Multi-Omics Analysis of Survival-Related Splicing Factors and Identifies CRNKL1 as a Therapeutic Target in Esophageal Cancer

Abstract

Background: RNA alternative splicing represents a pivotal regulatory mechanism of eukaryotic gene expression, wherein splicing factors (SFs) serve as key regulators. Aberrant SF expression drives oncogenic splice variant production, thereby promoting tumorigenesis and malignant progression. However, the biological functions and potential targets of SFs remain largely underexplored. Methods: Through multi-omics analysis, we identified survival-related splicing factors (SFs) in esophageal cancer and elucidated their biological regulatory networks. To further investigate their downstream splicing targets, we combined alternative splicing events resulting from SF knockdown with those specific to esophageal cancer. Finally, these splicing events were validated through full-length RNA sequencing and confirmed in cancer cells and clinical specimens. Result: We identified six SFs that are highly expressed in esophageal cancer and correlate with poor prognosis. Further analysis revealed that these factors are significantly associated with immune infiltration, cancer stemness, tumor heterogeneity, and drug resistance. CRNKL1 was identified as a hub SFs. The target genes and pathways regulated by these SFs showed substantial overlap, suggesting their coordinated roles in promoting cancer stemness and metastasis. Specifically, alternative splicing of key markers, such as CD44 and CTTN, was regulated by most of these SFs and correlated with poor prognosis. Conclusions: Our study unveils six survival-related SFs that contribute to the aggressiveness of esophageal cancer and CTTN and CD44 alternative splicing may act as common downstream effectors of survival-related SFs. This study provides mechanistic insights into SF-mediated tumorigenesis and highlight novel therapeutic vulnerabilities in esophageal cancer.

Keywords: alternative splicing; esophageal cancer; splicing factors.

Figure 1

Workflow of data acquisition and analysis in this study.

Figure 2

(A) Kaplan–Meier plot showing the survival analysis of six survival-related splicing factors (SFs) in esophageal cancer, based on TCGA database. (B) mRNA expression level of six survival-related SFs in normal and esophageal cancer tissues based on TCGA database. (C) The abnormal expression of survival-related SFs in esophageal cancer based on proteomics data. (D,E) Aberrant expression of SNRPB2 and CRNKL1 in esophageal cancer tissues. (F) Protein–protein interaction (PPI) network analysis using the STRING database demonstrated that five of the six survival-related SFs were functionally correlated. Gene Ontology (GO) enrichment analysis indicated that these SFs are involved in U2-type spliceosome assembly and mRNA splicing. (G) Analysis of mRNA expression levels of the survival-related SFs in the TCGA database showed significant correlations among the SFs. The thickness of lines represents the strength of correlation between SFs, with the correlation coefficient indicated along the lines. *** p < 0.001.

Figure 3

(A) Negative correlation between survival-related splicing factors (SFs) and levels of tumor-infiltrating CD8⁺ T cells. (B) Positive correlation between survival-related SFs and cancer-associated fibroblasts (CAFs). (C) Aberrant expression of survival-related splicing factors (SFs) in cancer cells shown by single-cell RNA sequencing. (D) Negative correlation between SNRPB2 expression and tumor-infiltrating CD8⁺ T cells, based on immunohistochemical (IHC) staining. Arrows indicate tumor-infiltrating CD8⁺ T cells. (E) Positive correlation between survival-related SFs and representative chemokines. (F) Positive correlation between survival-related SFs and tumor purity.

Figure 4

(A) Positive correlation between survival-related SFs and cancer stemness, as assessed by RNAss estimation. (B,C) Positive correlations between survival-related SFs and MATH (B) and ploidy (C). (D) Representative splicing factor CRNKL1 is correlated with resistance to cisplatin, oxaliplatin, and 5-fluorouracil in esophageal cancer. (E) Survival-related SFs are associated with resistance to tyrosine kinase inhibitors.

Figure 5

(A) Identification of hub survival-related splicing factors (SFs) using Cytoscape PPI score. (B) Correlation between CRNKL1 expression and clinical stages of esophageal cancer (ESCA). (C) Validation of CRNKL1 knockdown in KYSE30 and KYSE410 cells by Western blotting. (D–G) CRNKL1 knockdown significantly impairs cell migration in KYSE30 and KYSE410 cells, as demonstrated by transwell assays (D,E) and wound healing assays (F,G). (H) CRNKL1 knockdown significantly affects cytoskeleton remodeling in KYSE30 cells. (I) CRNKL1 knockdown significantly impairs cancer stemness in KYSE30 and KYSE410 cells, as demonstrated by tumor sphere formation experiments. (J,K) CRNKL1 knockdown significantly affects tumor proliferative ability, as indicated by proliferation curves (J) and colony formation assays (K). * p < 0.05, ** p < 0.01, *** p < 0.001.

Figure 6

(A) Workflow for deciphering SFs-driven esophageal cancer-specific splicing targets. (B) Statistics of survival-related SFs-driven splicing targets in ESCA. (C) KEGG enrichment analysis of pathways enriched in SF-driven splicing targets. Blue and red rectangles showed cytoskeletal pathways and stemness cell pathways. (D) Common splicing targets driven by survival-related SFs in esophageal cancer. (E) KEGG enrichment analysis of common splicing targets.

Figure 7

(A) Workflow for exploring common downstream splicing targets in ESCA. (B) Alternative splicing (AS) of CTTN exon 11 and its different isoforms. (C–E) Changes in CTTN exon 11 AS in SFs-knockdown HeLa cells (C), the TCGA database (D), and full-length RNA-seq of paired ESCA tissues (E). (F) AlphaFold3-predicted interaction between CRNKL1 and CTTN exon 11 RNA. (G–J) Knockdown of CRNKL1 significantly promotes CTTN exon 11 skipping in KYSE30 (G) and KYSE410 (H) cells; conversely, CTTN exon 11 inclusion was observed in both ESCA cell lines (I) and clinical specimens (J). (K) CTTN exon 11 inclusion is significantly correlated with poor prognosis in ESCA. (L) Illustration of different CD44 isoforms. (M–O) Changes in CD44v8 AS in SFs-knockdown HeLa cells (M), the TCGA database (N), and full-length RNA-seq of paired ESCA tissues (O). (P) Knockdown of CRNKL1 significantly promotes CTTN exon 11 skipping in KYSE30 cells. (Q) CD44v8 inclusion significantly correlated with poor prognosis in ESCA. * p < 0.05, ** p < 0.01, *** p < 0.001.