MicroRNA-144-3p targets relaxin/insulin-like family peptide receptor 1 (RXFP1) expression in lung fibroblasts from patients with idiopathic pulmonary fibrosis

Abstract

The hormone relaxin is considered a potential therapy for idiopathic pulmonary fibrosis (IPF). We have previously shown that a potential limitation to relaxin-based IPF therapy is decreased expression of a relaxin receptor, relaxin/insulin-like family peptide receptor 1 (RXFP1), in IPF fibroblasts. The mechanism that down-regulates RXFP1 in IPF remains unclear. To determine whether microRNAs (miRs) regulate RXFP1 gene expression, here we employed a bioinformatics approach to identify miRs predicted to target RXFP1 and identified a putative miR-144-3p target site in the RXFP1 mRNA. In situ hybridization of IPF lung biopsies revealed that miR-144-3p is expressed in fibroblastic foci. Furthermore, we found that miR-144-3p is up-regulated in IPF fibroblasts compared with lung fibroblasts from healthy donors. Transforming growth factor β increased miR-144-3p expression in both healthy and IPF lung fibroblasts in a SMAD family 2/3 (SMAD2/3)-dependent manner, and Jun proto-oncogene AP-1 transcription factor subunit (AP-1) was required for constitutive miR-144-3p expression. Overexpression of an miR-144-3p mimic significantly reduced RXFP1 mRNA and protein levels and increased expression of the myofibroblast marker α-smooth muscle actin (α-SMA) in healthy lung fibroblasts. IPF lung fibroblasts transfected with anti-miR-144-3p had increased RXFP1 expression and reduced α-SMA expression. Of note, a lentiviral luciferase reporter carrying the WT 3' UTR of RXFP1 was significantly repressed in IPF lung fibroblasts, whereas a reporter carrying a mutated miR-144-3p-binding site exhibited less sensitivity toward endogenous miR-144-3p expression, indicating that miR-144-3p down-regulates RXFP1 in IPF lung fibroblasts by targeting its 3' UTR. We conclude that miR-144-3p directly represses RXFP1 mRNA and protein expression.

Keywords: epigenetics; fibroblast; idiopathic pulmonary fibrosis; lung disease; lung fibroblasts; miR-144-3p; microRNA (miRNA); myofibroblast; post-transcriptional regulation; pulmonary fibrosis; relaxin; relaxin/insulin-like family peptide receptor 1 (RXFP1); transforming growth factor β (TGF-B).

Figure 1.

TGFβ is associated with decreased half-life of RXFP1 in human lung fibroblasts. GAP scores were calculated on the IPF patients in the LTRC dataset, and the patients were assigned a GAP stage from 1 to 3. Gene expression by microarray for RXFP1 (A), RXFP2 (B), and RXFP3 (C) were plotted as a function of increasing GAP stage. For GAP 1, n = 45; GAP 2, n = 75; and GAP 3, n = 14. Data were analyzed by Kruskal-Wallis, and p values are indicated in the panel. D, donor (n = 3), and E, IPF (n = 3) lung fibroblasts were processed for quantitative RT-PCR for RXFP1 at several time points following incubation with TGFβ and actinomycin D. Data represent the percentage of relative abundance of RXFP1 mRNA remaining compared with the 0-h time point, and the best fit curves for RXFP1 decay were plotted. F, dark line indicates data representing relative abundance of RXFP1 expression of RXFP1 mRNA remaining compared with the 0-h time point for donor fibroblasts. Dotted line indicates the percentage of relative abundance of RXFP1 expression in IPF fibroblasts at the indicated time points normalized to the levels of RXFP1 mRNA of donor fibroblasts at the 0-h time point. Data represent the relative abundance of miR-144-3p (ΔΔC_t) remaining compared with the 0-h time point, and the best fit curves for miR-144-3p decay were plotted.

Figure 2.

miR-144-3p targets RXFP1 in human lung fibroblasts. A, the seed region of hsa–miR-144-3p predicted to target the 3′ UTR of human RXFP1. B and C, donor and IPF lung fibroblasts were treated with miR-144-3p mimic or scrambled control. RNA was isolated and processed for qPCR for RXFP1 (B) or RXFP2 (C). Significantly higher basal levels of RXFP1 mRNA were detected in donor lung fibroblasts compared with IPF lung fibroblasts (p = 0.007, n = 3). Treatment of donor lung fibroblasts with the miR-144-3p significantly decreased RXFP1 expression (p = 0.0005, by two-way ANOVA and Tukey post hoc testing, n = 3). D, donor and IPF lung fibroblasts were transfected with 10 nm miR-144-3p mimic or control, and cells were lysed for immunoblotting for RXFP1, α-SMA, and β-actin. E, densitometry of immunoblotting presented in F. Data are expressed as mean ± S.D., normalized to β-actin. Less RXFP1 protein was detected in donor lung fibroblasts following transfection of miR-144-3p compared with scrambled control (p = 0.0016, n = 3). Significantly more RXFP1 was present in donor lung fibroblasts compared with IPF lung fibroblasts following transfection with the scrambled control (p = 0.0019, n = 3). Data were analyzed by two-way ANOVA and Tukey's post-hoc test. F, densitometry of α-SMA were normalized to β-actin. Transfection of miR-144-3p into donor lung fibroblasts significantly increased expression of α-SMA (p = 0.0117 for log-transformed data, n = 3, by two-way ANOVA followed by Tukey's post hoc testing).

Figure 3.

miR-144-3p targets RXFP1 in a dose-dependent manner. A, donor lung fibroblasts were transfected with increasing concentrations of miR-144-3p mimic (0.1, 1, 5, and 10 nm). miR-144-3p mimic decreased the expression of RXFP1 and up-regulated the expression of α-SMA in a dose-dependent manner. Densitometry of RXFP1 (B) and α-SMA (C) from A (data analyzed by one-way ANOVA, for RXFP1, p value for trend 0.0006, R² = 0.640) and for α-SMA (p value of log-transformed data for trend <0.02, R² = 0.51). D, top panel, prediction of the seed region of hsa–miR-144-3p WT and mismatch mutant targeting the 3′ UTR of human RXFP1. Bottom panel, representative image of gel contraction experiments of donor lung fibroblasts following infection with WT miR-144-3p, mismatch miR-144-3p, or GFP-expressing particles or incubation with TGFβ (n = 3). E, quantification of collagen gel areas was performed using ImageJ and plotted (p < 0.001, vehicle + miR-144-3p versus vehicle + GFP; p < 0.05, TGFβ + miR-144-3p versus TGFβ + GFP; ns, untreated GFP versus untreated mismatch group). Values are presented as mean ± S.D.

Figure 4.

Regulation of miR-144-3p expression. A, lung biopsies were obtained from IPF subjects (n = 5). In situ hybridization was performed as described employing probes directed against U6 (left), scrambled control (center), or miR-144-3p (right). Low power images are shown on the top. Yellow inset squares are magnified at the bottom. Green inset bar = 200 μm and red inset bar = 50 μm. B, donor and IPF lung fibroblasts were stimulated with and without TGFβ, and RNA was isolated for qPCR. Significantly higher levels (>70-fold) of miR-144-3p were detected in IPF lung fibroblasts compared with donor controls (p < 0.0001, n = 6). TGFβ stimulation increased expression of miR-144-3p in donor lung fibroblasts by ∼5-fold (p < 0.0001), whereas it increased by ∼1.4-fold (p < 0.0001) in IPF lung fibroblasts. Data were log-transformed and analyzed by two-way ANOVA followed by Tukey's post hoc test. C, donor lung fibroblasts were incubated with Smad2 and/or Smad3 siRNA and stimulated with TGFβ. TGFβ significantly increased miR-144-3p expression in the presence of the scrambled control (p < 0.0001, n = 6). Smad2 and/or Smad3 siRNA significantly decreased the effect of TGFβ on miR-144-3p expression (p < 0.005, n = 6). Data were analyzed by two-way ANOVA, followed by Tukey's post hoc test. D, IPF lung fibroblasts were incubated with Smad2 and/or Smad3 siRNA and stimulated with TGFβ. TGFβ significantly increased miR-144-3p expression in the presence of the scrambled control (p < 0.0001, n = 6). Smad2 and/or Smad3 siRNA significantly decreased the effect of TGFβ on miR-144-3p expression (p < 0.0001, n = 6). Data were analyzed by two-way ANOVA, followed by Tukey's post hoc test. E and F, quantitative RT-PCR for Smad2 (E) and Smad3 (F) for experiments in C and D. Data were analyzed by two-way ANOVA followed by Tukey's post hoc test.

Figure 5.

AP-1 complex is one of the key regulators of miR-144 locus. Quantitative RT-PCR for miR-144-3p following silencing of c-Jun and/or c-Fos in (A) donor or (B) IPF lung fibroblasts (data were analyzed by two-way ANOVA, followed by Tukey's post hoc test, n = 4). Experiments with donor and IPF lung fibroblasts were performed in parallel, and data from A and B were analyzed independently. C and D, quantitative RT-PCR for silencing of c-Jun (C) and c-Fos (D) from A and B, respectively. Data were analyzed by two-way ANOVA, followed by Tukey's post-hoc test, n = 4. E and F, quantitative RT-PCR for c-Jun (E) and c-Fos (F) between unstimulated donor and IPF lung fibroblasts. Data were analyzed by unpaired t test, n = 6.

Figure 6.

PMA and lentiviral overexpression of c-Jun induced expression of miR-144-3p. A, quantitative RT-PCR for miR-144-3p following incubation of donor lung fibroblasts with PMA, an activator of the AP-1 transcription factor (p < 0. 05, paired two-tailed t test, n = 6). B and C, quantitative RT-PCR for miR-144-3p (B) and miR-186-5p (C) following infection of donor lung fibroblasts with a lentiviral vector expressing GFP (control) or c-Jun (data were analyzed by paired t test, n = 5).

Figure 7.

Anti-miR-144-3p reversed the suppression of RXFP1 in IPF lung fibroblasts. A and B, quantitative RT-PCR for miR-144-3p (A) or RXFP1 (B) from donor lung fibroblasts after transfection with a miR-144-3p antagonist (anti-miR-144-3p) and stimulation with TGFβ (p = 0.0001 for log-transformed data for miR-144-3p and RXFP1, by two-way ANOVA followed by Tukey's multiple comparisons test, n = 6). C, immunoblots for RXFP1 from donor and IPF lung fibroblasts following transfection with anti-miR-144-3p or scrambled controls (n = 3). D, densitometry of band intensity of RXFP1 normalized to β-actin from donor (E) and IPF (F) lung fibroblasts (p = ns for donor; p < 0.0001 for IPF, n = 3, data were analyzed by paired t test). E, Bio-Plex 200 multiplex analysis of bronchoalveolar lavage for Relaxin (n = 18 healthy controls and n = 126 IPF patients, p = not significant, Kolmogorov-Smirnov test).

Figure 8.

Relaxin reverses contractile phenotypes associated with miR-144-3p. A, transfection of donor lung fibroblasts with miR-144-3p mimic to decrease RXFP1 levels followed by incubation with low and high concentrations of relaxin. Right panel, transfection of IPF lung fibroblasts with anti-miR-144-3p followed by incubation with low and high concentrations of relaxin. Cells were lysed for immunoblotting for RXFP1, phospho-MLC20, total MLC20, α-SMA, and β-Actin (n = 3). B-E, densitometry of immunoblotting presented in A. Data were expressed as mean ± S.D., normalized to β-Actin or total MLC20. B, between mimic and no mimic, significant effects (p < 0.05) were observed for vehicle, relaxin low, and relaxin high in donor lung fibroblasts. More α-SMA protein was detected in donor lung fibroblasts following transfection of miR-144-3p compared with scrambled control. C, in IPF lung fibroblasts, no significant effects were observed with either low or high dose relaxin between mimic or no mimic groups. D, significant effects between mimic and no mimic were observed only with high dose of relaxin (p < 0.05). Significantly more phospho-MLC20 was present in IPF lung fibroblasts compared with donor lung fibroblasts at baseline. E, significant effects were observed only for relaxin high dose in IPF lung fibroblasts following transfection with antagomir (p < 0.05; n = 3). Data were analyzed by two-way ANOVA followed by the Holm-Sidak test.

Figure 9.

miR-144-3p directly targets the 3′ UTR of RXFP1 mRNA. A, depiction of pmirGLO dual-luciferase reporter construct for the WT 3′ UTR of the RXFP1 seed region and mutations. B and C, 293T cells were plated 24 h prior to transfection. Cells were then co-transfected with 50 ng of either pmirGLO vector carrying WT (B) 3′ UTR of RXFP1 or mutated (C) 3′ UTR of RXFP1 with and without 100 nm miR-144-3p mimic or control mimic. miR-144-3p significantly repressed the luciferase activity of the reporter containing the WT 3′ UTR of RXFP1 (p < 0.005, unpaired t test, mean ± S.D., n = 3), whereas the mutated construct was insensitive in 293T cells (I, unpaired t test, mean ± S.D., n = 3). Luciferase activity was measured using Renilla luciferase as an internal control. D, lentiviral luciferase reporter construct carrying WT 3′ UTR of RXFP1 was used to transduce both donor and IPF lung fibroblasts. There was a reduction in luciferase activity with WT 3′ UTR of RXFP1 compared with mutated 3′ UTR of RXFP1. Also, the luciferase activity of WT 3′ UTR RXFP1 was highly suppressed in IPF lung fibroblasts compared with donor lung fibroblasts indicating higher levels of endogenous miR-144-3p in IPF lung fibroblasts compared with that of donor. Lentiviral Renilla luciferase vector was used as an internal control. Data shown are from a representative experiment (n = 3). E, depiction of mmu–miR-144-3p mimic and its modified version to enhance base pairing with mouse RXFP1 3′ UTR. There is a single nucleotide mismatch in mmu–miR-144-3p with mouse RXFP1 3′ UTR target region. E, primary MLF were isolated from C57/B6 mice (n = 2; repeated twice). MLF were transfected either with a mmu–miR-144-3p mimic or a modified mmus–miR-144-3p mimic (modified to match mouse RXFP1 target region) along with a negative control mimic. MLF were harvested to isolate total RNA and qRT-PCR was performed to determine the levels of mRXFP1, mPPIA, and mNFE2L2 (mNrf2).

Figure 10.

miR-144-3p controls fibrotic gene expression in lung fibroblasts. Donor and IPF lung fibroblasts (n = 4) were treated with miR-144-3p mimic or scrambled control and anti-miR-144-3p mimic or negative control inhibitor, respectively. RNA was isolated and processed for qPCR for (A, A′), COL1A2 (B, B′), COL3A1 (C, C′), ACTA2 (D, D′), and VCAN (E, E′) FN1. Transfection of mimic resulted in significantly higher basal levels of COL1A2, COL3A1, ACTA2, VCAN, and FN1 mRNA in donor lung fibroblasts compared with scrambled controls, whereas transfection of antagomir resulted in significantly lower basal levels of COL1A2, COL3A1, ACTA2, VCAN, and FN1 in IPF lung fibroblasts compared with a negative control inhibitor (paired, two-tailed Student's t test, n = 4).