PTBP1 Regulates DNMT3B Alternative Splicing by Interacting With RALY to Enhance the Radioresistance of Prostate Cancer

Abstract

Radiotherapy is a curative arsenal for prostate cancer (PCa), but radioresistance seriously compromises its effectiveness. Dysregulated RNA splicing factors are extensively involved in tumor progression. Nonetheless, the role of splicing factors in radioresistance remains largely unexplored in PCa. Here, 23 splicing factors that are differentially expressed between PCa and adjacent normal tissues across multiple public PCa databases are identified. Among those genes, polypyrimidine tract binding protein 1 (PTBP1) is significantly upregulated in PCa and is positively associated with advanced clinicopathological features and poor prognosis. Gain- and loss-of-function experiments demonstrate that PTBP1 markedly reinforces genomic DNA stability to desensitize PCa cells to irradiation in vitro and in vivo. Mechanistically, PTBP1 interacts with the heterogeneous nuclear ribonucleoproteins (hnRNP) associated with lethal yellow protein homolog (RALY) and regulates exon 5 splicing of DNA methyltransferase 3b (DNMT3B) from DNMT3B-S to DNMT3B-L. Furthermore, upregulation of DNMT3B-L induces promoter methylation of dual-specificity phosphatase-2 (DUSP2) and subsequently inhibits DUSP2 expression, thereby increasing radioresistance in PCa. The findings highlight the role of splicing factors in inducing aberrant splicing events in response to radiotherapy and the potential role of PTBP1 and DNMT3B-L in reversing radioresistance in PCa.

Keywords: DNMT3B; PTBP1; prostate cancer; radioresistance; splicing factor.

Figure 1

Elevated PTBP1 expression is closely associated with poor prognosis in prostate cancer. a) Venn diagram showing the overlapping splicing factors with dysregulated expression in prostate cancer (PCa) across the indicated public databases. GSA: Genome Sequence Archive, GEO: Gene Expression Omnibus database. TCGA: The Cancer Genome Atlas. b) Forrest plots of hazard ratios for the associations between 23 dysregulated splicing factors and progression‐free intervals (left) and disease‐free intervals (right) in the TCGA database. Prognostic splicing factors (red) with p <0.05 were considered statistically significant. c) Difference in the expression of PTBP1, POLR2H, NONO and BUB3 between low Gleason score (6‐7(3+4)) PCa tissues and high Gleason score (7(4+3)−10) PCa tissues in the TCGA database. d) Difference in the expression of PTBP1, POLR2H, NONO and BUB3 between low T stage (T2) PCa tissues and high T stage (T3‐4) PCa tissues in the TCGA database. e) Representative immunohistochemical images of PTBP1 in high and low Gleason score PCa tissues and normal tissues. Scale bar, 50 µm. f) Difference in the expression of PTBP1 between PCa tissues and normal tissues in cohort 1. g) Difference in the expression of PTBP1 between low Gleason score (6‐7(3+4)) PCa tissues and high Gleason score (7(4+3)−10) PCa tissues from cohort 1(left) and cohort 2 (right). h) Difference in the expression of PTBP1 between low T stage (T2) PCa tissues and high T stage (T3‐4) PCa tissues from cohort 1(left) and cohort 2 (right). i) Kaplan–Meier curves for the overall survival of PCa patients with high or low expression of PTBP1 in Cohort 1 (left) and Cohort 2 (right). j) Kaplan–Meier curves for the cancer‐specific survival of PCa patients with high or low expression of PTBP1 in Cohort 1 (left) and Cohort 2 (right). *p < 0.05, **p < 0.01, ***p < 0.001, ns. no significance. Student's t test.

Figure 2

PTBP1 facilitates prostate cancer cell resistance to radiotherapy by maintaining genomic stability a) Clonogenic survival in response to irradiation (0, 2, and 4 Gy) of prostate cancer (PCa) cells transiently transfected with the indicated siRNAs. b) Clonal survival of PCa cells transfected with PTBP1 or control plasmids in response to irradiation (0, 2, and 4 Gy). c,d) Images (c) and quantification (d) of cell apoptosis of PCa cells transfected with the indicated siRNAs and exposed to 4 Gy irradiation. e) Quantification of the cell apoptosis of PCa cells transfected with PTBP1 or control plasmids and exposed to 4 Gy irradiation. f,g) Representative images (left) and statistical analyses (right) of the comet assay results of the indicated DU145 (f) and PC‐3 (g) cells at 0, 1, and 24 h after 4 Gy irradiation. Scale bar, 20 µm. h,i) Protein levels of γ‐H2AX in the indicated DU145 (h) and PC‐3 (i) cells at the indicated times after 4 Gy irradiation were measured via western blot (left) and gray quantitative analysis (right). The data are presented as the means ± S.D.s of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001 by two‐tailed Student's t test or by two‐way ANOVA followed by Tukey's post‐hoc test where applicable. ns. no significance.

Figure 3

Overexpression of PTBP1 increases radioresistance of prostate cancer cells in vivo. a) Representative image of xenograft tumors in the indicated groups treated with or without 4 Gy irradiation. b) Tumor volume curves in the indicated groups following irradiation treatment, measured every 3 days. c) Statistical analysis of the tumor volume in the indicated groups. d) Representative images of hematoxylin‐eosin and immunohistochemical staining of PTBP1, ki‐67 and γ‐H2AX in xenograft tumors from the indicated groups; Scale bar, 50 µm. e–g) Statistical analyses of difference in PTBP1 (e), Ki‐67 (f) and γ‐H2AX (g) expression among the indicated groups. The data are presented as the mean ± S.D.s of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001 by two‐tailed Student's t test, or by two‐way ANOVA followed by Tukey's post‐hoc test where applicable. ns. no significance.

Figure 4

PTBP1 mediates the inclusion of exon 5 in DNMT3B pre‐mRNA. a) Quantitative analysis of the different alternative splicing events, including five categories, after PC‐3 cells were transfected with the indicated siRNAs. b) Scatterplots showing PTBP1 knockdown with the indicated siRNAs affecting skipped exon events. c) Intron exclusion (left) or inclusion (right) of candidate genes in PTBP1 knockdown prostate cancer (PCa) cells was validated by agarose gel electrophoresis of PCR products. d) Representative gel electrophoresis image of DNMT3B intron retention in PTBP1 overexpressing PCa cells. e) RT‐PCR‐based statistical analysis of the DNMT3B‐L/DNMT3B‐S ratio in PTBP1 knockdown or PTBP1 overexpressing PCa cells. f) Representative images of RNA‐FISH images showing the changes in DNMT3B‐L and DNMT3B‐S in PTBP1‐knockdown or PTBP1‐overexpressing PCa cells. Scale bar, 20 µm. g,h) Representative images (g) and statistical analysis (h) of DNMT3B pre‐mRNA enrichment on PTBP1 from the RIP assay in PCa cells using anti‐PTBP1 antibody. i) RNA‐pulldown assay followed by western blotting demonstrating the interaction of PTBP1 with the 4 indicated intron segments of DNMT3B pre‐mRNA. j Specific interactions of PTBP1 with 5‐1 intron segments of DNMT3B pre‐mRNA were validated by RNA pulldown assay followed by western blotting. K,l) Schematic diagram (k) and protein expression (l) of PTBP1 truncation variants lacking RRM1, RRM2, RRM3, or RRM4. m,n) Representative image (m) and statistical analysis (n) of RT‐PCR results of DNMT3B‐L/DNMT3B‐S PSI in DU145 cells with PTBP1 knockdown, and re‐overexpression of PTBP1 deletion mutants. The data are presented as the means ± S.D.s of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001 by two‐tailed Student's t test, or by two‐way ANOVA followed by Tukey's post‐hoc test where applicable. ns. no significance.

Figure 5

PTBP1 interacts with RALY to regulate DNMT3B splicing in prostate cancer cells a,b) Co‐immunoprecipitation analysis of the interaction between PTBP1 and RALY in prostate cancer (PCa) cells with an anti‐PTBP1 antibody (a) or anti‐RALY antibody (b). c) Immunofluorescence demonstrating the subcellular colocalization of PTBP1(red) and RALY (green) in PCa cells. Scale bar, 20 µm. d) Co‐immunoprecipitation analysis of the interaction between RALY and PTBP1 truncation variants with anti‐MYC antibody. e) Co‐immunoprecipitation analysis of the interaction between PTBP1 and RALY truncation variants with anti‐FLAG antibody. f) Clonogenic survival of PCa cells transiently transfected with the indicated siRNAs against RALY in response to irradiation (0, 2, and 4 Gy). g,h) Representative gel electrophoresis image (g) and statistical analysis (h) of the DNMT3B‐L/DNMT3B‐S ratio in RALY knockdown PCa cells by RT‐PCR. i,j) Representative images (i) and statistical analysis (j) of DNMT3B pre‐mRNA enrichment on RALY from RIP assay in PCa cells using anti‐RALY antibody. k) Statistical analysis of DNMT3B pre‐mRNA enrichment on PTBP1 from the RIP assay in RALY knockdown or control PCa cells using anti‐PTBP1 antibody. l) Statistical analysis of DNMT3B pre‐mRNA enrichment on RALY from the RIP assay in PTBP1 knockdown or control PCa cells using anti‐RALY antibody. The data are presented as the mean ± S.D.s of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001 by two‐tailed Student's t test, or by two‐way ANOVA followed by Tukey's post‐hoc test where applicable. ns. no significance.

Figure 6

PTBP1 increases radioresistance of prostate cancer cells via a DNMT3B‐L‐dependent manner. a,b) Representative gel electrophoresis image (a) and statistical analysis (b) of the overexpression efficiency of DNMT3B‐L and DNMT3B‐S in prostate cancer (PCa) cells. c) Clonogenic survival in response to irradiation (0, 2, and 4 Gy) of PCa cells transfected with the indicated DNMT3B isoforms or control plasmids. d,e) Representative gel electrophoresis image (d) and statistical analysis (e) of the efficiency of DNMT3B‐L knockdown in PCa cells. f) Clonogenic survival in response to irradiation (0, 2, and 4 Gy) of PCa cells transfected with DNMT3B‐L siRNA or control siRNA. g,h) Clonogenic survival of PCa cells treated as indicated and exposed to irradiation (0, 2, and 4 Gy). i–l) Representative images (left) and statistical analyses (right) of the comet assay results of the indicated DU145 (i, k) and PC‐3 (j, l) cells at 24 h after 4 Gy irradiation. Scale bar, 20 µm. The data are presented as the mean ± S.D.s of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001 by two‐tailed Student's t test, or by two‐way ANOVA followed by Tukey's post‐hoc test where applicable. ns. no significance.

Figure 7

PTBP1 modulates DUSP2 through DNMT3B‐L mediated DNA methylation a,b) The expression of the indicated target genes was detected via qPCR in PTBP1(a) or DNMT3B‐L (b) knockdown prostate cancer (PCa) cells. c,d) The expression of DUSP2 was verified in PCa cells subjected to the indicated treatments via western blot assay. e) Difference in DUSP2 expression between PCa tissues and normal tissues in the TCGA database. f) Kaplan–Meier curves for progression‐free survival of PCa patients with high or low expression of DUSP2 in TCGA database. g) Correlation between DUSP2 expression and the methylation level of the indicated methylation sites in the TCGA data. h,i) DUSP2 expression in PCa cells treated with 5‐Aza was detected by qPCR (h) and western blotting (i). j,k) Representative gel electrophoresis image (j) and statistical analysis (k) of the methylation of DUSP2 promoter in PCa cells treated as indicated. l,m) Clonal survival of PCa cells subjected to the indicated treatment followed by irradiation. n) Schematic model of the mechanism underlying the role of PTBP1 in PCa radioresistance. The data are presented as the mean ± S.D.s of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001 by two‐tailed Student's t test, or by two‐way ANOVA followed by Tukey's post‐hoc test where applicable. ns. no significance.