SRSF1 promotes vascular smooth muscle cell proliferation through a Δ133p53/EGR1/KLF5 pathway

Abstract

Though vascular smooth muscle cell (VSMC) proliferation underlies all cardiovascular hyperplastic disorders, our understanding of the molecular mechanisms responsible for this cellular process is still incomplete. Here we report that SRSF1 (serine/arginine-rich splicing factor 1), an essential splicing factor, promotes VSMC proliferation and injury-induced neointima formation. Vascular injury in vivo and proliferative stimuli in vitro stimulate SRSF1 expression. Mice lacking SRSF1 specifically in SMCs develop less intimal thickening after wire injury. Expression of SRSF1 in rat arteries enhances neointima formation. SRSF1 overexpression increases, while SRSF1 knockdown suppresses the proliferation and migration of cultured human aortic and coronary arterial SMCs. Mechanistically, SRSF1 favours the induction of a truncated p53 isoform, Δ133p53, which has an equal proliferative effect and in turn transcriptionally activates Krüppel-like factor 5 (KLF5) via the Δ133p53-EGR1 complex, resulting in an accelerated cell-cycle progression and increased VSMC proliferation. Our study provides a potential therapeutic target for vascular hyperplastic disease.

Figure 1. Increased SRSF1 expression in proliferating VSMCs in vivo and in vitro.

(a) Photomicrographs of haematoxylin/eosin-stained carotid arteries from sham-operated and balloon-injured rats. Immunohistochemical staining of vessels with specific anti-SRSF1 antibody revealed SRSF1 mainly in the neointima (N, neointima; M, media). Normal rabbit IgG served as a negative control. Scale bars, 50 μm. (b) Immunofluorescence double staining of injured carotid arteries with specific antibodies against SM-α-actin (green) or SRSF1 (red). Nuclei were stained with DAPI (blue), and yellow indicates their co-localization in the merged images. Scale bars, 50 μm. (c) Real-time PCR showing the mRNA levels of SRSF1 in carotid arteries at 1, 4, 7, 14 and 21 days after balloon injury; n=8 per group. (d) Representative western blots and averaged data showing SRSF1 and PCNA levels in rat carotid arteries at 1, 4, 7, 14 and 21 days after balloon injury; n=8 per group. (e) Real-time PCR data showing the mRNA levels of SRSF1 in HASMCs treated with Ang II (200 nM), serum (10% FBS) or PDGF-BB (10 μg l⁻¹) at 6, 12 and 24 h; n=8 per group. (f) Representative western blots and averaged data showing SRSF1 levels in HASMCs treated as in e; n=7 per group. *P<0.05, **P<0.01, one-way ANOVA (c–f). Data are mean±s.e.m. of five independent experiments (c–f).

Figure 2. SRSF1 deficiency in VSMCs inhibits neointima formation.

(a) PCR showing genotyping of WT (Flox^+/+Cre⁻) and Srsf1^−/− (Flox^+/+Cre⁺) mice. Band in the upper image indicates the product from SRSF1^flox/flox, and lower image indicates product from SM22α-Cre. (b) Representative western blots showing SRSF1 levels in vessel from smooth muscle cell-specific Srsf1 knockout (Srsf1^−/−) mice. (c,d) Representative photomicrographs of haematoxylin and eosin staining (scale bars, 50 μm) (c) and averaged data (d) of the neointimal area, neointima/media ratio, media area and circumference of external elastic lamina (EEL) of carotid arteries from Srsf1^−/− and WT control mice 14 and 28 days after wire injury (N, neointima; M, media); n=9 per group. (e,f) Representative photomicrographs of immunohistochemical staining (scale bar, 25 μm) (e) and averaged data (f) showing the percentages of PCNA-positive cells in carotid arteries from Srsf1^−/− and WT control mice 14 and 28 days after wire injury. Arrows indicate PCNA-positive cells (dark brown); n=9 per group. (g,h) Representative pictures (g) and averaged data (h) of re-endothelialization. Re-endothelialization was quantified in Evans blue-stained carotid arteries at 3 and 7 days after vascular injury. Blue staining indicates endothelial denudation. Scale bar, 1 mm; n=9 per group. **P<0.01, NS, not significant; Student’s t-test (d,f,h). Data are mean±s.e.m. of five independent experiments (d,f,h).

Figure 3. Inducible SMC-specific SRSF1 deficiency inhibits neointima formation.

(a) PCR showing genotyping of WT (Flox^+/+Cre^ERT2−) and Srsf1^−/−/Cre^ERT2 (Flox^+/+Cre^ERT2+) mice. Band in the upper image indicates the product from SRSF1^flox/flox, and lower image indicates product from SMA-Cre^ERT2. (b) Representative western blots showing SRSF1 levels in the vessel and liver from inducible smooth muscle cell-specific Srsf1 knockout (Srsf1^−/−/Cre^ERT2) mice. (c,d) Representative photomicrographs of haematoxylin and eosin staining (scale bars, 50 μm) (c) and averaged data (d) of the neointimal area, neointima/media ratio, media area and circumference of external elastic lamina (EEL) of carotid arteries from Srsf1^−/−/Cre^ERT2 and WT control mice 14 and 28 days after wire injury (N, neointima; M, media); n=9 per group. (e,f) Representative photomicrographs of immunohistochemical staining (scale bar, 25 μm) (e) and averaged data (f) showing the percentages of PCNA-positive cells in carotid arteries from Srsf1^−/−/Cre^ERT2 and WT control mice 14 and 28 days after wire injury. Arrows indicate PCNA-positive cells (dark brown); n=9 per group. (g,h) Representative pictures (g) and averaged data (h) of re-endothelialization. Re-endothelialization was quantified in Evans blue-stained carotid arteries at 3 and 7 days after vascular injury. Blue staining indicates endothelial denudation. Scale bar, 1 mm; n=9 per group. **P<0.01, NS, not significant; Student’s t-test (d,f,h). Data are mean±s.e.m. of five independent experiments (d,f,h).

Figure 4. SRSF1 enhances HASMC proliferation and neointima formation.

(a) Western blots showing SRSF1 protein in rat carotid arteries 4 days after Ad-SRSF1-GFP delivery; n=5. (b,c) Haematoxylin/eosin staining (scale bars, 50 μm) (b) and averaged data (c) of the neointimal area, neointima/media ratio, media area and circumference of external elastic lamina (EEL) of rat carotid arteries transfected with Ad-GFP or Ad-SRSF1 14 days postinjury (N, neointima; M, media); n=8 each. (d) PCNA staining (scale bar, 25 μm) (left) and the percentage of PCNA-positive cells (right) in rat carotid arteries infected with Ad-SRSF1 14 days postinjury. Arrows indicate PCNA-positive cells (dark brown); n=8 each. (e) SRSF1 expression in cultured HASMCs infected with Ad-GFP or Ad-SRSF1 (m.o.i. 50 and 100); n=5. (f) PCNA protein levels in HASMCs infected with Ad-GFP or Ad-SRSF1; n=5 each. (g,h) Cell counts (g) and MTT assays (h) of HASMCs infected with Ad-GFP or Ad-SRSF1 after Ang II or PDGF-BB stimulation for 4 days; n=12 each. (i) Representative images and migration distance from wound-healing assays in HASMCs infected with Ad-GFP or Ad-SRSF1; n=9 each. Scale bar, 200 μm. (j) SRSF1 expression in HASMCs infected with scrambled or SRSF1 siRNAs (si1 and si2); n=6 each. (k) PCNA levels in cultured HASMCs infected with scrambled or SRSF1 siRNAs after Ang II stimulation (24 h); n=5 each. (l) MTT assays of HASMCs infected with SRSF1 siRNAs after Ang II stimulation for 4 days; n=9 each. (m) Representative images and averaged data from wound-healing assays in HASMCs infected with SRSF1 siRNAs and treated with Ang II; n=12 each; scale bar, 200 μm. All adenoviral infection above is 100 m.o.i. for 48 h unless specified. Ang II is 200 nM and PDGF-BB is 10 μg l⁻¹. Scr indicates scrambled siRNA control. *P<0.05, **P<0.01, NS, not significant; Student’s t-test (c,d,i) or one-way ANOVA (f,h,j–m) or two-way ANOVA (g). Data are mean±s.e.m. of five (c,d,f,j,k) or four (g–i,l,m) independent experiments.

Figure 5. SRSF1 modulates Δ133p53 in regulating VSMC proliferation.

(a,b) Δ133p53 expression increases by Ad-SRSF1 infection in HASMCs at mRNA (a) and protein (b) level; n=7 each. (c,d) SiRNA-deletion of SRSF1 reduces mRNA (c) and protein (d) abundance of Δ133p53; n=5–6 each. (e) MΔ133p53 levels in vessels from Srsf1^−/− mice at 56 days after birth; n = 8 each. (f) Δ133p53 expression in cultured HASMCs by Ad-Δ133p53; n=5 each. (g–i) In HASMCs infected with Ad-GFP, Ad-Δ133p53, Ad-p53 or Ad-Δ40p53, PCNA expression (g), cell counts (h) and MTT assays (i) after Ang II or PDGF-BB stimulation for 4 days are examined; (g), n=5; (h,i), n=12. (j) Wound-healing assays in HASMCs infected with indicated adenovirus and treated Ang II; n=9 each; scale bar, 200 μm. All adenoviral infection above is 100 m.o.i. for 48 h unless specified. Ang II is 200 nM and PDGF-BB is 10 μg l⁻¹. **P < 0.01; one-way ANOVA (a–d,g,i,j), Student's t-test (e), or two-way ANOVA (h). Data are mean±s.e.m. of five (a–d,g) or four (e,h–j) independent experiments.

Figure 6. Δ133p53 promotes neointima formation after vascular injury.

(a,b) Hematoxylin/eosin staining (a) and averaged data (b) of the neointimal area, neointima/media ratio, media area, and circumference of external elastic lamina (EEL) of rat carotid arteries transfected with indicated adenovirus 2 weeks postinjury (N, neointima; M, media); n=8 each; scale bar, 50 μm. (c) PCNA staining and PCNA-positive cells in rat carotid arteries treated as in (a) 2 weeks postinjury. Arrows indicate PCNA-positive cells; n=8 each; scale bar, 25 μm. (d) Δ133p53 siRNAs (Δ133 si1 and Δ133 si2) knock down Δ133p53 level in HASMCs; n=8. (e–g) PCNA level (e), MTT assays (f) and wound-healing assays (g) in HASMCs infected with Δ133p53 siRNAs after Ang II stimulation; (e), n=8; (f), n=11; (g), n=9. (h) PCNA levels in HASMCs infected with Δ133p53 siRNAs in the presence or absence of Ad-SRSF1; n=6. (i) MTT assays showing SRSF1-mediated enhancement of proliferation was inhibited by Δ133p53 knockdown; n=12. (j) Averaged data from wound-healing assays in HASMCs infected with Δ133p53 siRNAs in the presence or absence of Ad-SRSF1 and treated with Ang II; n=8. All adenoviral infection in HASMCs above is 100 m.o.i. for 48 h. Ang II is 200 nM. Scr indicates scrambled siRNA. *P <0.05, **P<0.01, NS, not significant; one-way ANOVA (b–j), Data are mean±s.e.m. of five (b–e,h) or four (f,g,i,j) independent experiments.

Figure 7. SRSF1 promotes VSMC proliferation via KLF5–p21 signal.

(a) Venn diagram of modulated genes in HASMCs overexpressing Δ133p53 and treated with Ang II, showing 247 genes that fit the criteria of >0.5 FPKM (fragments per kilobase of exon per million fragments mapped) and an adjusted P-value of <0.05 (Wald tests implemented in the DESeq R package), commonly modulated in both conditions. (b,c) The mRNA levels (b) and protein abundance (c) of KLF5 in HASMCs with Δ133p53 overexpression or Δ133p53 knockdown; n=7–8 per group. (d,e) KLF5 expression in HASMCs with SRSF1 overexpression or SRSF1 knockdown at mRNA level (d) and protein level (e); n=6 per group. (f) Cell-cycle distributions in HASMCs infected with Ad-SRSF1 in the absence or presence of KLF5 siRNA stimulated with Ang II (200 nM) for 24 h; n=9 per group. (g) KLF5 siRNAs (KLF5 si1 and KLF5 si2) knock down KLF5 protein in HASMCs; n=8 per group. (h) Expression of p21 in HASMCs infected with Ad-SRSF1 or Ad-Δ133p53; n=5 per group. (i) p21 levels in HASMCs with SRSF1 knockdown or Δ133p53 knockdown; n=7 per group. (j) PCNA and p21 levels in HASMCs infected with KLF5 siRNAs in the absence or presence of Ad-SRSF1; n=7 per group. (k) KLF5 and p21 levels in HASMCs infected with Δ133p53 siRNAs (Δ133 si1 and Δ133 si2) in the presence or absence of Ad-SRSF1; n=7 per group. (l,m) KLF5 and p21 levels in cultured HASMCs infected with SRSF1 siRNAs (l) or Δ133p53 siRNAs (m) after Ang II stimulation (200 nM, 24 h); n=5–7 per group. (n) KLF5 and p21 levels in the arteries from Srsf1^−/− or WT control mice; n=7 per group. (o) KLF5 and p21 levels in rat carotid arteries transfected with Ad-SRSF1 1 week after balloon injury; n=10 per group. Scr indicates scrambled siRNA control; BI, balloon injury. *P<0.05, **P<0.01, one-way ANOVA (b–f,h–m,o) or Student’s t-test (n). Data are mean±s.e.m. of four (b,d) or five (c,e–o) independent experiments.

Figure 8. SRSF1-triggered activation of KLF5 requires EGR1 binding to Δ133p53.

(a) Co-immunoprecipitation (IP) assay identified interaction between Δ133p53 and EGR1. Cell lysates from HASMCs infected with Ad-β-gal or Ad-Δ133p53-myc were immunoprecipitated with antibody to myc (left) or EGR1 (right) and detected with western blots. Input lysates were loaded as control; n=5 per group. (b) Chromatin immunoprecipitation (ChIP) assays were performed to detect the recruitment of EGR1 protein to the promoter of KLF5 after overexpression of Δ133p53. HASMCs were infected with Ad-β-gal or Ad-Δ133p53-myc. Crosslinked DNA–protein lysates are immunoprecipitated with antibody to EGR1. The fragments containing consensus EGR1-binding motifs are PCR amplified from both samples. Four distinct EGR1-binding positions (P1, P2, P3 and P4) in KLF5 are shown. n=9 per group. (c) Representative western blots and averaged data showing the EGR1 levels in HASMCs infected with scrambled or two sets of EGR1 siRNAs (EGR1 si1 and EGR1 si2). (d) ChIP assays were performed in HASMCs infected with Ad-β-gal or Ad-Δ133p53-myc in the presence or absence of EGR1 siRNA1 (EGR1 si); n=9 per group. Crosslinked DNA–protein lysates from infected cells are immunoprecipitated with EGR1 antibody. The P1 and P2 promoter fragments (same as Fig. 7b) containing EGR1-binding motifs are PCR amplified from both samples. (e) Representative western blots and averaged data showing KLF5, p21 and PCNA levels in HASMCs transfected with EGR1 siRNAs with or without Ad-SRSF1 overexpression; n=7 per group. (f) Representative western blots and averaged data showing KLF5, p21 and PCNA levels in HASMCs transfected with EGR1 siRNAs with or without Ad-Δ133p53 overexpression; n=9 per group. (g,h) The proliferation enhancement induced by SRSF1 (g) or Δ133p53 (h) was attenuated by EGR1 siRNAs; n=12 per group. Scr indicates scrambled siRNA control. *P<0.05, **P<0.01, NS, not significant; Student’s t-test (b) or one-way ANOVA (d–h). Data are mean±s.e.m. of four (b,d) or five (e–h) independent experiments.

Figure 9. SRSF1 facilitates endothelia cell migration and proliferation.

(a) Representative pictures and averaged data of re-endothelialization. Re-endothelialization was quantified in Evans blue-stained carotid arteries at 3 and 7 days after vascular injury. Blue staining indicates endothelial denudation. n=9 per group. Scale bar, 10 mm. (b) mRNA levels of SRSF1 in human coranary artery endothelia cells (HCAECs) treated with Ang II at 6, 12 and 24 h; n=8 per group. (c) SRSF1 expression in HCAECs treated as in a; n=7 per group. (d) Overexpression of SRSF1 by Ad-SRSF1 infection in HCAECs; n=5 per group. (e) Δ133p53 levels in HCAECs infected with Ad-GFP or Ad-SRSF1; n=6 per group. (f) Overexpression of Δ133p53 in cultured HCAECs by Ad-Δ133p53 infection; n=5 per group. (g) Representative images and averaged data from wound-healing assays in HCAECs infected with Ad-GFP, Ad-SRSF1 or Ad-Δ133p53 and treated Ang II; n=9 per group. Scale bar, 200 μm. (h) Δ133p53 siRNAs (Δ133 si1 and Δ133 si2) deplete Δ133p53 protein in HCAECs; n=6 per group. (i) Representative images and averaged data from wound-healing assays in HCAECs infected with Δ133p53 siRNAs with or without Ad-SRSF1 overexpression and treated with Ang II; n=9 per group. Scale bar, 200 μm. (j,k) PCNA expression in cultured HCAECs infected with Ad-SRSF1 (j) or Ad-Δ133p53 (k); n=5 each. (l) MTT assays of HCAECs infected with Ad-GFP, Ad-SRSF1 or Ad-Δ133p53 after Ang II for 4 days; n=12 per group. (m) KLF5, p21, and PCNA levels in HCAECs infected with scrambled or Δ133p53 siRNAs in the presence or absence of Ad-SRSF1 overexpression; n=7 per group. (n) MTT assays showing the SRSF1-mediated enhancement of proliferation was inhibited by Δ133p53 knockdown in HCAECs; n=15 per group. Adenoviral infection above is 100 m.o.i. for 48 h. Ang II concentration is 200 nM. Scr indicates scrambled siRNA control. *P<0.05, **P<0.01, one-way ANOVA (a–c,e,g–n). Data are mean±s.e.m. of four independent experiments (a–c,e,g–n).

Figure 10. SRSF1 enhances human coronary artery SMC migration and proliferation.

(a) The mRNA levels of SRSF1 in human coronary artery smooth muscle cells (HCASMCs) treated with Ang II at 6, 12 and 24 h; n=8 each. (b) SRSF1 levels in HCASMCs treated as in a; n=9 each. (c) SRSF1 expression in cultured HCASMCs infected with Ad-GFP or Ad-SRSF1-GFP; n=5 each. (d) Δ133p53 expression in cultured HCASMCs infected with Ad-GFP or Ad-Δ133p53; n=5 each. (e) Δ133p53 levels in HCAECs infected with Ad-GFP or Ad-SRSF1; n=6 each. (f) Representative images and averaged data from wound-healing assays in HCASMCs infected with Ad-GFP, Ad-SRSF1 or Ad-Δ133p53 and treated with Ang II; n=9 each. Scale bar, 200 μm. (g,h) PCNA expression in cultured HCAECs infected with Ad-SRSF1 (g) or Ad-Δ133p53 (h); n=5 each. (i) MTT assays of HCASMCs infected with Ad-GFP, Ad-SRSF1 or Ad-Δ133p53 (m.o.i. 100, 48 h) after Ang II for 4 days; n=12 each. (j) KLF5, p21 and PCNA levels in HCASMCs infected with Δ133p53 siRNAs in the presence or absence of Ad-SRSF1 overexpression; n=7 each. (k) Δ133p53 levels in HCASMCs infected with Δ133p53 siRNAs (Δ133 si1 and Δ133 si2); n=7 each. (l) MTT assays showing the SRSF1-mediated enhancement of proliferation was inhibited by Δ133p53 knockdown in HCASMCs; n=15 each. (m) Averaged data from wound-healing assays in HCASMCs infected with Δ133p53 siRNAs with or without Ad-SRSF1 overexpression and treated with Ang II; n=5 each. (n) Schematic outline of SRSF1 -D133p53 -KLF5 axis in VSMC. Ang II, angiotensin II; SRSF1, serine/arginine-rich splicing factor 1; EGR1, early-growth-response gene 1; KLF5, Krüppel-like factor 5; VSMC, vascular smooth muscle cell. Scr indicates scrambled siRNA control. All adenoviral infection above is 100 m.o.i. for 48 h. Concentration of Ang II is 200 nM. Scr indicates scrambled siRNA control. *P<0.05, **P<0.01, one-way ANOVA (a,b,e–m). Data are mean±s.e.m. of four independent experiments (a,b,e–m).