SNRPD2-mediated regulation of DDX39B splicing promotes endometrial cancer progression by suppressing the activation of CTSC cryptic exons
Yingwei Li, Zhongshao Chen, Yanling Liu, Yuehan Gao, Yingying Pu, Qianqian Gao, Feng Gao, Ning Yang, Peng Li
TL;DR
This study shows that the protein SNRPD2 promotes endometrial cancer by regulating gene splicing, and targeting it could be a new treatment approach.
Contribution
The paper identifies a novel SNRPD2–DDX39B–CTSC regulatory axis in endometrial cancer progression.
Findings
SNRPD2 overexpression correlates with poor clinical outcomes in endometrial cancer.
SNRPD2 silencing reduces tumor growth and metastasis in endometrial cancer models.
SNRPD2 regulates DDX39B splicing, which in turn affects CTSC expression and cancer progression.
Abstract
Recent studies have reported the overexpression of Sm proteins in several cancers, suggesting their potential as therapeutic targets; however, the specific Sm family members involved in endometrial cancer and their mechanisms remain unclear. Here, we show that the Sm protein SNRPD2 is markedly upregulated in both fresh-frozen and formalin-fixed paraffin-embedded (FFPE) endometrial cancer specimens and that its overexpression correlates with poorer clinical outcomes. In vitro and in vivo functional assays demonstrate that silencing SNRPD2 suppresses endometrial cancer cell proliferation and metastasis. Specifically, antisense oligonucleotides (ASOs) targeting SNRPD2 markedly reduced tumor growth in a patient-derived xenograft (PDX) model. Mechanistic analyses reveal that SNRPD2 knockdown induces the retention of intron 5 in DDX39B, resulting in the production of a noncoding transcript…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8- —China Health Promotion Foundation (XH-C010)
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsRNA Research and Splicing · Cancer-related molecular mechanisms research · Cancer-related gene regulation
Introduction
Endometrial cancer is among the leading malignancies of the female reproductive system and is the most common gynecologic cancer in Western countries [1]. Its incidence is increasing, driven in part by increasing rates of obesity, longer life expectancy, and lower fertility [2]. Although early-stage endometrial cancer generally has a favorable prognosis (5-year survival of approximately 95%), the outlook for advanced cases is poor (5-year survival of only ~19%) [3]. The American Cancer Society estimates that 69,120 new endometrial cancer cases and 13,860 deaths will occur in 2025 [4], with a projected 40–50% increase in incidence by 2030, accompanied by a continuous increase in mortality [5, 6]. In China, the incidence and mortality are likewise increasing, with 81,964 new cases and 16,607 deaths reported in 2020, representing a substantial share of the global burden [7]. At present, the primary treatment modalities for endometrial cancer include hysterectomy, radiotherapy, chemotherapy, endocrine therapy and immunotherapy; however, endocrine approaches have limited efficacy, and hysterectomy is incompatible with preserving fertility [8]. These limitations highlight the need to identify new molecular targets and therapeutic strategies.
Pre‑mRNA splicing is a fundamental step in gene expression that removes introns and ligates exons to produce mature mRNAs [9]. Alternative splicing, which affects nearly all multi‑exon genes, increases proteomic diversity and regulates many cellular programs [10]. Splicing is performed by the spliceosome, a dynamic complex composed of five small nuclear ribonucleoproteins (snRNPs: U1, U2, U4, U5 and U6) and numerous non‑snRNP proteins [11]. The assembly and nuclear import of U snRNPs require the formation of a heptameric Sm protein ring (SNRPB, SNRPD1, SNRPD2, SNRPD3, SNRPE, SNRPF and SNRPG) on snRNAs [12]. Dysregulated splicing is a pervasive feature of cancer and contributes to malignant phenotypes, including altered proliferation, invasion, apoptosis, immune evasion and therapeutic resistance [13]. Components of the spliceosome are frequently deregulated in tumors and can act as oncogenic drivers. For example, SNRPB facilitates ovarian cancer progression by modulating exon skipping events in POLA1 and BRCA2 [14], SNRPC drives triple‑negative breast cancer via the TNFAIP2–Rac1–β‑catenin axis [15], and SNRPD1 silencing inhibits breast cancer cell proliferation by inducing G0/G1 arrest [16]. SNRPD3, in cooperation with MYCN, alters splicing programs in neuroblastoma [17], and SNRPE influences FGFR4 expression through alternative splicing in hepatocellular carcinoma (HCC) [11].
SNRPD2 is a core component of the spliceosomal machinery [18] and has emerged as a candidate oncogenic Sm protein in diverse malignancies. It is markedly upregulated in hepatocellular carcinoma and promotes tumorigenesis by regulating intron retention in DDX39A [19]. In breast cancer, SNRPD2 has been implicated in sister chromatid cohesion and cell proliferation [20]. In colorectal cancer, it disrupts PABPN1 phase separation and accelerates cell proliferation and migration through the alternative polyadenylation of CTNNBIP1 [21]. Pan‑cancer analyses further revealed widespread SNRPD2 overexpression and a significant association between high SNRPD2 expression levels and a poor prognosis for patients; notably, SNRPD2 depletion results in selective cytotoxicity toward cancer cells in several settings [22]. Additionally, SNRPD2 has been shown to regulate DNA damage and BRCA1/FANC splicing in HCC, with its depletion increasing the sensitivity of HCC to PARP inhibitors; pharmacological targeting of SNRPD2 acetylation using romidepsin in combination with olaparib has shown substantial therapeutic potential [18]. Despite these advances, the precise role and underlying mechanisms of SNRPD2 in endometrial cancer have not yet been elucidated.
In this study, we show that SNRPD2 expression is significantly upregulated at both the mRNA and protein levels in endometrial cancer and that high SNRPD2 expression correlates with poor clinical outcomes. Functional experiments demonstrate that SNRPD2 depletion suppresses the proliferation and metastasis of endometrial cancer cells in vitro and reduces tumor growth in a PDX model when SNRPD2 is targeted by ASOs. Mechanistically, SNRPD2 knockdown induces the retention of intron 5 in DDX39B, resulting in the production of a transcript subject to NMD and decreased DDX39B expression. The loss of DDX39B permits the activation of a cryptic exon (Exon 2_3) in CTSC, which introduces premature termination codons and triggers additional NMD-mediated degradation, resulting in reduced CTSC expression. Together, these findings define a novel SNRPD2–DDX39B–CTSC regulatory axis that supports malignant traits in endometrial cancer and identify SNRPD2 as a potential therapeutic target.
Materials and methods
Data mining and bioinformatics analysis
Differentially expressed genes (DEGs) between endometrial cancer and normal endometrial tissues were identified using Gene Expression Profiling Interactive Analysis 2 (GEPIA2) (http://gepia2.cancer-pku.cn/#index, accessed on 2025‑03) [23]. The mRNA expression data from The Cancer Genome Atlas and Genotype-Tissue Expression (TCGA-GTEx) were retrieved from the UCSC Xena platform (https://xenabrowser.net/hub/, accessed on 2024‑12) [24], and protein expression data for endometrial cancer were obtained from the Clinical Proteomic Tumor Analysis Consortium (CPTAC) portal (https://gdc.cancer.gov/about-gdc/contributed-genomic-data-cancer-research/clinical-proteomic-tumor-analysis-consortium-cptac, accessed on 2021‑05) [25]. The prognostic value of SNRPD2 was assessed using Kaplan–Meier Plotter (https://kmplot.com/analysis/, accessed on 2024‑06) [26] and the University of Alabama at Birmingham CANCER Data Analysis Portal (UALCAN) (https://ualcan.path.uab.edu/, accessed on 2024‑06) [27]. The Gene Ontology (GO) enrichment analysis was performed using the Database for Annotation, Visualization and Integrated Discovery (DAVID) (https://davidbioinformatics.nih.gov/, accessed on 2024‑01) [28]. Heatmaps were generated with TBtools [29]. Alternative splicing events were detected from mapped RNA‑seq binary alignment/map (BAM) files using replicate Multivariate Analysis of Transcript Splicing (rMATS) software [30], and splicing patterns at the DDX39B and CTSC loci were visualized in Sashimi plots using Integrative Genomics Viewer (IGV).
Patient samples
Clinical tissue samples, including fresh‑frozen samples and FFPE tissues, were collected from Qilu Hospital, Shandong University. The procurement and utilization of these samples were approved by the Ethics Committee of Qilu Hospital.
Immunohistochemical (IHC) staining
Paraffin-embedded tissue sections were deparaffinized in xylene and rehydrated through a graded ethanol series. Antigen retrieval was performed in Tris-EDTA buffer (pH 9.0). Endogenous peroxidase activity was quenched by incubating the sections with a hydrogen peroxide solution. After washes with phosphate-buffered saline (PBS), the sections were blocked using the appropriate buffer and incubated with the primary antibody overnight at 4 °C. Following the primary antibody incubation, a biotinylated secondary antibody was applied, and signal amplification was achieved with a streptavidin–horseradish peroxidase (HRP) complex. Immunoreactivity was visualized using freshly prepared 3,3’-diaminobenzidine tetrahydrochloride (DAB). Sections were rinsed, counterstained, dehydrated, and mounted for evaluation. SNRPD2 (Abcam, ab198296) and Ki-67 (Cell Signaling Technology, #9449) antibodies were used for detection. The intensity of immunohistochemical staining was quantified using the histochemistry score (H‑score) method, which incorporates both the proportion of positively stained tumor cells and the staining intensity.
Cell culture
Ishikawa cells were obtained from the European Collection of Authenticated Cell Cultures (ECACC, United Kingdom), and AN3CA and HEC-1A cells were purchased from the American Type Culture Collection (ATCC, USA). Ishikawa cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco, USA) supplemented with 10% fetal bovine serum (FBS; Gibco). AN3CA cells were maintained in minimum essential medium (MEM; Gibco) containing 10% FBS. HEC-1A cells were cultured in McCoy’s 5 A medium (Gibco) supplemented with 10% FBS. All the cultures were incubated at 37 °C in a humidified atmosphere containing 5% CO_2_. All cell lines were confirmed free of mycoplasma contamination and authenticated by short tandem repeat profiling.
siRNA and plasmid transfection, and lentivirus infection
siRNA oligonucleotides targeting SNRPD2, DDX39B, and CTSC were synthesized by GenePharma (Shanghai, China). Transient transfections were performed using Lipofectamine 2000 (Thermo Fisher Scientific) according to the manufacturer’s instructions. The sequences of all the siRNAs and antisense oligonucleotides are provided in Supplementary Tables 1 and 2.
The open reading frame of SNRPD2 was amplified, cloned, and inserted into the PCMV expression vector to generate the SNRPD2 overexpression construct. Short hairpin RNAs (shRNAs) targeting SNRPD2, DDX39B and CTSC were subsequently cloned and inserted into the pLKO.1‑TRC backbone. Lentiviral particles were produced by co-transfecting HEK293T cells with the transfer vector together with psPAX2 and pMD2.G packaging plasmids. Viral supernatants were collected, filtered and used to transduce endometrial cancer cells in the presence of polybrene when indicated. Stable cell lines were selected with puromycin and subsequently validated for altered target gene expression by quantitative real-time PCR (qPCR) and/or western blotting.
RNA isolation, qPCR, and western blotting
The detailed protocols for these assays are described in a previous study [31]. The primer sequences used in this study are listed in Supplementary Table 3. For western blot analyses, protein extraction and analysis were performed using a previously established protocol [31], with the modifications described below. Protein concentrations were determined using the MERCK Millipore BCA Protein Assay Kit (catalog no. 71285-3) according to the manufacturer’s instructions. Exactly 30 µg of total protein from each sample was mixed with 5× SDS-PAGE loading buffer, boiled for 10 min, and separated on gels consisting of a 12% resolving gel and a 5% stacking gel. The proteins were transferred onto PVDF membranes (Millipore) using a Bio-Rad semidry transfer apparatus under the following conditions: 15 V for 60 min.
The membranes were blocked with 5% non-fat dry milk in TBST for 1 h at room temperature and then incubated overnight at 4 °C with primary antibodies against SNRPD2 (Abcam, ab198296), DDX39B (Abcam, ab181059), CTSC (Abcam, ab314644), and β-actin (Sigma-Aldrich, A5441). After three washes with TBST, the membranes were incubated with the appropriate HRP-conjugated secondary antibodies (KPL; 074-1506 or 074-1806) for 1 h at room temperature. The protein bands were visualized using ECL substrate (Millipore, ORT2655 and ORT2755) and imaged with a GE Amersham Imager 600. A densitometric analysis of protein bands was performed using ImageJ software, and band intensities were normalized to those of the corresponding β-actin loading control. The uncropped bands for the western blotting included in this study have been provided in the Supplementary Materials.
Colony formation, methylthiazolyldiphenyl-tetrazolium bromide (MTT), 5-ethynyl-2’-deoxyuridine (EdU), and Transwell assays
The proliferative and metastatic abilities of endometrial cancer cells were evaluated using MTT, EdU, colony formation, and Transwell assays. The detailed protocols for these assays are described in a previous study [31]. Quantitative data from the colony formation and Transwell migration and invasion assays are presented as the means ± SEMs from three independent biological experiments, each performed with three technical replicates.
Nude mouse xenograft model and patient-derived xenograft (PDX) model
Nude mouse xenograft and PDX models were performed according to the protocols outlined in a previous study [31]. Mice were randomly allocated to either the experimental group or the control group. The animal studies were executed utilizing a single-blind experimental design.
RNA-seq analysis
Ishikawa cells were transiently transfected with siRNAs or si-NC, and total RNA was extracted using TRIzol reagent 48 h after transfection. RNA-seq was performed by BioSune Biotechnology Company (China). DEGs were identified under the criteria: |log₂ fold change | ≥0.58 and Padj <0.05.
RNA immunoprecipitation (RIP) assays
RIP assays were performed in Ishikawa cells overexpressing Flag-SNRPD2 using a Flag antibody. The detailed protocols for RIP are described in a previous study [32]. RIP qPCR primers were designed within regions common to all the transcript variants to assess total DDX39B expression. Specifically, the forward primer was located in exon 9, and the reverse primer was located in exon 10.
Statistical analysis
All the experiments were conducted with three biological replicates. The data are shown as the means ± SEMs. Statistical significance was evaluated using unpaired two-tailed Student’s t tests, and analyses were performed with GraphPad Prism 5. A P value of less than 0.05 was considered to indicate statistical significance.
Results
SNRPD2 is a crucial core splicing factor in endometrial cancer
We intersected upregulated genes from TCGA–Uterine Corpus Endometrial Carcinoma (TCGA-UCEC) (log_2_FC ≥ 1, q < 0.01) and CPTAC–Uterine Corpus Endometrial Carcinoma (CPTAC–UCEC) (P < 0.05) datasets with a curated list of 134 core splicing factors to identify the core splicing factors that are dysregulated in endometrial cancer. This analysis yielded six candidate splicing factors (Fig. 1A). A proteomic evaluation of the CPTAC–UCEC dataset revealed the consistent upregulation of these genes at the protein level (heatmap, Fig. 1B). The expression of six factors—NAA38, HSPA5, HSPA4, SNRPD2, SNRPF, and SNRPG—was significantly increased at the mRNA level in tumors compared with normal endometrium from TCGA–UCEC and TCGA–GTEx cohorts and at the protein level in the CPTAC–UCEC cohort (Fig. 1C–E). Notably, SNRPD2, SNRPF and SNRPG belong to the Sm protein family.Fig. 1SNRPD2 is identified as a crucial core splicing factors in endometrial cancer.A Venn diagram showing the intersection of differentially expressed genes (DEGs) from TCGA‑UCEC, CPTAC‑UCEC and a curated list of 134 core splicing factors to identify candidate splicing regulators. B Heatmap of the differential protein expression of six selected splicing factors in CPTAC–UCEC proteomic data (tumor n = 100, normal n = 31). C Differential expression of the six selected splicing factors in TCGA-UCEC RNA-seq dataset (tumor, n = 177; normal, n = 24). Statistical analysis was performed using a two-tailed unpaired Student’s t test; exact P values are indicated in the panel. D Differential expression of the six selected splicing factors in the combined TCGA-GTEx dataset (tumor n = 181, normal n = 23). Statistical analysis was performed using a two-tailed unpaired Student’s t test; exact P values are indicated in the panel. E The differential expression patterns of six selected splicing factors between endometrial cancer (n = 100) and normal endometrial tissues (n = 31) from CPTAC–UCEC. Statistical analysis was performed using a two-tailed unpaired Student’s t test; exact P values are indicated in the panel. F Kaplan–Meier analysis of overall survival stratified by SNRPD2 expression (high versus low; each group n = 271). Statistical significance was assessed by the log-rank test. G qPCR analysis of SNRPD2 mRNA levels in endometrial cancer tissues (n = 16) compared to normal endometrial tissues (n = 16) (n = 3 biologically independent samples). H Quantitative analysis of SNRPD2 protein expression levels in normal endometrial and endometrial cancer tissues, based on western blotting data. I Western blotting analysis of SNRPD2 protein levels in endometrial cancer tissues (n = 12) compared to normal endometrial tissues (n = 12). J Representative immunohistochemistry images showing SNRPD2 staining in tumor and normal endometrium. K Immunohistochemical H-score quantification of SNRPD2 expression in endometrial cancer (n = 15) and normal endometrium (n = 8). P values were obtained by unpaired t tests (C, D, E, G, H, K) or Log-rank test (F). *P < 0.05, **P < 0.01.
Given the role of Sm proteins in RNA splicing, we prioritized SNRPD2, SNRPF and SNRPG for further analysis. Kaplan–Meier survival analysis using the KMplotter database revealed that high SNRPD2 expression was strongly associated with shorter overall survival, whereas SNRPF and SNRPG expression were not significantly associated with prognosis (Fig. 1F and Supplementary Fig. 1A). An independent analysis using UALCAN corroborated the negative prognostic association of SNRPD2 (Supplementary Fig. 1B). CPTAC pan‑cancer data from UALCAN further indicated that the SNRPD2 protein is frequently upregulated in multiple tumor types (Supplementary Fig. 1C). Based on these results, we selected SNRPD2 for detailed functional studies.
We evaluated the expression levels of SNRPD2 in clinical samples to further confirm these results. qPCR and Western blotting of fresh‑frozen specimens from Qilu Hospital revealed significantly higher SNRPD2 mRNA and protein levels in endometrial cancer tissues compared with normal endometrium (Fig. 1G–I). IHC data from The Human Protein Atlas showed a higher SNRPD2 staining intensity in endometrial cancer (7/12 high, 5/12 medium) than in normal endometrial tissue, which predominantly exhibited low–medium staining (Supplementary Fig. 1D). Consistent with these findings, our IHC staining of FFPE samples revealed the predominant nuclear localization of SNRPD2, with a significantly higher H‑score in endometrial cancer tissues than in normal endometrial tissues (Fig. 1J, K).
Overall, these findings suggest that SNRPD2 is markedly upregulated in endometrial cancer and is associated with a poor prognosis, supporting a potential role for SNRPD2 in endometrial tumorigenesis.
SNRPD2 regulates the proliferation and metastatic potential of endometrial cancer cells
We silenced SNRPD2 in endometrial cancer cell lines using two independent siRNAs to investigate its biological function. The knockdown efficiency was confirmed by qPCR and Western blotting (Fig. 2A, B and Supplementary Fig. 2A). SNRPD2 depletion reduced cell viability, as determined by MTT assays (Fig. 2C), and decreased the colony formation capacity of Ishikawa, AN3CA, and HEC-1A cells (Fig. 2D and Supplementary Fig. 2B). Consistent with these results, EdU incorporation assays showed a reduction in DNA synthesis after SNRPD2 knockdown (Fig. 2E, F and Supplementary Fig. 2C). Transwell migration and invasion assays revealed decreases in both migratory and invasive behaviors in endometrial cancer cells following SNRPD2 silencing (Fig. 2G and Supplementary Fig. 2D–F).Fig. 2SNRPD2 regulates proliferation and metastatic phenotypes of endometrial cancer cells.A, B The knockdown efficiency of SNRPD2 in endometrial cancer cells was verified by qPCR and Western blotting assays (n = 3 biologically independent samples). C The MTT assay was used to assess the impact of SNRPD2 knockdown in Ishikawa, AN3CA, and HEC-1A cells on cell proliferation (n = 3 biologically independent samples). D Colony formation assays were performed to evaluate the effects of SNRPD2 knockdown on the viability of Ishikawa, AN3CA, and HEC-1A cells (n = 3 independent experiments, triplicate wells; mean ± SEM; two‑tailed unpaired Student’s t test). (E-F) EdU incorporation assays were conducted to evaluate the fraction of DNA-replicating cells after SNRPD2 silencing in Ishikawa and AN3CA cells (n = 3 biologically independent samples). G The Transwell migration and invasion assay was conducted to assess the impact of SNRPD2 silencing on the metastasis potential of Ishikawa cells (n = 3 independent experiments, triplicate wells; mean ± SEM; two‑tailed unpaired Student’s t test). H Representative photographs of subcutaneous xenograft tumors in nude mice, comparing the PLKO.1-shSNRPD2 and PLKO.1-control groups (n = 6 mice per group). I Comparison of tumor weight between the PLKO.1-shSNRPD2 and PLKO.1-control groups. J Comparison of tumor volume between the PLKO.1-shSNRPD2 and PLKO.1-control groups. K Immunohistochemical staining of SNRPD2 and Ki-67 in the subcutaneous xenograft tumors. P value was obtained by unpaired t test. *P < 0.05, **P < 0.01.
Ishikawa cells stably expressing the SNRPD2 shRNA or control vector were implanted subcutaneously into nude mice to assess the effects in vivo. After 15 days, tumors from the SNRPD2-knockdown group were smaller in volume and weight than those from the control group (Fig. 2H–J). IHC staining of the xenograft tissues confirmed that SNRPD2 expression was reduced and that the Ki-67 index was lower in the knockdown group than in the control group (Fig. 2K).
Together, these in vitro and in vivo data indicate that SNRPD2 promotes proliferation and metastatic traits in endometrial cancer cells, supporting a functional role for SNRPD2 in tumor progression.
ASO-mediated silencing of SNRPD2 suppresses the proliferation capacity in endometrial cancer cells
Compared with other classes of small molecules, ASOs offer high target specificity and reduced off-target effects, making them a viable strategy for silencing oncogenic mRNAs. We therefore designed ASOs targeting SNRPD2 and evaluated their effects on endometrial cancer.
Ishikawa, AN3CA and HEC-1A cells were treated with SNRPD2-specific ASOs. Western blot analysis confirmed a reduction in SNRPD2 protein levels after ASO treatment (Fig. 3A and Supplementary Fig. 2G). Functionally, SNRPD2 silencing decreased cell viability in MTT assays and reduced clonogenic potential in colony formation assays (Fig. 3B, C and Supplementary Fig. 2H), indicating an impaired proliferative capacity in vitro. We used a PDX model of endometrial cancer to assess antitumor activity in vivo. Compared with ASO-NC treatment, SNRPD2-ASO treatment resulted in a reduction in tumor volume and tumor weight (Fig. 3D–F). IHC staining of PDX tumors showed lower SNRPD2 expression and a decreased Ki-67 proliferation index in the SNRPD2-ASO group (Fig. 3G).Fig. 3ASO-mediated SNRPD2 silencing inhibits proliferation of endometrial cancer cells.A Western blotting analysis assessing the impact of SNRPD2-ASO on SNRPD2 expression in endometrial cancer cells. B, C The effects of SNRPD2-ASO on the proliferation and colony formation abilities of endometrial cancer cells were evaluated by MTT and clonogenic assays (n = 3 independent experiments, triplicate wells; mean ± SEM; two‑tailed unpaired Student’s t test). D Representative photographs of endometrial cancer PDX tumors excised from NCG mice in both the SNRPD2-ASO and ASO-NC groups (n = 4 mice per group). E, F Tumor volume and weight of endometrial cancer PDX tumors were measured and compared between the SNRPD2-ASO and ASO-NC groups (n = 4 mice per group). G Immunohistochemical staining of SNRPD2 and Ki-67 in the PDX tumors. P value was obtained by unpaired t test. *P < 0.05, **P < 0.01.
Together, these data demonstrate that ASO-mediated knockdown of SNRPD2 expression suppresses endometrial cancer cell proliferation both in vitro and in vivo, supporting the potential of SNRPD2-targeting ASOs as a therapeutic approach for endometrial cancer.
Identification of SNRPD2 downstream targets in endometrial cancer cells
We performed RNA-seq after transient SNRPD2 knockdown in Ishikawa cells to investigate the molecular mechanisms by which SNRPD2 contributes to endometrial cancer progression. The differential expression analysis identified 3549 DEGs ( | log_2_FC | ≥0.58, Padj <0.05), of which 2004 were downregulated and 1545 were upregulated (Fig. 4A). The GO analysis of these DEGs revealed enrichment in processes related to transcriptional regulation, cell adhesion, protein ubiquitination, signal transduction, transmembrane transport, lipid metabolism, and translation (Fig. 4B).Fig. 4. Identification of SNRPD2 downstream targets in endometrial cancer cells.A Heatmap showing DEGs following SNRPD2 knockdown in Ishikawa cells. B GO analysis of DEGs after SNRPD2 silencing in Ishikawa cells using DAVID tools. C AS profile comparing the twelve basic AS events derived from RNA-seq data following SNRPD2 knockdown in Ishikawa cells. D Statistical analysis of differential AS events performed by rMATS following SNRPD2 silencing in Ishikawa cells. E GO analysis of genes related to differential intron retention AS events following SNRPD2 knockdown in Ishikawa cells. F Venn diagram illustrating the overlap of genes related to intron retention AS, downregulated genes after SNRPD2 silencing, and genes with a correlation coefficient ≥0.35 from the SNRPD2 co-expression analysis in TCGA-UCEC data. G Heatmap of RNA-seq expression data showing the expression patterns of 24 selected candidate genes following SNRPD2 knockdown in Ishikawa cells. H Correlation analysis of SNRPD2 and the 24 candidate genes based on TCGA-UCEC data (tumor n = 181). I Sashimi plot of DDX39B generated from RNA-seq BAM data, illustrating splicing patterns following SNRPD2 depletion in Ishikawa cells. J RT–PCR analysis was conducted using primers spanning exon 5 and exon 6 to determine the retention of intron 5 (where intron 5 refers to the 5th intron in DDX39B-201) in DDX39B following SNRPD2 knockdown in endometrial cancer cells (n = 3 biologically independent samples). K qPCR analysis of the effect of SNRPD2 knockdown on DDX39B mRNA expression levels in Ishikawa, AN3CA, and HEC-1A cells (n = 3 biologically independent samples). L Western blotting analysis of the impact of SNRPD2 knockdown on DDX39B protein expression levels in Ishikawa, AN3CA, and HEC-1A cells. P value was obtained by unpaired t test. *P < 0.05, **P < 0.01.
Given the role of SNRPD2 in splicing, we next examined changes in alternative splicing using ASprofile and rMATS. The ASprofile analysis revealed altered frequencies across 12 defined splicing categories—single exon skipping (SKIP), multiple exon skipping (MSKIP), approximate exon skipping (XSKIP), multiple approximate exon skipping (MXSKIP), intron retention (IR), multiple intron retention (MIR), approximate intron retention (XIR), multiple approximate intron retention (XMIR), alternative exon ends (AE), approximate alternative exon ends (XAE), alternative first exon (AFE) and alternative last exon (ALE)—after SNRPD2 depletion, with notable increases in XAE, XIR, IR, MIR, and MSKIP events (Fig. 4C). rMATS detected 8,685 differential splicing events ( | IncLevelDifference | >0.1, FDR < 0.05), comprising exon skipping (ES, 68.94%), mutually exclusive exons (MXE, 11.53%), intron retention (IR, 9.21%), and alternative 3′/5′ splice site usage (A3SS/A5SS, 10.32%) (Fig. 4D). Considering that both ASprofile and rMATS consistently indicated increased intron retention, we focused our subsequent analyses on IR events. Using our selection criteria, 800 differential IR events were identified. The GO analysis of genes harboring such IR events revealed significant enrichment in chromatin remodeling, protein transport, DNA repair, cell cycle regulation, the DNA damage response, mRNA splicing, and methylation (Fig. 4E).
We first examined genes with differential IR events, downregulated DEGs (fold change ≤ 0.75, Padj <0.05), and genes positively coexpressed with SNRPD2 in TCGA–UCEC cohort (Pearson r ≥ 0.35) to systematically identify candidate downstream effectors of SNRPD2 in endometrial cancer. This intersection yielded 24 candidate genes (Fig. 4F), whose changes in expression following SNRPD2 knockdown are depicted in the corresponding heatmap (Fig. 4G). We performed a parallel integrative analysis of differential IR events, upregulated DEGs (fold change ≥1.3, Padj <0.05), and genes positively coexpressed with SNRPD2 in further efforts to capture the full spectrum of SNRPD2‑associated splicing alterations. This approach identified 18 additional candidate genes (Supplementary Fig. 3A, B), for which the correlation with SNRPD2 expression is shown in Supplementary Fig. 3C. Given that SNRPD2 functions as an oncogenic splicing factor, genes that were downregulated upon its depletion, particularly those implicated in IR, may also possess oncogenic potential. Such genes often encode proteins with well‑defined functional domains that can be effectively targeted by small molecules or monoclonal antibodies with limited effect on normal tissues, thus representing promising avenues for precision cancer therapy. Based on this rationale, subsequent analyses concentrated on those oncogenic candidates that were both suppressed upon SNRPD2 knockdown and involved in IR events, thereby refining the focus on splicing-regulated drivers of tumorigenesis.
A correlation analysis of TCGA–UCEC indicated that 22 of these 24 genes (all except EWSR1 and MTA1) had a Pearson correlation coefficient > 0.30 with SNRPD2; these genes were ranked by correlation strength (Fig. 4H). Sashimi plots generated from RNA-seq BAM files revealed that, among these candidates, only intron 5 of DDX39B (in the transcript DDX39B-201, Ensembl) showed clear and significant intron retention following SNRPD2 knockdown (Fig. 4I). We experimentally validated this intron retention event through RT–PCR analysis of DDX39B following SNRPD2 knockdown using primers spanning exon 5 and exon 6. As shown in Fig. 4J, the depletion of SNRPD2 led to a pronounced increase in the number of transcripts retaining intron 5, accompanied by a corresponding decrease in the number of transcripts with normal splicing of this intron. Moreover, SNRPD2 and DDX39B expression were positively correlated at both the mRNA and protein levels in TCGA–UCEC and CPTAC–UCEC datasets (correlation coefficients of 0.44 and 0.48, respectively; Supplementary Fig. 3D, E).
We measured DDX39B expression after SNRPD2 knockdown in endometrial cancer cells to validate that DDX39B is a functional downstream target of SNRPD2. qPCR and Western blotting showed reductions in both the DDX39B mRNA and protein levels upon SNRPD2 depletion (Fig. 4K, L and Supplementary Fig. 3F, G). Together, these data indicate that SNRPD2 regulates the intron retention of specific transcripts and that DDX39B is a key downstream target whose expression is reduced following SNRPD2 loss.
SNRPD2 regulates intron retention in DDX39B in endometrial cancer cells
We examined RNA-seq from SNRPD2-silenced Ishikawa cells to define how SNRPD2 affects DDX39B expression and observed the retention of intron 5 in the DDX39B transcript (visualized in IGV) (Fig. 5A). A review of Ensembl transcript annotations revealed that DDX39B-201, -202 and -214 are protein-coding isoforms, whereas DDX39B-215 retains intron 5 that introduces PTCs and therefore is predicted to be noncoding (Fig. 5A). An analysis of TCGA–UCEC transcript data indicated that DDX39B-201 is the predominant protein-coding isoform expressed in tumor and is expressed at higher levels than DDX39B-215 (Fig. 5B, C). CPTAC proteomic data further confirmed that the DDX39B protein abundance was higher in endometrial cancer tissue compared with normal endometrium (Fig. 5D).Fig. 5SNRPD2 regulates intron retention of DDX39B in endometrial cancer cells.A Information on the DDX39B-201 (ENST00000376177), DDX39B-202 (ENST00000396172), DDX39B-214(ENST00000458640), and DDX39B-215 (ENST00000462256) transcripts retrieved from the Ensembl database. B Differential expression levels of DDX39B-201, DDX39B-202, DDX39B-214, and DDX39B-215 in endometrial cancer tissues (n = 181) from the TCGA-UCEC database. C Differential expression levels of DDX39B-total, DDX39B-201, and DDX39B-215 in endometrial cancer tissues (n = 181) from the TCGA-UCEC database. D Differential expression analysis of DDX39B protein between normal endometrial (n = 31) and endometrial cancer (n = 100) tissues from the CPTAC–UCEC database. E Correlation analysis between SNRPD2 and DDX39B-215 expression using TCGA-UCEC data (tumor n = 181). F Differential expression analysis of DDX39B-201 across TCGA pan-cancer datasets from the GEPIA2 database. G Differential expression analysis of DDX39B-215 across TCGA pan-cancer datasets from the GEPIA2 database. H Schematic representation of the normal splicing transcript DDX39B-201 and the aberrant splicing transcript DDX39B transcripts with intron 5 retention (where intron 5 refers to the 5th intron in DDX39B-201). The schematic also illustrates the locations of qPCR and RT–PCR primers, as well as the structure of the minigene construct used in these assays. I qPCR was used to assess the proportion of DDX39B transcripts with intron 5 retention after SNRPD2 inhibition in endometrial cancer cells (n = 3 biologically independent samples). J RT–PCR analysis was conducted using intron-flanking primers to determine the retention of intron 5 (where intron 5 refers to the 5th intron in DDX39B-201) in DDX39B following SNRPD2 knockdown in HEK293T cells (n = 3 biologically independent samples). K The mRNA ratio of DDX39B transcripts with intron 5 retention to the normally spliced transcript DDX39B-201 bands from (J) was quantified using ImageJ analysis. L The mRNA expression levels of DDX39B were assessed in SNRPD2 knockdown cells co-transfected with siUPF1 (n = 3 biologically independent samples). M The mRNA expression levels of DDX39B were evaluated in SNRPD2 knockdown cells treated with CHX (n = 3 biologically independent samples). N RIP-qPCR assay was performed to assess the binding of SNRPD2 to DDX39B mRNA in SNRPD2-overexpressing Ishikawa cells using anti-Flag antibody (n = 3 biologically independent samples). The P values were calculated using the Pearson’s correlation test (E). P value was obtained by unpaired t test (B–D, I, K–N). *P < 0.05, **P < 0.01.
Across endometrial cancer samples, the expression of the intron-retaining isoform DDX39B-215 correlated negatively with that of SNRPD2 (Pearson r = −0.37), which is consistent with a model in which SNRPD2 loss promotes intron retention (Fig. 5E). The pan-cancer analysis revealed that the expression of the DDX39B-201 isoform was upregulated in most malignancies, whereas the expression of DDX39B-215 was generally downregulated in tumors compared with normal tissues (Fig. 5F, G and Supplementary Fig. 3H).
We directly tested whether SNRPD2 modulates intron 5 retention by designing qPCR primers specific for intron 5 and measuring the proportion of DDX39B transcripts with intron 5 retention to total DDX39B after SNRPD2 knockdown. SNRPD2 depletion increased the proportion of intron-retaining DDX39B transcripts (Fig. 5H, I). Moreover, we constructed a minigene containing exons 5–6 of *DDX39B-*201 and the intron 5 sequence and co-transfected it with the SNRPD2 siRNA. RT–PCR with exon 5/6-specific primers was performed to assess changes in the expression of DDX39B with intron 5 retention and the normally spliced DDX39B transcript (Fig. 5H). The abundance and relative proportion of this intron-retaining product increased after SNRPD2 silencing (Fig. 5J, K), confirming that SNRPD2 loss promotes intron 5 retention in the DDX39B pre-mRNA.
We next tested whether the intron-retaining DDX39B transcripts are degraded by NMD. Knockdown of UPF1, a core NMD factor, partially restored DDX39B mRNA levels in SNRPD2-silenced Ishikawa cells (Fig. 5L). Similarly, treatment with cycloheximide (CHX), which inhibits NMD, increased DDX39B mRNA abundance following SNRPD2 depletion (Fig. 5M). These results indicate that the retention of intron 5 produces NMD-sensitive DDX39B transcripts, resulting in reduced DDX39B protein expression upon SNRPD2 loss. Finally, RIP using Flag-tagged SNRPD2 in Ishikawa cells showed the enrichment of DDX39B mRNA in SNRPD2 immunoprecipitates, suggesting that SNRPD2 binds to the DDX39B mRNA in endometrial cancer cells (Fig. 5N).
Together, these data indicate that SNRPD2 maintains the proper splicing of DDX39B by preventing intron 5 retention. The loss of SNRPD2 promotes the production of a PTC-containing DDX39B isoform that is targeted by NMD, thereby reducing functional DDX39B expression levels in endometrial cancer cells.
Targeting DDX39B inhibits the proliferation and metastatic behavior of endometrial cancer cells
We evaluated the effects of DDX39B depletion in vitro and in vivo to determine the role of DDX39B in endometrial cancer. Transient transfection of a DDX39B-specific siRNA in endometrial cancer cells resulted in efficient knockdown at both the mRNA and protein levels, as confirmed by qPCR and western blotting (Fig. 6A, B and Supplementary Fig. 4A).Fig. 6. Targeting DDX39B inhibits the malignant progression of endometrial cancer.A, B The effectiveness of DDX39B knockdown was confirmed by qPCR and western blotting (n = 3 biologically independent samples). C Growth curve assay showing the effect of DDX39B knockdown on the proliferative capacity of Ishikawa, AN3CA, and HEC-1A cells (n = 3 biologically independent samples). D Colony formation assay assessing the impact of DDX39B inhibition on clonogenic ability in Ishikawa, AN3CA, and HEC-1A cells (n = 3 independent experiments, triplicate wells; mean±SEM; two‑tailed unpaired Student’s t test). E, F EdU incorporation assays to assess the proportion of DNA-replicating cells following SNRPD2 silencing in Ishikawa and AN3CA cells (n = 3 biologically independent samples). G Subcutaneous tumor formation assay to assess the influence of DDX39B silencing on tumor growth in endometrial cancer (n = 6 mice per group). H Comparison of tumor volume between the PLKO.1-shDDX39B and PLKO.1-ctrl groups (n = 6 mice per group). I Comparison of tumor weight between the PLKO.1-shDDX39B and PLKO.1-ctrl groups (n = 6 mice per group). J, K MTT and colony formation assays were performed to assess the role of DDX39B in SNRPD2-mediated tumor proliferation and clonogenic abilities in Ishikawa cells (n = 3 independent experiments, triplicate wells; mean ± SEM; two‑tailed unpaired Student’s t test). L Transwell assays were conducted to evaluate the role of DDX39B in SNRPD2-mediated tumor migration in Ishikawa cells (n = 3 independent experiments, triplicate wells; mean ± SEM; two‑tailed unpaired Student’s t test). P value was obtained by unpaired t test. *P < 0.05, **P < 0.01.
Functionally, MTT assays revealed a reduction in the proliferation rate of all three cell lines after DDX39B silencing (Fig. 6C). Colony formation assays showed a decrease in the clonogenic capacity following DDX39B knockdown (Fig. 6D and Supplementary Fig. 4B). Consistent with these findings, EdU incorporation assays revealed a lower fraction of S-phase cells in the Ishikawa and AN3CA cell lines following DDX39B depletion, reflecting impaired DNA replication (Fig. 6E, F). Furthermore, Transwell assays revealed that DDX39B deficiency suppressed both migration and invasion (Supplementary Fig. 4C–F).
We next assessed the oncogenic role of DDX39B in vivo using a subcutaneous xenograft model. Ishikawa cells with stable DDX39B knockdown (shRNA) produced tumors with smaller volume and weight compared with control (sh-Ctrl) xenografts (Fig. 6G–I), supporting a key role for DDX39B in tumor growth.
We tested whether DDX39B mediates the pro-tumorigenic effects of SNRPD2 by knocking down DDX39B in SNRPD2-overexpressing Ishikawa cells. The reduction in DDX39B expression abrogated the increases in proliferation, colony formation, migration driven by SNRPD2 overexpression, as shown by the results of the MTT, colony formation and Transwell assays (Fig. 6J–L and Supplementary Fig. 4G, H).
Together, these results indicate that DDX39B is required for the proliferative and metastatic characteristics in endometrial cancer cells and functions downstream of SNRPD2 to mediate its oncogenic activity.
CTSC functions as a key downstream target of DDX39B in endometrial cancer
We performed RNA sequencing of Ishikawa cells following DDX39B knockdown to identify the downstream effectors of DDX39B. The differential expression analysis identified 785 DEGs ( | log₂FC | ≥0.58, Padj <0.05), comprising 457 downregulated genes and 328 upregulated genes (Fig. 7A). The GO analysis of the downregulated genes showed that processes related to transcriptional regulation, protein stability, mRNA processing, cell proliferation, and protein glycosylation were enriched (Fig. 7B).Fig. 7CTSC is a key downstream target gene of DDX39B in endometrial cancer.A Heatmap displaying DEGs following DDX39B knockdown in Ishikawa cells. B GO analysis of downregulated DEGs after DDX39B silencing in Ishikawa cells using DAVID tools. C Statistical analysis of differential alternative splicing events conducted by rMATS following DDX39B knockdown in Ishikawa cells. D GO analysis of genes associated with differential exon skipping/inclusion events following DDX39B knockdown in Ishikawa cells. E Venn diagram illustrating the overlap of genes related to intron retention AS (n = 737) and downregulated genes (n = 457) after DDX39B silencing. F Heatmap of RNA-seq expression data showing the expression patterns of 19 key downstream genes following DDX39B knockdown in Ishikawa cells. G Analysis of differential mRNA expression levels of 19 key downstream genes in endometrial cancer tissues (n = 181) and normal endometrial tissues (n = 23) based on TCGA-UCEC data. H Analysis of differential protein expression levels of the key downstream genes in endometrial cancer tissues and normal endometrial tissues based on CPTAC–UCEC data (tumor n = 100, normal n = 31). I Correlation analysis between DDX39B expression levels and those of key downstream genes using CPTAC–UCEC data (tumor n = 100, normal n = 31). J Western blotting analysis of the impact of DDX39B knockdown on CTSC protein expression levels in Ishikawa, AN3CA, and HEC-1A cells. P value was obtained by unpaired t test. *P < 0.05, **P < 0.01.
Given that DDX39B functions in pre-mRNA splicing and RNA export, we assessed changes in alternative splicing using rMATS ( | IncLevelDifference | >0.1, FDR < 0.05). We identified 1087 differential splicing events, categorized as exon skipping/inclusion (ES/EI, 67.8%), MXE (7.0%), IR (10.4%), and A3SS/A5SS (14.6%) (Fig. 7C). Because ES/EI events predominated, subsequent analyses focused on genes affected by differential ES/EI following DDX39B depletion. The GO enrichment analysis of these genes indicated that they are involved in transcriptional regulation, phosphorylation, mRNA processing, microtubule organization, cell cycle control, autophagy, protein glycosylation, and homologous recombination repair (Fig. 7D).
We intersected downregulated DEGs with genes showing differential ES/EI to prioritize functionally relevant targets and identified 19 candidate downstream genes (Fig. 7E). RNA-seq heatmaps confirmed that all 19 genes were downregulated upon DDX39B knockdown (Fig. 7F). An analysis of TCGA–UCEC data showed that the expression of VDR, CTSC, P2RY6, TGM5, and R3HDM1 was significantly higher in tumors than in normal endometrium (Fig. 7G). CPTAC proteomic data further demonstrated that CTSC protein expression was higher in endometrial cancer tissue than in normal tissue, whereas the expression of several other candidates (RAP1B, TNRC6C, DAG1, LIX1L, and NFIC) was higher in the normal endometrium (Fig. 7H).
The correlation analysis using CPTAC–UCEC data revealed that CTSC protein levels were most strongly positively correlated with DDX39B protein expression (Pearson r = 0.41) among the candidates (Fig. 7I). The observed IncLevelDifference for CTSC was 0.138, indicating that DDX39B knockdown promotes exon inclusion in CTSC transcripts. Importantly, Western blotting confirmed that DDX39B silencing reduced CTSC protein levels in endometrial cancer cells (Fig. 7J and Supplementary Fig. 4I, J).
Collectively, these findings identify CTSC as a key gene regulated by DDX39B in endometrial cancer. The loss of DDX39B induces exon inclusion in CTSC transcripts and is accompanied by decreased CTSC protein expression, suggesting that altered CTSC splicing may contribute to the tumor-promoting effects of DDX39B.
The inhibition of CTSC reduces the proliferative potential of endometrial cancer cells
We silenced CTSC in Ishikawa, AN3CA, and HEC-1A cells using two independent siRNAs to investigate its role in endometrial cancer. Efficient knockdown at both the mRNA and protein levels was confirmed by qPCR and Western blotting (Supplementary Fig. 5A, B and Supplementary Fig. 6A). Functional assays showed that CTSC depletion impaired cell growth: MTT assays revealed a reduction in proliferation across all three cell lines, and colony formation assays revealed a decrease in clonogenic potential following CTSC knockdown (Supplementary Fig. 5C, D and Supplementary Fig. 6B).
We next assessed the effect of CTSC loss in vivo. Compared with control tumors, Ishikawa cells with stable, shRNA-mediated CTSC knockdown generated subcutaneous xenografts with smaller volumes and weights (Supplementary Fig. 5E–G), supporting a role for CTSC in tumor growth.
We determined whether CTSC mediates the pro-tumorigenic activity of DDX39B by knocking down CTSC in Ishikawa cells engineered to overexpress DDX39B. CTSC depletion attenuated the increased proliferation, colony formation and migratory/invasive behaviors driven by DDX39B overexpression, as shown by MTT, colony formation and Transwell assays (Supplementary Fig. 5H–K and Supplementary Fig. 6C, D).
Together, these data indicate that CTSC is required for the DDX39B-mediated promotion of proliferation and metastasis abilities in endometrial cancer cells and support CTSC as a potential therapeutic target.
Targeting DDX39B activates a cryptic exon in CTSC and reduces CTSC expression in endometrial cancer
We analyzed RNA‑seq BAM files and generated Sashimi plots in IGV to determine how DDX39B affects CTSC splicing. Knockdown of DDX39B in Ishikawa cells induced the inclusion of novel exonic sequences within the intron between CTSC exons 2 and 3, hereafter referred to as Exon 2_3 and Exon 2_3‑2 (Fig. 8A). The inclusion of these sequences was rare in control cells but became prominent upon DDX39B depletion.Fig. 8. Targeting DDX39B activates cryptic exon of CTSC in endometrial cancer.A Sashimi plot of CTSC generated from RNA-seq BAM data, illustrating splicing patterns following DDX39B depletion in Ishikawa cells. B Information on the CTSC-201 (ENST00000227266) and CTSC-216 (ENST00000677796) transcripts retrieved from the Ensembl database. C Information on the CTSC-202 (ENST00000393301), CTSC-203 (ENST00000524463), CTSC-205 (ENST00000528020), CTSC-206 (ENST00000529974), CTSC-216 (ENST00000677796), CTSC-217 (ENST00000677802), CTSC-218 (ENST00000677877), and CTSC-228 (ENST00000679032) transcripts retrieved from the Ensembl database. D Differential expression levels of CTSC-total and *CTSC-*201 in endometrial cancer tissues (n = 181) from the TCGA-UCEC database. E Differential expression levels of CTSC-201 between normal endometrial tissues (EM, n = 23) and endometrial cancer (UCEC, n = 181) tissues from the TCGA-UCEC database. F Differential expression levels of CTSC-201, CTSC-202, CTSC-203, CTSC-205, and CTSC-206 in endometrial cancer tissues (n = 181) from the TCGA-UCEC database. G Differential expression levels of CTSC-201, CTSC-202, CTSC-203, CTSC-205, and CTSC-206 in endometrial cancer cell lines (n = 28) from the CCLE-UCEC database. H Schematic representation of the normal splicing isoform CTSC-201 (CTSC-L) and the aberrant splicing isoform CTSC with inclusion of Exon 2_3 (CTSC-S). The diagram also highlights the locations of qPCR and RT–PCR primers, as well as the position of the PTCs. I qPCR analysis to evaluate the expression level of Exon 2_3 inclusion in CTSC following DDX39B inhibition in Ishikawa, AN3CA, and HEC-1A cells (n = 3 biologically independent samples). J Effect of Actinomycin D treatment on mRNA stability of endogenous CTSC-Total and CTSC-S transcripts in Ishikawa cells (n = 3 biologically independent samples). K Assessment of the mRNA stability of total CTSC expression in Ishikawa cells following DDX39B knockdown was conducted using Actinomycin D treatment (n = 3 biologically independent samples). L The mRNA expression level of CTSC was assessed in DDX39B knockdown cells co-transfected with siUPF1(n = 3 biologically independent samples). M The mRNA expression level of CTSC was evaluated in DDX39B knockdown cells treated with CHX (n = 3 biologically independent samples). N MTT assay was performed to evaluate the impact of CTSC-L and CTSC-S overexpression on cell proliferation in Ishikawa cells (n = 3 biologically independent samples). O Colony formation assay was conducted to compare the effects of CTSC-L and CTSC-S overexpression on the clonogenic ability of Ishikawa cells (n = 3 independent experiments, triplicate wells; mean ± SEM; two‑tailed unpaired Student’s t test). P Transwell assay was used to assess the effects of CTSC-L and CTSC-S overexpression on the metastatic potential of Ishikawa cells (n = 3 independent experiments, triplicate wells; mean ± SEM; two‑tailed unpaired Student’s t test). P value was obtained by an unpaired t test. *P < 0.05, **P < 0.01.
Ensembl annotation indicated that CTSC-201 (ENST00000227266) is the predominant transcript encoding the full‑length CTSC protein (463 amino acids) (Fig. 8B). In contrast, transcripts that include Exon 2_3 (for example, CTSC-216, ENST00000677796, and several other annotated isoforms) introduce PTCs and encode substantially truncated peptides (21–165 aa), predicting a loss of function and susceptibility to NMD (Fig. 8B, C). These observations identify Exon 2_3 as a cryptic exon whose inclusion disrupts the coding potential of CTSC.
An analysis of TCGA–UCEC and Cancer Cell Line Encyclopedia (CCLE) data revealed that CTSC-201 is the dominant isoform expressed in both primary tumors and cell lines, whereas CTSC isoforms harboring Exon 2_3 are expressed at very low levels (Fig. 8D–G), supporting the requirement for cryptic exon exclusion to maintain productive CTSC expression in endometrial tissue.
We validated the change in splicing by qPCR using primers targeting total CTSC and the Exon 2_3 inclusion event (Fig. 8H). DDX39B knockdown reduced total CTSC expression and increased the proportion of transcripts containing Exon 2_3 (designated CTSC-S) relative to the normally spliced transcript (CTSC-L) in endometrial cancer cells (Fig. 8I). To assess the stability of CTSC isoforms, we performed mRNA stability assays in Ishikawa cells by measuring endogenous transcripts. Due to sequence limitations, primers specific for CTSC-L spanning the exon 2–exon 3 junction could not be designed. Instead, two primer sets were used: one detecting all CTSC transcripts (CTSC-total), including CTSC-L, and another specifically amplifying CTSC-S by targeting cryptic exons within the intron between exons 2 and 3 as well as the adjacent exon 3. After treatment with actinomycin D to inhibit transcription, total RNA was collected at 0, 4, and 8 h. qPCR results revealed that the CTSC-S transcript decayed much faster than the total CTSC transcripts, indicating that CTSC-S is significantly less stable (Fig. 8J). Furthermore, depletion of DDX39B reduced the stability of total CTSC mRNA (Fig. 8K).
We tested whether the exon-inclusion transcripts are degraded by NMD by silencing UPF1 and treating cells with CHX. Both UPF1 knockdown and CHX treatment increased CTSC mRNA levels in DDX39B-silenced cells, indicating that the cryptic exon-containing transcripts are NMD targets (Fig. 8L, M).
Finally, we generated expression constructs for CTSC-L and CTSC-S and reintroduced them after the siRNA-mediated depletion of endogenous CTSC to assess the functional consequences of the two isoforms. Compared with the control, the overexpression of CTSC-L, but not CTSC-S, restored and increased cell proliferation and invasive behavior, indicating that only the full-length isoform is functionally active in promoting tumor phenotypes (Fig. 8N–P and Supplementary Fig. 6E, F).
Collectively, these results demonstrate that DDX39B suppresses the activation of the cryptic Exon 2_3 in the CTSC pre‑mRNA. The loss of DDX39B leads to cryptic exon inclusion, the generation of PTC-containing NMD-sensitive CTSC transcripts, reduced CTSC protein expression, and impaired CTSC-dependent oncogenic functions in endometrial cancer cells.
Discussion
Alternative splicing is a critical post‑transcriptional mechanism that increases transcript and protein diversity and plays a pivotal role in cancer biology. Given their frequent dysregulation in various malignancies, core Sm proteins of the spliceosome have emerged as promising therapeutic targets. Previous studies have shown that elevated Sm protein levels are common in cancer and that the depletion of SNRPD2 suppresses the proliferation of hepatocellular carcinoma (HCC) and breast cancer cells [19, 20]. In this study, we extend these findings to endometrial cancer: SNRPD2 is overexpressed in tumor tissues, and its high expression is associated with poor clinical outcomes. Functional assays confirmed that SNRPD2 silencing attenuates both the proliferative and metastatic capabilities of endometrial cancer cells both in vitro and in vivo.
Mechanistically, we demonstrate that SNRPD2 supports oncogenic splicing programs by maintaining the correct processing of the DDX39B transcript. The loss of SNRPD2 increases intron retention within DDX39B, generating PTC‑containing transcripts that are degraded via NMD, thereby reducing DDX39B expression. This mechanism is consistent with previous reports that SNRPD2 regulates intron retention in DDX39A [19]. DDX39B is a multifunctional RNA helicase that is required for mRNA splicing and export [33, 34]. Previous work has shown that DDX39B participates in diverse biological processes: for example, the modulation of FOXP3 splicing and regulatory T‑cell fate [35], protection against sorafenib‑induced ferroptosis through GPX4 splicing and export in HCC [36], and the promotion of metastasis via FUT3 splicing in colorectal cancer [37]. Our data establish that DDX39B serves as a crucial downstream effector of SNRPD2 in endometrial cancer and indicate that a reduction in its expression contributes to the antitumor effects of SNRPD2 depletion.
Further downstream, we identified CTSC as a functionally relevant splicing target of DDX39B. Knockdown of DDX39B activates a cryptic exon (Exon 2_3) within the CTSC pre‑mRNA. The inclusion of this exon introduces PTCs, generating truncated or nonfunctional isoforms and triggering NMD, thereby eliminating the expression of functional CTSC. CTSC has been implicated in multiple cancers and can promote tumor growth and metastasis through diverse pathways, including autophagy/ER stress in colorectal cancer [38], Wnt/β-catenin signaling in bladder cancer [39], TNF-α/p38 MAPK signaling in HCC [40], and the modulation of neutrophil infiltration in breast cancer metastasis [41]. Our experiments show that only the full-length CTSC isoform restores proliferation and invasive behavior, supporting the functional importance of correct CTSC splicing in endometrial cancer.
The activation of cryptic exons coupled to NMD has become an increasingly recognized consequence of splicing perturbations [42–44]. RNA‑binding proteins such as TDP‑43 normally safeguard transcript integrity by preventing toxic cryptic exon inclusion in healthy cells; the loss of these factors promotes aberrant splicing events that disrupt the expression of essential genes, as exemplified by STMN2 and UNC13A in neurodegenerative diseases [45]. Our findings suggest that DDX39B plays a similar protective role in maintaining CTSC expression: when DDX39B levels decrease, such as following SNRPD2 depletion, a cryptic exon becomes exposed, leading to the production of CTSC transcripts containing PTCs that are degraded via NMD and thereby compromising CTSC protein output.
These results have translational implications. First, core spliceosomal components such as SNRPD2 may be therapeutic targets in tumors that depend on increased splicing activity. ASOs and other splicing modulators are advancing clinically and can be used to target spliceosome components or their transcripts [46]. In our study, SNRPD2-targeting ASOs reduced tumor growth in a PDX model, providing proof‑of‑concept for this approach. Second, interventions that specifically correct pathogenic splicing alterations in key downstream effectors, such as splice‑switching oligonucleotides to block CTSC cryptic exon inclusion, may represent more selective strategies with reduced global effects on splicing. Finally, given that cancer‑associated splicing changes can generate tumor‑specific neoepitopes and remodel the tumor microenvironment, targeting splicing might deliver additional therapeutic benefits in combination with immunotherapies [47].
This study has several limitations. First, the SNRPD2–DDX39B–CTSC axis was validated in cell lines and xenograft/PDX models; studies in larger clinical cohorts and independent patient samples are needed to confirm its clinical relevance. Second, the precise molecular interactions by which SNRPD2 prevents DDX39B intron retention and by which DDX39B suppresses CTSC cryptic exons remain to be mapped. The application of high-throughput approaches such as RIP-seq and CLIP-seq would help to identify RNA–protein interaction sites and elucidate the regulatory mechanisms of SNRPD2 in splicing control. Notably, our analysis of IR events focused primarily on downregulated DEGs. The regulatory effects of SNRPD2 on IR events in upregulated DEGs, as well as its influence on other types of splicing (e.g., exon skipping), were beyond the scope of the present study but represent promising directions for future research. Furthermore, splicing factor regulation occurs within a complex molecular network, and multiple alternative mechanisms may also contribute to the malignant behavior promoted by SNRPD2 in endometrial cancer. Identifying such pathways will be important to ensure a balanced interpretation of our findings. Finally, the global transcriptomic and immunological consequences of targeting SNRPD2 or DDX39B require a comprehensive evaluation to anticipate potential toxicity and define therapeutic contexts in which the clinical benefit will outweigh the risks.
In summary, our work identifies SNRPD2 as an oncogenic spliceosomal factor in endometrial cancer that promotes tumor growth by maintaining DDX39B expression and thereby preventing cryptic exon activation in CTSC. Disruption of this splicing regulation reduces CTSC expression and impairs tumor phenotypes in endometrial cancer. In summary, our work reveals a previously unrecognized role of the SNRPD2–DDX39B–CTSC axis in endometrial cancer progression. Future studies should further dissect the broader molecular network involving SNRPD2 and assess the translational potential of targeting this pathway in clinically relevant settings.
Supplementary information
Supplementary data Original WB bands Uncropped bands for the splicing-PCR
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Li J, Li P, Brachtlova T, van der Meulen-Muileman IH, Dekker H, Kumar VS, et al. Evaluation of spliceosome protein Sm D 2 as a potential target for cancer therapy. Int J Mol Sci. 2024;25:13131.10.3390/ijms 252313131 PMC 1164271739684842 · doi ↗ · pubmed ↗
- 2Li Y, Chen Z, Gao Y, Liu Y, Gao Q, Pu Y, et al. SNRPB-mediated regulation of DDX 39A splicing promotes ovarian cancer progression by regulating alpha 6 integrin subunit expression. Oncogene. 2025;44:2170–85.10.1038/s 41388-025-03386-040216968 · doi ↗ · pubmed ↗