We have shown that microRNAs (miRNAs) are necessary for renin cell specification and kidney vascular development. Here, we used a screening strategy involving microarray and in silico analyses, along with in situ hybridization and in vitro functional assays to identify miRNAs important for renin cell identity. Microarray studies using vascular smooth muscle cells (SMCs) of the renin lineage and kidney cortex under normal conditions and after reacquisition of the renin phenotype revealed that of 599 miRNAs, 192 were expressed in SMCs and 234 in kidney cortex. In silico analysis showed that the highly conserved miR-330 and miR-125b-5p have potential binding sites in smoothelin (Smtn), calbindin 1, smooth muscle myosin heavy chain, α-smooth muscle actin, and renin genes important for the myoepithelioid phenotype of the renin cell. RT-PCR studies confirmed miR-330 and miR-125b-5p expression in kidney and SMCs. In situ hybridization revealed that under normal conditions, miR-125b-5p was expressed in arteriolar SMCs and in juxtaglomerular (JG) cells. Under conditions that induce reacquisition of the renin phenotype, miR-125b-5p was downregulated in arteriolar SMCs but remained expressed in JG cells. miR-330, normally absent, was expressed exclusively in JG cells of treated mice. In vitro functional studies showed that overexpression of miR-330 inhibited Smtn expression in SMCs. On the other hand, miR-125b-5p increased Smtn expression, whereas its inhibition reduced Smtn expression. Our results demonstrate that miR-330 and miR-125b-5p are markers of JG cells and have opposite effects on renin lineage cells: one inhibiting and the other favoring their smooth muscle phenotype.
- kidney microarrays
- JG cells
juxtaglomerular (JG) cells are renin-producing cells located in the afferent arteriole at the entrance to the glomerulus that originate from renin cells localized to the undifferentiated metanephric mesenchyme before the formation of blood vessels (20). Later in fetal life, renin-expressing cells can also be found in the wall of the afferent arteriole and large intrarenal arteries as well as in the glomerular mesangium and the kidney interstitium (6, 20). The distribution of these cells becomes progressively more restricted during ontogeny to finally acquire the typical JG localization found in the adult (7, 24). We have previously shown that in addition to JG cells, renin cells give rise to renal arteriolar smooth muscle cells (SMCs), interstitial pericytes, glomerular mesangial cells, and a subset of proximal tubular cells (19). Cells of the renin lineage exhibit a large degree of plasticity in response to environmental stimuli. In particular, an increased number of renin cells can be observed along the preglomerular arterioles, in the interstitium and inside the glomerulus, in a pattern resembling embryonic distribution, as a response to blood pressure and electrolyte homeostasis disturbances (5). These renin-producing cells are generated by dedifferentiation of arteriolar SMCs, mesangial cells, and interstitial cells from the renin cell lineage which reacquire the renin phenotype (19).
The mechanisms that control the differentiation and fate of renin cells are not well understood. Using a cell culture model, in which cells of the renin lineage are labeled with cyan fluorescent protein (CFP) and cells actively expressing renin are labeled with yellow fluorescent protein (YFP), we have previously demonstrated that arteriolar SMCs of the renin lineage reacquire the renin cell phenotype upon cAMP stimulation. This effect is mediated by increased histone H4 acetylation at the cAMP-responsive element in the renin promoter (17). In addition to this epigenetic/transcriptional control, we have also found that the renin cell phenotype is regulated by endogenous microRNAs (miRNAs), a group of small (22 nucleotides) noncoding RNAs that regulate gene expression at the posttranscriptional level (21). Experiments in which the gene that encodes the enzyme responsible for the production of mature miRNAs (Dicer) was conditionally deleted in renin-expressing cells showed a marked reduction in the number of renin cells accompanied by decreased renin expression, hypotension, and severe renal, vascular, and glomerular abnormalities, demonstrating that miRNAs are necessary for renin cell specification and normal development of the kidney vasculature.
In the present study, we hypothesized that specific miRNAs regulate the identity and fate of cells of the renin lineage. We conducted microarray analysis to screen for miRNAs with differential expression in 1) arteriolar SMCs induced to express renin in culture, and 2) kidney cortices of adult mice subjected to a treatment that induces the transformation of arteriolar SMCs into the renin phenotype. Selected miRNAs were further analyzed by RT-PCR and in situ hybridization. Finally, functional studies involving inhibition and overexpression of miRNAs were conducted to evaluate their effect on gene expression. We found that miR-330 and miR-125b-5p are markers of the JG cells and have opposite effects on their smooth muscle phenotype.
MATERIALS AND METHODS
All reagents were of reagent grade available from commercial sources.
All mouse work was conducted in accordance with the guidelines approved by the Council of the American Physiological Society and with federal laws and regulations. The protocol was reviewed and approved by the Animal Care and Use Committee of the University of Virginia. Mice were housed in a room with a 12:12-h light-dark cycle and given food and water ad libitum.
In vivo system.
C57BL/6 mice were used to isolate renal cortices for microarray and RT-PCR studies and to process kidney sections for in situ hybridization. To induce reacquisition of the renin phenotype in cells of the renin lineage in vivo, mice were administered a low-sodium diet (0.05%, Harlan, Madison, WI) plus captopril (Sigma, St. Louis, MO) in the drinking water (0.5 g/l) for 10 days.
In vitro system.
Cultured renal arteriolar SMCs were used for miRNA microarray analysis and in functional studies. This cell line was previously established from CFP/YFP mice as described (17). These cells are permanently labeled with CFP (a marker of cells of the renin lineage) and express renin upon stimulation with cAMP, a process that can be visualized by YFP expression driven by a Ren1c-YFP transgene (17). Cells were grown in DMEM/F12 supplemented with 10% heat-inactivated fetal bovine serum at 37°C in a humidified 95% air-5% CO2 atmosphere. To induce the acquisition of the renin phenotype, SMCs were treated with 10 μM forskolin (Sigma) and 100 μM IBMX (Sigma) for 24 h plus an additional treatment with 10 μM forskolin for 30 min before harvesting of the cells for microarray, RT-PCR, or functional studies.
To identify miRNAs that regulate the identity and fate of renin cells, we performed miRNA microarray analysis in kidney cortices from C57BL/6 mice and cultured vascular SMCs of the renin lineage (17) under normal conditions and after treatment to promote reacquisition of the renin phenotype. Kidney cortices were dissected from 3.5-mo-old wild-type nontreated (2 males, 1 female) and low-sodium diet plus captopril (low Na+C)-treated (2 males, 1 female) mice. CFP/YFP SMCs, at passage 7, were seeded onto 60-mm plates and grown to 80% confluence. Cells were then treated with 10 μM forskolin and 100 μM IBMX for 24 h plus an additional treatment with 10 μM forskolin for 30 min before harvest. RNA from kidney cortices and cells was prepared using the mirVana miRNA isolation kit (Applied Biosystems, Foster City, CA) according to the manufacturer's instructions. RNA was tested for purity by the ratio of absorbance at 260 and 280 nm and for integrity by Northern analysis. Two-color miRNA microarray analysis was performed by a service provider (LC Sciences; Houston, TX) using 10 μg RNA/sample. The microarray chips contained probes complementary to 599 mature miRNA transcripts listed in the Sanger miRBase database Release 11.0 (Chip ID M11.0). Data were corrected by subtracting background and normalized using a cyclic LOWESS (Locally-weighted Regression) method to remove system-related variations. A transcript was considered detectable when the signal intensity was higher than three times the background SD and the spot coefficient of variation (CV; SD/signal intensity) was <0.5. The ratio of the intensity values between control and experimental samples was transformed into log2 scale to show miRNAs that were up- or downregulated. Unsupervised hierarchical clustering and t-test analysis were performed using MeV software. The microarray data have been submitted to GEO.
miRNA target predictions.
To identify miRNAs with potential binding sites in genes affected during the reacquisition of the renin phenotype, we performed in silico analysis using the miRWalk database (4), which includes the miRWalk algorithm and eight other miRNA-target prediction programs: Diana-microT, miRanda, miRDB, PICTAR, PITA, RNA22, RNAhybrid, and TargetScan.
RNA extraction and RT-PCR.
To validate microarray miRNA expression data, we performed RT-PCR in kidney and SMC RNA samples. Total RNA was extracted from kidneys and cells using TRIzol Reagent (Invitrogen) according to the manufacturer's directions. Contaminating DNA was removed using the DNA-free kit (Ambion). To determine mRNA levels, cDNA was prepared from 2 μg of RNA using Moloney-murine leukemia virus reverse transcriptase and an oligo(dT)15 primer (both from Promega). For miRNAs, the NCode kit (Invitrogen) was used for polyadenylation and RT. RT and RT-PCR for 5S rRNA were performed using the miRCURY LNA microRNA PCR System and the primers provided (Exiqon). Quantitative real-time PCR was performed in a DNA Engine Opticon 2 system thermocycler (M. J. Research, Waltham, MA) using SYBR Green I (Invitrogen Molecular Probes) and Taq DNA polymerase (Promega) for regular gene expression or the Platinum SYBR Green SuperMix-UDG (Invitrogen) for miRNAs. Primers and PCR conditions were as follows: smoothelin qPCR: 5′-AACTGGCTACACTCTCAACAGCGA (forward; F); 5′-AAGGTGGCAGCCTTAATCTCCTGA (reverse; R), 94°C, 59.9°C, 72°C, 45 cycles; smooth muscle myosin heavy chain semiquantitative PCR: 5′-GGCTGGGGGCCGTAGAGTTATTGA (F); 5′-GAAGTGAACTGTGTGTCTGAGGTG (R), 94°C, 60°C, 72°C, 35 cycles; smooth muscle actin semiquantitative PCR: 5′-TATGTCGCTCTGGACTTTGAA (F); 5′-ACAGTTGTGTGCTAGAGACAG (R), 94°C, 62°C, 72°C, 33 cycles; GAPDH for qPCR and semiquantitative PCR: 5′-AACTTTGGCATTGTGGAAGGGCTC (F), 5′-ACCAGTGGATGCAGGGATGATGTT (R); 98°C, 56.5°C, 72°C, 25 cycles and 40 cycles, respectively; miR330: 5′-TCTCTGGGCCTGTGTCTTAGGCAA, 95°C, 62°C, 39 cycles; miR-125b-5p: 5′-TCCCTGAGACCCTAACTTGTGA, 95°C, 57°C, 72°C, 39 cycles; miR-322* 5′-AAACATGAAGCGCTGCAACAC, 95°C, 60°C; 40 cycles; miR-298: 5′-GGCAGAGGAGGGCTGTTCTTCCC, 95°C, 60°C; 40 cycles; and 5S rRNA: primer sequence from Exiqon, 95°C, 60°C; 40 cycles.
In situ hybridization.
To localize miRNAs in the kidney, we performed in situ hybridization in tissue sections of control mice and mice treated with low Na+C to induce reacquisition of the renin phenotype by arteriolar SMCs. Mice were perfused with 4% paraformaldehyde (PFA). Kidneys were immediately removed and fixed with 4% PFA for 24 h. In situ hybridization was performed on 7-μm-thick frozen sections. Detection of miRNAs was carried out as previously described (21) with modifications. Sections were postfixed in 4% PFA/PBS, sequentially washed with 0.85% NaCl, 70 and 95% ethanol, and dried. Hybridization was conducted at 37°C (for miR-125b-5p) or 45°C (for miR-330) for 18 h using 40 nM digoxygenin-labeled locked nucleic acid probe (Exiqon, Woburn, MA) specific for mouse miR-125b-5p (5′-TCACAAGTTAGGGTCTCAGGGA) or miR-330 (5′-GCCTAAGACACAGGCCCAGAGA) in 50% formamide, 5× SSC, 50 μg/ml tRNA, 1% SDS, and 5 μg/ml heparin. Sections were sequentially washed once with 5× SSC at hybridization temperature, three times with 0.2× SSC at 40–45°C for miR-125b-5p or 45°C for miR-330 and once with 0.2× SSC at room temperature. Sites of hybridization were detected using alkaline phosphatase-conjugated DIG antibody (Roche Diagnostics, Indianapolis, IN) at a 1:4,000 dilution, 4°C for 18 h, followed by BM Purple AP substrate color development (Roche). Negative controls were performed by omitting the probe and by using a commercial nontargeting miRNA probe. Detection of renin mRNA was carried out using 40 nM of a digoxygenin-labeled oligonucleotide probe (Eurofins MWG Operon, Huntsville, AL) specific for mouse Ren1 (5′-GTGTCAAAGATGACTTTGAAGGTCTG). Hybridization and washes were conducted at 40 and 50°C, respectively.
Cell transfection with miRNA precursor and inhibitor.
To assess the effect of miR-330 and miR-125b-5p on smooth muscle gene expression, we transfected cultured CFP/YFP SMCs with specific miRNA precursors and inhibitors. CFP/YFP SMCs were treated with 10 μM forskolin and 100 μM IBMX for 24 h to induce the renin phenotype or with DMSO alone (control). Cells were transfected with pre-miR precursors, anti-miR inhibitors, or nontargeting pre-miR and anti-miR negative controls (PM11180 and AM11180 for miR-330. PM10148 and AM10148 for miR-125b-5p; AM17110 and AM17010 for negative controls, Applied Biosystems Ambion) at a concentration of 50 nM as indicated in the manufacturer's instructions using Lipofectamine 2000 Reagent (Invitrogen). At indicated times, cells were harvested for RNA isolation.
Results are expressed as means ± SD. Statistical comparisons between groups were performed using a two-tailed Student's t-test.
To identify miRNAs involved in renin cell specification, we used a screening strategy involving multiple methodologies and strict exclusion criteria that allowed us to eliminate nonrelevant miRNAs: 1) the miRNAs had to be expressed in microarrays from kidney and in arteriolar SMCs of the renin lineage, 2) the miRNAs had to have predicted binding sites in the renin or smooth muscle genes, 3) the miRNAs had to be detectable by RT-PCR in kidney and arteriolar SMCs, 4) the miRNAs had to be expressed exclusively in JG cells or arteriolar SMCs of the kidney vasculature as assessed by in situ hybridization, and 5) the miRNAs had to show an effect on gene expression of renin cells in functional assays.
miRNA microarray expression profiling.
To identify miRNAs that regulate the identity and fate of renin cells, we performed miRNA microarray analysis in vascular SMCs of the renin lineage and kidney cortex under normal conditions and after treatment to promote reacquisition of the renin phenotype. We found that of 599 mature miRNAs in the microarray chip, 192 were expressed in SMC samples and 234 in kidney cortex samples. Differentially expressed miRNAs after the reacquisition of the renin phenotype are shown in Table 1 and as heat maps in Fig. 1, A and B. This analysis identified a total of 48 miRNAs differentially expressed in SMCs that had been induced to acquire the renin character (Fig. 1A). Twenty-two miRNAs were upregulated, and 26 were downregulated. In the kidney cortex, four miRNAs were differentially expressed after reacquisition of the renin phenotype (1 upregulated and 3 downregulated, Fig. 1B). Next, we focused on miRNAs whose expression in vascular SMCs and kidney has been validated by others by RT-PCR or other techniques. We found that miR-221, miR-222, miR-100, and miR-125b-5p were differentially expressed in our microarrays (Fig. 1C). On the other hand, miR-21, let-7d, miR-16, miR-30a, and miR-145, although expressed at moderate to high levels, did not show changes in expression in kidneys or SMCs stimulated to reacquire the renin phenotype. The raw data of the miRNA expression analysis are available at the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/; GSE31779 and GSE31780). Heat maps of all detectable miRNAs are available upon request.
In silico analysis of potential miRNA binding sites in genes affected by reacquisition of the renin phenotype.
The number of cells reacquiring the renin phenotype in the kidney and in SMC cultures represents a very low percentage of the whole cell population. As a result, some miRNAs specific to these cells, especially if expressed at low levels, might not be detectable in our microarrays. Therefore, in addition to the microarray studies, we carried out in silico alignment analysis to identify miRNAs with potential binding sites in genes affected during the reacquisition of the renin phenotype. We used the miRWalk database, which includes the miRWalk algorithm and eight other miRNA-target prediction programs, to identify miRNA binding sites in the smooth muscle genes smoothelin (Smtn), calponin 1, 2, and 3 (Cnn1, Cnn2, Cnn3), smooth muscle myosin heavy chain (Myh11), and smooth muscle α-actin (Acta2), and in the renin gene (Ren1). Our analysis showed that miR-330, which was detected in our microarrays at levels below threshold (i.e., signal <32), has one or more potential binding sites in Smtn, Cnn1, Myh11, Acta2, and Ren1 (Table 2). We also observed that miR-125b-5p, which as mentioned above was detected at high levels in our microarrays and was downregulated in SMCs stimulated to reacquire the renin phenotype (Table 1, Fig. 1, A and C), has potential binding sites in smooth muscle genes Smtn, Cnn1, and Myh11 (Table 2). Considering that both miR-330 and miR-125b-5p are highly evolutionarily conserved and that they have the potential to target several genes affected during reacquisition of the renin phenotype, we selected them for further analysis as described below. We also selected miR-322* and miR-298 as potential regulators of the contractile/endocrine character of renin cells because miR-322* was one of the few miRNAs differentially expressed in kidney (Fig. 1B and Table 1) and both miR-322* and miR-298 have potential binding sites in the 3′-untranslated region (UTR) of Ren1 (Table 2).
In summary, based on the combined results from the microarray and in silico analyses, we selected miR-330, miR-125b-5p, miR-322*, and miR-298 for subsequent experiments.
RT-PCR detection of expression of selected miRNAs.
To corroborate the microarray and the in silico data, we assessed the expression of selected miRNAs in kidney and SMCs by RT-PCR. Confirming the microarray and the in silico results, we found that miR-330, mir-125b-5p, miR-322* (Fig. 2), and miR-298 (not shown) were expressed in the kidney as well as in SMCs.
Localization of mir-330 and mir-125b-5p expression in the kidney.
To determine the distribution of miRNAs expressed in kidney and vascular SMCs, we performed in situ hybridization in kidney sections of control mice and mice treated with low Na+C to induce reacquisition of the renin phenotype by arteriolar SMCs. Using digoxygenin-labeled oligonucleotide LNA probes, we found that miR-330 and miR-125b-5p were expressed exclusively in cells of the renin lineage and that their expression pattern changed in response to stimuli that elicited reacquisition of the renin phenotype (Fig. 3). miR-330 was undetectable in control mice but localized exclusively to JG cells in mice treated with low Na+C. This staining pattern correlates with the low levels of miR-330 expression found in the microarrays. On the other hand, miR-125b-5p was expressed along the renal arterioles and in JG cells of untreated mice. Upon treatment with low Na+C, miR-125b-5p expression was significantly downregulated along the arterioles and remained confined to JG cells. In contrast to the defined localization of miR-330 and miR-125b-5p, miR-298 was weakly and diffusely expressed throughout the kidney, with poor expression in JG cells (not shown). Although there is species conservation of miR-298, the lack of specific staining revealed by in situ hybridization suggests that miR-298 may not be involved in the regulation of renin cell identity. miR-322*, which also showed low expression in the arrays, could not be detected above background levels by in situ hybridization. These results indicate that miR-330 and miR-125b-5p are specifically expressed in cells of the renin lineage and that their pattern of expression changed (miR-330 was upregulated and miR-125b-5p was downregulated) in response to stimuli that induce reacquisition of the renin phenotype.
Functional studies: mir-330 and mir-125b-5p regulation of smooth muscle gene expression.
Having established that miR-330 and miR-125b-5p are expressed specifically in cells of the renin lineage, we sought to determine their functional relevance. Because miR-330 and miR-125b-5p have putative binding sites in several smooth muscle genes (Table 2), we hypothesized that these miRNAs may regulate the contractile character of renin cells. We used inducible cultured SMCs as a model for reacquisition of the renin phenotype and Smtn expression as a readout for smooth muscle differentiation. As mentioned previously, physiological manipulations that induce renin synthesis (such as disturbances of fluid-electrolyte and/or blood pressure homeostasis) result in reacquisition of the renin phenotype along the afferent arterioles. As a consequence, smooth muscle proteins are downregulated while renin expression increases markedly along the renal arterioles. We have shown that this process can be replicated in vitro (17) by treating SMCs from the renin lineage with cAMP analogs, forskolin, or histone deacetylase inhibitors, conditions leading to cell dedifferentiation and to reexpression of renin. Here, we show that this process is accompanied by the downregulation of smooth muscle genes. As shown in Fig. 4A, the levels of Smtn, Myh11, and Acta2 mRNA significantly decreased when SMCs were treated with forskolin+IBMX to induce reacquisition of the renin phenotype.
Finally, we conducted functional studies to test the effect of miRNAs on smooth muscle gene expression. Figure 4B shows the effects of miR-330, miR-125b-5p or their inhibitors on the expression of Smtn. Transfection of SMCs with pre-miR-330 reduced Smtn expression to 0.64 ± 0.19 (P < 0.002) of controls. miR-330 inhibitor reverted Smtn expression to its control levels. On the other hand, transfection with pre-miR-125b-5p increased Smtn expression to 1.35 ± 0.19 (P < 0.01). miR-125b-5p inhibition reduced Smtn to 0.48 ± 0.18 (P < 0.01) of controls, a finding agreeing with our microarray data showing a 41% decrease in miR-125b-5p levels when SMCs are transformed into renin cells (Fig. 1 and Table 1). These results suggest that miR-330 inhibits and miR-125b-5p promotes the smooth muscle phenotype of renin cells.
Using a screening strategy involving microarray and in silico analyses, along with in situ hybridization and in vitro functional assays, we show here that miR-330 and miR-125b-5p are specifically expressed in cells of the renin lineage of the adult kidney and regulate smooth muscle gene expression crucial for the maintenance of the myoepithelioid renin cell phenotype.
It is well accepted that the acquisition and maintenance of a cell phenotype require a fine balance between expression and repression of many genes. In recent years, some of the pathways that govern the identity of renin cells have been elucidated (2, 8, 17). In particular, we have demonstrated that miRNAs are necessary for the maintenance of the renin cell phenotype and the normal maintenance of the renal vasculature (21). Given that to this date over a thousand miRNAs have been identified in the mouse genome and that each miRNA has wide-reaching effects on gene expression, i.e., where a single miRNA has the potential to influence the expression of hundreds of genes, the identification of specific miRNAs and their targets that are involved in the determination of the renin cell phenotype is a seemingly complex task. In this study, we found that of 599 mature miRNAs analyzed, 192 were expressed in SMCs and 234 in kidney cortex. Surprisingly, only a few of these miRNAs (22 in SMCs and 1 in the kidney) were differentially expressed with a value of P < 0.05 after the reacquisition of the renin phenotype. This can be explained by the fact that the number of cells reacquiring the renin phenotype in the kidney and in SMC cultures represents only a small percentage of the total cell population and that some miRNAs are expressed at very low levels. Therefore, by extending our analysis to include miRNAs differentially expressed with P < 0.1, we were able to identify miR-125b-5p as a potential candidate. Also, by conducting a thorough miRNA target analysis, we singled out miR-330, a miRNA expressed at very low levels in our microarrays, as a potential candidate as well. These results stress the importance of utilizing a combination of screening methodologies to avoid overlooking valuable information.
It is well accepted that there is a high degree of evolutionary conservation of miRNA genes and miRNA-mediated functions (12). Many miRNAs exhibit spatial, temporal, and tissue/cell specificity that results in their involvement in tissue morphogenesis, developmental timing, and cell differentiation (22). miR-330 and miR125b-5p are highly conserved among species. miR125b-5p, in particular, is found from invertebrates to humans. In addition to their highly conserved nature, these two miRNAs showed a highly localized pattern of expression, i.e., in cells of the afferent arterioles and JG cells. Moreover, their localization drastically changed in response to homeostatic disturbances. These observations suggest that miR-330 and miR125b-5p are likely to have an important role in the regulation of the phenotype of these cells. In fact, we found that these miRNAs have opposite effects on the smooth muscle character of renin cells under conditions that threaten homeostasis, as assessed by Smtn expression. Smoothelins are well established markers of contractile SMCs (25) and are particularly useful for detecting phenotypic changes in these cells. During the arterial remodeling that follows vascular injury or disease, SMCs undergo a process known as phenotypic modulation that is characterized by a dramatic increase in the rate of proliferation, migration, and synthesis of extracellular matrix proteins, and decreased expression of SMC-specific genes (16). Among them, Smtn is the most stringently regulated marker for the SMC switch to the synthetic phenotype, as its expression is rapidly and more extensively reduced compared with other smooth muscle genes (1). We found that overexpression of miR-330 inhibits Smtn expression in arteriolar SMCs. On the other hand, miR-125b-5p increased Smtn expression, whereas its inhibition reduced Smtn expression. miRNAs generally act as negative posttranscriptional regulators of mRNAs; therefore, one should look for changes in their expression that are opposite those of their predicted target mRNAs. We found two potential miR-330 binding sites in the Smtn gene that could explain a direct inhibitory effect of miR-330 on Smtn expression, one in the 3′-UTR and the other in the coding region. The stimulatory effect of miR-125b-5p on Smtn, on the other hand, is likely to be mediated by targeting a Smtn inhibitor gene or genes. Potential candidates are, for example, NF-κB and Elk1, which inhibit myocardin and are predicted targets of miR-125b-5p. It has been reported that miRNAs can also function to induce gene expression by binding to specific sites in the promoter of genes (18), but the absence of miR-125b-5p binding sites 10 Kb upstream of Smtn suggests that a direct stimulatory effect of miR-125b-5p on Smtn is unlikely.
Little is known about the function of miR-330 and its potential targets (13). The sequence for miR-330 lies in the first intron of Eml2, also known as EMAP, a microtubule-associated protein first described in the sea urchin (23) that alters the assembly dynamics of microtubules (10). Interestingly, renin release is highly dependent on microtubules (3). It is possible that miR-330 is transcribed from its host gene, Eml2, as it is generally accepted for intronic miRNAs. miR-330, in turn, may inhibit Eml2 expression through a binding site in the 3′-UTR of Eml2. It can be hypothesized that when homeostasis is threatened, in addition to inhibiting the contractile phenotype, miR-330 is induced in JG cells to control renin release through its regulation of Eml2. On the other hand, miR-330 transcription could be independent of Eml2 as potential RNA PolIII-regulatory elements have been found associated with miR-330 (14).
miR-125b-5p exists in two copies, miR-125b1 and miR-125b2, which in the mouse are localized to chromosomes 9 and 16 and have been linked to the control of cell fate and differentiation. miR-125b regulates the proliferation of hematopoietic stem cells and also controls cell fate during lymphoid development (15) and promotes neuronal differentiation (11). miR-125b-5p exists in a cluster with other miRNAs: let-7a2 and miR-100 for miR-125b1-5p and let-7c1 and miR-99a for miR-125b2–5p. Interestingly, all these miRNAs except for let-7a2 are expressed in our microarrays from SMCs and kidneys. In particular, miR-100, a mammalian target of rapamycin inhibitor in endothelial and vascular SMCs (9), showed a pattern of expression very similar to miR-125b-5p in SMCs.
The results from this and previous studies suggest that the pattern of gene expression in cells of the renin lineage is determined by the balance between the signals that drive renin cell identity and stimulate transcription of the renin gene and the spatial expression of miRNAs along the kidney vasculature and in JG cells. Therefore, cAMP, via CREB and associated coactivators, initiates and maintains renin gene transcription and acquisition of the renin cell identity while a set of microRNAs controls the expression of genes from other cell lineages, such as smooth muscle, thus maintaining the JG cell phenotype.
The findings reported here are consistent with a model in which miR-125b-5p and miR-330 have opposite effects on renin lineage cells in the kidney: one favoring and the other inhibiting the smooth muscle phenotype (Fig. 4C). Under normal conditions, miR-125b-5p is present in arteriolar SMCs to promote the smooth muscle phenotype and in JG cells to ensure the maintenance of their contractile function. When homeostasis is challenged, miR-125b-5p expression decreases in arteriolar SMCs to allow them to regain the renin phenotype but remains in the JG cell to ensure their contractile function. miR-330, normally absent in JG cells, becomes expressed under stress and inhibits their contractile characteristics, favoring the endocrine character of these cells. Therefore, miR-125b-5p and miR-330 act in concert to allow the plastic changes in arteriolar SMCs upstream from the glomerulus which, under physiological stress, can transiently regain the embryonic renin phenotype and ensure the permanent dual contractile/endocrine character of JG cells so crucial for the rapid control of body fluids and blood pressure homeostasis.
Further studies involving genetic ablation of miR-330 and miR-125b-5p are necessary to determine the molecular network responsible for their opposite effect on the contractile character of the cells during the reacquisition of the renin phenotype.
This work was supported by National Institutes of Health Grants to R. A. Gomez (HL096735 and HL066242) and M. L. S. Sequeira-Lopez (DK075481).
No conflicts of interest, financial or otherwise, are declared by the authors.
Author contributions: S.M., M.C.M., E.S.P., and R.A.G. analyzed data; S.M., M.C.M., E.S.P., and R.A.G. interpreted results of experiments; S.M., M.C.M., and E.S.P. prepared figures; S.M. drafted manuscript; S.M., M.C.M., M.L.S.S.-L., E.S.P., and R.A.G. edited and revised manuscript; S.M., M.C.M., M.L.S.S.-L., E.S.P., and R.A.G. approved final version of manuscript; M.C.M. and E.S.P. performed experiments; M.L.S.S.-L., E.S.P., and R.A.G. provided conception and design of research.
We thank Kimberly Hilsen-Durette for technical assistance with the mouse work.
- Copyright © 2012 the American Physiological Society