Therapy-induced stress response is associated with downregulation of pre-mRNA splicing in cancer cells

Abstract

Background: Abnormal pre-mRNA splicing regulation is common in cancer, but the effects of chemotherapy on this process remain unclear.

Methods: To evaluate the effect of chemotherapy on slicing regulation, we performed meta-analyses of previously published transcriptomic, proteomic, phosphoproteomic, and secretome datasets. Our findings were verified by LC-MS/MS, western blotting, immunofluorescence, and FACS analyses of multiple cancer cell lines treated with cisplatin and pladienolide B.

Results: Our results revealed that different types of chemotherapy lead to similar changes in alternative splicing by inducing intron retention in multiple genes. To determine the mechanism underlying this effect, we analyzed gene expression in 101 cell lines affected by ɣ-irradiation, hypoxia, and 10 various chemotherapeutic drugs. Strikingly, оnly genes involved in the cell cycle and pre-mRNA splicing regulation were changed in a similar manner in all 335 tested samples regardless of stress stimuli. We revealed significant downregulation of gene expression levels in these two pathways, which could be explained by the observed decrease in splicing efficiency and global intron retention. We showed that the levels of active spliceosomal proteins might be further post-translationally decreased by phosphorylation and export into the extracellular space. To further explore these bioinformatics findings, we performed proteomic analysis of cisplatin-treated ovarian cancer cells. Finally, we demonstrated that the splicing inhibitor pladienolide B impairs the cellular response to DNA damage and significantly increases the sensitivity of cancer cells to chemotherapy.

Conclusions: Decreased splicing efficiency and global intron retention is a novel stress response mechanism that may promote survival of malignant cells following therapy. We found that this mechanism can be inhibited by pladienolide B, which significantly increases the sensitivity of cancer cells to cisplatin which makes it a good candidate drug for improving the efficiency of cancer therapy.

Keywords: Alternative splicing; Cell cycle; Chemotherapy; DNA damage response; Pladienolide B; Spliceosome.

Fig. 1

Analysis of pre-mRNA splicing changes in cancer cell lines and xenografts after chemotherapy. a Enrichment analysis of genes with common differential splicing events induced by therapy. Left side: lung adenocarcinoma patient-derived xenografts (PDX) treated with carboplatin, docetaxel, afatinib, BEZ235, BKM120, DAPT, erlotinib, tivantinib, and selumetinib; dataset GSE69405. Right side: A375, A549, H3122, N87, PC9, RT112 cell lines treated with erlotinib, crizotinib, trametinib, lapatinib, vemurafenib, BGJ398; dataset GSE89127. The STRING database was used for Gene Ontology Biological Processes analysis. p value is indicated with a color scale. b Summary of alternative splicing events observed in GSE89127 and GSE69405 datasets (before slash: total number of splicing events; after slash: common splicing events appeared in at least half of the samples). Events: SE—skipped exon, A5SS—alternative 5′ splice site, A3SS—alternative 3′ splice site, RI—retained intron, MXE—mutually exclusive exons. c Scatter plot representing the intron retention (upper panel) and exon skipping (lower panel) events detected in the GSE89127 dataset before and after chemotherapy. Splicing events in spliceosomal genes are illustrated with a dark-blue color. d Sashimi plots for the splicing factor RBM6 in untreated cancer cells (dark blue) and in cancer cells that were treated with different chemotherapeutic drugs (light blue). The inclusion level (IncLevel) indicates the splicing status of the intron. e Heat map demonstrating the changes in the expression of spliceosomal genes (Z-score) after chemotherapy. Clusterization of expression data was made before scaling data (Z-score transformation)

Fig. 2

Changes in gene expression in cancer cells in response to different stress conditions. a Workflow of the meta-analysis of mRNA microarray gene expression data. b Heat maps of selected pathways obtained by KEGG (top panel) and Reactome (bottom panel) enrichment analyses of up and downregulated genes in each of the six meta-analyses of gene expression

Fig. 3

Different types of clusterization show concerted changes in the expression of spliceosomal and cell cycle genes. a–с Time clusterization of gene expression data (right panel) and subsequent pathway enrichment analysis (left panel) of clusters with highly represented spliceosomal genes. The clusters were constructed based on the following datasets: a E-GEOD-8057, b E-GEOD-18494, and c E-GEOD-59861. Blue lines represent z-score values of gene expression in each cluster. The red line is a mean value of z-score values of a cluster. d Graph representing the common transcription factor SOX2 that may induce concerted changes in the expression of pairs of mitotic and splicing genes after a course of chemotherapy. Solid black lines connect a pair of co-expressed genes, and red lines connect transcription factors with their target genes. Additional file 4: Figure S2A, B shows similar graphs for the transcription factors GFI1B and TARDBP

Fig. 4

Changes in proteomic profiles induced by different stress conditions. a Comparison of proteins that were phosphorylated in a breast cancer cell line after gamma-irradiation [49] (blue); in an osteosarcoma cancer cell line after gamma-irradiation [48] (orange) and in a melanoma cancer cell line after neocarzinostatin [50] (red). The number of spliceosomal proteins in a given sector is shown after the slash. b Venn diagram of upregulated proteins in therapy-induced secretomes of ovarian cancer cells (blue circle), glioblastoma cells (orange circle), and ovarian cancer ascites obtained from patients after the course of chemotherapy (red circle) [51, 52]. The number of spliceosomal proteins in a given sector is shown after the slash. c Venn diagram representing the proteins identified in SKOV3 cells before (blue) and after (red) cisplatin treatment. d Results of the enrichment analysis of proteins (from “C”) for which abundance was decreased by more than twofold after chemotherapy. e Results of Fisher’s test of the intersection between differentially secreted proteins (derived from data we reported in [51, 52]) and the hits from siRNA screening (based on the study by Paulsen et al. [53]). f Intersection of the lists of the spliceosomal genes with a decrease in expression (green circle, according to our meta-analyses of microarray data), intron retention in transcripts (blue circle, according to our meta-analyses of RNA-Seq data), upregulated secretion of the corresponding proteins (orange circle, according to our previous proteomic data) and upregulated protein phosphorylation (red circle, according to the analysis of phosphoproteomics data) observed after treatment with chemotherapeutic drugs

Fig. 5

Pladienolide B impairs pre-mRNA splicing and increases the sensitivity of cancer cells to cisplatin. a Viability assay of SKOV3, A549, HepG2, and HT29 cells that were pretreated with 2 nM pladienolide B (2 days) following treatment with different concentrations of cisplatin (4 days). b FACS analysis of caspase 3/7 and SYTOX staining of A549 cells treated with 0.5 nM pladienolide B, 10 μM Cisplatin or both drugs together. c FACS analysis of phospho ATM staining of SKOV3 cells that were cultivated with different concentrations of pladienolide B (2 days) and subsequently treated with 10 μM Cisplatin (1 day). d Representative immunofluorescence images of SKOV3 cells stained for phosphoATM (S1981) (green) and with DAPI (blue) after treatment with 10 μM Cisplatin in the presence or absence of 0.5 nM Pladienolide B. Scale bar: 50 μm. e Enrichment analysis of genes affected by differential splicing events before and after treatment with splicing inhibitors: pladienolide B (upper part; E-GEOD-67770) and spliceostatin A (lower part; GSE72156). The STRING database was used for Gene Ontology Biological Processes analysis