CSTF3 contributes to platinum resistance in ovarian cancer through alternative polyadenylation of lncRNA NEAT1 and generating the short isoform NEAT1_1

Abstract

Platinum-based chemotherapy is the standard postoperative adjuvant treatment for ovarian cancer (OC). Despite the initial response to chemotherapy, 85% of advanced OC patients will have recurrent disease. Relapsed disease and platinum resistance are the major causes of death in OC patients. In this study, we compared the global regulation of alternative polyadenylation (APA) in platinum-resistant and platinum-sensitive tissues of OC patients by analyzing a set of single-cell RNA sequencing (scRNA-seq) data from public databases and found that platinum-resistant patients exhibited global 3' untranslated region (UTR) shortening due to the different usage of polyadenylation sites (PASs). The APA regulator CSTF3 was the most significantly upregulated gene in epithelial cells of platinum-resistant OC. CSTF3 knockdown increased the sensitivity of OC cells to platinum. The lncRNA NEAT1 has two isoforms, short (NEAT1_1) and long (NEAT1_2) transcript, because of the APA processing in 3'UTR. We found that CSTF3 knockdown reduced the usage of NEAT1 proximal PAS to lengthen the transcript and facilitate the expression of NEAT1_2. Downregulation of the expression of NEAT1 (NEAT1_1/_2), but not only NEAT1_2, also increased the sensitivity of OC cells to platinum. Overexpressed NEAT1_1 reversed the platinum resistance of OC cells after knocking down CSTF3 expression. Furthermore, downregulated expression of CSTF3 and NEAT1_1, rather than NEAT1_2, was positively correlated with inactivation of the PI3K/AKT/mTOR pathway in OC cells. Together, our findings revealed a novel mechanism of APA regulation in platinum-resistant OC. CSTF3 directly bound downstream of the NEAT1 proximal PAS to generate the short isoform NEAT1_1 and was conducive to platinum resistance, which provides a potential biomarker and therapeutic strategy for platinum-resistant OC patients.

Fig. 1. ScRNA-seq exploratory data analysis for platinum-resistant and sensitive OC tissues.

A Composition of the four patient scRNA-seq datasets by tSNE plot. B Individual tumor cells were assigned to five cell types. C Annotation validation of epithelial cells (EPI) using the highly expressed epithelial marker genes EPCAM, KRT8, KRT18, and KRT19. D Cells were assigned to seven types based on the expression of variable genes and visualized using tSNE. E The proportion of cell types for platinum-resistant and platinum-sensitive patients. F The proximal peak usage index of platinum-resistant and platinum-sensitive epithelial cells. G Transcripts with significant 3 ‘UTR length changes between resistant and sensitive epithelial cells. H The distribution of APA peaks. I Differential expression analysis of APA primary regulators in platinum-resistant epithelial cells compared with sensitive cells.

Fig. 2. CSTF3 facilitated platinum resistance in OC cells in vitro.

A Construction of platinum-resistant OC cell lines A2780-DDP and OVCAR3-DDP cells. Western blot (B) and RT‒qPCR (C) were performed to detect the expression of CSTF3 in A2780-DDP, OVCAR3-DDP and corresponding parental cells. D CSTF3 protein was measured to confirm the knockdown efficiency in A2780-DDP and OVCAR3-DDP cells. E Cell viability was evaluated to measure the IC50 of platinum after knockdown of CSTF3 and treatment with gradient concentrations of platinum in A2780-DDP and OVCAR3-DDP cells. F A2780-DDP and OVCAR3-DDP cells were treated with 100 μM and 40 μM platinum for 6 h respectively. Colony formation assay was performed with control and CSTF3 knockdown A2780-DDP and OVCAR3-DDP cells treated with or without platinum. G Western blot was performed to verify CSTF3 knockdown efficiency in WT A2780 and OVCAR3 cells. H A2780 and OVCAR3 cells with CSTF3 knockdown were treated with different concentrations of platinum, and cell survival was determined. I A2780 and OVCAR3 cells were treated with 20 μM and 5 μM platinum for 6 h respectively. Colony formation assays were performed in A2780 and OVCAR3 cells treated with or without platinum after CSTF3 knockdown. J FLAG-NC and FLAG-CSTF3 were stably transfected into A2780 and OVCAR3 cells. Then, cell viability curves (K) and colony formation assays (L) were performed to measure the sensitivity of cells to platinum after overexpression of CSTF3. *P < 0.05, **P < 0.01, ***P < 0.001 vs. shNC.

Fig. 3. CSTF3 facilitated platinum resistance in OC cells in vivo.

OVCAR3 cells transfected with CSTF3 knockdown or negative control constructs were injected subcutaneously into BALB/c nude mice. The xenografted mice were treated with or without DDP (4 mg/kg) through intraperitoneal injection once every 6 days. A Photo of representative tumors from six groups of xenografted nude mice on the 22nd day. B The growth of tumor volumes was monitored every 3 days beginning on the 7th day. C Tumor weights are presented as the mean ± SD from five mice per group. D, E HE staining was performed to evaluate tissue morphology, and IHC was performed to visualize Ki-67- and Caspase-3-positive staining in xenografted tumors. *P < 0.05, **P < 0.01, ***P < 0.001 vs. shNC.

Fig. 4. CSTF3 bound and regulated the length of the 3’UTR of lncRNA NEAT1 in OC cells.

A CSTF3 knockdown induced an APA shift of target genes in A2780 cells, left: scatterplot of the 3’UTR alteration in CSTF3 knockdown cells when compared with the control cells, right: the usage of proximal and distal PAS in CSTF3 knockdown cells. B Detection of APA events in CSTF3 knockdown OVCAR3 cells, combining 3’UTR alteration and usage of PAS. C Filtering out the directly bound target genes of CSTF3 by eCLIP-seq. D The consensus sequences of CSTF3 binding sites detected by HOMER motif analysis with eCLIP-seq data. E Distribution of CSTF3-targeted transcript types (left) and CSTF3 binding sites within different regions (right) as identified through eCLIP-seq. F The Venn diagram screened the candidate target genes using CSTF3 eCLIP-seq, A2780 and OVCAR3 PAS-seq data. G IGV genome browser showing the PAS usage of the NEAT1 3′UTR by eCLIP-seq. H Schematic diagram for primer design of short and long NEAT1 transcripts to validate the NEAT1 APA regulation by CSTF3. I Histogram showing the relative expression of the NEAT1 isoform with the distal PAS (dPAS) relative to that with the proximal PAS (pPAS) when CSTF3 was downregulated in A2780, A2780-DDP, OVCAR3 and OVCAR3-DDP cells compared with the negative control. J CLIP-qPCR detected the interaction between CSTF3 and NEAT1. K RT-qPCR was used to detect the relative expression of NEAT1 in A2780-DDP, OVCAR3-DDP and corresponding parental cells. *P < 0.05, **P < 0.01, ***P < 0.001.

Fig. 5. The short 3’UTR isoform of NEAT1 promotes platinum resistance in OC cells.

A The IC50 values were evaluated when silencing the expression of NEAT1 and NEAT1_2 in A2780, OVCAR3 and corresponding platinum-resistant cells compared with the negative control. B A2780, OVCAR3, A2780-DDP and OVCAR3-DDP cells were treated with 20 μM, 5 μM, 100 μM and 40 μM platinum for 6 h respectively. Colony formation assay was performed to measure the sensitivity of the drug when silencing the expression of NEAT1 and NEAT1_2 in A2780, OVCAR3 and corresponding platinum-resistant cells treated with or without platinum. C, D The IC50 values and colony formation were evaluated when NEAT1_1 was overexpressed in A2780, OVCAR3 and platinum-resistant cells with CSTF3 knockdown treated with platinum or not. *P < 0.05, **P < 0.01, ***P < 0.001. ns not significant.

Fig. 6. The change of paraspeckles formation in A2780 and OVCAR3 cells when downregulating the expression of CSTF3, NEAT1 or NEAT1_2.

Immunofluorescence assays were performed with CSTF3 knockdown in A2780 (A) and OVCAR3 (B) cells, and images of DAPI (blue), CSTF3 (green), PSPC1 (red, left) and SFPQ (red, right) were obtained. C Immunofluorescence of A2780 and OVCAR3 cells transfected with si-NEAT1, si-NEAT1_2 and the negative control. DAPI (blue), PSPC1 (green) and SFPQ (red) were detected.

Fig. 7. CSTF3 and NEAT1_1 regulate the PI3K/AKT pathway.

A Overlap of upregulated and downregulated genes in A2780 and OVCAR3 cells when CSTF3 was knocked down. B KEGG enrichment analysis for overlap of upregulated and downregulated genes. C Overlap of differentially expressed genes (DEGs) in OVCAR3 cells when NEAT1 and NEAT1_2 were silenced. D KEGG analysis of DEGs for silencing NEAT1. E KEGG analysis of DEGs for silencing NEAT1_2. F Histogram of KEGG analysis for the independent DEGs list with NEAT1 and NEAT1_2 silenced. G KEGG enrichment analysis of the individual DEGs list with NEAT1 silenced. H KEGG enrichment analysis of the individual DEGs list with NEAT1_2 silenced. I The levels of PI3K/AKT/mTOR pathway-related proteins were detected by western blot when CSTF3 was downregulated in A2780 and OVCAR3 cells. J Western blot was used to measure the levels of PI3K/AKT/mTOR pathway-related proteins with NEAT1 or NEAT1_2 siRNA transfected into A2780 and OVCAR3 cells.

Fig. 8

Diagram of the regulatory mechanism and function underlying CSTF3 mediating the platinum resistance of OC cells.