Role of the splicing factor SRSF4 in cisplatin-induced modifications of pre-mRNA splicing and apoptosis

Abstract

Background: Modification of splicing by chemotherapeutic drugs has usually been evaluated on a limited number of pre-mRNAs selected for their recognized or potential importance in cell proliferation or apoptosis. However, the pathways linking splicing alterations to the efficiency of cancer therapy remain unclear.

Methods: Next-generation sequencing was used to analyse the transcriptome of breast carcinoma cells treated by cisplatin. Pharmacological inhibitors, RNA interference, cells deficient in specific signalling pathways, RT-PCR and FACS analysis were used to investigate how the anti-cancer drug cisplatin affected alternative splicing and the cell death pathway.

Results: We identified 717 splicing events affected by cisplatin, including 245 events involving cassette exons. Gene ontology analysis indicates that cell cycle, mRNA processing and pre-mRNA splicing were the main pathways affected. Importantly, the cisplatin-induced splicing alterations required class I PI3Ks P110β but not components such as ATM, ATR and p53 that are involved in the DNA damage response. The siRNA-mediated depletion of the splicing regulator SRSF4, but not SRSF6, expression abrogated many of the splicing alterations as well as cell death induced by cisplatin.

Conclusion: Many of the splicing alterations induced by cisplatin are caused by SRSF4 and they contribute to apoptosis in a process requires class I PI3K.

Figure 1

Platinum-based chemotherapeutic agents affect MDM2 and VEGF pre-mRNA splicing. A-D: MCF7 and Ishikawa cells were treated with the indicated concentrations of cisplatin and harvested after 24 hours. E: MCF7 cells treated with cisplatin (50 μM) were harvested at the indicated times. F: The indicated cells were treated with cisplatin (50 μM) for 24 hours. G: MCF7 cells were treated with oxaliplatin at 50 μM and harvested after 24 hours. Analysis of the splicing of MDM2 (A,B,E-G) or of VEGF-A (C,D) transcripts was performed by end-point RT-PCR and acrylamide gel electrophoresis as detailed in Methods. Illustrated gels are representative of three independent experiments. FL: Full Length; ALT1: Splice variant of MDM2.

Figure 2

Analysis of selected splicing events modified by cisplatin in MCF7 cells. End-point RT-PCR and analysis of amplification products by acrylamide gel electrophoresis were performed on control and cisplatin (50 μM, 24 hours)-treated MCF7 cells to validate inclusion/exclusion events detected by RNA-Seq and SpliceSeq analysis. Cisplatin-treatment induced the inclusion of HNRNPDL exon 8 (E8; ***p = 0.0008), MTA1 exons 3–4 (E3-4; *p = 0.047) and NFE2L1 exon 5 (E5; **p = 0.008), and the exclusions of CSDE1 exons 2–4 (delta E2-4; ***p = 0.00003), HNRNPDL exon 6 (delta E6; **p = 0.0045), EIF4A2 exon 4 (delta E4; * = 0.02), TMPO exons 6–8 (delta E6-8; **p = 0.009), AMZ2 exon 3 (delta E3; ***p = 0.001), STRAP exon 2 (delta E2; p = 0.059) and MAGOH exon 3 (delta E3; * = 0.04). Graphs show the mean and SD and are representative of at least 3 independent experiments.

Figure 3

Lack of contribution of ATM, ATR and DNA-PK pathways in cisplatin-induced splicing. AT5BIVA (ATM deficient, A) and MO59J (DNA-PK deficient, B) cells were treated with cisplatin (50 μM, 24 hours) and analysed for alternative splicing events in HRNPDL pre-mRNA. (*p ≤ 0.05; **p ≤ 0.01); C-D: MCF7 cells were treated with ATM inhibitor (50 μM) or DNA-PK inhibitor (NU7026; 25 μM) three hours prior to treatment with cisplatin (50 μM for 24 hours); E-F: MCF7 cells were pre-treated with caffeine (5 mM) and with DNA-PK inhibitor (NU7026; 25 μM) for three hours prior to treatment with cisplatin (50 μM) (E: HNRNPDL-E6 p = 0.13; F: AMZ2: p = 0.49). G-L: MCF7 cells were treated with wortmannin (100 nM and 500 nM) or PX866 (500 nM), three hours prior to treatment with cisplatin (50 μM, 24 hours). Modifications of alternative splicing were evaluated for G: HNRNPDL-E6 *p = 0.02; H: HNRNPDL-E8 *p = 0.02; I: AMZ2 **p = 0.004, ***p = 0.0005; J: MDM2. Similar modification was observed with PX866 and illustrated for K: HNRNPDL-E6 *p = 0.05; L: AMZ2 *p = 0.02. Alternative splicing was evaluated by end-point RT-PCR and acrylamide gel electrophoresis. Each bar shows the mean with SD of at least three independent experiments.

Figure 4

Involvement of PI3K pathway, but not Akt, in cisplatin-induced splicing. A-F: MCF7 cells were treated with TGX221 (5 μM; A-B) or IC87114 (10 μM; C-D) three hours prior to treatment with cisplatin (50 μM, 24 hours). E-H: MCF7 cells were treated with triciribine (20 μM) or MK2206 (2 μM) three hours before cisplatin treatment (50 μM; 24 hours). Alternative splicing of exon 6 of HNRNPDL (B,D,F,H) and exon 3 of AMZ2 (A,C,E,G) was evaluated by RT-PCR. RT-PCR products were fractionated by gel electrophoresis.

Figure 5

SRSF4 is involved in cisplatin-induced splicing. A-B: MCF7 cells were transfected with siRNA targeting SRSF4 or/and SRSF6. Cells were harvested three days post-transfection and SRSF4 and SRSF6 mRNA levels were measured by end-point RT-PCR to control the efficiency of the siRNA. C-F: the histograms and errors bars represent mean and SD, respectively, illustrating the inter-experiment differences in the percentage of exon inclusion (n = 5 to 7). However, the statistics were made on the fold change measured in each independent experiment. MCF7 cells transfected with control siRNA (siSCR) or siRNA targeting SRSF4, SRSF6 or both were treated with cisplatin. The splicing of MDM2 (C), HNRNPDL-E6 (**p = 0.012; ***p = 0.0013) (D), HNRPDL-E8 (**p = 0.0163) (E) and AMZ2 (***p = 0.0002, **p = 0.001) (F) was evaluated by end-point RT-PCR.

Figure 6

SRSF4 contributes to cisplatin-induced cell death. A-D: Apoptosis was measured by FACS after annexin V/propidium iodide staining of siSCR (A,B) or siSRSF4 (C,D) transfected MCF-7 cells untreated (A,C) or treated with cisplatin (B,D; 50 μM, 24 hours). The percentages in each quartile are mean values calculated from three independent experiments. E. Quantification of data from trypan blue exclusion and FACS analysis. Histograms indicate cell death in siSRSF4 transfected MCF7 cells treated with cisplatin as compared to death in cells transfected with control siRNA (siSCR) taken as 100%. Data were corrected for cell death measured in untreated cells. Each bar shows the mean with SD of three independent experiments.