miR-483 Targeting of CTGF Suppresses Endothelial-to-Mesenchymal Transition: Therapeutic Implications in Kawasaki Disease

Abstract

Rationale: Endothelial-mesenchymal transition (EndoMT) is implicated in myofibroblast-like cell-mediated damage to the coronary arterial wall in acute Kawasaki disease (KD) patients, as evidenced by positive staining for connective tissue growth factor (CTGF) and EndoMT markers in KD autopsy tissues. However, little is known about the molecular basis of EndoMT involved in KD.

Objective: We investigated the microRNA (miRNA) regulation of CTGF and the consequent EndoMT in KD pathogenesis. As well, the modulation of this process by statin therapy was studied.

Methods and results: Sera from healthy children and KD subjects were incubated with human umbilical vein endothelial cells. Cardiovascular disease-related miRNAs, CTGF, and EndoMT markers were quantified using reverse transcriptase quantitative polymerase chain reaction, ELISA, and Western blotting. Compared with healthy controls, human umbilical vein endothelial cell incubated with sera from acute KD patients had decreased miR-483, increased CTGF, and increased EndoMT markers. Bioinformatics analysis followed by functional validation demonstrated that Krüppel-like factor 4 (KLF4) transactivates miR-483, which in turn targets the 3' untranslated region of CTGF mRNA. Overexpression of KLF4 or pre-miR-483 suppressed, whereas knockdown of KLF4 or anti-miR-483 enhanced, CTGF expression in endothelial cells in vitro and in vivo. Furthermore, atorvastatin, currently being tested in a phase I/IIa clinical trial in KD children, induced KLF4-miR-483, which suppressed CTGF and EndoMT in endothelial cells.

Conclusions: KD sera suppress the KLF4-miR-483 axis in endothelial cells, leading to increased expression of CTGF and induction of EndoMT. This detrimental process in the endothelium may contribute to coronary artery abnormalities in KD patients. Statin therapy may benefit acute KD patients, in part, through the restoration of KLF4-miR-483 expression.

Clinical trial registration: URL: http://www.clinicaltrials.gov. Unique identifier: NCT01431105.

Keywords: CTGF; EndoMT; Kawasaki disease; atorvastatin; endothelial dysfunction; miR-483; microRNA.

© 2016 American Heart Association, Inc.

Figures

Figure 1. Elevated CTGF in KD sera and sera-treated ECs

(A) Circulating levels of CTGF were quantified by ELISA in age-similar healthy controls (HC, n=6), KD subjects (n=20). (B, C) HUVECs were treated with medium containing 15% sera from HC (n=6) and KD (n=28) subjects. After 96 hr, CTGF mRNA levels were measured by qPCR (B, all samples were normalized to HUVECs without patient sera treatment) and protein levels in ECs were measured by Western blot (C, HC, n=6 and KD, n=12). The right panel of (C) is the statistical analysis of immunoblots. (D, E) HUVECs were treated with TNFα (100 ng/mL), IL-6 (100 ng/mL), or left untreated (Ctrl) for 72 hr. CTGF mRNA and protein levels were measured by qPCR and Western blotting, respectively. The bar graphs are mean±SEM. *P<0.05 compared with respective control group.

Figure 2. CTGF induction by KD sera is inversely associated with KLF4

(A, B) HUVECs were treated with individual KD sera (n=28) for 96 hr as described in Fig. 1B. KLF4 mRNA and protein were measured by qPCR and Western blotting, respectively. (C, D) HUVECs were incubated with HC sera pooled from 5 individuals, and transfected with control siRNA (Ctrl siRNA) or KLF4 siRNA for 48 hr. (E, F) HUVECs were treated with HC or KD sera (pooled from 8 patients) for 24 hr, followed by infection with Ad-null or Ad-KLF4 for an additional 48 hr. Levels of CTGF and KLF4 mRNA were measured using qPCR (in C, E) and protein using Western blotting (in D, F). Data in C-F are mean±SEM from 3 independent experiments involving different batches of cells but the same pooled HC or KD sera. *P<0.05 when compared with respective controls or between indicated groups.

Figure 3. KLF4 transactivates miR-483

(A) Bioinformatics prediction of KLF4 binding sites in the promoter region of the IGF2-miR-483 and putative miR-483 targeting sites in the 3′UTR of human CTGF mRNA. (B) HUVECs were treated with patients’ sera as described in Fig. 1B, and (C) CD31+ MPs were isolated from sera from KD patients (n=10) and age-matched HC (n=5). All samples were normalized to the subject with the highest miR-483 expression. (B, C) The miR-483 level was detected by qPCR. (D, E) HUVECs were infected with Ad-null or Ad-KLF4 for 72 hr. In (D), ChIP assay was performed using anti-KLF4 and HUVEC extracts. The enrichment of KLF4 binding to the putative binding sites in the promoter region of IGF2-miR-483 was quantified by qPCR. In (E), the mRNA levels of KLF4, IGF2, miR-483, and CTGF were detected by qPCR. (F) Lung ECs were isolated from EC-KLF4−/− or EC-KLF4-Tg mice (n=3 per group), mRNA and miR-483 levels were quantified by qPCR. Data in (D, E) are mean±SEM from 3 independent experiments, and comparisons were made with control groups arbitrarily set as 1. *P<0.05 when compared between indicated groups.

Figure 4. miR-483 targets CTGF in ECs

(A–C) HUVECs were transfected with miR-483 mimic (pre-483) or inhibitor (anti-483) for 48 hr. In (C), HUVECs were pre-treated with or without pooled HC sera or infected with or without Ad-KLF4 for 48 hr. CTGF mRNA and protein were detected and quantified. (D) HUVECs were transfected with pre-483 or pooled HC sera with anti-483 (HC+Anti-483) for 48 hr followed by Ago1 immunoprecipitation. Ago1-associated miR-483 and CTGF mRNA were quantified by qPCR. (EG) BAECs transfected with Luc-CTGF (WT) or Luc-CTGF (MT1&2, 3) were co-transfected with pre-483 in (E), pooled HC sera with anti-483, or treated with pooled KD sera as indicated in (G). Luciferase activity was measured with that of Renilla as transfection control. (H, I) HUVECs were treated with pooled sera from HC or KD patients for 48 hr before transfection with pre-control or pre-483. CTGF expression was determined by qPCR and Western blot. Controls for the miRNA mimic (Pre-Ctrl) or miRNA inhibitor (Anti-Ctrl) were used. Data are mean±SEM from 3 independent experiments, and comparisons were made with control groups arbitrarily set as 1. *P<0.05 when compared between indicated groups.

Figure 5. Atorvastatin decreases CTGF expression in ECs by activating the KLF4-miR-483 axis

(A, B) HUVECs were treated with pooled KD sera for 48 hr followed by atorvastatin treatment with indicated dosage for an additional 48 hr. The changes in miR-483, CTGF, and KLF4 were determined by qPCR and Western blotting. (C) BAECs were transfected with Luc-CTGF (WT) or Luc-CTGF (MT1,2, and MT3) before atorvastatin treatment (5 μM) for 48 hr. Luciferase activity was measured. (D) Comparison of circulating levels of CTGF and miR-483 in KD patients between acute phase (pre-treatment) and 6 weeks post-treatment with standard treatment (IVIG+ASA+infliximab) with or without atorvastatin (0.125–0.25 mg/kg/day, n=5/group). Control subjects were matched for age, illness day and Z-worst (Supplemental Table II). (E, F) HUVECs were treated with pooled patients’ sera from acute phase (pre-treatment) and 6 weeks post-treatment. The changes in miR-483, CTGF, and KLF4 levels were determined by qPCR and Western blot. Data are mean±SEM from 3 independent experiments (A-C) or from 5 subjects per group (D–F). *P<0.05 compared between the indicated group.

Figure 6. Atorvastatin and miR-483 alleviate KD sera-induced EndoMT

Expression levels of EndoMT markers were quantified by qPCR. In (A) HUVECs were treated with individual sera from HC (n=6) or KD (n=29) patients as described in Fig. 1B. (B) HUVECs were treated with HC sera (pool from 5 subject) for 48 hr and transfected with Ctrl siRNA or KLF4 siRNA for 48 hr. In (C, D) HUVECs were treated with KD sera (pooled from 8 patients) for 48 hr before transfection with pre-Ctrl, pre-483, or atorvastatin for an additional 48 hr. Data are mean±SEM from 3 independent experiments. *P<0.05 compared with the control or indicated group.

Figure 7. The EC-derived miR-483 inversely correlates with CAA in KD subjects and a schematic illustration of KD- and statin-modulated KLF4-miR-483-CTGF pathway in EndoMT

(A, B) miR-483 expression was determined by qPCR in (A) HUVECs treated with sera from HC (n=6), KD CAA− (n=10), and KD CAA+ (n=18) patients as described in Fig. 1B and in (B) respective patient sera (all samples were normalized to the HC subject with the lowest miR-483 expression). *P<0.05 when compared between indicated groups. (C) Spearman’s correlation between miR-483 level in CD31+ MPs and Z-score. Data were from 10 acute patients (closed triangles) and 5 convalescent patients (open triangles). (D) Schematic illustration of KD- and statin-modulated KLF4-miR-483-CTGF pathway in EndoMT. In healthy endothelial cells, KLF4 binds to the promoter region of miR-483 and transcriptionally induces miR-483. miR-483 in turns binds to the 3′UTR of CTGF mRNA to inhibit CTGF expression. In KD, the KLF4-miR-483 axis is suppressed in ECs and CD31+-MPs have decreased miR-483 enrichment. As a result, CTGF is de-suppressed, thereby inducing EndoMT (the transition of endothelial cells into myofibroblast-like cells) and contributing to coronary artery abnormalities. Such pathological modulation may be reversed by statin therapy, which restores the KLF4-miR-483 axis.