Identification of SNRNP200 as a crucial regulator that fuels glycolytic TNBCs in vitro and in vivo

Using weighted gene coexpression network analysis (WGCNA), we identified 1230 genes showing significant and coordinated alterations in the MPS2 subtype (Supplementary Fig. S2a). Among these genes, 8 overlapped with 42 upregulated core spliceosome proteins, including 2 U5 snRNP components (SNRNP200 and EFTUD2) and 1 U4/U6.U5 tri-snRNP component (USP39), indicating a regulatory role for U5 snRNPs in the glycolytic subtype (Fig. ​(Fig.2a).2a). To explore metabolic regulation by spliceosome components in MPS2, we investigated their response to glucose variations. A screen of 14 core spliceosome proteins revealed that U5 snRNP component levels were significantly sensitive to glucose fluctuations in MPS2 cells (MDA-MB-231 and BT-549), whereas mRNA levels remained stable (Fig. ​(Fig.2b2b and Supplementary Fig. S2b, c). Notably, USP39 levels varied at both the RNA and protein levels under different glucose conditions, suggesting distinct regulatory mechanisms (Supplementary Fig. S2d). Glucose deprivation experiments confirmed the impact of glucose availability on U5 snRNPs, with the reintroduction of glucose significantly increasing protein levels (Fig. ​(Fig.2c2c and Supplementary Fig. S2e). Moreover, under high-glucose conditions (25mM), the half-life of U5 snRNP proteins was extended (Supplementary Fig. S2f). In MPS2 breast cancer cells, there was a consistent and significant increase in the expression of these crucial proteins, with SNRNP200 exhibiting the most pronounced fluctuations in response to glucose concentrations (Fig. 2b, c and Supplementary Fig. S2b–f). By contrast, the U5 snRNP protein levels of the TNBC subtypes MPS1 (MDA-MB-468) and MPS3 (HCC1143) exhibited minor changes in response to glucose fluctuations compared with those of MPS2 (Supplementary Fig. S2g–i), indicating subtype-specific metabolic dependencies. The core U5 snRNP components were consistently upregulated in MPS2 TNBC cells and across TNBC patient cohorts (Supplementary Fig. S3a, b). Multiplex immunohistochemistry (mIHC) also confirmed the upregulation of U5 snRNP proteins in MPS2 TNBCs (Supplementary Fig. S4a). Gene set variation analysis (GSVA) further revealed positive correlations between carbohydrate and nucleic acid metabolism pathways (dominant features of MPS2) and the RNA abundance of U5 snRNP components, whereas lipid metabolism was negatively correlated (Fig. ​(Fig.2d2d and Supplementary Fig. S3c). These findings suggest that glucose-stabilized U5 snRNP proteins, coupled with increased RNA levels, contribute to the dysregulated metabolism observed in the MPS2 subtype of TNBC.

Open in a separate window

Fig. 2

SNRNP200 is a crucial regulator of glycolytic TNBCs, promoting tumor proliferation both in vitro and in vivo.

a Venn diagram depicting the overlap between the MPS2 hub genes and the 42 spliceosome genes whose protein levels were upregulated in TNBC. b MDA-MB-231 cells were cultured in media supplemented with glucose at the indicated concentrations for 16h. The cell lysates were subjected to immunoblotting. c MDA-MB-231 cells were glucose-starved for 12h and then stimulated with glucose (25mM) for the indicated times. The cell lysates were subjected to immunoblotting. d Correlation coefficient heatmap illustrating associations between normalized mRNA expression levels of upregulated U5 snRNP genes in tumors and enrichment scores of metabolic pathways via Spearman’s correlation analysis in the FUSCC-TNBC cohort (*P<0.05; **P<0.01). e Diagram depicting the U5 snRNP core components. f Western blot analysis of SNRNP200, EFTUD2, and PRPF8 expression levels in control and SNRNP200-knockdown MDA-MB-231 cells. g MDA-MB-231 cells were immunoprecipitated with an anti-SNRNP200 antibody, with IgG serving as a negative control. The IP data were analyzed by western blot analysis (on the left). qPCR analysis of the U4, U5, and U6 snRNA levels in the input, RNase A⁻, and RNase A⁺ groups. The relative U4, U5, and U6 snRNA levels were normalized to that of the input using the 2^−Ct method (on the right). qPCR analysis data are presented as the mean±SEM. h CCK-8 proliferation assay in control and SNRNP200-knockdown MDA-MB-231 cells. i Schematic outline showing the Snrnp200-targeted ASO treatment timeline: 4T1 mouse breast cancer cells were subcutaneously injected into BALB/c mice. When the tumors reached 50–100mm³, the mice were treated with ASO-Snrnp200 (5mg/kg subcutaneous injection, twice a week) or PBS (50µL, subcutaneous injection, twice a week) for 2 weeks (n=5 mice/group). Tumor growth in different groups. The data are presented as the mean±SEM. j A representative image of 4T1 tumors illustrating the effect of ASO-Snrnp200 treatment (n=5 mice/group). k Western blot analysis of mouse Snrnp200 protein expression in tumor tissues from 4T1 cell-derived xenografts. The data are representative of three independent biological replicates. For g–i, the data were compared using Student’s t-test if the data in each group were normally distributed: ***P<0.001; ns not significant, P>0.05.

The U5 snRNP is crucial for spliceosome activation and active site formation because its catalytic function is linked to the PRPF8, SNRNP200, and EFTUD2 proteins²³. In SNRNP200-knockdown cells, the PRPF8 and EFTUD2 protein levels decreased without affecting the corresponding mRNA levels, indicating that SNRNP200 stabilizes the U5 complex independently of transcription (Fig. ​(Fig.2f2f and Supplementary Fig. S5a–c). Immunoprecipitation (IP) assays confirmed that stable protein‒protein interactions between SNRNP200, PRPF8, and EFTUD2 were unaffected by U5 snRNA digestion (Fig. ​(Fig.2g2g and Supplementary Fig. S5d). Cycloheximide (CHX) treatment further confirmed that SNRNP200 knockdown led to shortened half-lives of PRPF8 and EFTUD2. Introduction of an SNRNP200 mutant (SNRNP200^Mut) resistant to targeting by SNRNP200 sgRNA-1 restored protein levels and extended half-lives under high glucose conditions, confirming the role of SNRNP200 in U5 complex stability (Supplementary Fig. S5e–g). Furthermore, compared with normal breast epithelial cells (MCF10A), TNBC cells, particularly those with glycolytic subtypes (Hs-578T, BT-549, and MDA-MB-231), presented significantly elevated SNRNP200 mRNA and protein levels (Supplementary Fig. S5h, i). In clinical samples, SNRNP200 was markedly upregulated in glycolytic TNBC subtypes and correlated with poorer overall survival (Supplementary Fig. S5j, k).

To investigate the biological function of SNRNP200, we conducted SNRNP200 knockdown experiments in MPS2 cells via CRISPR-Cas9 technology. Colony formation, EdU incorporation, and CCK-8 assays revealed significant inhibition of MPS2 cell proliferation in vitro following SNRNP200 knockdown (Fig. ​(Fig.2h2h and Supplementary Fig. S6a–e). Flow cytometry analysis revealed increased early- and late-stage apoptosis in SNRNP200-knockdown MPS2 cells (Supplementary Fig. S6f, g). Conversely, SNRNP200 knockdown in MCF10A cells resulted in minimal growth reduction, highlighting the selective vulnerability of glycolytic cancer cells to SNRNP200 depletion (Supplementary Fig. S6h–j). To explore SNRNP200 as a potential therapeutic target for glycolytic TNBC, we tested a panel of ASO agents targeting mouse Snrnp200. Among the three designed ASOs, ASO-2 achieved the greatest reduction in Snrnp200 expression, whereas control ASOs had no effect on Snrnp200 levels in a murine MPS2 cell line (4T1; Supplementary Fig. S6k, l). Treatment of 4T1 cells with ASO-Snrnp200 significantly inhibited cell growth and increased apoptosis in vitro (Supplementary Fig. S6m–p). To evaluate the in vivo efficacy of ASO-Snrnp200, we generated 4T1 xenograft tumors by injecting 5×10⁵ 4T mouse breast cancer cell lines into mice treated with ASO-Snrnp200, which exhibited a substantial reduction in tumor growth compared with that in PBS-treated controls (Fig. 2i, j). On-target pharmacodynamic activity was confirmed by reduced Snrnp200 protein expression in tumors (Fig. ​(Fig.2k2k).

Elevated glucose levels trigger SNRNP200 K1610 acetylation, safeguarding it from proteasomal degradation

Under high-glucose conditions, we observed an increase in the SNRNP200 protein level and the level of stable RNA (Fig. 2b, c and Supplementary Fig. S2b, c). Given the roles of the ubiquitin-proteasome system and autophagy in cellular degradation, we investigated the degradation pathway of SNRNP200. In vivo, ubiquitylation assays in MPS2 cells revealed a marked decrease in SNRNP200 ubiquitylation under high-glucose conditions (Fig. ​(Fig.3a).3a). Treatment with the proteasome inhibitor MG132 or the histone deacetylase inhibitor trichostatin A (TSA) increased SNRNP200 protein levels (Fig. ​(Fig.3b3b and Supplementary Fig. S7a). However, autophagy inhibitors such as bafilomycin A1 (Baf-A1) and ammonium chloride (NH₄Cl), as well as the autophagy inducer rapamycin (Rapa), did not affect SNRNP200 levels in MPS2 cells. These findings collectively indicate that the degradation of SNRNP200 primarily occurs through the ubiquitin-proteasome pathway rather than the autophagy-lysosome pathway. Moreover, TSA treatment significantly augmented SNRNP200 acetylation, concurrently impeding its ubiquitylation (Fig. 3c, d). IP and liquid chromatography-mass spectrometry (IP-LC-MS) data identified PCAF as a potential acetyltransferase responsible for SNRNP200 acetylation (Fig. ​(Fig.3e).3e). The interaction between SNRNP200 and PCAF was further validated through IP assays (Fig. ​(Fig.3f3f and Supplementary Fig. S7b). Moreover, PCAF overexpression enhanced SNRNP200 acetylation and inhibited its ubiquitylation, whereas endogenous PCAF depletion led to a notable decrease in SNRNP200 acetylation, a significant reduction in the half-life of SNRNP200, and a subsequent decrease in the protein level of SNRNP200 (Fig. 3f–k). Notably, in vitro experiments indicated that glucose did not directly influence the interaction between PCAF and SNRNP200 but likely affected the accessibility of other contributing factors (Fig. ​(Fig.3l).3l). Recent studies have shown that acetylation is directly linked to acetyl-CoA and is affected by acetyl-CoA availability^24,25. Consistently, we observed significant fluctuations in the intracellular acetyl-CoA levels corresponding to changes in the glucose concentration, which is consistent with the observed pattern of SNRNP200 acetylation (Fig. 3l, m).

Open in a separate window

Fig. 3

Elevated glucose levels trigger PCAF-mediated acetylation of SNRNP200.

a HEK293T cells were transfected with Myc-tagged SNRNP200 and HA-Ub as indicated and treated with glucose at either 2.5mM or 25mM. The cells were harvested for ubiquitylation analysis. b Immunoblot analysis of SNRNP200, EFTUD2, and PRPF8 protein levels in BT-549 cells treated with or without nicotinamide (NAM, 5mM, 6h), TSA (10μM, 16h), Baf-A1 (200nM, 16h), NH₄Cl (20mM, 16h), or Rapa (1μM, 16h). MG132 treatment (10μM, 12h) was used as a positive control. c, d After treatment with or without NAM, TSA, or MG132, the ubiquitylation levels (c) and acetylation levels (d) of SNRNP200 in HEK293T cells were measured by immunoblotting with the indicated antibodies. e Endogenous PCAF was immunoprecipitated from HEK293T cells with an anti-SNRNP200 antibody and analyzed by LC-MS. The data show MS2 spectra for a signature peptide of PCAF. f Myc-SNRNP200 was cotransfected with Flag-tagged PCAF into HEK293T cells. Acetylation was determined by immunoblotting. Co-IP assays were performed to determine the interaction between SNRNP200 and PCAF. g Flag-tagged PCAF was cotransfected with Myc-tagged SNRNP200 and HA-tagged ubiquitin. The ubiquitylation of SNRNP200 was determined by IP-western blotting with an anti-HA antibody. h Western blot analysis of PCAF expression levels in control and PCAF-depleted BT-549 cells. i, j BT-549 cells maintained in 25mM glucose were transfected with siPCAF or the control and treated with CHX as previously described. The endogenous SNRNP200, EFTUD2, and PRPF8 proteins were analyzed by immunoblotting (i) and quantified against actin (j). The data for the relative SNRNP200 levels are presented as the mean±SEM. k BT-549 cells were transfected with siPCAF or the control. The acetylation levels of SNRNP200 were analyzed by immunoblotting. l BT-549 cells were treated with glucose at the indicated concentrations. The acetylation levels of SNRNP200 were analyzed by immunoblotting. Co-IP assays were performed to determine the dynamic interactions between PCAF and SNRNP200. m Acetyl-CoA levels were measured in BT-549 cells treated with glucose at the indicated concentrations. The relative acetyl-CoA levels are presented as the mean±SEM. For j and m, the data were compared using Student’s t-test: ns not significant, P>0.05; ***P<0.001.

IP-LC-MS data identified lysine 1610 (K1610) as the primary acetylation site on SNRNP200 (Fig. ​(Fig.4a).4a). K1610 is located within the DEXH box domain (residues 1310–1897) and is highly conserved across species (Fig. 4b, c). TSA treatment increased acetylation and decreased the ubiquitylation of wild-type (WT) SNRNP200 but not of the K1610R (unacetylated) or K1610Q (acetylation mimicked) mutants (Fig. ​(Fig.4d4d and Supplementary Fig. S7c, d). High glucose stabilized WT SNRNP200 but not the K1610R or K1610Q mutants (Fig. ​(Fig.4e).4e). To determine the deacetylase responsible for SNRNP200, we investigated three prominent deacetylases and found that histone deacetylase 5 (HDAC5) specifically facilitated SNRNP200 deacetylation (Fig. ​(Fig.4f).4f). HDAC5 overexpression decreased SNRNP200 acetylation and increased ubiquitylation (Fig. 4f, g). Endogenous IP assays confirmed the interaction between SNRNP200 and HDAC5 in MPS2 tumor cells (Supplementary Fig. S7e). Moreover, HDAC5 exhibited a greater affinity for SNRNP200 exposed to lower glucose concentrations (2.5mM), with diminished binding occurring as the glucose concentration increased (Fig. ​(Fig.4g).4g). RNF123 was identified as a potential SNRNP200-interacting protein via IP-MS (Supplementary Fig. S7f). This interaction was validated in HEK293T and MPS2 cells (Supplementary Fig. S7g, h). Lowering glucose increased SNRNP200 ubiquitylation and its association with RNF123 (Supplementary Fig. S7i). RNF123 depletion decreased SNRNP200 ubiquitylation under low-glucose conditions, extending the half-life of SNRNP200 (Fig. 4h–j and Supplementary Fig. S7j). Notably, low glucose effectively promoted K48-type polyubiquitylation, whereas monoubiquitylation or the non-degradative K63-type polyubiquitylation of SNRNP200 was not observed (Supplementary Fig. S7k). These findings indicated that glucose protected SNRNP200 from proteasomal degradation.

Open in a separate window

Fig. 4

The acetylation of SNRNP200 at lysine 1610 safeguards it from proteasomal degradation.

a Endogenous SNRNP200 was immunoprecipitated from HEK293T cells and analyzed by LC-MS. The data show MS2 spectra for a signature peptide acetylated at K1610. b Alignment of the SNRNP200 protein sequence across different species. c A secondary structure model of SNRNP200 was visualized via SWISS-MODEL⁶⁸. The K1610 residue is highlighted with a different color. d HEK293T cells were transfected with Myc-tagged SNRNP200 WT, K1610R, or K1610Q plasmids for 36h with or without TSA. The acetylation levels of SNRNP200 were quantified against immunoprecipitated Myc-tag and are presented as the mean±SEM. e HEK293T cells were transfected with the indicated plasmids and treated with glucose at either 2.5mM or 25mM. Ubiquitylation analysis was revealed by immunoblotting. f HDAC1, HDAC2, and HDAC5 were overexpressed in HEK293T cells treated with or without TSA. The acetylation levels of SNRNP200 were determined by immunoblotting. Co-IP assays were performed to determine the interaction between SNRNP200 and HDAC5. g HEK293T cells were maintained in a medium supplemented with either 2.5mM or 25mM glucose. The SNRNP200 ubiquitylation levels were determined via IP-western blotting. Co-IP assays were performed to determine the dynamic interactions between SNRNP200 and HDAC5. h BT-549 cells were transfected with siRNF123 or the control. The levels of ubiquitinated SNRNP200 were analyzed by immunoblotting. i, j BT-549 cells maintained in 2.5mM glucose were transfected with siRNF123 or the control and treated with CHX as previously described. The endogenous SNRNP200, EFTUD2, and PRPF8 proteins were analyzed by immunoblotting (i) and quantified against actin (j). The data for the relative SNRNP200 levels are presented as the mean±SEM. k Working model illustrating the mutually exclusive acetylation and ubiquitylation of SNRNP200 in glycolytic TNBCs. For d and j, the data were compared using Student’s t-test: ns not significant, P>0.05; *P<0.05; **P<0.01.

In summary, our data revealed that increased glucose uptake in glycolytic subtypes elevates acetyl-CoA levels, maintaining SNRNP200 K1610 acetylation through PCAF. This prevents HDAC5-mediated deacetylation and RNF123-mediated K48-type polyubiquitylation, protecting SNRNP200 from proteasomal degradation (Fig. ​(Fig.4k4k).

SNRNP200 enhances RNA splicing in metabolic enzyme-encoding genes with weak 5’ splice sites

We employed RNA sequencing (RNA-seq) analysis to investigate SNRNP200-mediated splicing regulation within the glycolytic subtype of TNBC. SNRNP200 knockdown led to significant alterations in 2742 AS events across 2134 genes (FDR<0.05 and ΔPSI>0.02), including skipped exons (SEs), alternative 5’ splice sites (A5SSs), alternative 3’ splice sites (A3SSs), retained introns (RIs), and mutually exclusive exons (MXEs; Supplementary Fig. S8a). Notably, 292 genes presented two or more AS events, highlighting the impact of SNRNP200 on transcriptome diversity. A greater proportion of RI events were upregulated than other AS types, affecting introns enriched in genes linked to cell cycle regulation, DNA/RNA processing, and cellular metabolism (Fig. ​(Fig.5a5a and Supplementary Fig. S8b and Table S3). Specifically, metabolic enzymes involved in glycolysis and nucleotide metabolism displayed altered splicing patterns upon SNRNP200 depletion. For example, GAPDH, a key glycolysis enzyme responsible for converting glyceraldehyde 3-phosphate to 1,3-diphosphoglycerate, exhibited altered RNA splicing following SNRNP200 knockdown, resulting in the retention of introns 4 and 5. Similarly, GSS transcripts encoding glutathione (GSH) synthase were enriched in intron 4 after SNRNP200 depletion (Fig. ​(Fig.5b).5b). Semiquantitative RT-PCR validated the role of SNRNP200 in modulating the splicing of representative genes (Fig. ​(Fig.5c5c and Supplementary Fig. S8c).

Open in a separate window

Fig. 5

SNRNP200 enhances RNA splicing in genes encoding metabolic enzymes with weak 5’ splice sites.

a Enrichment networks of intron-retained transcripts in control and SNRNP200-knockdown MDA-MB-231 cells were analyzed and visualized by Metascape. Representative terms with kappa similarities above 0.3 formed a network and were depicted using Cytoscape software. b Sashimi plots depicting the RIs of GAPDH (top) and GSS (bottom). The number of RI reads is indicated (control sgRNA in blue, SNRNP200 sgRNA in red). c Representative RT-PCR validation of SNRNP200-regulated RI events in MDA-MB-231 cells. The structure of each isoform is illustrated in the diagrams. Individual data points are presented (n=3; *P<0.05; *P<0.01; **P<0.001; Student’s t-test). d A 5’ splice site strength analysis of RI events was carried out. Non-retained introns (NRIs) within the same set of genes were used for comparison (Wilcoxon rank-sum test). e Motif enrichment analysis of the same set of genes revealed that the most frequently identified motifs aligned with the consensus 5’ splice site sequences for both the RI and NRI. f The GC compositions of the RI and NRI were calculated by dividing the GC content of each intron by the average of its adjacent exons and compared by the Wilcoxon rank-sum test.

We further investigated features common to transcripts with increased RI events, focusing on differences in splice site strength between retained and properly spliced introns post-SNRNP200 depletion. Our analysis revealed that introns with inefficient splicing tended to have weaker 5’ splice sites (P value=0.0002, Wilcoxon test; Fig. ​Fig.5d),5d), whereas distinctions at the 3’ splice site were less pronounced (P value=0.2932, Wilcoxon test; Supplementary Fig. S8d). Motif enrichment analysis supported these findings (Fig. ​(Fig.5e5e and Supplementary Fig. S8e). Additionally, compared with non-RIs, RIs presented a slightly greater GC content, resembling the GC levels observed in adjacent exons (Fig. ​(Fig.5f).5f). Taken together, these findings underscore the critical role of U5 snRNP components, particularly SNRNP200, in regulating the splicing patterns of metabolic enzyme-encoding genes characterized by weak 5’ splice sites.

SNRNP200 enhances glycolysis and glutathione metabolism via RNA splicing

To elucidate the role of SNRNP200 in metabolic regulation through RNA splicing, we conducted untargeted metabolomic analysis on WT and SNRNP200-knockdown cells. Depletion of SNRNP200 led to significant reductions in metabolites essential for glycolysis and glutathione metabolism, including lactic acid, oxidized glutathione, UDPG, NAD (nicotinamide adenine dinucleotide), and SAM (Fig. ​(Fig.6a6a and Supplementary Table S4). These metabolites were particularly enriched in glycolytic subtypes (Fig. ​(Fig.1f),1f), highlighting the pivotal role of SNRNP200 in these pathways. Additionally, the levels of specific lipids, such as diacylglycerol (DAG), phosphatidylethanolamine (PE), phosphatidylserine (PS), and phosphatidylcholine (PC), increased following SNRNP200 knockdown (Fig. ​(Fig.6b6b and Supplementary Table S5). Pathway-based differential abundance (DA) analysis confirmed alterations in metabolic pathways due to SNRNP200 depletion (Supplementary Fig. S9a).

Open in a separate window

Fig. 6

SNRNP200 enhances glycolysis and glutathione metabolism via RNA splicing.

a Volcano plot of the polar metabolites profiled in control and SNRNP200-knockdown MDA-MB-231 cells. Significantly differentially abundant metabolites are color-coded by individual category. b Volcano plot of the lipid profiles of control and SNRNP200-knockdown MDA-MB-231 cells. Metabolites with significant differences are color-coded according to their respective categories. Furthermore, the proportions of five distinct lipid categories in SNRNP200-knockdown MDA-MB-231 cells were assessed. The data were statistically analyzed using the chi-square test (***P<0.001). c Images of the RIs of ALDOA, GAPDH, and GSS. d Western blot analysis of the expression levels of SNRNP200, ALDOA, GAPDH, and GSS in control and SNRNP200-knockdown MDA-MB-231 and BT-549 cells. e Comparison of the ECAR between control and SNRNP200-knockdown MDA-MB-231 cells. f Diagram summarizing metabolic genes involved in glycolysis, the TCA cycle, and glutamate metabolism. The diagram visually depicts normalized metabolite levels (boxes) and their respective upstream metabolic enzymes (circles) undergoing intron retention induced by SNRNP200 ablation. UDP uridine 5’-diphosphate, G-1-P glucose 1-phosphate, G-6-P glucose 6-phosphate, Gal-1-P galactose 1-phosphate, Ri-5-P ribulose 5-phosphate, R-5-P ribose 5-phosphate, PEP phosphoenolpyruvate, α-KG alpha-ketoglutarate, G-3-P glycerol 3-phosphate, 1,3-BPG glyceric acid 1,3-biphosphate, NADP nicotinamide adenine dinucleotide phosphate, FA fatty acids, GL glycerolipids, GP glycerophospholipids, SP sphingolipids, ST sterol lipids.

We further explored whether the retention of introns in genes encoding key enzymes had a discernible impact on cellular metabolites. SNRNP200 knockdown led to intron retention in ALDOA (intron 9), GAPDH (introns 4 and 5), and GSS (intron 4), causing frameshift mutations and premature stop codons (Fig. ​(Fig.6c6c and Supplementary Fig. S9b). This resulted in reduced protein levels of ALDOA, GAPDH, and GSS (Fig. ​(Fig.6d).6d). Clinically, U5 snRNP SNRNP200 expression was positively correlated with ALDOA, GAPDH, and GSS in TNBC patient cohorts (Supplementary Fig. S9c). Metabolites in the glycolytic and glutathione pathways, such as lactate, glutamic acid, and glutathione, were significantly downregulated following SNRNP200 depletion (Fig. ​(Fig.6f).6f). The extracellular acidification rate (ECAR) also decreased in SNRNP200-deficient cells (Fig. ​(Fig.6e).6e). Additionally, introns within the genes encoding metabolic enzymes of lipid pathways, specifically diacylglycerol kinase (DGKZ) and phosphatidylserine decarboxylase (PISD), were retained after SNRNP200 knockdown. This retention might inhibit the transformation of metabolites, leading to the accumulation of PS and DAG (Supplementary Fig. S9d). These results indicate that SNRNP200 regulates the RNA splicing of specific metabolic enzyme-encoding genes, thereby influencing cellular metabolism.

To further examine whether the tumor-promoting function of SNRNP200 relies on its regulation of enzymatic activities and cellular metabolism, we conducted a rescue assay. Restoring ALDOA, GAPDH, or GSS expression in SNRNP200-knockdown cells partially restored lactic acid and glutathione levels (Supplementary Fig. S10a–c). Cell function assays revealed that restoring these enzymes partially rescued the proliferation inhibition caused by SNRNP200 knockdown (Supplementary Fig. S10d–f). Flow cytometry revealed a partial reduction in early- and late-stage apoptosis in SNRNP200-knockdown MPS2 cells upon enzyme restoration (Supplementary Fig. S10g). These results confirm that the tumor-promoting function of SNRNP200 is linked to its regulation of enzymatic activities and cellular metabolism.

ASO-Snrnp200 enhances immunotherapy efficacy in glycolytic MPS2 tumors

Extensive research has explored the intricate link between immunological responses and metabolic status. Recent studies have elucidated the pivotal role of regulatory T (Treg) cells, which function as metabolic conductors, in shaping immune responses via nutritional cues²⁶. Glutathione has emerged as a crucial factor in preserving Treg functionality, whereas lactic acid in highly glycolytic tumor microenvironments promotes PD-1 expression in Treg cells, contributing to resistance to PD-1 blockade therapy and inhibiting tumor-infiltrating CD8⁺ T cells^27,28. Here, we investigated whether SNRNP200 affects immune functions by modulating lactic acid and glutathione levels and whether ASO-Snrnp200 treatment enhances immunotherapy sensitivity in glycolytic TNBC. In MPS2 tumor cells (4T1), ASO-Snrnp200 reduced both the intracellular and extracellular lactic acid and glutathione levels in vitro (Supplementary Fig. 11a, b). Consistently, ASO-Snrnp200 treatment significantly decreased Snrnp200 expression and lactic acid and glutathione levels in 4T1 xenograft tumors in vivo (Fig. 7a–c and Supplementary Fig. S11c). Compared with the LDH inhibitor FX-11 plus anti-PD-1, co-administration of ASO-Snrnp200 and anti-PD-1 markedly inhibited tumor growth (Fig. ​(Fig.7d7d and Supplementary Fig. S11d, e). To further elucidate the immune response following treatment, we employed flow cytometry analysis (Supplementary Fig. S11f, g). The combined therapy yielded a pronounced immune response characterized by increased proportions of tumor-infiltrating CD8⁺ T cells (Fig. ​(Fig.7e),7e), accompanied by elevated PD-1 expression within these cells (Fig. ​(Fig.7g).7g). Moreover, CD8⁺ T cells in the combination therapy group presented increased IFN-γ and granzyme B production (Fig. ​(Fig.7h7h and Supplementary Fig. S12a, b). The combination therapy downregulated FOXP3 and PD-1 expression in Tregs (Fig. 7f, g) and reduced the expression of activation markers (CTLA-4, ICOS, and GITR; Fig. ​Fig.7i7i and Supplementary Fig. S12c–e). mIHC staining of tumors for CD8⁺ T cells and Tregs consistently verified these observations (Fig. ​(Fig.7j).7j). Our findings underscore the potential of ASO-Snrnp200 to increase immunotherapy efficacy, highlighting its role in modulating tumor microenvironment dynamics, which are beneficial for therapeutic outcomes.

Open in a separate window

Fig. 7

ASO-Snrnp200 synergistic immunotherapy amplifies the antitumor response in glycolytic tumors.

a Schematic outline showing Snrnp200-targeted ASO treatment or LDH inhibition combined with anti-PD-1 antibody treatment of tumors: 5×10⁵ 4T1 mouse breast cancer cells were subcutaneously injected into BALB/c mice. When the tumors reached 50–100mm³, the mice were treated with ASO-Snrnp200 (5mg/kg subcutaneous injection, twice a week, n=6 mice/group), FX-11 (2mg/kg daily i.p. injection, daily), or PBS (50µL, i.p. injection, daily) for 2 weeks combined with an isotype control or anti-PD-1 antibody (10mg/kg, i.p. injection, twice a week). b, c Relative lactic acid (b) and GSH (c) levels in the six treatment groups. The data are presented as the mean±SEM. d Tumor growth in different groups. The data are presented as the mean±SEM. e, f Primary tumors from 4T1 model mice were harvested for flow cytometry to determine the percentages of CD8⁺ T cells among CD3⁺ T cells (e) and of Treg cells among CD4⁺ T cells (f). g The expression of PD-1 by CD8⁺ T cells (top) and Treg cells (bottom) in the tumor microenvironment was examined. Representative histograms and summary data are shown. The data are presented as the mean±SEM. h Representative histogram plots showing GZMB⁺ and IFN-γ⁺ cells among CD8⁺ T cells. i Representative histogram plots showing CTLA4⁺, ICOS⁺, and GITR⁺ cells among Treg cells. j Representative multiplexed immunohistochemistry images of PD1⁺CD8⁺ T cells (top) and PD1⁺CD4⁺FOXP3⁺ T cells (bottom) in the TME. Scale bars, 100μm or 40μm. For b–g, the data were compared using Student’s t-test if the data in each group were normally distributed (n=6 mice/group; *P<0.05; **P<0.01; ***P<0.001; ns not significant, P>0.05).

Human and mouse TNBC cell lines

The human breast cancer cell lines MCF-7, BT474, AU565, HCC1954, SK-BR-3, BT-20, BT-549, HCC1143, Hs578T, MDA-MB-231, and MDA-MB-468; the human breast epithelial cell line MCF10A and the human embryonic kidney cell line HEK293T were obtained from the American Type Culture Collection (ATCC). All TNBC breast cancer cell lines were classified into TNBC subtypes. The 4T1 mouse TNBC cell line was obtained from Y. Kang’s laboratory (Princeton University, USA). All the above cells were identified via cell line DNA profiling (short tandem repeat, STR) and passed routine cell line quality control tests (e.g., morphology, cell viability, and mycoplasma). MCF10A cells were maintained in DMEM/F12 (BasalMedia, L310KJ) supplemented with 5% horse serum (Gibco, SR0035), 1% penicillin-streptomycin (BasalMedia, S110JV), 10μg/mL insulin (BasalMedia, S450J7), 100ng/mL cholera toxin (Sigma-Aldrich, 227036), 0.5μg/mL hydrocortisone (Sigma-Aldrich, 3867) and 20ng/mL EGF (PeproTech, 100-47). The other cell lines were cultured according to standard conditions. Other cell lines were maintained in high-glucose DMEM supplemented with 10% fetal bovine serum (Gibco, 16000044) and 1% penicillin-streptomycin. All the cell lines were passaged fewer than 30 times and cultured in a humidified incubator containing 5% CO2 at 37°C. TSA (10μM, MedChemExpress, HY-15144) and NAM (5mM, MedChemExpress, HY-B0150) were added to the culture media 16h and 6h before harvest, respectively. Glucose-free medium was prepared with DMEM (Gibco, 11966) supplemented with glucose (Sigma-Aldrich, G7528) as indicated.

Xenograft models

All animal experiments were performed according to protocols approved by the Research Ethical Committee of FUSCC. Six- to seven-week-old female BALB/c mice were used in this study. The mice were housed under a 12-h light/dark cycle at ambient temperature (20–22°C) and humidity (60±10%) with free access to standard rodent diet and water in individually ventilated cages in a specific pathogen-free facility to monitor health status via culture, serum and microscopic examination.

Dimensionality reduction analysis

PCA determines the variable combinations that contribute the most to data variance and is accomplished via the singular value decomposition (SVD) method given by the prcomp function from the R package stats (version 3.4.1)⁶⁴. The overall contribution of each variable (splicing event or gene) to the data variance along specified main components was calculated via fviz_contrib from factoextra (version 1.0.5). t-SNE was used for dimensionality reduction and visualization of high-dimensional data via the Rtsne (version 0.15) package⁶⁵.

WGCNA

First, we calculated the Pearson correlation coefficient between all gene pairs across samples. Next, the correlation coefficients were transformed into an adjacency matrix with a soft threshold of β=4. The matrix was then transformed into a topological matrix with the topological overlap measure (TOM), which describes the degree of association between genes. A 1-TOM was used as the distance to cluster the genes, after which a dynamic pruning tree was constructed, and the 84 MPS2-related modules were identified by setting the merging threshold function at 0.25. We computed the eigengene for each module, defined as the first principal component of the module representing the overall expression level of the module, as well as module membership (MM). We quantified associations of individual genes with our clinical variables by defining gene significance as (the absolute value of) the correlation between the gene and the trait. For each module, we also defined a quantitative measure of MM as the correlation of the module eigengene and the gene expression profile. This allowed us to quantify the similarity of all genes on the array to every module. Based on the genes of significantly related modules (yellow modules: cor=0.87, P<0.001), two genes with weights>0.02 were used to construct the network.

Correlations between core spliceosome genes and metabolic pathways

GSVA, performed via the R package GSVA (version 1.36.2), was utilized to calculate the enrichment score for each pathway in every sample or cell via log₂-transformed FPKM⁶⁶. The correlation between log₂ expression levels of 42 genes coding for core spliceosome components and metabolic pathways based on the standard error of the enrichment score was tested via Spearman’s rho by cor.test from stats (Version 4_4.0.3).

Western blot analysis

Proteins were extracted via Tissue Protein Extraction Reagent (T-PER, ThermoFisher Scientific, 78510) in combination with proteinase and phosphatase inhibitors (Sigma-Aldrich, 5892970001). The protein concentration was then determined via BCA reagent (Solarbio, PC0020), and the sample was boiled for 10min in an SDS-PAGE loading buffer. Equal concentrations of proteins were electrophoresed on SDS-PAGE gels and transferred to PVDF membranes. The membranes were blocked with 5% skim milk for 60min before being probed with primary antibodies for 12–16h at 4°C. After thorough washing with TBST, the membranes were incubated for 1h with a goat anti-rabbit antibody (Cell Signaling Technology, 7074) or a goat anti-mouse antibody (Cell Signaling Technology, 7076) at room temperature. The targeted protein signals were identified via an enhanced chemiluminescence substrate (Pierce Biotechnology) and visualized utilizing Image Lab Software (Bio-Rad). Triplicates of each experiment were run. A list of the primary antibodies used for immunoblotting is provided in Supplementary Table S8.

SNRNP200 knockdown

For CRISPR/Cas9-mediated knockdown of human SNRNP200, first, cells transfected with lentivirus containing lentiCas9-Blast (Addgene, 52962) were selected with blasticidin (InvivoGen, ant-bl-10p) for at least 2 weeks. Next, the sgRNA lentivirus was introduced into Cas9 cells. Forty-eight hours after lentiviral sgRNA transfection, stably integrated cells were selected with 1–2µg/mL puromycin (InvivoGen, ant-pr-1) for 7 days. These cells were used for subsequent assays (RNA-seq, qPCR, Western blot, etc.). The knockdown efficiency was confirmed by western blot analyses. The SNRNP200-sgRNA sequences used were as follows: SNRNP200-sg1: ACCCGCCGGGATGAACCCAC; and SNRNP200-sg2: TAGGCGAAGAAAGCGTGATG.

RNA treatment in endogenous IP assays

For RNA treatment in the endogenous IP assay, 3×10⁷ breast cancer cells were collected and lysed with predetermined IP buffer (50mM Tris-HCl pH 8.0, 100mM NaCl, 1mM EDTA, and 1% NP-40). After a 1-h rotation at 4°C, the lysates were subjected to centrifugation at 12,000rpm for 5min at 4°C. A 10μL aliquot of each sample was set aside as input for RNA extraction, whereas another 10μL was designated as input for western blot analysis. The lysates were then categorized into three distinct groups: the IgG, RNase A⁻ (no RNase A treatment), and RNase A⁺ (with RNase A treatment, 100μg/mL, ThermoFisher Scientific, R1253) groups. Afterward, 30μL of lysate was removed from both groups for RNA extraction. The remaining samples were incubated with the following antibodies for the IP assay: anti-IgG (Cell Signaling Technology, 2729) and anti-SNRNP200 (Abcam, ab176715). After rotating overnight at 4°C, the samples were centrifuged to precipitate the beads. After washing with IP buffer five times, the beads were boiled for 10min with 40µL of 1× loading buffer for each sample. Finally, the samples were analyzed by western blot analysis as previously described. After cDNA synthesis, the U4, U5, and U6 snRNA levels in the input, RNase A⁻ and RNase A⁺ groups were detected using RT-qPCR. The relative U4, U5, and U6 snRNA levels of the RNase A⁻ and RNase A⁺ groups were normalized to those of the input group via the 2^−CT method.

In vitro cell viability assays

For the cell proliferation assay, cells were seeded in quintuplicate in a 96-well plate (1×10³cells/well) and cultivated in an incubator containing 5% CO₂ at 37°C. Cell viability was measured via a cell counting kit‐8 (CCK8; MedChemExpress, HY-K0301), and the highest and lowest values were excluded. For the EdU incorporation assay, an EdU kit (RiboBio, C10310) was used to detect the proliferation of breast cancer cells. In brief, treated cells (100μL of a 2×10⁴cells/mL suspension) were seeded into 96-well plates, and 50mM EdU was added at 37°C for a 2-h incubation. After fixation with 4% paraformaldehyde and permeabilization with 1% Triton X-100, 100μL of click reaction cocktail was added to the treated cells, and the cells were then stained with Hoechst 33342. Images of the immunostained sections were captured and quantitatively evaluated with a fluorescence microscope (Olympus). For the colony formation assay, cells were seeded in triplicate in a 6-well plate (1×10³cells/well) and cultivated under normal conditions. After two weeks, the colonies were fixed in methanol and stained with 0.05% crystal violet for 30min. Colonies with more than 50 cells were counted.

Flow cytometry measurement of cell apoptosis

The cells of interest were fixed and stained with Annexin V binding buffer (BD Biosciences, 556454) according to the manufacturer’s instructions. The apoptotic cells were then identified and analyzed with FlowJo software.

IP-LC-MS

HEK293T cells transfected with the pCMV6-SNRNP200-Myc plasmid were lysed and immunoprecipitated with Protein A/G magnetic beads (Bimake, B23202) in IP buffer. The eluted FLAG peptide was resolved by SDS-PAGE and Coomassie blue staining. Lysates from HEK293T cells transfected with control pCMV6-Entry (Origene, PS100001) served as controls. Protein bands specific to Myc-SNRNP200-transfected cells were digested with trypsin for sequencing by MS analysis (Novogene Bioinformatics Institute, Beijing, China). Protein identification was performed via Proteome Discoverer 2.2 (PD2.2, ThermoFisher Scientific) in the human RefSeq protein database (National Center for Biotechnology Information).

In vivo ubiquitylation assays

For the analysis of SNRNP200 ubiquitylation, HEK293T cells were transfected with plasmids, including pCMV6-SNRNP200-Myc, pCMV6-SNRNP200-K1610Q-Myc, pCMV6-SNRNP200-K1610R-Myc, pRK5-HA-Ubiquitin-WT (Addgene, 17608), pRK5-HA-Ubiquitin-K63 (Addgene, 17606), pRK5-HA-Ubiquitin-K48 (Addgene, 17605), pCMV-RNF123-FLAG, pCMV-HDAC5-FLAG, and pcdna3.4-KAT2B-FLAG, as indicated. Prior to being harvested for IP experiments, HEK293T cells were pretreated with 10μM MG132 (MedChemExpress, HY-13259) for 10–12h to inhibit the proteasome pathway. After 48h of transfection, the cells were lysed and subsequently incubated with anti-FLAG (Cell Signaling Technology, 14793) or anti-Myc-Tag (Cell Signaling Technology, 2278) antibodies overnight at 4°C, followed by further incubation with protein A/G magnetic beads (Bimake, B23202) for an additional 8h at 4°C. After five washes, ubiquitinated SNRNP200 was detected by immunoblotting with an anti-HA antibody (Abcam, ab9110). To detect the endogenous ubiquitylation of SNRNP200, BT-549 cells were transfected with pCMV6-SNRNP200-Myc plasmids along with pRK5-HA-Ubiquitin-WT (Addgene, 17608), and subjected to siRNF123 or control treatment. Additionally, the cells were exposed to the proteasome inhibitor MG132 (10μM, MedChemExpress, HY-13259) for 10h. The cellular proteins were then subjected to IP using an anti-SNRNP200 antibody (Abcam, ab176715), followed by immunoblotting with an anti-HA antibody (Abcam, ab9110) for analysis.

In vivo acetylation assays

For SNRNP200 acetylation analysis, HEK293T cells were transfected with the indicated plasmids. and pretreated with or without TSA (10μM, MedChemExpress, HY-15144) for 16h or NAM (5mM, MedChemExpress, HY-B0150) for 6h before being harvested for IP experiments. The level of acetylated SNRNP200 was determined by immunoblotting with an anti-acetyl-lysine antibody (Abcam, ab190479).

siRNAs and transfection

Specific siRNAs targeting PCAF and RNF123, along with corresponding negative control siRNAs (siNCs), were obtained from Tsingke Biotechnology (Shanghai, China). The siRNA duplexes were introduced into cells via Lipofectamine 2000 transfection reagents (Invitrogen, 11668019), according to the manufacturer’s guidelines. The knockdown efficiency was assessed through immunoblotting analysis at 48h posttransfection. The siRNA sequences are listed in the Supplementary Table S6.

Acetyl-CoA measurement

After 16h of exposure to glucose at various concentrations (0mM, 2.5mM, 7.5mM, 12.5mM, and 25mM), 1×10⁷ BT-549 cells were rapidly frozen in liquid nitrogen (liquid N₂) and then pulverized. Acetyl-CoA content assessment was conducted following the manufacturer’s guidelines (Sigma Aldrich, MAK039) using a fluorescence assay. The fluorescence intensity was measured (λ_ex=535/λ_em=587nm) in black, 96-well flat-bottom plates with clear bottoms. Standard curves were generated using 0.02mM acetyl-CoA standard solutions with volumes of 0μL, 10μL, 20μL, 30μL, 40μL, and 50μL.

Splicing assays

Total RNA was extracted from the cells using TRIzol (Life Technologies, 15596018), and reverse transcription was carried out to generate complementary DNA (cDNA) using the HiScript III 1st Strand cDNA Synthesis Kit (Vazyme, R312). Primers were designed using SnapGene software (version 7.0) to amplify transcripts near alternatively spliced exons. For subsequent DNA gel electrophoresis separation, real-time PCR was performed using 2× Phanta Max Master Mix (Vazyme, #P515), while qPCR was carried out with SYBR qPCR Master Mix (Vazyme, #Q311). In this work, the relative expression level of each RNA was quantified using the 2^−ΔΔCT method, with ACTB serving as the internal standard. For the RI analyses, the cDNA samples were subjected to qPCR analyses with primers designed for the retained introns (In) and the adjacent exons (All). The percentage of transcripts with RIs was determined using the formula PSI=2^{CT(All)−CT(In)}. Each experiment was conducted in triplicate. The primers used for the RT-PCR and qPCR analyses are summarized in Supplementary Table S7.

Analysis of splice site and GC content

Using the MaxEntScan algorithm, the strength of splice sites was computed. A standard collection of background sequences was used to perform motif enrichment analysis with the R program ggseqlogo (version 0.1). To account for splicing signals, GC contents in introns versus neighboring exons were compared by deleting 20 nucleotides from either end of the introns and 3 nucleotides from the exons.

RNA-seq

RNA extraction was completed as described previously. RNA-seq library construction was performed using the VAHTS mRNA-seq V2 Library Prep Kit for Illumina (Vazyme, NR612) according to the manufacturer’s instructions, followed by double-ended 150nt sequencing using the Illumina NovaSeq 6000 platform. The FASTQ data were mapped to the human genome (hg38) using TopHat2, and index generation was performed using SAMtools and Picard.

Metabolite extraction

MDA-MB-231 cells were frozen in liquid nitrogen until extraction. One milliliter of extraction solution (acetonitrile:methanol:water=2:2:1) containing 2mg/mL L-2-chlorophenylalanine as an internal standard was added to the samples. After 30s of vortexing, the samples were homogenized at 40Hz for 4min and sonicated for 10min in an ice-water bath. The homogenization and sonication cycles were repeated three times. The samples were subsequently incubated at −40°C for 1h and centrifuged at 10,000rpm for 15min at 4°C. A total of 800mL of the supernatant was transferred to a fresh tube and dried in a vacuum concentrator at 37°C. The dried samples were reconstituted in 200mL of 50% acetonitrile by sonication on ice for 10min. The constitution was then centrifuged at 13,000rpm for 15min at 4°C, and 75mL of the supernatant was transferred to a fresh glass vial for LC/MS analysis.

LC-MS/MS analysis

Ultrahigh-performance liquid chromatography (UHPLC) separation was performed via an Agilent 1290 Infinity UHPLC System with a UPLC BEH amide column. The mobile phase consisted of a mixture of 25mmol/L ammonium acetate, 25mmol/L ammonia hydroxide in water (pH=9.75) (A), and acetonitrile (B). The column temperature was 25°C, the autosampler temperature was 4°C, and the injection volume was 2mL (pos) or 2mL (neg).

A 6550 QTOF mass spectrometer (Agilent Technologies) was used for full-scan MS1 data acquisition in the 60–1200Da scan range. The ESI source conditions included a gas temperature of 250°C, a gas flow of 16L/min, a sheath gas temperature of 350°C, a sheath gas flow of 12L/min, a nebulizer at 20psi, a fragmentor at 175V, and a capillary voltage of 3000V.

For MS/MS spectra, a Triple TOF 6600 mass spectrometer (AB Sciex) was employed in information-dependent acquisition (IDA) mode with Analyst TF version 1.7 (AB Sciex). The 12 most intense precursor ions with intensities above 100 were selected for MS/MS analysis at a collision energy (CE) of 30eV. The cycle time was 0.56s. The ESI source conditions included gas 1 at 60psi, gas 2 at 60psi, the curtain gas at 35psi, the source temperature at 600°C, the declustering potential at 60V, and the ion spray voltage floating (ISVF) at 5000V (positive) or −4000V (negative).

DA score

The DA score reflects a pathway’s tendency to exhibit higher metabolite levels than those in the control group. It is determined by initially subjecting all pathway metabolites to a nonparametric DA test (Mann–Whitney U-tests with Benjamini–Hochberg correction). The DA score was calculated based on the significantly increased or decreased metabolite levels, and the DA score was defined as:

Thus, the DA score varies from −1 to 1. A score of −1 indicates that all metabolites in a pathway decreased in abundance, whereas a score of 1 indicates that all metabolites increased.

ECAR assays

The ECAR of the MDA-MB-231 cells was quantified using a Seahorse XFe96 Extracellular Flux Analyzer (Agilent Technologies). MDA-MB-231 cells were seeded at 10,000 cells per well in a 96-well XF cell culture microplate (Agilent Technologies) with the corresponding growth medium and incubated overnight in a 5% CO2 incubator. Approximately 1h before the Seahorse Analyzer readings, the cells were washed and incubated in serum-free, bicarbonate-free XF base medium in a non-CO2 incubator. An XF Glycolysis Stress Test Kit (Agilent Technologies, 103020-100) was used to detect the ECAR according to the manufacturer’s instructions. The ECAR was determined after 12 cycles (3min of mixing and 3min of mixing). The baseline measurements were the average of the last three readings before oligomycin addition (1mM final concentration).

ASO delivery

All ASOs were designed and synthesized by RiboBio, China. The oligonucleotide concentration was determined with a Nanodrop spectrophotometer. For in vitro ASO delivery, 1×10⁵ 4T1 mouse breast cancer cells were seeded in six-well plates and transfected with ASOs at various concentrations (0nM, 10nM, 20nM, 30nM, 50nM, and 100nM) with 5μL of Lipofectamine 3000. For viability assays, the protein was isolated after 48h for protein knockdown assessment. For in vivo ASO delivery, ASOs were delivered by subcutaneous injection at a dose of 5mg/kg. PBS was injected as the control.

In vivo mouse studies

Six- to seven-week-old female BALB/c mice were obtained from Shanghai Jihui Laboratory Animal Care Co., Ltd. and maintained under pathogen-free conditions. All animal experiments were performed according to protocols approved by the Research Ethical Committee of FUSCC. A total of 5×10⁵ 4T1 mouse breast cancer cells were injected subcutaneously into the mammary fat pad region of each mouse. Tumor size was measured twice weekly using a caliper. For the Snrnp200-targeted ASO monotherapy assay for the 4T1 cell-derived xenograft model, the mice were divided into two groups: (1) the vehicle group (PBS, 50µL, subcutaneous injection twice a week) and (2) the ASO-Snrnp200 group (5mg/kg, subcutaneous injection twice a week). For the combination therapy assay, the mice were injected with 4T1 cells and randomly divided into six groups: (1) vehicle (50µL of PBS, injected intraperitoneally daily) plus isotype (IgG2a, 10mg/kg injected intraperitoneally twice a week, Bio X Cell, BE0089); (2) vehicle (50µL of PBS, injected i.p. daily) plus anti-PD-1 (RMP1-14, 10mg/kg injected i.p. twice a week, Bio X Cell, BE0146); (3) anti-LDH (FX-11, 2mg/kg injected i.p. daily, MedChemExpress, HY-16214) plus isotype (IgG2a, 10mg/kg injected i.p. twice a week); (4) anti-LDH (FX-11, 2mg/kg injected i.p. daily) plus anti-PD-1 (RMP1-14, 10mg/kg injected i.p. twice a week); (5) ASO-Snrnp200 (5mg/kg, subcutaneous injection twice a week) plus isotype (IgG2a, 10mg/kg injected i.p.twice a week); and (6) ASO-Snrnp200 (5mg/kg, subcutaneous injection twice). The tumor volume in mm³ was calculated using the formula: tumor volume = 0.5×L×W², where L is the longest dimension and W is the perpendicular dimension.

Flow cytometry analysis

For flow cytometry analysis of in vivo experiments, mouse tumors were quickly excised and then mechanically dissociated with scissors in sterile PBS. Tumor fragments were digested in serum-free RPMI+1mg/mL collagenase D (Roche, 11088866001)+1mg/mL collagenase I (Sigma-Aldrich, C0130)+20mg/mL DNase I (Roche, 10104159001) for 30–60min at 37°C with rotation to promote dissociation. Single-cell suspensions were passed through a 70mM cell strainer. The red blood cells in the tumor samples were then lysed with red blood cell lysis buffer (eBioscience, 00-4300-54) for 5min at room temperature. Lysis reactions were quenched by the addition of 20mL of PBS, and the samples were subsequently centrifuged at 300×g for 5min at 4°C. The cells were then washed in D-PBS and stained with fluorescently labeled antibodies against the following surface proteins at a 1:100 dilution in Cell Staining Buffer (BioLegend, 420201) for 30min at 4°C: PE-conjugated anti-mouse CD45 (ThermoFisher Scientific, 12-0451-82), CD3e (ThermoFisher Scientific, 17-0032-82), CD4 (BD Biosciences, 550954), CD25 (BD Biosciences, 553071), CD8a (ThermoFisher Scientific, 11-0081-82), CTLA-4 (BD Biosciences, 553720), PD-1 (BD Biosciences, 562523), and GITR (BD Biosciences, 558119). For all the intracellular protein analyses, the cells were then fixed and permeabilized by using fixation buffer (BioLegend, 420801) and intracellular staining permeabilization wash buffer (BioLegend, 421002) according to the manufacturer’s instructions. Permeabilized cells were then incubated with fluorescently labeled antibodies against FOXP3 (BD Biosciences, 560401), ICOS (BD Biosciences, 552146), granzyme B (BioLegend, 372214), and IFN-γ (BioLegend, 505850) for 30min at 4°C. A CytoFLEX S flow cytometer (Beckman Coulter) was used for flow cytometry data acquisition, and the data were analyzed with FlowJo software (version 10.5.3).

mIHC analysis

As described previously, antibodies against SNRNP200 (Abcam, ab241589), EFTUD2 (Abcam, ab188327), PRPF8 (Abcam, ab79237), FASN (Abcam, ab128870), LDH (Abcam, ab52488), FOXP3 (Servicebio, GB112325), CD4 (Servicebio, GB15064), PD-1 (Cell Signaling Technology, 84651), and CD8a (Cell Signaling Technology, 98941) were used following the manufacturer’s instructions (Panovue, 0060000050). In brief, tissue sections were rehydrated in ethanol and deparaffinized in xylene. After a 15-min microwave antigen retrieval step in hot citric acid buffer (pH 6.0), endogenous peroxidase activity was quenched with 3% H₂O₂ for 10min, and nonspecific binding sites were blocked with goat serum for 10min. Following incubation with primary antibodies in a humidified room temperature environment for 1h, the samples were treated with secondary horseradish peroxidase-conjugated polymers. To visualize each target, fluorescein tyramide signal amplification staining (1:100) was applied. An antigen retrieval step using microwave heating with citric acid buffer (pH 6.0) was performed before proceeding to the next stage to remove excess antibodies. Finally, antifade mounting media was used to cover the sections, and DAPI (Sigma-Aldrich, D5942) was utilized to visualize the cell nuclei. Whole-slide scanning fluorescence images were scanned via an Olympus VS200 and analyzed via OlyVia software (version 3.4.1). A list of the primary antibodies is provided in Supplementary Table S8.

Lactic acid and glutathione content assays

For the lactic acid content assay of in vivo experiments, mouse tumors were detected with a lactic acid content detection kit (Solarbio, BC2235) according to the manufacturer’s instructions. For the glutathione content assay, the tumors were frozen in liquid nitrogen. GSH and GSSG assays were performed, and the GSH/GSSG ratio was calculated using a GSH/GSSG detection assay (Beyotime, S0053) according to the manufacturer’s instructions.

CIBERSORT

The “Cell type Identification by Estimating Relative Subsets of RNA Transcripts (CIBERSORT)” calculated with the “kappa” function in R was used to calculate the abundance of 22 types of immune cell subsets in each sample⁶⁷.

AUCell

The AUCell R package (version 1.24.0) was utilized to assess RNA splicing activity in individual breast cancer cells. Initially, a set of 42 spliceosome genes upregulated in glycolytic TNBCs was used as input to derive a spliceosome gene score. The AUC was used to estimate the proportion of highly expressed genes within the set for each cell. Cells expressing a greater number of genes from this set demonstrated elevated AUC values. The ‘AUCell_exploreThresholds’ function determines the threshold for defining the active state of the current gene set. Then, UMAP embedding for cell clustering was color-coded based on the AUC of each cell.

This work received funding from the National Natural Science Foundation of China (81872137, 82072917, 82272957, and 82303735), the Shanghai Science and Technology Innovation Action Plan (22DZ2204400), and the Ministry of Science and Technology of China (2023YFF0600039, 2023YFF1205003, and 2023YFC3402504, the National Key R&D Program of China). The funders were not involved in the study design, data collection and analysis, publication decisions, or manuscript preparation.