Small RNA Sequencing of Human Urinary Extracellular Vesicles Reveals Association of High‐Sodium Diet With Renal Proinflammatory Pathways

Abstract

Background

High sodium intake is associated with arterial hypertension and cardiovascular disease, through mechanisms that go beyond hemodynamic changes, including endothelial dysfunction, oxidative stress, and induction of a proinflammatory milieu. The aim of this study was to assess the role of dietary sodium modulation on renal pathophysiology through evaluation of small RNA cargos on urinary extracellular vesicles.

Methods

Fourteen high‐risk normotensive subjects with normal kidney function were prospectively enrolled to undergo a low‐sodium diet followed by a high‐sodium diet (HSD). The urinary extracellular vesicles were isolated from a 24‐hour urine collection at the end of each diet phase and profiled by small RNA sequencing. Selected differentially expressed miRNAs were validated in human proximal tubular cell line (human kidney 2 cells) to assess miRNA‐mRNA target interactions.

Results

We identified 111 small RNA species, of which 30 were significantly different between the low‐sodium diet and HSD. Bioinformatic network analysis showed that pathways related to the innate and adaptive immune system, interleukin and interferon signaling were enriched in the HSD, whereas pathways related to PPARα (peroxisome proliferator‐activated receptor α) regulation were enriched in the low‐sodium diet. In human kidney 2 cells, the inhibition of miR‐320b, downregulated in the HSD, increased ICAM‐1 (intercellular adhesion molecule 1), with renal proinflammatory effects. The inhibition of miR‐10b‐5p, downregulated in the low‐sodium diet, increased PPARα, which has an antifibrotic and anti‐inflammatory role in the kidney.

Conclusions

Small RNA characterization from extracellular vesicles revealed that an HSD is associated with proinflammatory changes, potentially contributing to sodium‐induced low‐grade renal inflammation.

Nonstandard Abbreviations and Acronyms

EV

extracellular vesicles

HSD

high‐sodium diet

ICAM‐1

intercellular adhesion molecule 1

LSD

low‐sodium diet

MCP‐1

monocyte chemotactic protein‐1

PPAR

peroxisome proliferator‐activated receptor

uEV

urinary extracellular vesicle

Clinical Perspective.

What Is New?

• Bioinformatic network analysis and in vitro experiments revealed upregulation of pathways related to the immune system and interferon signaling after a high‐sodium diet and PPAR (peroxisome proliferator‐activated receptor)‐mediated anti‐inflammatory pathways after a low‐sodium diet.

• The miRNA‐mediated activation of proinflammatory pathways after high sodium consumption could play an important role for the development of hypertension and renal damage

What Are the Clinical Implications?

• Future longitudinal studies should prospectively evaluate the long‐term effects of dietary sodium modulation on small RNA expression and renal physiology and the potential prognostic role of miRNAs to predict the development of arterial hypertension.

Excessive dietary sodium intake contributed to an estimated 1.89 million deaths worldwide in 2019, with the associated health burden steadily rising over the past 3 decades. ¹ High sodium consumption is associated with increased blood pressure (BP) levels, ² , ³ cardiovascular events, ³ and chronic kidney disease. ¹ Strong evidence shows that reducing sodium intake lowers BP and cardiovascular risk, leading international guidelines to recommend a daily sodium intake of <2 g (<5 g of salt). ⁴ , ⁵ However, adherence to a low‐sodium diet (LSD) remains low, with a global average sodium intake of 4 g per day (10 g of salt), significantly exceeding the recommended target. Significant evidence, derived from preclinical research in animal models, ⁶ , ⁷ showed that the long‐term deleterious effects of a high‐sodium diet (HSD) are only partially explained by the hemodynamic changes induced by sodium load. These effects also involve endothelial dysfunction, increased oxidative stress, and the promotion of a proinflammatory environment. ⁶ , ⁷ In particular, low‐grade renal inflammation appears to play a pivotal role in the long‐term deleterious effects of an HSD. ⁶ , ⁷ Preclinical studies suggested that an HSD could directly activate the adaptive and innate immune system at the renal level, increasing the infiltration of inflammatory cells and cytokine production. ⁷

Extracellular vesicles (EVs) are bilayer membrane‐bound structures released by all cell types, which carry nucleic acids (such as small noncoding RNAs, mRNAs, and DNA fragments), proteins, lipids, and metabolites. ⁸ EVs can be isolated from cell cultures, biofluids, ⁸ and tissues, ⁹ and their cargos reflect the activation state of parental cells. EVs allow cell‐to‐cell paracrine and endocrine communication, influencing the content and biological processes of the recipient cells. ⁸ , ¹⁰

In the presence of an intact glomerular filtration barrier, circulating EVs cannot cross the glomerular membrane pores. ¹¹ Therefore, urinary EVs (uEVs) are secreted primarily from the cells of the nephron lumen and the urinary tract. ¹¹ , ¹² This feature makes the uEVs an optimal noninvasive tool to characterize biological processes of kidney, bladder, and urogenital tract. ¹²

miRNAs are small noncoding RNAS abundantly released into EVs that complementarily bind the 3′ untranslated region of target mRNAs, regulating mRNA degradation and translation. ¹³ miRNAs transported within EVs are stable molecules, able to target distant sites and regulate gene expression. ⁸

The aim of this study was to investigate whether high sodium consumption is associated with pathways related to inflammatory processes. For this purpose, small noncoding RNAs were evaluated in EVs isolated from the 24‐hour urine home collections of normotensive subjects (with normal kidney function and high risk of hypertension development) who underwent a controlled dietary sodium intervention, including a sodium‐restricted and sodium‐loaded phase. After the identification of differentially expressed miRNAs among the 2 dietary regimens, we identified mRNA targets and performed a bioinformatic network analysis to identify the pathways uniquely enriched following each diet. We then selected specific miRNA‐mRNA target pairs, considered as potential drivers of sodium‐mediated effects and validated the miRNA‐mRNA target regulation using human tubular cells in vitro.

METHODS

Primary data, supplemental tables, and summary statistics have been deposited in the following repository: https://github.com/CentroIpertenUnito/uEVSodium_smallRNAseq. The data discussed in this publication have been deposited in the National Center for Biotechnology Information's Gene Expression Omnibus and are accessible through Gene Expression Omnibus Series accession number GSE288935 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE288935). ¹⁴ The complete data set that supports the findings of this study is available from the corresponding author upon reasonable request.

Participants Selection

The study complied with the Declaration of Helsinki. Participants for this study were included from a prospective study assessing hormonal mechanisms of hypertension (NCT03484130), conducted at Brigham and Women's Hospital, Boston, Massachusetts, and approved and monitored by the human research and ethics committee of MassGeneral Brigham. Written informed consent was obtained from all the participants who were recruited between 2018 and 2022, according to the following criteria: age 35 to 70 years, an estimated glomerular filtration rate >60 mL/min per 1.73 m², an untreated BP 120 to 135/75 to 85 mm Hg, and at least 1 condition among body mass index ≥25 kg/m², family history of hypertension before the age of 60 years in a parent or sibling, and diabetes with a hemoglobin A1c <9%. If systolic BP was 115 to 120 mm Hg or diastolic BP was 70 to 75 mm Hg, the participants had to fulfill at least 2 criteria among those described above (Data S1). Professional dieticians prepared the LSD first, which included prepared meals for 5 to 7 days designed to include ≈10 to 40 mEq/d of sodium (0.23–0.92 g/d), ≈50–75 mEq/d of potassium, and ≈600 mg/d of calcium. Once the LSD was completed, participants completed 5 to 7 days of the HSD with a target >200 mEq/d (or 4.6 g/d) of sodium, maintaining ≈50 to 75 mEq/d of potassium and ≈600 mg/d of calcium (Figure 1). The high‐sodium phase involved dietary counseling by professional dieticians to increase dietary sodium intake with sodium‐rich foods and in addition sodium‐rich broth packets when participants' daily sodium intake was not considered high enough, targeting a daily intake of 200 mEq/d. The 24‐hour urine volume and creatinine were measured on every occasion to ensure that parameters were reflective of a 24‐hour period. Office systolic BP and diastolic BP were measured at the end of each dietary phase using an automated and unattended technique with the average of triplicate measurements calculated. ¹⁵ A 24‐hour urine home collection was obtained at the end of each dietary condition. All blood samples were collected in the morning at 8 am. Plasma aldosterone concentrations were measured by ELISA (IBL International [catalog RE52301]; Hamburg, Germany). The lowest reportable concentration was 0.2 ng/dL with a dynamic range of 0.57 to 100 ng/dL, interassay coefficient of variation=8.6% to 9.4%, intra‐assay coefficient of variation=5.6% to 9.7%. Plasma renin activity was measured by ELISA (IBL‐America; Minneapolis, MN; interassay coefficient of variation=4.8%–7.1%, intra‐assay coefficient of variation=6.3%–8.7%). Further details on participants selection are provided in Data S1.

Figure 1. Design of the study.

We prospectively recruited 14 subjects with high‐risk normotension who underwent 5 to 7 days of a low‐sodium diet (<40 mmol/day or 0.92 g/day) or high‐sodium diet (>200 mmol/day or 4.6 g/day). At the end of each diet, we isolated uEVs from 24‐hour urine collection through serial centrifugation and ultracentrifugation. We then performed small RNA sequencing of the uEVs and bioinformatic analysis to select the miRNA‐mRNA target for in vitro validation. We finally validated miRNA‐target regulation in human tubular cell lines (human kidney 2 cells). EV indicates extracellular vesicle; and uEVs, urinary extracellular vesicles.

EVs Isolation and Characterization

uEV isolation and characterization complied with recommendations of the International Society for Extracellular Vesicles. ¹² Detailed description of EV characterization is provided in Data S2 through S4.

uEVs were analyzed for diameter and quantified using Nanosight LM10 (NanoSight, Salisbury, UK). To provide the particle concentration and size distribution profiles, nanoparticle tracking analysis version 3.1 software (NanoSight) was used for both data acquisition and analysis. The presence of EV‐specific markers (CD63, flotillin‐1) and cell‐specific marker (calnexin) was assessed by Western blotting. We assessed EV integrity and morphology by transmission electron microscopy (detailed description of EV characterization is provided in Data S2 through S4).

RNA Isolation and Library Preparation for Small RNA Sequencing

Details on RNA isolation are reported in Data S5. Small RNA transcripts were converted into barcoded cDNA libraries as previously described. ¹⁶ Briefly, library preparation was performed with the NEBNext Multiplex Small RNA Library Prep Set for Illumina (protocol E7330; New England BioLabs). For each library, 6 μL of RNA (minimum 35 ng−maximum 100 ng) were used in all the experimental procedures as starting material. A unique indexed primer was used for each library so that all the libraries could be pooled into 1 sequencing lane. Multiplex adaptor ligations, reverse transcription primer hybridization, reverse transcription reaction, and polymerase chain reaction amplification were performed according to the protocol for library preparation. After polymerase chain reaction amplification, the cDNA constructs were purified with the QIAQuick Polymerase Chain Reaction Purification Kit (Qiagen, Germany) following the modifications suggested by the NEBNext Multiplex Small RNA Library Prep Protocol E7330 and loaded on the Bioanalyzer 2100 (Agilent Technologies) using the DNA High Sensitivity Kit (Agilent Technologies) according to the manufacturer's protocol. Libraries were pooled together equimolar (90 nM for each library) in 36‐plex and further purified with a gel size selection.

A final Bioanalyzer 2100 run with the High‐Sensitivity DNA Kit (Agilent Technologies) allowed the final analysis of DNA libraries on size, purity, and concentration. The obtained libraries were subjected to the Illumina sequencing pipeline, passing through clonal cluster generation on a single‐read flow cell and 75 cycles sequencing‐by‐synthesis on the NextSeq500 sequencer (Illumina).

Statistical Analysis

Raw reads adapter clipping was performed with the Cutadapt software (version 1.18). ¹⁷ Reads longer than 14 nucleotides were mapped to a small noncoding RNA (sncRNA) reference with the bwa alignment software (version 0.7.17‐r1188), ¹⁸ using the mem algorithm and a seed length of 10. Only alignments without mismatches or indels were considered, and those with the highest quality were used to assign each read to a unique sncRNA. Thus, sncRNAs were quantified for each sample after normalization for number of reads in each library and then merged into a single count matrix, setting missing sncRNAs to 0. Differential expression analysis was performed with the DESeq2 Bioconductor's package (version 1.22.2) ¹⁹ using paired comparison, after correction for 24‐hour urinary volume. For each model, samples with missing covariates were dropped, and only sncRNAs, where at least 70% of the remaining samples had counts >5, were tested. sncRNAs were considered significantly associated with a condition or a trend if their P value, after adjustment for multiple testing by False Discovery Rate, was below the 0.05 threshold. Normalized expression levels were obtained from DESeq2 for plotting, because absolute counts from sequencing are not directly comparable between different samples otherwise.

Pathway and Network Analyses

Extended methods of pathway and network analysis as well as statistical analysis are available in Data S6 and S7. Briefly, validated miRNA‐mRNA target genes of differentially expressed miRNAs were obtained by MiRTarBase (retrieved in July 2023) and used for the following analysis. ²⁰ Enriched Reactome 2022 pathways were generated by testing MiRTarBase‐derived mRNA targets with EnrichR. ²¹ The pathway network and cluster analysis were designed with Gephi 0.9.2. The nodes represent enriched pathways, and the node size is proportional to −log (q value) of each pathway. Connection thickness is proportional to the Jaccard Index, which estimates pathway similarity between nodes. Unbiased clustering of pathways was obtained through modularity assessment. ²² miRNA‐mRNA target analysis was performed using differentially downregulated miRNAs for each condition and respective mRNA targets.

In Vitro Experiments

Human kidney 2 cells (ATCC, CRL‐2190–passage 7–10) were transfected with mirVana miRNA inhibitors for hsa‐miR‐320b and hsa‐miR‐10b‐5p (Applied Biosystems, number 4464084) with Lipofectamine 2000 Reagent (Invitrogen, number11668027). Details on cell culture miRNAs, mRNAs, and protein quantification (by Western blot and flow cytometry) are provided in Data S3 and S8 through S10.

RESULTS

Subject Characteristics

Participants had a mean age of 59±8 years, a high average body mass index of 31.0±4.4 kg/m², and 24‐hour ambulatory blood pressure values within the normal range (Table). Twenty‐four‐hour urinary sodium excretion showed adherence to the prescribed diet (median 183.0 [108.7–255.6] mmol/d in the HSD versus 14.6 [7.7–24.2] mmol/d in the LSD]. Aldosterone and plasma renin activity were appropriately low on the HSD and increased on the LSD.

Table 1.

Clinical and Biochemical Characteristics of Enrolled Subjects (N=14)

Variable	Value
Age, y	59±8
Sex (%)
Women	6 (42.9)
Men	8 (57.1)
BMI, kg/m2	31.0±4.4
ABPM 24‐h SBP, mm Hg	116±10
ABPM 24‐h DBP, mm Hg	69±9

	LSD	HSD	P value
24‐h urinary sodium, mmol/d	17.3±10.6	178.9±76.2	<0.001
Urinary spot sodium/creatinine, mmol/L per mg/dL	12.5±8.1	116.6±40.7	<0.001
24‐h urinary volume, mL	1723±591	1974±641	0.182
SBP, mm Hg	119±10	115±9	0.291
DBP, mm Hg	78±6	70±9	<0.001
PRA seated, ng/mL per h	3.25 (1.88–6.15)	0.40 (0.10–0.85)	0.001
Aldosterone seated, ng/dL	24.5±11.0	9.0±4.2	<0.001
Serum creatinine, mg/dL	0.88±0.16	0.80±0.17	0.006
eGFR, mL/min per 1.73 m2	86±12	93±13	0.011
Serum sodium, mmol/L	138±2	139±2	0.013
Serum potassium, mmol/L	4.4±0.2	4.2±0.2	0.033
Weight, kg	85.2±12.5	86.7±12.5	<0.001

The table shows the clinical characteristics of recruited subjects and biochemical features following an LSD and HSD. Variables are reported as mean±SD, median (interquartile range), or absolute number (percentage), as appropriate. Differences were considered significant when P<0.05. Significance was calculated with a paired t test for normally distributed variables and Wilcoxon signed rank test for nonnormally distributed variables. ABPM indicates ambulatory blood pressure monitoring; BMI, body mass index; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; HSD, high‐sodium diet; LSD, low‐sodium diet; PRA, plasma renin activity; and SBP, systolic blood pressure.

uEVs Characterization and Small RNA Sequencing

To investigate the enrichment of small RNA cargos in uEVs after the LSD and HSD, we isolated uEVs from 24‐hour urine home collection through serial centrifugation and 3‐hour ultracentrifugation (Figure 1). We characterized uEV size, distribution, morphology, and protein content through nanoparticle tracking analysis, Western blot, and transmission electron microscopy as shown in Figure S1. The mean size of uEVs of characterized samples was 216±20 nm in the LSD and 218±31 nm in the HSD, and the mean concentration was 1.3×10¹¹±4.1×10¹⁰ and 7.5×10¹⁰±6.0×10¹⁰ particles/mL in the LSD and HSD, respectively.

Small RNA sequencing identified 111 small RNAs (Figure 2A), including 30 differentially expressed between the 2 diets (Figure 2C and Table S1): 21 piwi‐interacting RNAs, 8 miRNAs, and 1 ribosomal RNA (Figure 2B). Focusing on miRNA expression, 63 miRNAs were identified in uEVs, including 8 miRNAs differentially expressed in the 2 dietary conditions (Figure 2D). miR‐320a‐3p, miR‐99a‐5p, miR‐320b, and miR‐221‐3p were significantly upregulated under the LSD (Figure 2E); let‐7f‐5p, miR‐10a‐5p, miR‐10b‐5p, and miR‐27b‐3p were significantly upregulated under the HSD (Figure 2F). The bioinformatic analysis of miRNA expression in the healthy human kidney (retrieved from the DIANA‐microRNA Tissue Expression Database), showed that 62 out of 63 miRNAs detected in uEVs were moderately high or highly abundant in kidney tissue (Figure S2; Tables S2 and S3). Notably, miR‐10b‐5p, which was upregulated in the HSD, was the most abundant miRNA both in kidney tissue and uEVs. The correlation matrix (Figure S3A; Table S4) shows a moderate positive correlation between natural logarithm 24‐hour urinary sodium excretion and miRNAs upregulated in the HSD (let‐7f‐5p, miR‐10a‐5p, miR‐10b‐5p) and a trend toward a significant moderate negative correlation with miR‐320b (which was downregulated in the HSD). The scatterplot matrix (Figure S3B) showed a significant inverse correlation between miR‐320b and miR‐10b‐5p, suggesting a parallel and dose‐dependent increase of proinflammatory pathways and a reduction of anti‐inflammatory pathways.

Figure 2. Small RNA sequencing profiles in urinary extracellular vesicles of 14 subjects after high‐ and low‐sodium diets.

A and B, The ring chart shows the proportion of all the expressed (A) and differentially expressed small RNAs (B) highlighted for type. C and D, Volcano plots reporting expression levels of all small RNAs (C) and only miRNAs (D) expression levels when the HSD and LSD were compared. The dashed line is set at q value=0.05. Significantly differentially enriched small RNAs or miRNAs are highlighted in yellow (with violet border). The y axis is expressed as −log (q value); the x axis is expressed as log₂ (fold change). E and F, Dot and box plots of significantly differentially expressed miRNAs upregulated in the LSD (E) and HSD (F). EV indicates extracellular vesicle; HSD, high‐sodium diet; and LSD, low‐sodium diet.

Pathway‐Network Analysis

To determine the biological role of differentially expressed miRNAs, we used MiRTarBase to predict validated mRNA targets and considered the canonical function in miRNAs of negative translational regulators of mRNAs. Therefore, for the downregulated miRNAs, we expected an upregulation of the relative targets. miRNAs downregulated in the HSD targeted 1083 mRNAs (Table S5), whereas miRNAs downregulated in the LSD targeted 1321 mRNAs (Table S6). To identify biological processes enriched in each dietary condition, we used the Reactome 2022 database for pathway enrichment analysis. For clarity of representation, we showed, for each dietary condition, the pathways associated with downregulated miRNAs, thereby associated with predicted upregulated mRNA targets (Figures 3A and 3B; Tables S7 and S8). In the HSD, we observed an enrichment of 326 pathways [−log (q value) >1.5] (Table S7), including those related to translation regulation, immune system, and senescence (Figure 3A). In the LSD, we identified an enrichment of 39 pathways [−log (q value) >1.5] (Table S8), including some related to p53 regulation of cell cycle, white adipocyte differentiation and erythropoietin, platelet‐derived growth factor, and epidermal growth factor receptor signaling (Figure 3B).

Figure 3. Bioinformatic network analysis of predicted differentially regulated mRNA targets of differentially enriched miRNAs.

A and B, Pathway enrichment analysis conducted with EnrichR showing the enriched terms for miRNA target genes expected to be upregulated in the high‐sodium diet (A) or in the low‐sodium diet (B). Adjustment for multiple testing of the P value was performed with Benjamini‐Hochberg to calculate q values; the −log (q value) was calculated for pathway bar graph and network analysis. Pathways uniquely enriched in each condition were considered for pathway and network analysis. C and D, Network cluster analysis showing pathways enriched in the high‐ (C) and low‐sodium diet (D). The node size is proportional to –log (q‐value) of each pathway. For clarity of representation, pathways with –log (q value) <1.5 were excluded in the low‐sodium diet network and –log (q value) <2.0 in the high‐sodium diet. Connection thickness is proportional to the Jaccard Index. The main functional communities were named on the basis of the principal enriched pathways for each cluster. E and F, miRNA‐mRNA target networks showing miRNA‐target interactions in the high‐ (E) and low‐sodium diet (F). Yellow dots indicate miRNAs downregulated in the high‐sodium diet (E) and red dots in the low‐sodium diet (F). Blue dots indicate mRNA targets and are connected to regulating miRNAs. mRNA targets involved in immune‐related clusters (adaptive immune system, innate immune system, interleukin signaling, and interferon signaling) are highlighted in green in the high‐sodium diet (E). mRNA targets involved in PPAR‐related clusters are highlighted in green in the low‐sodium diet (F). In the high‐sodium diet (E), the connection between miR‐320b and ICAM‐1 was added manually since being validated by a previously published study. ²⁶ EGFR indicates epidermal growth factor receptor; ER, endoplasmic reticulum; ESR, estrogen receptor; FLT3, Fms‐like tyrosine kinase 3; GTP, guanosine triphosphate; HOX, homeobox; ICAM‐1, intercellular adhesion molecule 1; MAPK, mitogen‐activated protein kinase; PDGF, Platelet‐Derived Growth Factor; PI3K/akt, phosphoinositide 3‐kinase/protein kinase B; PPAR, peroxisome proliferator‐activated receptor; ROBO, roundabout; RT, regulatory target; SMAD, suppressor of mothers against decapentaplegic; SUMO, small ubiquitin‐like modifier; TGF‐β, transforming growth factor‐beta; TP53, tumor protein P53; and TR, transcriptional regulation.

To further investigate the biological effects of differentially enriched mRNA targets, we analyzed the interactions between significantly enriched pathways through networks‐cluster analysis. In the HSD, we identified an enrichment of pathway clusters related with cell cycle, translation regulation, transforming growth factor β pathways, and several clusters related to the immune system (adaptive and innate immune system, interleukin signaling, and interferon signaling) (Figure 3C; Table S7). In the LSD, we identified clusters of pathways related to signaling of erythropoietin and platelet‐derived growth factor, p53 transcriptional regulation, and PPAR (peroxisome proliferator‐activated receptor) transcription pathways (Figure 3D; Table S8).

miRNA‐mRNA Target Networks

To select potential mRNA targets responsible for the effects of the HSD at the kidney level, we built an miRNA‐mRNA target network, highlighting miRNA‐mRNA targets involved in specific pathways of interest. In these networks, each differentially expressed miRNA was connected to predicted and validated mRNA targets. Considering that previous preclinical studies suggested that low‐grade renal inflammation may favor the development of salt‐sensitive hypertension, ²³ we leveraged the miRNA‐mRNA target networks for the detection of mRNA targets involved in immune system regulation. In the HSD, we highlighted mRNA targets involved in immune‐related clusters (adaptive immune system, innate immune system, interleukin signaling, and interferon signaling) (Figure 3E; Figure S4; Table S9).

Considering the anti‐inflammatory role of PPARα at the renal level, ²⁴ we highlighted the mRNA target involved in PPAR transcription cluster in the LSD (Figure 3F; Figure S5; Table S10). mRNA targets related to immune system regulation were uniformly connected to all the miRs differentially expressed in both dietary conditions, suggesting that the enrichment of these pathways was not exclusive of 1 or few miRNAs. Hence, we identified 2 potential mRNA targets in the HSD and LSD as candidates for in vitro validation. ICAM‐1 (intercellular adhesion molecule 1) (Figure 3E; Figure S4) is a validated target of miR‐221 in human cholangiocytes ²⁵ and of miR‐320b in human endothelial cells ²⁶ (although this last validation was not present in MiRTarBase). PPARα, which is predicted as upregulated in the LSD (Figure 3F; Figure S5), is a validated target of miR‐10b in human hepatocytes. ²⁷ The 3′‐UTR‐specific targeting was previously demonstrated for miR320b/ICAM‐1 and miR10b‐5p/PPARα by luciferase assays. ²⁶ , ²⁷

In Vitro Studies

Because both ICAM‐1 and PPARα were validated as mRNA targets in nonrenal cells, we designed a proof‐of‐concept in vitro study of both mRNA targets in human kidney 2 cells. The transfection of miR‐320b inhibitor in human kidney 2 cells resulted in an efficient knockdown of miR‐320b (Figure 4A), 2.17‐fold increase of ICAM‐1 mRNA expression (Figure 4B), and 2.21‐fold increase in ICAM‐1 protein (Figure 4C), as assessed by flow‐cytometry. The transfection of miR‐10b‐5p inhibitor in the same cells resulted in an efficient knockdown of miR‐10b‐5p (Figure 4D), with consequential nonsignificant modification of PPAR‐α mRNA (Figure 4E), but 2.85‐fold increase in PPAR‐α protein (Figure 4F; Figure S6), assessed by Western blot analysis.

Figure 4. In vitro validation of miRNA‐target regulation in HK‐2 cells.

A through C, Relative expression levels of miR‐320b (A), mRNA of ICAM‐1 (B), and protein levels of ICAM‐1 (C) in HK‐2 cells after treatment with miR‐320b inhibitor. D through F, Relative expression levels of miR‐10b‐5p (D), mRNA of PPARα (E), and protein levels of PPARα (F) in HK‐2 cells after treatment with miR‐10b‐5p inhibitor. Expression levels were normalized to miR‐16‐5p expression (a.u.) for miRNAs, to GAPDH expression (a.u.) for mRNA. Protein abundance was normalized for vinculin expression (a.u.). A, B, D, and E, Quantified by quantitative real‐time polymerase chain reaction, median fluorescence of each experiment was normalized by Z score (a.u.). C and F, Protein expression was quantified by flow‐cytometry and Western blot respectively. Representative Western blot analysis of PPARα and vinculin is presented in Figure S6. In all experiments, the control group was treated with the empty vehicle (Lipofectamine 2000) alone. Data are represented with dot plots. Each experiment was repeated twice with 3 technical replicates (n=6). Statistical differences were assessed by the use of an unpaired t test. *P<0.05. a.u. indicates arbitrary unit; Ctr, control; GAPDH, glyceraldehyde 3‐phosphate dehydrogenase; HK‐2, human kidney 2; ICAM‐1, intercellular adhesion molecule 1; and PPARα, peroxisome proliferator‐activated receptor α.

DISCUSSION

In a cohort of normotensive individuals with normal kidney function who underwent 2 dietary sodium interventions, we demonstrated for the first time the impact of dietary sodium modulation on the small‐RNAome of uEVs. Bioinformatic analyses revealed that the miRNAs differentially expressed between the 2 dietary conditions targeted genes enriched in pathways related to immune system activation in the HSD and anti‐inflammatory pathways in the LSD. Human proximal tubular cells were then used in vitro to validate the miRNA‐mRNA targets of interest. The results showed a negative regulation of ICAM‐1 by miR‐320b (upregulated in LSD) and PPARα by miR‐10b‐5p (upregulated in the HSD), suggesting a proinflammatory role of uEVs miRNA cargo during HSD.

In the past 3 decades, many studies suggested that long‐term sodium excess induces hypertension through several mechanisms that go beyond short‐term hemodynamic effects. ⁶ , ⁷ Clinical studies using a relatively short‐term regimen of the HSD in individuals with normotension (with timing and sodium intake similar to our study) showed nonsignificant modifications of BP levels. ²⁸ , ²⁹ However, the earliest alterations of endothelial function were evident even after 1 week of an HSD. ³⁰ Results from preclinical studies showed that the relationship between an HSD and cardiovascular risk involves a multiple array of pathways, including aging, endothelial dysfunction, redox signaling, and immune system activation. ⁶ , ³¹ The effects of an HSD on target organs are mediated by multiple factors: high extracellular sodium concentration, ⁷ alterations of hormonal mechanisms, including regulation of the renin‐angiotensin‐aldosterone system, sympathetic nervous system activation, ⁶ and calcium metabolisms. ³²

In animal models, sodium excess stimulates systemic innate immune response, with increased circulating neutrophils ³³ and monocytes. ³³ , ³⁴ Activated antigen presenting cells polarize T cells toward an IL (interleukin)‐17‐producing phenotype, with the activation of an adaptive immune system and migration of inflammatory cells to target organs. ⁷ , ³⁵ In the present study, the miRNA set found altered after the HSD was connected with a series of gene targets that, in a network analysis, showed an enrichment of both adaptive and innate immune system regulation, corroborating the findings from animal models after high sodium intake. ⁷ , ³³ , ³⁵ Moreover, we identified significant enrichment in pathways related to interferon signaling, aligning with previous findings in a murine model where high sodium conditions activated interferon‐γ‐secreting regulatory T cells, triggering a proinflammatory response. ³⁶

The examination of human kidneys from patients with hypertension showed an accumulation of monocyte, macrophage, and dendritic cells in the corticomedullary junction and into the deeper regions of the medulla, where sodium concentrations are markedly increased. ³⁷ The local complex interplay between innate and adaptive immunity further stimulates the proinflammatory milieu induced by sodium excess. ³⁶ Infiltration of inflammatory cells in the kidney is therefore crucial for the development of low‐grade renal inflammation induced by sodium.

In this study, we showed that miR‐320b is downregulated in uEVs after the HSD and miR‐320b inhibition increases ICAM‐1 mRNA and protein expression in the human tubular cell line. In the kidney, ICAM‐1 is expressed in the renal vascular endothelium and in proximal tubular epithelium, especially under pathological stimuli ³⁸ , ³⁹ and is associated with interstitial infiltration of inflammatory cells. ³⁸ , ⁴⁰ The administration of an HSD increases leukocyte adhesion and the renal expression of MCP‐1 (monocyte chemotactic protein‐1) and ICAM‐1, ³⁵ through mechanisms that are independent from sodium‐induced hypertension. ³⁵ Moreover, ICAM‐1 expression is induced by interferon‐γ. ²⁵ Considering the observed enrichment of interferon signaling following a high‐sodium diet, we speculate that multiple mechanisms may work synergistically to increase ICAM‐1 expression in the kidneys in response to sodium load.

In the LSD, the down‐regulated miRNAs targeted a set of genes enriched in pathways related to white adipocyte differentiation and PPAR transcriptional regulation. In the kidney, PPARα is expressed in medullary thick ascending limbs and the proximal tubular epithelium where it plays a critical role in metabolic regulation and anti‐inflammatory activity. ²⁴ In a transgenic mouse model, the increased expression of proximal tubule PPARα reduced adhesion molecules, proinflammatory cytokines, and the infiltration of inflammatory mononuclear cells. ⁴¹ On the other side, diabetic PPARα‐knockout mice display increased renal macrophage infiltration and more severe diabetic renal disease than diabetic wild‐type mice. ⁴² In our study, we showed that miR‐10b‐5p was downregulated after an LSD and that this miRNA inhibition increases the expression of PPARα protein in human proximal tubular cells, potentially contributing to the anti‐inflammatory pathways of the LSD. It should be noticed that miR‐10b‐5p inhibition increased PPARα protein, without significant modification of PPARα mRNA. This finding is consistent with the results of a previous study that showed that miR‐10b regulates PPARα posttranscriptionally in human hepatocytes, without significant modification of mRNA levels. ²⁷

Beyond the anti‐inflammatory effects, PPARα also has nephroprotective properties through regulation of renal fibrosis. ⁴¹ In particular, PPARα reduces tubulointerstitial fibrosis through reduction of proximal tubule expression of transforming growth factor β. ⁴¹ Intriguingly, in our study, several transforming growth factor β‐related pathways were enriched in the HSD. We can speculate that a relative reduction of PPARα activity in HSD can further enhance transforming growth factor β activation leading to interstitial fibrosis with sodium excess.

The strengths of the study are the careful selection of patients, the tight surveillance of dietary compliance, and the within‐patient comparison of the biomolecular effects of the 2 diets, avoiding potential confounding factors of between‐patient variability. Another strength is that the HSD was defined according to the recommendations of international guidelines, ⁴ , ⁵ with values of sodium intake being similar to the sodium intake in the Western diet. ⁴³

A limitation of the study is the relatively small sample size; larger studies are needed to validate these findings and enhance their generalizability, including different cohorts in terms of age, sex, and comorbidities. Additionally, it is possible that, beyond the designed extreme dietary sodium manipulation, other uncontrolled confounding factors due to environmental, dietary, or behavioral factors could have also influenced the findings. Moreover, further longitudinal studies should assess the long‐term effects of dietary sodium modulation on small RNA expression and renal physiology. In regard to the proof‐of‐concept in vitro validation, we acknowledge that the study design does not allow a complete understanding of the complex mechanistic link between miR‐320b and miR‐10b‐5p regulation of ICAM‐1 and PPARα and sodium‐induced renal inflammation. Future studies should investigate these pathways further in vitro and in vivo. Finally, it should be considered that we used uEVs as a readout of biological processes performed in cells from the urogenital tract. This approach is the most used and widely accepted, considering that EV cargos are actively loaded into EVs by parental cells. ⁸ However, it could not be excluded that the incorporation of miRNAs into uEVs serves as a mechanism of downregulating miRNA expression within the cell.

In summary, the analysis of the small‐RNAome of human uEVs revealed that an HSD is associated with the enrichment of proinflammatory pathways that potentially contribute to the development of hypertension and kidney damage. Our findings unravel an important role of miRNAs in the regulation of critical gene targets, that suggests upregulation of ICAM‐1 in an HSD and upregulation of PPARα in an LSD.

Sources of Funding

A.V. was supported by the National Institutes of Health award R01 DK115392.

Disclosures

P.M. reports consulting fees unrelated to the contents of this work from Diasorin. A.V. reports consulting fees from Corcept, HRA Pharma, and Mineralys, unrelated to the current work. The remaining authors have no disclosures to report.

Supporting information

Data S1

Tables S1–S10

Figures S1–S6

References 44–46