This article is mentioned in:

Introduction

Hepatocellular carcinoma (HCC) is the most prevalent subtype of liver cancer and mostly develops from hepatocytes. From 2006 to 2009, the age-standardized mortality rate of HCC in the United States showed an annual percentage change of 4.1%, followed by a slower yet sustained increase of 1.8% from 2009 to 2022, demonstrating a consistent upward trend in annual mortality, which is projected to continue rising through 2040. Furthermore, alcohol-related liver disease (ALD) is anticipated to become the leading cause of HCC mortality by 2026(1). A recent 2024 study from India indicates that the incidence, prevalence and mortality rates of HCC are higher in males. Specifically, the lowest crude incidence rate and age-standardized incidence rate are 1.6 and 1.9 for males, whereas they were 1.1 and 1.3 for females, respectively. However, the annual rate of change is more significant in females, with a rate of 0.72 (95% CI, 0.23-1.31), compared with 0.50 (95% CI, 0.11-0.99) in males. The incidence of HCC associated with hepatitis B virus (HBV) is also decreasing, whilst the incidence of HCC linked to alcohol consumption and metabolic dysfunction-associated steatotic liver disease is increasing. A comparative analysis of risk factors contributing to disability-adjusted life years and HCC-related mortality indicates that alcohol consumption is the most significant factor, followed by drug abuse and smoking, highlighting the growing impact of these lifestyle-related factors on HCC trends (2). HCC malignancy is marked by its high mortality rates. An analysis of data from the World Health Organization (WHO) mortality database, covering 112 countries across five continents, revealed a significant increase in global liver disease-related mortality rates. The age-standardized mortality rate rose from 103.4 per 1,000,000 individuals (95% CI, 88.16–118.74) in 1990 to 173.0 per 1,000,000 individuals (95% CI, 155.15–190.95) in 2021. Projections suggest that, despite population aging and growth being key driving factors, liver disease-related mortality rates are expected to decline by 2050(3). The pathogenesis of HCC is multifactorial, with prominent reported contributors including chronic hepatitis B or C virus infections, a spectrum of liver diseases, including alcoholic liver disease, non-alcoholic fatty liver disease and cirrhosis, exposure to aflatoxins and specific genetic predispositions (4-6). Current therapeutic options encompass surgical resections, liver transplantation, ablation techniques, chemoembolization and targeted systemic therapies (such as Sorafenib and Lenvatinib) (7-9). However, a retrospective study conducted in Spain in 2024 reported that 58% patients with HCC experienced recurrence after liver resection, with 35% of these recurrences classified as aggressive recurrences (10). Due to the aggressive nature of HCC and its tendency for recurrence, the overall prognosis for patients remains poor. Additionally, a previous study analyzing HCC data over the past 10 years found that the 1-year overall survival rate of fibrolamellar HCC was higher compared with conventional HCC, but no significant differences were observed at 3 and 5 years (11). Therefore, understanding the molecular and genetic underpinnings of HCC remains of high importance to identify potential biomarkers and therapeutic targets. Traditional biomarkers and pathways have been extensively studied. α-fetoprotein is a commonly used biomarker for liver cancer diagnosis, whereas key signaling pathways, such as Wnt/β-catenin and PI3K/AKT, are involved in the proliferation, survival and metastasis of liver cancer cells (12,13). Therefore, to improve the clinical outcomes for patients with HCC, it is imperative to discover novel diagnostic markers, creative treatment approaches and enhanced prognostic tools.

Crystallin αB (CRYAB) is a small heat shock protein that primarily functions as a molecular chaperone to protect cells from stress-induced damage (14,15). Such function has garnered attention in the cancer research field to explore its roles in tumorigenesis. Huang et al (16) previously revealed that CRYAB has a role in facilitating macrophage M2 polarization through the AKT1/mTOR signaling pathway to mediate an anti-inflammatory role in liver ischemia-reperfusion injury. In a previous review, Zhang et al (17) reported that CRYAB can inhibit apoptosis by interacting with pro-apoptotic proteins and regulating various signaling pathways, including PI3K/AKT, Raf/MEK/ERK and ERK1/2/FOS-related antigen-1/Slug signaling pathways. CRYAB also induces epithelial-mesenchymal transition (EMT) through the ERK1/2/Fra-1/Slug signaling pathway, promoting HCC progression and associating with poor prognosis (17).

Nuclear factor IA (NFIA) is a transcription factor that has been reported to be involved in a diverse range of cellular processes, including development, differentiation and response to environmental signals, such as extracellular matrix changes and specific signaling molecules within the tumor microenvironment (18). Its aberrant expression and function have been associated with various types of cancers (19). In particular, the role of NFIA in glioblastoma is multifaceted, since it can not only directly promote the proliferation and survival of tumor cells, but it also suppress their apoptosis by interacting with NF κB p65 to form a positive feedback loop (20). In esophageal squamous cell carcinoma (ESCC), NFIA may promote tumor progression downstream of microRNA (miR)-29a regulation. In this axis, NFIA downregulation reduces the activity of the Notch signaling pathway, particularly by lowering the expression of hairy and enhancer of split-1, which in turn enhances the proliferation and migration of ESCC cells (21). In uroepithelial bladder cancers, NFIA expression has been found to be decreased with increasing tumor stage and grade, especially in muscle-invasive cancers, which may indicate its inhibitory role in tumor aggressiveness (22).

However, to the best of our knowledge, the specific association of NFIA and CRYAB with HCC has not been comprehensively explored in the literature. Investigating these proteins may reveal novel molecular pathways. In addition, understanding the expression patterns and function of NFIA and CRYAB may lead to the discovery of novel biomarkers for early detection and prognostic prediction, in addition to novel therapeutic targets. Researching these molecules may provide deeper mechanistic insights into the molecular biology of HCC, aiding in the improvement of diagnosis, prognosis and treatment.

As HCC is a major health concern, there is escalating interest in unveiling its intricate molecular underpinnings. Therefore, in the present study, the roles and regulatory mechanisms of NFIA and CRYAB in HCC were assessed by investigating their gene expression profile, prognostic implications and intracellular functionality. This type of exploration aimed to enhance the understanding of its potential impact on HCC progression and treatment pathways.

Materials and methods

Data sources and differential expression analysis

The analysis in the present study was conducted using the following two main datasets: Gene Expression Omnibus (GEO; https://ww.ncbinlm.nih.gov/geo/) database for GSE113996 (23,24) dataset and The Cancer Genome Atlas (TCGA; https://tcga-data.nci.nih.gov/tcga) dataset for liver hepatocellular carcinoma (LIHC) (25). They offer extensive, well-annotated datasets that facilitate the robust bioinformatics analysis of HCC.

TCGA dataset consisted of 371 LIHC samples and 50 control samples, whilst the GSE113996 dataset included 20 cancer samples (liver cancer tissues) and 20 control samples (adjacent non-tumor tissues). Gene expression profiles were normalized and analyzed for differentially expressed gene (DEG) expression in TCGA-LIHC and GSE113996 datasets using the ‘limma’ package (26) in R software (version 3.4.1; https://www.r-project.org). To identify the differentially expressed genes (DEGs), Log2 transformation was applied to fold change (FC) values, whereas log10 transformation to P values was used for visualization purposes in the volcano plot. The criteria used were FC <0.77 considered to be downregulated DEGs whereas FC >1.3 was considered to be upregulated. Significance was determined using P<0.05.

Enrichment analysis of overlapping genes

To identify the overlapping genes among DEGs in TCGA and GSE113996 datasets, the ‘VennDiagram’ package (27) from the R software was used. This package can effectively identify overlapping DEGs across multiple datasets, thereby providing a clear visual representation of these overlaps. Subsequently, for the analysis of these overlapping genes, Biological Process (BP) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were conducted using the ‘clusterProfiler’ R package (version 3.14.0; Bioconductor: https://bioconductor.org/packages/release/bioc/html/clusterProfiler.html) (28). Adjusted P<0.05 was established as the significance criterion for enriched terms.

Protein-protein interaction (PPI) and survival analysis of overlapping genes

To investigate the functional relationships among the overlapping genes, a PPI analysis was conducted using the Search Tool for the Retrieval of Interacting Genes (STRING; version 11.0; https://string-db.org/) database, leveraging its comprehensive repository of known and predicted protein interactions to elucidate the functional associations among these genes. The resulting PPI network was then visualized using the Cytoscape software (version 3.9.0; http://www.cytoscape.org) (29,30). Subsequently, Kaplan-Meier (KM) survival analysis was conducted on the overlapping genes to analyze the impact of high or reduced expression of genes on the overall survival (OS) probability of patients with LIHC. For each gene, high and low expression levels were defined based on the median expression value within the dataset, with samples above the median classified as ‘high expression’ and those below as ‘low expression’. The log-rank test was used to evaluate the statistical significance of the survival curves (31).

Identification of signature prognostic genes in the risk model

To identify key genes associated with LIHC prognosis, Least Absolute Shrinkage and Selection Operator (LASSO) regression analysis was performed using the ‘glmnet’ package in R software (version 4.1.2; https://cran.r-project.org/package=glmnet), which employed a 10-fold cross-validation to ascertain the optimal λ value (32). In this study, the ‘optimal λ value’ was defined as the value that minimized the mean squared error (MSE) across the folds during cross-validation, balancing bias and variance to enhance model stability and prevent overfitting. This approach ensured the selection of the coefficients that can most accurately predict patient outcomes in the TCGA-LIHC dataset. TCGA-LIHC samples were divided into high-risk groups and low-risk groups based on the expression of the selected genes in LIHC, with risk scores calculated using the ‘survival’ package (33) in R (version 2.41.3; https://cran.r-project.org/package=survival). The survival time and risk score of each sample were then compared, in addition to the expression patterns of the characteristic genes. KM survival analysis was conducted utilizing the log-rank test to identify survival differences between the high-risk and low-risk groups. Additionally, to determine the robustness and accuracy of the predictive model, a receiver operating characteristic (ROC) analysis was performed using the ‘survivalROC’ package in R (Version 1.0.3; https://CRAN.R-project.org/package=survivalROC) (34). The predictive power of the model was quantified by calculating the area under the ROC curve (AUC). AUC values close to 1 would indicate optimal predictive ability. P<0.05 were set as the threshold for statistical significance in all analyses.

Cell culture

Human liver cancer cell lines (MHCC97H, Hep3B, Huh7 and HepG2) were selected for the present study, which were procured from BeNa Culture Collection; Beijing Beina Chunglian Institute of Biotechnology. The huma liver immortalized cell line THLE-2, which was utilized as the control in the present study, were sourced from the Institute of Virology, Chinese Academy of Medical Sciences (Beijing, China) (35).

All of the cell lines were grown in DMEM (Gibco; Thermo Fisher Scientific, Inc.), which was supplemented with 1% penicillin-streptomycin and 10% FBS (Gibco; Thermo Fisher Scientific, Inc.). The cells were cultivated in a humidified atmosphere with 5% CO2 at 37˚C under carefully monitored circumstances.

Transfection

For the manipulation of gene expression in MHCC97H and Huh7 cell lines, specific transfection assays were performed. Expression of the CRYAB gene was knocked down using small interfering RNA (si)-CRYAB (Shanghai GenePharma Co., Ltd.), with a non-targeting siRNA (si-NC; Shanghai GenePharma Co., Ltd.) serving as the negative control. Simultaneously, the NFIA gene was knocked down using si-NFIA and overexpressed using an overexpression vector [pcDNA3.1 (+), Invitrogen; Thermo Fisher Scientific, Inc.]. An empty pcDNA3.1(+) vector was used as the control for overexpression studies. All transfections were conducted using Lipofectamine 3000 (Invitrogen; Thermo Fisher Scientific, Inc.) (36) according to the manufacturer's protocols. For each transfection reaction, 50 nM siRNA or 2 µg of plasmid DNA was used per well in a 6-well plate. The transfection mixture was incubated at 37˚C for 24 h. After transfection, cells were allowed to recover in complete medium for an additional 24 h at 37˚C before subsequent experimentation. Sequence of siRNA targeting NFIA (si-NFIA): 5'-GCCGUGAAGGAUGAAUUGCUA-3'. Sequence of siRNA targeting CRYAB (si-CRYAB): 5'-GCAGGCCCAAAUUAUCAAGCTT-3'. Sequence of non-targeting control siRNA (si-NC): 5'-GCUUCGCCGCCGCCGUAGUUA-3'.

Reverse transcription-quantitative PCR (RT-qPCR)

The expression of CRYAB in liver cancer cells (MHCC97H, Hep3B, Huh7 and HepG2) and liver immortalized cells (THLE-2) was assessed using RT-qPCR. Following the manufacturer's protocol, total RNA was extracted from cells using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc.). Reverse transcription was conducted using the PrimeScript™ RT Reagent Kit (Takara Bio, Inc.), with the following temperature protocol: 37˚C for 15 min for reverse transcription, followed by 85˚C for 5 sec for enzyme inactivation. For qPCR, the specific primer sequences used were as follows: CRYAB forward (F), 5'-AACGGCCTGGGTGGATAGAAG-3' and reverse (R), 5'-CAGTACTCACTGAGCTGCTCTT-3'. For GAPDH, which served as the internal control, the primer sequences F, 5'-GCACCGTCAAGGCTGAGAAC-3' and R, 5'-TGGTGAAGACGCCAGTGGA-3'. Using the SYBR™ Green Universal Master Mix (cat. no. 4309155; Applied Biosystems; Thermo Fisher Scientific, Inc.) and the ABI 7500 Real-Time PCR System (Applied Biosystems; Thermo Fisher Scientific, Inc.), RT-qPCR was performed. The thermocycling protocol for qPCR was as follows: Initial denaturation at 95˚C for 10 min, followed by 40 cycles of 95˚C for 15 sec and 60˚C for 1 min. CRYAB expression was normalized to GAPDH as the internal standard, which was then quantified using the 2-ΔΔCq approach (37).

Western blotting (WB)

Following the manufacturer's protocols, proteins were extracted from liver cancer cells (MHCC97H, Hep3B, Huh7 and HepG2) and huma liver immortalized cells (THLE-2) using RIPA buffer (Beyotime Institute of Biotechnology). The protein concentration was quantified using the BCA protein assay kit (Thermo Fisher Scientific, Inc.). Proteins (20-30 µg) were separated by SDS-PAGE on a 10-12% gel and transferred onto PVDF membranes (MilliporeSigma). PVDF membranes were then blocked with 5% non-fat dry milk in TBS-0.1% Tween-20 (TBS-T) for 1 h at room temperature. Membranes were incubated overnight at 4˚C with the primary antibodies: CRYAB at a dilution of 1:1,000 (cat. no. ab76467; Abcam); NFIA at a dilution of 1:1,000 (cat. no. ab228897; Abcam) or GAPDH (cat. no. 2118L; Cell Signaling Technology, Inc.) at a dilution of 1:1,000. After washing three times with TBS-T, membranes were incubated with HRP-conjugated secondary antibodies at a dilution of 1:2,000 (cat. no. ab6721; Abcam) for 1 h at room temperature. An enhanced chemiluminescence substrate (Pierce™ ECL Western Blotting Substrate; Thermo Fisher Scientific, Inc.) was used to observe protein bands. Band intensities were quantified with ImageJ software (version 1.48v; National Institutes of Health) for accurate comparison.

Cell viability assay

Following transfection, cell proliferation was evaluated using the Cell Counting Kit-8 (CCK-8) from Dojindo Molecular Technologies, Inc. Cells were seeded at a density of 2x103 cells per well into 96-well plates and cultured at 37˚C for 1, 2, 3 and 4 days. At each time point, 10 µl CCK-8 reagent was added into each well and incubated for 2 h at 37˚C in a humidified incubator with 5% CO2. The absorbance was measured at 450 nm using a microplate reader (Kehua Technologies, Inc.).

JASPAR database

The JASPAR database (https://jaspar.genereg.net/) (38) was utilized to identify potential transcription factor binding sites within the CRYAB promoter region. JASPAR, an open-access database, offers a comprehensive collection of curated, non-redundant transcription factor binding profiles sourced from various species (39). By inputting the specific DNA sequence of the CRYAB promoter region (Gene ID: 1410; positions 344-353) into the JASPAR human-specific database and setting the relative profile score threshold to 80%, potential transcription factors that can bind to this region were predicted.

Luciferase activity assay

To assess the transcriptional activity at the binding sites, a luciferase activity assay was conducted. Reporter constructs based on the pGL3-Basic vector (Promega Corporation) were used to examine this activity. Specifically, 293T cells, either with or without overexpression/knockdown of NFIA (untreated cells served as the control group), were co-transfected with pGL3-CRYAB wild-type (WT) or pGL3-CRYAB mutant (MUT) plasmids. All transfections were conducted using Lipofectamine 3000 (Invitrogen; Thermo Fisher Scientific, Inc.). For each transfection, 1 µg plasmid DNA was used per well. The transfection reaction was incubated at 37˚C for 6 h. The luciferase activity in the cells was measured 24 h after transfection using the Dual-Luciferase Reporter Assay System (Promega Corporation). The assays were conducted in accordance with the manufacturer's recommended protocol. Firefly luciferase activity was normalized to Renilla luciferase activity to minimize the effects of variations in experimental conditions. To measure the luciferase activity a luminometer (PerkinElmer, Inc.) was utilized.

TIMER database analysis

The TIMER database (Version 1.0 1; https://cistrome.shinyapps.io/timer/) (40) was used to study the expression profile of NFIA in various cancers. This database provides a comprehensive platform, primarily providing transcriptomic data from TCGA project, enabling the analysis of gene expression across various cancer types. In the present study, the TIMER database (Version 1.0.1; https://cistrome.shinyapps.io/timer/) was used to analyze the expression of NFIA in different cancer types. Specifically, the expression data of the NFIA gene were queried and set the default search criteria and thresholds. The TIMER platform demonstrated the expression levels of NFIA in various cancers by providing transcriptome data from TCGA project. The differential expression of NFIA in LIHC and its corresponding normal tissues were specifically focused upon to investigate its potential role in liver cancer.

Transwell assay

Transwell chambers (Corning, Inc.) with inserts with a pore size of 8.0 µm were used for the invasion and migration assays. For the migration assay, MHCC97H and Huh7 cells were transfected with siRNAs targeting CRYAB, NFIA, both CRYAB and NFIA or a negative control (si-NC) and then seeded into the upper chamber without any coating at an absolute cell density of 1x105 cells per chamber, suspended in serum-free DMEM. The lower chamber was supplemented with culture medium containing 10% FBS to serve as a chemoattractant. For the invasion assay, the upper chamber was pre-treated with Matrigel (cat. no. 354234; Corning, Inc.) at a concentration of 1 mg/ml and the coating was performed at 37˚C for 30 min. to replicate the cellular barriers encountered during invasion. After an incubation period of 24 h at 37˚C, cells that did not undergo migration or invasion were delicately eliminated using a cotton swab. Cells that had successfully migrated or invaded to the lower surface of the membrane were then fixed using 4% paraformaldehyde for 20 min at room temperature. Subsequently, these cells were subjected to DAPI staining at a final concentration of 1 µg/ml for visualizing the nuclei, with staining performed at room temperature for 10 min. The number of migrated or invaded cells was then counted in five randomly selected fields of view per chamber under a fluorescence microscope, facilitating the comparison between different treatment groups.

Statistical analysis

The statistical analysis was conducted using R software, version 3.6.0, released in 2019, in conjunction with the ‘limma’ package of the same version for the normalization and analysis of gene expression profiles. The ‘VennDiagram’ package (version 1.6.13) and the ‘clusterProfiler’ package (version 3.14.0) were used for enrichment analysis.

For in vitro experimental data, GraphPad Prism (version 7.0; Dotmatics), was utilized to perform the necessary statistical tests and generate graphical representations. Data from the experiments were presented as mean ± SD and subjected to an unpaired t-test for two-group comparisons or a one-way ANOVA for comparisons among multiple independent groups. Tukey's test was used as a post hoc analysis for ANOVA results. P<0.05 was considered to indicate a statistically significant difference. Each experiment was repeated at least three times to ensure accuracy and reproducibility.

Results

Identification and enrichment analysis of 42 overlapping genes

From the GSE113996 and TCGA-LIHC datasets, DEGs were identified. Specifically, TCGA dataset revealed 9,078 upregulated DEGs and 1,084 downregulated DEGs (Fig. 1A). By contrast, the GSE113996 dataset showed 60 upregulated and 124 downregulated DEGs (Fig. 1B). An overlap analysis of DEGs from both datasets yielded 42 overlapping genes, including 36 overlapping upregulated DEGs and six overlapping downregulated DEGs (Fig. 1C). Utilizing the ‘ClusterProfiler’ tool for functional pathway enrichment analysis, these 42 genes were found to serve a significant role in various BPs, notably including ‘cholesterol biosynthesis (GO: 0006695)’, ‘sterol biosynthetic process (GO: 0016126)’, ‘regulation of lipid metabolic process (GO: 0018216)’, and ‘regulation of primary metabolic process (GO: 0080090)’ (Fig. 1D). KEGG analysis further indicated their enrichment in pathways including ‘glycerolipid metabolism’, ‘peroxisome proliferator-activated receptor (PPAR) signaling’ and ‘vitamin B6 metabolism’ (Fig. 1E).

Figure 1 Comprehensive analysis of differentially expressed genes and pathway enrichment. Volcano maps of DEGs based on the (A) TCGA and (B) GSE113996 datasets. Gray scatter points represent genes with insignificant differences in expression, whilst orange and green scatter points represent genes with significant differences in expression. X-axis shows log2 fold-change and Y-axis shows -log10 P-value. (C) Venn diagram showing the overlapping DEGs between TCGA-LIHC and GSE113996 datasets. Chord diagram showing enrichment analysis of (D) BP and (E) KEGG. The right ring of the outermost circle shows 10 BP terms, each term is identified by a color. The left ring of the outermost circle shows the gene column, the gradient color indicates the gene differential expression fold change. Inner circles show associations between genes and terms. BP, biological process; DEGs, differentially expressing genes; TCGA, The Cancer Genome Atlas; up, upregulated; down, downregulated.

PPI network and survival analysis of the overlapping genes

Using the Cytoscape software, a PPI network analysis of the 42 overlapping genes was performed, yielding a network of 36 nodes and 45 edges (Fig. 2A). A KM survival analysis on these genes was then performed after categorizing the samples into high-risk and low-risk cohorts, with samples above the median risk score classified as high-risk and those below as low-risk. Notably, nine genes displayed significant associations with OS (Fig. 2B-J). Elevated expression of aldo-keto reductase family 1, member B10 (AKR1B10), aldo-keto reductase 1B15 (AKR1B15), CRYAB, dihydrolipoamide S-acetyltransferase (DLAT), stanniocalcin 2 (STC2), heat shock protein family B member 8 (HSPB8), NAD(P)-dependent steroid dehydrogenase-like (NSDHL) and squalene epoxidase (SQLE) were found to associate with reduced OS, whilst increased IFITM1 expression showed a superior OS. These data suggest the prognostic importance of these genes in HCC.

Figure 2 PPI network analysis and prognostic significance of overlapping genes in liver cancer. (A) PPI network representation of 42 overlapping genes. Nodes represent proteins or protein domains, while edges represent interactions between these proteins, with edge thickness representing the confidence of the interaction. Kaplan-Meier survival curves for the identified prognostically significant genes: (B) AKR1B10, (C) AKR1B15, (D) CRYAB, (E) DLAT, (F) HSPB8, (G) IFITM1, (H) NSDHL, (I) SQLE and (J) STC2. The x-axis represents survival time in months and the y-axis represents survival probability. HR represents the risk of an event in the high-risk group compared to the low-risk group. Log-rank P-value indicates the statistical significance of the difference in survival between high-risk and low-risk groups. PPI, protein-protein interaction; HR, hazard ratio; AKR1B10, aldo-keto reductase family 1, member B10; AKR1B15, aldo-keto reductase 1B15; CRYAB, Crystallin αB; DLAT, dihydrolipoamide S-acetyltransferase; HSPB8, heat shock protein family B member 8; IFITM1, interferon induced transmembrane protein 1; NSDHL, NAD(P)-dependent steroid dehydrogenase-like; SQLE, squalene epoxidase; STC2, stanniocalcin 2.

Prognostic risk analysis and identification of signature prognostic genes

The Lasso regression model was then parameterized, which were used to develop a risk score model to predict patient survival, leveraging the expression and regression analyses of the prognostic genes (Fig. 3A and B). In total, seven significant genes were identified at λmin=0.019. This λmin value was selected based on 10-fold cross-validation within the Lasso regression model, aiming to minimize the MSE and avoid overfitting. Utilizing the risk score formula of riskscore=0.0125 x CRYAB + 0.1125 x AKR1B15 + 0.0019 x HSPB8 + 0.1906 x STC2 + 0.2135 x NSDHL + 0.1357 x DLAT + (-0.0349) x IFITM1, risk scores were assigned based on the gene expression levels, which were then used to categorize the patients into stratified groups of either high or low risk using the median risk score as the cut-off. Fig. 3C displayed the risk score distribution between these groups, revealing both the survival status and duration for the different risk classifications. Notably, a greater number of fatalities occurred in the high-risk group, where seven significant prognostic genes emerged. These genes, displayed in the heatmap in Fig. 3C, exhibit distinct expression patterns between high-risk and low-risk groups. Specifically, IFITM1 is highly expressed in the low-risk group, but shows low expression in the high-risk group. By contrast, the remaining six genes (CRYAB, AKR1B15, HSPB8, STC2, NSDHL and DLAT) are highly expressed in the high-risk group and show low expression in the low-risk group. Survival analysis highlighted a markedly reduced OS in the high-risk group compared with that in its low-risk counterpart (Fig. 3D). Furthermore, ROC curve analysis verified the viable predictive ability of the present model on the 1-year survival rate, with the highest AUC value of 0.76 (Fig. 3E).

Figure 3 Risk prediction model analysis and characterization of prognostic genes. (A) LASSO regression analysis highlighting the coefficients of the eight prognosis-associated genes against the L1 Norm. (B) Cross-validation utilized for tuning parameter determination in the LASSO regression model. The x-axis represents the log(λ) value and the y-axis indicates partial likelihood deviance. (C) Risk score distribution of patients with liver cancer, segregated into low-risk and high-risk categories. The upper scatterplot elucidates the association between risk scores and patient survival status and duration. The lower plot shows a heatmap of z-scores of normalized expression levels of the prognostic genes across the risk stratifications. The x-axis in the upper scatterplot shows ‘patient’ identifiers, each representing an individual's risk score and survival data. In the lower heatmap below, the x-axis shows the ‘genes’ involved in liver cancer prognosis, with their expression levels depicted across the different risk categories. (D) Kaplan-Meier survival analysis contrasting the overall survival between low- and high-risk cohorts. The x-axis chronicles the time in years, whereas the y-axis plots the overall survival probability. (E) Receiver operating characteristic curves demonstrating the predictive performance of the prognostic signature at 1-year, 3-year, and 5-year intervals. HR, hazard ratio; AUC, area under the curve; LASSO, Least Absolute Shrinkage and Selection Operator; NSDHL, NAD(P)-dependent steroid dehydrogenase-like; IFITM1, interferon induced transmembrane protein 1; DLAT, dihydrolipoamide S-acetyltransferase; STC2, stanniocalcin 2; HSPB8, heat shock protein family B member 8; AKR1B15, aldo-keto reductase 1B15; CRYAB, crystallin αB.

Inhibition of HCC cell proliferation, migration and invasion after CRYAB knockdown

CRYAB expression in the liver cancer cell lines and human liver cell line THLE-2 was first measured using RT-qPCR and WB (Fig. 4A and B). Elevated CRYAB expression was found in the liver cancer cell lines compared with that in THLE-2 cells, especially in MHCC97H and Huh7 cells. After knocking down CRYAB expression, both RT-qPCR and WB analysis showed that CRYAB expression was significantly reduced compared with that in the si-NC group (Fig. 4C and D). Further functional assays demonstrated the role of CRYAB in HCC cells. CCK-8 assay revealed a significant decrease in cell proliferation after CRYAB knockdown compared with that in the si-NC group (Fig. 4E and F). In addition, Transwell assays showed that CRYAB knockdown significantly inhibited cell migration and invasion compared with that in the si-NC group in both MHCC97H and Huh7 cell lines (Fig. 4G-J). Notably, CRYAB promoted HCC cell proliferation, migration, and invasion. These observations suggest that CRYAB knockdown not only inhibited its expression, but also significantly inhibited liver cancer cell invasion.

Figure 4 Regulation of liver cancer cell proliferation, migration and invasion by CRYAB knockdown. (A) RT-qPCR and (B) WB analysis of CRYAB expression in THLE-2 cells and 4 liver cancer cell lines (HepG2, MHCC97H, Huh7 and Hep3B). *P<0.05 vs. THLE-2. (C) RT-qPCR and (D) WB were used to evaluate the efficiency of CRYAB knockdown in MHCC97H and Huh7 cells. *P<0.05 vs. THLE-2. Cell Counting Kit-8 assay of the effect of CRYAB knockdown on (E) MHCC97H and (F) cell proliferation. *P<0.05 vs. si-NC. Transwell assay to assess the effect of CRYAB knockdown on migration and invasion of (G) MHCC97H cells and (H) were semi-quantified (scale bars, 50 µm). Transwell assay to assess the effect of CRYAB knockdown on migration and invasion of (I) Huh7 cells and (J) were semi-quantified. *P<0.05 vs. si-NC. CRYAB, Crystallin αB; RT-qPCR, reverse transcription-quantitative PCR; WB, western blotting; si, small interfering; NC, negative control; OD, optical density.

NFIA regulates CRYAB transcriptional activity in HCC cells

Utilizing the TIMER database to evaluate NFIA expression levels across diverse types of cancers, the differential expression patterns of this gene among the various malignancies were assessed. Notably, compared with that in the control samples, NFIA demonstrated significantly reduced expression in LIHC tumor specimens (Fig. 5A). To further investigate the functional relevance of NFIA, its knockdown and overexpression were performed in MHCC97H and Huh7 cells. WB analysis revealed that upon NFIA knockdown, there was a marked decrease in the NFIA protein expression levels but those of CRYAB were markedly increased. This inverse correlation suggests a repressive regulatory effect of NFIA on CRYAB expression. However, when NFIA was overexpressed, a marked decline in CRYAB protein expression was evident (Fig. 5B). Furthermore, using the JASPAR software, potential NFIA binding sites were identified within the CRYAB promoter region at positions 344-353 relative to the transcription start site (TSS), pinpointing a promising NFIA binding site for CRYAB transcriptional initiation (Fig. 5C). To ascertain the implications of NFIA binding to this CRYAB promoter, luciferase reporter gene assays were next conducted. Luciferase reporter vectors pGL3-CRYAB-WT and pGL3-CRYAB-MUT, which mirrored the wild-type and mutant CRYAB binding sites, respectively, were constructed (Fig. 5C). NFIA overexpression significantly amplified the luciferase activity at the CRYAB binding site, whilst NFIA knockdown significantly mediated the opposite effects. This supports the regulatory role of NFIA in suppressing CRYAB transcription. These aforementioned effects did not occur on the pGL3-CRYAB-MUT sequences. These findings suggest the pivotal role of NFIA in modulating CRYAB expression in liver cancer cells by directly interacting with the CRYAB promoter.

Figure 5 NFIA activates CRYAB gene expression by binding to its promoter region. (A) Expression level analysis of NFIA in pan-cancer using TIMER database. *P<0.05, **P<0.01 and ***P<0.001. (B) Western blotting analysis of NFIA and CRYAB expression in liver cell lines after NFIA overexpression (using NFIA overexpression plasmid, labeled as ‘over-NFIA’) and knockdown. ‘Vector’ refers to the control plasmid used in the overexpression experiments, whilst ‘si-NC’ represents the negative control for siRNA knockdown experiments. (C) Upper figure shows JASPAR software detecting potential NFIA binding sites in the upstream genome sequence of the CRYAB coding region, lower figure shows the structure diagram of reporter genes based on the pGL3-Basic vector in luciferase detection. Dual-luciferase reporter assay in 293T cells measuring changes in wild-type and mutant luciferase activities after NFIA (D) overexpression or (E) knockdown. *P<0.05 vs. over-NC or si-NC. TPM, transcript per million; CRYAB, Crystallin αB; NFIA, nuclear factor IA; WT, wild-type; MUT, mutant; over-, overexpression plasmid; si-small interfering; NC, negative control.

Interaction between NFIA and CRYAB in regulating liver cancer cell proliferation

To elucidate the functional roles of NFIA and CRYAB in modulating liver cancer cell proliferation, invasion and migration, CCK-8 and Transwell assays were performed across different experimental groups (Fig. 6A-F). Upon the knockdown of NFIA expression, the liver cancer cells exhibited a marked enhancement in their proliferative, migratory and invasive capacities compared with those in the si-NC group. By contrast, concomitant CRYAB knockdown expression led to a significant reversal in these aforementioned effects mediated by NFIA alone, thereby suggesting that NFIA likely acts as a suppressor of tumor cell proliferation, migration and invasion, while CRYAB may serve to counterbalance this effect. This suggests the potential roles NFIA and CRYAB serve in the cellular dynamics and malignancy of HCC.

Figure 6 Role of NFIA and CRYAB in liver cancer cell proliferation, migration and invasion. Cell Counting Kit-8 was used to measure (A) MHCC97H and (B) Huh7 cell proliferation after the knockdown of NFIA and/or CRYAB. *P<0.05 vs. si-NC. Transwell assay was used to measure cell migration and invasion after combined knockdown of NFIA and/or CRYAB in (C) MHCC97H cells, (D) which was semi-quantified. Transwell assay was used to measure cell migration and invasion after combined knockdown of NFIA and/or CRYAB in (E) Huh7 cells, (F) which was semi-quantified. Scale bars, 50 µm. *P<0.05 vs. si-NC. #P<0.05 vs. si-NFIA. CRYAB, Crystallin αB; NFIA, nuclear factor IA; si, small interferin; NC, negative control.

Discussion

HCC represents a significant global health challenge. Despite the availability of various diagnostic tools (such as ultrasound, CT scan and MRI) and tumor marker analyses (such as α-fetoprotein levels), early and precise detection of this cancer remains elusive due to their limited sensitivity and specificity (41-44). The emergence of certain molecular biomarkers holds potential for advancements in the diagnostic accuracy of HCC. Yang et al (45) previously found that DEAD-box 20 (DDX20) helicase is overexpressed in HCC, which associated with poorer survival (45). DDX20 affects the tumor microenvironment by interacting with miR-324-5p, affecting immune cell dynamics and promoting macrophage differentiation. This interaction also highlights the high dependence on EGFR signaling, that is, EGFR activation supports DDX20-mediated tumor promotion. A previous study showed that the reduction of runt-related transcription factor 1 (RUNX1) expression can promote HCC carcinogenesis. Specifically, the reduction of RUNX1 has been shown to promote EMT by reducing E-cadherin expression and increasing that of vimentin and MMP2, thereby enhancing tumor invasiveness and metastatic potential. In addition, the reduction of RUNX1 lead to reduced angiogenesis by downregulating VEGF expression, impairing the ability of tumors to form new blood vessels necessary for growth and survival in the tumor microenvironment (46). Other studies have indicated RUNX1 and its association with the upregulation of collagen 4A1 (COL4A1), thereby stimulating HCC cell proliferation, emphasizing its promise as a molecular candidate for HCC treatment (47,48). Therefore, exploring and understanding these molecular markers of HCC remain in high demand, not only for enhancing diagnostic precision but also for paving the way for targeted therapeutic strategies for HCC management.

From the analysis of both GSE113996 and TCGA-LIHC datasets, DEGs that were significantly associated with essential BPs were identified. Notably, the regulation of cholesterol biosynthesis and lipid biosynthesis were found to be key pathways enriched by these genes. This is in line with findings on apolipoprotein H (APOH) in HCC, emphasizing the role of lipid biosynthesis in tumor progression, especially in HBV-related cases. APOH can regulate lipid metabolism and the tumor microenvironment by promoting macrophage infiltration, contributing to HCC progression through altered lipid biosynthesis and microenvironment remodeling (49). Apart from providing crucial membrane components for rapid tumor cell division and proliferation, this pathway also produces a range of signaling molecules, including sphingomyelin and phosphatidylinositol oxides (50). Additionally, the involvement of DEGs in the PPAR signaling pathway is consistent with the findings of Lv et al (51), who revealed the potential of PPARγ as a favorable prognostic indicator in bladder cancer. Prognostic analysis, particularly in the context of complex diseases such as HCC, is paramount for tailoring personalized treatment plans and predicting patient outcomes. Among the seven prognostic markers identified by survival analysis (IFITM1, DLAT, NSDHL, STC2, HSPB8, AKR1B15 and CRYAB), CRYAB was selected as the hub gene for further study. Previous reports have shown that CRYAB can serve a role in other cancers, such as glioblastoma, breast cancer and lung cancer (17,52), underscoring its potential significance in oncology. However, its specific role in HCC remains unclear, which is the focus of the present study.

CRYAB is a protein that has been investigated in previous cancer studies. In colorectal cancer (CRC), CRYAB has been demonstrated to facilitate the formation and maintenance of cancer stem cells through the Wnt/β-catenin signaling pathway, thereby enhancing their self-renewal and metastatic potential, underscoring its significance as a molecular target for CRC therapeutic strategies (53). Additionally, owing to its role in immune cell infiltration, CRYAB has been documented to serve as a tumor suppressor in CRC, offering potential as a predictive biomarker for tumor progression and patient prognosis (54). In bladder cancer, CRYAB serves a suppressive role in cell migration, invasion through the PI3K/AKT and ERK signaling pathways (55). In addition, the CRYAB C-802G G allele has been associated with an increased risk of breast cancer, highlighting its potential as an early detection marker in a clinical setting (56). In the present study, CRYAB was found to promote HCC cell proliferation, migration and invasion, suggesting its important role in HCC pathogenesis.

NFIA is garnering interest due to its implications in a diverse range of cancer types, such as glioma, breast cancer and CRC, highlighting its potential as a broad-spectrum factor in oncological research and treatment strategies. In ovarian cancer, Sry-box transcription factor 9 has been reported to promote metastasis by upregulating NFIA and activating the Wnt/β-catenin signaling pathway, suggesting a potential role for NFIA as a therapeutic target and diagnostic marker (57). Additionally, NFIA has been associated with the progression of CRC, as it is regulated by the long intergenic non-coding RNA 00511/miR-29c-3p axis, where LINC00511 serves as a competing endogenous RNA (ceRNA) to upregulate NFIA expression by sponging miR-29c-3p, thereby promoting cell proliferation, metastasis and stemness in CRC tissues and cells (58). In non-small cell lung cancer (NSCLC), NFIA has been noted to augment radiosensitivity by reducing of AKT and ERK phosphorylation, providing a potential avenue for enhancing the efficacy of radiotherapy for patients with NSCLC (59). According to results from the present study, NFIA was found to exert a modulatory effect on CRYAB expression in HCC cells by directly binding to the CRYAB promoter. Furthermore, the interplay between NFIA and CRYAB significantly influenced the regulation of HCC cell proliferation, migration and invasion capacities. Specifically, simultaneous knockdown of NFIA and CRYAB reduced the enhanced cell activity observed when NFIA was knocked down alone, highlighting the role of CRYAB in mediating the effects of NFIA on HCC progression. This is consistent with the findings of Zhu et al (60), where NFIA was found to directly bind to the CRYAB promoter in prostate cancer. In CRC, elevated CRYAB expression was previously found promote tumorigenesis and metastasis by enhancing cell proliferation, migration and invasion capabilities (61). This is similar to the findings in the present study, where knockdown of CRYAB in liver cancer cells led to a significant reduction in cell proliferation, migration and invasion. Although NFIA expression was associated with tumor characteristics, it was not significantly associated with patient survival according to another previous study, suggesting that it may not be an independent prognostic factor (22). Collectively, findings from the present study shed light on the complex interplay between CRYAB and NFIA in liver cancer, offering novel insights into their potential roles as prognostic indicators and therapeutic targets.

The present in vitro study implicates NFIA and CRYAB in HCC progression, suggesting their potential as targets for novel tailored therapies. However, a number of limitations persist in the present study. First, the limited use of distinct cell lines and the number of biological replicates may impact the robustness of our findings and further studies with greater experimental diversity and sample sizes are needed to confirm these observations. Second, potential confounding variables, such as slight variations in cell culture conditions, differences in transfection efficiency and inherent biological variability among HCC cell lines, could influence the observed effects of NFIA and CRYAB modulation, despite our efforts to standardize experimental procedures. Additionally, the in vitro setup of the present study may not fully reflect tumor microenvironment complexity in vivo. In vivo studies are required to confirm the present results for broader genomic and proteomic studies to uncover additional pathways in HCC. Additionally, the influence of dietary components, such as resveratrol, on gene expression and the impact of pharmacological interventions on gene regulation must be considered to develop a comprehensive understanding of HCC's molecular landscape. The present findings point to several paths for future research. Clinical trials are needed to assess the efficacy of therapies that target NFIA and CRYAB. Understanding factors that affect NFIA and CRYAB expression may lead to personalized medicine and the development of biomarkers to predict patient responses to targeted therapies is a priority. In summary, the present study established a foundation for understanding the roles of NFIA and CRYAB in liver but also indicates the necessity for further research to bring these insights into clinical application.

From the GSE113996 and TCGA-LIHC datasets, 42 overlapping genes involved in fundamental metabolic processes lipid production and cholesterol. Among these, nine genes (AKR1B10, AKR1B15, CRYAB, DLAT, HSPB8, IFITM1, NSDHL, SQLE and STC2) showed a strong association with OS in HCC. Specifically, CRYAB promoted HCC cell proliferation, migration and invasion. In addition, NFIA was found to modulate CRYAB expression by binding to its promoter, and luciferase reporter assay results showed reduced activity after NFIA knockdown. NFIA knockdown significantly affected HCC cell dynamics, promoting cell proliferation, invasion and migration. These effects were attenuated by simultaneous CRYAB knockdown, highlighting the regulatory role of CRYAB in mediating the effects of NFIA on HCC cell behavior. Results from the present study suggest the significance of the NFIA and CRYAB interplay in HCC progression, providing potential therapeutic targets for intervention.

Acknowledgements

Not applicable.

Funding

Funding: The present study was supported by the Minsheng Research Project of Pudong New Area Science and Technology Development Fund (grant no. PKJ2022-Y39), Health Science and Technology Project of Pudong New Area Health Commission (grant no. PW2023A-33), Outstanding Clinical Discipline Project of Shanghai Pudong New Area (grant no. PWYgy2021-09), Famous Doctor of Yunnan ‘Xingdian Talent Support Program’ (grant no. XDYC-MY-2022-0032), Yunnan Fundamental Research Project (grant no. 202201AS070002), Yunnan Science and Technology Commission from Yunnan provincial Science and Technology Department and Kunming Medical University (grant no. 02401AY070001-105), Kunming University of Science and Technology & the First People's Hospital of Yunnan Province Joint Special Project on Medical Research (grant no. KUST-KH2023022Y).

Availability of data and materials

The data generated in the present study may be requested from the corresponding author.

Authors' contributions

YJ, PH, JH and HS took part in the conception and design of the study. YJ, PH took part in the acquisition of data. JH and WG performed the in vitro assays. YJ, PH, YD and WG took part in analysis and interpretation of data. YD and WG took part in statistical analysis. JH and HS took part in the revision of manuscript for important intellectual content. All authors read and approved the final version of the manuscript. YJ and HS confirm the authenticity of all the raw data.

Ethics approval and consent to participate

Not applicable.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References