inclusion and exclusion criteria, 866 HBV-HCC patients were ultimately included in the current study and were randomly divided into a discovery dataset and a replication dataset in a 1:1 ratio. Their age varied from 20 to 81 years (median age of 47 years old). All these patients were followed up every three months in the first two years after hepatectomy and followed up every six months in the following years via phone until the death of the patient or March 2020. The endpoint of the current study was overall survival (OS), defined as the time from hepatectomy to death or the last effective follow-up. All the above-mentioned basic characteristics of patients are defined as co-variables to be used for subsequent multivariable analyses. The current study has been approved by the Institutional Review Board of Guangxi Medical University Cancer Hospital (Approval number: LW2023189), and written informed consent has been obtained from all participants.

Genotyping, Imputation, Gene and SNP Selection

The whole-genome DNA from the peripheral blood of each HCC patient was extracted by using blood DNA extraction kit (Concert, Xiamen, China). Genotyping was performed by using the Illumina Infinium® Global Screening Assay genotyping chip (GSA, GSAMD-24v1-0, Illumina, San Diego) at Genenergy Biotechnology (Shanghai, China) using the Illumina iScan System. SNP imputation was performed by a web-based platform Michigan Imputation Server (Minimac3) according to the International 1000 Genomes projects (1000G, Phase 3 v5) by using the Minimac3 software.

We selected the genes involved in the p53 pathway using the keyword “p53” by searching the Molecular Signatures Database (http://software.broadinstitute.org/gsea/msigdb/index.jsp). This database uses gene set enrichment analysis methods to construct a gene set of multiple genes with similar positions or functions from multiple perspectives such as gene location, function, metabolic pathways, and target binding, in order to compensate for the deficiency that a single gene cannot explain disease phenomena. After excluding duplicated genes, genes on the sex chromosome, and genes absent in the human genome reference (hg19), we identified a total of 70 candidate genes for further analysis (Table S1). We extracted relevant SNPs within ± 2kb flanking regions of candidate genes from the Chinese Han population in Beijing (CHB) in the 1000 Genomes Project using Plink (version 1.09) (http://pngu.mgh.harvard.edu/purcell/plink/) for analysis. We screened SNPs that met the following criteria: (1) call rate ≥ 95%, (2) MAF ≥ 5%, and (3) HWE test P ≥ 1.0×10−6. Ultimately, a total of 4698 SNPs from the discovery dataset were available for further analysis.

Functional Prediction and Expression Quantitative Trait Loci (eQTL) Analysis

Furthermore, the linkage disequilibrium (LD) analysis between the significant SNPs was analyzed by using Haploview software with r2 threshold of 0.8, and functional prediction of the significant SNPs was carried out by using various web-based tools, including Regulome DB v2.2 (https://www.regulomedb.org/regulome-search/), 3DSNP v2.0 (https://www.omic.tech/3dsnpv2/), HaploReg v4.2 (https://pubs.broadinstitute.org/mammals/haploreg/haploreg.php) and VannoPortal (http://www.mulinlab.org/vportal). Among those significant SNPs in LD, the predicted functional SNPs will be chosen for further analysis. The correlation between the independently functional SNPs and their respective mRNA expression levels were evaluated by using eQTL analysis. The mRNA expression data from the genotype-tissue expression (GTEx) project, 1000 Genomes Project and TCGA dataset were used. GTEx v8 has 208 normal liver tissues and 607 whole blood cells from donors, and 1000 Genomes has 358 Europeans and 76 Han Chinese in Beijing, China (CHB), were used in this analysis. Additionally, we also compared the mRNA expression levels between tumor and adjacent normal tissues by using the TCGA data from the UALCAN database (https://ualcan.path.uab.edu/analysis.html) as well as the mutation rates in the identified genes in the tumor tissues by using the cBioPortal database (https://www.cbioportal.org/). The Haploview software was used to draw a Manhattan plot.

Statistical Analysis

In the single-locus analysis, we used multivariate Cox proportional hazards regression analysis with adjusted for age, sex, smoking and drinking status, AFP level, cirrhosis, cancer embolus, and BCLC stage to evaluate the associations between the SNPs in 70 genes of the p53 pathway and OS of the HBV-HCC patients in additive genetic models. Univariable and multivariable Cox proportional hazards regression analyses were also used to calculate the hazard ratio (HRs) and their 95% confidence interval (CI) for each of SNPs. Multiple testing corrections were performed by using false-positive report probability (FPRP)25 and the Bayesian false discovery probability (BFDP)26 as recommended. We chose BFDF with a cut-off value of 0.80 and a prior probability of 0.10 for the multiple testing corrections, because many SNPs tested in the single-locus analysis were in LD or in the same block as a result of imputation.26 Replication of the SNPs found in the discovery dataset was conducted by using the replication dataset, and the independent SNPs were identified by using multivariable stepwise Cox regression analysis with the adjustments for age, sex, smoking and drinking status, AFP level, cirrhosis, cancer embolus, and BCLC stage. We used bootstrap method for 1000 times random sampling to verify that the results of SNPs were not accidental.27 Kaplan–Meier curves analysis was also performed to assess survival for the combined risk genotypes. Subgroup analysis was conducted to explore the potential interactions between the combined risk genotypes and each of covariables on the HBV-HCC survival. All these statistical analyses were conducted by using R software (versions 3.1.3 and 4.1.2) and PLINK (version 1.9), and two-sided P < 0.05 was considered statistically significant.

Results Association Between SNPs in p53 Pathway Genes and OS of the HBV-HCC Patients

The current study design is illustrated in Figure 1, a total of HBV-HCC 866 patients were included in the current study and were randomly divided into a discovery dataset and a replication dataset. The specific characteristics of the patients are shown in Table S2. In the discovery dataset, a total of 4698 SNPs (295 genotyped and 4403 imputed SNPs) in 70 genes of the p53 pathway were identified. In the single-locus analysis, the associations between these SNPs and the survival of patients with HBV-HCC were evaluated through multivariable Cox regression analysis with adjustment for the co-variables. After multiple testing correction, 236 SNPs were found to be significantly associated with the OS of patients with HBV-HCC (P < 0.05, BFDP < 0.8). Further replication identified six SNPs that remained statistically noteworthy (P < 0.05, BFDP < 0.8, Table 1 and Figure 2). Among these six SNPs, four were located in CD82, and two were in PMAIP1. The results are summarized in the Manhattan plots (Figure S1).

Table 1 Association Between Six Significant SNPs and OS of HBV-HCC in Multivariable Cox Regression Models for the Single-Locus Analysis

Figure 1 Flow chart of the study design.

Figure 2 Forest plot of the association between six significant SNPs and OS of HBV-HCC in multivariable Cox regression models.

Functional Prediction and LD Analysis

In order to identify potentially functional SNPs among these six identified SNPs, we used bioinformatics tools, including RegulomeDB v2.2, 3DSNP v2.0, HaploReg v4.2 and VannoPortal to explore their potential functions. As shown in Table 2, the four SNPs in the CD82 gene were all located in intron regions. Among them, the RegulomeDB v2.2 rank and 3DSNP v2.0 score of rs7925603 are 2b and 28.45, respectively; and the rs7925603 was located in the enhancer histone marks and DNase hypersensitive sites (IPSC, KID, BLD), and it may affect nine motifs and multiple-histone modifications. The two SNPs in the PMAIP1 both were downstream gene variants, among which the RegulomeDB v2.2 rank and 3DSNP v2.0 scores of rs4396625 were 4 and 85.26, respectively, and the rs4396625 located in the promoter histone marks, enhancer histone marks and DNase hypersensitive sites, and it may affect various histone modifications. In addition, we conducted LD analysis on SNPs in CD82 and PMAIP1. It was found that there was a high LD between CD82 rs744631 and rs7925603, CD82 rs7129118 and rs7925603, and PMAIP1 rs4396625 and rs4396625 (r2 > 0.80, Figure S2). Based on functional prediction and LD analysis, we selected functional SNPs rs7925603 and rs4396625 for further analysis.

Table 2 Functional Annotation of the Six SNPs

Independent SNPs and OS of HBV-HCC in the Whole Dataset

To evaluate the independent effects of two functional SNPs (ie, rs7925603 and rs4396625) on the survival of HBV-HCC patients, a stepwise multivariable Cox regression analysis was conducted using the whole dataset of 866 HBV-HCC. As a result, the CD82 rs7925603 A > G (HR = 1.24, 95% CI = 1.07–1.44 and P = 0.004) and PMAIP1 rs4396625 A > T (HR = 0.78, 95% CI = 0.66–0.92 and P = 0.004) were found to be independently associated with the OS of HBV-related HCC patients in the presence of covariables (ie, age, sex, smoking, drinking, AFP, Cirrhosis, Cancer embolus, and BCLC stage) (Table 3).

Table 3 Stepwise Cox Regression Analysis of Clinical Factors and SNPs in the Whole Dataset

Associations Between CD82 rs7925603 and PMAIP1 rs4396625 Genotypes with HBV-HCC OS

As shown in Table 4, the CD82 rs7925603 G allele was significantly associated with a poor survival for HCC patients, while the PMAIP1 rs4396625 T allele was significantly associated with a better for HCC patients, in a dose–response manner in the additive models (both Ptrend = 0.003). In the dominant genetic model, CD82 rs7925603 AG/GG was associated with a poor survival compared with AA genotypes (HR = 1.27, 95% CI = 1.04–1.54, P = 0.019), while PMAIP1 rs4396625 AT/TT was associated with a better survival in HBV-HCC patients compared with AA genotypes (HR = 0.75, 95% CI = 0.62–0.92, P = 0.006).

Table 4 Associations Between Genotypes of Two Independent SNPs and OS of HBV-HCC

Subsequently, to verify that the results of these two SNPs were not accidental, we used bootstrap method for 1000 times random sampling and calculated HR values by using multivariable Cox regression analysis. The results showed that the HR values followed a normal distribution, and their 95% CI included the HR values of these two SNPs, which indicated that the results were reliable (Figure S3).

To analyze the accumulative effect of the two independent SNPs on survival of HBV-HCC patients, their risk genotypes (rs7925603 AG/GG and rs4396625 AA) were classified into a genetic score, the number of risk genotypes (NRGs), to categorize all HBV-HCC patients into three groups with 0, 1, and 2 NRGs. Multivariate analysis showed that a high NRG was significantly associated with a poor survival in a dose–response manner (Ptrend < 0.001) with adjustments for the covariables (Table 4). Additionally, all the patients were divided into low-risk (0–1 NRG) and high-risk (two NRGs) groups. Compared with the low-risk group, the high-risk group had a significantly poor survival for OS (HR = 1.38, 95% CI = 1.13–1.69, P = 0.002). Kaplan–Meier curves were constructed to represent the associations between risk genotypes and the OS of patients with HBV-HCC (Figure 3).

Figure 3 Kaplan–Meier curves of combined protective genotypes and OS of HBV-HCC in the whole dataset. (A) Kaplan–Meier curves of NRGs; (B) Kaplan–Meier curves of low-risk group (0–1 risk genotype) and high-risk group (2 risk genotypes).

Stratified Analysis Between Risk Genotypes with HBV-HCC OS

In further stratified analysis, we investigated whether the combined risk genotypes interplayed with all the covariables on the survival of HBV-HCC patients in the whole dataset. It was shown that among subgroups of patients aged less than or equal to 47 years, male, never drinking, never smoking, AFP levels less than or equal to 400 ng/mL, liver cirrhosis, no cancer embolus, and BCLC stage, the high-risk group had a poor survival for OS. In addition, we did not find any interaction between the combined risk genotypes and each of these covariables on HBV-HCC survival, except for sex (P = 0.011, Table 5).

Table 5 Stratified Analysis of Combined Risk Genotypes and OS of HBV-HCC

The eQTL Analysis and Differential mRNA Expression Analysis

To explore the association between the two independently functional SNPs and the respective mRNA expression, we performed the eQTL analysis. In the GTEx project dataset, for CD82 rs7925603, it was found that the G allele was significantly associated with lower CD82 mRNA expression levels in whole blood cells (P = 0.034, Figure 4A), while there was no significant association was observed in normal liver tissues (P = 0.246, Figure S4A). For another SNP PMAIP1 rs4396625, the T allele was significantly associated with higher mRNA expression levels in whole blood cells (P < 0.001, Figure 4C) instead of normal liver tissues (P = 0.378, Figure S4D). We also performed the eQTL analysis by using data from the 1000 Genomes Project. However, in both European descendants and CHB population, no significant correlations were found between both CD82 rs7925603 G and PMAIP1 rs4396625 T allele with the mRNA expression levels of corresponding genes (both P > 0.05, Figure S4B, C, E and F). Besides, the two SNPs information were not available in TCGA dataset.

Figure 4 The e-QTL analysis of significant SNPs CD82 rs7925603 A > G and PMAIP1 rs4396625 A > T, and differential mRNA expression analysis of HCC from the UALCAN using the TCGA data. (A) The association between CD82 rs7925603 A > G and mRNA expression levels of CD82 in whole blood samples from the GTEx database; (B) CD82 mRNA expression levels were down-regulated in tumor tissues but not statistically significant from the UALCAN using the TCGA data; (C) The association between PMAIP1 rs4396625 A > T and mRNA expression levels of PMAIP1 in whole blood samples from the GTEx database; (D) PMAIP1 mRNA expression levels were down-regulated in tumor tissues from the UALCAN using the TCGA data.

Subsequently, we further compared the mRNA expression levels of these two genes between HCC tissues and adjacent normal tissues by using the TCGA database. It was found that the expression levels of PMAIP1 mRNA in HCC tissue were significantly lower than in adjacent normal tissues (P = 0.001, Figure 4D). Similar trend was also observed in CD82, but there was no statistically significant (P = 0.444, Figure 4B). In addition, we used the TCGA data to investigate the somatic mutations of CD82 and PMAIP1 in HCC tissues. It was found that the mutation rates of these two genes in HCC tissues were both less than 1.5% (Figure S5), indicating that somatic mutations of these two genes may play a minor role in their expression levels.

Discussion

In the current study, we identified two novel SNPs (ie, CD82 rs7925603 A > G and PMAIP1 rs4396625 A > T) in the p53 pathway genes that were significantly associated with the survival of HBV-HCC patients. Specifically, the CD82 rs7925603 G allele was associated with a poor survival, while the PMAIP1 rs4396625 T allele was associated with a better survival of HBV-HCC patients. Additionally, their combined risk genotypes were associated with poor HBV-HCC survival. Further eQTL analysis revealed that the rs7925603 G allele was significantly correlated with reduced mRNA expression levels of CD82, while the rs4396625 T allele was significantly correlated with elevated mRNA expression levels of PMAIP1. The CD82 rs7925603 A > G and PMAIP1 rs4396625 G > T change may have an effect on its binding to enhancer histone marker, promoter histone marker, motif modification, and histone modifications. Collectively, our study provided some evidence for the associations between genetic variants in the p53 pathway genes and the prognosis of HBV-HCC patients. These potentially functional SNPs may serve as prognostic markers for HBV-HCC, which may be beneficial for guiding personalized management and treatment of HBV-HCC patients.

CD82, also known as KAI1, is situated on the human chromosome 11p11.2. It belongs to the transmembrane 4 superfamily and has been implicated in various biological processes including signal transduction, adhesion, proliferation, migration, movement, protein trafficking, and aggregation.28,29 For example, CD82 functions as a tumor suppressor by inhibiting tumor invasion and metastasis, generally expressed in normal tissues and downregulated in various cancers including HCC.30–33 Noticeably, the CD82 promoter contains a binding motif of transcription factor p53, which activates the CD82 gene at the transcriptional level through binding by p53 to the specific site in its 5’upstream region; thus, the expression of CD82 is positively regulated by p53, but the loss of p53 function may down-regulate the expression of CD82, promoting tumor metastasis and leading to a poor survival of cancers.30,34 Another study found that the HBx protein, a component of HBV, induced CD82 promoter methylation and inhibited CD82 expression at the transcriptional level, accelerating the progression of HCC.35 These findings suggested that CD82 may have inhibit the progression of HBV-HCC. Consistent with previous studies, we found that the mRNA expression levels of CD82 in HCC tissues were higher than that in normal adjacent tissues, and the rs7925603 G allele was associated with lower CD82 mRNA expression levels.

Another p53 pathway-related gene, PMAIP1 identified by the rs4396625 SNP, is located on chromosome 18q21.32 and belongs to the pro-apoptotic subfamily of the B-cell lymphoma-2 (Bcl-2) protein family. It only contains the BCL-2 homology domain 3 (BH3-only) and is mainly located in mitochondria, regulating apoptosis,36 and it is a type of programmed cell death, which is essential for organogenesis during embryonic development and tissue homeostasis in the adults.37PMAIP1 mainly antagonizes the activity of the anti-apoptotic protein Mcl-1 in the Bcl-2 family, promoting the activation of BAX/BAK protein on the surface of mitochondrial membrane, thereby exerting a pro-apoptotic effect.38,39 Multiple studies have shown that PMAIP1 can be activated and regulated by p53-dependent pathways, playing an important role in cell apoptosis, the occurrence and development of various malignant tumors, and their treatment.39–41 For instance, low PMAIP1 expression in tumor samples was correlated with a worse survival in a clinical trial for relapsed/refractory B-cell lymphoma using tandem CD19/20 CAR T-cells.42 In liver cancer cells,43PMAIP1 specifically induced cell apoptosis and inhibited the migration and invasion of HepG2 cells, while having minimal effect on the survival of normal liver HL-7702 cells. Here, our results indicated that the PMAIP1 rs4396625 T allele was associated with higher PMAIP1 mRNA expression levels, and the PMAIP1 mRNA expression levels were significantly lower in HCC tissues compared with adjacent normal tissues in TCGA dataset. Furthermore, we also found that the PMAIP1 rs4396625 T allele is located in promoter and enhancer histone marks, and it may affect the modifications of various histones, such as H3K27ac, H3K4me1, H3K79me2, etc. All these provide strong biological evidence to support our findings of the association between the PMAIP1 rs4396625 SNP and survival of patients with HBV-HCC.

However, there were several limitations in the current study. Firstly, the above findings may not be generalizable to other populations, because we have enrolled participants from southern China only. Secondly, the sample size of the study population was limited, which did not provide us opportunity to examine subgroup effects. Thirdly, we cannot exclude selection bias of the patients as they all originated from a single institution, where some clinical information was not made available, such as treatment and detailed testing results of hepatitis B infection. Fourthly, this study was unable to collect more comprehensive clinical data, such as detailed treatment information of patients. This study failed to adjust for these confounders, which may influence the associations observed. Finally, the exact molecular mechanisms of the identified independent SNPs that may influence survival of HBV-HCC patients remain unclear. Therefore, additional comprehensive mechanistic studies are needed to validate our findings.

Conclusion

The current study has identified two independent SNPs that were significantly associated with the OS of HBV-HCC. These novel SNPs were predicted to have potentially biological functions in modifying targeted gene expression. Our results provide a rationale for further exploring the roles of survival-related SNPs in p53 pathway genes in the progression of HBV-HCC.

Data Sharing Statement

The data that support the findings of this study are available on request from the corresponding authors. The data are not publicly available due to privacy or ethical restrictions.

Ethics Statement

The current study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Review Board of the Guangxi Medical University Cancer Hospital (Approval number: LW2023189).

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Acknowledgments

We would like to thank all study participants, researchers and clinicians who contributed to this study.

Funding

The current study was supported by the Key R&D Program of Scientific Research and Technical Development Project of Qingxiu District (2022009); Joint Project on Regional High-Incidence Diseases Research of Guangxi Natural Science Foundation (2023GXNSFBA026091); Guangxi Medical and health appropriate Technology opening and popularization and application project (S2022112); Youth Program of Scientific Research Foundation of Guangxi Medical University Cancer Hospital (YQJ2022-5); Guangxi University Young and middle-aged Teachers’ Basic Research Ability Improvement Project (2021KY0112); Key Laboratory of Early Prevention and Treatment for Regional High-Frequency Tumor, Ministry of Education (GKE-KF202102).

Disclosure

The authors report no conflicts of interest in this work.

References

1. Sung H, Ferlay J, Siegel RL, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71(3):209–249. doi:10.3322/caac.21660

2. Rumgay H, Ferlay J, De Martel C, et al. Global, regional and national burden of primary liver cancer by subtype. Eur J Cancer. 2022;161:108–118. doi:10.1016/j.ejca.2021.11.023

3. Maucort-Boulch D, De Martel C, Franceschi S, et al. Fraction and incidence of liver cancer attributable to hepatitis B and C viruses worldwide. Int, J, Cancer. 2018;142(12):2471–2477. doi:10.1002/ijc.31280

4. Lin J, Zhang H, Yu H, et al. Epidemiological characteristics of primary liver cancer in mainland china from 2003 to 2020: a representative multicenter study. Front Oncol. 2022;12:906778. doi:10.3389/fonc.2022.906778

5. Zeng H, Chen W, Zheng R, et al. Changing cancer survival in China during 2003-15: a pooled analysis of 17 population-based cancer registries. Lancet Glob Health. 2018;6(5):e555–e67. doi:10.1016/s2214-109x(18)30127-x

6. Shindoh J, Hasegawa K, Einou Y, et al. Risk factors of post-operative recurrence and adequate surgical approach to improve long-term outcomes of hepatocellular carcinoma. HPB. 2013;15(1):31–39. doi:10.1111/j.1477-2574.2012.00552.x

7. Liu F, Luo L, Liu Z, et al. A genetic variant in the promoter of CD46 is associated with the risk and prognosis of hepatocellular carcinoma. Mol, Carcinog. 2020;59(11):1243–1255. doi:10.1002/mc.23252

8. Song S, Shi Y, Wu W, et al. Reticulon 3-mediated Chk2/p53 activation suppresses hepatocellular carcinogenesis and is blocked by hepatitis B virus. Gut. 2021;70(11):2159–2171. doi:10.1136/gutjnl-2020-321386

9. Feitelson MA, Zhu M, Duan LX, et al. Hepatitis B x antigen and p53 are associated in vitro and in liver tissues from patients with primary hepatocellular carcinoma. Oncogene. 1993;8(5):1109–1117.

10. Unsal H, Yakicier C, Marçais C, et al. Genetic heterogeneity of hepatocellular carcinoma. Proc Natl Acad Sci U S A. 1994;91(2):822–826. doi:10.1073/pnas.91.2.822

11. Wang XW, Forrester K, Yeh H, et al. Hepatitis B virus X protein inhibits p53 sequence-specific DNA binding, transcriptional activity, and association with transcription factor ERCC3. Proc Natl Acad Sci U S A. 1994;91(6):2230–2234. doi:10.1073/pnas.91.6.2230

12. Harris CC. Structure and function of the p53 tumor suppressor gene: clues for rational cancer therapeutic strategies. J Natl Cancer Inst. 1996;88(20):1442–1455. doi:10.1093/jnci/88.20.1442

13. Bond GL, Hu W, Bond EE, et al. A single nucleotide polymorphism in the MDM2 promoter attenuates the p53 tumor suppressor pathway and accelerates tumor formation in humans. Cell. 2004;119(5):591–602. doi:10.1016/j.cell.2004.11.022

14. Thomas AF, Kelly GL, Strasser A. Of the many cellular responses activated by TP53, which ones are critical for tumour suppression? Cell Death Differ. 2022;29(5):961–971. doi:10.1038/s41418-022-00996-z

15. Liu Y, Su Z, Tavana O, et al. Understanding the complexity of p53 in a new era of tumor suppression. Cancer Cell. 2024;42(6):946–967. doi:10.1016/j.ccell.2024.04.009

16. Hernández BORRERO LJ, El-Deiry WS. Tumor suppressor p53: biology, signaling pathways, and therapeutic targeting. Biochim Biophys Acta Rev Cancer. 2021;1876(1):188556. doi:10.1016/j.bbcan.2021.188556

17. Wang H, Guo M, Wei H, et al. Targeting p53 pathways: mechanisms, structures, and advances in therapy. Signal Transduct Target Ther. 2023;8(1):92. doi:10.1038/s41392-023-01347-1

18. Candi E, Agostini M, Melino G, et al. How the TP53 family proteins TP63 and TP73 contribute to tumorigenesis: regulators and effectors. Hum Mutat. 2014;35(6):702–714. doi:10.1002/humu.22523

19. Tam V, Patel N, Turcotte M, et al. Benefits and limitations of genome-wide association studies. Nat Rev Genet. 2019;20(8):467–484. doi:10.1038/s41576-019-0127-1

20. Gallagher MD, Chen-plotkin AS. The post-GWAS era: from association to function. Am J Hum Genet. 2018;102(5):717–730. doi:10.1016/j.ajhg.2018.04.002

21. Zhang K, Meng Y, Cao X, et al. Genetic variants in p53 signaling pathway genes predict chemotherapy efficacy in colorectal cancer. Cancer Med. 2019;8(7):3428–3436. doi:10.1002/cam4.2215

22. Huang Q, Liu Y, Qiu M, et al. Potentially functional variants of MAP3K14 in the NF-κB signaling pathway genes predict survival of HBV-related hepatocellular carcinoma patients. Front Oncol. 2022;12:990160. doi:10.3389/fonc.2022.990160

23. Lin Q, Qiu M, Wei X, et al. Genetic variants of SOS2, MAP2K1 and RASGRF2 in the RAS pathway genes predict survival of HBV-related hepatocellular carcinoma patients. Arch Toxicol. 2023;97(6):1599–1611. doi:10.1007/s00204-023-03469-5

24. Commission General Office of National Health Commission. General Office of National Health. Standard for diagnosis and treatment of primary liver cancer (2022 edition). J Clin Hepatol. 2022;38(02):288–303.

25. Wacholder S, Chanock S, Garcia-Closas M, et al. Assessing the probability that a positive report is false: an approach for molecular epidemiology studies. J Natl Cancer Inst. 2004;96(6):434–442. doi:10.1093/jnci/djh075

26. Wakefield J. A Bayesian measure of the probability of false discovery in genetic epidemiology studies. Am J Hum Genet. 2007;81(2):208–227. doi:10.1086/519024

27. Bradley Efron. Bootstrap methods: another look at the jackknife. In: Breakthroughs in Statistics: Methodology and Distribution. Springer; 1992:569–593.

28. Zhou L, Yu L, Wu S, et al. Clinicopathological significance of KAI1 expression and epithelial-mesenchymal transition in non-small cell lung cancer. World J Surg Oncol. 2015;13(1):234. doi:10.1186/s12957-015-0657-8

29. Kim JS, Kim HK, Lee J, et al. Inhibition of CD82 improves colitis by increasing NLRP3 deubiquitination by BRCC3. Cell Mol Immunol. 2023;20(2):189–200. doi:10.1038/s41423-022-00971-1

30. Tennis MA, Van Scoyk MM, Freeman SV, et al. Sprouty-4 inhibits transformed cell growth, migration and invasion, and epithelial-mesenchymal transition, and is regulated by Wnt7A through PPARgamma in non-small cell lung cancer. Mol Cancer Res. 2010;8(6):833–843. doi:10.1158/1541-7786.Mcr-09-0400

31. Liu WM, Zhang XA. KAI1/CD82, a tumor metastasis suppressor. Cancer Lett. 2006;240(2):183–194. doi:10.1016/j.canlet.2005.08.018

32. Lee J, Byun HJ, Lee MS, et al. The metastasis suppressor CD82/KAI1 inhibits fibronectin adhesion-induced epithelial-to-mesenchymal transition in prostate cancer cells by repressing the associated integrin signaling. Oncotarget. 2017;8(1):1641–1654. doi:10.18632/oncotarget.13767

33. Yan W, Huang J, Zhang Q, et al. Role of metastasis suppressor KAI1/CD82 in different cancers. J Oncol. 2021;2021:9924473. doi:10.1155/2021/9924473

34. Mashimo T, Watabe M, Hirota S, et al. The expression of the KAI1 gene, a tumor metastasis suppressor, is directly activated by p53. Proc Natl Acad Sci U S A. 1998;95(19):11307–11311. doi:10.1073/pnas.95.19.11307

35. Yu G, Bing Y, Li W, et al. Hepatitis B virus inhibits the expression of CD82 through hypermethylation of its promoter in hepatoma cells. Mol Med Rep. 2014;10(5):2580–2586. doi:10.3892/mmr.2014.2495

36. Oda E, Ohki R, Murasawa H, et al. Noxa, a BH3-only member of the Bcl-2 family and candidate mediator of p53-induced apoptosis. Science. 2000;288(5468):1053–1058. doi:10.1126/science.288.5468.1053

37. Tuzlak S, Kaufmann T, Villunger A. Interrogating the relevance of mitochondrial apoptosis for vertebrate development and postnatal tissue homeostasis. Genes Dev. 2016;30(19):2133–2151. doi:10.1101/gad.289298.116

38. Elkholi R, Floros KV, Chipuk JE. The Role of BH3-only proteins in tumor cell development, signaling, and treatment. Genes Cancer. 2011;2(5):523–537. doi:10.1177/1947601911417177

39. Guikema JE, Amiot M, Eldering E. Exploiting the pro-apoptotic function of NOXA as a therapeutic modality in cancer. Expert Opin Ther Targets. 2017;21(8):767–779. doi:10.1080/14728222.2017.1349754

40. Yang X, Gao Y, Yan Q, et al. Transcriptome analysis of ochratoxin A-induced apoptosis in differentiated caco-2 cells. Toxins. 2019;12(1):23. doi:10.3390/toxins12010023

41. Yu Z, Zhang Y, Shao S, et al. Identification of CDCA2 as a diagnostic and prognostic marker for hepatocellular carcinoma. Front Oncol. 2021;11:755814. doi:10.3389/fonc.2021.755814

42. Yan X, Chen D, Wang Y, et al. Identification of NOXA as a pivotal regulator of resistance to CAR T-cell therapy in B-cell malignancies. Signal Transduct Target Ther. 2022;7(1):98. doi:10.1038/s41392-022-00915-1

43. Zhao T, Zhao C, Lu Y, et al. Noxa and Puma genes regulated by hTERT promoter can mitigate growth and induce apoptosis in hepatocellular carcinoma mouse model. J Cancer. 2022;13(6):2001–2013. doi:10.7150/jca.70282