Drs. Maurer, S. Gay, and O. Distler have filed an application for a patent for the use of miR-29 as a novel pharmacologic treatment of scleroderma.
MicroRNA-29, a key regulator of collagen expression in systemic sclerosis
Article first published online: 3 MAR 2010
Copyright © 2010 by the American College of Rheumatology
Arthritis & Rheumatism
Volume 62, Issue 6, pages 1733–1743, June 2010
How to Cite
Maurer, B., Stanczyk, J., Jüngel, A., Akhmetshina, A., Trenkmann, M., Brock, M., Kowal-Bielecka, O., Gay, R. E., Michel, B. A., Distler, J. H. W., Gay, S. and Distler, O. (2010), MicroRNA-29, a key regulator of collagen expression in systemic sclerosis. Arthritis & Rheumatism, 62: 1733–1743. doi: 10.1002/art.27443
- Issue published online: 28 MAY 2010
- Article first published online: 3 MAR 2010
- Accepted manuscript online: 3 MAR 2010 12:00AM EST
- Manuscript Accepted: 23 FEB 2010
- Manuscript Received: 30 MAR 2009
- Olga Mayenfisch Foundation
- Foundation for the Medical Research University of Zurich
- Swiss National Science Foundation
- European Union Sixth Framework Programme (project AutoCure)
- Seventh Framework Programme (project Masterswitch)
- Zurich Center of Integrative Human Physiology
- DFG. Grant Number: DI 1537/2-1
- Interdisciplinary Center of Clinical Research in Erlangen. Grant Number: A20
- Career Support Award of Medicine from the Ernst Jung Foundation
- EULAR Orphan Disease Programme grant
- Top of page
- PATIENTS AND METHODS
- AUTHOR CONTRIBUTIONS
To investigate the role of microRNA (miRNA) as posttranscriptional regulators of profibrotic genes in systemic sclerosis (SSc).
MicroRNA, which target collagens, were identified by in silico analysis. Expression of miRNA-29 (miR-29) was determined by TaqMan real-time polymerase chain reaction analysis of skin biopsy and fibroblast samples from SSc patients and healthy controls as well as in the mouse model of bleomycin-induced skin fibrosis. Cells were transfected with precursor miRNA (pre-miRNA)/anti-miRNA of miR-29 using Lipofectamine. Collagen gene expression was also studied in luciferase reporter gene assays. For stimulation, recombinant transforming growth factor β (TGFβ), platelet-derived growth factor B (PDGF-B), or interleukin-4 (IL-4) was used. The effects of inhibiting PDGF-B and TGFβ signaling on the levels of miR-29 were studied in vitro and in the bleomycin model.
We found that miR-29a was strongly down-regulated in SSc fibroblasts and skin sections as compared with the healthy controls. Overexpression in SSc fibroblasts significantly decreased, and accordingly, knockdown in normal fibroblasts increased, the levels of messenger RNA and protein for type I and type III collagen. In the reporter gene assay, cotransfection with pre-miR-29a significantly decreased the relative luciferase activity, which suggests a direct regulation of collagen by miR-29a. TGFβ, PDGF-B, or IL-4 reduced the levels of miR-29a in normal fibroblasts to those seen in SSc fibroblasts. Similar to human SSc, the expression of miR-29a was reduced in the bleomycin model of skin fibrosis. Inhibition of PDGF-B and TGFβ pathways by treatment with imatinib restored the levels of miR-29a in vitro and in the bleomycin model in vivo.
These data add the posttranscriptional regulation of collagens by miR-29a as a novel aspect to the fibrogenesis of SSc and suggest miR-29a as a potential therapeutic target.
Systemic sclerosis (SSc) is a multisystemic fibrotic disorder with high morbidity and mortality rates (1). The progressive replacement of normal tissue by collagen-rich extracellular matrix leads to impairment and, ultimately, to failure of affected organs. Fibroblasts, the cellular key players of this process, are activated by profibrotic cytokines and growth factors, such as interleukin-4 (IL-4), transforming growth factor β (TGFβ), and platelet-derived growth factor B (PDGF-B). Despite the ongoing research in identifying stimuli that contribute to the continuous activation of fibroblasts in SSc, the intracellular molecular events that induce and sustain the status of activation have yet to be elucidated.
MicroRNA (miRNA) have recently been discovered to be a new class of posttranscriptional repressors of gene expression (2). Up to date, more than 700 of these small, noncoding, evolutionarily conserved RNAs have been experimentally validated in humans. In silico analyses estimate that miRNA account for 2–3% of the human genome (3). The term miRNA refers to the mature molecule of 19–25 nucleotides in length, which is generated in a multistep process in the nucleus and in the cytoplasm. The miRNA mediate the posttranscriptional regulation of gene expression by binding to partially complementary sites in the 3′-untranslated region (3′-UTR) of target messenger RNA (mRNA). The formation of miRNA–mRNA duplexes leads to mRNA degradation or translational repression (4).
Since miRNA are predicted to regulate ∼30% of the genes in humans (5), dysregulation of miRNA seems a likely scenario for the development of diseases such as cancer (6) or neurologic (7), cardiovascular (8), or autoimmune (9) disorders.
There is rapidly growing evidence from in vitro studies and studies of animal models that the modification of altered miRNA levels with chemically modified anti-miR or miRNA mimics may provide a novel, specific, molecular-based therapy (10–13).
The role of miRNA in the pathogenesis of SSc has not been addressed so far. Thus, the objectives of our present study were to identify differentially expressed miRNA in patients with SSc, to investigate the regulation of selected miRNA by profibrotic mediators, and to characterize the therapeutic effects of miRNA modulation on collagen production.
PATIENTS AND METHODS
- Top of page
- PATIENTS AND METHODS
- AUTHOR CONTRIBUTIONS
Patients, biopsy specimens, and cell culture.
Skin biopsy specimens were obtained by punch biopsy from 12 SSc patients and 11 healthy donors. Skin fibroblasts were expanded by outgrowth culture in Dulbecco's modified Eagle's medium (DMEM; Gibco Invitrogen) as described previously (14). Cells from passages 3–8 were used for the experiments. All patients fulfilled the criteria for SSc as described by LeRoy et al (15), and the study was approved by the local ethics committee. (Detailed characteristics of the 12 SSc patients, as well as additional information concerning the methodology used in these studies, are available upon request from the authors.)
Reagents and stimulation assays.
After 24 hours in serum-reduced DMEM (0.5% fetal calf serum), cells were stimulated for 24 and 48 hours with TGFβ (10 ng/ml), PDGF-B (40 ng/ml), or IL-4 (10 ng/ml) (R&D Systems). In subsets of experiments, dermal fibroblasts were incubated for 24 hours with imatinib mesylate that had been dissolved in 0.9% NaCl to a final concentration of 1.0 μg/ml (stock concentration 10 mg/ml; Novartis) (14). Myofibroblast differentiation was induced by stimulating SSc fibroblasts with TGFβ (2 ng/ml) for 4 days (16).
Cells were transfected with synthetic precursor miRNA (pre-miR), with inhibitors of miR-29 (anti-miR), or with negative controls (Pre-miR/Anti-miR Negative Control #1; Ambion/Applied Biosystems) at a final concentration of 100 nM with the use of Lipofectamine 2000 reagent (Invitrogen). At 72 hours after transfection, cellular lysates were collected to analyze the expression of types I and III collagen. Transfection efficiency was controlled by TaqMan-based real-time polymerase chain reaction (PCR). Additionally, normal skin fibroblasts were treated with TGFβ, PDGF-B, or IL-4 for 48 hours after overnight serum starvation, then transfected with pre-miR-29a or negative control, and lysed after 24 hours to measure COL1A1 and COL3A1 mRNA.
In another subset of experiments, SSc skin fibroblasts were transfected with small interfering RNA (siRNA) against c-abl, PDGF receptor β (PDGFRβ), and c-kit (Qiagen) at a final concentration of 50 nM and lysed after 24 hours to examine the expression of miR-29.
Computational prediction analysis of miRNA targeting collagen genes.
RNA isolation and quantitative real-time PCR analysis of miRNA.
Skin biopsy specimens (0.5 cm2) were homogenized with TissueLyser (Qiagen). A mirVana miRNA Isolation kit and a RecoverAll Total Nucleic Acid Isolation kit were used for isolation of total RNA (Ambion/Applied Biosystems). Specific single TaqMan miRNA assays (Ambion/Applied Biosystems) were used to measure the expression levels of selected miRNA in a model 7500 real-time PCR system analyzer (Applied Biosystems). Expression of the U6B small nuclear RNA (RNU6B) or snoRNA-202 was used as endogenous control to normalize the data for human and murine samples, respectively. For relative quantification, the comparative threshold cycle (Ct) method was used (18). To measure the expression of collagen genes, total RNA was isolated using an RNeasy Mini kit (Qiagen). Gene expression was quantified using SYBR Green real-time PCR, as previously described (14).
Western blot analysis for type Iα2 and type III collagen.
Equivalent amounts of protein from each cell lysate were separated by sodium dodecyl sulfate–10% polyacrylamide gel electrophoresis and electrotransferred onto nitrocellulose membranes, as described previously (14). Polyclonal antibodies against type Iα2 and type III collagen (Santa Cruz Biotechnology) and horseradish peroxidase–conjugated goat anti-rabbit secondary antibodies (Jackson ImmunoResearch) were used. Semiquantitative analysis based on densitometry was performed using AlphaImager software.
Luciferase reporter assay for targeting the 3′-UTR of COL3A1.
To validate the idea that miR-29a/29b/29c directly regulate the expression of type III collagen by binding to the complementary seed matches within the 3′-UTR of collagen genes, we performed a luciferase reporter gene assay. The human embryonic kidney cell line HEK 293 was used in these experiments, and the results were confirmed with primary dermal fibroblasts. (Data available upon request from the authors.) For the experiment, a COL3A1 3′-UTR segment of 973 bp that contained 2 conserved binding sites for miR-29 was amplified by PCR from human genomic DNA and inserted into the pGL3 control vector (Promega) by using the Xba I site immediately downstream of the stop codon of the luciferase. HEK 293 cells were transfected with the pGL3 control 3′-UTR of the COL3A1 construct (0.15 μg/well) using Lipofectamine (Invitrogen). For normalization, a control vector containing Renilla luciferase (pRL-SV40 vector, 0.8 μg/well; Promega) was added. For each well, a 100 nM concentration of pre-miR-29a, pre-miR-29b, pre-miR-29c, or scrambled control (Ambion/Applied Biosystems) was used.
Firefly and Renilla luciferase activities were measured in the cell lysates by the Dual-Luciferase Reporter Assay system (Promega) 24 hours after transfection. For the 3′-UTR of COL1A1 and COL1A2, conserved seed matches for miR-29 were likewise predicted (19).
Bleomycin-induced skin fibrosis in mice.
Skin fibrosis was induced in 6-week-old female DBA mice by local intracutaneous injection of bleomycin as described previously (14, 20, 21). The animal protocol was approved by the local animal care and use committee. Briefly, the mice were challenged every other day for 6 weeks with intracutaneous injections of bleomycin (100 μl dissolved in 0.9% NaCl of a 0.5 mg/ml concentration) in defined areas of 1.5 cm2 on the upper back. One group of mice was treated with intraperitoneal injections of imatinib 150 mg/kg/day for the last 3 weeks to assess the impact of imatinib on miR-29 levels in established fibrosis. Another group was treated with 0.9% NaCl instead of imatinib for the last 3 weeks. The respective control group for bleomycin exposure received intracutaneous injections of 100 μl of 0.9% NaCl for 6 weeks instead of bleomycin and additionally received 0.9% NaCl for the last 3 weeks. After 6 weeks, animals were asphyxiated with CO2. The skin was removed and processed for further analysis.
GraphPad Prism software was used for statistical analyses. Normal distribution of the data was confirmed using the Kolmogorov-Smirnov test. Student's unpaired t-test was used for unrelated parametric data. Data are expressed as the mean ± SEM. P values less than 0.05 were considered statistically significant.
- Top of page
- PATIENTS AND METHODS
- AUTHOR CONTRIBUTIONS
Identifying miRNA-targeting collagen genes.
Biocomputational prediction algorithms are well-established screening tools for miRNA profiling (22). Among other miRNA, such as the let-7 family, miR-26, and miR-155, the miR-29 family members (miR-29a/29b/29c) were identified as potential posttranscriptional regulators of collagen genes. The miR-29 family was selected for further analysis because it was predicted to have the best seed matches with the collagens (http://www.targetscan.org) and because it is broadly conserved among most vertebrates (23). MicroRNA-29a and miR-29b1 are clustered together and are located in chromosome 7 in humans, whereas the miR-29b2/miR-29c cluster is found on chromosome 1 (http://www.mirbase.org). Other predicted profibrotic targets of miR-29a included, for example, PDGF-B, PDGFRβ, Sp-1, thrombospondin, and secreted protein, acidic and rich in cysteine (SPARC).
Down-regulation of baseline miR-29 expression in SSc.
Given that most miRNA act as posttranscriptional repressors and that in SSc, the expression of collagen genes is remarkably up-regulated, we expected a down-regulation of the miR-29 family in SSc patients as compared with healthy controls. Therefore, we analyzed the expression of the miR-29 family and of type I and type III collagen in skin fibroblasts and biopsy samples from SSc patients and healthy controls. We found that all members of the miR-29 family were expressed in skin fibroblasts from SSc patients, with no differences in disease subsets (limited versus diffuse SSc) or disease stages (early [<2 years] versus late [>2 years]). (Data on the expression of miR-29 in different disease subsets and stages of SSc are available upon request from the authors.)
Within the miR-29 family, miR-29a showed the highest level of expression (mean ± SEM ΔCt 17.4 ± 0.2), followed by miR-29c (ΔCt 22.8 ± 0.1), and miR-29b (ΔCt 27.1 ± 0.3) or miR-29c (ΔCt 22.8 ± 0.1) (Figure 1A). However, compared with normal skin fibroblasts, the baseline levels of miR-29 were consistently down-regulated in SSc fibroblasts: by a mean ± SEM of 60 ± 15% for miR-29a (P = 0.004), 45 ± 15% for miR-29b (P = 0.001), and 58 ± 7% for miR-29c (P < 0.0001) (Figure 1B). In the same SSc fibroblasts, this was paralleled by increased levels of COL1A1 and COL3A1 mRNA (mean ± SEM 1.5 ± 0.2–fold and 1.6 ± 0.4–fold, respectively; P = 0.06) in SSc fibroblasts as compared with healthy control fibroblasts (Figure 2A).
Next, we investigated the expression of the miR-29 family in skin samples from SSc patients. First, the expression of miR-29 in tissue extracts isolated from paraffin-embedded skin sections was analyzed, and similar to the findings in SSc fibroblasts, decreased levels of the miR-29 family members were observed. In particular, miR-29a was down-regulated in the dermis of SSc patients as compared with healthy controls (by 79 ± 9%; P < 0.0001) (Figure 1C).
To exclude the possibility that paraffin processing influenced the content of miRNA in the tissue, we additionally investigated the expression of the miR-29 family in fresh skin biopsy samples. Consistent with the previous results, miR-29a was found to be down-regulated by 42 ± 8% (P = 0.0011) in fresh skin biopsy samples from SSc patients as compared with healthy controls (Figure 1D). In the same SSc skin biopsy samples, this was again paralleled by increased levels of COL1A1 and COL3A1 mRNA (8.0 ± 2.5–fold [P = 0.09] and 2.7 ± 0.6–fold [P = 0.03], respectively) as compared with healthy control skin in the same samples (Figure 2B).
Taken together, the miR-29 family members and miR-29a in particular were consistently expressed at a remarkable down-regulation in skin fibroblasts and skin tissue sections from SSc patients as compared with healthy controls.
Altered collagen expression following overexpression and knockdown of miR-29.
We next aimed to analyze the functional effects of altered miR-29 expression on fibrogenesis. According to our hypothesis, the restoration of miR-29 levels would decrease collagen levels in SSc fibroblasts, whereas down-regulation of miR-29 in normal skin fibroblasts would induce an SSc-like phenotype, with an increase in the expression of collagen.
In SSc fibroblasts, transfection with pre-miR-29a/29b/29c increased the levels of the respective miRNA by a mean ± SEM of 147.4 ± 26.9–fold, 1,625 ± 1,271–fold, and 57.5 ± 0.4–fold. Transfection with the respective pre-miR did not substantially alter the levels of the other miR-29 family members. Knockdown with anti-miR reduced the expression of the respective miRNA to a mean ± SEM of 0.08 ± 0.05–fold for miR-29a, 0.1 ± 0.04–fold for miR-29b, and 0.5 ± 0.04–fold for miR-29c. Except for knockdown of miR-29a, which also affected the levels of miR-29b, the effects were specific.
In all gain-and-loss of function experiments, miR-29a, as compared with miR-29b/29c, showed the strongest effects. The expression of types I and III collagen decreased by a mean ± SEM of 66 ± 10% (P = 0.0003) and 65 ± 8% (P < 0.0001), respectively, on the mRNA level, as analyzed by TaqMan-based real-time PCR (Figure 3A), and by 37 ± 5% (P < 0.0001) and 45 ± 9% (P = 0.0006), respectively, on the protein level, as analyzed by Western blotting (Figure 3B). Knockdown of miR-29a in normal skin fibroblasts increased the expression of types I and III collagen by 1.8 ± 0.5–fold and 2.0 ± 0.6–fold, respectively, on the mRNA level (Figure 3C) and by 2.1 ± 0.3–fold (P = 0.03) and 1.8 ± 0.4–fold, respectively, on the protein level (Figure 3D).
Overexpression of miR-29b in SSc fibroblasts down-regulated the expression of mRNA for type I and type III collagen by 35 ± 9% (P = 0.01) and 40 ± 15% (P = 0.03), respectively, and reduced the level of type Iα2 collagen protein by 44 ± 22% (P = 0.05). The expression of type III collagen protein did not change significantly. Similar to the results with miR-29a, knockdown of miR-29b in healthy dermal fibroblasts increased the expression of types I and III collagen mRNA, but the effects at the protein level were less distinct. (Data showing the effects of miR-29b on the expression of collagen are available upon request from the authors.)
Enforced expression of miR-29c in SSc fibroblasts decreased type I collagen mRNA by 62 ± 14% (P = 0.001) and type III collagen mRNA by 61 ± 14% (P = 0.002), whereas on the protein level, expression of types Iα2 and III collagen was reduced by 14 ± 7% (P = 0.05) and 30 ± 5% (P = 0.0002), respectively. The effects of miR-29c knockdown in healthy control fibroblasts on the expression of types I and III collagen were similar to the effects of the other miR-29 family members examined. (Data showing the effects of overexpression and knockdown of miR-29c on collagen expression are available upon request from the authors.)
COL3A1, a direct target of miR-29.
The observed reduction in collagen expression by miR-29 family members could be caused indirectly, for example, by a modification of profibrotic cytokines and growth factors, or directly, by direct posttranscriptional regulation of collagen gene expression. To assess whether there was direct regulation of collagen genes, we created a luciferase reporter gene system by cloning a part of the 3′-UTR of COL3A1 with respective binding sites for miR-29a, miR-29b, and miR-29c into the pGL3 control vector (Figure 4A).
All miR-29 family members were expressed in HEK 293 cells, but at lower baseline levels as compared with the expression in skin fibroblasts (mean ± SEM ΔCt 22.87 ± 0.25 for miR-29a, 28.09 ± 1.9 for miR-29b, and 24.32 ± 0.44 for miR-29c). Transfection of HEK 293 cells with pre-miR was specific and had no substantial effects on the other family members. In HEK 293 cells, transfection with pre-miR-29a/29b/29c increased the levels of the respective miRNA by a mean ± SEM of 510 ± 189–fold, 5,517 ± 1,597–fold, and 420 ± 152–fold. Knockdown of the miR-29 decreased their expression to 0.36 ± 0.07–fold for miR-29a, 0.13 ± 0.09–fold for miR-29b, and 0.06 ± 0.02–fold for miR-29c. Except for knockdown of miR-29c, which also affected the levels of miR-29b and, to a lesser extent, miR-29a, the effects were specific.
In the reporter gene assay, cotransfection of HEK 293 cells with pre-miR-29a and the luciferase reporter plasmid containing the 3′-UTR of COL3A1 (Figure 4A) decreased the relative luciferase activity by a mean ± SEM of 47 ± 10% (P = 0.007) compared with cells transfected with the plasmid and scrambled controls (Figure 4B). Accordingly, cotransfection with anti-miR-29a increased the relative enzyme activity by 20 ± 6% (P = 0.01) (Figure 4C). Similar effects were observed for miR-29b and miR-29c. Cotransfection with pre-miR-29b reduced the relative luciferase activity by 67 ± 10% (P < 0.0001) (Figure 4B), and cotransfection with anti-miR-29b increased the relative enzyme activity by 30 ± 7% (P = 0.008) (Figure 4C). Cotransfection with pre-miR-29c decreased the relative luciferase activity by 70 ± 6% (P < 0.0001) (Figure 4B), whereas cotransfection with anti-miR-29c did not show significant effects (Figure 4C). In summary, these findings prove that miR-29 directly regulates collagen production.
MicroRNA-29, a downstream mediator of profibrotic molecules.
To explore factors that potentially down-regulate miR-29 levels in SSc, we analyzed the effects of major profibrotic cytokines and growth factors. Treatment of normal skin fibroblasts with profibrotic cytokines induced an SSc-like phenotype, with down-regulation of miR-29a to an extent similar to that seen in SSc fibroblasts. The most pronounced effects were observed for miR-29a after 48 hours of treatment, with a reduction in the levels of mRNA for TGFβ by 80 ± 8% (P < 0.0001), for PDGF-B by 70 ± 20% (P = 0.003), and for IL-4 by 74 ± 11% (P < 0.0001) (Figure 5A). The levels of miR-29b and miR-29c were also reduced, but to a lesser extent. (Data showing miR-29b and miR-29c as downstream mediators of profibrotic molecules are available upon request from the authors.)
To confirm miR-29a as a downstream mediator of the profibrotic effects of these cytokines, we investigated whether rescuing cytokine-treated normal fibroblasts by overexpression of miR-29a could decrease the levels of collagens induced by these cytokines. Stimulation with TGFβ, PDGF-B, or IL-4 increased the levels of COL1A1 and COL3A1 mRNA in nonrescued scrambled controls, but this effect was abolished by overexpression of miR-29a (Figure 5B). Thus, it seems likely that miR-29 is indeed a downstream mediator of the profibrotic effects of the cytokines we examined.
We next analyzed whether treatment with TGFβ, PDGF-B, or IL-4 could further decrease the reduced levels of miR-29 in SSc fibroblasts. In contrast to healthy fibroblasts, no additional decrease in miR-29a/29b/29c levels was observed. Accordingly, transdifferentiation of SSc fibroblasts into myofibroblasts by long-term stimulation with TGFβ for 4 days did not further reduce the levels of miR-29 (data not shown). The recombinant proteins used in these experiments were biologically active, as proven by the strong decrease in miR-29 in healthy fibroblasts and by an induction in downstream molecules, such as type I collagen, in both healthy and SSc fibroblasts. A potential explanation for this finding was that miR-29 levels in SSc fibroblasts were already maximally suppressed and therefore could not be further decreased by profibrotic cytokines. To further analyze this hypothesis, SSc fibroblasts were incubated with the tyrosine-kinase inhibitor imatinib, which abrogates TGFβ and PDGF-B signaling (14, 21). Indeed, treatment of SSc fibroblasts with imatinib increased the levels of miR-29a by a mean ± SEM of 40 ± 7% (P = 0.002), miR-29b by 60 ± 3% (P = 0.05), and miR-29c by 30 ± 6% (Figure 5C).
To further investigate which of the imatinib targets mediates the observed effects, we performed selective knockdown experiments for c-abl, PDGFRβ, and c-kit in SSc fibroblasts and analyzed the levels of miR-29. Inhibition of c-abl and PDGFRβ, but not c-kit, increased the levels of miR-29a by a mean ± SEM of 9.1 ± 4.6–fold and 5.6 ± 3.7–fold (P = 0.03), respectively, as compared with the scrambled controls (Figure 5D). Similar effects were observed for miR-29c and for miR-29b (data available upon request from the authors). These results suggest abrogation of PDGFRβ and c-abl signaling as the main mechanism of how imatinib restores the levels of miR-29 in SSc fibroblasts.
MicroRNA-29 in the murine model of bleomycin-induced skin fibrosis.
Finally, we wanted to investigate whether the reduced levels of miR-29 in human SSc are mimicked in the SSc model of bleomycin-induced skin fibrosis in mice. Similar to the findings in human SSc (Figure 1B), the expression of all miR-29 family members in the skin of mice with bleomycin-induced fibrosis was down-regulated after 6 weeks of bleomycin challenge: miR-29a was reduced by a mean ± SEM of 34 ± 19% (P = 0.02), miR-29b by 50 ± 27% (P = 0.01), and miR-29c by 69 ± 11% (P < 0.0001) (Figure 6A).
We next used imatinib to analyze the effects of impaired PDGF-B and TGFβ signaling on the levels of miR-29 in vivo. To mimic the clinical situation, we used a modified bleomycin model in which fibrosis is established before treatment with imatinib is started (24). Remarkably, imatinib treatment of mice with bleomycin-induced fibrosis increased the expression of miR-29a by a mean ± SEM of 2.0 ± 1.4–fold (P = 0.003) as compared with NaCl control treatment. The levels of the other miR-29 family members showed a similar, but not significant, up-regulation (Figure 6B). These results suggest that even in preestablished fibrosis, abrogation of PDGFRβ and c-abl/TGFβ signaling could increase the levels of miR-29a. As compared with normal controls (i.e., mice treated with saline instead of bleomycin for 6 weeks and then treated with NaCl instead of imatinib during the last 3 weeks), the levels of miR-29a, but not miR-29b or miR-29c, were normalized (data not shown).
- Top of page
- PATIENTS AND METHODS
- AUTHOR CONTRIBUTIONS
The direct posttranscriptional regulation of collagens by miR-29 adds a novel aspect to the complex network of factors that regulate the expression of collagen genes in SSc at the level of transcription. This includes deregulated stimulatory (e.g., Sp-1, Smad3, p300/CREB binding protein) and inhibitory transcription factors and cofactors (e.g., Smad7, Fli-1, peroxisome proliferator–activated receptor γ), as well as several stimulatory cytokines and growth factors (25). The relative importance of these factors in the pathogenesis of SSc is unclear. Similarly, there is limited knowledge about the extent to which down-regulated repressors and up-regulated stimulators of collagen transcription interplay. Of note, miR-29a as a posttranscriptional repressor, which is induced by the stimulators mentioned above, acts downstream of most of the previously identified profibrotic molecules.
Thus, the main hypothesis derived from the present study is that targeting miR-29 family members as posttranscriptional regulators could be a potent antifibrotic approach. This needs to be tested in future studies. Specifically, the effects of the modification of miR-29 levels on the synthesis of extracellular matrix should be analyzed in several animal models of SSc covering different aspects of the disease. Optimally, miR-29–knockout mice could be created and potentially crossed with genetic fibrotic animal models or challenged with profibrotic stimuli. If the findings of these experiments turned out to be promising, modification of miR-29 levels could be adapted to humans and the first clinical proof-of-concept studies and studies on toxicity could be performed.
So far, miRNA-based therapeutic approaches in animal studies of other diseases have been promising, since toxic effects were not observed (10–13). Accordingly, our experiments with the restoration of miR-29 in skin fibroblasts did not have overt toxic effects on cell survival and proliferation (data not shown). Our results also demonstrated that the bleomycin model of skin fibrosis, although it does not cover all of the features of human SSc, represents one of the potential in vivo models in which to test the different strategies of miR-29 restoration in SSc because it displayed the reduced levels of miR-29 that are seen in humans with SSc. Abrogation of PDGFRβ and, eventually, TGFβ signaling by administration of imatinib increased the levels of miR-29 in vivo. The direct therapeutic modulation of miR-29 might have even more pronounced effects. Whether imatinib also restores miR-29 levels in human SSc should be analyzed in the ongoing imatinib trials.
Interestingly, miR-29 might additionally influence the production of other profibrotic molecules, such as PDGF-B, PDGFRβ, thrombospondin, and SPARC, since in silico analyses have shown these molecules to be predicted targets of miR-29. The ability of miR-29 to coordinate such a broad spectrum of profibrotic molecules illustrates the potential of miRNA to control complex molecular interactions. Further research should include the analysis of miRNA expression profiles in SSc biologic samples, as well as the identification of miRNA that regulate other key molecules, such as growth factors, profibrotic cytokines, and matrix metalloproteinases.
The precise mechanisms that lead to the down-regulation of miR-29, especially miR-29a, in SSc remain to be determined, but our study showed that TGFβ, PDGF-B, and IL-4, all of which are major profibrotic mediators in SSc, decreased the expression of miR-29a in normal skin fibroblasts. Moreover, the stimulatory effects of these cytokines on collagen synthesis could be reduced by rescuing cells with miR-29a, suggesting that the reduced levels of miR-29a occur downstream of these cytokines and that miR-29 at least partially mediates their profibrotic effects. Accordingly, inhibition of TGFβ and PDGF-B signaling by imatinib or, more specifically, by knockdown of c-abl or PDGFRβ restored the decreased levels of miR-29 in SSc fibroblasts. Since these major profibrotic factors are not specific for SSc, the miR-29 family could also play a role in other fibrotic diseases. Indeed, there is evidence that miR-29b plays a role in cardiac fibrosis (26), and miR-29c was recently found to be involved in the mechanisms of fibrosis during the development of nasopharyngeal carcinomas (24).
Additional factors potentially involved in the down-regulation of miR-29 in SSc fibroblasts may include intrinsic mechanisms, because the low expression of miR-29 family members was maintained from early to late cell culture (passages 3–8) without additional profibrotic stimuli and because stimulation with the profibrotic cytokines did not further decrease the levels of miR-29 in SSc fibroblasts.
Epigenetic dysregulation of gene expression may also be part of this scenario. Fli-1, a negative regulator of collagen synthesis, has been proven to be epigenetically silenced in SSc fibroblasts by DNA methylation and histone deacetylation (27). Changes in the expression of histone deacetylases have also been found to play a role in the profibrotic mechanisms of SSc (28, 29). Thus, epigenetic silencing of miR-29 might also contribute to the activated state of SSc fibroblasts. Recently, a study of skeletal myogenesis and the pathogenesis of rhabdomyosarcoma demonstrated that in rhabdomyosarcoma cells and primary tumors with impaired differentiation, miR-29 was epigenetically silenced by an activated NF-κB/YY-1 pathway (30). On the other hand, DNA methyl transferases are experimentally validated targets of miR-29 (31), and thus, it might also be hypothesized that miR-29 plays a role in the epigenetic regulation of profibrotic genes in SSc.
So far, the available data on miR-29 indicate a general role in the development of fibrosis not limited to SSc. To evaluate in SSc whether low miR-29 levels correlate with individual profibrotic activity, as assessed by the modified Rodnan skin thickness score (MRSS) or by disease stage (early versus late), longitudinal, rather than cross-sectional, studies with larger cohorts of patients will be needed, since in our study, patients with high MRSS scores and/or early disease were underrepresented and, in general, the patients were only assessed at a single time point. In addition, it seems that there is some organ specificity in terms of which member of the miR-29 family has a predominant action. This fits very well with the findings of previous studies demonstrating that fibroblasts from different anatomic sites showed remarkable differences in their patterns of gene expression (32).
In conclusion, our study shows that in SSc miR-29a directly regulates collagen expression at the posttranscriptional level, which is a novel aspect of the complex regulatory network of fibrosis in SSc. Furthermore, our data indicate that the dysregulation of miR-29a is mediated, at least in part, by profibrotic cytokines and growth factors. Based on the encouraging results from miRNA-based therapeutic approaches in animal studies (10–12, 33) and given the fact that the first human trial targeting a liver-specific miRNA involved in the replication of the hepatitis C virus has been launched (34), the development of strategies to maintain the expression or to prevent the repression of miR-29a in SSc appears to be a promising future therapeutic strategy.
- Top of page
- PATIENTS AND METHODS
- AUTHOR CONTRIBUTIONS
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. O. Distler had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Maurer, Stanczyk, J. H. W. Distler, S. Gay, O. Distler.
Acquisition of data. Maurer, Akhmetshina, Trenkmann, Brock.
Analysis and interpretation of data. Maurer, Stanczyk, Jüngel, Akhmetshina, Kowal-Bielecka, R. E. Gay, Michel, S. Gay, O. Distler.
- Top of page
- PATIENTS AND METHODS
- AUTHOR CONTRIBUTIONS
We thank Ferenc Pataky and Peter Künzler for excellent technical support.
- Top of page
- PATIENTS AND METHODS
- AUTHOR CONTRIBUTIONS