Dysregulated inclusion of BOLA3 exon 3 promoted by HNRNPC accelerates the progression of esophageal squamous cell carcinoma

Abstract

Dysregulated RNA splicing events produce transcripts that facilitate esophageal squamous cell carcinoma (ESCC) progression, but how this splicing process is abnormally regulated remains elusive. Here, we unveiled a novel alternative splicing axis of BOLA3 transcripts and its regulator HNRNPC in ESCC. The long-form BOLA3 (BOLA3-L) containing exon 3 exhibited high expression levels in ESCC and was associated with poor prognosis. Functional assays demonstrated the protumorigenic function of BOLA3-L in ESCC cells. Additionally, HNRNPC bound to BOLA3 mRNA and promoted BOLA3 exon 3 inclusion forming BOLA3-L. High HNRNPC expression was positively correlated with the presence of BOLA3-L and associated with an unfavorable prognosis. HNRNPC knockdown effectively suppressed the malignant biological behavior of ESCC cells, which were significantly rescued by BOLA3-L overexpression. Moreover, BOLA3-L played a significant role in mitochondrial structural and functional stability. E2F7 acted as a key transcription factor that promoted the upregulation of HNRNPC and inclusion of BOLA3 exon 3. Our findings provided novel insights into how alternative splicing contributes to ESCC progression.

Keywords

esophageal squamous cell carcinoma / alternative splicing / exon skipping / BOLA3 / HNRNPC

1 Introduction

Esophageal cancer is a common malignant cancer originating from the mucosal epithelium of the esophagus. In 2020, esophageal cancer had 604 000 new cases and resulted in 544 000 deaths worldwide, ranking 8th and 6th among cancer types, respectively [1]. Esophageal squamous cell carcinoma (ESCC) is the major histologic subtype in China. It is characterized by insidious onset and aggressive progression, often resulting in delayed diagnosis and necessitating multiple interventions, such as surgery, radiotherapy, and immunotherapy [2]. The five-year survival rate of ESCC is approximately 20%–30% because of surgical complications, resistance to radiotherapy and chemotherapy, and limited responsiveness to immunotherapy [3, 4]. Therefore, oncogenic molecular events underlying ESCC and effective therapeutic targets are urgently needed.

Alternative splicing (AS) is a critical step in the posttranscriptional processing of pre-mRNA, generating diverse transcripts through the inclusion or exclusion of multiple exons or other sequences [5]. AS greatly expands the diversity of the transcriptome and proteome, thereby ensuring normal cell differentiation and organ development [6]. AS regulation relies on the interplay between regulatory splicing factors and intrinsic cis-acting RNA elements. The dysregulation of AS leads to alterations in the function, activation, and expression of specific splicing transcripts, contributing to pathological conditions, including carcinogenesis. These splicing aberrations frequently drive malignant phenotypes, such as cell proliferation, invasive metastasis, vascular proliferation, and resistance to radiotherapy [7, 8]. Therapeutic strategies targeting oncogenic aberrant spliced transcripts and regulatory factors are promising routes for precise cancer treatment [7, 9].

AS events occur approximately 20% more frequently in tumor tissues than in normal tissues [10]. In ESCC, the AS profile reveals that exon skipping (ES) is the predominant abnormal AS category, and ES dysregulation correlates with poor prognosis. The aberrant expression of splicing factors are associated with unfavorable prognosis in ESCC [11]. Many splicing factors induce oncogenic splicing switching by participating in regulating splicing sites and tendencies. Exploring the mechanism of key AS events promoting ESCC facilitates the discovery of potential biomarkers and therapeutic strategies.

BOLA3 belongs to the Bola family, which localizes in mitochondria and participates in iron–sulfur and glycine metabolism [12, 13]. The skipping of BOLA3 exon 3 generates two coding transcripts: long-form BOLA3(BOLA3-L; NM-212552.3) and short-form BOLA3 (BOLA3-S; NM-001035505.2). Compared with BOLA3-L, BOLA3-S exhibits exon 3 skipping and exon 4 forward termination. BOLA3 is closely associated with the occurrence and progression of various tumors, including hepatocellular carcinoma and ovarian cancer, and exhibits potential as a diagnostic biomarker [14–16]. However, the mechanism underlying the BOLA3 oncogenic characteristics remains unclear.

HNRNPC is a core member of the heterogeneous nuclear ribonucleoprotein family, which plays a crucial role in RNA splicing regulation [17]. HNRNPC is dysregulated in multiple types of cancer and identified as an oncogenic factor through RNA processing, which involves protein interaction, RNA splicing manipulation, and m⁶A RNA modification [18]. In non-small cell lung cancer (NSCLC), HNRNPC interacts with KHSRP to activate the IFN-α/JAK/STAT1 axis, thereby facilitating NSCLC metastasis [19]. HNRNPC interacts with circ-FIRRE and promotes ESCC progression by stabilizing GLI2 mRNA [20]. By participating in ALKBH5-mediated m⁶A modification, HNRNPC promotes the maturation of DDX58 and facilitates the progression of head and neck squamous cell carcinoma [21]. In this work, we investigated AS events in ESCC and identified the excessive inclusion of BOLA3 exon 3 as a key dysregulated ES event. We further conducted a comprehensive functional analysis of BOLA3-L and BOLA3-S and verified whether HNRNPC is a splicing factor mediating exon 3 skipping. Moreover, we uncovered the role of E2F7 in transcriptional upregulation of HNRNPC and inclusion of BOLA3 exon 3. Our findings elucidated a novel oncogenic mechanism of the HNRNPC/BOLA3 splicing axis in ESCC.

2 Materials and methods

2.1 Patients and tumor samples

In our study, we collected 21 pairs of ESCC and adjacent normal tissue samples from patients who underwent radical resection at Changhai Hospital between April 2021 and December 2021. The normal tissues were located at least 3 cm away from the tumor tissues. We sequenced the transcriptomes of 11 tissue pairs, designating this set as the “CH-Cohort”. An additional 10 pairs of tissues were used for Western blot validation and DNA electrophoresis. All specimens were obtained from primary ESCC patients who had not previously undergone chemotherapy or surgery. Tissue samples were immediately snap-frozen in liquid nitrogen and stored at −80 °C. The study received approval from the Ethics Committee of Changhai Hospital (Approval No. CHEC2023-018).

2.2 Cell culture

The human esophageal epithelial cell (HEEC), 293T, and ESCC cell lines (KYSE30, KYSE150, KYSE410, KYSE510, ECA109, TE-1) were purchased from the Chinese Academy of Sciences (Shanghai, China) and Procell Life Science & Technology (Wuhan, China). All cells were cultured in RPMI1640 (Gibco) supplemented with 10% fetal bovine serum (FBS, Gibco) and 1% penicillin–streptomycin (Gibco). The cell lines were cultured in a humidified incubator at 37 °C with 5% CO₂.

2.3 Lentiviral, plasmid, and siRNA transfection

BOLA3-L- and HNRNPC-silencing lentiviral plasmids (ShBOLA3-L1/2 and ShHNRNPC-1/2), BOLA3-L- and BOLA3-S-overexpressing lentiviral plasmid (OE-BOLA3-L, OE-BOLA3-S), and their negative controls (Vector, Sh-NC) were obtained from OBiO Technology (Shanghai). ESCC cell lines (KYSE150 and KYSE510) were infected with lentivirus for 24 h. Stable transfected cell lines were screened out with a culture medium containing puromycin (2.5 μg/mL, Merck Millipore) for 7 days. Plasmids containing Flag-tagged wild-type (WT) HNRNPC (HNRNPC-WT) and HNRNPC domain deletion mutants (HNRNPC-ΔRRM) were constructed. The BOLA3 minigenes, designated as BOLA3 E(2‒4), were created by amplifying genomic DNA sequences that encompass exons 2 through 4, along with 150 base pairs flanking each side of the introns 2 and 3 within the BOLA3 gene.These fragments were then cloned into the pcDNA3.1 vector separately. The small interfering RNA (siRNA) oligos against E2F7 and the plasmids above mentioned were commercially synthesized by OBiO Technology (Shanghai, China). The transfection of plasmids, siRNA, and negative controls into KYSE150 and KYSE510 cells was performed using lipofectamine 3000 (Invitrogen). Samples were harvested after 24 h of transfection for further research.

2.4 RNA isolation and real-time quantitative PCR (RT-qPCR)

TRIzol reagent (Invitrogen) was used to extract the total RNA from the fresh-frozen tissues and cell lines. Then, 1 μg of RNA was reverse transcribed into cDNA using the Hifair II 1st Strand cDNA synthesis kit (YEASEN Biotechnology). Real-time quantitative PCR (RT-qPCR) was carried out in triplicate with Hieff UNICON qPCR SYBR Green Master Mix (YEASEN) on a LightCycler 480 II System (Roche). The 2^−ΔΔCT method was used in RT-qPCR data analysis. GAPDH was used as an internal control. To facilitate the detection of BOLA3-L and BOLA3-S expression, PCR using Premix Taq (Takara) was applied to amplify the specific transcript, and the DNA levels were determined by gel electrophoresis. We developed a BOLA3 E(2–4) primer specifically covering the exons 2–4 segment. All primer sequences are listed in Table S1.

2.5 RNA sequencing analysis

Total RNA isolated from CH-Cohort, KYSE150 cells transfected with Sh-NC, ShHNRNPC-1, and ShBOLA3-L1 lentivirus was subjected to paired-end RNA sequencing (RNA-seq) using an Illumina HiSeq 6000 system according to the manufacturer’s instructions. Next-generation sequencing was performed by Oebiotech Co. Ltd. (Shanghai). Genes with log₂FC (fold change) ≥1 and P < 0.05 indicated differentially expressed genes. On the basis of the hypergeometric distribution, significantly enriched terms were screened with Gene Ontology (GO) and KEGG pathway enrichment analyses, and the AS of differentially regulated transcripts isoforms or exons was analyzed with rMATS. AS events with ΔPSI ≥0.01 and P < 0.05 were considered dysregulated splicing events. We downloaded the expression and clinical data of GSE53246 from the Gene Expression Omnibus (GEO) database for further analysis. The series of our raw sequence data was deposited in the GEO database under the accession number GSE235538.

2.6 TCGA RNA-seq data analysis

RNA-seq data (RPKM) and clinical characteristics of esophageal cancer and normal samples were obtained from the TCGA database. After excluding the data of patients without complete clinical information, we performed survival regression analysis using the “survival” package to identify coding genes associated with overall survival (OS) based on hazard ratio (HR) and P value. The expression and association with esophageal cancer prognosis of AS events were analyzed by “oncosplicing” web tool based on PSI data originating from TCGA [22]. The boxplot shows the distribution of PSI in esophageal cancer samples compared with adjacent normal samples. The optimal cutoff was predicted using survival data by the “surv_cutpoint” function in the R package “survminer”.

2.7 Western blot analysis

Protein extraction was performed using RIPA lysis buffer, followed by separation on SDS-PAGE. The SDS-PAGE was transferred to a PVDF membrane and blocked with 5% skim milk powder at 37 °C for 60 min. The PVDF membrane was incubated with primary antibodies at 4 °C overnight and then incubated with appropriate HRP-labeled secondary antibodies. An ECL system was used in the development of the films. The primary antibodies against BOLA3 (A15985), HNRNPC (11760-1-AP), Flag (66008-4-Ig), and β-tubulin (10068-1-AP) were purchased from ABclonal Technology and Proteintech.

2.8 Immunohistochemistry (IHC)

The ESCC tissue microarray (TMA) was purchased from Shanghai Outdo Biotech. Two independent pathologists performed IHC and used a modified Histo-score. The ESCC and TMA were incubated with HNRNPC antibodies. The proportion of positively stained cells and the intensity score were scored. A final score was then calculated by multiplying the two scores.

2.9 Immunofluorescence staining

The cells were fixed in 4% paraformaldehyde and rinsed with phosphate-buffered saline (PBS) buffer three times. Cells were then permeabilized using 0.1% Triton X-100 and blocked with 5% goat serum. The cells were incubated with BOLA3 overnight at 4 °C. Then, the cells were incubated with fluorescent secondary antibodies and 4′,6-diamidino-2-phenylindole (DAPI). Mitotracker was used to mark mitochondrion. Immunofluorescence analysis was conducted on a confocal laser scanning microscope.

2.10 Luciferase reporter assay

Fragments containing the HNRNPC-promoter regions (WT) or mutants of the predictive E2F7 binding site (mutation, Mut) were inserted upstream of the firefly luciferase coding sequences in the pGL4.1-basic reporter plasmid. The binding of E2F7 to HNRNPC promoter was assessed after esophageal cancer cells were cotransfected with the E2F7 overexpression or negative control, and the HNRNPC-WT or HNRNPC-Mut vectors using a Lipofectamine 3000 kit. A Renilla luciferase reporter vector was cotransfected to normalize transfection efficiency. They were commercially synthesized by OBiO Technology (Shanghai). The hTFtarget website was used to predict the potential binding motif [17]. Luciferase activities were detected after transfection for 2 days with a dual-luciferase reporter assay system (E1980, Promega).

2.11 Cell proliferation

Cell proliferation assays were conducted by using a 96-well plate, and 100 µL cell suspensions containing 3000 cells subjected to different interventions were added to each test well and incubated at 37 °C and 5% CO₂ for 0, 24, 48, and 72 h. At each time point, 10 μL of cell counting kit-8 (CCK-8) solution was added, and the cells were incubated for 1.5 h in an incubator. Then, the absorbance at 450 nm was detected with an Infinite 200 spectrometer.

2.12 Colony formation

Colony formation assays were performed on six-well plates. For each cell type, 1000 cells and a normal medium were added to each well, cultured in an incubator for 10 days, and fixed with 4% paraformaldehyde. The cells were then stained with crystal violet and imaged.

2.13 EdU assay

The transfected cells were seeded into 96-well plates (1 × 10⁴ cells) and cultured overnight in a 37 °C incubator. According to the instructions of the BeyoClick EdU cell proliferation kit with Alexa Fluor 488 (Beyotime), EdU detection was performed. Then, EdU-positive cells were stained with Azide 488 and Hoechst 33342. Images from three random fields of vision were obtained under a microscope, and the percentage of EdU-positive cells was calculated.

2.14 Wound healing assay

Esophageal cancer cells of different types were seeded in six-well plates, scratched with a sterile 200 μL pipette tip, and then cultured in a serum-free medium. Cells were photographed after incubation for 0 and 24 h, and the width of the wounds was measured.

2.15 Transwell assay

The invasion and migration ability of cells were assessed. A Transwell chamber was placed on a 24-well plate. In the invasion test, matrigel was laid in the upper chamber. No matrix was coated in the upper chamber for the migration assay. Then, 200 µL of serum-free medium containing 1 × 10⁵ cells were seeded into the upper chamber, and 700 μL of medium containing 20% FBS was added to the lower chamber. The Transwell chamber was placed in a 37 °C incubator for 48 h. The cells on the lower surface of the chamber were fixed with methanol and stained with crystal violet (0.1%). After thorough washing with PBS, migrated cells were imaged and quantified with ImageJ.

2.16 Xenograft tumor formation assay

A xenograft tumor formation assay was performed on 4-week-old female nude mice (BALB/c) purchased from Beijing Weitong Lihua. The animal experiments were approved by the animal care and use committee of Changhai Hospital and complied with its guidelines and policies. We constructed stable lentivirus transfection cell lines in KYSE150 cells. The cells were subcutaneously injected into each nude mouse. After 3 weeks, the mice were euthanized and dissected. The tumor was photographed and weighed, and tumor volume was calculated: tumor volume = length*width²/2.

2.17 RNA immunoprecipitation assay (RIP)

The interaction between HNRNPC protein and BOLA3 mRNA was analyzed through RNA immunoprecipitation (RIP) assays. The RIP kit (Bes5101) was used according to the manufacturer’s instructions. PCR and gel electrophoresis were used to detect the levels of BOLA3-L and BOLA3-S mRNA in the immunoprecipitated complex.

2.18 Chromatin immunoprecipitation (ChIP)

Chromatin immunoprecipitation (ChIP) assay was conducted using a ChIP kit (Bersinbio) according to the manufacturer’s instructions. Briefly, KYSE150 cells were cross-linked at room temperature with 1% formaldehyde and 1.5 mmol/L ethylene glycol-bis. The cross-linked chromatin was sonicated and subjected to immunoprecipitation with an E2F7 antibody. The protein–DNA complexes detached from the antibodies, freeing DNA fragments. PCR-amplified DNA products were detected using gel electrophoresis.

2.19 Transmission electron microscopy

Cells were fixed with 0.05 M cacodylate buffer containing 2.5% glutaraldehyde and 2% formaldehyde, immersed in 1% osmium tetroxide, rinsed with phosphate buffer, dehydrated with gradient concentrations of ethanol, and embedded in epoxy resin. Then, the samples were sliced into 70–80 nm sections with an EM UC7 ultramicrotome and stained with uranyl acetate and lead citrate. The sections were observed using a transmission electron microscope at 80 kV.

2.20 Statistical analysis

All data presented as histograms represent a mean value ± SD of the total number of three independent experiments. Statistical analysis was performed by two-way ANOVA and Student’s t-test at a significance level of P < 0.05. Analysis of molecular expression correlation analysis was performed using Pearson correlation analysis. Survival curves were generated by Kaplan–Meier methods, and comparisons were performed using the log-rank test. The results were visualized using GraphPad Prism9 and R software, and statistical significance was analyzed.

3 Results

3.1 Landscape of AS events in ESCC

To characterize AS events linked to ESCC progression, we analyzed the whole transcriptome data of 11 sample pairs, including ESCC tumors and paired adjacent normal tissues (CH-Cohort). We observed a large number of dysregulated AS events, and ES had the highest dysregulation frequency (Fig.1). Key dysregulated ES events contributing to ESCC progression were identified, and candidate genes were prioritized according to their ΔPSI values. In the CH-Cohort, we set the threshold at ΔPSI ≥0.1 and P < 0.05 for the screening of candidate genes. In the TCGA database, we used the “oncosplicing” tool with a threshold of ΔPSI > 0 and P < 0.05 to screen dysregulated ES genes. By intersecting the results, we identified 579 dysregulated ES genes (Fig.1). GO pathway analysis revealed their involvement in biological processes, such as cytoskeleton organization, actomyosin structure organization, and positive regulation of basement membrane assembly (Fig.1).

3.2 BOLA3-L was highly expressed in ESCC and associated with poor prognosis of patients with ESCC

We then investigated the prognostic data from the TCGA database and identified 355 OS-related coding genes (Table S2). By intersecting these genes with dysregulated ES events within the threshold (HR > 1, P < 0.05), we identified 12 genes with 15 key dysregulated ES events (Fig.1). Among these, the BOLA3-ES54017 event caught our attention due to its expression, PSI value, and RT-PCR validation. Human BOLA3 contains 4 exons, and the ES event leads to long (L) or short (S) isoforms with the inclusion or exclusion of the 89 bp exon-3, respectively (Fig.1).

To assess the expression levels of BOLA3 isoforms in ESCC samples, we performed RT-PCR and gel electrophoresis using a primer pair spanning the alternative exon-3. The tumor samples exhibited a higher ratio of BOLA3-L than the control samples, consistent with the high PSI in the ESCC tumors of CH-Cohort (Fig.1, 1G, and S1A). We further conducted electrophoresis to verify the ESCC cell line (Figs. S1B, and S1C). The elevated expression levels of BOLA3-L and BOLA3 were observed in various ESCC cell lines compared with HEEC. In TCGA, the PSI of BOLA3 in the esophageal cancer samples was higher than that in the normal tissues (Fig.1). Survival analysis revealed that BOLA3-L was significantly associated with poor survival in patients with esophageal cancer (Fig.1). These findings suggested that the dysregulated BOLA3 ES is a novel prognostic factor for patients with ESCC and plays a critical role in ESCC progression.

3.3 BOLA3-L promoted the malignant behavior of ESCC cells in vitro and in vivo

To investigate the role of BOLA3-L in ESCC tumorigenesis, we specifically downregulated BOLA3-L expression in two ESCC cell lines (KYSE150 and KYSE510), using ShRNAs targeting the exon 3 of BOLA3, while keeping the BOLA3-S isoform unaffected (Fig.2 and 2B). This knockdown led to a reduction in BOLA3 protein levels (Fig.2). ESCC cells with BOLA3-L knockdown exhibited considerable inhibition in growth and colony formation compared with control cells (Fig.2 and 2E). Additionally, EdU staining assays demonstrated a significant decrease in the proportion of EdU(+) cells in the BOLA3-L knockdown group, indicating suppressed proliferation (Fig.2). Moreover, wound healing and Transwell assays revealed limited migration and invasion ability in ESCC cells with stable BOLA3-L knockdown (Fig.2 and 2H).

Subsequently, we investigated the role of BOLA3-L in tumorigenesis by using in vivo xenograft models. KYSE150 cells with BOLA3-L knocked down through shRNA targeting or control shRNA (sh-NC) were separately injected into nude mice. The BOLA3-L knockdown group exhibited considerably reduced tumor growth rate and tumor weight compared with the control group (Fig.2). These findings indicated that the knockdown of BOLA3-L resulted in reduced malignancy in ESCC cells.

To validate our findings, we generated stable ESCC cell lines overexpressing BOLA3-L (OE-BOLA3-L), BOLA3-S (OE-BOLA3-S), or an empty vector (Vector, control). PCR and gel electrophoresis confirmed the elevated mRNA expression of BOLA3-L or BOLA3-S (Fig.3). Western blot analysis further confirmed the overexpression of BOLA3 at the protein level (Fig.3). Functional assays demonstrated that BOLA3-L overexpression greatly enhanced ESCC cell proliferation and colony formation, whereas BOLA3-S overexpression had no such effect compared with the control groups (Fig.3 and 3D). Consistently, the BOLA3-L-overexpression group exhibited a significant increase in EdU(+) cells, indicating enhanced cell proliferation (Fig.3). Moreover, wound healing and Transwell assays revealed that BOLA3-L overexpression promoted the migration and invasion abilities of ESCC cells compared with the BOLA3-S overexpression or the empty vector (Fig.3 and 3G). These oncogenic effects of BOLA3-L overexpression were opposite to the effects observed in the ESCC cells with BOLA3-L knockdown, confirming the essential role of BOLA3-L in the malignant behavior of ESCC cells.

Additionally, we independently injected KYSE150 cells with stable BOLA3-L or BOLA3-S overexpression or an empty control vector into nude mice. We observed a high tumor growth rate and increase in tumor weight in the BOLA3-L-overexpression group, whereas BOLA3-S overexpression had no significant effect compared with the vector group (Fig.3). Furthermore, after knocking down BOLA3-L in KYSE150 cells, we overexpressed BOLA3-L and BOLA3-S transcripts to observe the rescue function of malignant phenotypes. We observed that the overexpression of BOLA3-S did not lead to an increase in proliferation (Fig. S2A and S2B) and migration (Fig. S2C) capabilities after BOLA3-L knockdown in vitro. Conversely, BOLA3-L overexpression resulted in the substantial enhancement of proliferation and migration capabilities. Moreover, the same rescuing trend was observed in the in vivo xenograft models (Fig. S2D). BOLA3-L overexpression in BOLA3-L knockdown xenografts partially restored tumorigenesis. However, the xenografts with BOLA3-S overexpression and BOLA3-L knockdown did not exhibit such effects. These data strongly support the pro-tumorigenic function of BOLA3-L in ESCC cells, which was not observed with BOLA3-S. Collectively, our results suggest that BOLA3-L plays an essential role in the progression of ESCC.

3.4 HNRNPC was a splicing factor of BOLA3 and highly expressed HNRNPC correlated with poor prognosis of ESCC

To identify transacting factors that potentially regulate the AS of BOLA3 exon 3, we analyzed the transcriptome data of the CH-Cohort. We identified 118 upregulated splicing factors (Table S3). By intersecting these factors with TCGA OS-related genes, we identified 13 dysregulated splicing factors as key candidates (Fig.4). We further visualized their mRNA expression levels in the CH-Cohort using a heatmap (Fig.4). Next, we analyzed PAR-CLIP sequencing data obtained from the POSTAR3 database and predicted a total of 16 RNA-binding proteins (RBPs) that potentially bind to BOLA3 pre-mRNA (Table S4). To identify the candidate RBPs, we intersected the key dysregulated splicing factors with the predicted RBPs and screened out HNRNPC as the key splicing factor (Fig.4). We further validated that HNRNPC was overexpressed in ESCC in CH-Cohort (Fig.4). Additionally, a significant correlation was observed between the PSI of BOLA3 exon 3 and the expression of HNRNPC in patients with ESCC (Fig.4). Analysis of TCGA data revealed high expression of HNRNPC in esophageal cancer (Fig.4), which correlated with poor prognosis in esophageal cancer patients (Fig.4). Consistently, elevated expression of HNRNPC was observed in ESCC samples from the GSE53624 data set, and it was associated with poor prognosis (Fig.4 and 4I). Furthermore, the protein expression of HNRNPC was higher in ESCC cell lines than in HEEC. The expression of HNRNPC was elevated in ESCC compared with adjacent normal tissues validated by Western blot (Fig.4). IHC analysis further confirmed that ESCC tissues had higher HNRNPC expression levels than the adjacent normal tissues in TMA (Fig.4). Kaplan–Meier survival analysis demonstrated that a high level of HNRNPC protein was associated with poor prognosis in patients with ESCC (Fig.4).

3.5 Knockdown of HNRNPC reduced BOLA3 exon 3 inclusion and inhibited tumorigenesis in ESCC

We attained stable HNRNPC knockdown in ESCC cells by using shRNAs (ShHNRNPC-1 and ShHNRNPC-2). The reduced mRNA and protein levels of HNRNPC were confirmed through RT-qPCR and Western blot, respectively (Fig.5 and 5B). We observed exon 3 exclusion, which led to the production of the BOLA3-S isoform (Fig.5). We also observed that after HNRNPC knockdown, no considerable change occurred in the overall protein levels of BOLA3 (Fig. S3). These findings demonstrated that HNRNPC promoted the inclusion of BOLA3 exon 3. Subsequently, we investigated the effect of HNRNPC on tumorigenesis. Cell growth assays revealed a substantial decrease in the proliferation of ESCC cells upon HNRNPC knockdown (Fig.5). EdU assays indicated a significant decline in the proportion of EdU(+) cells in the HNRNPC knockdown group (Fig.5). Colony formation assays demonstrated a decrease in the number of focal adhesions when HNRNPC was knocked down (Fig.5). Additionally, wound healing (Fig.5) and Transwell assays (Fig.5) revealed the weakened migration and invasion abilities of ESCC cells in the HNRNPC knockdown group. To further validate these observations, we independently implanted KYSE150 cells with stable knocked down HNRNPC or an empty control into nude mice. Notably, the HNRNPC knockdown group ShHNRNPC-1 exhibited a lower tumor growth rate and reduced tumor weight (Fig.5). Collectively, these data confirmed the protumorigenic function of HNRNPC in ESCC cells, emphasizing its potential role in the progression of ESCC.

4 HNRNPC promoted BOLA3 exon 3 inclusion by interacting with BOLA3 mRNA via the RRM domain

To investigate the AS events regulated by HNRNPC in ESCC cells, we identified a total of 11 668 differentially spliced events (ΔPSI ≥ 0.01, P < 0.05) in response to HNRNPC knockdown in KYSE150 cells. Most of the events were ES events (8723; Fig.6). The Sashimi plot confirmed the exclusion of exon 3 in BOLA3 upon HNRNPC knockdown (Fig.6). We further explored the binding and regulatory relationship between HNRNPC and BOLA3. RIP-PCR analysis revealed considerably higher BOLA3-L and BOLA3-S mRNA expression levels in the HNRNPC precipitates than in the IgG control (Fig.6), indicating the binding of HNRNPC protein with BOLA3-L and BOLA3-S mRNA. To validate the regulatory mechanism of HNRNPC on the AS of BOLA3, we designed plasmids expressing HNRNPC-WT or a mutant lacking the RNA-binding domain (HNRNPC-ΔRRM) in 293T cells and confirmed their expression by Western blot (Fig.6). Subsequently, we introduced a plasmid carrying the analyzed AS of exon 3 into 293T cells with HNRNPC knockdown or rescue (Fig.6, top). We observed that the BOLA3 minigene with exon 3 inclusion was predominant in 293T cells, while HNRNPC knockdown in 293T cells markedly decreased this inclusion, resulting in the short isoform (Fig.6, bottom, lanes 2 vs. 1). Notably, introducing the HNRNPC-WT plasmid in the cells markedly restored exon 3 inclusion, leading to the formation of the long isoform (Fig.6, bottom, lanes 4 vs. 3), whereas the HNRNPC-ΔRRM plasmid did not show the same effect (Fig.6, bottom, lanes 5 vs. 4). These findings demonstrated that HNRNPC-WT promoted BOLA3 exon 3 inclusion, whereas HNRNPC-ΔRRM did not. The promoting effect of HNRNPC on BOLA3 exon 3 inclusion was impaired when the RNA-binding domain was deleted. These observations suggested that HNRNPC promoted the inclusion of exon 3 in BOLA3 by interacting with BOLA3 mRNA via the RRM domain.

4.1 The overexpression of BOLA3-L partially rescued the antitumorigenic effects of HNRNPC knockdown

To elucidate the functional relationship between BOLA3-L and HNRNPC in tumorigenesis, we generated stable ESCC cell lines with HNRNPC knockdown and concomitant overexpression of BOLA3-L or BOLA3-S. We observed that HNRNPC knockdown disrupted the formation of BOLA3-L in the ESCC cells. This effect decreased the number of surviving colonies and the ratio of Edu(+) proliferating cells and was partially restored by overexpressing BOLA3-L but not BOLA3-S (Fig.6 and 6G). Similarly, Transwell assays revealed a rescue effect on the migration ability of ESCC cells with HNRNPC knockdown upon overexpression of BOLA3-L, but not BOLA3-S (Fig.6). To further validate these findings, we employed an in vivo mouse model. We observed that xenografts with HNRNPC knockdown displayed a reduction in size and weight compared with the control group. Moreover, BOLA3-L overexpression in HNRNPC knockdown xenografts partially restored tumorigenesis. However, the xenografts with BOLA3-S overexpression and HNRNPC knockdown did not exhibit such effects (Fig.6). These observations revealed that HNRNPC regulated the exon 3 skipping of BOLA3, mediating its tumorigenic functions in ESCC cells.

4.2 Identifications of the downstream pathways and functional activities of BOLA3-L

To investigate the signaling pathways associated with BOLA3-L in ESCC, we performed RNA-seq analysis after BOLA3-L knockdown in KYSE150 cells. Validation of the reduced inclusion of exon 3 in BOLA3 was confirmed through Sashimi plot visualization of RNA-seq reads (Fig.7). Differential expression analysis revealed 113 upregulated and 163 downregulated genes (log₂FC ≥ 1, P < 0.05), and a heatmap was generated to display their expression patterns (Fig.7). GO analysis indicated that the dysregulated genes were primarily involved in biological processes, such as cell–cell signaling, keratinization, and amyloid-beta clearance (Fig.7). Additionally, KEGG analysis demonstrated that the affected targets of BOLA3-L were functionally associated with the Hippo and WNT Pathways (Fig.7). Given that BOLA3 is a mitochondrial protein involved in maintaining the stability of the mitochondrial respiratory chain, we performed immunofluorescent staining and confocal microscopy to validate the localization of BOLA3-L. Our findings showed the colocalization of the mitotracker and the majority of BOLA3 protein in the Sh-NC group. However, a considerable reduction in BOLA3 protein intensity with mitotracker was observed upon BOLA3-L knockdown (Fig.7). Furthermore, functional assays revealed that knocking down BOLA3-L led to reduced JC-1 polymerization intensity and triggered mitochondrial depolarization, as shown by the JC-1 assay (Fig.7). Electron microscopy examinations demonstrated the emergence of partial ridge disappearance, swelling, and rupture in mitochondria when BOLA3-L was knocked down in the ESCC cells (Fig.7). These observations indicated that BOLA3-L played important roles in multiple cellular pathways that regulate mitochondrial homeostasis.

4.3 E2F7 activated the transcription of HNRNPC and promoted BOLA3 exon 3 inclusion

Additionally, we investigated the mechanism underlying the elevated expression of HNRNPC in ESCC. We analyzed the data from TCGA and CH-Cohort RNA-seq. By taking the intersection of the dysregulated transcription factors of CH-Cohort and TCGA poor OS–related genes, we obtained two key transcription factors: E2F7 and ASCL-2 (Fig.8). E2F7 was overexpressed in esophageal cancer in the CH-Cohort and TCGA data sets (Fig.8 and 8C), and the overexpression of E2F7 was associated with poor prognosis in esophageal cancer in TCGA (Fig.8). Correlation analysis revealed the positive relationship of E2F7 expression with the expression of HNRNPC and PSI of BOLA3 exon 3 in the CH-Cohort (Fig.8 and 8F). Knocking down E2F7 with siRNA reduced the expression of E2F7 and HNRNPC at the mRNA and protein levels (Fig.8–8I). Additionally, the inclusion of BOLA3 exon 3 decreased after E2F7 knockdown (Fig.8). These findings suggested that E2F7 regulate the transcription of HNRNPC. The potential binding motif of E2F7 was predicted using the JASPAR online database (Fig.8, top). We obtained the potential binding motif table by predicting the E2F7 binding sites of the HNRNPC promoter with the “htfTarget” tool (Table S5). By referencing the TOP5 binding sequence based on the binding score, we identified the TTTCCCCGCCCCCGCGCA as a key binding sequence (Fig.8, bottom). ChIP assay of HNRNPC, followed by PCR and gel electrophoresis, confirmed that HNRNPC was directly bound to the promoter of HNRNPC. To confirm the transcriptional function of this binding sequence, we subcloned the WT and mutant forms of HNRNPC and transfected them with overexpressed E2F7 and control plasmids for luciferase assays. The results showed that E2F7 enhanced the luciferase activity of the reporter vector carrying the WT binding site but not the mutant, highlighting the critical transcriptional function of the key binding sequence of E2F7 with HNRNPC (Fig.8). We conducted experiments to investigate the impact of overexpressing HNRNPC or BOLA3-L following E2F7 knockdown on the malignant phenotype. After inducing E2F7 knockdown, we observed a decrease in the proliferation (Fig. S4A) and migration (Fig. S4B) phenotype of the cells compared with the control group. Subsequently, upon the overexpression of HNRNPC, the malignant phenotype was partially restored. This effect indicated that HNRNPC can rescue the malignant phenotype. Furthermore, when E2F7 was knocked down, the overexpression of BOLA3-L resulted in the recovery of proliferation (Fig. S4C) and migration (Fig. S4D) phenotype, whereas the overexpression of BOLA3-S did not exhibit malignant phenotype restoration.

In summary, this study uncovered the HNRNPC/BOLA3 splicing axis that promoted ESCC development. HNRNPC bound to BOLA3 mRNA via its RRM domain to enhance the inclusion of BOLA3 exon 3. E2F7 transcriptionally upregulated the expression of HNRNPC and drove ESCC through the BOLA3-L transcript by regulating multiple oncogenic pathways (Fig.9).

5 Discussion

Aberrant AS events are crucial promoters in ESCC carcinogenesis. Addressing the key splicing events driving ESCC is essential to the discovery of specific targeted inhibitors. Here, we discovered that the exon 3 of BOLA3 is highly included in ESCC tumor tissues and demonstrated that the long transcript BOLA3-L, rather than BOLA3-S, promotes the oncogenesis of ESCC cells in vitro and in vivo. Moreover, HNRNPC is verified as an RBP for BOLA3 and plays an oncogenic role in ESCC. HNRNPC binds to BOLA3 mRNA and promotes BOLA3 exon 3 inclusion forming BOLA3-L. Furthermore, E2F7 enhances HNRNPC expression through transcriptional regulation, and BOLA3-L might maintain the structural and functional stability of the mitochondria of ESCC cells.

ES is a crucial process in AS that involves the removal of exon sequences. Excessive exons skipping or inclusions often lead to the imbalanced expression of different transcripts and functional domain changes in protein products [23, 24]. Abnormal ES events in ESCC frequently emerge, further generating cancer-promoting transcripts. We identified ES-based differential AS events mainly involving various cellular pathways and physiologic functions, such as intercellular adhesion, energy metabolism, ubiquitination, and mitochondrial autophagy [7]. Abnormal ES generates oncogenic transcripts that exhibit specific activity changes, promoting malignant phenotypes. For instance, the members of the Lysyl oxidase enzyme family, such as Lysyl oxidase-like 2 (LOXL2), play a role in tumor invasion and metastasis by remodeling the tumor microenvironment. The deletion of exon-specific nucleotide sequences in LOXL2Δ72 and LOXL∆13 alters their structural domains. Compared with WT LOXL2, LOXL2Δ72 and LOXL∆13 exhibit lower oxidase activity but higher performance in enhancing ESCC migration and invasion [25, 26]. Additionally, splice variant 1 (SV1) generated from multiple ES events in the growth hormone-releasing hormone gene acts as a hypoxia-driven pro-oncogenic promoter. SV1 activates muscle-type phosphofructokinase through the nuclear factor-κB pathway, further enhancing glycolytic metabolism and promoting ESCC development [27]. Investigations into ES pathways have expanded our understanding of the molecular mechanisms underlying ESCC progression and the potential targets for oncogenetic pathway inhibitors.

BOLA3 is highly expressed in ovarian cancer, hepatocellular carcinoma, and other tumors, and is associated with poor prognosis [15, 16]. By interacting with NFU1, BOLA3 promotes the transfer of Fe-S from the synthesis phase to the chaperone protein transport phase and protects the [4Fe-4S] cluster from oxidative damage [12]. In our study, we identified the pro-oncogenic function of BOLA3-L. No similar functions were observed in BOLA3-S. BOLA3-L promotes the stability of the mitochondrial respiratory chain and enhances the efficiency of respiration [28]. We discovered that the crucial oncogenic mechanism of BOLA3 proceeds through BOLA3-L with exon 3 and BOLA3-L primarily contributes to mitochondrial stability and influenceds multiple cancer-related pathways, including the Hippo and WNT pathways. Our results indicate that the aberrant skipping of BOLA3 exon 3 promotes ESCC progression, and targeting BOLA3-L transcript offers a new route for developing cancer biomarkers and treatments.

Aberrant AS events are mainly derived from disordered regulatory events, including abnormal expression of splicing factors, mutations in splicing regulatory element (SRE), and mutations in chromatin modifiers [29]. Splicing factors mainly include two major families: the serine/arginine-rich proteins and HNRNPs [30]. In ESCC, the disordered expression of splicing factors leads to aberrant expression ratio, stability, and activity of downstream transcripts, initiating a pro-oncogenic splice switch [31–33]. SRSF1 is a core splicing factor in the splicing process and is upregulated in ESCC. By binding to LncRNA DGCR5, SRSF1 promotes the aberrant splicing of myeloid leukemia-1 protein (MCL-1), leading to the excessive expression of MCL-1L with anti-apoptotic properties in ESCC cells [34]. USP39 promotes Rictor splicing and transcript maturation, further accelerating ESCC development through the mTOR pathway [33]. HNRNPF regulates fibroblast growth factor receptor 2 (FGFR2) AS by promoting the conversion of FGFR2IIIb to FGFR2 IIIc, which activates epithelial–mesenchymal transition (EMT) in ESCC [32]. Through the unique RRM and functional domains, splicing factors can bind to SRE and promote the assembly of spliceosome complexes. The characteristic of the SRE sequence and expression of splicing factors further modulate splicing tendency.

The aberrant expression of HNRNPC has been observed in multiple cancer types and contributes to malignant progression through various mechanisms [18]. In ESCC, HNRNPC enhances ZEB1 and ZEB2 mRNA stability by binding to LBX2-AS1, thereby promoting ESCC cell migration and EMT [35]. In pancreatic cancer, HNRNPC downregulates the antimetastatic isoform TAF8-L and increases the prometastatic isoform TAF8-S. m⁶A site mutation in TAF8 reduces the level of interaction between HNRNPC and TAF8, resulting in a decrease in TAF8-S expression [36]. In our study, we identified HNRNPC as a pro-oncogenic factor that is transcriptionally upregulated by E2F7, suggesting its essential role in ESCC development. E2F7, known as a hypoxia-induced transcription factor, promotes the upregulation of QKI and enhances circBCAR3 biogenesis, thereby driving the development of esophageal cancer [37]. The impact of hypoxia on the HNRNPC/BOLA3 splicing axis requires further investigation. Additionally, we observed that the tumor suppressive effect of HNRNPC knockdown could be partially rescued by overexpressing BOLA3-L, highlighting the key role of the HNRNPC/BOLA3 splicing axis in aberrant splicing mechanisms during ESCC progression. Furthermore, we found that HNRNPC promotes BOLA3 exon 3 inclusion by binding to BOLA3 mRNA through its RRM domain. The interaction between HNRNPC and BOLA3 modulates the splicing switch, providing a novel therapeutic target for developing strategies against ESCC.

In conclusion, our study revealed an excessive inclusion of BOLA3 exon 3 in ESCC and the highly expressed long isoform BOLA3-L associated with poor prognosis. HNRNPC plays a crucial role as an ESCC promoter by binding to BOLA3 mRNA via its RRM domain and enhancing exon 3 inclusion. Furthermore, E2F7 transcriptionally upregulates the expression of HNRNPC and BOLA3-L, driving ESCC progression through multiple pathways. These findings highlight the importance of the HNRNPC/BOLA3 splicing axis, offering novel insights into potential diagnostic markers and splicing-regulating therapeutic strategies for ESCC.

References

[1] Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, Bray F. 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

[2] Zhou C, Zhang J. Immunotherapy-based combination strategies for treatment of gastrointestinal cancers: current status and future prospects. Front Med 2019; 13(1): 12–23

[3] Zeng H, Chen W, Zheng R, Zhang S, Ji JS, Zou X, Xia C, Sun K, Yang Z, Li H, Wang N, Han R, Liu S, Li H, Mu H, He Y, Xu Y, Fu Z, Zhou Y, Jiang J, Yang Y, Chen J, Wei K, Fan D, Wang J, Fu F, Zhao D, Song G, Chen J, Jiang C, Zhou X, Gu X, Jin F, Li Q, Li Y, Wu T, Yan C, Dong J, Hua Z, Baade P, Bray F, Jemal A, Yu XQ, He J. Changing cancer survival in China during 2003–15: a pooled analysis of 17 population-based cancer registries. Lancet Glob Health 2018; 6(5): e555–e567

[4] Shah MA, Kennedy EB, Alarcon-Rozas AE, Alcindor T, Bartley AN, Malowany AB, Bhadkamkar NA, Deighton DC, Janjigian Y, Karippot A, Khan U, King DA, Klute K, Lacy J, Lee JJ, Mehta R, Mukherjee S, Nagarajan A, Park H, Saeed A, Semrad TJ, Shitara K, Smyth E, Uboha NV, Vincelli M, Wainberg Z, Rajdev L. Immunotherapy and targeted therapy for advanced gastroesophageal cancer: ASCO guideline. J Clin Oncol 2023; 41(7): 1470–1491

[5] Pajares MJ, Ezponda T, Catena R, Calvo A, Pio R, Montuenga LM. Alternative splicing: an emerging topic in molecular and clinical oncology. Lancet Oncol 2007; 8(4): 349–357

[6] Baralle FE, Giudice J. Alternative splicing as a regulator of development and tissue identity. Nat Rev Mol Cell Biol 2017; 18(7): 437–451

[7] Bradley RK, Anczuków O. RNA splicing dysregulation and the hallmarks of cancer. Nat Rev Cancer 2023; 23(3): 135–155

[8] Xie D, Wang Z, Sun B, Qu L, Zeng M, Feng L, Guo M, Wang G, Hao J, Zhou G. High frequency of alternative splicing variants of the oncogene Focal Adhesion Kinase in neuroendocrine tumors of the pancreas and breast. Front Med 2023; 17(5): 907–923

[9] Desterro J, Bak-Gordon P, Carmo-Fonseca M. Targeting mRNA processing as an anticancer strategy. Nat Rev Drug Discov 2020; 19(2): 112–129

[10] Kahles A, Lehmann KV, Toussaint NC, Hüser M, Stark SG, Sachsenberg T, Stegle O, Kohlbacher O, Sander C, Cancer Genome Atlas Research Network, Rätsch G. Comprehensive analysis of alternative splicing across tumors from 8705 patients. Cancer Cell 2018; 34(2): 211–224.e6

[11] Sun JR, Kong CF, Lou YN, Yu R, Qu XK, Jia LQ. Genome-wide profiling of alternative splicing signature reveals prognostic predictor for esophageal carcinoma. Front Genet 2020; 11: 796

[12] Uzarska MA, Nasta V, Weiler BD, Spantgar F, Ciofi-Baffoni S, Saviello MR, Gonnelli L, Mühlenhoff U, Banci L, Lill R. Mitochondrial Bol1 and Bol3 function as assembly factors for specific iron-sulfur proteins. eLife 2016; 5: e16673

[13] Tajima K, Ikeda K, Chang HY, Chang CH, Yoneshiro T, Oguri Y, Jun H, Wu J, Ishihama Y, Kajimura S. Mitochondrial lipoylation integrates age-associated decline in brown fat thermogenesis. Nat Metab 2019; 1(9): 886–898

[14] Wang XF, Lei W, Liu CM, Yang J, Zhu YH. BOLA3 is a prognostic-related biomarker and correlated with immune infiltrates in lung adenocarcinoma. Int Immunopharmacol 2022; 107: 108652

[15] Wang D, Wang Z, Tao Y. Prognostic values of BolA family member expression in hepatocellular carcinoma. BioMed Res Int 2022; 2022: 8360481

[16] Zhu M, Xiao S. Expression profiles and prognostic values of BolA family members in ovarian cancer. J Ovarian Res 2021; 14(1): 75

[17] Xie W, Zhu H, Zhao M, Wang L, Li S, Zhao C, Zhou Y, Zhu B, Jiang X, Liu W, Ren C. Crucial roles of different RNA-binding hnRNP proteins in stem cells. Int J Biol Sci 2021; 17(3): 807–817

[18] Mo L, Meng L, Huang Z, Yi L, Yang N, Li G. An analysis of the role of HnRNP C dysregulation in cancers. Biomark Res 2022; 10(1): 19

[19] Yan M, Sun L, Li J, Yu H, Lin H, Yu T, Zhao F, Zhu M, Liu L, Geng Q, Kong H, Pan H, Yao M. RNA-binding protein KHSRP promotes tumor growth and metastasis in non-small cell lung cancer. J Exp Clin Cancer Res 2019; 38(1): 478

[20] Zhou Y, Xue X, Luo J, Li P, Xiao Z, Zhang W, Zhou J, Li P, Zhao J, Ge H, Tian Z, Zhao X. Circular RNA circ-FIRRE interacts with HNRNPC to promote esophageal squamous cell carcinoma progression by stabilizing GLI2 mRNA. Cancer Sci 2023; 114(9): 3608–3622

[21] Jin S, Li M, Chang H, Wang R, Zhang Z, Zhang J, He Y, Ma H. The m⁶A demethylase ALKBH5 promotes tumor progression by inhibiting RIG-I expression and interferon alpha production through the IKKε/TBK1/IRF3 pathway in head and neck squamous cell carcinoma. Mol Cancer 2022; 21(1): 97

[22] Zhang Q, Liu W, Zhang HM, Xie GY, Miao YR, Xia M, Guo AY. hTFtarget: a comprehensive database for regulations of human transcription factors and their targets. Genom Proteom Bioinf 2020; 18(2): 120–128

[23] Ma X, Zhang W, Zhang R, Li J, Li S, Ma Y, Jin W, Wang K. Overexpressed long noncoding RNA CRNDE with distinct alternatively spliced isoforms in multiple cancers. Front Med 2019; 13(3): 330–343

[24] Leclair NK, Brugiolo M, Urbanski L, Lawson SC, Thakar K, Yurieva M, George J, Hinson JT, Cheng A, Graveley BR, Anczuków O. Poison exon splicing regulates a coordinated network of SR protein expression during differentiation and tumorigenesis. Mol Cell 2020; 80(4): 648–665.e9

[25] Zou HY, Lv GQ, Dai LH, Zhan XH, Jiao JW, Liao LD, Zhou TM, Li CQ, Wu BL, Xu LY, Li EM. A truncated splice variant of human lysyl oxidase-like 2 promotes migration and invasion in esophageal squamous cell carcinoma. Int J Biochem Cell Biol 2016; 75: 85–98

[26] Jiao JW, Zhan XH, Wang JJ, He LX, Guo ZC, Xu XE, Liao LD, Huang X, Wen B, Xu YW, Hu H, Neufeld G, Chang ZJ, Zhang K, Xu LY, Li EM. LOXL2-dependent deacetylation of aldolase A induces metabolic reprogramming and tumor progression. Redox Biol 2022; 57: 102496

[27] Xiong X, Ke X, Wang L, Yao Z, Guo Y, Zhang X, Chen Y, Pang CP, Schally AV, Zhang H. Splice variant of growth hormone-releasing hormone receptor drives esophageal squamous cell carcinoma conferring a therapeutic target. Proc Natl Acad Sci USA 2020; 117(12): 6726–6732

[28] Cameron JM, Janer A, Levandovskiy V, Mackay N, Rouault TA, Tong WH, Ogilvie I, Shoubridge EA, Robinson BH. Mutations in iron-sulfur cluster scaffold genes NFU1 and BOLA3 cause a fatal deficiency of multiple respiratory chain and 2-oxoacid dehydrogenase enzymes. Am J Hum Genet 2011; 89(4): 486–495

[29] Bradley RK, Anczuków O. RNA splicing dysregulation and the hallmarks of cancer. Nat Rev Cancer 2023; 23(3): 135–155

[30] Pereira B, Billaud M, Almeida R. RNA-binding proteins in cancer: old players and new actors. Trends Cancer 2017; 3(7): 506–528

[31] Kidogami S, Iguchi T, Sato K, Yoshikawa Y, Hu Q, Nambara S, Komatsu H, Ueda M, Kuroda Y, Masuda T, Mori M, Doki Y, Mimori K. SF3B4 plays an oncogenic role in esophageal squamous cell carcinoma. Anticancer Res 2020; 40(5): 2941–2946

[32] Wang X, Sun M, Gao Z, Yin L, Pu Y, Zhu Y, Wang X, Liu R. N-nitrosamines-mediated downregulation of lncRNA-UCA1 induces carcinogenesis of esophageal squamous by regulating the alternative splicing of FGFR2. Sci Total Environ 2023; 855: 158918

[33] Zhao Y, Geng H, Liu G, Ji Q, Cheng X, Li X, Liu W, Thorne RF, Zhang R, Liu X. The deubiquitinase USP39 promotes ESCC tumorigenesis through pre-mRNA splicing of the mTORC2 component rictor. Front Oncol 2021; 11: 667495

[34] Duan Y, Jia Y, Wang J, Liu T, Cheng Z, Sang M, Lv W, Qin J, Liu L. Long noncoding RNA DGCR5 involves in tumorigenesis of esophageal squamous cell carcinoma via SRSF1-mediated alternative splicing of Mcl-1. Cell Death Dis 2021; 12(6): 587

[35] Zhang Y, Chen W, Pan T, Wang H, Zhang Y, Li C. LBX2-AS1 is activated by ZEB1 and promotes the development of esophageal squamous cell carcinoma by interacting with HNRNPC to enhance the stability of ZEB1 and ZEB2 mRNAs. Biochem Biophys Res Commun 2019; 511(3): 566–572

[36] Huang XT, Li JH, Zhu XX, Huang CS, Gao ZX, Xu QC, Zhao W, Yin XY. HNRNPC impedes m⁶A-dependent anti-metastatic alternative splicing events in pancreatic ductal adenocarcinoma. Cancer Lett 2021; 518: 196–206

[37] Xi Y, Shen Y, Wu D, Zhang J, Lin C, Wang L, Yu C, Yu B, Shen W. CircBCAR3 accelerates esophageal cancer tumorigenesis and metastasis via sponging miR-27a-3p. Mol Cancer 2022; 21(1): 145

RIGHTS & PERMISSIONS

Higher Education Press