Downregulation of exosomal miR-192-5p and miR-204-5p in subjects with nonclassic apparent mineralocorticoid excess

Background The “nonclassic” apparent mineralocorticoid excess (NC-AME) has been identified in approximately 7% of general population. This phenotype is characterized by low plasma renin activity (PRA), high serum cortisol (F) to cortisone (E) ratio, low cortisone, high Fractional Excretion of potassium (FEK) and normal-elevated systolic blood pressure (SBP). An early detection and/or identification of novel biomarkers of this phenotype could avoid the progression or future complications leading to arterial hypertension. Isolation of extracellular vesicles, such as exosomes, in specific biofluids support the identification of tissue-specific RNA and miRNA, which may be useful as novel biomarkers. Our aim was to identify miRNAs within urinary exosomes associated to the NC-AME phenotype. Methods We perform a cross-sectional study in a primary care cohort of 127 Chilean subjects. We measured BP, serum cortisol, cortisone, aldosterone, PRA. According to the previous reported, a subgroup of subjects was classified as NC-AME (n = 10). Urinary exosomes were isolated and miRNA cargo was sequenced by Illumina-NextSeq-500. Results We found that NC-AME subjects had lower cortisone (p < 0.0001), higher F/E ratio (p < 0.0001), lower serum potassium (p = 0.009) and higher FEK 24 h (p = 0.03) than controls. We found miR-204-5p (fold-change = 0.115; p 0.001) and miR-192-5p (fold-change = 0.246; p 0.03) are both significantly downregulated in NC-AME. miR-192-5p expression was correlated with PRA (r = 0.45; p 0.028) and miR-204-5p expression with SBP (r = − 0.48, p 0.027) and F/E ratio (r = − 0.48; p 0.026). Conclusions These findings could support a potential role of these miRNAs as regulators and novel biomarkers of the NC-AME phenotype.


Introduction
Arterial hypertension (AH) currently affects approximately 40% of adults worldwide and is a complex, multifactorial disease. This disease affects many metabolic systems in its pathogenesis, mainly the renal, vascular, and endocrine systems. Approximately 15% of essential hypertensive patients may be associated with endocrine pathologies related to mineralocorticoid receptor (MR) activation, such as primary aldosteronism and deficits in 11β-hydroxysteroid dehydrogenase type 2 (11β-HSD2) enzyme activity. Severe deficiency in 11β-HSD2 is also known as apparent mineralocorticoid excess (AME) and results from inefficient metabolism of cortisol (F) to its inactive metabolite cortisone (E), leading to the activation of the mineralocorticoid pathway by cortisol [1][2][3]. by lower renin and a higher urinary potassium excretion rate) [4], suggesting the existence of a distinctive phenotypical spectrum in these subjects, which may be prevalent in a primary care cohort. In this study, we did not identify a genetic cause associated to that phenotype, which led us to hypothesize that epigenetic modifications (e.g. miRNAs affecting MR pathway genes) might be responsible for the onset and progression of hypertension of this phenotype, the nonclassic AME. miRNAs are small noncoding RNA molecules that are approximately 21 to 23 bp long and they regulate target mRNAs through either translational repression, mRNA destabilization or a combination of both mechanisms. A single miRNA can regulate hundreds of genes, and collectively, miRNAs may regulate approximately 50-60% of the total transcriptome [5,6]. The miRNAs influence gene expression both within their parental cells and promote intercellular communication by being transferred to other cells through nanovesicles-called exosomeswhere they regulate cellular processes in the recipient cell [7][8][9]. Exosomes are small extracellular vesicles (EVs) with a size of 50-150 nm originating from endosomes. They are released from all cell types with a specific cargo (RNA, lipids and protein). Exosome cargo may mirror the physiological state or metabolic change of the cells of origin [10,11]. Exosomes are isolated from different biofluids by ultracentrifugation (UCF) technology and could be potential biomarkers by itself or by their cargo for a variety of pathophysiological conditions, such as arterial hypertension [12][13][14].
The identification of miRNAs, RNA or proteins within exosomes associated to metabolic changes could be very informative of the local cellular physiology and pathophysiology. Thus, different translational approaches have been developed to identify specific miRNAs and proteins that are associated with pathophysiological conditions. The aim of the current study was to study the miRNA profile obtained from urinary exosomes from subjects with the nonclassic AME (NC-AME) phenotype.

Subjects
A total of 396 Chilean subjects between 10 and 65 years old and both genders were invited to participate from two primary care centers in Santiago, Chile. All subjects have a similar socioeconomic status and ethnicity, and they declared had not ingested licorice. For the current study, we excluded subjects with pre-puberal stage (Tanner < 2), kidney disease, uncontrolled diabetes mellitus, liver failure and heart failure. Subjects using glucocorticoids, oral contraceptives or spironolactone were excluded due to the known effects these have on adrenal hormones levels in plasma. Subjects with previously diagnosed secondary causes of hypertension, such as familial hyperaldosteronism, classic apparent mineralocorticoid excess (AME), hypercortisolism, and renovascular disease, were also not included in this study. Subjects with glucocorticoid remediable aldosteronism (GRA) (with positive molecular diagnosis), primary aldosteronism (PA) (aldosterone-renin ratio > 25 and PRA < 1 ng/ml*h) or potential hypercortisolism (urinary free cortisol > 110 μg/24 h [15]) were also excluded from this study. Patients who were using antihypertensive drugs that affect the renin-angiotensin-aldosterone system (RAAS), such as beta-blockers, angiotensin-converting enzyme (ACE) inhibitors, angiotensin II receptor blockers, and diuretics, or those who received amlodipine or doxazosin for at least 4 weeks for blood pressure control due to the neutral effect over the RAAS were also excluded from this study. The protocol followed in this study was in agreement with the guidelines of the Declaration of Helsinki and was approved by the Ethical Committee of the Faculty of Medicine, Pontificia Universidad Catolica de Chile (CEC-MEDUC#14-268). A written informed consent was signed by all participants.

Clinical characteristics
All subjects underwent a complete physical examination performed by trained endocrinologists of the Pontificia Universidad Catolica de Chile. The subjects' heights were measured using a wall-mounted Harpenden stadiometer (Holtain). Three BP measurements were obtained from the right arm at consecutive 5-minute intervals using an oscillometric method (Dinamap CARESCAPE V100, GE Healthcare, Medical Systems Information Technologies, Milwaukee, WI) with the subjects in a seated position. Hypertension was diagnosed when blood pressure was higher than 130/80 mm Hg in adults [16] and adolescents [17] after at least 2 different measurements.

Biochemical assays
After overnight fasting, basal blood samples were obtained between 08:00 and 10:00 AM. Serum aldosterone (SA) and plasma renin activity (PRA) were measured by radioimmunoassay using a commercial kit (Coat-A-Count Kit; Siemens, Los Angeles, CA and DiaSorin, Stillwater, MN, respectively). At the same time, spot and 24-h urine samples were collected. Serum and urinary cortisol and cortisone were quantified using LC-MS/MS and levels were validated according to the parameters suggested by the Food and Drug Administration (FDA) and Clinical and Laboratory Standard Institute (CLSI) using deuterated internal standards of cortisol and cortisone (cortisol-d4 and cortisone-d2) in an Agilent 1200 series HPLC equipment coupled to an ABSciex 4500-QTrap mass spectrometer.

Identification of NC-AME subjects
After exclusion criteria were applied, we perform the study with 127 subjects (age 10-65 years old). In all subjects we analyzed serum cortisone and the cortisol to cortisone ratio (F/E) [4]. Subjects with both cortisone lower than percentile 25th and a serum cortisol to cortisone ratio higher than percentile 75th were classified as subjects suspected of nonclassic AME (NC-AME) [4]. We identify 10 subjects with NC-AME suspicion, corresponding to the 7.8% of studied subjects (Table 1). We performed a cross-sectional study in NC-AME subjects compared with a control group (N = 15 subjects) similar in age, gender, body mass index and urinary sodium excretion (Table 1).

Exosomes isolation
Approximately 13 ml of spot urine was collected from each subject in the morning (08:00-10:00 AM

Electron microscopy
Exosomal shape and size was determined by transmission electron microscopy (TEM). For this, 15 μl of an exosome pellet was added onto a carbon-coated cooper grid (300 mesh) for 1 min and stained with 2% uranyl acetate for 1 min. Grids were visualized at 80 kV in a Phillips Tecnai transmission electron microscope.

Nanoparticle Tracking Analysis (NTA)
NTA measurements were performed using a NanoSight NS500 instrument (Malvern, UK) with the NanoSight NTA 3.0 nanoparticle tracking and analysis software (Version Build 0064) as previously described [18]. Three videos of 30 s were processed and analyzed. A minimum of 200 completed tracks per video was collected for each analyzed sample. NTA post-acquisition settings were optimized and kept constant between samples and each video was then analyzed to determine the mean, mode, and median particle size, together with an estimated number of particles per ml of plasma. A spreadsheet (Excel, Microsoft Corp., Redmond, Washington) was automatically generated, recording the concentration of each particle size. 100 nm polystyrene latex microspheres (Malvern NTA 4088) were routinely analyzed to confirm the instrument performance.

Exosomal RNA isolation
RNA was isolated by organic extraction using the Trizol ® reagent according to the manufacturer's instructions. The SPECTROstar Nano Microplate Reader (BMG LABTECH) spectrophotometer was used to quantify the RNA concentration. Following a cleanliness check and blank measurement using RNase-free water, 2 μL of each sample was pipetted on to a microdrop well on an LVis plate. The RNA concentration was measured using the MARS Data Analysis microplate reader software.

Next generation sequencing
Sequencing libraries were generated using the TruSeq 1 Small-RNA Library Prep Kit (Illumina, San Diego, Ca, USA) according to manufacturer's instructions. A total of 100 to 300 ng of exosomal RNA was used as input for library preparation. These RNA samples were barcoded by ligation with unique adaptor sequences to allow pooling of samples into groups of 24. Subsequently, these ligated samples were reverse transcribed, PCR amplified, and size selected using gel electrophoresis. Finally, DNA libraries were eluted from the extracted gel pieces overnight in 200 μL nuclease free water. The elution, containing the pooled DNA library, was further processed for cluster generation using the NextSeq 500 high output kit for 75 cycles and sequencing using the Illumina Next-Seq 500 sequencing platform.

Identification of miRNAs in raw sequencing data
Initially, raw FASTQ files were processed to remove barcode and adaptor sequences. Subsequently, these files were analyzed using the miRDeep2 program to identify known miRNAs [19]. The miRDeep2 algorithm requires a genomic index and miRNA database to perform analysis. The human genome (version 19) prebuilt index was obtained from the bowtie website (http://bowti e-bio. sourc eforg e.net/index .shtml ). The miRNA reference database (version 20) was obtained from the miRBase website (http://www.mirba se.org/) [20]. Sequencing data have been deposited in the GEO database with the accession number GSE138556.

Taqman quantitative real-time PCR assay
To validate the RNA sequencing data, we performed a Taqman qRT-PCR analysis. Reverse transcription was performed for each specific miRNA using the Taqman MicroRNA RT kit (Applied Biosystems, Foster City, CA). The expression of the RNU6 snRNA was used as an internal normalization control. The expression levels of the miRNAs were evaluated with the TaqMan MicroRNA Assay kit in the RotorGene 6000 thermocycler (Corbett Research, Sydney, Australia).

Bioinformatic analyses
Gene target identification for identified miRNAs was performed using the miRwalk software. Candidate miR-NAs identified from the sequencing data were imported into miRwalk. A total of 4 miRNA gene target databases (miRWalk, TargetScan, miRanda, and RNA22) were provided to miRwalk for analysis. Subsequently, gene targets were filtered to identify those RNAs that are targeted by the same miRNA within at least two separate databases. From this selection, genes shown to be regulated by at least two miRNAs were extracted and subjected to gene ontology analysis. MirPath v.3 analysis of downregulated miRNAs (Diana Tools Software) was performed to find pathways in KEGG and the gene-ontology (GO) enrichment analyses (Additional file 1).

Statistical analysis
Data are presented as percentages for categorical variables. Variables with a non-normal distribution are reported as median and interquartile ranges [Q1-Q3]. In situations where a variable was not normally distributed, a bootstrapping procedure with 1000 iterations was performed. We evaluated the differences between NC-AME subjects and control subjects by t-test with bootstrapping in non-normal variables. A 2-sided p value of < 0.05 was considered to be statistically significant. We also perform univariate and multivariate regression analyses, using stepwise selection, to find predictive variables of miRNA expression. All analyses were performed using SPSS 20 and the GraphPad Prism v5.0 software. Differential expression and statistical analysis of sequencing data was performed by the DESeq 2 package in R [21]. This package uses a generalized linear model to perform differential expression. Statistical analysis and significance were calculated using a Wald test and adjusted for multiple testing using the Benjamini and Hochberg procedure.

Clinical and biochemical analyses
From the 127 subjects studied, we identified 10 subjects that fulfilled the criteria for NC-AME (7.8%), which were compared with 15 control subjects matched by age, gender and BMI. The NC-AME group included 5 adolescents and 5 adults, and the control group included 7 adolescents and 8 adults. Baseline characteristics of both groups are shown in Table 1

Exosome isolation and characterization
The morphological characteristics of the urinary exosomes (i.e. round donut shape and 50-150 nm diameter) were observed by transmission electron microscopy (TEM) and are shown in Fig. 1a. The presence of the known exosome markers CD63 and TSG101 were confirmed by immunoblot analysis as shown in Fig. 1b. We found that 73% of the urinary nanovesicles with a size between 50 and 150 nm in both groups of subjects (NC-AME and controls) and 27% corresponded to nanovesicles with a size greater than 150 nm (Fig. 1c). The percentage of nanovesicles smaller than 50 nm was less than 1%.

Urinary exosome microRNA profiling and data analysis
To identify the expression profile of miRNA in urinary exosomes, we used a high-throughput sequencing technology, RNA-sequencing. From the 2822 described miR-NAs, we detected 355 miRNAs in urinary exosomes, of which 170 miRNAs showed a fold-change higher than 1 (or Log2 > 0) and 185 miRNAs showed a fold-change lower than 1 (or Log2 < 0) in NC-AME compared to control subjects. However, only two miRNAs were found to be significantly downregulated (p < 0.05). Statistical analyses indicated that the expression of hsa-miR-204-5p (fold change = 0.115; p < 0.05) and hsa-miR-192-5p (fold change = 0.246; p < 0.05) were downregulated in NC-AME subjects as compared to the control group (Table 2, Fig. 3).

Bioinformatic studies for target prediction of miR-192-5p and miR-204-5p
Bioinformatics analyses were performed using four databases for target prediction (miRWalk, TargetS

Association of miRNAs and clinical-biochemical parameters
Association studies of miRNA expression and clinical and biochemical characteristics of all subjects showed that hsa-miR-192-5p expression correlates positively with PRA (r = 0.45; p = 0.028) and hsa-miR-204-5p expression displayed an inverse correlation with SBP (r = − 0.48, p = 0.027) and serum cortisol to cortisone ratio (r = − 0.48; p = 0.026) ( Table 3). A linear regression supports a model where the expression of hsa-miR-204-5p is predicted by DBP, serum cortisol to cortisone ratio, FEK, and gender (R 2 = 0.86). Similarly, multivariate analyses support a model where expression of miR-192-5p can be predicted by serum aldosterone levels and urinary sodium excretion (R 2 = 0.37).

Discussion
We identified 355 miRNAs in urinary exosomes, of which only 2 were significant downregulated in NC-AME subjects compared with controls ( Table 2, Fig. 3). By Taqman RT-qPCR we confirmed that miR-192-5p was downregulated, meanwhile miR-204-5p have a trend to a lower expression in NC-AME subjects.
Previous studies indicate that miR-192-5p and miR-204-5p are highly expressed in the kidney [22,23]. MIR-192 is expressed at higher levels in the renal cortex than in the medulla [24] and is 20-fold higher in the proximal tubules than in the glomeruli. miR-192 is involved in regulation of sodium transport in renal epithelial cells [25]. A recent study by Baker et al. showed low expression of miR-192-5p in kidney biopsy specimens from patients with hypertensive nephrosclerosis and hypertension [26]. Reduced expression of miR-192-5p is associated with an increase in Na/K-ATPase function (ATP1B1 gene), which contributes to hypertension and kidney injury [26]. Similarly, it has been shown that loss of miR-192-5p is associated with fibrogenesis in diabetic nephropathy [27]. All of these examples highlight the role of miR-192-5p in the renal system, which could be useful as a biomarker for some types of kidney diseases, especially in AH. Bioinformatics studies shown that miR-192-5p could regulate genes related to both small GTPase mediated signal transduction (CUL3, ARHGAP1, ARHGAP36, ARH-GEF39) and sodium transport (ATP1A2, SCL5A12), which have been previously related to the mineralocorticoid receptor [28][29][30][31] and sodium/potassium exchange [26,32,33] pathways (Table 4), suggesting a role in the etiopathogenesis of arterial hypertension.
The miR-204-5p is also highly expressed in kidney tissues and has been shown to be downregulated in advanced diabetic nephropathy biopsies [34]. Other studies have observed a reduction of miR-204-5p expression in epithelial cells associated with reduced expression of claudins 10, 16 and 19, suggesting a critical, albeit indirect, a role of this miRNA in maintaining epithelial cell function [35]. Using mirWalk, we found that miR-204-5p could potentially regulate genes downstream MR-activation related to sodium transmembrane transport (NEDD4, ATP1A2, ATP2B4, WNK3), cellular response to hormone stimuli (NEDD4, ATP1A2, NR3C1, NR3C2,  YWHAG), and genes that regulate of molecular function in the cell (ATP2B4, ATP1A2, NEDD4, WNK3, ARH-GEF37, ARHGEF26, ARHGAP30, YWHAG), suggesting a potential role for miR-204-5p in renal pathways associated with sodium/potassium exchange (Table 4). Recently, has been reported that miR-204 is a critical regulator of de novo DNA methylation, through affecting the DNA methyltransferase 3-alpha (DNMT3a) [36]. In this way, we speculate that low expression of miR-204 seen in NC-AME, could be associated with higher expression of DNMT3a and hypermethylation the HSD11B2 promoter [37,38], decreasing the HSD1B2 expression, and lately affecting the cortisol to cortisone metabolism (Table 4).
Our results also show that both miR-192-5p and miR-204-5p could regulate ATP1A2 expression and it has been previously shown that this α2-isoform of the Na/K-ATPase pump mediates ouabain-induced hypertension in mice and increased vascular contractility in vitro [33]. Association studies indicated that miR-192-5p expression is correlated with PRA which suggest being a potential biomarker of MR activation, also can be predicted by aldosterone and sodium urinary excretion, which is in agreement with a previous report by Elvira-Matelot et al. that showed renal miR-192-5p expression is decreased by aldosterone infusion [39]. Similarly, hsa-miR-204-5p expression was negatively associated with serum cortisol to cortisone and SBP, which highlights it as a potential biomarker in NC-AME subjects. However, further research should be performed to validate these miRNAs as potential biomarker of NC-AME and regulator of key genes in MR pathway [34,[40][41][42].
To the best of our knowledge, this is the first study addressing the role of urinary exosomes and their miRNA content in subjects with nonclassic AME. We found that urinary exosomes from NC-AME subjects have lower expression of two miRNAs compared with controls subjects. Nevertheless, there are a few limitations in our study. First, the sample size of patients and controls was relatively small, and these findings should be validated in larger cohorts to better evaluate the sensitivity and specificity of miR-204-5p and miR-192-5p as potential biomarkers of nonclassic AME. Second, the predicted miRNA gene targets that were identified using bioinformatics methods would need to be validated in vitro, and the role of these miRNAs in the regulation of cellular pathways will need to be further investigated.

Conclusions
The present study showed a lower expression of miR-192-5p and miR-204-5p in urinary exosomes from NC-AME compared with control subjects, which are associated with low PRA and high cortisol to cortisone ratio. Moreover, our results shown that NC-AME is also present in normotensive subjects (Table 1), which open new biomedical challenges aimed to identification of novel and early biomarkers, preventive actions and also potential second hits associated to NC-AME phenotype.
We suggest these miRNAs may have a potential role as early biomarkers and may regulators of the mineralocorticoid activity in NC-AME subjects, which will be useful to may uncover and understand the mechanisms associated to this phenotype.