Serum Alpha-1-Acid Glycoprotein-1 and Urinary Extracellular Vesicle miR-21-5p as Potential Biomarkers of Primary Aldosteronism

Primary aldosteronism (PA) is the most common cause of secondary hypertension and reaches a prevalence of 6-10%. PA is an endocrine disorder, currently identified as a broad-spectrum phenotype, spanning from normotension to hypertension. In this regard, several studies have made advances in the identification of mediators and novel biomarkers of PA as specific proteins, miRNAs, and lately, extracellular vesicles (EVs) and their cargo. Aim To evaluate lipocalins LCN2 and AGP1, and specific urinary EV miR-21-5p and Let-7i-5p as novel biomarkers for PA. Subjects and Methods A cross-sectional study was performed in 41 adult subjects classified as normotensive controls (CTL), essential hypertensives (EH), and primary aldosteronism (PA) subjects, who were similar in gender, age, and BMI. Systolic (SBP) and diastolic (DBP) blood pressure, aldosterone, plasma renin activity (PRA), and aldosterone to renin ratio (ARR) were determined. Inflammatory parameters were defined as hs-C-reactive protein (hs-CRP), PAI-1, MMP9, IL6, LCN2, LCN2-MMP9, and AGP1. We isolated urinary EVs (uEVs) and measured two miRNA cargo miR-21-5p and Let-7i-5p by Taqman-qPCR. Statistical analyses as group comparisons were performed by Kruskall-Wallis, and discriminatory analyses by ROC curves were performed with SPSS v21 and Graphpad-Prism v9. Results PA and EH subjects have significantly higher SBP and DBP (p <0.05) than the control group. PA subjects have similar hs-CRP, PAI-1, IL-6, MMP9, LCN2, and LCN2-MMP9 but have higher levels of AGP1 (p <0.05) than the CTL&EH group. The concentration and size of uEVs and miRNA Let-7i-5p did not show any difference between groups. In PA, we found significantly lower levels of miR-21-5p than controls (p <0.05). AGP1 was associated with aldosterone, PRA, and ARR. ROC curves detected AUC for AGP1 of 0.90 (IC 95 [0.79 – 1.00], p <0.001), and combination of AGP1 and EV-miR-21-5p showed an AUC of 0.94 (IC 95 [0.85 – 1.00], p<0.001) to discriminate the PA condition from EH and controls. Conclusion Serum AGP1 protein was found to be increased, and miR-21-5p in uEVs was decreased in subjects classified as PA. Association of AGP1 with aldosterone, renin activity, and ARR, besides the high discriminatory capacity of AGP1 and uEV-miR-21-5p to identify the PA condition, place both as potential biomarkers of PA.


INTRODUCTION
The etiology of arterial hypertension (AHT) is unknown in more than 80-90% of cases, which is named essential hypertension (EH). One third of EH has been suggested to be associated with endocrine disorders (1). Primary aldosteronism (PA) is an endocrine disorder, currently identified as a broad-spectrum phenotype, spanning from normotension (4% prevalence) to hypertension (10% prevalence) (2)(3)(4)(5)(6). PA is characterized by an inappropriately high circulating aldosterone independent of known physiological regulators such as renin, angiotensin II, potassium, and sodium status (e.g., high saline intake) (7). The diagnosis of PA is relevant, not only for its association to high blood pressure but also for the harmful effects in extra-renal tissues, generally associated with the mineralocorticoid receptor (MR) activation by aldosterone which induces inflammation (8,9), tissue remodeling, and fibrosis (8,(10)(11)(12)(13)(14), affecting the renal, heart, the vascular system (endothelial cells and smooth muscles cells), the immune system (15) and the adipose tissue (16).
The current study aimed to evaluate lipocalins LCN2 and AGP1, and specific urinary extracellular vesicles miR-21-5p and Let-7i-5p as novel biomarkers of primary aldosteronism.

METHODS
This study used the serum, plasma, and urine samples from a biobank obtained from a cohort of 206 adult Chilean subjects of both genders, between 18 and 65 years old. The subjects were recruited from outpatient centers associated with the UC-Christus Health Network in Santiago, Chile, following the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of the Faculty of Medicine of the Pontificia Universidad Catoĺica de Chile (Certification CEC-MEDUC 12-207 and 14-268, and updated by CEC-MEDUC 190823001 and 200619004).
All subjects had a sodium diet ad libitum and declared that they did not ingest any herbal products or extreme diets during the month prior to the analysis. Subjects with a BMI >30 kg/m 2 , kidney disease, diabetes mellitus, liver, and heart failure were excluded. Subjects using glucocorticoids, contraceptives, or some interfering drugs, such Ag-II-receptor blockers (ARB), ACE-Inhibitors (ACEI), and spironolactone (MR antagonist), were also excluded.
After exclusion criteria were applied, 132 subjects were included in the study. The subjects were classified as normotensive controls (CTL), have clinical and biochemical parameters in the normal range, essential hypertensives (EH) according to the 2017 ACC/AHA Guidelines for High Blood Pressure (52), and subjects having a positive screening for PA (Aldosterone >9ng/dl, PRA <1 ng/ml*h), according to The Endocrine Society 2016 guidelines (53) and Vaidya et al. (5,(54)(55)(56). All studied subjects (PA, EH, and CTL) have a clinical record including medical history and physical examination, as well a biochemical profile, creatinine, electrolytes, aldosterone, plasma renin activity (PRA), serum, and 24-hour and morning urine samples. Aldosterone and PRA were measured by immunoassay using a commercial kit (DiaSorin, Stillwater, MN). Urine samples for uEV isolation were stored at -80C with a 1X protease inhibitor cocktail (Roche, USA).

Evaluation of the Parameters Associated With Inflammation, Endothelial, and Renal Damage in PA Subjects
The inflammatory status of all subjects was evaluated by measuring hs-CRP with a nephelometric assay (BN ProSpec Systems; Siemens Healthcare Diagnostics Products, Marburg, Germany) and IL-6 by an ELISA with commercial reagents and standards (D6050, R&D Systems, Minneapolis, MN), according to the manufacturer's protocols. Endothelial damage was evaluated by surrogate markers such as PAI-1, MMP9, and MMP2 activities. PAI-1 was determined by ELISA (HYPHEN BioMed, Neuville sur Oise, France), and MMP9 and MMP2 activities by zymography, as previously described (57). Early renal damage was evaluated with 24-hour urine albuminuria to creatinine ratio (UACR). Albumin is measured by a turbidimetric immunoassay (Roche, Germany), and urine creatinine was measured with a colorimetric assay (Roche, Indianapolis, IN) in a Hitachi Automatic Analyzer 7600 (Roche/Hitachi, Kobe, Japan). Plasma and urinary electrolytes (sodium and potassium) were evaluated with methods previously described (58).

Isolation and Characterization of Urinary Extracellular Vesicles From PA Subjects
Urinary EVs (uEVs) were isolated by a sequential ultracentrifugation protocol previously described by Barros et al. (26). Urinary creatinine was used to normalize samples of uEVs (59,60). Isolated uEVs were characterized as previously described (26,31) and according to the International Society for Extracellular Vesicles guidelines (27) using transmission electron microscopy (TEM), nanoparticle tracking analysis (NTA), and western blot with characteristic EV proteins (61).
TEM was performed with 15 ml of uEVs suspension were absorbed onto a 200 mesh carbon-coated copper grid for 1 min. Samples were negatively stained with 2% uranyl acetate solution for 1 min. Grids were visualized in a Phillips Tecnai transmission electron microscope at 80 kV and images were acquired using a SIS-CCD Camera Megaview G2 (62). The concentration and size of uEVs were determined by nanotracking analyses (NTA) performed in a low-volume flow cell (LVFC) of a NanoSight NS300 and NTA 3.2 software (Malvern Instruments Ltd, Malvern, UK). Camera level and detection threshold was optimized to identify individual particles and minimum background noise during recordings (camera level = 12-14; detection threshold = 3-5; flow speed = 50). Particles were tracked by passing a laser beam through the liquid sample and the scattered light was detected and captured in short videos by a sCMOS camera (3 videos of 20 seconds each). The Brownian motion of particles was determined, and the distance moved by the detected particles will be used to calculate the diameter (mean and mode size) and concentration of vesicles using the Stokes-Einstein equation (63).

Urinary EV RNA Isolation
RNA from the extracellular vesicle was isolated by organic extraction using the Trizol ® reagent according to the manufacturer's instructions. Two microliters of each RNA sample were pipetted on the NanoQuant Plate ™ of the Infinite ® M200 PRO spectrophotometer (TECAN; Männedorfl; Switzerland) to quantify the RNA concentration (A260 nm) and purity (A260/A280 nm ratio) using Tecan i-control ™ software.

Expression of miR-21-5p and Let-7i-5p in Urinary Extracellular Vesicles
Reverse transcription of miRNA samples was performed using the TaqMan ™ Advanced miRNA cDNA Synthesis Kit (A28007), according to the manufacturer's instructions. The expression of miRNAs (Hsa-miR-21-5p and Hsa-let-7i-5p) were evaluated with TaqMan ™ Advanced miRNA Assay (A25576) and the TaqMan ™ Fast Advanced Master Mix (4444557, Applied Biosystems) in the RotorGene 6000 thermocycler (Corbett Research, Sydney, Australia). The amplification reactions were performed as follows: Enzyme activation at 95°C for 20 seconds and 40 cycles of 95°C for 3 seconds, anneal/extend at 60°C for 30 seconds. RNU6 snRNA was used as an internal normalization control (TaqMan ™ MicroRNA Assay, ID001973). The fold changes of miRNA expression were calculated using the relative cycle threshold (2−DDCt) method and further normalized by the spot urinary creatinine. Unpaired Kruskal-Wallis test was performed to identify differences in PA patients versus EH and healthy controls.

Statical Analyses
Clinical, biochemical, and expression data are expressed as median [Q1-Q3]. Data normality was determined by Kolmogorov-Smirnov test. For parametric and non-parametric comparisons between two sets of data, an unpaired Student t-test or a Mann-Whitney test were performed. To assess differences between groups of data and an independent variable, a one-way Analysis of Variance (ANOVA) or Kruskal Wallis was performed using a Tukey or Dunn post hoc test, respectively. Associations were performed by linear regression by Pearson or Spearman regression according to data normality.
Receiver operating characteristic (ROC) analysis was used to test the ability of lipocalins (LCN2, AGP1) and uEV-associated miRNAs (miR-21-5p and Let-7i-5p) to discriminate PA patients from EH and control subjects. A p value < 0.05 was considered statistically significant. Data were analyzed using GraphPad Prism v9.1 (GraphPad, CA, USA) or SPSS v21 (IBM, USA) software.

Clinical and Biochemical Characteristics of Subjects With PA
We identified 11 PA subjects (8.3%) in our cohort of study according to the PA criteria described in the Methods section. Clinical and biochemical baseline characteristics are shown in Table 1 Figure 2). No differences were found in plasma and urinary sodium and potassium electrolytes, nor in the fractional excretion of potassium (FEK) or the fractional excretion of sodium (FENa) in PA, EH, and controls ( Table 1).   Table 3). Total LCN2 was found to be higher in EH with respect to the control group, meanwhile, PA was similar to EH but did not reach a significant difference when compared to the control ( Figure 4). We observed significant associations of AGP1 with Aldosterone (rho = 0.34, p <0.05), with PRA (rho = -0.44, p <0.01) and with ARR (rho = 0.38; p <0.05) ( Figure 5).

Characterization and Quantification of Urinary Extracellular Vesicles
We isolated uEVs from all subjects in this study. Figure 6 shows a representative image of isolated uEVs with a donut-shape morphology by TEM ( Figure 6A), a characteristic plot size/ concentration from NTA with the main peak near to 150 nm ( Figure 6B), and the western-blot of EV markers CD9 and TSG101 ( Figure 6C). No differences were observed in concentration, mean and mode size of uEVs measured by NTA in PA, EH, and controls ( Table 4 and Figure 7).

Expression of miR-21-5p and Let-7i-5p in Urinary Extracellular Vesicles
We identified a low expression of miR-21-5p in uEVs in PA and EH in the control group. No difference was detected of miR-21-5p between the PA and EH group ( Table 5). Concerning Let-7i- 5p, we did not detect any differences between all groups ( Table 5 and Figure 8).

DISCUSSION
In the present study, we evaluated the concentration of lipocalins AGP1 and LCN2, and the expression of miR-21-5p and Let-7i-5p in uEVs as potential biomarkers of PA. We observed a higher concentration of AGP1 in PA subjects, which is associated with the critical variables used to screen PA, as plasma aldosterone, renin, and ARR. Further to these novel findings, we noted a low expression of miR-21-5p in PA subjects, which is interesting since it supports a combinate model for the identification of PA conditions. We suggest that both AGP1 and miR-21-5p are associated with the pathogenic course of the primary aldosteronism and can be useful in the design of a novel diagnostic algorithm for PA. There is also a widely accepted   consensus regarding a positive screening for PA is an ARR >30 ng/dL per ng/mL/h, with suppressed renin (PRA<1.0 ng/mL*h or DRC <10 uUI/ml) and an aldosterone concentration >15 ng/ dL. Some studies have been identified that can improve the detection of milder forms of primary aldosteronism when using less conservative ARR thresholds with suppressed renin activity and plasma aldosterone levels >9 ng/dL (5,53,54), which is in agreement with the outcome of this study. We found similar levels of hs-CRP and IL-6 as markers of inflammation in PA subjects, which were similar to EH and control groups, according to previous studies (12,64). Similarly, endothelial damage markers (PAI-1, MMP9, and MMP2) and renal function markers (Urinary albumin (UACR)) do not show any significant changes in PA compared with EH and controls. It suggests these subjects, currently classified as subclinical PA (54,65) do not have chronic inflammation, vascular compromise, or renal function impairment as is seen in overt or classic PA. Hence, is highly necessary novel and sensitive biomarkers aimed to detect subclinical PA and avoid complications associated with the renal and extra-renal effects reported in classic PA.
This perspective is the first to report findings that show a higher serum AGP1 concentration in PA than EH and controls subjects (Figure 4). We also observed a significant association of AGP1 with classic screening parameters for PA (e.g., aldosterone,  PRA, and ARR) ( Figure 5). Moreover, we found by discriminative analyses by ROC curves that AGP1 can identify PA from HE&CTL subjects with high sensitivity and specificity. All these results suggest that circulating AGP1 protein is a novel and potential biomarker of PA, which was also suggested for AGP1 protein in urinary exosomes (26). Since AGP1 is a protein from the family lipocalin associated with the acute phase response with immunomodulatory properties (66,67), affected by glucocorticoids (68)(69)(70) and mineralocorticoids (71), we suggest that AGP1 expression is modified by high aldosterone    levels through MR activation, having a dual role as a potential biomarker of PA, and possible mediator of the tissue response to high aldosterone. Further clinical and animal model or in-vitro studies using MR antagonists should be performed to support this hypothesis. Similarly, we measured free LCN2 and LCN2 conjugated with matrix metalloproteinase 9 protein as a potential biomarker of PA (72). We found an increase only in total LCN2 (the sum of free and complexed LCN2) in EH, but it did not reach a significant difference in PA when compared with the control or EH. LCN2 is a proinflammatory molecule upregulated in obese individuals or patients with cardiometabolic syndrome, as also has been described in classic PA (73,74) and is suggested as an MR sensitive protein (75,76). LCN2 expression is influenced by several factors including obesity, salt intake, aging, infection, and inflammatory status (72,74,(77)(78)(79). Since these subjects have a middle or subclinical PA, with no clear evidence of inflammation, renal/vascular damage (Figure 3), or concomitant cardiometabolic disease, we hypothesize the LCN2 fails to increase in these PA subjects since they require a concomitant hit as inflammation (78,79), obesity (high adipose tissue) (74), or high salt intake (72) to increase the circulating LCN2 levels.
We studied the urinary extracellular vesicles as a source of potential biomarkers for PA (26,30). In the present study, uEVs  showed similar particle concentration and size in PA subjects with respect to EH and controls ( Table 4). Previous studies in PA show similar findings in uEVs concentration (26) but differ from studies in circulating EVs in PA (75,76), where they reported an increased concentration of circulating EVs in the serum of PA patients when compared with essential hypertensives and attributed it to an enhanced biological response of the endothelium to aldosterone in vivo (75), which has also been observed in vitro (80)(81)(82). These differences could be related to the PA classification, overt PA versus subclinical PA, and also the different biofluids used to quantify the impact of high aldosterone in EV concentration, serum versus urine, which is associated with distinct mechanisms and the rates of EV shedding that have different tissues (e.g., vascular endothelium vs renal epithelia). Based on previous reports, we measured the expression of two miRNA in urinary EVs, miR-21-5p, and Let-7i-5p, as potential biomarkers of PA. We observed that uEV-associated miR-21-5p expression in uEVs from PA were lower than controls ( Figure 8) and similar to EH, however a trend to lower levels was observed in PA. This result suggests that uEV-miR-21-5p is downregulated and associated with pathophysiological mechanisms depending on both high BP and PA conditions. miR-21-5p expression is regulated by cytokines, inflammatory modulators (e.g., NF1, AP1), and steroids. Downregulation of miR-21-5p would affect the downstream target genes related with inflammation (83) as IL-1B gene, aldosterone effect as NEDD4, YWHAZ, SCL12A2 genes, and fibrotic processes (42,84) as COL1A and COL4A1 genes ( Table 6). Prospective animal models and in vitro studies with miR-21-5p are necessary to gain depth of understanding about the role of this miRNA in high aldosterone conditions in renal epithelia, as occurs in PA.
With respect to uEV-associated Let-7i-5p, we did not observe any differences in Let-7i-5p expression in all groups. Let-7i has been found in either urine (31) and plasma exosomes (44) and is associated with RAAS, mediating inflammation and fibrosis, in both in vitro models and experimental models of kidney disease (51,85). Let-7i regulates downstream target genes TGFBR1, IL6, IL10, COL1A1, COL3A1, DNMT3A, NEDD4, ORM1, VIM, FN1, ACTIN, SCL12A1, and YWHAZ, among others (85-87) ( Table 6). In the current study, we did not find differences in inflammation parameters, and were unable to measure other important parameters related to fibrosis in these PA subjects, such as the procollagen type 1 protein (PINP, COL1A1).
The ROC curves analyses with AGP1 and miR-21-5p as significant variables associated with PA subjects, support a simple (AGP1) or combinate model (AGP1 + miR-21-5p) to discriminate PA with significant AUC of 90% or 94%, respectively. This AUC is similar to previous reports on AGP1 in uEVs (92%), which support free or uEV-associated AGP1 as potential biomarkers of PA (26).
In summary, we found higher levels of serum AGP1 and lower uEV-miR-21-5p expression in subjects classified as PA with respect to EH and controls. Besides the high discriminatory capacity identified by ROC curves, the association of AGP1 with aldosterone, PRA, and ARR, place both as potential biomarkers of PA. Further studies examining the possible role of AGP1 and miR-21-5p as a mediator of the pathogenic course of PA are encouraged.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Unidad de Ética y Seguridad de la Investigacioń, CEC-SaludUC, Pontificia Universidad Catoĺica de Chile. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
CC, AT-C, and CF contributed to conception and design of the study. CC and CF provide with the contribution of patients or study material. CC, AT-C, JP performed the collection and/or obtaining of results, organized the database, and performed the statistical analysis. CC and AT-C achieve the analysis and interpretation of the results. CC wrote the first draft of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version.