SEARCH

SEARCH BY CITATION

Keywords:

  • microRNA;
  • TGF-β1;
  • fibrosis;
  • gene expression;
  • translational repression;
  • wound healing

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Mechanisms of miR-mediated regulation
  5. Transforming growth factor beta-1
  6. Transforming growth factor beta-1 synthesis: importance of post-transcriptional regulation
  7. MiRs as effectors of key transitions in cellular phenotype in tissue fibrosis
  8. Tissues
  9. Conclusions and future perspectives
  10. Acknowledgments
  11. Author contribution statement
  12. References

MicroRNAs are short noncoding RNA regulators that repress synthesis of their targets post-transcriptionally. On average, each microRNA is estimated to regulate several hundred protein-coding genes, and about 60% of proteins are thought to be regulated by microRNAs in total. A subset of these genes, including the key profibrotic cytokine transforming growth factor beta-1 (TGF-β1), exhibits particularly strong levels of post-transcriptional control of protein synthesis, involving microRNAs and other mechanisms. Changes in microRNA expression pattern are linked to profound effects on cell phenotype, and microRNAs have an emerging role in diverse physiological and pathological processes. In this review, we provide an overview of microRNA biology with a focus on their emerging role in diseases typified by organ fibrosis.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Mechanisms of miR-mediated regulation
  5. Transforming growth factor beta-1
  6. Transforming growth factor beta-1 synthesis: importance of post-transcriptional regulation
  7. MiRs as effectors of key transitions in cellular phenotype in tissue fibrosis
  8. Tissues
  9. Conclusions and future perspectives
  10. Acknowledgments
  11. Author contribution statement
  12. References

The haploid human genome is composed of over 3 giga base pairs of DNA. While only approximately 1.5% of this sequence encodes proteins, it now appears that more than 95% is transcribed into RNA [1, 2]. Identifying the significant components of this widespread genomic expression in health and disease presents a considerable challenge. Numerous functional non-protein-coding RNA species have recently been identified (Table 1), introducing new levels of complexity to the understanding of pathological and physiological processes [3], and representing potential novel targets for disease intervention.

Table 1. Classes of regulatory RNA
TypePrincipal site of actionRole
Small nuclear RNA (snRNA)UbiquitousRNA splicing
Small nucleolar RNA (snoRNA)UbiquitousChemical modifications of RNA (ie methylation)
MicroRNA (miR)UbiquitousTranslational repression and transcript degradation
Piwi-interacting RNA (piRNA)TestisSilencing of retrotransposons and other genetic elements
Long noncoding RNA (lncRNA)UbiquitousTranscriptional and post-transcriptional gene regulation. Inactivation of X chromosome and regulation of telomerase activity
Endogenous siRNA (endosiRNA)OocytePost-transcriptional gene silencing during oocyte development

MicroRNAs (miRs), one such class of noncoding (nc) RNAs, are small, endogenous, post-transcriptional regulators of gene expression. The last decade has revealed a central role for miRs in diverse cellular processes, including proliferation, differentiation, and death. miRs were first identified in the nematode Caenorhabditis elegans, where they were found by Victor Ambros's group to control developmental timing [4]. Additional miRs were subsequently identified that regulate C. elegans development, and the finding that one of these, let-7, was conserved in mammals led to appreciation that miRs might be an important class of regulator of gene expression in higher organisms [5].

The number of miR transcripts known to be encoded by the human genome now exceeds 1000. Some of these transcripts are widely expressed, while the synthesis of others is restricted by developmental stage or organ. The expression of the majority of human genes is mediated by a combination of trans-acting regulators, and miRs are now second only to transcription factors in importance in this regard [6].

Recent estimates suggest that selective pressure on more than 60% of human protein-coding genes has maintained pairing to miRs, which regulate the expression of these messenger RNAs post-transcriptionally [7]. Individual miR families may have a multiplicity of targets, and the average number of conserved sites has been estimated as more than 500 [7]. The effects of one miR family, and indeed one miR sequence, can therefore be both profound and widespread.

Differences in miR expression profiles, together with the functional impact of these miRs on expression of their target genes, serve to reinforce tissue boundaries. Recent human, chimpanzee, and rhesus macaque high-throughput sequencing data show that miR expression varies more between different tissues than between different species [8]. In addition, these authors suggest that the effect of miRs on global gene expression may be more far-reaching than previously predicted, due to their regulation of transcription factor expression [8].

This review concerns miRs, but it is important to remember that they represent only the best characterized of several classes of ncRNAs that may act in a coordinated manner to define protein expression post-transcriptionally (Table 1). Long noncoding RNAs are another emerging class of post-transcriptional regulator, which may act as natural antisense RNAs. Transcribed at protein-coding genomic loci, from the opposite genomic DNA strand to the messenger RNA (mRNA) sequence, these antisense transcripts can regulate expression in a cell-specific manner. An example relevant to fibrosis is the regulation of hyaluronan synthase 2 (HAS2) gene expression via its interaction with natural antisense RNA, HAS2-AS1 [9]. This can result in either up- or down-regulation of HAS2 transcript expression in a cell-specific manner, and hyaluronan production may be modulated bidirectionally by these mRNA–ncRNA interactions [9, 10]. Synthesis of hyaluronan, an extracellular matrix glycosaminoglycan, is up-regulated in liver, lung, kidney, and peritoneal fibrosis [11-13].

Mechanisms of miR-mediated regulation

  1. Top of page
  2. Abstract
  3. Introduction
  4. Mechanisms of miR-mediated regulation
  5. Transforming growth factor beta-1
  6. Transforming growth factor beta-1 synthesis: importance of post-transcriptional regulation
  7. MiRs as effectors of key transitions in cellular phenotype in tissue fibrosis
  8. Tissues
  9. Conclusions and future perspectives
  10. Acknowledgments
  11. Author contribution statement
  12. References

Like mRNAs, miRs are transcribed by RNA polymerase II and may be derived from independent genes or from introns of protein-coding genes. The primary microRNA transcripts (pri-miRs) share important post-transcriptional 5′- and 3′-end processing events with mRNAs, namely capping and polyadenylation. pri-miRs contain one or more stem-loop secondary structures, from which the mature miRs are derived. These secondary structure elements are recognized by members of the RNase III family of enzymes, including Drosha and Dicer. Drosha, in association with DGCR8 and p68, acts in the nucleus to cleave an ∑70-nucleotide pre-miRNA from the pri-miR, which is then exported to the cytoplasm by Exportin 5. Dicer cleaves an ∼22-nucleotide miR/miR* duplex from the pre-miR. The strand corresponding to the mature miR is incorporated into the miR-induced silencing complex (miRISC) and the other strand is degraded. Preferential selection of one strand is typically observed. The other strand may be referred to as the passenger strand, or miR* sequence. However, both strands may exhibit miRISC incorporation and miR activity, and the identity of the strand with greatest probability of selection may switch between organisms through evolution [14].

Mammalian miRs identify their targets through imperfect base-pairing with sequences in the 3′-UTR of target mRNAs. Continuous base-pairing of miR nucleotides 2–8 (seed region) is typically required for efficient targeting. The extent of post-transcriptional repression of a given mRNA by a given miR does not appear to rest only on their relative abundance, but rather reflects an interaction of miRs with other mechanisms of post-transcriptional control. For example, access of the miRISC to target sites in the mRNA may be blocked by binding of other RNA binding proteins, as is seen in the interaction of bicaudal with the miR-17 binding site in the 3'-UTR of polycystin 2 [15].

Core components of the miRISC include Argonaute proteins (AGO1–4), which interact directly with the miR, and glycine-tryptophan protein (GW182), which facilitates deadenylation of mRNA targets. The precise mechanism of miR-mediated translational repression has still to be elucidated, but a consensus is now emerging (reviewed in refs 16 and 17). Translation of mRNA into protein involves (i) initiation, in which eukaryotic initiation factors bind to the 5′-UTR and the small ribosomal subunit is recruited and scans to the start codon; (ii) elongation, in which amino acids are added to the nascent polypeptide chain; and (iii) termination, in which the polypeptide separates from the mRNA. Recent evidence indicates that translational repression occurs at the initiation step [18-20], rather than by stalled elongation or premature termination. The repressed mRNA is then deadenylated [16] and potentially stored in a repressed state. Ultimately, the deadenylated transcripts are decapped and degraded by the 5′–3′ exonuclease XRN1 pathway [16].

In order for miRs to function efficiently, it has been suggested that they require correct subcellular localization, with co-localization of miRISC components. One potential candidate site for miR-mediated repression is the discrete cytoplasmic foci known as processing bodies (P-bodies) [21, 22]. P-bodies function as sites of translational repression and degradation of mRNA, and are enriched in enzymes responsible for deadenylation, decapping, and decay, together with miRs and repressed mRNAs. P-bodies are dynamic in nature, with protein, mRNA, and miRs transiting in and out, and their number varies depending on the cellular state. P-bodies are induced by forms of stress that lead to polysome disassembly. Components of the miRISC are detected in P-bodies, including AGO proteins and GW182 [21, 22], and disruption of miR biogenesis or depletion of GW182 results in their disappearance [23-25]. A positive correlation also exists between the degree of miR translational repression and the accumulation of target mRNA transcripts in P-bodies [26, 27], but miR-mediated repression may be seen in their absence [24, 25]. Overall, P-bodies are intimately linked to miR-mediated translational repression and are postulated to play an important role in certain cellular states, such as stress responses.

Although the majority of proteins are probably post-transcriptionally regulated to some extent, there are paradigm genes for which post-transcriptional regulation is particularly important. Transforming growth factor beta-1 (TGF-β1) is one such gene and will be used in this review to highlight particular mechanisms of post-transcriptional control.

Transforming growth factor beta-1

  1. Top of page
  2. Abstract
  3. Introduction
  4. Mechanisms of miR-mediated regulation
  5. Transforming growth factor beta-1
  6. Transforming growth factor beta-1 synthesis: importance of post-transcriptional regulation
  7. MiRs as effectors of key transitions in cellular phenotype in tissue fibrosis
  8. Tissues
  9. Conclusions and future perspectives
  10. Acknowledgments
  11. Author contribution statement
  12. References

TGF-β1 is of central importance in wound healing, fibrosis, and in the negative regulation of inflammation. It directs key cellular processes including proliferation, survival, differentiation, motility, and adhesion. Precise control of TGF-β1 expression is required for normal embryogenesis. Increased expression of TGF-β1 is a key driver of scarring in most human diseases typified by fibrogenesis. Dysregulated TGF-β1 synthesis is an important component of carcinogenesis and metastasis [28]. Consistent with these roles, increased TGF-β1 synthesis is seen in the majority of human diseases in which fibrosis is a dominant part of pathology, and in their respective animal models. Its role in these processes is covered in detail in many of the other reviews in this series.

The general importance of TGF-β1 beyond its role as a key profibrotic cytokine is illustrated by the phenotype of knockout mice. Mice null for TGF-β1 have a high rate of intrauterine death. A proportion survive due to transfer of maternal TGF-β1 across the placenta, but die of a systemic autoimmune-like disease in which tumour necrosis factor and interleukin-1 levels are markedly elevated [29]. Mice heterozygous for TGF-β1 deletion have a normal phenotype at birth, but exhibit increased epithelial cell proliferation, and have a greatly increased risk of carcinoma formation when challenged with carcinogens [30].

The mechanisms by which cells read and respond to TGF-β1 family signals have been subject to many excellent reviews, notably that by Massague et al. [31]. The principal effectors by which TGF-β1 family proteins elicit their actions are the Smad proteins. The core architecture of Smad signalling involves the phosphorylation of R-Smads (Receptor-regulated Smad proteins) followed by the formation of R-Smad/co-Smad complexes, which accumulate in the nucleus and alter gene transcription. In the case of TGF-β1, the predominant R-Smads are Smad2 and Smad3, which combine with the co-Smad, Smad4.

The details by which Smad complexes alter gene transcription are more complex. A cell typically responds to TGF-β1 by increasing and repressing the transcription of hundreds of different genes, with the direction and magnitude of such responses dependent on phenotype and context of the cell. The core Smad binding element (SBE) in DNA is beguilingly simple, comprising the sequence CAGA. However, the energy of the Smad–SBE interaction is not sufficient alone to engender strong binding. Rather, Smad–DNA interaction may depend on cofactor engagement, and the requirement for precise location on the DNA in relation to the core Smad binding element to allow the complete Smad/cofactor/DNA complex to form affords a great deal of specificity.

Transforming growth factor beta-1 synthesis: importance of post-transcriptional regulation

  1. Top of page
  2. Abstract
  3. Introduction
  4. Mechanisms of miR-mediated regulation
  5. Transforming growth factor beta-1
  6. Transforming growth factor beta-1 synthesis: importance of post-transcriptional regulation
  7. MiRs as effectors of key transitions in cellular phenotype in tissue fibrosis
  8. Tissues
  9. Conclusions and future perspectives
  10. Acknowledgments
  11. Author contribution statement
  12. References

Regulatory sites including AP-1 and Sp-1 binding sites have been identified in the promoter region for the TGF-β1 gene [32, 33], and the AP-1 sites have been shown to mediate autoinduction of TGF-β1 transcription [33]. Post-transcriptional regulation is also key to TGF-β1 synthesis. Post-transcriptional regulation was highlighted as likely in the original description of TGF-β1 sequence [34] based on the presence of long 5'- and 3'-UTRs, and subsequent studies reinforced this, based on disparities between the expression of TGF-β1 mRNA and protein across numerous tissues [30]. Specific examination of the rate of synthesis of TGF-β1 protein from its mRNA shows that it is inherently poorly translated, in vitro and in vivo [35, 36], and that increased translation can be stimulated by specific cues [37-39].

Multiple mechanisms for translational control of TGF-β1 have been identified. The 5'-UTR of TGF-β1 is inherently inhibitory to translation [40] and contains a regulatory element to which proteins, including YB-1, bind and alter translational efficiency by stabilizing/destabilizing secondary structure formation [41, 42]. The 3'-UTR of TGF-β1 contains an alternate polyadenylation site, leading to transcripts with two different 3'-UTR lengths. The predominant transcript in adult human tissues is the short UTR transcript, which is targeted by miR-744 [43]. MiR-663 has significant similarity to miR-744 and is also identified as targeting TGF-β1 [44], although these experiments were performed with an intermediate UTR length that does not directly reflect an expressed transcript. Direct comparison of miR-663 with miR-744 suggests that miR-663 has a lesser effect on TGF-β1 synthesis [43].

Taken together, these studies show that TGF-β1 is extensively post-transcriptionally regulated by its 5'- and 3'-UTRs, including repression of TGF-β1 synthesis by miR-744 and miR-663. These mechanisms have been analysed in isolation, and an important area for future work may be analysis of the coordinate interaction of 5'- and 3'-mechanisms of translational control.

MiRs as effectors of key transitions in cellular phenotype in tissue fibrosis

  1. Top of page
  2. Abstract
  3. Introduction
  4. Mechanisms of miR-mediated regulation
  5. Transforming growth factor beta-1
  6. Transforming growth factor beta-1 synthesis: importance of post-transcriptional regulation
  7. MiRs as effectors of key transitions in cellular phenotype in tissue fibrosis
  8. Tissues
  9. Conclusions and future perspectives
  10. Acknowledgments
  11. Author contribution statement
  12. References

TGF-β1-induced profibrotic changes in cell phenotype are accompanied by significant alterations in miR expression profile [45-49]. Many of these changes may reflect transcriptional regulation by Smad proteins binding to the promoter regions for regulated miRs in combination with cofactors but, intriguingly, TGF-β1 has been shown to post-transcriptionally regulate a subset of miRs by enhanced pri- to pre-miR processing by Drosha [50, 51]. TGF-β1-activated R-Smads translocate to the nucleus and recognize specific pri-miRs by a consensus sequence (5′-CAGAC-3′), not dissimilar to a DNA Smad binding element, present on the stem of the stem-loop secondary structure containing the mature miR [51]. The R-Smad–pri-miR interaction is mediated via the amino-terminus MH1 domain of the R-Smad [51]. The R-Smads associate with the Drosha/DGCR8/p68 microprocessor complex, facilitating the cleavage of the pri- to pre-miR [50, 51]. MiRs post-transcriptionally regulated by TGF-β1 include miR-21, -105, -199a-5p, -215, -421, 509-5p, and -600 [51]. In vascular smooth muscle cells, the resultant induction of miR-21 was shown to initiate a contractile phenotype by targeting programmed cell death 4 (PDCD4), a negative regulator of smooth muscle contractile gene expression [50]. The relative significance of post-transcriptional regulation of miR processing compared to regulation of pri-miR transcription in the downstream actions of TGF-β1 otherwise remains to be elucidated. Overall, however, the changes in miR expression mediated by TGF-β1 appear to be a core part of transitions in cellular phenotype engendered by profibrotic signals, including epithelial to mesenchymal transition (EMT), myofibroblast differentiation, and control of macrophage phenotype.

Epithelial to mesenchymal transition

In health, most tissues contain small numbers of resident fibroblasts. In diseases characterized by fibrosis, increased numbers of fibroblasts are seen, which drive the fibrotic process via synthesis and alteration of the constitution of the extracellular matrix, contractility, and secretion of cytokines and other intermediaries that alter the phenotype of cells of epithelial and other origin. The increase in fibroblast number stems from a combination of proliferation of the resident fibroblasts, recruitment of bone marrow-derived fibroblast precursors, and acquisition of a fibroblastic phenotype by resident cells of other lineages. The most clearly recognized form of the latter is the process of epithelial to mesenchymal transition (EMT), by which epithelial cells may become fibroblasts [52]. EMT is a key process in embryogenesis, and is probably also an important component of carcinogenesis and metastasis. It is likely also to play an important role in fibrotic diseases, although there is controversy as to whether the changes in epithelial phenotype often seen in fibroproliferative diseases represent a true, complete EMT or lesser alterations in cell connectivity and polarity, and gene expression [53].

Several miRs are strongly implicated in EMT processes, most notably the miR-200 family, which comprises five miRs transcribed from two genomic loci: miR-200a, miR-200b, miR200c, miR-141, and miR-429. The miR-200 family members target ZEB1 and ZEB2, which encode transcriptional repressors of E-cadherin, and miR-200 expression strongly marks out epithelial versus mesenchymal phenotype in a panel of 60 cancer-derived cell lines [47]. Overexpression of miR-200 family members enhances E-cadherin expression and hinders TGF-β1-mediated EMT in vitro [45, 46], while enforced miR-200 family expression induces a change in phenotype from mesenchymal to epithelial in the 4T07 mouse carcinoma cell line [46] and in cells exhibiting ectopic expression of the protein tyrosine phosphatase Pez [45]. MiR-192 also targets ZEB1 and ZEB2, and loss of miR-192 is linked to EMT and fibrogenesis in the kidney [54, 55]. Down-regulation of miR-192 by TGF-β1 occurs at the level of transcription, secondary to inhibition of binding of the transcription factor HNF (hepatocyte nuclear factor) to the miR-192 promoter [56]. HNF exhibits a restricted tissue expression pattern that mirrors that of miR-192 and its family members miR-194 and miR-215 [56]. p53 induces expression of miR-200 and miR-192, leading to repression of EMT, suggesting that these miRs may act together to maintain an epithelial phenotype [57]. These data provide compelling evidence that miRs are of central importance in cell fate decisions on the mesenchymal–epithelial axis.

Myofibroblast differentiation

Transition of fibroblasts and fibroblast-like cells to an activated myofibroblastic phenotype characterized by expression of alpha-smooth muscle actin (α-SMA) and EDA-fibronectin is a key stage in large organ fibrosis. It is now becoming clear that a multiplicity of miRs are involved in this process, and below we have summarized several well-characterized instances in which miRs are implicated in phenotypic differentiation to myofibroblasts.

TGF-β1-driven phenotypic conversion of MRC-5 human fetal lung fibroblasts leads to up-regulated miR-21 expression in the resultant myofibroblasts [58]. This effect is mediated via down-regulation of expression of programmed cell death 4 (PDCD4), itself a negative regulator of myofibroblast-specific proteins such as α-SMA [58]. MiR-21 is overexpressed in myofibroblasts following treatment of primary rat adventitial fibroblasts with TGF-β1, and the resultant decreased PDCD4 expression also leads to up-regulated JNK/c-Jun activity [59].

In the presence of TGF-β1, expression of miR-29a, -29b, and -29c is down-regulated in IMR-90 human fetal lung fibroblasts and collagen expression increases [60]. TGF-β1 also down-regulates miR-29 expression in murine hepatic stellate cells (HSCs) [61]. Transfection of human, mouse, and rat HSCs with synthetic nuclear receptor farnesoid X receptor (FXR) ligand GW4064 up-regulated expression of miR-29a, but not in cells from FXR null mice [62]. Exposure of rat HSCs to GW4064 also resulted in down-regulation of the expression of collagen, elastin, and fibronectin genes [62].

In activated HSCs isolated from livers in a mouse model of carbon tetrachloride (CCl4)-induced hepatic fibrogenesis, peroxisome proliferator-activated receptor gamma (PPARγ) expression is decreased following down-regulation of miR-132 expression [63]. Decreased miR-132 up-regulates expression of methyl CpG binding protein 2 (MeCP2), which acts in concert with histone-lysine N-methyltransferase EZH2 to block expression of PPARγ via transcriptional and translation repression [63]. By contrast, miR-132 expression by saphenous vein-derived pericyte progenitor cells mediates protection in the myocardium following mouse myocardial infarction, due in part to reduced myofibroblast formation through inhibition of the miR-132 targets MeCP2 and p120Ras-GTPase activating protein [64].

Macrophage phenotype

Macrophages play important and diverse functions in fibroproliferative diseases. Functionally distinct subpopulations of macrophages are central to scarring and in resolution following injury, in the kidney [65], liver [66], and lung [67]. Bone marrow-derived and peripheral blood monocytes have a distinct miR expression profile [68, 69], and recent studies show that manipulating individual miRs can profoundly alter macrophage phenotype. MiR-155 is up-regulated in primary murine macrophages by pro-inflammatory stimuli [70] and its enforced expression restricts a monocytic-like cell line to macrophage differentiation [71]. In contrast, miR-124 is linked to the quiescent, resident macrophage-like phenotype in microglia, and miR-124 repression leads to a pro-inflammatory state, while transduction of inflammatory macrophages with miR-124 induces their quiescence [72]. These data point to the importance of miRs in determining macrophage phenotype and suggest that this will be a fruitful area for further research in understanding the biology of fibrosis.

Tissues

  1. Top of page
  2. Abstract
  3. Introduction
  4. Mechanisms of miR-mediated regulation
  5. Transforming growth factor beta-1
  6. Transforming growth factor beta-1 synthesis: importance of post-transcriptional regulation
  7. MiRs as effectors of key transitions in cellular phenotype in tissue fibrosis
  8. Tissues
  9. Conclusions and future perspectives
  10. Acknowledgments
  11. Author contribution statement
  12. References

MiR expression profiling across a broad range of cells and tissues indicates distinct expression patterns in different lineages and organs [68]. Intriguingly, it is estimated that more than 97% of mature miRs detected originate from approximately 300 miR genes in a given species, suggesting that a limited pool of highly expressed miRs plays a dominant gene regulatory role [68]. There are an increasing number of studies recording changes in tissue miR expression in disease states in various organs. In aggregate, these data are heterogeneous and complex, but it is possible to discern certain commonalities of miRs dysregulated in experimental models and conditions characterized by (i) increased TGF-β1 and (ii) fibrosis. A summary of these studies follows for the kidney, heart, liver, lung, and skin (Figure 1 and Table 2). We have had to omit many excellent studies because of limitations of space and apologize to authors whose work has not been cited.

image

Figure 1. A schematic summary of studies implicating specific miRs in regulation of fibrosis in the kidney, heart, liver, lung, and skin.

Download figure to PowerPoint

Table 2. MicroRNAs implicated in fibrosis. (NB. Not all targets are experimentally validated and thus may be indirect)
MicroRNATissueSpeciesEffectTargetReference
let-7dLungHumanAnti-fibroticHMGA2 [129-132]
miR-1HeartRodentProfibrotic [107]
miR-15b, -16LiverRodentAnti-fibroticBcl2, Cyclin D1 [122, 123]
miR-19bLiverRodent, humanAnti-fibroticTGFRII[120]
miR-21Heart, lung, skin, kidneyRodent, humanProfibroticSpry1, PTEN, Smad7, PPARα [96-99, 107-109, 125, 134]
 KidneyRodentAnti-fibroticPTEN[100]
miR-24HeartRodent, humanAnti-fibroticFurin[113]
miR-29Heart, liver, lung, skin, kidneyHuman, rodentAnti-fibroticCollagen, MMP, Fos [60, 61, 101-103, 112, 121, 126, 129, 133, 134]
miR-29cKidneyRodentProfibroticSpry1[104]
miR-30HeartRodentAnti-fibroticCollagen, Ctgf[110]
miR-101aHeartRodentAnti-fibroticFos[114]
miR-133aHeartHuman, rodentAnti-fibroticCollagen, Ctgf [110, 111]
miR-145SkinHumanAnti-fibrotic [134]
miR-146aLiverRodentAnti-fibroticSmad4[124]
miR-150LiverRodentAnti-fibroticMyb[49]
miR-155LungHumanProfibroticFGF7[127]
miR-192KidneyRodent, humanAnti-fibroticZEB1, ZEB2 [54, 55]
 Kidney (glomerular / mesangial cell only)RodentProfibroticZeb1, Zeb2 [89-92]
miR-194LiverRodentAnti-fibroticRac1[49]
miR-196aSkinHumanAnti-fibroticCollagen[135]
miR-200 familyLung, kidneyRodent, humanAnti-fibroticZeb1, ZEB2 [45-47, 87, 88, 128]
 Kidney

(glomerular / mesangial cell only)

RodentProfibroticZeb1, Zeb2[89]
miR-355Liver Anti-fibrotic [119]

In considering changes in expression pattern from these studies that may be a consequence of TGF-β1 and fibrosis, it is important to be aware of some limitations to current techniques. Restricted access to truly normal control tissue is a major limitation: studies of human tissue are often reliant on macroscopically normal tissue obtained, for example, at surgery for removal of a solid organ tumour. Such tissue may contain unrecognized alterations in miR profile, and may additionally be subject to changes in miR expression during the surgical procedure. A distinctive pattern of miR response to hypoxia is present in many cells [73], and even short periods of interruption to blood supply during surgery, leading to warm ischaemia, may thus alter basal miR profile. Another important limitation is that miR expression studies in fibrosis have to date considered expression at the whole organ level, while these tissues represent a complex admixture of cells of highly different phenotypes and functions. A key future development may be the ability to refine techniques such that it is standard to analyse miR expression in single cells, or populations of similar phenotype, isolated from whole organs.

Kidney

The varied and emerging functions of miRs in the kidney have been the subject of recent, excellent reviews [74-77]. Conditional deletion of Dicer and Drosha, key elements of the miR processing machinery, has been used in studies of the overall importance of miRs in renal physiology and development. Targeted disruption of miR processing in this fashion to specialized cells of the kidney including the podocyte and the juxtaglomerular cells leads to failure of the specialized phenotype of these cells, culminating in glomerulosclerosis and renal failure in the case of the podocyte [78-81], and hypotension and renal fibrosis in the case of the juxtaglomerular cells [82]. Interestingly, in contrast, deletion of Dicer in proximal tubular cells has no apparent ill effects, but protects the kidney against ischaemia–reperfusion injury [83]. The lack of a phenotype in this model likely reflects PEPCK-driven recombination and dicer deletion occurring after renal development had completed. Conditional dicer deletion in nephron progenitors prevents renal development, emphasizing the importance of miRs in nephrogenesis [84, 85]. Combined analysis of miR and protein expression patterns in renal cortex and medulla identified miRs differentially highly expressed in kidney cortex and in medulla, and demonstrated reciprocal expression of predicted targets [86].

As described above, miR-192 and miR-200 families may play a core role in the maintenance of epithelial phenotype, and loss of their expression is linked to EMT in cancer and fibrosis. Both families are highly expressed in the kidney. Decreased expression of miR-192 is seen in advanced diabetic nephropathy and correlates with extent of renal fibrosis [55]. This likely reflects the role of miR-192 in the maintenance of epithelial phenotype in tubular epithelial cells [55]. Renal expression of miR-200 family members is decreased in mouse models of renal scarring, and by TGF-β1 in rat proximal tubular cells [87]. Injection of miR-200b precursor antagonizes ZEB1 and ZEB2 induction and collagen synthesis, and ameliorates fibrosis in this model [88]. However, in the mesangial cells of the glomerulus, ZEB1 and ZEB2 have important anti-fibrotic roles, and the miR-192 and −200 families have profibrotic actions. MiR-192 expression is increased in mouse glomeruli in the early phase of renal injury, secondary to increased expression in mesangial cells, where miR-192 facilitates profibrogenic effects including collagen synthesis [89-91]. Inhibition of miR-192 attenuates proteinuria and attenuates matrix deposition in nephropathy secondary to streptozotocin-induced diabetes in the mouse [92]. The miR-200 family also represses ZEB1 and ZEB2 and promotes matrix deposition in mesangial cells [89]. We have recently reviewed the pleiotropy of miR-192 action in the kidney elsewhere [93]. Beyond their role in disease, it is emerging that these miRs play important roles in renal physiology. MiR-192 and -200b are central to the distal nephron's handling of sodium [94] and tonicity [95].

MiR-21 is identified in other organs, notably the heart, as a key profibrotic miR and there is evidence supporting a profibrotic role for miR-21 also in the kidney [96-99]. However, again there are important differences in the glomerulus, with decreased miR-21 expression in the glomerulus of diabetic mice and enforced mesangial cell miR-21 expression protective in this context [100].

The miR-29 family is highly expressed in the kidney, and miR-29a is down-regulated by TGF-β1 in proximal tubular cells in vitro, facilitating expression of its targets Col4a1 and Col4a2, favouring enhanced matrix deposition [101]. MiR-29b is up-regulated in the medulla of normal rats by a high salt diet, but not in Dahl salt-sensitive rats, which sustain salt-induced hypertension and renal injury [102]. MiR-29b targets multiple collagens and MMPs, suggesting that failure to up-regulate miR-29b facilitates matrix accumulation in salt-induced renal injury [102]. Unilateral ureteric obstruction (UUO) is associated with decreased miR-29 expression in the mouse kidney, while in mouse embryonic fibroblasts in vitro, TGF-β1-dependent Smad3 nuclear accumulation leads to Smad3 binding to the miR-29 promoter and transcriptional repression of this miR [103]. Smad3-deficient mice retain miR-29 at control levels following UUO and are protected from renal fibrosis [103]. Interestingly, however, miR-29c is increased in the glomeruli of db/db mice, and miR-29c antisense oligonucleotide decreased albuminuria and matrix accumulation in db/db mice [104]. MiR-29c induces podocyte apoptosis and extracellular matrix accumulation, involving Spry1 repression facilitating increased Rho kinase activity [104].

Heart

Overwhelming evidence implicates miRs in cardiac development and pathology. Conditional knockout of Dicer in mouse myocardium results in hypertrophy, ventricular fibrosis [105], dilated cardiomyopathy, and eventual heart failure [106]. Microarray analysis identified changes in miR expression in mouse cardiac hypertrophy, particularly the induction of miR-21 and repression of miR-1 [107]. MiR-21 is selectively expressed in cardiac fibroblasts and increased expression is detected in heart failure [108]. MiR-21 targets Spry1, leading to enhanced fibroblast proliferation and survival [108]. MiR-21 also targets the negative regulator of phospho-inositol 3-kinase (PI3K), Pten, in a mouse model of ischaemia–reperfusion injury in which increased PI3K signalling induced extracellular matrix remodelling [109]. MiR-133a and miR-30 have also been implicated in myocardial extracellular matrix remodelling in a rodent model; these two miRs are down-regulated in ventricular hypertrophy and regulate connective tissue growth factor (Ctgf), and subsequently collagen production [110, 111].

Several miRs have been identified as dysregulated in myocardial infarction, including the down-regulation of miR-29 [112], miR-24 [113], and miR-101a [114]. Several potential targets of miR-29 have been identified including collagen [112]. MiR-24 was identified to target Furin transcripts encoding a protease involved in activation of latent TGF-β1 [113]. In vivo lentiviral expression of miR-24 improved heart function and reduced fibrosis in myocardial infarction [113]. MiR-101 targets transcripts of Fos, encoding the transcription factor C-Fos, and suppresses proliferation and extracellular matrix deposition by cardiac fibroblasts [114]. In vivo adenovirus expression of miR-101a improved cardiac function in a rodent model of myocardial infarction [114].

Liver

Liver fibrosis occurs in most chronic liver diseases and the main instigator is myofibroblastic differentiation of hepatic stellate cells (HSCs) [115]. Microarray analysis identified numerous miRs dysregulated in both murine models of liver fibrosis and human samples, and several demonstrated parallel expression profiles including miR-199a, -200a, and -200b [116]. Changes in miR expression observed in chronic liver injury due to hepatitis and in liver cancer are beyond the scope of this article and have recently been reviewed [117]. Changes in miR expression are linked to HSC myofibroblastic differentiation, apoptosis, and migration [118, 119]. In vitro overexpression of miR-19b, -29, -150, -194, and -355 inhibited cell migration and reduced both the myofibroblastic marker α-SMA and collagen [49, 61, 119-121]. The TGF-β receptor II was validated as a target of miR-19b [120] and C-Fos as a putative target of miR-29b [121]. It is hypothesized that TGF-β1 and NF-kappaB-dependent down-regulation of miR-29 family members in HSCs results in myofibroblastic differentiation and extracellular matrix deposition [61].

Resolution of fibrosis involves the apoptosis of myofibroblastic cells. In vitro overexpression of miR-15b and -16 inhibited proliferation and induced apoptosis via the de-repression of Bcl-2 and reduction in cyclin D1, and subsequent activation of caspases 3, 8, and 9 [122, 123]. In vitro overexpression of miR-146a suppresses TGF-β1-induced proliferation and increased apoptosis, via the targeting of Smad4 [124].

Lung

The bleomycin-induced mouse model of pulmonary fibrosis identified dysregulation of several miRs, including miR-21 [125], -29 [60, 126], -155 [127], and -200 [128]. MiR-21 was up-regulated in both experimental and idiopathic pulmonary fibrosis, primarily in myofibroblastic cells [125]. MiR-21 was postulated to target the inhibitory Smad, Smad7, thus enhancing TGF-β1 signalling [125]. In vivo antagonism of miR-21 decreased the severity of experimental fibrosis in mice [125]. The down-regulation of miR-29, via TGF-β1/Smad3 signalling [60, 126], was correlated with changes in extracellular matrix deposition and remodelling [60]. MiR-29 is also down-regulated in human idiopathic pulmonary fibrosis [129]. Up-regulation of miR-155 correlates with the extent of pulmonary fibrosis, which it may promote via targeting of keratinocyte growth factor (FGF7) [127]. MiR-200 family members were down-regulated in both experimental and idiopathic pulmonary fibrosis [128], in keeping with their key role in EMT described above. Also down-regulated in idiopathic pulmonary fibrosis is let-7d, via TGF-β1/Smad3 signalling [129]; let-7d targets transcripts of HMGA2 (high-mobility group AT-hook 2) [130], an effector of TGF-β1-mediated EMT [131, 132].

Skin

The dysregulation of miRs has also been identified in fibroproliferative diseases of the skin including scleroderma and keloids. MiR-29 was down-regulated in scleroderma biopsy specimens, explanted fibroblasts, and bleomycin-induced skin fibrosis [133]. Independently, the down-regulation of miR-29 was confirmed along with miR-145 and the up-regulation of miR-21 [134]. MiR-29 targets collagen [133, 134] and was shown to be down-regulated in response to TGF-β1 and PDGF [133]. MiR-196a was down-regulated in explanted keloid fibroblasts and it was shown to directly target transcripts of COL1A1 and COL3A1 [135].

Conclusions and future perspectives

  1. Top of page
  2. Abstract
  3. Introduction
  4. Mechanisms of miR-mediated regulation
  5. Transforming growth factor beta-1
  6. Transforming growth factor beta-1 synthesis: importance of post-transcriptional regulation
  7. MiRs as effectors of key transitions in cellular phenotype in tissue fibrosis
  8. Tissues
  9. Conclusions and future perspectives
  10. Acknowledgments
  11. Author contribution statement
  12. References

There is increasing understanding of the mechanisms of miR-mediated regulation and their role in tissue fibrosis. Interestingly, some miRs are emerging that appear to have a pivotal role in regulating fibrosis in multiple tissues. Two notable examples are miR-21 [96-100, 107-109, 125, 134] and miR-29 [60, 61, 101-103, 112, 121, 126, 129, 133]. In contrast, the role of other miRs appears highly context-specific, with differing importance and effects in different tissues and cells. Relevant examples include miR-192 in the kidney [54, 55, 89-92], miR-24 in the heart [113], miR-19b in the liver [120], and let-7d in the lung [129, 130].

At present, there are few effective treatment options for tissue fibrosis, and options for the monitoring of progression of fibroproliferative diseases are often limited, outside of repeated invasive tissue sampling by biopsy. This lack of good biomarkers limits the ease with which potential new therapies may be tested. MiRs show promise as biomarkers, due to their stability in cells and circulation [136], and ease of quantification in various body fluids such as serum, plasma, and urine [137-140].

Manipulation of specific miRs in target tissues also holds promise as a new approach to therapy. Many in vivo experiments, some of which have been highlighted in this review, demonstrate the potential of antisense oligonucleotides to antagonize miRs [92, 104, 112, 125] and the use of viral-expression systems to overexpress miRs [100, 113, 114, 120, 121]. To this end, the tissue specificity of a particular miR can be exploited, as exemplified by the role of miR-122 in progression of viral hepatitis. MiR-122 is liver-specific, constitutes ∼70% of all miR transcripts in the liver [141], and may aid viral replication by enhancing ribosome occupation of viral RNA and thus translation [142]. This has formed the basis of a successful preclinical trial of miR-122 antagonism to block hepatitis viral replication [143].

Such miR therapeutics are in their infancy, with more research required to understand the mechanisms of action and roles in pathology, along with better target prediction and validation to understand the downstream pathways. However, the increasing understanding of the role of miRs in diverse disease states has brought new insights into these disorders, together with the promise of new treatment approaches for a class of conditions responsible for much human suffering, that are at present very hard to treat.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Mechanisms of miR-mediated regulation
  5. Transforming growth factor beta-1
  6. Transforming growth factor beta-1 synthesis: importance of post-transcriptional regulation
  7. MiRs as effectors of key transitions in cellular phenotype in tissue fibrosis
  8. Tissues
  9. Conclusions and future perspectives
  10. Acknowledgments
  11. Author contribution statement
  12. References