Epigenetic regulation of miR-17~92 contributes to the pathogenesis of pulmonary fibrosis

Abstract

Rationale: Idiopathic pulmonary fibrosis (IPF) is a disease of progressive lung fibrosis with a high mortality rate. In organ repair and remodeling, epigenetic events are important. MicroRNAs (miRNAs) regulate gene expression post-transcriptionally and can target epigenetic molecules important in DNA methylation. The miR-17~92 miRNA cluster is critical for lung development and lung epithelial cell homeostasis and is predicted to target fibrotic genes and DNA methyltransferase (DNMT)-1 expression.

Objectives: We investigated the miR-17~92 cluster expression and its role in regulating DNA methylation events in IPF lung tissue.

Methods: Expression and DNA methylation patterns of miR-17~92 were determined in human IPF lung tissue and fibroblasts and fibrotic mouse lung tissue. The relationship between the miR-17~92 cluster and DNMT-1 expression was examined in vitro. Using a murine model of pulmonary fibrosis, we examined the therapeutic potential of the demethylating agent, 5'-aza-2'-deoxycytidine.

Measurements and main results: Compared with control samples, miR-17~92 expression was reduced in lung biopsies and lung fibroblasts from patients with IPF, whereas DNMT-1 expression and methylation of the miR-17~92 promoter was increased. Several miRNAs from the miR-17~92 cluster targeted DNMT-1 expression resulting in a negative feedback loop. Similarly, miR-17~92 expression was reduced in the lungs of bleomycin-treated mice. Treatment with 5'-aza-2'-deoxycytidine in a murine bleomycin-induced pulmonary fibrosis model reduced fibrotic gene and DNMT-1 expression, enhanced miR-17~92 cluster expression, and attenuated pulmonary fibrosis.

Conclusions: This study provides insight into the pathobiology of IPF and identifies a novel epigenetic feedback loop between miR-17~92 and DNMT-1 in lung fibrosis.

Figure 1.

Decreased expression of the miR-17∼92 cluster in human idiopathic pulmonary fibrosis. (A) Expression of each microRNA (miRNA) from the miR-17∼92 cluster was determined by quantitative real-time polymerase chain reaction from control (n = 10), >80% FVC (n = 7), 50–80% FVC (n = 8), and <50% FVC (n = 9) lung tissue samples. Data were normalized to miR-191. Relative copy number (RCN, 2^−dCT × 100) was determined. Data are expressed as the average RCN ± SD, *P < 0.018 compared with control tissue. Comparison of the mild disease (80% FVC) with control tissue for miR-19b and -92, P = 0.1325 and 0.1320, respectively. (B) In situ hybridization was performed using LNA-modified DNA probes for let-7c (positive control) and miR-19b and -20a at a magnification of ×400. Scrambled probes were used as a negative control. The arrows denote positively stained cells (dark blue/purple). Shown are representative images (n = 3 per group).

Figure 2.

Introduction of miR-17∼92 in idiopathic pulmonary fibrosis (IPF) lung fibroblasts induces phenotypic and molecular changes. Normal and IPF lung fibroblasts were transfected with an expression vector containing the miR-17∼92 cluster. Empty vector-transfected cells (vector) also served as a negative control. (A) After 24 hours, cells were stained with phalloidin and DAPI 0.5 μg/ml then photographed using an Olympus inverted fluorescent at ×10 magnification. Shown are representative images from three independent experiments. (B) To quantity actin fibers, the red fluorescence of phalloidin was quantitated and indicated as a percentage compared with cell number as enumerated by DAPI-positive nuclei. Shown is the average intensity ± SD. Significant decrease was apparent in miR-17∼92 transfected cells compared with vector-only transfected IPF cells, *P = 0.0327; differences in phalloidin staining for IPF vector-transfected IPF cells compared with untransfected (P = 0.2549) or vector-transfected (P = 0.3534) normal lung fibroblasts. (C) RNA was isolated and quantitative real-time polymerase chain reaction was performed for the indicated genes. Data were normalized using glyceraldehyde phosphate dehydrogenase as a housekeeping control. Data are expressed as the average relative copy number (RCN) ± SD from three experiments. Comparison was made between vector-only transfected IPF cells to the miR-17∼92 transfected IPF cells, *P < 0.004. COL = collagen; CTGF = connective tissue growth factor; VEGF = vascular endothelial growth factor.

Figure 3.

Increased DNA methylation of the miR-17∼92 promoter in idiopathic pulmonary fibrosis (IPF). (A) DNA was isolated from control lung tissue (n = 3) and individuals with IPF (n = 3 per severity group). Samples were selected from the cohort used in Figure 1. Data are presented as the average percent of unmethylated or methylated DNA of the miR-17∼92 promoter ± SD. Statistical difference was determined by comparing the % unmethylated with % methylated for each severity group, *P < 0.003. (B) DNA was isolated from normal or IPF fibroblasts. DNA methylation for the miR-17∼92 promoter was determined. Shown is the percent DNA methylation, n = 6 (*P < 0.0001 normal fibroblast compared with IPF fibroblast for unmethylated or methylated DNA).

Figure 4.

Treatment of idiopathic pulmonary fibrosis (IPF) lung fibroblasts with 5′-aza-2′-deoxycytidine liberates microRNA (miRNA) expression and down-regulates target mRNAs. IPF lung fibroblast cell line was either treated with the vehicle control dimethyl sulfoxide (DMSO) or treated with 0.5 μM 5′-aza-2′-deoxycytidine (5′-Aza) for 24 hours. RNA was isolated and subjected to quantitative real-time polymerase chain reaction analysis for (A) miRNA or (B) fibrotic genes. Data were normalized to (A) RNU48 or (B) CAP-1 expression. Shown is the average relative copy number (RCN) ± SD from two independent experiments. COL = collagen; CTGF = connective tissue growth factor; VEGF = vascular endothelial growth factor.

Figure 5.

DNA methyltransferase (DNMT)-1 is altered in idiopathic pulmonary fibrosis (IPF) and a target of the miR-17∼92 cluster. (A) DNMT-1 expression was examined by quantitative real-time polymerase chain reaction from a select cohort of tissue samples shown in Figure 1 (n = 3, per each disease severity group and control tissue samples). Using glyceraldehyde phosphate dehydrogenase as an endogenous control, the average relative copy number (RCN) ± SD was calculated. Significance *P < 0.05 is shown. (B) IPF lung fibroblasts cells were transfected with or without miR-19b or 20a antagomir to knock down (KD) their expression. After treatment with 5′-aza-2′-deoxycytidine (5′ Aza) for 24 hours, RNA was extracted and DNMT-1 expression quantitated by quantitative real-time polymerase chain reaction. DNMT-1 expression was normalized to glyceraldehyde phosphate dehydrogenase in vehicle-treated untransfected cells. Shown is the average fold-change ± SD (n = 3). Significance for 5′ Aza–treated cells compared with 5′Αza/miR-19b KD cells is *P < 0.001. No clinical difference was apparent between 5′ Αza–treated cells and 5′ Αza/miR-20a KD cells. (C) The wild-type (WT) or mutated (mut) 3′ untranslated region for DNMT-1 was cloned into the pGL-3 Firefly luciferase vector and transfected in the presence of the indicated microRNAs (miRNAs) in HEK 293 cells. Irrelevant scrambled miRNA served as a control. Cells were also cotransfected with the Renilla luciferase construct, pRL-TK. Luciferase production was measured first for the Firefly luciferase followed by the Renilla luciferase from the culture supernatant after 24 hours. Firefly luciferase was normalized to the Renilla luciferase. Shown is the average luciferase production from eight independent experiments (± SD). WT DNMT-1 was compared with the mutated DNMT-1 for each specified miRNA, *P < 0.001.

Figure 6.

In vivo treatment of mice with 5′-aza-2′-deoxycytidine attenuates bleomycin-induced fibrotic gene expression. Mice were treated with 0.035 U/kg twice weekly for 4 weeks with bleomycin (n = 4) or phosphate-buffered saline (PBS) alone (n = 4). A set of bleomycin-treated mice were injected with 0.156 mg/kg/wk of 5′-aza-2′-deoxycytidine intraperitoneally for the last 2 weeks of bleomycin treatment (n = 4) and designated as Bleomycin (4 wk) 5′ Aza (2 wk). As a control, mice were treated with PBS alone for 4 weeks or PBS for 2 weeks and then 5′-aza-2′-deoxycytidine for 2 weeks without bleomycin (5′Aza [2 wk]). (A) RNA was extracted from the lung tissue and microRNA (miRNA) expression was examined by quantitative real-time polymerase chain reaction and normalized using snoRNA-202 expression. Shown is the average relative copy number (RCN) ± SD, *P value < 0.001 compared with bleomycin only treated mice. (B) Lung tissue was section and paraffin-embedded then subjected to trichrome staining to detect collagen deposition. The larger image is at ×10 magnification, whereas the inset is a ×4 magnification. Shown are representative images from a mouse from each group. (C) Trichrome sections were blindly assessed by a board-certified pathologist. The average arbitrary score ± SD is shown (n = 4 per group). Significance compared with bleomycin-only treated mice, *P < 0.0075 and **P = 0.1348. (D) RNA was subjected to quantitative real-time polymerase chain reaction for the indicated genes and (E) DNA methyltransferase (DNMT)-1 expression. CAP-1 served as an endogenous control for normalization for fibrotic genes and DNMT-1. Data are expressed as the relative copy number (RCN) ± SD, *P value < 0.001 compared with bleomycin-treated mice. (F) DNA were isolated and subjected to analysis for promoter DNA methylation of the miR-17∼92 cluster. Data presented is the average % unmethylated and % methylated, n = 3 mice per treatment group. *P value < 0.001 compared with bleomycin-only treated mice. COL = collagen; CTGF = connective tissue growth factor; VEGF = vascular endothelial growth factor.