MicroRNA-192 in diabetic kidney glomeruli and its function in TGF-β-induced collagen expression via inhibition of E-box repressors

Abstract

Key features of diabetic nephropathy (DN) include the accumulation of extracellular matrix proteins such as collagen 1-α 1 and -2 (Col1a1 and -2). Transforming growth factor β1 (TGF-β), a key regulator of these extracellular matrix genes, is increased in mesangial cells (MC) in DN. By microarray profiling, we noted that TGF-β increased Col1a2 mRNA in mouse MC (MMC) but also decreased mRNA levels of an E-box repressor, δEF1. TGF-β treatment or short hairpin RNAs targeting δEF1 increased enhancer activity of upstream E-box elements in the Col1a2 gene. TGF-β also decreased the expression of Smad-interacting protein 1 (SIP1), another E-box repressor similar to δEF1. Interestingly, we noted that SIP1 is a target of microRNA-192 (miR-192), a key miR highly expressed in the kidney. miR-192 levels also were increased by TGF-β in MMC. TGF-β treatment or transfection with miR-192 decreased endogenous SIP1 expression as well as reporter activity of a SIP1 3′ UTR-containing luciferase construct in MMC. Conversely, a miR-192 inhibitor enhanced the luciferase activity, confirming SIP1 to be a miR-192 target. Furthermore, miR-192 synergized with δEF1 short hairpin RNAs to increase Col1a2 E-box-luc activity. Importantly, the in vivo relevance was noted by the observation that miR-192 levels were enhanced significantly in glomeruli isolated from streptozotocin-injected diabetic mice as well as diabetic db/db mice relative to corresponding nondiabetic controls, in parallel with increased TGF-β and Col1a2 levels. These results uncover a role for miRs in the kidney and DN in controlling TGF-β-induced Col1a2 expression by down-regulating E-box repressors.

Diabetic nephropathy (DN) is the most common cause of kidney failure in patients with diabetes mellitus. The major characteristics of DN include glomerular basement-membrane thickening, mesangial expansion and hypertrophy, and an accumulation of extracellular matrix (ECM) proteins (1). Evidence shows that transforming growth factor β1 (TGF-β) levels are increased under diabetic conditions in renal cells, including mesangial cells (MC), can up-regulate ECM proteins such as collagens (2, 3), and also can promote MC survival and oxidant stress (4).

To date, Smad transcription factors have been shown to be the major effectors of TGF-β signaling (5, 6). Collagen 1-α 1 and -2 (Col1a1 and -2) and other ECM genes are regulated in MC by TGF-β via Smads (7, 8). The regulation of collagen by TGF-β in MC also is mediated by mitogen-activated protein kinases (MAPKs) such as p38 and ERKs (9–11). However, the molecular mechanisms by which TGF-β regulates ECM genes still are not understood fully. The collagen gene has E-box elements in the far upstream enhancer region (12, 13). An E-box repressor, δEF1, is a key inhibitor of E-cadherin (14) and E2-box transcription factors such as Nkx2.5 (12). Moreover, it is a known repressor of collagen type 1 and type 2 genes in other cells (12, 13), but its role in MC is not known.

During microarray profiling, we observed that TGF-β leads to a marked down-regulation of δEF1 expression in MC. In this article, we demonstrate that this effect can lead to increased collagen expression via relief of repression at the E-box elements in the collagen gene. Furthermore, we found that TGF-β also down-regulates the expression of Smad-interacting protein 1 (SIP1), another E-box repressor belonging to the same family as δEF1 (15).

MicroRNAs (miRs) are short noncoding RNAs of ≈22 nt that recently have been shown to play important roles in mammalian gene expression (16–19). They induce posttranscriptional gene repression by blocking protein translation (by binding to the 3′ UTR of their target genes) or by inducing mRNA degradation and therefore have the potential to play central roles in physiological and pathological conditions. The field of miRs now is exploding with researchers racing to discover their gene targets and disease relevance. Recent evidence suggests that miRs may regulate the expression of key genes relevant to cancer and potentially other diseases (17, 19). However, to date nothing is known regarding miRs in the context of diabetic complications such as DN. A recent report showed that at least 5 miRs (miR-192, -194, -204, -215, and -216) are highly and quite exclusively expressed in the kidney (20), but their functions or targets are not yet known. In this article, we identify SIP1 as a target of miR-192. We also demonstrate a previously uncharacterized mechanism for TGF-β-mediated collagen regulation involving a cross-talk between E-box repressors (δEF1 and SIP1) and miRs (miR-192) that could be relevant to the pathogenesis of diseases such as DN.

Results and Discussion

Down-Regulation of δEF1 by TGF-β in MMC.

To uncover previously uncharacterized pathways regulated by TGF-β, we profiled genes regulated by TGF-β in mouse MC (MMC) by using TGF-β BMP signaling pathway microarrays (Fig. 1A). Although, as anticipated, TGF-β could increase the expression of key genes associated with the pathogenesis of DN, such as Col1a1, Col1a2, and plasminogen activator inhibitor 1, several other previously unidentified genes with varying expression levels also were detected (Fig. 1 A and B). One of the target genes down-regulated by TGF-β in the microarray was δEF1 (Fig. 1 A, C, and D), a repressor of E-box transcription factors (21). Follow-up verification of the microarray data demonstrated that serum depletion decreased Col1a2 expression but increased δEF1 mRNA levels, and these events were reversed by TGF-β treatment (Fig. 1 B and C). Significant decrease in δEF1 mRNA levels occurred 6 h after TGF-β treatment (Fig. 1C), whereas a significant increase in Col1a2 was observed by 24 h (Fig. 1B). Similar to δEF1 mRNA, δEF1 protein levels also were differentially regulated by serum depletion and TGF-β (Fig. 1D). Because E-boxes are located in the far upstream enhancer region of collagen genes (12, 13), these results suggest that TGF-β may induce collagen expression by decreasing δEF1 expression and thus relieving δEF1-mediated repression at the E-box elements.

Fig. 1.

Profiling genes regulated by TGF-β in MMC. (A) Representative Oligo GEArray profiling results. Intensities of several spots were up-regulated or down-regulated by treating serum-depleted MMC with TGF-β (10 ng/ml for 24 h) (Top). Close-up magnified view of regions containing Col1a2 and δEF1 genes are shown in Middle and Bottom and demonstrate that TGF-β-treated MMC have marked increase in the intensities of Col1a2 spots but decrease in δEF1 spots (each region has four replicate spots). (B and C) Verification of Col1a2 and δEF1 expression levels by real-time qPCR. Col1a2 gene was decreased by serum depletion (SD) relative to cells growing in serum (NT, no treatment), whereas TGF-β treatment significantly increased its expression by 24 h but not at 6 h (B). Conversely, δEF1 expression was significantly increased by serum depletion, and this was reversed by TGF-β, which down-regulated δEF-1 by 6 and 24 h. ∗, P < 0.05; #, P < 0.01. TGF6, TGF-β for 6 h; TGF24, TGF-β for 24 h. (D) Western blot analysis of δEF1 protein levels. Similar to the mRNA data, serum depletion increased δEF1 protein expression, whereas TGF-β (24 h) reversed (decreased) the expression.

δEF1 Negatively Regulates E-Box Enhancer of Col1a2 Gene.

We next performed chromatin immunoprecipitation (ChIP) analysis and confirmed that δEF1 occupancy at the E-boxes in the upstream enhancer region of Col1a2 gene [schematically shown in Fig. 2A, details provided in supporting information (SI) Fig. 6] is markedly enhanced in serum-depleted MMC and is reversed by TGF-β (Fig. 2B). TGF-β treatment significantly decreased δEF1 occupancy at the collagen enhancer in serum-depleted cells even by 6 h, and this effect was sustained up to 24 h.

Fig. 2.

δEF1 negatively regulates the enhancer region of Co1a2 gene. (A) Schematic diagram of the far upstream region of Col1a2 gene. Three E-boxes are noted in the enhancer region located 16-kb upstream of the Col1a2 gene transcription start site. This region was amplified by PCR using primers indicated by arrows. Details of the enhancer sequence are shown in SI Fig. 6. (B) ChIP assay to evaluate δEF1 occupancy at the enhancer region containing E-boxes in the Col1a2 gene. After IP with δEF1 antibody, target ChIPed DNA amplification (with primers shown in A) was noted only in serum-depleted MMC (Upper). Preimmune IgG served as negative control. DNA without IP was used as input control. Serum depletion increased δEF1 expression and its Col1a2 gene enhancer occupancy. Bar graph (Lower) shows δEF1 binding quantified by real-time qPCR. Relative amount was calculated by the ratio of ChIPed DNA/input control DNA. A 6-fold increase in δEF1 occupancy was observed in serum-depleted cells (SD) compared with serum-treated controls (NT). TGF-β treatment gradually decreased the amount of δEF1 at the promoter from 6 to 24 h (TGF6 and TGF24). ∗, P < 0.05; #, P < 0.01. (C) The same enhancer DNA fragment with E-boxes was cloned into the pGL3-promoter driven by SV40 promoter. TGF-β treatment increased luciferase activity (∗, P < 0.05). (D) δEF1 shRNA increased transcription from the Col1a2 promoter E-box elements. δEF1 shRNA cotransfected with the reporter construct pGL3-promoter-Col1a2-Ebox caused a significant increase in luciferase activity (∗, P < 0.05). (E) Down-regulation of SIP1 in response to TGF-β. SIP1 mRNA expression levels were quantified by real-time qPCR. Similar to δEF1, SIP1 expression was increased by serum depletion (SD), but TGF-β treatment reversed this by decreasing expression from 6 h posttreatment up to 24 h (#, P < 0.01). (F) Scheme showing that TGF-β inhibition of δEF1, a repressor of the E-box in the Col1a2 promoter, can increase collagen gene expression. Other negative factors or E-box repressors (SIP1) also may cooperate to regulate the enhancer elements in the Col1a2 gene.

We cloned the upstream Col1a2 enhancer E-box elements into a luciferase reporter vector and noted that TGF-β significantly increased luciferase activity of this reporter in MMC (Fig. 2C). To determine whether this effect of TGF-β is caused by its down-regulation of δEF1, we next examined whether short hairpin RNA (shRNA)-mediated knockdown of δEF1 could mimic the effects of TGF-β. We observed significant decrease of endogenous δEF1 by transfection with δEF1 shRNA (90% of mRNA and 70% of protein) in 48 h (SI Fig. 7). δEF1 shRNA significantly increased luciferase activity of the vector containing Col1a2 gene E-box elements (Fig. 2D) in MMC, similar to the effect of TGF-β (Fig. 2C). This observation further supports our hypothesis that δEF1 is a negative regulator of the Col1a2 gene enhancer and that TGF-β-mediated down-regulation of δEF1 is a key mechanism for up-regulating the collagen gene. These results suggest that, apart from the activation of positive factors Smads and MAPKs during collagen expression, TGF-β-mediated repression of the negative regulator δEF1 may be a previously uncharacterized mechanism by which TGF-β increases collagen expression in MCs under conditions such as diabetes associated with enhanced TGF-β levels. Furthermore, δEF1 also can bind to Smad proteins (22) and act as a repressor associated with CtBP (22, 23). Thus, the down-regulation of δEF1 by TGF-β also may activate the Smad pathway. δEF1 therefore may affect not only the far upstream E-box enhancer but also Smad-responsive elements (7) present in the proximal promoter of the Col1a2 gene.

SIP1, Another E-Box Repressor, also Is Negatively Regulated by TGF-β.

Interestingly, the induction of the luciferase activity by δEF1 shRNA, although significant, was relatively modest (Fig. 2D). We therefore hypothesized that other negative factors may associate with E-box elements in the Col1a2 gene and cooperate with δEF1. One potential candidate is a gene belonging to the same family as δEF1, namely, SIP1, an E-box repressor (also called ZEB2 or Zfhx1b) and closely related to δEF1 (also called ZEB1 or Zfhx1a) (15). Both δEF1 and SIP1 can interact with the corepressor CtBP (22, 23). We therefore tested whether TGF-β can regulate SIP1 in a manner similar to δEF1. Fig. 2E confirms that serum depletion of MMC significantly induced SIP1 expression, and TGF-β treatment reversed this in the same pattern as with δEF1 (Fig. 1C). These results support the concept that SIP1 may indeed be a coregulator of E-box elements along with δEF1 in the Col1a2 enhancer (Fig. 2F).

SIP1 3′ UTR Is a Potential Target of Kidney-Specific miR-192 and miR-215.

Importantly, further support for SIP1 as a candidate in this context was obtained by our observation that the 3′ UTR of the SIP1 gene is a potential target site of two miRs (miR-192 and miR-215) that were shown to be highly expressed in the kidney (20). Thus, by computational miR target predictions from miR databases [Memorial Sloan–Kettering Cancer Center, New York, NY (http://cbio.mskcc.org) and MIRANDA and TargetScan at Wellcome Trust Sanger Institute, Cambridge, U.K. (http://microrna.sanger.ac.uk/index.shtml)], we determined that SIP1 is a potential target of miR-192 and miR-215 because a sequence in its 3′ UTR has a perfect match to the seed region of these miRs (SI Fig. 8).

miR-192 Is Expressed in MMC and Up-Regulated by TGF-β.

To test the potential functions of the renal miR-192 and miR-215 in regulating SIP1 expression, we first examined the expression of these miRs in mouse kidney tissues and MMC by a real-time quantitative PCR (qPCR) method (24). We first examined miR-194 because it is highly expressed in the kidney and liver (20, 25) and confirmed that the qPCR expression pattern was similar to that found in Northern blot analyses (high levels in liver and kidney but not in spleen tissues) and thus verified the reliability of our qPCR method to monitor miR expression (Fig. 3A). (Dissociation curves and amplification curves of miR-194 standards and kidney samples are shown in SI.) We next examined the expression of miR-192 and miR-215. miR-192 was highly expressed in spleen, kidney, and MMC and relatively less expressed in liver (Fig. 3B). miR-215 was expressed almost exclusively in the kidney (Fig. 3C). miR-192 was the only miR that was expressed in MMC (Fig. 3B). Interestingly, the expression of miR-192 was decreased by serum depletion and increased by TGF-β (Fig. 3E), a pattern that clearly is opposite to that of its target SIP1 regulation (Fig. 2E). Both miR-192 and miR-215 have similar sequences and are likely to target the same genes. However, unlike miR-192, the expression of miR-215 was induced by serum depletion and decreased by TGF-β (Fig. 3F). miR-194 was up-regulated by serum depletion but remained unchanged by TGF-β treatment (Fig. 3D). These patterns are different from that of miR-192. Therefore, SIP1 is most likely regulated by miR-192 but not miR-215 or miR-194.

Fig. 3.

Expression of miRs in mouse tissues and MMC and regulation by TGF-β. (A) Verification of miR-194 expression in mouse tissues by using real-time qPCR. miR-194 expression was high in mouse liver and kidney tissue but undetected in the spleen. (B) miR-192 was highly expressed in mouse kidney, spleen, and MMC, with lower levels in liver. (C) miR-215 was expressed exclusively in the kidney with very little in other tissues or MMC. (D) miR-194 expression was very low in MMC, but serum depletion and TGF-β treatment both increased its expression. (E) Serum depletion decreased miR-192 in MMC, whereas TGF-β increased miR-192 expression from 6 to 24 h (∗, P < 0.05). (F) Serum depletion increased miR-215 expression, which was reversed by TGF-β.

The molecular mechanisms by which miR-192 is induced by TGF-β needs further investigation. However, a highly conserved upstream ets-1 protooncogene binding site has been reported in the miR-192 promoter (20). In response to serum, ets-1 forms a ternary complex with serum response factor. TGF-β also induces ets-1 expression (26). Our data showing that serum depletion decreased whereas TGF-β increased miR-192 expression (Fig. 3E) suggests that this regulation may be mediated by the ets-1 site in the miR-192 promoter. ets-1 also is essential for the development of kidneys, integrity of glomeruli, and expression of matrix metalloproteinase (27). Thus, miR-192 may be induced by TGF-β by means of ets-1 under diabetic conditions.

SIP1 3′ UTR Is a Bona Fide Target of miR-192 but Not miR-215.

In the next step, we performed experiments to confirm that SIP1 is indeed a target of miR-192. The sequence of the miR target region in the SIP1 3′ UTR is conserved among human, mouse, and rat (Fig. 4A and SI Fig. 8). We constructed a reporter vector containing the SIP1 3′ UTR immediately downstream of the luciferase gene but upstream of poly(A) signal (Fig. 4 A and B). Double-stranded RNA oligonucleotides (dsRNA) mimicking mature miR-192 (Mimic) were cotransfected with this reporter in MMC. miR-192 Mimic significantly inhibited SIP1 3′ UTR vector luciferase activity (by ≈90% relative to that with negative control Mimic) but not in the two control constructs, namely, control without 3′ UTR (Fig. 4B, first pair of bars) or that with the SIP1 3′ UTR in the antisense (AS) orientation (Fig. 4B, third pair of bars). These results confirmed that the SIP1 3′ UTR sequence is recognized by miR-192 and that SIP1 is a true target of miR-192. On the other hand, miR-215 Mimic could not decrease luciferase activity of this same SIP1 sense 3′ UTR construct (SI Fig. 11). These results, coupled with earlier data (Fig. 3 E and F) showing a different expression pattern of miR-215 relative to miR-192 in MMC treated with TGF-β, indicate that miR-192, but not miR-215, recognizes the 3′ UTR sequence in the SIP1 gene and thereby can decrease SIP1 expression. Furthermore, because TGF-β increased miR-192 expression (Fig. 3E), we tested whether TGF-β can decrease luciferase activity of the SIP1 3′ UTR reporter. Fig. 4C confirms that this is indeed the case. TGF-β had no effect in control or AS vector-treated cells.

Fig. 4.

Regulation of SIP1 gene by miR-192 via the 3′ UTR. (A) Alignment of hsa-miR-192 and mmu-miR-192 with human SIP1 3′ UTR and mouse SIP1 3′ UTR based on software from the Memorial Sloan–Kettering Cancer Center (http://cbio.mskcc.org). Seven nucleotides in the 5′ region of miR-192 (human and mouse) contain a perfect match with the 3′ UTR sequence of the human and mouse SIP1 genes. Numbers next to the SIP1 3′ UTR sequence are relative to the start sites of the 3′ UTR. Sequences of these regions were 100% conserved in human, mouse, and rat (SI Fig. 8). (B) miR-192-mediated inhibition of luciferase reporter activity in vector containing the SIP1 3′ UTR sequence. The SIP1 3′ UTR luciferase reporter (0.5 μg) was cotransfected with either the miR-192 Mimic (+miR-192 Mimic) or the negative control mimic (NC Mimic) in MMC. A significant decrease in luciferase activity with miR-192 mimic was observed with the SIP1 3′ UTR-S (sense) construct (∗, P < 0.05) but no change with the control constructs (without SIP1 3′ UTR, control) or with SIP1 3′ UTR AS construct (AS). (C) TGF-β treatment decreased the reporter activity of Luc-SIP1 3′ UTR construct (∗, P < 0.05). No change in luciferase activity was seen with control plasmid without 3′ UTR or plasmid with AS SIP1 3′ UTR. (D) miR-192 inhibitor increased luciferase activity of Luc-SIP1 3′ UTR construct (#, P < 0.01). (E) TGF-β treatment reversed the induction of luciferase activity by the miR-192 inhibitor (∗, P < 0.05). MMC were transfected with either Luc-SIP1 3′ UTR or miR-192 inhibitor (5 nM and 10 nM) and treated with TGF-β. (F) miR-192 decreased SIP1 expression. TCMK-1 cells were transfected with miR-192 Mimic and SIP1 mRNA expression quantified by real-time qPCR (∗, P < 0.05). (G) Cooperation between miR-192 and δEF1 shRNA at the E-box region. miR-192 alone slightly increased Col1a2-Ebox-luc activity. δEF1 shRNA caused a significant increase of reporter activity (∗, P < 0.05). Cotransfection of miR-192 and δEF1 shRNA lead to significant enhancement of reporter activity over their individual contributions (#, P < 0.01).

Because miR-192 Mimic could inhibit SIP1 3′ UTR luciferase reporter activity, we next examined whether a miR-192 inhibitor can have the opposite effect in this system (Fig. 4D). A 5 nM concentration of miR-192 inhibitor (2′-O-methyl-modified RNA oligonucleotides designed AS to the mature miR) demonstrated a 20-fold increase of luciferase activity that peaked at 10 nM. These results suggest that the miR-192 inhibitor, by arresting miR-192 function, can increase reporter activity. The increase in reporter activity induced by miR-192 inhibitor also was inhibited by TGF-β (Fig. 4E), suggesting that TGF-β-mediated increase in miR-192 expression can compete with and overwhelm the effects of the miR-192 inhibitor. These results further confirm that, in response to TGF-β, miR-192 can negatively regulate SIP1 gene by targeting its 3′ UTR sequence.

We then confirmed that exogenous miR-192 (Mimic) can reduce endogenous SIP1 mRNA expression (Fig. 4F), suggesting that miR-192 reduces the expression of its target SIP1 most likely via mRNA degradation. Mammalian miRs generally function by binding to their complementary sequences in the 3′ UTRs of target genes, thereby inhibiting translation of these targets. Although these miRs reduce protein expression (translational inhibition), it now is evident that this also can occur by mRNA degradation and destabilization (28, 29). Recent investigations indicate that miRs can recruit their target mRNAs to P-bodies (that concentrate enzymes involved in mRNA degradation) and inhibit protein translation by preventing ribosomal initiation or by slow degradation of mRNAs (30). One report revealed that miR-125b promotes mRNA deadenylation and enhanced degradation of its target (31). Thus, miRs can reduce target protein or mRNA expression or both. These reports support our data demonstrating that miR-192 decreases SIP1 mRNA levels and that SIP1 mRNA reduction by TGF-β coincides with an increase in miR-192 expression.

Cooperation of miR-192 and δEF1 shRNA in Up-Regulating Col1a2.

The next step was to correlate these observations to collagen regulation. Because miR-192 decreased SIP1 expression, the effect of miR-192 on Col1a2-Ebox-luciferase reporter activity in MMC was examined. Fig. 4G shows that miR-192 alone slightly increased reporter activity of this Col1a2 upstream promoter construct. δEF1 shRNA caused a modest but significant increase in reporter activity in MMC, as shown earlier in Fig. 2D. However, interestingly, when miR-192 and δEF1 shRNA were transfected together in MMC, there was a significant synergistic enhancement of reporter activity, which supports our hypothesis that both SIP1 and δEF1 induce repression at this E-box site and that a knockdown of both induces a much greater increase in Col1a2 promoter activation than either one alone (Fig. 4G).

Increased Expression of miR-192 in Diabetic Kidney Glomeruli.

Importantly, we next examined the in vivo relevance in mouse models of type 1 [streptozotocin (STZ)-injected] and type 2 (db/db) diabetes. Sieved renal glomeruli from the diabetic mice had significantly higher levels of miR-192 (Fig. 5 C and F) in parallel with increased levels of TGF-β (Fig. 5 A and D) and Col1a2 (Fig. 5 B and E) relative to the respective control mice. As mentioned earlier, profibrotic TGF-β levels are increased under diabetic conditions in renal cells, including MC, and can up-regulate ECM proteins such as collagens (2, 3). Our results showing parallel increases in TGF-β, collagen, and miR-192 in diabetic glomeruli suggest that TGF-β-induced miR-192 is responsible, at least in part, for increased Col1a2 expression not only in vitro in MC but also in vivo. Because miR-192 was increased in tissues from both type 1 and type 2 diabetic mice, hyperglycemia may be a common factor.

Fig. 5.

Expression of TGF-β, Col1a2, and miR-192 levels in glomeruli from mouse models of type 1 and type 2 diabetes. Total RNAs from normal and diabetic mouse kidney glomeruli were extracted, and the expression levels of TGF-β1, Col1a2, and miR-192 examined by real-time qPCR. (A–C) RNA from type 1 diabetic model (streptozotocin-treated) and control mice. (D–F) RNA from type 2 diabetic model (db/db) and control (db/+) mice. Significant increase (#, P < 0.01) of the expression levels of TGF-β1, Col1a2, and miR-192 was observed in diabetic glomeruli relative to healthy control. (G) A previously uncharacterized mechanism for TGF-β-induced Col1a2 gene expression. Diabetic condition increases TGF-β levels in kidney glomeruli and MC. TGF-β relieves repression at the E-box region of Col1a2 gene by reducing the expression of two E-box repressors, δEF1 (by unknown mechanisms) and SIP1 (by increasing miR-192 expression, possibly via Ets-1), to increase Col1a2 gene expression.

Together, we have discovered a mechanism for TGF-β-mediated collagen regulation that involves cross-talk between E-box repressors and miRs. Thus, TGF-β induced down-regulation of SIP1 (via miR-192) and δEF1 (via as-yet-unknown mechanisms) can cooperate to enhance Col1a2 expression via derepression at E-box elements (Fig. 5G). This finding could be related to kidney dysfunction under pathological conditions such as diabetes and also tubulointerstitial disease, in general, where TGF-β levels are enhanced. Our observations provide the first functional role for a miR expressed in the kidney. Small noncoding RNAs and miRs such as miR-192 and their inhibitors may be targets for diseases such as DN and other diabetic complications.

Methods

Cell Culture.

Primary MMC were isolated and cultured as described in ref. 4. Mouse kidney epithelial TCMK-1 cells (gift from Y. Matsuno and A. D. Riggs, City of Hope National Medical Center) were cultured in DMEM (Irvine Scientific, Santa Ana, CA) supplemented with 10% FBS. Recombinant human TGF-β1 was from R & D Systems (Minneapolis, MN).

Mouse Models.

All animal studies were conducted in accordance with a protocol approved by the Institutional Animal Care Committee. Induction of diabetes by STZ injections in C57BL/6 mice was carried out as described in ref. 4. Mice were used 7 weeks after the onset of diabetes. db/db mice and genetic control db/+ mice (10 weeks old) were obtained from The Jackson Laboratory (Bar Harbor, ME). Glomeruli were sieved from renal cortical tissue as described in ref. 4.

Microarray Experiments.

Oligo GEArray Mouse TGF-β BMP Signaling Pathway Microarrays (OMM-035) were used for expression profiling (SuperArray Bioscience Corporation, Frederick, MD) in conjunction with the TrueLabeling-AMP linear RNA amplification kit, according to manufacturer's protocols. Expression profiles from array experiments were analyzed by using GEArray expression analysis suite (SuperArray).

Real-Time qPCR.

Real-time qPCR was performed by using SYBR Green PCR Master Mix and the 7300 real-time PCR system (Applied Biosystems, Foster City, CA) as reported in ref. 4. PCR primer sequences were Col1a2, 5′-CAGAACATCACCTACCACTGCAA-3′ and 5′-TTCAACATCGTTGGAACCCTG-3′; δEF1, 5′-CATTTGATTGAGCACATGCG-3′ and 5′-AGCGGTGATTCATGTGTTGAG-3′; SIP1, 5′-CCCTTCTGCGACATAAATACGA-3′ and 5′-TGTGATTCATGTGCTGCGAGT-3′; and TGF-β, 5′-CAACGCCATCTATGAGAAAACC-3′ and 5′-AAGCCCTGTATTCCGTCTCC-3′.

Real-time qPCR also was used to detect the miRs, as reported in ref. 24. In brief, 0.5 μg of total RNA was used for cDNA synthesis primed by particular miR-specific primer with a tail sequence recognized by the universal primer. cDNA was amplified by using the miR-specific reverse and universal primers. Amplification of a single fragment was confirmed by a dissociation curve (SI Fig. 9), and good correlation between standards and threshold-cycle values was observed (SI Fig. 10). Temperature for annealing was optimized depending on the amplification of true targets. 5S RNA served as internal control (Ambion, Austin, TX). Sequences of miR-specific primers and reverse primers were miR-194, 5′-CATGATCAGCTGGGCCAAGATCCACATGGAGT-3′ and 5′-TGTAACAGCAACTCCA-3′; miR-192, 5′-CATGATCAGCTGGGCCAAGATGTCAATTCATA-3′ and 5′-CTGACCTATGAATTG-3′; and miR-215, 5′-CATGATCAGCTGGGCCAAGAGTCTGTCAATTC-3′ and 5′-ATGACCTATGAATTG-3′.

Western Blot Analysis.

Western blot analysis was performed as described in ref. 4. Antibodies against δEF1 and actin were from Santa Cruz Biotechnology (Santa Cruz, CA) and Cell Signaling (Beverly, MA), respectively.

ChIP Assays.

ChIP assays were performed according to our protocols (32). Briefly, MMC were serum-depleted and treated with TGF-β for 24 h and then fixed with formaldehyde. The cross-linked chromatin was sheared and immunoprecipitated by using antibodies against δEF1 or IgG (negative control). “ChIPed” DNA was purified and used as template for real-time qPCR by using primers spanning E-box regions (Fig. 2A). ChIP PCR primer sequences were 5′-CTAAATGCACAGTTCCTCCATGTGTTTAG-3′ and 5′-GGTGAAGTGTCCCACCAAATGGCACACC-3′. An aliquot of the formaldehyde-fixed DNA also was reverse-cross-linked and used as input control. Amplified PCR fragment was cloned into pCR3.1 (Invitrogen, Carlsbad, CA) to confirm the sequence (SI Fig. 6).

Plasmids.

To construct the Col12a-Ebox-luc plasmid, the same fragment used in the ChIP assay was excised from the pCR3.1 vector and cloned into the pGL3-promoter (Promega, Madison, WI). For SIP1 3′ UTR reporters, luciferase cDNA was excised from pGL3-basic (Promega) by digestion with HindIII and XbaI restriction enzymes. The fragment then was cloned into HindIII–XbaI site in pcDNA3.1/Zeo(+) (Invitrogen) to yield pcDNA3.1/luc3 vector. The XbaI site-tagged potential target sequences in the 3′ UTR of SIP1 gene (sense, 5′-CTAGTTTTGCATTCATTTAATTTTAGGTCAA-3′; AS, 5′-CTAGTTGACCTAAAATTAAATGAATGCAAAA-3′) were cloned into the XbaI site of pcDNA3.1/luc3. Plasmid containing sense SIP1 3′ UTR sequence was used as a reporter (pcDNA3.1/luc-SIP1 3′S), whereas the plasmid containing the 3′ UTR sequence in the AS orientation was used as a negative control (pcDNA3.1/luc-SIP1 3′AS).

Luciferase Assays.

MMC were transfected with luciferase reporter plasmids by using FuGENE 6 (Roche, Indianapolis, IN) and treated with TGF-β as reported in ref. 4. For miR cotransfections, XtremeGENE (Roche) was used according to the manufacturer's protocols. After 24 h, luciferase activities were measured as described in ref. 4. For miR and shRNA experiments, cells were treated for 48 h.

shRNA Against δEF1.

A U6 promoter-derived shRNA expression vector targeting δEF1 was constructed by using our protocol (33). Briefly, U6 promoter region was amplified by PCR using 5′ U6 primer and δEF1 target reverse primer 5′-GGACTCGAGAAAAAACCAGCAGACCAGACAGTATTACTACACAAATAATACTGTCTGGTCTGCTGGCGGTGTTTCGTCCTTTCC-3′ (the target sequence is underlined). The amplified fragment was cloned into pCR3.1EGFP (33). A plasmid expressing scrambled shRNA (pCR3.1EGFP-Scr) was used as control (33).

miRNAs.

Oligonucleotides representing the miR Mimics, the negative control 1 for Mimics, the miR-192 inhibitor (2′-O-methyl-modified RNA oligonucleotides designed AS to the mature miR), and the negative control 1 for the inhibitor all were obtained from Dharmacon (Lafayette, CO). Negative control sequences are based on Caenorhabditis elegans miRNA cel-miR-67.

Abbreviations

DN

diabetic nephropathy

MC

mesangial cells

MMC

mouse MC

SIP1

Smad-interacting protein 1

miR

microRNA

ECM

extracellular matrix

Col1a1 and -2

collagen 1-α 1 and 2

shRNA

short hairpin RNA

qPCR

quantitative PCR

AS

antisense

STZ

streptozotocin.