ESRP1 regulates alternative splicing of CARM1 to sensitize small cell lung cancer cells to chemotherapy by inhibiting TGF-β/Smad signaling

ESRP1 positively correlates with the survival of patients with SCLC

To clarify the role of ESRP1 in SCLC, we examined its clinical relevance in patients with SCLC. We first analyzed the differential expression of ESRP1 in the tumor and adjacent normal tissues from five patients by Western blot, which revealed that the expression of ESRP1 was significantly decreased in SCLC tumor tissues as compared with corresponding adjacent normal tissues (Figure 1A). Then, we investigated levels of ESRP1 mRNA expression in 56 human SCLC and 13 normal lung tissue samples by qRT-PCR. The results of this analysis revealed lower ESRP1 mRNA expression in SCLC tissues than in normal tissues (Figure 1B). Moreover, analysis of the results showed that low expression levels of ESRP1 was significantly associated with smoking history, extensive disease, and worse status in patients with SCLC (Table 1). Kaplan-Meier analysis revealed that low ESRP1 mRNA levels in SCLC tissues correlated with reduced overall survival (Figure 1C). The univariate analysis showed that ESRP1 expression levels correlated with disease stage and survival in SCLC, whereas the multivariate analysis suggested that ESRP1 was not an independent predictor of prognosis in these patients (Table 2). Overall, the above analysis indicated that ESRP1 was down-regulated in SCLC chemoresistant cells and tumor tissues, and that the expression of ESRP1 was correlated with the overall survival of SCLC patients.

ESRP1 is reduced in chemoresistant SCLC cells and inhibits chemoresistance of SCLC cells in vitro and in vivo

Previous research by our group showed that the expression of circular RNA cESRP1 in SCLC chemoresistant cells was significantly lower than that in chemosensitive cells and suggested that cESRP1 promotes chemosensitivity of SCLC by inhibiting the TGF-β signaling pathway [20]. Therefore, we speculated whether the ESRP1 protein formed by the same precursor RNA may also participate in the chemoresistance of SCLC. To confirm this hypothesis, we evaluated the mRNA and protein expression levels of ESRP1 in SCLC chemoresistant and chemosensitive cells. We found decreased expression of ESRP1 in H69AR and H446DDP cells compared with H69 and H446 cells, this result was in concordance with their higher IC₅₀ values for chemotherapeutic drugs (Figure 2A–2C).

In order to clarify the role of ESRP1 in chemoresistance of SCLC, we transfected H69AR and H446DDP cells with lentivirus to stably overexpress ESRP1. The qRT-PCR and western blot assays were performed to judge vector transfection efficiency (Figure 2D). CCK8 assays were conducted to detect the chemosensitivity of SCLC cells to cisplatin and etoposide. The results showed that overexpression of ESRP1 increased chemosensitivity, with significant decreases in the IC₅₀ values (Figure 2E). In addition, we also examined the effects of ESRP1 on cell apoptosis and cell cycle exposure to chemotherapeutic drugs. Compared with the control group, overexpression of ESRP1 significantly increased cell apoptosis and cell cycle arrest (Figure 2F, 2G).

Then we further verified the effect of ESRP1 on chemoresistance of SCLC in the parental sensitive H69 and H446 cells. We established stable ESRP1 knockdown in H69 and H446 cells, and through qRT-PCR and western blot experiments verified the knockdown efficiency of ESRP1 (Figure 2H). Contrary to the results of chemoresistant cells, decreased expression of ESRP1 resulted in promotion of SCLC chemoresistance (Figure 2I). Similarly, we confirmed the effect of ESRP1 knockdown on cell apoptosis and cell cycle. Contrary to the upregulation of ESRP1, the cell apoptosis and cell cycle arrest were reduced significantly (Figure 2J, 2K).

To further assess whether ESRP1 influenced chemoresistance of SCLC in vivo, we subcutaneously transplanted H69AR and H69 with altered ESRP1 expression into nude mice (Figure 3A, 3D). Consistent with our in vitro findings, overexpression of ESRP1 significantly reduced tumor volume and weight compared with those of the paired control group (Figure 3B, 3C), whereas knockdown of ESRP1 led to the opposite results (Figure 3E, 3F). These results indicate that ESRP1 promotes chemosensitivity and inhibits the growth of SCLC in vitro and in vivo.

ESRP1 mediates chemoresistance of SCLC by regulating alternative splicing of CARM1

To clarify the mechanism by which ESRP1 reversed SCLC chemoresistance, we conducted mRNA transcriptome sequencing with H69AR cells that stably overexpressed ESRP1. We identified 283 ESRP1-regulated AS events and found that various types of AS could be regulated by ESRP1, including skipped exon (SE), alternative 5′ss exon (A5SS), alternative 3′ss exon(A3SS), retained intron (RI), and mutually exclusive exons (MXEs) (Figure 4A and Supplementary Table 2). Then, we analyzed cellular functions of ESRP1-regulated AS events using gene ontology (GO) and found that ESRP1 affected genes in the RNA processing pathway, including mRNA splicing and mRNA metabolic processes (Figure 4B). Moreover, ESRP1 targets were also enriched with drug resistance-related functions, such as transmembrane receptor protein tyrosine kinase signaling pathway and Rho GTPase binding. Intriguingly, several ESRP1-regulated AS events were found to regulate the response to the chemical stimulus pathway. Although this enrichment of response to chemical stimulus was slightly below our significance cut-off, the changes in exon inclusion ratios were fairly large and, therefore, may have significant functional effects on consequences of drug resistance. Many of the ESRP1-regulated splicing targets were functionally connected to well-linked interaction networks, as judged by the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) (Figure 4C). As expected, a large subgroup of ESRP1 targets contained genes involved in RNA processing. Surprisingly, the other two subgroups included many genes involved in cell migration and cell cycle. Taken together, these results suggest that the biological processes affected by ESRP1 are related to apoptosis, proliferation, migration, and tumorigenesis. We subsequently validated mRNA-seq results by measuring splicing change in five newly identified targets that were selected arbitrarily to include genes with a drug resistance-related function (Figure 4E). We confirmed that ESRP1 either positively or negatively controlled all endogenous AS events tested and that the relative changes in exon inclusion ratio obtained from RT-PCR were highly consistent with mRNA-seq (Figure 4D). Taken together, these data imply that ESRP1 may be a key regulator that controls cancer progression.

We then explored the downstream splicing targets of ESRP1. Studies show that CARM1 methylated MED12 can increase chemosensitivity of breast cancer cells [26]. Wang et al. have confirmed that there are two major splice variants of CARM1 produced by AS, which are the full-length transcript of CARM1 (CARM1FL) and the deletion of exon 15 transcript (CARM1ΔE15). It has been confirmed that the deletion of exon 15 of CARM1 could block the self-methylation of CARM1, which in turn affects the transcriptional regulation of CARM1 on estrogen receptor-α (ERα). The CARM1 self-methylation site R511 is located on the fifteenth exon, and the self-methylation defect of CARM1 reduces the activity of CARM1 on the methylation of some protein substrates [27]. Therefore, we hypothesized that ESRP1 may regulate AS of CARM1 to affect SCLC chemoresistance. We performed a RIP assay using ESRP1 monoclonal antibody and found that ESRP1 enriched the pre-mRNA of CARM1 (Figure 4F). We detected the content of two splice isoforms of CARM1 in SCLC cells, and found that the content of CARM1FL in chemoresistant cells was significantly higher than that in chemosensitive cells, whereas the content of CARM1ΔE15 was completely opposite; this suggested that the alternative splicing of CARM1 is related to the chemoresistance of SCLC (Figure 4G). The RT-PCR results showed that ESRP1 increased the proportion of CARM1ΔE15 and decreased the proportion of CARM1FL (Figure 4H).

Since ESRP1 reduced the ratio of CARM1FL to CARM1ΔE15, in order to further examine the different roles of CARM1FL and CARM1ΔE15 in regulating chemoresistance of SCLC, we designed specific siRNAs targeting CARM1FL and CARM1ΔE15 transcripts respectively. We transfected siRNA in H69AR cells and verified the knockdown efficiency by qRT-PCR (Figure 5A). The CCK8 assay showed that the IC50 value was significantly decreased after CARM1FL or CARM1 knockdown, whereas the IC50 value did not change significantly after CARM1ΔE15 down-regulation (Figure 5B). These results suggested that knockdown of CARM1FL inhibits chemoresistance, CARM1ΔE15 may have no effect on chemoresistance of SCLC. To further demonstrate that ESRP1 affects chemoresistance of SCLC by regulating the content of different transcripts of CARM1, we performed rescue experiments in H69AR and H446DDP cells. We overexpressed CARM1FL or CARM1ΔE15 by using transfection plasmids and confirmed overexpression efficiency by qRT-PCR (Figure 5C and Supplementary Figure 1A). The CCK8 assays were performed in ESRP1-overexpressing cells with up-regulation of CARM1FL or CARM1ΔE15. The results showed significantly decreased IC50 values in ESRP1 overexpressing cells compared with the control group; upregulation of CARM1FL in ESRP1 overexpressing cells could reverse the decrease in IC50 values mediated by ESRP1 overexpression, whereas upregulation of CARM1ΔE15 cannot reverse the IC50 values (Figure 5D and Supplementary Figure 1B). Next, we conducted apoptosis and cell cycle cytometry analyses. We found that increased cell apoptosis and cell cycle arrest in ESRP1 overexpressing cells, and upregulation of CARM1FL in ESRP1 overexpressing cells could rescue the increase in cell apoptosis and cell cycle arrest regulated by ESRP1 overexpression. Similarly, upregulation of CARM1ΔE15 in ESRP1 overexpressing cells cannot rescue these (Figure 5F, 5G and Supplementary Figure 1C, 1D). In addition, western blot experiments showed that the changes to the expression levels of apoptotic proteins PARP and BAX are consistent with the results of flow cytometry (Figure 5E). These results demonstrate that ESRP1 affects chemoresistance of SCLC by reducing the proportion of CARM1FL, whereas CARM1ΔE15 has no effect on chemoresistance.

CARM1 activates TGF-β/Smad signaling pathway by regulating Smad7 arginine methylation CARM1 belongs to the PRMT proteins family, PRMTs can methylate arginine residues in histones, epigenetically control the expression of a series of genes, and can also modify non-histone proteins, including signal-conducting systems, to regulate their functions [28, 29]. Now that CARM1 has been shown to be regulated by ESRP1 and affects chemoresistance in SCLC, then we need to explore the mechanism through which CARM1 works. We used the pMeS database to predict genes that may undergo arginine methylation and found that there were 17 arginine methylation sites in the sequence of Smad7 (Table 3). Research by Katsuno et al. found that PRMT1 can methylate Smad6 and Smad7, promote the separation of Smad6/Smad7 from the receptor, and activate BMP-induced Smad1/Smad5 and TGF-β-induced Smad2/Smad3 [25, 30]. Furthermore, our group’s previous research already has shown that Smad7 can participate in chemoresistance in SCLC [20]. Therefore, we speculated that CARM1 could regulate arginine methylation of Smad7 to activate the TGF-β/Smad signaling pathway thereby promoting the chemoresistance of SCLC. To confirm the relationship between CARM1 and Smad7, we performed a Co-IP assay in H69AR cells and found that CARM1 physically interacted with Smad7 (Figure 6A). Then we used methyltransferase inhibitor ADOX to treat H69AR cells and observed the IC50 values by CCK8 assay (Figure 6B). The results showed that the IC50 value significantly decreased after inhibiting the effect of methyltransferase, thereby suggesting that CARM1 indeed promoted SCLC chemoresistance by methylating Smad7.

In order to explore the relationship between chemoresistance and EMT phenotype, we first detected the basic morphology of chemoresistant cell H446DDP and chemosensitive cell H446. The results showed that H446 cells showed an epithelial phenotype, whereas H446DDP cells showed more of a mesenchymal phenotype (Figure 6C). Moreover, we analyzed the expression levels of EMT markers in chemoresistant cells and chemosensitive cells via immunofluorescence, we found that the expression levels of the mesenchymal markers vimentin and N-cadherin in chemoresistant cells H69AR and H446DDP were significantly higher than those of chemosensitive cells, whereas the expression level of epithelial marker E-cadherin was completely opposite (Figure 6D). We subsequently conducted rescue experiments to explore the mechanism by which the ESRP1/CARM1 axis involved in regulating the TGF-β/Smad signaling pathway and EMT. Overexpression of ESRP1 reduced the expression of phosphorylated-Smad3(p-Smad3), vimentin and N-cadherin; silencing of ESRP1 resulted in the opposite effects (Figure 6E). Overexpression of CARM1FL increased the expression of p-Smad3, EMT-inducing TF snail and reduced the expression of E-cadherin, EMT-inhibiting TF OVOL2, but the results could be rescued by addition of ADOX; Overexpression of CARM1ΔE15 had no effect on the protein expression (Figure 6F). Finally, we overexpressed ESRP1 in combination with upregulation of CARM1FL or CARM1ΔE15 in cells. We found that compared with overexpression of ESRP1 alone, simultaneous overexpression of ESRP1 and CARM1FL increased the expression of p-Smad3, vimentin, and N-cadherin. However, this effect was reversed after the addition of ADOX. Overexpression of ESRP1 and CARM1ΔE15 at the same time had no significant effects on the TGF-β/Smad signaling pathway and EMT compared with ESRP1 overexpression alone (Figure 6G). These results were consistent with previous functional experimental results, and proved that ESRP1 inhibits the activation of TGF-β/Smad signaling pathway by reducing the content of CARM1FL and thereby enhancing chemosensitivity of SCLC.

Clinical samples

Fifty-six tumor tissue sections were collected from the First Affiliated Hospital of Hebei North University and the Minzu Hospital of Guangxi Zhuang Autonomous Region that had been approved for use by the ethics committee. RNA extraction was done from paraffin-embedded tissue samples using the formalin-fixed paraffin-embedded (FFPE) RNA kit (Gibco, Guangzhou, China) to assess ESRP1 gene expression lever.

Cell culture and reagents

Human chemosensitive SCLC cell lines (H69 and H446) and chemoresistant cells (H69AR) were purchased from the American Type Culture Collection (ATCC, USA). Human chemoresistant cell line (H446DDP) was induced by continuous cisplatin treatment of H446 cells in our laboratory. All cells were cultured in RPMI1640 basal medium (Gibco, Guangzhou China) plus 10% calf serum at 37° C, in a constant temperature incubator, containing 5% CO₂.

Adenosine dialdehyde (ADOX) was purchased from Selleck Chemicals (Houston, Texas, USA); ADOX is an adenosine analog that inhibits methyltransferase in cells. Cells were treated with 5 μM ADOX for 48 hours and then related experiments were performed.

RNA isolation and real-time polymerase chain reaction (qRT-PCR)

Extraction of total RNA from cells and FFPE tissues was done using TRIzol (Invitrogen, USA) and RNeasy FFPE Kit (Qiagen). NanoDrop 2000 (Thermo) was used to detect the concentration of extracted RNA. Next, qRT-PCR was performed with an ABI Illumina instrument (Foster, USA) using SYBR Green (Tiangen). The sequences of all qPCR primers are shown in the Supplementary Materials (Supplementary Table 1).

Western blot

Total protein was extracted from cells using RIPA buffer (Biyuntian, China). The concentration of the extracted protein was detected using a BCA protein quantification kit (Biyuntian, China). Equivalent amounts of cell protein lysates were electrophoresed on 10% SDS-polyacrylamide gels. The membrane was blocked with 5% bovine serum albumin (BSA) for 2 hours at room temperature and then incubated overnight with a primary antibody at 4° C. The membrane was incubated with horseradish peroxidase-conjugated anti-rabbit or anti-mouse secondary antibody (EarthOx, USA) for 1 hour. The immune complexes were detected by chemiluminescence (ECL).

siRNA and plasmid transfection

We transiently transfected cells with siRNA or plasmid using Lipofectamine 3000 (Thermo Scientific), P-3000 (Thermo Scientific) and OPTI-MEM (Invitrogen) to knockdown or overexpress CARM1, according to the manufacturer’s instructions.

Cell counting kit-8 assays (CCK-8)

SCLC cells were cultured in 96-well plates with 1–2×10⁴ cells per well. The cells were treated with cisplatin (Shandong) or etoposide (Jiangsu) for 24 hours, then 10 μL CCK8 reagent was added and absorbance was detected at 450 nm. According to the OD value measured, the 50% inhibitory concentration (IC₅₀ value) of the chemotherapeutic drug was calculated.

Tumor xenograft experiments

Animal experiments for this study were performed in accordance with the institutional guidelines of Guangdong Province and the Use Committee for Animal Care. Female BALB/c nude mice aged 3–4 weeks were purchased from the Experimental Animal Center of Southern Medical University (Guangdong, China). The cell suspension was adjusted to 1×10⁷ cells per 200 μL phosphate-buffered saline (PBS). 200 μL cell suspension was injected subcutaneously on the flank of each nude mouse separately. Fourteen days after subcutaneous tumor implantation, nude mice were injected intraperitoneally with chemotherapeutic drugs (cisplatin 2.5 mg/kg, day 1 and etoposide 4 mg/kg, days 1–3) or saline. The tumor volume of nude mice was measured every 3 days. Nude mice were sacrificed after two cycles of drug injection.

Flow cytometry and cell cycle analysis

Cells were treated with cisplatin or etoposide for 24 hours and then collected for further assays. Cell apoptosis was detected using the Annexin V APC/PI apoptosis kit (BestBio, China). In the cell cycle experiment, cells were fixed with 75% ethanol for 4 hours, and then we used RNase (Solarbio) and PI (Sigma, USA) to detect the cell cycle. All samples were analyzed by the CytoFLEX flow cytometer (Beckman).

Immunofluorescence assays

Cells were seeded on slides for 24 hours. The cells fixed with 4% paraformaldehyde for 10 min and permeabilized with 0.5% Triton X-100 for 10 min at room temperature. Then, the cells were rinsed in PBS three times for 5 min each time and incubated with primary antibody overnight at 4° C. After three washings with PBS, the cells were incubated with fluorescent secondary antibody at room temperature for 1 hours. Then, the cells incubated with DAPI for 10 min at 37° C. Finally, we added an anti-fluorescence quencher on the slide and images were acquired using a confocal microscope.

Co-immunoprecipitation assay (Co-IP)

We performed Co-IP assay, according to the operating instructions of the Co-IP kit (Thermo Scientific). The immunoprecipitated proteins were detected by western blots.

RNA immunoprecipitation assay (RIP)

We performed the RIP assay, according to the operating instructions of the Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore).

Nucleic acid electrophoresis

Agarose (Biyuntian, China), 50× TAE buffer (Biyuntian), and 10000× Gel Red (Biyuntian) were used to prepare an electrophoresis solution. The nucleic acid which was amplified by qRT-PCR was mixed with 6× loading buffer (Biyuntian), and electrophoresis was performed at a constant voltage of 100 V in electrophoresis solution. A fluorescence picture was taken by electrophoresis.

Statistical analysis

Statistical analysis of data was done using GraphPad Prism 7.0 and SPSS. Data are expressed as mean ± standard deviation (SD). Analysis of differences between data was done by two independent sample t-tests or one-way ANOVA. The association between ESRP1 expression and clinical features was analyzed by Chi-Square test or Fisher’s exact test. The survival curve was assessed by Kaplan-Meier analysis. P<0.05 was considered statistically significant.