Regulation of the IGFBP-5 and MMP-13 genes by the microRNAs miR-140 and miR-27a in human osteoarthritic chondrocytes

Many factors contribute to the overall degradation of cartilage observed in osteoarthritis (OA), either directly or indirectly by modulating anabolic factors. Examples of such molecules are the matrix metalloprotease (MMP)-13 and the insulin-like growth factor binding protein (IGFBP)-5. MMP-13 is a well known key player in cartilage biology and OA pathology because of its capacity to degrade, in addition to collagens, a wide range of matrix components [1‐6]. Although a large number of factors including pro-inflammatory cytokines, growth factors, and fibronectin fragments have been reported to regulate MMP-13 expression [5, 7, 8], further knowledge about its regulation is needed in order to identify factors that could specifically inhibit this MMP while sparing others and, as such, avoid the unwanted side effects observed with broad spectrum MMP inhibitors [9, 10]. IGFBPs are proteins known to modulate the availability/activity of the anabolic factor IGF-1. Evidence has shown that in the joint, IGFBP-5 plays an important storage role for IGF-1 [11]. Furthermore, results from a study using an OA dog model demonstrated that increasing IGFBP-5 concentration led to an increased level of IGF-1 and was associated with a reduction in cartilage destruction [12]. Despite its regulatory role in cartilage, the regulation of human IGFBP-5 itself has not yet been investigated in this tissue or in chondrocytes.

Although MMP-13 promoter regulation has been the subject of many publications [13‐16], there is no report on the role of 3'-untranslated regions (3'-UTRs) on either its regulation or that of IGFBP-5. Since microRNAs (miRNAs) act on this region and are important regulators of gene expression, we investigated whether MMP-13 and IGFBP-5 are the targets of specific miRNAs.

miRNAs are small non-coding RNAs (20-25 nucleotides) naturally produced by the cells. They are derived from primary miRNA transcripts (70-100 nucleotides) that are processed in the nucleus to precursor miRNAs (pre-miRNAs) by the ribonuclease Drosha [17]. The pre-miRNAs are then transported into the cytoplasm where they are further processed into miRNAs by the ribonuclease Dicer [18]. The miRNAs play a role in gene silencing by regulating the stability or translational efficiency of target messenger RNA (mRNA). Depending on the degree of base pairing between the miRNA and the target mRNAs, the miRNAs either repress translation (imperfect pairing) or cleave the mRNAs (perfect pairing) [19]. Pairing generally occurs in the 3'UTR of the mRNAs. Another mechanism of miRNA-mediated mRNA degradation may involve AU-rich elements (AREs), which are located in the 3'-UTR of unstable mRNAs [20].

Several hundred miRNAs have been identified so far and initial studies have linked specific miRNAs to different tissues, developmental processes, and pathologies such as cancer [21‐23]. Although algorithms are used to predict potential mRNA targets, only a few miRNAs have been validated and assigned to specific mRNAs.

The cellular outcomes of miRNA-mediated gene regulation are complex, as some miRNAs decrease while others increase cell growth, and still others increase the level of apoptosis [22]. However, because of their role, miRNAs may represent another avenue for therapeutic intervention in arthritic diseases.

The importance of miRNAs in joint pathologies and in inflammatory events has been addressed only recently. Stanczyk et al [24] reported that the expression of miR-155 and miR-146a was increased in synovial fibroblasts from rheumatoid arthritis (RA) patients as compared to OA. The miR-146 was also found to be up-regulated in peripheral blood mononuclear cells [25] and in synovial tissues [26] from RA patients. Moreover, the expression of miR-146 and miR-155 was also shown to be up-regulated by bacterial endotoxins and the pro-inflammatory cytokines interleukin-1β (IL-1β) and tumor necrosis factor-α (TNF-α) [27]. In a recent study [28], miR-146a was reported to be expressed mostly in OA cartilage showing mild scores and its expression was stimulated in normal chondrocytes by IL-1β. These findings thus suggest that some miRNAs could be of importance in the inflammatory events of arthritis.

There have been few reports on the role of miRNAs in cartilage biology. Tuddenham et al [29] reported the presence of miR-140 in cartilaginous tissues of the developing mouse and showed that this miRNA targeted the mouse histone deacetylase 4 mRNA. Kobayashi et al [30] showed that Dicer, an enzyme involved in the miRNA pathway, was essential for chondrocyte function in mice; the growth plates from Dicer-null mice demonstrated a progressive reduction in the proliferating pool of chondrocytes, leading to severe skeletal growth defects and premature death of the mice. The miR-199* was also recently shown to control chondrogenesis in mice, via direct targeting to Smad1 [31]. In humans, there is to our knowledge only one study in which the miRNAs have been profiled comparing normal and OA cartilage [32]. In this study it was found that 16 miRNAs were differentially expressed when OA was compared to normal cartilage. Moreover, comparison with clinical data revealed that some miRNAs (miR-22 and miR-33) and proteins, peroxisome proliferator-activated receptor-α (PPARα), bone morphogenic protein-7 (BMP-7) and IL-1β, highly correlate with body mass index. Moreover, as mentioned above, miR-146a has also been found expressed in a subset of OA cartilage [28].

We thus investigated whether MMP-13 and IGFBP-5 are regulated by miRNAs in human OA chondrocytes. We identified the miRNAs miR-140 and miR-27a as regulators of these two genes and studied their expression and regulation in normal and OA human chondrocytes. This study provides a more comprehensive understanding of the overall regulation of MMP-13 and IGFBP-5.

The objective of this study was to complement the data on MMP-13 and IGFBP-5 regulation at the gene expression level by determining if miRNAs could affect the regulation of these genes and, if so, to identify and validate those miRNAs. Understanding the regulation of these factors is of great importance and could provide a new basis for the rationalization of a therapeutic strategy. Since several reports on miRNA profiling human cartilage [32], cancer [23] and general human tissues [21, 36] have already been published, we chose to follow up on MMP-13 and IGFBP-5 and focus our research on the expression and regulation of miR-140 and miR-27a, as these miRNAs were identified with high prediction by the five computational programs used as possible regulators of both MMP-13 and IGFBP-5 expression.

Many factors contribute to the overall degradation of cartilage in OA. MMP-13 is well known to be up-regulated and to play a major role in the pathophysiological process of OA [1, 4, 5]. On the other hand, the exact function of IGFBP-5 in cartilage is not totally understood, but it is suggested to play a role as facilitator of IGF-1 availability in the tissue. Indeed, IGFBP-5 has been shown to associate with extracellular matrix macromolecules where it is protected from degradation and acts as a local reservoir for IGF-1 [11]. In this bound state, its affinity for IGF-1 is decreased when compared to the soluble state [11], indicating that IGFBP-5 would facilitate the delivery of this growth factor to its specific cell surface receptors. In OA, decreased levels of IGFBP-5 would diminish the capacity of the extracellular matrix to act as a reservoir for IGF-1; the free IGF-1 could then be sequestered by other IGFBPs, such as IGFBP-3 known to be increased in OA [37], resulting in its reduced bio-availability.

Data showed that the IGFBP-5 expression level was significantly decreased in human OA chondrocytes. This concurs with results from a study on another articular cell, the human subchondral bone osteoblast, in which the IGFBP-5 was down-regulated in OA [38]. However, it is in contrast to earlier studies reporting increased expression/production of IGFBP-5 in OA cartilage/chondrocytes [39, 40]. Previous data [41], including our own [37], have shown that the IGFBP-5 protein basal level in human OA chondrocytes was undetectable or very low and subjected to the action of proteases. It is possible that variations may have occurred depending on the culture conditions as well as the variability of the different measurement methods used in those studies (immunohistochemistry, Western blot, semi-quantitative PCR). Here, we used more sensitive methods including quantitative PCR and a specific ELISA.

As IGFBP-5 is essential for maintaining IGF-1 anabolic activity, knowing the factor(s) responsible for its decreased expression level is important. Findings in the literature indicate that IGFBP-5 protein could be degraded by proteases and the serine protease Complement 1s, an enzyme present in the OA joint, was recently identified as being responsible for the cleavage of IGFBP-5 [41]. However, there are very few reports on gene expression. The present study showed that the IGFBP-5 gene expression is down-regulated by miR-140. This appears to be a direct effect, as IGFBP-5 is regulated as early as 24 hours post-treatment by the pre- and anti-miR-140. Although data showed that miR-140 is a regulatory factor of IGFBP-5, this does not imply that it is the only factor to down-regulate IGFBP-5, as miR-140 is also decreased in OA. Indeed, each gene is regulated by a variety of factors, some stimulatory and others suppressive. The decreased expression of IGFBP-5 in OA is the outcome of the interplay between these factors in which miR-140 plays a role.

Interestingly, MMP-13 and bFGF, which were also predicted to be miR-140 targets, were not affected by this miRNA. These latter data indicate that corroboration is necessary to conclude the specificity of a predicted target for a given miRNA.

Recent studies reported the role of some miRNAs in MMP regulation. For example, Stanczyk et al [24] demonstrated that the over-expression of miR-155 in RA synovial fibroblasts induced the repression of MMP-3 but not of MMP-13. However, MMP-3 has not yet been validated as a direct target of miR-155. On the other hand, Jones et al [42] recently reported that miR-9 could modulate MMP-13 expression, and Yamasaki et al [28] found, in OA cartilage, an association between the decreased expression of miR-146a and the increased MMP-13 expression level. Again, MMP-13 as a direct target of these miRNAs was not validated. Indirect regulation of MMP activity by miRNAs has also been recently reported in cancer cells [43]; miR-21 is over-expressed in glioblastomas and targets tissue inhibitor metalloprotease-3, a regulator of some MMPs. Moreover, on chondrocytes, miR-22 was shown to act on MMP-13 but through an effect on two other factors, PPARα and BMP-7 [32]. Thus, the control of gene expression by miRNAs can be both direct and indirect.

In this study, we show that MMP-13, as well as IGFBP-5, are likely indirect targets of miR-27a. Pre-miR-27a did not affect expression and anti-miR-27a treatment started to up-regulate transcription at 48 hours post-treatment, an increase which became significant after 72 hours. Of note, another gene predicted to be a target of miR-27a, IL-10, was not affected by either this pre- or anti-miRNA.

Data on MMP-13 and IGFBP-5 indicate that miR-27a affects the expression of another factor (or factors), which in turn acts on these two genes. It is likely that the factor is a stimulatory regulator of both IGFBP-5 and MMP-13 expression as they are affected only by the anti-miR-27a and not by the pre-miR-27a. The anti-miRNA would antagonize the inhibitory effect of miR-27a on the stimulatory factor resulting in its increased expression, which, in turn, would affect IGFBP-5 and MMP-13.

Although the identification of the miR-27a-targeted intermediate factor is currently ongoing, the computational programs have identified only a few miR-27a target genes that could have the potential to code for MMP-13 regulatory factors, and include PPARγ and Smad2. However, as the activation of PPARγ inhibits rather than stimulates MMP-13 expression [44], Smad2 is a more likely candidate. Although the IGFBP-5 promoter has been cloned and sequenced [45, 46] it has not been fully characterized. However, our results show that TGF-β strongly stimulates IGFBP-5 expression, and Smad2 is implicated in TGF-β signaling [47] TGF-β has also been reported to up-regulate MMP-13 expression [8, 48] and data further showed that the TGF-β-induced MMP-13 production in human OA chondrocytes was triggered by Smad proteins [49]. However, given the large number of potential miR-27a targets, the possibility that miR-27a targets two different regulatory factors for MMP-13 and IGFBP-5 is also considered.

Even though stimulators of IGFBP-5 were found in this study and include the cytokines TNF-α, IFN-γ and IL-10, and the growth factor TGF-β, they do not seem to be sufficient to maintain normal IGFBP-5 levels in OA chondrocytes, as the level of IGFBP-5 was significantly reduced in the diseased cells. This could be explained by the fact that OA chondrocytes do not produce these cytokines at high levels [50], in addition to the slightly increased miR-140 expression following TNF-α treatment. However, because of the differential role of TGF-β in the regulation of IGFBP-5 and miR-140, the low level of IGFBP-5 in OA chondrocytes was surprising. TGF-β significantly reduced miR-140 expression levels at the same time as strongly up-regulating IGFBP-5. Thus, TGF-β would act in two different ways to increase IGFBP-5: directly on its transcription, possibly at the promoter level, and indirectly by decreasing miR-140, a transcription inhibitor.

Reports on miRNA regulation/modulation are scarce. Our study is the first to show the regulation of the two miRNAs, miR-140 and miR-27a, in OA chondrocytes. Data show that TGF-β, known to be increased in OA cartilage [51, 52], is a candidate for the reduced expression of miR-140 in these cells. As for miR-27a, among the factors tested, only IFN-γ and IL-10 decreased its expression. However, as mentioned above, these two factors are not found at high levels in OA [50] which could explain that only a slight reduction in miR-27a was found in OA chondrocytes.

This study is the first to report on the regulation of miRNAs in OA chondrocytes. We show that miR-140 levels are decreased in OA and that the decrease could be attributed to the growth factor TGF-β. Data also suggest that miR-140 could act directly on decreasing IGFBP-5 expression but that miR-27a indirectly decreases both MMP-13 and IGFBP-5. These findings add another level of complexity to the overall regulation of MMP-13 and IGFBP-5, two factors involved in the OA pathological process. Data acquired from the study of these two genes could be the first step in generating more comprehensive knowledge of the regulation of other MMPs and IGFBPs involved in OA, which could open up novel avenues in therapeutic strategies targeting this disease.