Gene expression data were sourced from GSM2343352 and GSM2343354. The analysis and visualization of gene expression were conducted using the GEPIA (Gene Expression Profiling Interactive Analysis) online tool (http://gepia.cancer-pku.cn) and the ggplot2 package in R (https://www.r-project.org/). Survival analysis was performed and visualized using the “survival” package (https://cistrome.shinyapps.io/timer/). Raw data were compared and quantified using RSEM 1.3.3 and Bowtie2 software (27, 28). Alternative splicing was analyzed using rMATS 4.1.2 software, with visualization of these results facilitated by rmats2sashimiplot (29). Differentially expressed genes (DEGs) was identified using the “DESeq2” package (https://bioconductor.org/packages/release/bioc/html/DESeq2.html), and Gene Ontology (GO) functional enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were conducted using the “cluster Profiler” package (https://bioconductor.org/packages/release/bioc/html/clusterProfiler.html).

Hep G2 cells and SMMC 7721 were obtained from the Cell Bank of the Chinese Academy of Sciences and cultured according to the study’s protocol. These cells were maintained in DMEM enriched with 10% heat-inactivated FBS, penicillin (100 units/mL), and streptomycin (100 μg/mL) in a 37°C incubator with a humidified atmosphere of 5% CO₂. Cell viability was assessed using the MTT assay. For this assay, cells were seeded in 96-well plates and treated with the appropriate reagents. After treatment, 100 μL of MTT solution (5 mg/ml) was added to each well, and the plates were incubated for 4 h at 37°C. Subsequently, 100 μL of DMSO (Sangong, Shanghai) was used to dissolve the formazan crystals formed by the cells. The optical density was then measured at 492 nm using a PL-9602 Enzyme Labeling Instrument (Perlong Medical, Beijing, China). Cell viability was calculated as a percentage relative to the control.

The hnRNPA1 overexpression plasmid hnRNPA1 (NM_002136) pcDNA3.1-3xFlag-C was purchased from Youbio (Changsha, Hunan). For the construction of the pCMV-Myc-ZNF207 mutant plasmid, primers ZNF207-Bal I, ZNF207-Kpn I, and ZNF207-OE-F/R were designed ( Supplementary Table S1 ). Overlapping PCR was performed using ZNF207-Bal I and ZNF207-OE-R, as well as ZNF207-Kpn I and ZNF207-OE-F, under the following conditions: 95°C for 3 min, followed by 29 cycles of 95°C for 30 s, 58°C for 30 s, and 72°C for 10 s, with a final extension at 72°C for 5 min. The PCR products were then combined and used as a template for a subsequent round of PCR with ZNF207-Bal I-F and ZNF207-Kpn I-R. The pCMV-Myc plasmid served as the template for these PCR amplifications. Next, 20 μL of the PCR product was mixed with 1 μL of Bal I and 1 μL of Kpn I (Thermo Fisher Scientific, USA), and the mixture was incubated at 37°C for 14 h. Homologous recombination was then carried out at the mutant sites of the overlapping products using reagents from T4 DNA Ligase (Thermo Fisher Scientific, USA). Finally, 5 μL of the recombinant plasmid was transformed into Top10 competent cells and cultured on plates containing ampicillin overnight. The correctly constructed plasmid was confirmed through nucleic acid sequencing.

We collected tumor samples from patients. Briefly, the hnRNPA1 antibodies (1:100) were applied to stain each tissue sample. Subsequently, the indicated HRP-conjugated secondary antibody was used in incubating the slice at 37 °C for 45 min. We then used diaminobenzidine (DAB)solution (Immunoway, China) to stain the slice, and we counterstained the nuclei with Harris’hematoxylin. The percentage of specifically positive staining of tumor cells was classified with the following grades: 0 (<5%), 1 (6%-25%), 2 (26%-50%), 3 (51%-75%), and 4 (>75%). The final score was expressed by multiplying the staining intensity and the percentage of specifically positive staining tumor cells.

RNAiso easy (Takara, Japan) was used to extract total RNA according to the manufacturer’ instructions. cDNA was synthesized from 1 μg of total RNA using NovoScripts plus All-in-one 1st Strand cDNA Synthesis Super Mix (E401, Novoprotein, Suzhou, China), following the manufacturer’s instructions. Real-time RT-PCR analysis was performed on a CFX Connect PCR detection system (Bio-Rad, Germany) using NovoStart Universal Fast SYBR qPCR Super Mix (E047, Novoprotein, Suzhou, China), according to the manufacturer’s instructions. The reaction was performed on CFX Conect (Bio-Rad, Singapore) with conditions: 95°C for 2 min, 38 cycles of 95°C for 5 s, and 60°C for 15s. The mRNA expression (folds) was calculated using the 2^−ΔΔCt method.

The siRNAs of hnRNPA1 and control were purchased from Jikai Gene (Suzhou, China). Sequences are detailed in Supplementary Table S1 . Cells grown in 6-well plates were transfected with a final concentration of 75 pmol. Lipofectamine 2000 (Invitrogen) was used for all transfection assays according to the instructions. The efficiency of all transfections was examined by RT-qPCR. Relative sequences are listed in Supplementary Table S1 .

SMMC 7721 or Hep G2 were seeded onto a 6-well plate at a density of 5 × 10⁵ per well 36 h before transfection with si-hnRNPA1. To analyze apoptosis transfected si-hnRNPA1, cells were collected and double-stained with annexin V and propidium iodide using the Annexin V-PI apoptosis detect in Kit (Vazyme, Nanjing, China), following the manufacturer’s instructions. The flow data were selected using BD FACSDiva v.7 and analyzed using FlowJo v.7.6. The gating strategies are shown in Supplementary Figure S5 .

For the cell clone formation assay, 1,000 transfected cells were seeded in 6-well plates. After four days, the cells were fixed with 4% paraformaldehyde (PFA) for 30 min. Following fixation, cells were rinsed with tap water and stained with a 0.1% crystal violet solution for 1 h. The plates were rinsed again, and photographs were taken for documentation.

For the wound healing assay, cells were plated in 6-well plates and allowed to grow until they reached 100% confluence. A scratch was then made on the plate using a 10 μL pipette tip, and the cells were subsequently cultured in serum-free DMEM medium. For each sample, three random areas were selected and observed under 10 x 10 magnification at 0 and 24 h, respectively. The relative migration distance of the cells was measured to assess their migration capability.

Treated cells were lysed with RIPA lysis buffer supplemented with a protease inhibitor and phosphatase inhibitor (ABclonal, China) for 30 min on ice. The protein samples were collected after centrifugation at 12,000 rpm at 4°C for 10 min. Equal amounts of protein were loaded and separated on 6%~15% SDS-PAGE gels and transferred onto polyvinylidene difluoride (PVDF) membranes (Merck Millipore Ltd. IPVH00010, Darmstadt, Germany). The membranes were blocked with 5% milk blocking solution at room temperature for 1-2 h, washed in TBST buffer, and incubated overnight at 4°C with primary antibodies of p-mTOR, mTOR, p-PI3K and PI3K (1:1000, ZEN-BIOSCIENCE, China), p-Akt, Akt, hnRNPA1 (1:1000, Immunoway, China), β-actin, β-tubulin and GAPDH (1:2000, ABclonal, China), Bax, Bcl2, Caspase 3 and cleaved Caspase3 (1:1000, HuaBio, China). After washing with TBST 3 times, the membranes were incubated with secondary antibodies conjugated with horseradish peroxidase (HRP) (1:10000, Abbkine, China) at room temperature for 1 h. The membrane blots were detected by using an enhanced chemiluminescence (ECL) kit (NCM Biotech, China). All gray analyses for protein blots were performed with ImageJ software.

Gel electrophoresis at 5% concentration was performed using the following components: 6.5 mL of ddH₂O, 1.25 mL of 40% acrylamide, 2 mL of 5 X TBE, 100 μL of 10% ammonium persulfate (AP, Sangon, China), and 8 μL of TEMED. After the gel solidified, DNA samples from the transfected cells were loaded, and the electrophoresis was conducted at 125 V and 125 mA for 60 minutes. Subsequently, the gel was immersed in 4S Red stain (Sangon, China) and allowed to develop under a UV lamp for one hour.

Cells were lysed by multimeric lysate contained with 1 M KCl, 50 mM MgCl₂, 100 mM HEPEs-NaOH (PH 7), 5% NP-40, 1mM DTT, 200 units/mL RNase Inhibitor (Sangon Biotech, B600478), Protease Inhibitor Cocktail (MedChemExpress, HY-K0010). Then, protein A/G magnetic beads (Selleck, B23201) preincubated with IgG (negative control) or antibody specific for HNRNPA1 (Abclonal, A11564) were incubated with lysates at 4°C overnight. After washing five times, Eluted RNAs were purified by Proteinase K (Sangon Biotech, A610451) The enrichment of ZNF207 was detected by qRT-PCR.

Data are presented as means ± standard error (SD). Comparisons among groups were performed using one-way ANOVA followed by Dunnett’s test using GraphPad Prism 9.0 with different letters representing different significances: a: p< 0.05, b: p< 0.01, c: p< 0.005, d: p< 0.001 (n = 3).

Using bioinformation tool, we analyzed gene expression profiles from 374 cancer tissue samples and 50 normal tissue samples extracted from The Cancer Genome Atlas (TCGA) database. Our study identified a significant elevation in hnRNPA1 expression in different cancers such as colorectal adenocarcinoma (COAD), diffuse large B-cell lymphoma (DLBC), and liver hepatocellular carcinoma (LIHC) ( Figure 1A ). Kaplan-Meier survival curves further indicated a negative correlation between hnRNPA1 expression and survival rates in HCC patients ( Figure 1B ). Specifically, the mRNA expression level of hnRNPA1 was increased in the HCC cell (SMMC 7721 and Hep G2) compared with normal liver cells (L02 and LX-2), also the mRNA expression level of hnRNPA1 was increased in other cancer cells (U251, Hela and B16-F10) (p < 0.001) ( Figure 1C ). The result of western blot analysis confirmed that hnRNPA1 expression was predominantly elevated in HCC cells (SMMC 7721, MCHH 97H and Hep G2) in comparison to non-HCC cells (Hela, B16-F10 and U251) and normal liver cells (L-02 and LX-2) (p < 0.05) ( Figure 1D ; Supplementary Figure S1 ). The size difference in hnRNPA1 protein originates from the alternative splicing in the C-terminus. Immunohistochemistry performed on HCC patient tissues validated the significant elevation of hnRNPA1 in tumor tissues compared to adjacent non-tumor tissues (p < 0.001) ( Figure 1E ).

To elucidate the role of hnRNPA1 in facilitating the progression of HCC and its underlying molecular mechanisms, we engineered siRNAs targeting hnRNPA1 specifically ( Supplementary Table S1 , Supplementary Figure S2A ). MTT and clonogenic assays revealed that silencing hnRNPA1 significantly reduced the proliferation of Hep G2 cells (p < 0.05), ( Supplementary Figures S2B, C ). Additionally, wound healing assays demonstrated that hnRNPA1 knockdown markedly decreased the migration of Hep G2 cells (p < 0.05, Supplementary Figure S2D ). Previous research indicates that hnRNPA1 suppresses pro-apoptotic proteins, thereby influencing tumor growth (30). Western blot analysis confirmed that hnRNPA1 knockdown reduced Bcl2 levels and enhanced the expression of the pro-apoptotic protein Bax and caspase-3 activation ( Supplementary Figure S2E ).

Subsequently, we assessed hnRNPA1 expression in Hep G2 and SMMC 7721 cells ( Figure 2A ; Supplementary Figure S3A ). MTT and clonogenic assays showed that overexpressing hnRNPA1 significantly increased cell proliferation in both Hep G2 and 7721 cells, significance is only evaluated between knockdown and overexpressing cells ( Figures 2B, C ). The above results showed that cell proliferation was no significant difference within 24h. The result of wound healing assays showed that overexpression of hnRNPA1 led to a significantly higher migration rate compared to the control within 24h ( Figure 2D ). CCND1 (Cyclin D1), MPP2 (Membrane palmitoylated protein 2), Bax, Bcl2, HIF1A and VEGFA are transcriptionally regulated molecules that are related to cell proliferation, apoptosis and cancer cell invasion and metastasis processes. The results of qPCR showed that knockdown-hnRNPA1 dramatically lowered mRNA levels of these molecules in Hep G2 ( Supplementary Figure S3B ).

To further explore the relationship between hnRNPA1 and HCC, we performed an in-depth RNA-seq analysis of hnRNPA1-knockdown in Hep G2 cells using common data. The results identified 2,641 DEGS from TCGA, GSM2343352 and GSM2343354, consisting of 1,312 up-regulated genes and t. 1,329 down-regulated genes. Changes in down-regulated genes were ETV4 (ETS Variant Transcription Factor 4) and ETV5 (ETS Variant Transcription Factor 5) are related to transcriptional regulation, EML4 (EMAP like 4), BCL2 and Myc are related to cancer development. And changes in up-regulated genes were CDH1 (E-cadherin), NOTCH1 and IL21R were related to the normal morphology of cells ( Figure 3A ). The results of RT-qPCR were consistent with the results of DEG analysis ( Figure 3B ). GO analysis of the identified DEGs showed that several biological processes (BP) related to tumorigenesis, the apoptosis signaling pathway, and the negative regulation of cell proliferation were all affected after hnRNPA1 knockdown ( Figure 3C ). KEGG pathway analysis revealed significant differences in the expression of protein processing in endoplasmic reticulum, ubiquitin mediated proteolysis, and endocytosis ( Figure 3D ).

It has been reported that hnRNPA1 plays a crucial role in regulating many cellular AS events. To explore the specific AS events influenced by hnRNPA1 that affect the proliferation of Hep G2 cells, we processed enhanced ultraviolet cross-linking and immunoprecipitation (eCLIP) and KD-RNA-Seq data from eight samples. We established a genome index with the human genome GRCh38 as the reference, setting the read segment length at 99 for get bam files (31). The files on exon skipping events calculated by rMATS software were analyzed using the maser package in R software and rmats2sashimiplot software, and the occurrence of ZNF207 exon skipping in the control group and the knockdown hnRNPA1 experimental group was compared to obtaining the regulatory information of hnRNPA1 on alternative splicing of ZNF207. This analysis identified 75,465 exon skipping (SE) events. Among these, 22 were significantly noteworthy, with ZNF207 displaying the most pronounced change in Delta PSI, showing a decrease of 0.278 (avg reads > 100, FDR > 0.05, Delta PSI = 0.1) ( Figure 4A ). We also noted a substantially increased frequency of exon skipping in ZNF207. Initially, eCLIP data confirmed the binding of hnRNPA1 to ZNF207 mRNA ( Figure 4B ). RNA immunoprecipitation (RIP) results also revealed that hnRNPA1 could bind to ZNF207 in Hep G2( Figure 4C ). Detailed examination indicated that skipping exon 9 in ZNF207 is a specific event, leading to the generation of transcript B (NM00103229.3) ( Figure 4D ). Alternative splicing involves the change of isoforms. Visualization of this AS event revealed elevated IncLevel values for the skipped exon which represented an increased expression of transcript B, which showed by yellow area ( Figure 4E ). The results of TBE blot showed that ratios of transcript A/B were decreased, which indicated the expression of transcript B was higher in Hep G2 ( Figure 4F ; Supplementary Figure S4A ). Additionally, the result of RT-qPCR showed that the expression of ZNF207 transcript B was increased after hnRNPA1 knockdown ( Figure 4G ). This suggested that hnRNPA1 regulate the expression of exon-skipping transcripts of the ZNF207 gene in Hep G2. Skipping of exon 9 in ZNF207 produced transcripts of varying lengths, so the mRNA length changes after splicing.

The findings above demonstrated that hnRNPA1 knockdown impacted AS in Hep G2 cells. To assess the influence of ZNF207 exon 9 splice isoforms on the proliferation and migration of HCC cells, we synthesized mutant plasmids for ZNF207-short and ZNF207-long using overlapping PCR and whole-genome synthesis ( Supplementary Figures S4B, C ). We then evaluated the impact of these mutant plasmids on cell proliferation and migration. Our results indicated that both hnRNPA1 knockdown and ZNF207-long suppressed HCC cell proliferation whereas ZNF207-short enhanced these processes ( Figures 5A, B ). Wound healing assays further confirmed that knockdown of hnRNPA1 and ZNF207-long hindered HCC cell migration, while ZNF207-short facilitated it ( Figure 5C ). Additionally, flow cytometry analysis revealed that knockdown of hnRNPA1 and ZNF207-long increased apoptosis, while ZNF207-short reduced it ( Figure 5D ; Supplementary Figure S5A ). Western blot analysis showed that knockdown of hnRNPA1 and ZNF207-long upregulated the pro-apoptotic protein Bax and downregulated the anti-apoptotic protein Bcl2, thereby activating caspase-3 and enhancing apoptosis ( Figure 5E ; Supplementary Figure S5B ).

Since excessive activation of the PI3K/Akt/mTOR pathway is known to inhibit caspase family proteins, reducing apoptosis, we explored the impact of the ZNF207 splice isoform on this signaling pathway in Hep G2 or SMMC 7721 cells using Western blot analysis. Results indicated that ZNF207-short enhanced the activation of PI3K, leading to its conversion to phosphorylated 3,4,5-phosphatidylinositol triphosphate, which subsequently activated Akt and mTOR, fostering cell proliferation ( Supplementary Figures S6A, B ). Conversely, the knockdown of hnRNPA1 and ZNF207-long reduced the phosphorylation of PI3K, Akt, and mTOR, with the changes in Akt and mTOR phosphorylation being statistically significant (p < 0.05).

As mTOR is a key downstream effector of the PI3K/Akt pathway and plays a crucial role in cell growth and proliferation, its inhibition, such as by rapamycin, is often targeted in cancer therapy due to the pathway’s frequent dysregulation in tumors. After transfecting Hep G2 cells with ZNF207-long and ZNF207-short for 36 h, the cells were treated with 8 μM rapamycin for 24 h. Western blot analysis showed that rapamycin effectively inhibited mTOR phosphorylation. Following the rapamycin treatment, there was a slight decrease in PI3K/Akt activity, but the difference in the PI3K/Akt signaling pathway expression between cells transfected with ZNF207-short and those with inhibited mTOR activity was not significant ( Figures 6A, B ).