miR-324-5p and miR-30c-2-3p Alter Renal Mineralocorticoid Receptor Signaling under Hypertonicity

The Mineralocorticoid Receptor (MR) mediates the sodium-retaining action of aldosterone in the distal nephron, but mechanisms regulating MR expression are still poorly understood. We previously showed that RNA Binding Proteins (RBPs) regulate MR expression at the post-transcriptional level in response to variations of extracellular tonicity. Herein, we highlight a novel regulatory mechanism involving the recruitment of microRNAs (miRNAs) under hypertonicity. RT-qPCR validated miRNAs candidates identified by high throughput screening approaches and transfection of a luciferase reporter construct together with miRNAs Mimics or Inhibitors demonstrated their functional interaction with target transcripts. Overexpression strategies using Mimics or lentivirus revealed the impact on MR expression and signaling in renal KC3AC1 cells. miR-324-5p and miR-30c-2-3p expression are increased under hypertonicity in KC3AC1 cells. These miRNAs directly affect Nr3c2 (MR) transcript stability, act with Tis11b to destabilize MR transcript but also repress Elavl1 (HuR) transcript, which enhances MR expression and signaling. Overexpression of miR-324-5p and miR-30c-2-3p alter MR expression and signaling in KC3AC1 cells with blunted responses in terms of aldosterone-regulated genes expression. We also confirm that their expression is increased by hypertonicity in vivo in the kidneys of mice treated with furosemide. These findings may have major implications for the pathogenesis of renal dysfunctions, sodium retention, and mineralocorticoid resistance.


Introduction
The Mineralocorticoid Receptor (MR), encoded by the NR3C2 gene, belongs to the nuclear receptor superfamily and mediates aldosterone action by stimulating transcription of target genes involved in the control of sodium reabsorption in tight epithelia [1], such as the α subunit of the Epithelial Na + Channel [2][3][4], the Serum and glucocorticoid-regulated kinase 1 (Sgk 1) [5,6], or the Glucocorticoid-induced leucine zipper transcription factor (Gilz) [7]. Thus, MR controls salt and water balance and participates in regulating blood pressure. This MR signaling is not functional in newborns due to low renal MR expression at birth, which accounts for the physiologic partial aldosterone resistance of human newborns, reminiscent of the MR haploinsufficiency reported in autosomal dominant pseudohypoaldosteronism [8]. Moreover, many studies showed that the MR plays a pivotal role in the pathogenesis of several diseases, including heart failure, hypertension [9], and nephropathies [10]. However, mechanisms controlling renal MR expression remain poorly understood. At the transcriptional level, the human NR3C2 gene is driven by two alternative promoters, the proximal P1 promoter, which is transcriptionally active in all MR target tissues, and the distal P2 promoter, which is weaker and transcriptionally active in the central nervous system [11,12]. Beyond these transcriptional regulatory mechanisms, MR activity and signaling are also modulated by post-translational modifications such as phosphorylation, SUMOylation, ubiquitylation, and acetylation [1,13]. In the distal nephron, MR is mainly expressed in the cortex, where the tubular lumen is rather hypotonic, while its expression is low in the medulla [14,15], which is hypertonic. Previously, we have demonstrated that variations of extracellular tonicity, prevailing in nephron segments, modulate MR expression through post-transcriptional mechanisms involving the recruitment of RNA Binding Proteins (RBPs) [16]. Indeed, we showed that hypertonicity compromises MR signaling by recruiting Tis11b (Tetradecanoyl phorbol acetate inducible sequence 11b), which degrades MR mRNA [17], while HuR (Human antigen R) favors MR mRNA stability [18] and edits MR mRNA [19], thus enhancing MR signaling under hypotonicity. In this context, we could not exclude another working hypothesis involving the recruitment of microRNAs (miRNAs), another class of post-transcriptional regulators. These endogenous short non-coding RNAs [20,21], which negatively regulate gene expression mainly by targeting the 3 -Untranslated Region (3 -UTR) of target transcripts, are involved in various biological processes [22][23][24]. The miRNAs-target gene interaction constitutes a complex network. Indeed, overexpression of a particular miRNA affects the expression of hundreds of target transcripts [25,26]. Conversely, one specific transcript is generally regulated by several miRNAs [27]. Furthermore, it was shown that miRNAs can be secreted in plasma or urine and might thus constitute promising biomarkers, as well as potential novel therapeutic targets for renal dysfunction [28,29]. Indeed, accumulating evidence now suggests that miRNAs exert key roles not only in development but also in renal diseases [30][31][32], especially those affecting MR signaling [33]. For instance, putative binding sites for miRNAs were identified in the 3 -UTR of transcripts related to the reninangiotensin-aldosterone system [34]. miR-34c-5p was shown to play an important role in aldosterone-induced fibrosis in the kidney [35]. miRNAs were also shown to impact ionic transport in response to aldosterone [36][37][38][39][40] or MR expression in other pathophysiological situations [34,[41][42][43][44]. However, no study reported the role of miRNAs on MR expression in response to tonicity. We have identified miR-324-5p and miR-30c-2-3p, the expression of which is upregulated by hypertonicity in murine principal KC3AC1 cells. Herein, we describe how they functionally interact with MR transcript, in cooperation with RBPs, and subsequently affect its expression and MR signaling. In vivo, we have further demonstrated that their expression is significantly increased in the kidneys of mice treated with furosemide, a diuretic known to increase the relative luminal tonicity in the distal nephron.

Total RNAs Isolation
RNAs, isolated from cultured cells or from mouse kidneys, were extracted with NucleoSpin ® miRNA kit (Macherey Nagel, Hoerdt, France) according to the manufacturer's protocol. This kit facilitates the isolation of both miRNAs and large RNAs from each sample for use in RT-qPCR, Taqman Low-Density Array analysis, and miRNAs-sequencing.

Taqman Low Density Array (TLDA)
TLDA allows testing of a large subset of miRNAs by RT-qPCR in a single assay. First, 600 ng of miRNAs from isotonic or hypertonic medium-incubated KC3AC1 cells were subjected to reverse transcription (RT) using the TaqMan ® miRNA reverse transcription kit and Megaplex TM RT primers Pool Set v2.0 (Pool A and B) according to the manufacturer's protocol (Applied Biosystems, Villebon-sur-Yvette, France). An amplification step was then performed in the Service of Molecular Genetics, Pharmacogenetics, and Hormonology of Bicetre Hospital (France). RT products were diluted with RNAse-free water, combined with TaqMan ® 2X Universal PCR Master Mix, No AmpErase UNG then loaded into four TaqMan ® rodent microRNA array v2.0 (Applied Biosystems, Villebon-sur-Yvette, France: one Array A and one Array B for isotonic or hypertonic condition), which are 384-well microfluidic cards embedded TaqMan probes in each well for different mature rodent miRNAs enabling accurate quantification of 585 miRNAs for mice for each condition. snoRNA 202 transcript was used for normalization. Quantitative PCR was performed according to the manufacturer's instructions and ran on a 7900 HT Fast Real-Time PCR system with 384-Well Block Module (Applied Biosystems, Villebon-sur-Yvette, France). Results were analyzed on SDS 2.3 software (Applied Biosystems, Villebon-sur-Yvette, France) and expressed as 2 −∆∆Ct values with ∆Ct = Ct miRNA − Ct snoRNA 202 .

miRNAs-Sequencing (miRNAs-Seq)
miRNAs-Seq was performed to profile different miRNAs under isotonic or hypertonic medium-incubated KC3AC1 cells. An amount of 200 ng total RNAs (containing miRNAs) per sample (n = 4 per condition) was used to conduct miRNAs library preparation and sequencing at iGenSeq Platform (Institut du Cerveau et de la Moelle Epinière, Paris, France). The concentration and size distribution of all samples were determined by 4200 TapeStation (Agilent, Les Ulis, France). Library preparation for sequencing was performed on the RNA fragment under 200 bases by using the QIAseq ® miRNA Library kit according to the manufacturer's protocols (Qiagen, Villebon-sur-Yvette, France). Briefly, a pre-adenylated DNA adapter was ligated to the 3 -end of all mature miRNAs; an RNA adapter was ligated to the 5 -end of mature miRNAs. Ligated miRNAs were reverse transcribed to cDNA using a reverse transcription (RT) primer with integrated Unique Molecular Indices (UMIs), which are molecular tags used to detect and quantify unique mRNA transcripts. After reverse transcription, a cleanup of the cDNA was performed using a streamlined magnetic bead-based method before and after library amplification. Next-generation Sequencing was performed in a single-end mode, 1 × 75 bp, on the NextSeq 500 Illumina System platform with approximately 32 million reads per sample, following the manufacturer's protocols. Reads quality was assessed using FastQC on the FASTQ file. Alignment was performed against the mouse genome (GRCh38/mm10). Raw miRNA counts and differential expression were analyzed using R/Bioconductor package DESeq2. An adjusted p-value < 0.05 with Benjamini & Hochberg False Discovery Rate method was considered statistically significant.

miRNAs RT-qPCR
For miRNAs RT-qPCR, a multiplex reverse transcription protocol was performed to reverse transcribe simultaneous multiple miRNAs (12 miRNAs maximum) into cDNA from 200 ng of miRNAs using the TaqMan ® microRNA Reverse Transcription kit (Applied Biosystems, Villebon-sur-Yvette, France) and diluted (1:100 in 1X TE) specific stem-loop primers for corresponding miRNAs (Applied Biosystems, Villebon-sur-Yvette, France). RT products were used for quantification of individual miRNAs that were run on a QuantStudio 6 Flex (Applied Biosystems, Villebon-sur-Yvette, France) using specific miRNAs TaqMan probes (Life Technologies, Villebon-sur-Yvette, France) and TaqMan Universal PCR Master Mix, No AmpErase UNG, 2X (Life Technologies, Villebon-sur-Yvette, France). A list of specific miRNA stem-loop primers and TaqMan probes is presented in Supplementary Table S1. The specificity of each miRNA TaqMan probe was confirmed by cloning each amplicon into the pGEMT-easy plasmid (Promega, Charbonnière, France) and sequencing. The efficiency of qPCR was controlled before analysis. Relative expression in each miRNA was normalized to the housekeeping miRNA (mmu-miR-16 or snoRNA 202), where control condition values were arbitrarily set at 100%.

mRNAs RT-qPCR
For mRNAs RT-qPCR, after DNAse I treatment (Biolabs, Évry-Courcouronnes, France), 1 µg of total RNAs was reverse transcribed using the High-Capacity cDNA reverse transcription kit (Life Technologies, Villebon-sur-Yvette, France). Samples were analyzed by RT-qPCR using the Power SYBR ® Green PCR Master Mix (Life Technologies, Villebon-sur-Yvette, France) with the primers indicated in Supplementary Table S2A and run on QS6 Real-Time PCR System (Life Technologies, Villebon-sur-Yvette, France). The specificity of each primer set was confirmed by cloning each amplicon into the pGEMT-easy plasmid (Promega, Charbonnière, France) and sequencing. qPCR Efficiency was at least 95% in all experiments. Relative expression in each sample was calculated as a ratio (amol of specific gene/fmol of 18S or amol of 36b4), where the control condition values are arbitrarily set at 100%.

Plasmid Constructs
Wild type murine Elavl1 (HuR) 3 -UTR (mHuR 3 -UTR) was amplified by PCR from cDNA using specific primers presented in Supplementary Table S2B, in which SpeI and HindIII restriction sites were introduced to facilitate cloning into the pMIR-Report TM plasmid (Life Technologies, Villebon-sur-Yvette, France). pMIR-mMR-3 -UTR plasmid and Tis11b-encoding pTarget vector were generated as previously described [18]. All plasmids were sequenced to verify nucleotide identity at the Eurofins Genomics platform (Cologne, Germany).

Luciferase Assays
To study the functional interaction between miRNAs and Nr3c2 (MR) or Elavl1 (HuR) 3 -UTR, HEK 293T cells were transfected, using Lipofectamine 2000, with the luciferase constructs (40 ng/well of 96-well plates) with either increasing concentrations of Mimics or negative control Mimics (CTR Mimics) or with a maximal effective concentration of Inhibitors or negative control Inhibitors (CTR Inhibitors) (Life Technologies, Villebon-sur-Yvette, France, see sequences in Supplementary Table S3. To evaluate the cooperativity between miRNAs and the Tis11b RNA-Binding Protein on Nr3c2 (MR) and Elavl1 (HuR) 3 -UTR, the Tis11b-encoding pTarget vector (10 ng/well of 96-well plates) was cotransfected into HEK 293T together with pMIR-mMR-3 -UTR (40 ng) or pMIR-mHuR-3 -UTR (40 ng) with or without Mimics or control Mimics. Cell lysates were collected 24 h post-transfection to measure luciferase activities. Luciferase activities were normalized to β-galactosidase activities as previously described [17].

Establishment of KC3AC1 Clones Stably Transduced with Lentivirus Expressing Inducible miR-324-5p
Production of HIV-1 U3-SIN lentivirus particles was carried out at the Viral Vector and Gene Transfert Platform (Structure Federative de Recherche Necker, UMS 24, UMS 3633, Faculté de Santé, Université Paris Cité, France) by cotransfection into HEK 293T cells of the miR-324-5p inducible lentiviral plasmid (ShMIMIC Inducible miR-324-5p, #GSH11929-224640456, Dharmacon, Cambridge, United Kingdom) or the control inducible lentiviral plasmid (SMARTvector Inducible Non-targeting mCMV-TurboRFP, #VSC1165, Dharmacon, Cambridge, United Kingdom) with the vesicular stomatitis G protein (VSV-G) envelope, and other packaging plasmids using the standard calcium phosphate transfection method. After production, functional lentiviral titers were determined in hHTC 116 cells by puromycin selection or Red Fluorescent protein Fluorescence-Activated Cell Sorting analysis 48 h after induction by doxycycline (Sigma, Saint-Quentin-Fallavier, France) at 1 µg/mL. The harvested supernatant containing ShMIMIC miR-324-5p or scrambled lentiviral particles was transduced into renal KC3AC1 cells at a multiplicity of infection (MOI) of 20. The selection of transduced cells was performed by adding 1.5 µg/mL of puromycin (InvivoGen, Toulouse, France) 48 h after transduction. After 5 days of selection, cells were subjected to serial dilutions in 96-well plates and maintained in culture for up to 15 days until the expansion of isolated clones. After 48 h incubation, doxycycline (1 µg/mL), KC3AC1 clones overexpressing miR-324-5p could be visualized thanks to the observation of the Red Fluorescent Protein (excitation at 553 nm/emission at 574 nm) on BX73 optical microscope (Olympus). Finally, miRNAs RT-qPCR analyses allowed selection of the clone Sm-A1, which stably overexpresses scrambled control miRNAs, and the clone Sh-H8, which, on the opposite, stably overexpresses the miR-324-5p. Sm-A1 and Sh-H8 clones were cultured in the same medium as the parental KC3AC1 cells with puromycin (1 µg/mL). To analyze the impact of miR-324-5p overexpression on MR signaling, Sh-A1 and Sh-H8 clones cultured on filters onto 12 well-plates (1.2 × 10 6 cells/filter) were deprived in minimal medium lacking dexamethasone, EGF, DCC concurrently subjected to doxycycline (1 µg/mL) induction for 48 h then stimulated with 10 nM aldosterone for 1 h. Total RNAs were thereafter extracted then relative MR target gene expression was analyzed by RT-qPCR.

Western Blot Analyses
Forty micrograms of extracted proteins were subjected to SDS-PAGE and processed for detection of MR protein together with α-tubulin protein for loading normalization. The signal fluorescence intensity was determined with an Odyssey ® (Li-Cor, Homburg, Germany). Antibody sources and dilutions are given in Supplementary Table S4.

Investigations in Mice
Mice received an intraperitoneal injection of furosemide (40 mg/kg, Renaudin, Itxassou, France) for 4 h. Animals (n = 5-6 per group) were euthanized, and kidneys were collected and snap-frozen in liquid nitrogen for subsequent analyses. Mice were bred according to the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH Publication No. 85-23, revised 1996). The animal facility was granted approval (no. C94-043-12) by the Ministère de l'Agriculture, France. All procedures were approved by the local ethic committee CAPSud (N • 2012-021).
As expected, aldosterone stimulated the expression of Tsc22d3 (Gilz), a classical MR target gene, when cells were transfected with CTR Mimics. However, this effect was reduced when cells were transfected with 30c-2-3p Mimics ( Figure 5F). Similar results were obtained for the kinase Sgk1. Aldosterone induced an increase in Sgk1 expression, but this stimulatory effect was weaker in the presence of 10 nM 30c-2-3p Mimics ( Figure 5G). Taken together, our findings highlight a critical role for miR-30c-2-3p in compromising MR signaling following the reduction in MR expression in KC3AC1 cells.

Impairment of MR Signaling in KC3AC1 Cells Stably Expressing miR-324-5p
Because transient transfection with high concentrations of 324-5p Mimics had no impact on MR expression, we established a stably transduced renal KC3AC1 cell line with lentivirus expressing inducible miR-324-5p to evaluate the impact of its overexpression on MR expression and signaling. KC3AC1 cells were transduced with scrambled or miR-324-5p ShMIMIC lentiviral particles. After puromycin selection (1.5 µg/mL) and doxycycline induction (1 µg/mL), cells were subjected to serial dilutions to finally establish two cellular clones: Sm-A1 clone expressing scrambled miRNAs ( Figure 6, left panels) and Sh-H8 clone expressing miR-324-5p under doxycycline induction ( Figure 6, right panels). Given that expression of miRNAs was coupled to that of the red fluorescent protein, we used fluorescence microscopy and RT-qPCR to select transduced clones that overexpressed either scrambled miRNAs or miR-324-5p (Supplementary Figure S2). Left panels in Figure 6A indicate that incubation of Sm-A1 clone for 48 h with doxycycline had, as expected, no impact on miR-324-5p expression. Therefore, Nr3c2 (MR) and Elavl1 (HuR) expressions were not impaired ( Figure 6B,C). Conversely, miR-324-5p overexpression (×80) ( Figure 6D) induced a significant 20 % decrease in Nr3c2 (MR) and Elavl1 (HuR) transcripts in Sh-H8 clone ( Figure 6E,F). Finally, we determined the impact of miR-324-5p overexpression on MR expression and signaling: Sm-A1 clone and Sh-H8 clone were cultivated for 48 h in a minimal medium then stimulated for 1 h with 10 nM aldosterone. As expected, aldosterone induced a similar 20% increase in Tsc22d3 (Gilz) expression when the Sm-A1 clone was incubated in the absence or presence of doxycycline ( Figure 6G). However, this aldosteroneinduced expression of Tsc22d3 (Gilz) was reduced when the Sh-H8 clone was incubated in the presence of doxycycline ( Figure 6H). Altogether, these findings demonstrate that miR-324-5p also impairs MR signaling in KC3AC1 renal cells.    ) and Elavl1 (HuR) (C,F) expression were determined by RT-qPCR, 48 h after doxycycline induction (1 µ g/mL) in KC3AC1 clones stably transduced with lentiviral particles expressing inducible scrambled miRNAs (Sm-A1 clone, left panels) or miR-324-5p (Sh-H8 clone, right panels). Data are means ± SEMs from two independent experiments performed in six replicates (n = 12). NS = not significant, *** p < 0.001, **** p < 0.0001 compared to the condition in absence of doxycycline, arbitrarily set at 100%. (G, H) Sm-A1 and Sh-H8 clones were deprived in minimal medium for 48 h and incubated for 48 h with 1 µ g/mL doxycycline. Thereafter, Sm-A1 and Sh-H8 clones were stimulated for 1 h with 10 nM aldosterone and Tsc22d3 (Gilz) expression (G,H) was quantified by RT-qPCR where basal Gilz expression in renal cells, in the absence of DOX and aldosterone, was arbitrarily set at 100%. Data are means ± SEMs from three independent experiments performed in six replicates per condition (n = 18). NS = not significant. ** p < 0.01 *** p < 0.001 **** p < 0.0001 compared to the condition without aldosterone stimulation, arbitrarily set at 100%. § § § indicate p < 0.001 between the two conditions indicated by line. Figure 6. Impairment of MR signaling in KC3AC1 cells stably expressing miR-324-5p. miR-324-5p (A,D), Nr3c2 (MR) (B,E) and Elavl1 (HuR) (C,F) expression were determined by RT-qPCR, 48 h after doxycycline induction (1 µg/mL) in KC3AC1 clones stably transduced with lentiviral particles expressing inducible scrambled miRNAs (Sm-A1 clone, left panels) or miR-324-5p (Sh-H8 clone, right panels). Data are means ± SEMs from two independent experiments performed in six replicates (n = 12). NS = not significant, *** p < 0.001, **** p < 0.0001 compared to the condition in absence of doxycycline, arbitrarily set at 100%. (G,H) Sm-A1 and Sh-H8 clones were deprived in minimal medium for 48 h and incubated for 48 h with 1 µg/mL doxycycline. Thereafter, Sm-A1 and Sh-H8 clones were stimulated for 1 h with 10 nM aldosterone and Tsc22d3 (Gilz) expression (G,H) was quantified by RT-qPCR where basal Gilz expression in renal cells, in the absence of DOX and aldosterone, was arbitrarily set at 100%. Data are means ± SEMs from three independent experiments performed in six replicates per condition (n = 18). NS = not significant. ** p < 0.01 *** p < 0.001 **** p < 0.0001 compared to the condition without aldosterone stimulation, arbitrarily set at 100%. § § § indicate p < 0.001 between the two conditions indicated by line.

Expression of miR-324-5p and miR-30c-2-3p Are Increased in Kidneys of Mice Treated with Furosemide
To address the physiologic relevance of data obtained in KC3AC1 cells, we used kidney samples of mice treated for 4 h with furosemide, an NKCC2 inhibitor. In the cortex, basolateral fluid is maintained isotonic by vascularization, while luminal fluid, delivered at the end of the distal loop, is hypotonic due to the removal of salt in this segment impermeable to water. Thus, furosemide, by acutely inhibiting salt reabsorption in thick limbs, leads to an increase in the osmolality of fluid delivered to the distal nephron and to a decrease in MR expression ( Figure 7A), as previously published [17]. Thus, we demonstrated that miR-324-5p and miR-30c-2-3p expression were dramatically increased in the kidneys of these furosemide-treated mice ( Figure 7B,C), thus demonstrating, in vivo, that increased luminal tonicity, enhanced miR-324-5p and miR-30c-2-3p expressions. However, this increase in miRNAs expression in the whole kidney may be the result of changes not only in the mineralocorticoid sensitive distal nephron but also in other segments in the kidney, where MR and miRNAs may be co-expressed. We, therefore, established the quantitative expression profile of miRNAs (miR-324-5p and miR-30c-2-3p) along the nephron of three adult mice by performing glomeruli and nephron segments microdissection: the following structures were carefully microdissected according to morphologic and topographic criteria: glomeruli (Glo), proximal convoluted tubules (PCT), medullary and cortical thick ascending limb of Henle's loop (mTAL and cTAL), connecting tubules (CNT), cortical and outer medullary collecting duct (CCD and OMCD) (Supplementary Figure S3A). This manual method of tubule selection allowed us to separate renal segments with a minimum of contamination, if any, of one type with another [48]. As shown in Supplementary Figure  S3B below, relative MR expression is, as expected, weak in the glomeruli (Glo), the proximal tubule (PCT), and the Henle's loop (mTAL and cTAL). By contrast, and as expected, MR expression is highest in the aldosterone-sensitive distal nephron (CNT, CCD, OMCD). This quantitative expression profile of miRNAs along the nephron, although difficult to implement, enabled us to confirm that miR-324-5p and miR-30c-2-3p are co-expressed in the aldosterone-sensitive distal nephron together with MR but they are also expressed in other segments such as Glo and PCT (Supplementary Figure S3C,D).
increased luminal tonicity, enhanced miR-324-5p and miR-30c-2-3p expressions. However, this increase in miRNAs expression in the whole kidney may be the result of changes not only in the mineralocorticoid sensitive distal nephron but also in other segments in the kidney, where MR and miRNAs may be co-expressed. We, therefore, established the quantitative expression profile of miRNAs (miR-324-5p and miR-30c-2-3p) along the nephron of three adult mice by performing glomeruli and nephron segments microdissection: the following structures were carefully microdissected according to morphologic and topographic criteria: glomeruli (Glo), proximal convoluted tubules (PCT), medullary and cortical thick ascending limb of Henle's loop (mTAL and cTAL), connecting tubules (CNT), cortical and outer medullary collecting duct (CCD and OMCD) (Supplementary Figure S3A). This manual method of tubule selection allowed us to separate renal segments with a minimum of contamination, if any, of one type with another [48]. As shown in Supplementary Figure3B below, relative MR expression is, as expected, weak in the glomeruli (Glo), the proximal tubule (PCT), and the Henle's loop (mTAL and cTAL). By contrast, and as expected, MR expression is highest in the aldosterone-sensitive distal nephron (CNT, CCD, OMCD). This quantitative expression profile of miRNAs along the nephron, although difficult to implement, enabled us to confirm that miR-324-5p and miR-30c-2-3p are co-expressed in the aldosterone-sensitive distal nephron together with MR but they are also expressed in other segments such as Glo and PCT (Supplementary Figure S3C,D).

Discussion
We previously showed that MR expression is regulated by variations of extracellular tonicity prevailing in nephron segments [16] and highlighted the predominant role of RBPs [17][18][19]. Herein, we demonstrate that miRNAs also modulate MR expression in vitro and in vivo. Using high throughput screening approaches, we identified miR-324-5p and miR-30c-2-3p, expression of which is up-regulated under hypertonicity in renal cells and in kidneys of furosemide-treated mice, whereas MR expression is, in parallel, reduced. Renal cells resist hypertonicity by recruiting TonEBP (TonE Binding Protein), which belongs to the Rel factors family [49,50]. This transcription factor binds Tonicity response Elements (TonE), located in promoter regions of target genes such as Aldose Reductase [51] or Zfp36l1 gene, which encodes Tis11b [17]. Jaspar software [52] identified TonE consensus sequence TGGAAANNYNY in regulatory regions of murine mir324 and mir30c2 gene (Supplementary Figure S4), suggesting that TonEBP may regulate the expression of these miRNAs ( Figure 8). Moreover, we demonstrate that these miRNAs functionally interact with murine Nr3c2 (MR) 3 -UTR, which is conserved among species and presents with the highest number of putative binding sites for miRNAs compared to other transcripts of the renin-angiotensin-aldosterone system [34]. Indeed, transfection of Mimics for both miRNAs decreased Nr3c2 (MR) 3 -UTR-driven luciferase activity, an effect that was blunted by the corresponding Inhibitors. Given that cooperativity between RBPs and miRNAs had previously been reported [47], we transfected 324-5p or 30c-2-3p Mimics together with Tis11b in HEK 293T cells and demonstrated that they cooperate to decrease Nr3c2 (MR) 3 -UTR-driven luciferase activity. This latter result could be explained by the location of AREs in Nr3c2 (MR) 3 -UTR, on which Tis11b binds, located close to binding sites for miR-324-5p and miR-30c-2-3p, thus facilizing MR-destabilizing action of Tis11b [17]. Indeed, Grimson et al. demonstrated the functional importance of the environment surrounding miRNAs binding sites for their destabilizing action [53]. We also showed that miR-324-5p and miR-30c-2-3p could directly affect the stability of the Nr3c2 transcript (MR) and/or indirectly destabilize the Elavl1 transcript (HuR), a master RBP, which was shown to enhance MR expression and signaling under hypotonicity [18] (Figure 8). Such a complex interplay between these two classes of regulators is now coming to light [47]. Next, we analyzed the impact of miR-30c-2-3p and miR-324-5p overexpressions on MR expression in KC3AC1 cells. miR-30c-2-3p is more effective than miR-324-5p in destabilizing Nr3c2 (MR) 3 -UTR since its overexpression, following transfection of low concentrations of Mimics, led to a significant decrease in MR expression, both at the messenger (25%) and protein (50%) levels, compromising MR signaling with blunted responses in terms of aldosterone-regulated gene expression (Figure 8). In contrast, overexpression of miR-324-5p did not affect MR expression even with high concentrations of Mimics. Consequently, the use of lentivirus expressing miR-324-5p has proven to be an efficient strategy to overexpress this miRNA in fully differentiated renal principal KC3AC1 cells, which fully express functional MR when high confluency is reached. Thus, we showed that the Sh-H8 clone presented with a 20 % decrease in MR expression at the mRNA level responsible for a reduced aldosteroneinduced expression of Tsc22d3 (Gilz), thus demonstrating that this miRNA also impairs MR signaling (Figure 8). This difference in efficacy could be explained by the number of binding sites for these miRNAs: there are two binding sites for miR-30c-2-3p at nucleotide +219 and +1279, whereas only one binding site for miR-324-5p is located at nucleotide +1131. Indeed, Doench et al. reported that additivity between several identical binding sites for a miRNA might have a higher impact on its repressive action [54]. Another possible explanation relies on the location of the miR-30c-2-3p binding sites, which are close to the stop codon, thus accelerating MR protein degradation [53]. Although the decrease in MR expression was relatively small (50% decrease in protein level), this action of miRNAs on the fine posttranscriptional control of MR expression is reminiscent of the role of miRNAs as buffers against variation in gene expression, as described by Ebert et al. [55]. Our study highlighted the role of these miRNAs, which could impair MR signaling in KC3AC1 cells. miR-30c-2-3p belongs to the miR-30 family involved in various renal diseases [56,57]. In addition, it has been reported that levels of miR-30 family members (including miR-30a, miR-30c, and miR-30e) increased in plasma of contrast-induced nephropathy (CIN) rats, suggesting that the miR-30 family might serve as early biomarkers for CIN [58]. Finally, miR-30c was also shown to be involved the in regulation of tubular cells apoptosis in cisplatin-induced nephrotoxicity [59]. Thus, our results add to the diversity of miR-30 family actions in renal pathophysiology. miR-324-5p is located on chromosome 11 and inhibits diverse functions in cancer. A recent study demonstrated that miR-324-5p plays a protective role against hepatocellular carcinoma [60]. Another study indicated significant upregulation of miR-324-5p in lung cancer cells and demonstrated that it promotes proliferation [61]. Although few data are available on its action in the kidney, we could not exclude the role of this miRNA in other MR target tissues. Indeed, miR-324-5p was demonstrated to contribute to seizure onset by repressing the expression of Kv4.2, a potassium channel, in the central nervous system [62], where MR plays a neuroprotective role [63,64]. Moreover, miR-324-5p was shown to inhibit mitochondrial fission, apoptosis, and myocardial infarction [65] through downregulation of Mtfr1, whereas MR is expressed in cardiac myocytes, and its activation is associated with myocardial hypertrophy and fibrosis [66,67]. To address the physiologic relevance of our study, we treated mice for 4 h with furosemide, a diuretic known to increase the relative luminal hypertonicity and decrease MR expression [17] and showed that miR-324-5p and miR-30c-2-3p expressions were increased in kidneys of these animals by 2-to 4-fold, respectively. This increase was higher than that observed in KC3AC1 cells, suggesting that their expression may also be increased in other nephron segments. Finally, our group has identified a restricted temporal window during renal development where MR signaling is ineffective due to specific downregulation of MR expression in the perinatal period compared to the lung, where MR expression is constant [68]. In this context, it would be interesting to evaluate whether such regulatory mechanisms are implicated in the regulation of MR expression within the perinatal period. tochondrial fission, apoptosis, and myocardial infarction [65] through downregulation of Mtfr1, whereas MR is expressed in cardiac myocytes, and its activation is associated with myocardial hypertrophy and fibrosis [66,67]. To address the physiologic relevance of our study, we treated mice for 4 h with furosemide, a diuretic known to increase the relative luminal hypertonicity and decrease MR expression [17] and showed that miR-324-5p and miR-30c-2-3p expressions were increased in kidneys of these animals by 2-to 4-fold, respectively. This increase was higher than that observed in KC3AC1 cells, suggesting that their expression may also be increased in other nephron segments. Finally, our group has identified a restricted temporal window during renal development where MR signaling is ineffective due to specific downregulation of MR expression in the perinatal period compared to the lung, where MR expression is constant [68]. In this context, it would be interesting to evaluate whether such regulatory mechanisms are implicated in the regulation of MR expression within the perinatal period. TonEBP binds TonE elements located in the regulatory regions of mir324 and mir30c-2 genes which encode miR-324-5p and miR-30c-2-3p, respectively. Hypertonicity stimulates transcription of pri-miRNAs of miR-324-5p and of miR-30c-2-3p and their maturation. Thereafter, these miRNAs can repress MR expression directly by interacting with Nr3c2 (MR) 3′-UTR or indirectly by interacting with Elavl1 (HuR), an RNA Binding Protein, which was shown to stabilize Nr3c2 (MR) transcript under hypotonicity [19]. Moreover, these miRNAs can cooperate with Tis11b, an RNA Binding TonEBP binds TonE elements located in the regulatory regions of mir324 and mir30c2 genes which encode miR-324-5p and miR-30c-2-3p, respectively. Hypertonicity stimulates transcription of pri-miRNAs of miR-324-5p and of miR-30c-2-3p and their maturation. Thereafter, these miRNAs can repress MR expression directly by interacting with Nr3c2 (MR) 3 -UTR or indirectly by interacting with Elavl1 (HuR), an RNA Binding Protein, which was shown to stabilize Nr3c2 (MR) transcript under hypotonicity [19]. Moreover, these miRNAs can cooperate with Tis11b, an RNA Binding Protein, which was previously shown to accelerate destabilization of MR transcripts [18]. Thus, hypertonicity may compromise MR signaling.

Conclusions
Altogether, our study has deciphered a novel regulatory mechanism involving the recruitment of miRNAs in the control of renal MR expression under hypertonicity, resulting in the alteration of MR signaling with a potential impact on sodium reabsorption. Our findings may also account for (patho)physiological situations affecting MR signaling in nephropathies [10], such as diabetic nephropathies or partial aldosterone resistance in newborns [8]. Given that miRNAs are also secreted in body fluids [69], these miRNAs may be quantified in plasma and urine of patients presenting with an altered salt and water balance or with diabetic nephropathy and might be used as non-invasive biomarkers to follow the evolution of renal diseases.