3.1 Expression, prognosis, immunological and mutational analyses of ERCC3 across cancers and its high expression in HCC are strongly associated with poor prognosis

To investigate the mRNA levels of ERCC3 in 33 different tumors and normal tissues (Fig. 1A), the levels of ERCC3 in breast cancer (BLCA), bile duct cancer (CHOL), colon cancer (COAD), head and neck cancer (HNSC), kidney chromophobe (KICH), kidney clear cell carcinoma (KIRC), liver cancer (LIHC), lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), and stomach cancer (STAD) tissues were significantly different from those in normal tissues (P < 0.0001), and ERCC3 was found to be overexpressed in most tumors.

Fig. 1

Expression of ERCC3 in HCC and its relationship with clinicopathological features. A Differential expression of ERCC3 in pancancer tumor tissues and normal tissues in the TCGA database. B Expression of ERCC3 in HCC tissues and normal liver tissues in the TCGA-LIHC cohort. C Expression of ERCC3 in HCC tissues and paracancerous tissues in the TCGA-LIHC cohort. D Expression of ERCC3 in HCC tissues and normal liver tissues in the GSE64041 cohort. E Expression of ERCC3 in HCC tissues and paracancerous tissues in the GSE64041 cohort. K‒M survival curves for OS (F), PFS (G), RFS (H), and DSS (I) in patients with high and low ERCC3 expression in the TCGA-LIHC cohort. (J) Kaplan‒Meier survival curves for OS in patients with high and low ERCC3 expression in the ICGC-LIRI-JP cohort. K ROC curves to evaluate the diagnostic efficacy of ERCC3 in the TCGA-LIHC dataset. Differences in ERCC3 expression were statistically significant in patients stratified by sex (L), tumor stage (M), tumor grade (N), T stage (O), alpha-fetoprotein level (P), and body mass index (Q) (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001)

Bioinformatics analysis of 424 samples from the TCGA-LIHC cohort (Fig. 1B) revealed that ERCC3 mRNA expression was notably greater in HCC tissues (374 patients) than in normal tissues (50 patients), which was consistent with the results of the pancancer analysis. Figure 1C shows that ERCC3 expression in paracancerous tissues (50 patients) was greater than that in normal tissues (50 patients). The differential expression of ERCC3 between HCC tissues and normal liver tissues was further validated in the GSE64041 cohort (Fig. 1D, E).

K‒M survival curves were generated to assess the relationship between ERCC3 gene expression and the prognosis of patients with HCC (Fig. 1F–I), and the results revealed four prognostic indicators, OS (HR = 1.82, 95% CI: 1.26–2.63, P = 0.0012), PFS (HR = 1.78, 95% CI: 1.32–2.4, P = 0.00012), RFS (HR = 1.72, 95% CI: 1.24–2.4, P = 0.0011), and DSS (HR = 2, 95% CI: 1.25–3.2, P = 0.0033). Survival of HCC patients with high ERCC3 expression was significantly lower than that of patients with low ERCC3 expression, and there was a negative correlation between high expression of ERCC3 and good prognosis of patients with hepatocellular carcinoma.

In the ICGC-LIRI-JP dataset, the correlation between ERCC3 expression and patient survival in the TCGA-LIHC dataset was further validated, and the study demonstrated that the survival rate of patients with high ERCC3 expression was notably lower than that of patients with low ERCC3 expression (Fig. 1J). In addition, ROC curves were used to explore the diagnostic value of ERCC3 for HCC, and the results of ROC curve analyses of the TCGA-LIHC dataset at 1 year (area under the curve (AUC) = 0.69), 3 years (AUC = 0.61), and 5 years (AUC = 0.59) showed that ERCC3 had a greater prognostic accuracy (Fig. 1K).

To investigate the differences in ERCC3 expression between normal liver tissues and HCC tissues, we collected and collated data on different clinical characteristics recorded in the TCGA-LIHC dataset, including age, sex, T stage, stage, grade, body mass index (BMI) and the alpha-fetoprotein index (AFP). The results showed that ERCC3 expression was closely associated with the clinical parameters of HCC, including sex, T stage, stage, grade, BMI and AFP (Fig. 1L–Q). The high expression of ERCC3 in HCC patients is strongly associated with poor prognosis, suggesting that targeting ERCC3 may serve as a promising therapeutic strategy.

3.2 The ERCC3 expression level can be an independent predictor for HCC patients

To investigate the ability of clinical prognostic characteristics to predict the prognosis of HCC patients, one-way Cox regression analysis was performed, and the results showed that T stage (HR = 2.13, 95% CI: 1.48–3.05), stage (HR = 2.09, 95% CI: 1.43–3.06) and ERCC3 expression (HR = 1.6, 95% CI: 1.2–2.14) were the candidate factors for patient survival (Fig. 2A). The results of multivariate Cox regression analysis indicated that high ERCC3 expression might be an independent predictor of poor prognosis in patients with HCC (HR = 1.47, 95% CI: 1.08 ~ 2.01) (Fig. 2B). Data related to candidate factors were combined to create nomograms with high total scores to predict 1-, 3-, and 5-year OS in TCGA-LIHC patients (Fig. 2C). Based on the nomogram for classifying patients into high- and low-scoring groups, the Kaplan‒Meier survival curve results revealed a markedly poorer prognosis in the high-scoring group (Fig. 2D). To further assess the accuracy of this nomogram, ROC prognostic accuracy analysis was performed. The area under the ROC curve (AUC) values of the nomogram were 0.71 at 1 year, 0.68 at 3 years, and 0.64 at 5 years (Fig. 2E), indicating the high accuracy of the nomogram. Decision curve analysis (Fig. 2F), which assesses the clinical decision value by calculating the area and distance from the null horizontal axis of each clinical feature, showed that the nomogram had better clinical utility than other metrics in predicting the survival of patients with HCC (final curve). The calibration curves at 1, 3, and 5 years showed that the OS outcomes predicted by the nomogram were in good agreement with the actual patient conditions (Fig. 2G-J).

Fig. 2

ERCC3 may serve as an independent predictor of survival in HCC patients. Univariate (A) and multivariate (B) analyses of ERCC3 expression versus clinicopathological characteristics. C Nomogram of ERCC3 expression versus clinicopathological characteristics developed for predicting the probability of OS. D Kaplan‒Meier survival curves for OS in patients in the high and low Nomotu subgroups. E ROC curves for 1-, 3- and 5-year OS for patients stratified according to ERCC3 expression. F Decision curve analysis (DCA) curves for the clinical use of nomotu. G Calibration curves of the nomograms for 1-, 3- and 5-year OS. Local zoom calibration curves for the 1-year (H), 3-year (I) and 5-year (J) OS nomograms

3.3 Relationship between ERCC3 gene expression and the immune microenvironment in HCC

We used the single-sample gene set enrichment analysis (ssGSEA) algorithm to calculate enrichment scores for the predicted pathways of the immune cycling pathway and positive immune checkpoint blockade (ICB)-related signaling pathways. The associations between ERCC3 and the predicted immunotherapy pathways were visualized using the R package ggcor (Fig. 3A). High expression of ERCC3 correlated more significantly with cancer cell antigen release (step 1) (P = 0.19), cancer antigen presentation (step 2) (P = 0.21), and T-cell recognition of cancer cells (step 6) (P = 0.12) in the steps of the immune cycle, showing a negative correlation with some of the immune cell correlations in the steps of immune cell transfer to tumors (step 4). High expression of ERCC3 correlated markedly with most of the included ICB-related signals (Fig. 3B) but poorly with the proteasome system, APM signaling pathway, and SLE pathway, whereas it correlated with the cell cycle, DNA replication, Fanconi anemia pathway, homologous recombination, microRNAs in cancer, mismatch repair, nucleotide excision repair, oocyte meiosis, p53 signaling pathway, progesterone-mediated oocyte maturation, proteasome, spliceosome and viral oncogenic pathways. High ERCC3 expression was associated with the abundance of immune cells and inflammatory indicators in the tumor microenvironment.

Fig. 3

The ERCC3 gene is correlated with immune function. Correlations between the ERCC3 gene and cancer immune cycle steps (A) and the ICB response signaling pathway (B). Correlations between the ERCC3 gene and immune-related functions (C) and the immune cell infiltration score (D) (*P < 0.05, **P < 0.01, ***P < 0.001)

Analysis of ERCC3 gene expression in relation to immune cell type (Fig. 3C) and immune function (Fig. 3D) revealed that ERCC3 was significantly and positively associated with activated dendritic cells (aDCs), natural killer cells (NK cells), and T regulatory cells (Tregs); was notably and positively associated with major histocompatibility complex class I molecules (MHC class I), which are involved in immune function; and was significantly and negatively associated with cytolytic activity. A prolonged state of chronic inflammation may lead to abnormal NK cell function and elevated levels of NK cells along with inhibition of their cytolytic activity.

3.4 Role of ERCC3 gene expression in immunotherapy-treated HCC

Cancer immunotherapy has revolutionized cancer treatment, and the importance of immune infiltration assessment continues to increase with the development of novel immunotherapeutic agents. To analyze the distribution and association of the relative content of tumor-infiltrating immune cells in the TCGA-LIHC cohort, the degree of immune cell infiltration for each sample was determined by the CIBERSORT, EPIC, IPS, QUANTISEQ, TIMER and XCELL algorithms. The low-risk group appeared to have greater levels of immune penetration than did the high-risk group (Supplementary Fig. 1). Immunomodulatory genes (including immune checkpoint genes and immune cell marker genes) play a key role in tumor immunotherapy, and Spearman correlation analysis was performed to reveal the relationship between ERCC3 expression and 43 immunomodulatory genes in HCC. Immunomodulatory gene expression was higher in the high ERCC3 group compared to the low ERCC3 group, and ERCC3 gene expression was proportional to the expression of immunomodulatory genes, with the exception of CD244, CD40, KIR3DCL, LAG3, TMIGD2, and TNFSF14 (Fig. 4A). The age, sex, T stage, tumor stage, tumor grade, body mass index and alpha-fetoprotein index data of patients in the TCGA-LIHC cohort were compared at the top of the heatmap. Spearman rank correlation analysis was performed to assess the significant positive correlation of ERCC3 with common immune checkpoint proteins (PD1, PDL1, PDL2, and CTLA4) (Supplementary Fig. 2A–D). ERCC3 may affect immune function by upregulating immune checkpoints associated with targeted therapies to influence immune function and ultimately promote HCC progression.

Fig. 4

Prediction of immunotherapy outcome in patients with high and low ERCC3 gene expression. A Heatmap of the correlation between ERCC3 expression and the expression of different immunomodulatory genes. B Differences in tumor purity, stromal scores, immune scores and estimated scores between the high and low ERCC3 expression groups. C Differences in TIDE scores, dysfunction scores, exclusion scores, and MSI scores between patients with high and low ERCC3 gene expression. D Correlation between IPS and ERCC3 expression in patients with high and low ERCC3 gene expression. E Expression levels of ERCC3 in different cell types in different HCC datasets (*P < 0.05, **P < 0.01, ***P < 0.001)

The immune score, immunological score (P < 0.05) and estimation score (P < 0.05) were greater in the low-risk group (Fig. 4B), suggesting that the total immune level and immunogenicity of patients in this group were greater. In addition, the tumor cell purity was positively correlated (P < 0.05). To determine the effects of differences in ERCC3 expression on immunotherapy response, we performed tumor immune dysfunction and rejection (TIDE) analysis. According to the results of the study, patients had lower TIDE scores (Fig. 4C) and better response to immunotherapy in the high-risk group. In addition, the IPS helps to screen patients who are sensitive to immunotherapy. In our study, IPS and blocker scores were greater in the low-risk subgroup than in the high-risk subgroup, suggesting that low-risk patients may be more susceptible to treatment with immune checkpoint inhibitors (ICIs) and derive more significant benefits (Fig. 4D).

We used the TISCH2 database to analyze the comparative expression of ERCC3 in different immune cells derived from different HCC datasets. The results showed that ERCC3 expression was greater in CD4Tcon, regulatory T cells (Tregs), Tprolif, CD8T and CD8TEX cells in the LIHC GSE98638 dataset (Fig. 4E).

3.5 Mutation correlation analysis of the ERCC3 high- and low-expression groups in HCC

Assessing representative gene variants in the high- and low-risk groups, the top five mutation frequencies in the high-risk group (Fig. 5A) were TP53 (39%), TTN (30%), CTNNB1 (27%), MUC16 (18%), and OBSCN (14%). The top 5 genes with the highest mutation frequency (Fig. 5B) in the low-risk group were CTNNB1 (27%), TTN (24%), TP53 (21%), ALB (16%), and MUC16 (16%). The most common type of mutation was missense mutation, and the correlation of the top 20 genes in the high-risk and low-risk groups was also visualized, as shown in the following figure (Supplementary Fig. 2E, F), with mutational co-occurrence between some genes (P < 0.05). The significance of TP53, SPEG and LAMA5 differential mutations was greatest in the high-expression group (Fig. 5C).

Fig. 5

ERCC3 expression and mutation correlation analysis. The 20 genes with the highest mutation frequencies in the high-expression group (A) and low-expression group (B). C Forest plot showing different gene mutations in HCC patients in the high- and low-expression groups (*P < 0.05, **P < 0.01, ***P < 0.001)

3.6 High expression of ERCC3 promotes the proliferation and migration of HepG2 cells sensitive to tozasertib and I-BRD

To investigate the effect of ERCC3 overexpression on the cellular phenotype of HCC cell lines, ERCC3-overexpressing cell lines were constructed to infer the biological function of the ERCC3 gene in HCC cell lines. Bioinformatics and Western blot analyses revealed that in all HCC cell lines, the expression of ERCC3 was relatively low in the HepG2 cell line (Supplementary Fig. 3A, B). HepG2 cells with ERCC3 overexpression were used as the experimental group, and the NC group included HepG2 cells infiltrated with LV5-NC viral medium. The functional role of ERCC3 in HepG2 cells was explored. qRT‒PCR and Western blot experiments were performed to determine the efficiency of ERCC3 overexpression (Fig. 6A). Using the RTCA xCELLigence system, it was shown that ERCC3 overexpression promoted HepG2 cell viability (Supplementary Fig. 4, Fig. 6B). A scratch assay demonstrated that wound healing in the ERCC3 overexpression group was markedly faster than that in the control group (Fig. 6C), and a Transwell migration assay was performed to further assess the effect of ERCC3 on the migration ability of the cells (Fig. 6D). The number of HepG2 cells migrating across the polycarbonate membrane to the lower layer in the ERCC3 overexpression group was significantly greater than that in the control group, and ERCC3 overexpression notably improved the migration ability of HCC cells. ERCC3 overexpression significantly reduced the number of cells in the G1 phase, did not markedly change the number of cells in the S phase, and increased the number of cells in the G2 phase compared with that in the control group (Fig. 6E), it indicates that ERCC3 overexpression increases the proportion of cells in G2 phase, the proportion of cells in G1 phase, and more cells enter the active cell cycle state, which promotes cell division and proliferation, thus altering the cell cycle process and promoting cell proliferation, which in turn affects the cell cycle progression.

Fig. 6

ERCC3 gene overexpression promotes HepG2 cell proliferation and migration. A qRT‒PCR and Western blotting were used to determine the efficiency of ERCC3 overexpression. B Cell proliferation curve. C Wound healing assay. D Transwell assays were used to detect the migration ability of ERCC3-overexpressing cells. E Cell cycle analysis (*P < 0.05, **P < 0.01, ***P < 0.001)

Using the CellMiner database, our drug sensitivity analysis revealed that ERCC3 expression was significantly associated with the IC50 of multiple drugs, and among the top 4 drugs with the most significant correlations (Fig. 7A), as ERCC3 expression increased, cancer cell sensitivity to Osimertinib (cor = −0.660), Tozasertib (cor = −0.629), Cediranib (cor = −0.625), and I-BRD9 (cor = −0.605 drugs) increased (P < 0.001).

Fig. 7

RNA-seq analysis revealed an immunomodulatory role for ERCC3. A Relationship between ERCC3 expression and sensitivity to four drugs (osimertinib, tozasertib, cediranib, and I-BRD9). B Differences in the half inhibitory concentrations (IC50s) of osimertinib, tozasertib, cediranib and I-BRD in the ERCC3-overexpressing group and control group. C GO enrichment P value histogram. D KEGG enrichment P value histogram. E Positive regulation of cell proliferation and Focal adhesion pathway codifferentiation genes. F Correlation between RAC2 and ERCC3 expression. G RT‒qPCR validation of the RAC2 gene (****P < 0.0001)

To assess the changes in the sensitivity of HepG2 cells to the drug candidates and identify potential drug targets, drug sensitivity experiments were performed to validate the therapeutic effects of Osimertinib, Tozasertib, Cediranib and I-BRD9 on HCC tumor growth. The results showed that the sensitivity of cancer cells to Tozasertib (NC: 5589 nM, ERCC3 OE: 3476 nM) and I-BRD (NC: 35693 nM, ERCC3 OE: 8919 nM) increased with increasing ERCC3 gene expression (Fig. 7B), and the effect was significant. Osimertinib (NC: 1696 nM, ERCC3 OE: 2707 nM) and Cediranib (NC: 6268 nM, ERCC3 OE: 5242 nM) differed from the results of the raw letter analyses in the presence of a variety of biological factors and individual differences, and more cellular models will need to be utilised to validate the therapeutic effects of the drugs on HCC cells.

3.7 Differential gene function analysis and validation via transcriptome sequencing of ERCC3-overexpressing HepG2 cells

To investigate the molecular mechanisms by which the overexpression of ERCC3 affects HepG2 cells, we performed RNA sequencing on ERCC3-overexpressing HepG2 cells and control HepG2 cells. A differential gene volcano plot (Supplementary Fig. 5) revealed 599 differentially expressed genes (P < 0.05 and |log2FC|> 1), 239 upregulated genes and 360 downregulated genes.

The results of GO enrichment analysis of differentially expressed genes in the ERCC3 overexpression group showed (Fig. 7C, Supplementary Fig. 6A) that ERCC3 overexpression affects the molecular functions of protein kinase binding, receptor binding, growth factor activity, cytokine activity, and other molecules in HCC cells. Effects on cellular components such as the plasma membrane, extracellular region, extracellular space, cell surface, collagen-containing extra-articular matrix, and basolateral plasma membrane. These genes were involved in biological processes such as positive transcriptional regulation of the RNA polymerase II promoter, positive regulation of cell proliferation, coagulation, negative regulation of endopeptidase activity, and cholesterol biosynthetic processes.

Analysis of differential gene KEGG enrichment in the ERCC3 overexpression group showed (Fig. 7D, Supplementary Fig. 6B) that ERCC3 overexpression affects HCC cell mineral uptake, complement and coagulation cascades, osteoblast differentiation, the IL-17 signaling pathway, growth hormone synthesis, secretion and action, and the PPAR signaling pathway organic systems. Regulation of metabolic pathways such as steroid hormone biosynthesis, metabolic effects of cytochrome P450 on exogenous drugs, terpene skeleton biosynthesis, and primary bile acid biosynthesis. It is associated with human T-cell leukemia virus infection, lipid and atherosclerosis, insulin resistance, and bladder cancer epidemic formation. These proteins are involved in environmental information processing, such as the koK signaling pathway, cytokine‒cytokine receptor interactions, the TNF signaling pathway, the AMPK signaling pathway, viral proteins interacting with cytokines and cytokine receptors, and ECM-receptor interactions. Involvement in cellular processes such as adhesion plaque formation and iron death. RNA sequencing revealed the different functions of ERCC3 in HepG2 cell lines, which mainly involve cytokine and signal transduction-related processes, as a pro-carcinogenic gene that promotes cancer cell development by affecting cell proliferation.

qRT‒PCR validation of key genes in the cell proliferation pathway was performed to determine the accuracy of the transcriptome sequencing results. Venn diagram analysis (Fig. 7E) revealed that the RAC2 gene is a codifferential gene involved in the positive regulation of cell proliferation and the focal adhesion pathway, validating the transcriptome sequencing results. Bioinformatics analysis revealed a significant positive correlation between ERCC3 and RAC2 expression (Fig. 7F). The qRT‒PCR results showed that the expression of the RAC2 gene was markedly upregulated in the ERCC3-overexpressing HepG2 cell line (Fig. 7G).