MicroRNAs and the Glomerulus

Abstract

microRNAs (miRNAs) are short non-coding RNAs regulating gene expression at the post-transcriptional level by blocking translation or promoting cleavage of their target mRNAs. Increasing evidence shows that miRNAs play central roles in gene transcription, signal transduction and pathogenesis of human diseases. Diabetic nephropathy (DN) is a severe microvascular complication that can lead to end-stage renal disease. Increased expansion (hypertrophy) and accumulation of extracellular matrix (ECM) proteins such as collagen (fibrosis) in the glomerular mesangium along with glomerular podocyte dysfunction are major features of DN. Profiling of miRNAs and study of their functions in renal glomeruli can provide critical new information to advance our knowledge of DN as well as other kidney diseases and thereby uncover much needed new therapeutic targets. In this review, we summarize the biogenesis of miRNAs and their functions in the glomerulus, with particular emphasis on glomerular mesangial cells and podocytes related to the pathogenesis of DN.

Introduction

The kidney is a relatively small but very important organ due to its innumerable functions in various critical physiological processes including blood pressure regulation, maintenance of acid-base and electrolyte balance, and filtration of the blood to remove wastes. As such, renal dysfunction can lead to several diseases including diabetic nephropathy (DN), a debilitating microvascular complication of diabetes that can result in end-stage renal disease. Furthermore, patients with DN are highly susceptible to macrovascular diseases such as atherosclerosis, hypertension and stroke and there is a close correlation between chronic kidney disease (CKD) and cardiovascular disease (CVD) [1].

The kidneys are bean-shaped organs (Figure1A) with each being surrounded by a transparent fibrous renal capsule. There are two key regions in the kidney, the outer renal cortex and the inner medulla. Nephrons, the urine producing structural and functional units of the kidneys, span the cortex and medulla. Each nephron consists of a renal corpuscle composed of a capillary tuft (the glomerulus), surrounded by the glomerular (Bowman’s) capsule and a tubule. Glomeruli are located in the cortex and are the primary sites of filtration (Figure 1B). Blood enters into the glomerulus via the afferent arteriole, circulates inside and exits via the efferent arteriole (Figure 1C). Glomeruli are connected to the proximal tubule which eventually leads to the collecting tubule via the distal tubule and many loops and turns. Podocytes or visceral epithelial cells wrap around the glomerular capillaries and play critical roles in the regulation of filtration. Podocyte dysfunction can lead to proteinuria as seen in DN.

Figure 1. Functions of miRNAs in glomeruli related to the progression of diabetic nephropathy.

A,B,D,E) Representative examples of Hematoxylin and Eosin (H&E) staining of mouse kidney (C57/Bl6 strain) sections. A) Low magnification picture of whole kidney (courtesy of Dr. Ivan Todorov). Scale bar indicates 1mm. B) High magnification power picture of kidney cortex section. Arrows indicate glomeruli. Scale bar indicates 100μm. C) Schematic structure of the glomerulus. Mesangial cells, Podocytes and Endothelial cells are shown in blue, green and gray, respectively. D & E) Higher power representative picture of a glomerulus from normal non-diabetic mouse and from streptozotocin-injected mouse (model of type 1 diabetes), respectively. Scale bar indicates 20μm. F & G) Signal transduction pathways mediated by miRNAs in mesangial cells and podocytes related to the progression of DN. Please see main text for details.

Hypertrophy and expansion in the glomerular mesangium and tubular compartments, along with podocyte dysfunction and accumulation of ECM proteins (fibrosis), are major features of DN, and contribute to renal failure in diabetes mellitus [2-3]. Enhanced expression of Transforming growth factor-beta1 (TGF-β) in renal cells plays a key role in glomerular mesangial hypertrophy and fibrosis under diabetic conditions by inducing the expression of ECM proteins such as collagen and fibronectin [2-9]. Although an antibody against TGF-β was effective to some extent in animal models of DN [10] and some trials targeting TGF-β itself and with antifibrotic drugs are ongoing [11], such approaches could have limitations due to the multifunctional nature of TGF-β. Furthermore, while several studies have investigated the signaling mechanisms leading to DN [2], the molecular mechanisms are still not well understood. A further investigation of molecular events downstream of TGF-β signaling in renal cells can uncover new targets to more effectively prevent or treat DN.

Interesting reports showed that mice with podocyte-specific deletion of Dicer or Drosha, key enzymes involved in microRNA (miRNA) biogenesis, display progressive glomerular and tubular damage along with proteinuria and other podocyte defects [12-15]. Mice with Dicer deletion in the entire nephron lineage showed defects in kidney development [16-17]. miRNAs function in kidney development even in frogs [18]. Furthermore, of particular interest are reports demonstrating that a cluster of miRNAs are highly expressed in the kidney, and that there are key differences in the miRNA expression profile in renal cortex versus medulla [19-20]. In recent years, the functional roles of key miRNAs in glomerular cells have been extensively studied in TGF-β actions and diabetic kidney disease [5, 8, 21-26]. These results suggest that specific miRNAs in glomerular cells such as podocytes, mesangial and endothelial cells, as well as other renal cells may have critical biological functions and could therefore be exploited as novel targets for kidney diseases.

microRNAs

miRNAs are short (~21 nucleotides) non-coding RNAs that regulate gene expression generally by translational inhibition or cleavage of their target RNAs (Figure 2)[27-28]. miRNAs are initially transcribed as long primary miRNAs (Pri-miRNAs) which are processed to a stem-loop (hairpin) structure fragment termed precursor-miR (Pre-miRNA) in the nucleus by an RNase III enzyme, Drosha, in collaboration with the double strand RNA binding protein, DGCR8[27-28]. For certain miRNAs, this process can also be enhanced by Smads in response to TGF-β or bone morphogenetic proteins (BMPs) [29-30], or by p53[31]. Pre-miRNAs (~70 nucleotides) are then exported to the cytoplasm by Exportin-5 and further cleaved to miRNA duplexes by another RNase III family enzyme, Dicer [27-28]. The miRNA duplexes are then unwound and one strand, termed “mature miRNA” or guide strand, which has complementarity to target mRNAs, is loaded into the RNA-induced silencing complex (RISC) which contains Argonaute 2 (Ago2), Dicer and TRBP proteins [27-28]. If the complementarity of miRNAs to target RNAs is perfect, RISC cleaves the target mRNA (classical RNA interference, RNAi). However, most miRNAs and their target RNAs have some mismatches and RISC induces translational repression of target genes by hybridizing of miRNAs to their 3′ untranslated region (UTR) [27-28]. Although complete complementarity is not required for miRNA-mediated regulation of a target transcript, the “seed sequence”, namely seven key nucleotides of the miRNA 5′ sequence, is critical. miRNAs inhibit the initiation and elongation steps of translation to reduce protein expression [27-28]. They can also repress gene expression by sequestering targeted mRNAs to cytoplasmic mRNA processing bodies (P-bodies) for degradation. In addition to their role in such post-transcriptional repression, miRNAs are also implicated in transcriptional gene silencing by targeting the promoter regions [32]. Therefore, miRNAs regulate gene expression through translational repression, target mRNA degradation or transcriptional inhibition. There is increasing interest in miRNAs since they can affect more than 60% of protein-coding genes [28]. Furthermore, evidence shows that they are dysregulated in various diseases and affect the expression of several disease related genes. Moreover, these small RNAs have now been identified in plasma and urine of human patients and animal models [33-35]. Thus miRNAs are attractive candidate biomarkers and therapeutic targets for human diseases such as DN.

Figure 2. Biogenesis of microRNAs and their functions.

MicroRNA transcripts initially originate as primary miRNAs (Pri-miRNAs) that are then processed into Pre-miRNAs by the Drosha enzyme, which are further cleaved to result in double-strand RNA duplexes. The miRNA duplexes are then unwound by the action of a second enzyme, Dicer, and the mature miRNA guide strand is loaded into the RISC complex. miRNAs in the RISC complex then guide the recognition of target RNAs to induce their downregulation depending on the type of complementarity. Please see main text for details. RISC, RNA-induced silencing complex; DGCR8, DiGeorge syndrome critical region gene 8, an essential cofactor for Drosha; TRBP, transactivating response RNA-binding protein, a cofactor for Dicer; UTR, untranslated region; P-body, processing body. Adapted from Kato et al. [5, 21].

miRNAs regulating fibrosis in glomerular mesangial cells and other renal cells

Among the miRNAs highly expressed in the kidney [19-20], several key miRNAs (miR-192, miR-200b, miR-200c, miR-216a and miR-217) were found to be higher in renal glomeruli of mouse models of diabetes [type 1(streptozotocin (STZ)-induced) and type2 (db/db)] compared to the corresponding controls[5, 8, 22-23, 25-26]. Some or all of these miRNAs were also increased by TGF-β or high glucose (HG) in mouse mesangial cells (MMC) and in human MC [22-26]. Although in vitro treatments do not fully mimic DN in vivo, similarities in the profile of induced miRNAs have been noted most likely because diabetic stimuli and factors associated with DN like HG, growth factors and advanced glycation end products can induce TGF-β in renal cells. Furthermore, increased expression of these miRNAs have also been reported in kidneys from patients with hypertensive nephrosclerosis, IgA nephropathy and lupus nephritis[36-38], as well as some other animal models of kidney injury[22, 39-41].

Interestingly, miR-192 targets Zeb1 and Zeb2 which are now widely recognized as general E-box repressors [42-44], and increases Collagen type I alpha2 (Col1a2) gene expression in MMC, demonstrating that increased miR-192 in diabetic conditions induces fibrosis by inhibiting E-box repressors (Figure 1F)[8]. Other MC miRNAs have also been implicated in renal fibrosis associated with DN. miR-377 was shown to induce fibronectin (ECM protein) expression via downregulation of manganese superoxide dismutase and p21-activated kinase in MCs(Figure 1F)[25]. Upregulation of miR-21 by diabetic conditions or TGF-β was also associated with MC and renal fibrosis or injury [45-48], although a report showed that miR-21 levels were decreased in glomeruli from db/db mice [49]. Down-regulation of the miR-29 family, which targets various collagens in multiple cell types, was reported to induce fibrosis in hypertensive and other renal injury models (Figure 1F)[50-52].

The miR-200 family members are separated into two clusters based on genomic structures. They are located in the introns of their host RNAs and also regulated by the miR-192 targets Zeb1/2 through E-boxes in the promoter of the host gene [23, 53-54]. miR-200 family also targets Zeb1/2 at their upstream E-boxes to auto-upregulate their own expression and can thereby accelerate the signaling pathways leading to collagen expression and renal fibrosis[23].

In MC, diabetic conditions induce TGF-β1 through several transcriptional mechanisms including binding of E-box activators, Upstream Stimulatory Factors USFs, to an E-box element in the TGF-β1 promoter [23, 55-57]. Interestingly, TGF-β1 gene itself is upregulated by TGF-β1 in MMC through the same promoter E-box element and recent evidence shows this autoregulation is also mediated by the E-box repressors Zeb1/2 targeted by the miRNA circuit, miR-192 to miR-200b/c [23]. miR-200a/c were reported to be induced by TGF-β or BMPs in pulmonary smooth muscle cells[29]. miR-200b was upregulated not only in glomeruli from db/db mice but also in endothelial cells and podocytes treated with HG [22-23]. miR-200 family was reported to be both increased and decreased in Unilateral ureter obstruction (UUO) mouse model of kidney fibrosis[51, 58]. Interestingly, Wang et al. reported that in mouse DN and tubular epithelial cells treated with TGF-β there is a down-regulation of miR-200a which targets TGF-β receptor 2, suggesting another accelerating signaling loop via miR-200a [59]. These studies also demonstrate key cell-specific differences in miRNA-200 responses under disease states.

miR-216a and miR-217 are located in the second intron of the mouse non-coding RNA RP23-298H6.1-001(RP23) and their expression depends on this host non-coding RNA[5]. RP23 expression is regulated by the E-box cluster in its promoter and by miR-192 through Zeb1/2, suggesting signal amplification via miRNAs (Figure 1F)[5]. miR-216a can also increase collagen expression by another parallel mechanism involving the inhibition of Ybx1, an RNA binding protein and a component of P-bodies [26]. Taken together, miRNA circuits initiated by miR-192 can amplify and fine-tune downstream pathways to further augment collagen expression and mesangial fibrosis related to the pathogenesis of DN [5, 8, 23, 26].

miRNAs in glomerular mesangial cell hypertrophy

Renal glomerular expansion (hypertrophy) is a major feature of DN and contributes to kidney failure. Figure 1 D&E show representative examples demonstrating that glomeruli from diabetic mice are significantly larger than those from non-diabetic animals. Akt kinase activation is related to cancer development and diabetes by regulating several downstream cellular targets such as mTOR, GSK-3β and FoxO proteins [60]. Since these downstream targets are major regulators of protein synthesis, apoptosis and cell survival, Akt activation is a major mediator of cellular hypertrophy under disease conditions. One report suggested that TGF-β activates phosphatidylinositol-3-kinase (PI3K)/Akt pathway via direct physical interaction of the TGF-β receptor and PI3K [61]. However, the mechanisms, especially long-term chronic activation of Akt by TGF-β were not fully understood.

As described above, two miRNAs (miR-216a and miR-217) were upregulated in MMC treated with TGF-β and in glomeruli from diabetic mice. Interestingly, both these miRNAs target phosphatase and tensin homologue (PTEN), an upstream inhibitor of Akt, demonstrating that these two miRNAs upregulated by TGF-β can activate Akt and downstream signaling through inhibition of PTEN (Figure 1F)[5]. This is also supported by data showing that miR-21 targets PTEN, and thereby contributes to renal pathology [46]. Thus, Akt activation by miRNA circuits induced by TGF-β could be a novel mechanism for mesangial hypertrophy in diabetes. In addition, the miR-200 family has been reported to control PI3K signaling by targeting Fog2, an inhibitor of PI3K[62]. As PI3K/Akt activation in diabetic kidney cells induces hypertrophy and fibrosis [4-6, 26], upregulation of miR-200 family members may also activate Akt through inhibition of Fog2 to promote mesangial hypertrophy.

Interestingly, p53 and miR-192 regulate each other to form another signaling loop [63-65]. p53 regulates multiple genes related to cell cycle, such as p21 (cyclin-dependent kinase inhibitor), suggesting that inhibition of cell cycle progression by miRNA actions may induce hypertrophy in glomeruli. Since Eukaryotic translation Elongation factor 1-alpha 1 (EF1α) is also upregulated by p53[66], increased levels of EF1α might enhance protein synthesis and hypertrophy in the diabetic glomeruli. Furthermore, recent reports showed that miR-200 family is also regulated by p53 [67-68].

Therefore, these miRNA-mediated amplifying circuits encompassing miR-192, miR-200 family, miR-216a and miR-217 in conjunction with other miRNAs can serve as signal modulators to accelerate fibrosis and hypertrophy in TGF-β-mediated chronic diseases like DN (Figure 1F).

miRNAs in glomerular endothelial cells and podocytes

As noted earlier, mice with podocyte-specific deletion of Drosha or Dicer displayed severe renal phenotype and proteinuria[12-15], suggesting a significant role for miRNAs in podocytes. It was reported that miR-93 levels were lower in glomeruli obtained from diabetic db/db mice relative to control mice, and also in HG treated podocytes and renal microvascular endothelial cells. This corresponded to increases in Vascular Endothelial Growth Factor-A, a target of miR-93 (Figure 1G)[24]. miR-200b and miR-29c were upregulated along with miR-192 in glomeruli from db/db mice, as well as in endothelial cells and podocytes treated with HG. MiR-29c activates Rho kinase by targeting Spry1 and contributes to DN (enhanced ECM accumulation and podocyte apoptosis)[22]. db/db mice injected with 2′-O-methyl antisense oligonucleotides targeting miR-29c showed reduction in the progression of DN [22].

Regulation of miR-192 by TGF-β and in other models of kidney injury

There is much interest in evaluating the processes regulating the expression of miRNAs under disease conditions. Two major mechanisms may be involved in TGF-β induced upregulation of miR-192. One is the classical mechanism through Smad3[40] and another involving p53[63-65], since the miR-192 promoter has Smad and p53 binding sequences. Subsequently, miR-192 can in turn induce miR-200b/c probably through E-boxes in their promoters [23, 53-54].

However, it is well known that TGF-β effects are cell-type specific and context dependent. Evidence shows that the expression of miR-192 and miR-200 family members are decreased in cancer or kidney-derived epithelial cell lines treated with TGF-β[54, 69-70]. In the NRK52E epithelial cell line, one report showed that miR-192 levels were increased at 1-24 hr [40], while another showed that miR-192 levels were decreased by TGF-β at 3 days[71]. It is possible that there is a biphasic regulation in which miR-192 and 200 family may be induced by TGF-β1 at early time points (6-24 hours) but downregulated later, although such a regulation in vivo in DN is more difficult to envision. Since accumulation of ECM proteins in fibrotic events in MC and tubular cells are well-established features of DN [2], the upregulation of collagen by TGF-β1 via increase in miR-192 and miR-200b in MC could be a key feature of DN and potentially more relevant during the early stages of diabetic kidney injury. Furthermore, diabetes affects multiple cells in glomeruli as well as other renal compartments and can regulate a spectrum of miRNAs with cell-type specific targets and responses related to DN. Clearly, such miRNA profiles can vary in other models of renal injury.

As described above, miR-192 is regulated by Smad3 and p53 [40, 64-65]. Mutations in p53 and Smad genes have been observed in cancer cell lines and other immortalized cell lines. Mutant p53 changes the response to TGF-β via Smad interaction[72]. Kidney injury and fibrosis were ameliorated in p53 or Smad3 knockout mice[40, 73-74]. Recent papers also showed that p53 is a major regulator of epithelial to mesenchymal transition in cancer through miR-192 and miR-200 family [67-68]. Therefore, p53 and Smad genes may be key factors dictating TGF-β responses and hence the effects of TGF-β on miR-192, miR-200 family and other miRNAs in primary MC may be different from those reported in cancer or immortalized kidney cell lines. This might account for some of the cell-type discrepancies in TGF-β1 responses, as well as for the disease specific effects of miRNAs.

A recent report also showed that low-dose paclitaxel, a cancer drug, could prevent renal fibrosis in the remnant kidney model by down-regulating miR-192 [41]. In addition, in models of ischemia reperfusion, better survival was noted in tubular specific Dicer knockout mice that had lower levels of miR-192, and lower miR-192 was also noted in other studies [75-76]. Although there are both similarities and discrepancies in miRNA profiles observed in various models of renal injury, inhibition of miR-192 seems beneficial in several models as mentioned above [41, 75].

miRNAs as biomarkers and therapeutic targets

The rapid recent technological advances in miRNA detection and quantification, including microarrays, quantitative PCRs and next generation sequencing, have spurred tremendous interest in developing miRNAs as therapeutic targets and potential biomarkers for human diseases. Recent reports showed that circulating miRNAs in blood are sensitive biomarkers for cancer, tissue injury and heart failure[33-34]. Levels of miRNAs in the urinary sediment of patients with IgA nephropathy have been examined [35]. These results support the notion that profiling circulating miRNAs in biofluids could lead to the identification of sensitive and precise diagnostic biomarkers of various human diseases in general and renal disorders in particular. Since miRNAs can be quantitatively detected in urine and plasma, they could serve as new biomarkers for the early detection of diabetic renal injury, a major goal for the clinical nephrologist. Clinical trials with anti-miRNAs for hepatitis C and other diseases are already being developed [77-78]. Thus, it is possible that anti-miRNA therapies to target key circulating or locally expressed miRNAs could also be developed for the treatment of acute and chronic renal disorders. Since miR-192 can control several renal miRNAs related to DN, it is a good candidate target to evaluate for DN treatment. Recently, specific and efficient reduction of miR-192 was observed in vivo in normal mice injected with locked nucleic acid (LNA)-modified antimiR-192 (LNA-antimiR-192)[5]. LNA-antimiR-192 also reduced downstream miRNAs (miR-216a, miR-217 and miR-200 family) and functional indices of renal fibrosis and hypertrophy, such as collagens and TGF-β1 expression and Akt activation in these mice, similar to the effects in cultured MMC[5, 23]. Apart from miR-192, targeted inhibition or overexpression of other key renal miRNAs such as miR-200 family members, miR-29c, miR-29b or miR-21 have shown promise in in vitro and in vivo models of DN and other renal diseases, as described earlier. Given the tremendous activity in this field in recent years, several other renal miRNAs are expected to be identified in the near future that can also be evaluated as diagnostic biomarkers or potential therapeutic targets.

Closing Remarks

Renal glomeruli are relatively small structures but are powerhouses with critical roles in kidney function. Emerging evidence shows that even tinier molecules like miRNAs can play significant modulatory roles in the biology and pathology of glomerular function and renal pathogenesis. Major future advances are anticipated in this field including the identification of additional miRNA modulators of various renal functions, as well as development of miRNA based biomarkers and therapies for various renal diseases.