Proteomic Profile of Urinary Extracellular Vesicles Identifies AGP1 as a Potential Biomarker of Primary Aldosteronism

Abstract

Context

Primary aldosteronism (PA) represents 6% to 10% of all essential hypertension patients and is diagnosed using the aldosterone-to-renin ratio (ARR) and confirmatory studies. The complexity of PA diagnosis encourages the identification of novel PA biomarkers. Urinary extracellular vesicles (uEVs) are a potential source of biomarkers, considering that their cargo reflects the content of the parent cell.

Objective

We aimed to evaluate the proteome of uEVs from PA patients and identify potential biomarker candidates for PA.

Methods

Second morning spot urine was collected from healthy controls (n = 8) and PA patients (n = 7). The uEVs were isolated by ultracentrifugation and characterized. Proteomic analysis on uEVs was performed using LC-MS Orbitrap.

Results

Isolated uEVs carried extracellular vesicle markers, showed a round shape and sizes between 50 and 150 nm. The concentration of uEVs showed a direct correlation with urinary creatinine (r = 0.6357; P = 0.0128). The uEV size mean (167 ± 6 vs 183 ± 4nm) and mode (137 ± 7 vs 171 ± 11nm) was significantly smaller in PA patients than in control subjects, but similar in concentration. Proteomic analysis of uEVs from PA patients identified an upregulation of alpha-1-acid glycoprotein 1 (AGP1) in PA uEVs, which was confirmed using immunoblot. A receiver operating characteristic curve analysis showed an area under the curve of 0.92 (0.82 to 1; P = 0.0055).

Conclusion

Proteomic and further immunoblot analyses of uEVs highlights AGP1 as potential biomarker for PA.

urine, vesicles, proteomics, hypertension, aldosterone-to-renin ratio

Primary aldosteronism (PA) is the most common cause of secondary hypertension, affecting approximately 10% of hypertension patients (1-3). PA occurs due to inappropriately high aldosterone production (4), and unregulated activation of the mineralocorticoid receptor (MR) by the excess of aldosterone, affecting renal sodium and water reabsorption (5). PA is associated with vascular and cardiac abnormalities, high blood pressure (1, 6), inflammation (7, 8), and fibrosis (9) that worsen PA prognosis and increase cardiovascular risk (10).

The prevalence of PA could be higher than previously thought, including a relatively common continuum of renin-independent aldosterone secretion that spans from normotension to overt PA (11). Recent studies have found that the prevalence of overt PA in normotensive individuals is 11.3%, compared with 21.6% and 22% in stage 1 and 2 of hypertension, respectively (12). In general, PA screening criteria (5) include sustained hypertension, hypokalemia, and a high aldosterone-to-renin ratio (ARR) (13, 14). However, the ARR thresholds used in clinical practice could be lowered, depending on the aldosterone and renin values and the age of the subject (3, 5, 8), which has been shown to have poor sensitivity and negative predictive value for overt PA (12). Early diagnosis and treatment is critical to prevent the deleterious long-term outcomes associated with PA, which not only include high blood pressure, but also MR activation-dependent end-organ damage in the cardiovascular, renal, and immune systems (15-18), as well as the adipose tissue (19), promoting a proinflammatory state (20).

The latter highlights the importance of identifying novel PA biomarkers that contribute to detect the pathophysiological changes associated with high mineralocorticoid activity. In this way, some peptides have been found to be increased in PA, including the high-sensitivity C-reactive protein (hs-CRP), plasminogen inhibitor activator-1, matrix metallopeptidase 9, and malondialdehyde (7, 8). Furthermore, markers for kidney function, including free cystatin-C, neutrophil gelatinase associated lipocalin (NGAL) (21-25), osteopontin (26), and also some microRNAs (27) are suggested to be associated with PA.

Extracellular vesicles (EVs) are promising biomarkers of disease in cancer, metabolic disorders, and cardiovascular diseases (28-30). EVs are biologically active and important signaling vehicles, and regulators of local and distant cell-cell communication (31), that carry proteins, lipids, RNA, and microRNA. All cell types release EVs into the extracellular space and biofluids (eg, plasma, cerebrovascular fluid, breast milk, urine, and saliva), with a cargo that reflects the activation status and phenotype of the parent cell (31).

Urinary EVs (uEVs), mainly originated from cells lining the renal tubules, are a potential source of biomarkers for renal pathologies (32-35), protected from RNases (36, 37) and proteases (38) proper from the urinary tract, which also regulate renal physiological and pathophysiological processes (39). The abundance of renal proteins in uEVs has shown to reflect their abundance in the kidney (40), turning them into useful tools for the diagnosis of salt-wasting tubulopathies (33). In addition, activation of the renin-angiotensin-aldosterone system (RAAS) has been reported to modulate the uEV proteome (41-43), and PA patients have shown to carry increased levels of serum EVs (44).

Considering the scarce information regarding biomarkers or potential mediators associated with renal and extrarenal damage induced by chronic aldosteronism, the current study aims to identify by a proteomic approach novel peptides or proteins present in uEVs from PA patients in order to identify potential biomarker candidates for PA.

Materials and Methods

Subjects

We recruited 132 individuals from 18 to 65 years of age from 2 primary care centers in Santiago, Chile; subjects included both genders and had similar socio-economic status and ethnicity. Subjects had ad libitum sodium diet, no potassium supplementation, and declared no ingestion of herbal products or extreme diets the month prior to the analyses. Those with body mass index (BMI) > 35 kg/m², kidney disease, diabetes mellitus, liver and heart failure, hypercortisolism, apparent mineralocorticoid excess, renovascular disease, and familial hyperaldosteronism were excluded. Subjects using glucocorticoids, contraceptives, or patients who were taking antihypertensive drugs that affect the renin angiotensin aldosterone system (RAAS), such as β-blockers, angiotensin-converting enzyme (ACE) inhibitors, angiotensin II receptor blockers, diuretics or spironolactone, underwent a washout for at least 15 days by withdrawal or replacement with doxazosin (5). PA diagnosis was performed following the Endocrine Society guidelines (5), using an elevated ARR of >25 (ng/dL)/(ng/mL*h) (45) and suppressed plasma renin activity (PRA; <1 ng/mL*hour) as the diagnostic criteria. Identified PA subjects (n = 7) were compared to normotensive healthy controls (n = 8), who were similar in age, BMI, and gender. Informed consent was obtained from all participants according to the guidelines of the Declaration of Helsinki and the Ethics Committee of the Faculty of Medicine, Pontificia Universidad Católica de Chile (CEC-MEDUC#180309002).

Blood Pressure Measurements and Laboratory Tests

Blood and second morning urine samples were collected between 8:00 and 10:00 am after overnight fasting, and a 30-minute resting period. After collection, urine samples (15 mL) were mixed with 1× cOmplete Protease Inhibitor Cocktail (Roche; Basel, Switzerland) before freezing at −80 °C. Serum aldosterone and plasma renin activity (PRA) were measured by radioimmunoassay using a commercial kit (Coat-A-Count Kit, Siemens, CA, USA; DiaSorin, MN, USA, respectively). Urinary creatinine was determined using the Jaffe method on an automated equipment (Roche Modular Analysis; Mannheim, Germany). In order to evaluate inflammation, hs-CRP was measured using the turbidimetric method (Hitachi, Tokyo, Japan) with both intraassay and interassay coefficient of variation of 3.5% within the normal range of less than 3 mg/L.

Urinary Extracellular Vesicles Isolation

The uEVs were isolated using differential ultracentrifugation as previously described (46). In brief, urine samples (12 mL) were centrifuged at 17 000g, at 4 °C for 15 minutes using a Sorvall WX80+ Floor Ultra Centrifuge (Thermo Scientific, NC, USA) with a TH-660 swinging bucket rotor. The supernatant (SN1) was stored at room temperature, and the pellet resuspended in 50 mL of 3.2M dithiothreitol (DTT), and 200 mL of isolation solution (10mM triethanolamine and 250mM sucrose) for 2 minutes at room temperature. Following, the mix was vortexed and centrifuged at 17 000g, 4 °C for 15 minutes. The resultant supernatant was collected and mixed with SN1 and centrifuged at 200 000g for 90 minutes at 20 °C. The final pellet was resuspended in 100 μL of phosphate-buffered saline (PBS) and stored at −80 °C. Isolated uEVs were characterized following the International Society for Extracellular Vesicles guidelines (47), using transmission electron microscopy, immunoblotting, and nanoparticle tracking analysis. Urinary creatinine was used to normalize samples for immunoblotting as previously described (43).

Transmission Electron Microscopy

Transmission electron microscopy (TEM) was performed to verify uEV shape and size. Briefly, uEV suspensions were adsorbed onto carbon-coated copper grids for 10 minutes and stained with 2% (w/v) uranyl acetate solution for 1 minute. Grids were visualized in a Philips Tecnai transmission electron microscope at 80 kV and images were acquired using a SIS-CCD Camera Megaview G2 (48). Micrographs were analyzed using ImageJ (NIH, USA (49)).

Nanoparticle Tracking Analysis

Nanoparticle tracking analysis (NTA) was performed using a NanoSight NS300 and NanoSight NTA 3.2 software (Malvern Instruments Ltd, Malvern, UK). Samples were analyzed using a low volume flow cell, an automatic syringe pump with flow speed = 50, and camera level 10, with detection threshold of 5. Particles were tracked with a laser beam and scattered light was captured using a sCMOS camera (3 videos of 30 seconds each). The Brownian motion of particles was determined and used to calculate the diameter (mean and mode size) and concentration of vesicles (50).

Immunoblotting of EV Markers and Alpha-1-Acid-Glycoprotein (AGP1)

The uEVs resuspended in Laemmli buffer were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membranes (Bio-Rad, CA, USA), blocked with 5% (w/v) skim milk in PBS-Tween20 (PBS-T) 0.3% (v/v) for 1 hour and incubated with primary rabbit monoclonal anti-TSG101 (1:5000 Ab125011, Abcam, MS, USA, RRID:AB_2889037 (51)), mouse monoclonal anti-CD63 (1:200 sc5275, Santa Cruz, TX, USA, RRID:AB_627877 (52)), rabbit monoclonal anti-CD9 (1:500 CD9[D801A] #13174; Cell Signaling Technology, MA, USA, RRID:AB_2798139 (53)), rabbit polyclonal anti-alpha 1 acid glycoprotein (AGP1) (1:500 Ab200732, Abcam, MS, USA, RRID:AB_2889037 (54)) overnight at 4 °C. After washing, membranes were incubated with horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG (1:2000 #7074S, Cell Signaling Technology, MA, USA, RRID:AB_2099233 (55)) or rabbit anti-mouse IgG-HRP antibodies (1:10000; ab6728; Abcam, USA, RRID:955440 (56)) for 1 hour at room temperature. Proteins were detected using chemiluminescence (West Femto reagent, ThermoFisher, USA) in a Chemi-Doc MP imaging system (Bio-Rad, CA, USA). Normalized uEVs from PA subjects and controls were used to confirm and validate AGP1 abundance by immunoblot and further quantified using ImageJ (NIH (49)).

Mass Spectrometry of Urinary Extracellular Vesicles

The protein concentration of uEVs was determined using the bicinchoninic acid method (57), and the uEV proteome was analyzed by liquid chromatography–mass spectrometry (LC-MS). The uEV extracts from PA patients (n = 7) and controls (n = 8) were precipitated with acetone to concentrate the samples. In brief, the samples were mixed with 200 µL 8M urea (in 0.1M HEPES pH 8.5) in the presence of 0.01M DTT and loaded onto Centrifugal Filters (Microcon #MRCF0R030; Merck; Darmstadt, Germany) and centrifuged for 15 minutes at 14 000g. The filters were washed once with 8M urea (in 0.1M HEPES pH 8.5) followed by an incubation with 100 µL of fresh 0.05M iodoacetamide and 8M urea (in 0.1M Tris pH 8.5) for 20 minutes in the dark. The filters were centrifuged and washed 3 times with 8M urea (in 0.1M HEPES pH 8.5), followed by 3 washes with 0.05M ammonium bicarbonate. Overnight digestion was performed in a wet chamber at 37 °C with a 1/100 trypsin-to-protein ratio. After overnight digestion, samples were acidified with trifluoroacetic acid and desalted using stage tips (58).

The digested peptides were injected into an Easy-nLC1000 (Thermo Fisher Scientific Inc, IL, USA) connected online to an LTQ-Orbitrap-Fusion mass spectrometer (Thermo Fisher Scientific Inc, IL, USA) with a developing gradient from 7% to 30% buffer B (80% acetonitrile, 0.1% formic acid in water) for 214 minutes before washes at 60% then 95% buffer B, for a total data collection time of 240 minutes. Thermo raw files were searched against the curated UniProt human proteome database with MaxQuant versions 1.6.0.1 (59) and its integrated search engine Andromeda with default settings. In addition, match between runs, label-free quantification (LFQ) and iBAQ were enabled. Prior to downstream analysis, proteins flagged as reverse or contaminant hits were filtered out. Proteins were analyzed using the LFQ values obtained from MaxQuant with Perseus software (version 1.6.5.0) (60, 61). To identify differentially expressed proteins, a volcano plot was built using the Log₂ LFQ fold change between patients (PA) and controls (X axis) and −Log₁₀ of the P value (Y axis) and analyzed using a 2-side unpaired t test.

Data and Statistical Analysis

Clinical data are presented as the median and interquartile range (percentile [25^th-75^th]). uEV data and LFQ values are presented as mean ± standard error of the mean. Data normality was determined using the Kolmogorov-Smirnov test. For parametric and nonparametric comparisons between 2 sets of data, an unpaired Student t test or a Mann-Whitney test was performed, respectively. To evaluate the diagnostic potential of proteins detected in uEVs, a receiver operating characteristic (ROC) curve was generated using the Log₂ LFQ values. Sensitivity and specificity corresponding to the largest area under the ROC curve (AUC) were used to assess the diagnostic accuracy of uEV proteins in patients and controls. ROC was expressed as AUC and 95% Confidence Interval (CI 95%). A P value of <0.05 was considered statistically significant. Data were analyzed using GraphPad Prism v8.2 software (GraphPad, CA, USA). * P < 0.05; ** P < 0.01.

Results

Clinical and Biochemical Characteristics of Subjects With PA

We identified 7 PA subjects (5.3%) in our study cohort. Clinical and biochemical baseline characteristics are detailed in Table 1. The ARR (28.8 [25.7-49.8] vs 5.1 [3.2-7.7]; P < 0.01), and serum aldosterone levels (15.4 [8.6-20.4] ng/dL vs 8.0 [5.4-12.3] ng/dL; P < 0.05) were significantly higher, while the PRA was significantly lower (0.40 [0.20-0.60] ng/mL*h vs 1.40 [1.20-3.07] ng/mL*h; P < 0.01) in the PA group compared with the healthy controls. In addition, systolic (158.7 [139.0-189.7] mmHg vs 114.5 [105.8-115.5] mmHg; P < 0.01) and diastolic (98.0 [89.7-125.7] mmHg vs 71.4 [69.2-74.1] mmHg, P < 0.01) blood pressure was higher in the PA group compared with healthy controls. No differences were observed between PA patients and healthy controls regarding the fractional excretion of potassium (FEK) nor the fractional excretion of sodium (FENa) (Table 1). In addition, hs-CRP levels were measured in PA and controls and no differences were observed between groups (0.85 [0.51-1.70] vs 0.56 [0.38-0.99] mg/L; P 0.369).

Table 1.

Clinical and Biochemical Characteristics of the Studied Subjects

Variable	Control	PA	P value	Test
Gender (male %)	50%	57%	NS	Chi-square
Age	40.9 [28.9–52.4]	52.4 [39.2–58.4]	0.0553	Student’s t test
BMI, kg/m2	25.2 [23.5–26.0]	28.4 [23.7–29.2]	0.072	Mann-Whitney
SBP, mmHg	114.5 [105.8–115.5]	158.7 [139.0–189.7]	<0.0001**	Student’s t test
DBP, mmHg	71.4 [69.2–74.1]	98.0 [89.7–125.7]	0.0002**	Student’s t test
Aldosterone, ng/dL	8.1 [5.37–12.3]	15.4 [8.6–20.4]	0.043*	Mann-Whitney
PRA, ng/mL*h	1.40 [1.20–3.07]	0.40 [0.20–0.60]	0.001**	Mann-Whitney
ARR, (ng/dl)/(ng/mL*h)	5.1 [3.2–7.7]	28.8 [25.7–49.8]	0.0003**	Mann-Whitney
Serum sodium, mEq/L	140.5 [139.0–141.8]	140.0 [138.0–143.0]	0.97	Student’s t test
Serum potassium, mEq/L	4.1 [3.9–4.3]	3.9 [3.7–4.4]	0.34	Student’s t test
Urinary sodium, mEq/24h	86.0 [42.5–104.0]	108.0 [63.0–198.0]	0.41	Student’s t test
Urinary potassium, mEq/L	23.8 [16.8–35.2]	40.50 [30.8–54.2]	0.08	Student’s t test
FEK, %	6.2 [5.9–9.0]	12.2 [6.6–13.5]	0.07	Mann-Whitney
FENa, %	0.7 [0.6–0.9]	0.8 [0.5–1.2]	0.48	Mann-Whitney
Hs-CRP, mg/L	0.56 [0.38–0.99]	0.85 [0.51–1.70]	0.37	Mann-Whitney

Data are presented as median [interquartile range].

Abbreviations: ARR, aldosterone/renin ratio; DBP, diastolic blood pressure; F, females; FEK, fractional excretion of potassium; FENa, fractional excretion of sodium; Hs-CRP, high sensitive c-reactive protein; M, males; PRA, plasma renin activity; SBP, systolic blood pressure.

*Different from control, P < 0.05; ** P < 0.01, n = 15, unpaired Student t test or Mann-Whitney test.

Characterization and Quantification of Urinary EVs

Isolated uEVs showed a donut or cup-shape morphology by TEM (Fig. 1A), and positive expression for the canonical EV markers CD9, CD63, and TSG101 (Fig. 1B). Particle size distribution determined by NTA ranged from 50 to 400 nm, with a mean size of 176 ± 4 nm, a modal size of 155 ± 8 nm and a concentration of 1.2 × 10⁹ ± 1.6 × 10⁸ particles/mL (Fig. 1C). The uEVs showed a direct positive correlation with urinary creatinine (Rho = 0.6357; P = 0.0128) (Fig. 2A). When PA patients and healthy controls were separated, uEVs from PA patients had a significantly smaller mean (167 ± 6 vs 183 ± 4nm; P = 0.03) and modal (137 ± 7 vs 171 ± 11nm; P 0.02) size (Fig. 2B). In addition, the creatinine-normalized uEV particle concentration was not different between PA patients and healthy controls (1 × 10⁹ ± 1.4 × 10⁸ vs 8.6 × 10⁸ ± 1.2 × 10⁸ particles/urinary creatinine) (Fig. 2C).

Figure 1.

Characterization of urinary extracellular vesicles (uEVs). A, Representative transmission electron micrograph of urinary extracellular vesicles. B, Representative Western blot for extracellular vesicle markers CD9, CD63, and TSG101 in uEVs. C, Size distribution plot from u-EVs using a NanoSight NS300 instrument. In A, scale bar = 200nm; black arrows are showing the uEVs.

Figure 2.

PA condition modifies the size but not the release of urinary extracellular vesicles (uEVs) determined by nanoparticle tracking analysis. A, Correlation of uEV particles/mL with urinary creatinine in healthy controls and PA patients uEVs (n = 15, r = 0.6357, P = 0.01). B, Mean and Mode size of uEVs from healthy controls and PA patients. C, uEVs concentration in particles/mL/mg of creatinine from healthy controls and PA patients. *Corresponds to a significant P value (P < 0.05).

Proteomic Profile of uEVs

The proteomic profile of uEVs identified approximately 1500 proteins, with about 750 detected in all 15 samples (see Supplementary material (62)). A volcano plot comparing protein expression in uEVs from healthy controls and PA patients revealed that alpha-1-acid-glycoprotein (AGP1), also known as oromusocoid protein 1 (ORM1), was significantly upregulated (2.43-fold change; P < 0.001)(Fig. 3).

Figure 3.

Protein analysis from urinary extracellular vesicles (uEVs). Volcano plot showing differential expression of uEVs proteins between healthy controls and PA patients. Each dot represents a protein, highlighting AGP1 as unique significant upregulated protein (2.43-fold change; P < 0.001). Vertical dotted lines represent a fold change of 2 (right) or −2 (left). Horizontal dotted line represents a P value of 0.01.

AGP1 Detection in uEVs by Immunoblot

The presence of AGP1 in uEVs was confirmed using immunoblot with uEV samples obtained from subjects with PA and healthy controls. We observed an increase in the abundance of AGP1 in uEVs from PA subjects compared with control subjects (1.89 ± 0.22 vs 1.00 ± 0.02 relative units, P = 0.028), showing a band close to 55 kDa. The exosomal marker TSG101 was identified at 45 kDa (Fig. 4).

Figure 4.

AGP1 detection in urinary extracellular vesicles (uEVs) by immunoblot. Alpha-1-acid glycoprotein 1 abundance was validated using immunoblot in uEVs from subjects with primary aldosteronism (PA) and healthy control subjects. A representative immunoblot (left) and the corresponding quantification (right) is presented showing AGP1 in uEVs obtained from 4 PA and 4 control subjects. RU: relative units. *Corresponds to a significant P value (P < 0.05).

Potential Diagnostic Value of AGP1

We performed an ROC curve analysis in order to evaluate the potential of AGP1 protein levels in uEVs to discriminate the presence of the PA condition. A simple logistic regression analysis revealed an AUC of 0.928 (0.780-1.000) (P = 0.0055), and Log₂LFQ AGP1 levels higher than 31.9 showed a sensitivity of 85.7% and a specificity of 87.5% to identify PA (Fig. 5).

Figure 5.

Receiver operating characteristic (ROC) curve for AGP1. ROC curve for uEV levels of AGP1 in subjects with PA and healthy controls.

Discussion

In the present study, we observed a significant positive correlation between concentration of uEVs and urinary creatinine, also smaller uEVs and an upregulation of AGP1 protein in uEVs from PA patients compared with healthy controls.

The correlation observed between uEVs and urinary creatinine further validates the use of this parameter for uEVs normalization as proposed by others (43). To the extent of our knowledge, this is the first description of smaller urinary EVs in PA patients compared with healthy controls. Furthermore, there is no previously reported evidence describing size differences in uEVs from patients with essential hypertension compared with healthy controls. The latter suggests that size differences could be a reflection of extracellular volume changes or an enrichment of uEVs in specific proteins with ion and water channel activities (eg, aquaporins) (63). Previous reports have shown that stimuli such as the presence of high concentrations of oxalate in renal epithelial cells can affect EV morphology, release, and cargo (64). Therefore, it might be plausible that aldosterone in the renal epithelia selectively favors the release of a specific subset of uEVs and so the study of uEVs from essential hypertension patients is highly encouraged, to determine whether uncontrolled hypertension or aldosterone itself is driving these changes.

In order to elucidate potential mechanisms of the different uEV sizes observed in PA patients, we propose to initially determine the effect of sodium and water status on uEV size, considering that changes in urine concentration or osmolality could modify EV size. Additionally, the proteomic analysis of EVs from in vitro studies using renal epithelial cells exposed to aldosterone or uEVs from DOCA-SALT mice models using size-based approaches (eg, size-exclusion chromatography) could provide additional insight on the mechanisms and determinants of uEVs size. The latter size-based approaches detect potential differences masked by the batch effect of analyzing the whole spectrum of vesicles.

In addition, it remains to be elucidated whether the effect of aldosterone on either the quantity or the cargo of EVs is due to the direct stimulation of aldosterone on EV biogenesis, release, or a paracrine action associated with the progression of the pathological phenotype (eg, hypertension, inflammation) observed in PA (29). The latter is especially relevant considering that angiotensin II, a known stimulator of aldosterone release, is able to increase EV release in renal proximal tubule cells (65). In our study, we did not observe significant differences in creatinine-corrected uEV concentration between PA patients and healthy controls, nor did we see a correlation between uEVs and circulating aldosterone, PRA, or the ARR. Burrello et al identified a higher level of circulating serum EVs in PA patients, when compared with both hypertensives and healthy controls, that was attributed to an enhanced biological response of the endothelium to aldosterone levels, as well as the effect of EV on endothelial function (44). Supporting this observation, aldosterone has been shown to stimulate EV release in endothelial cells in vitro (66), and to activate endothelial exocytosis (67). These differences suggest that the mechanisms triggering EV release by high aldosterone in the circulation are different from the mechanisms regulating EV release by the renal epithelia.

The current LC-MS spectrum of the uEV proteome was in agreement with other reports of mass spectrometry findings from studies of uEVs (38, 68). Previous studies showed that the cargo of EVs from the renal tubules is modified under different pathological conditions (eg, inflammation, hypoxia) (69). Therefore, considering the chronic proinflammatory state associated with PA, and the increased renal inflammation reported in these patients (7), it seems plausible that PA may also influence the selective cargo of uEVs. Although we did not observe increased levels of hs-CRP in PA patients, this suggests that hs-CRP is a valuable marker for acute rather than chronic inflammation, with PA patients primarily affected by chronic inflammation. We have previously reported such observations (7), where no differences in hs-CRP levels were found between PA patients and essential hypertensive patients, which has also been confirmed by other groups (70). Hence the identification and implementation in clinical practice of novel sensitive biomarkers of the chronic inflammatory phenotype reported in PA is encouraged.

In the current study, we showed that AGP1, also called orosomucoid-1 (ORM1) was significantly upregulated in PA patients. Conversely, we did not find significant changes in the abundance of known MR-activated sodium channels or proteins (eg, sodium chloride co-transporter [NCC], epithelial sodium channel, sodium hydrogen exchanger 3). Previous reports, in this respect, only show a discrete presence of AGP1 in uEVs using N-linked-glycoproteomics (71), and the renal sodium channel NCC and phospho-NCC by immunoblot in PA models (40, 42). To the extent of our knowledge, this is the first report showing AGP1 in the cargo of uEVs associated with PA.

A denaturing immunoblot was performed to validate the presence of AGP1 in uEVs.

The immunoblot of uEVs revealed a band of around 55 kDa corresponding to AGP1 (Fig. 4). Although the predicted molecular weight of AGP1 is close to 23 kDa, the observed difference is explained by the multiple N-glycosylation sites reported for AGP1 in 5 asparagine residues (72, 73) (Supplementary material (62)), which also have shown to change in response to different pathological conditions (74, 75).

AGP1 is an acute-phase protein associated with inflammation and with the lipocalin family, known to be rapidly increased by proinflammatory cytokines, glucocorticoids (72), and inflammatory conditions such as sepsis (76), and it is considered a cardiovascular risk factor (77). Previous reports have shown that the promoter region of ORM1 contains binding sites for C/EBP-α (ORegAnno record OREG1303317) transcription factors that overlap with binding sites for glucocorticoid response elements, which is especially relevant for mineralocorticoid-mediated regulation, considering that not only the glucocorticoid receptor, but also the MR binds glucocorticoid response elements in promoter regions of crucial genes for sodium transport such as SGK1 (78) and the ATP1A1 (Na/K ATPase α1 subunit) genes (79). This suggests that AGP1 upregulation in PA patients might be a consequence of mineralocorticoid activation mediated by the aldosterone-MR adduct. Additionally, in vitro aldosterone stimulation of cardiomyocytes induced ORM1 gene expression (80), which was inhibited by spironolactone (an MR antagonist) (81), demonstrating a MR-dependent-regulation of ORM1 expression. Also, in vivo evidence from cardiomyocyte-targeted MR overexpression and treatment with aldosterone identified ORM1 as an aldosterone-regulated gene in cardiomyocytes (82) and potentially in other tissues, resembling previous reports of neutrophil gelatinase associated lipocalin (NGAL or LCN2), a lipocalin and acute-phase protein that is increased under aldosterone stimulation and that has been proposed as a potential PA biomarker (21).

All this evidence points toward the fact that AGP1 in uEVs might be an indicator of renal mineralocorticoid activity. Considering the proinflammatory state associated with PA, the upregulation of an acute-phase response protein that elevates during inflammation such as AGP1, highlights its potential as an indicator of mineralocorticoid activity both at the systemic and renal level. The latter was supported by the discriminatory analysis performed by ROC curve (Fig. 5), which showed that AGP1 could be a novel biomarker candidate for PA diagnosis.

In summary, we identified changes in the size and cargo of uEVs from PA patients. Proteomic analysis of uEVs identifies the protein AGP1 as a potential biomarker for PA, which was validated using immunoblot, suggesting that the effects of elevated aldosterone levels on uEVs might be related with increased renal inflammation. Further studies using a validation cohort are encouraged to confirm the validity of AGP1 as a biomarker for PA.

Abbreviations

Abbreviations
 
• AGP1

  alpha-1-acid glycoprotein 1

 
• ARR

  aldosterone-to-renin ratio

 
• AUC

  area under the curve

 
• BMI

  body mass index

 
• EV

  extracellular vesicle

 
• hs-CRP

  high-sensitivity C-reactive protein

 
• LC-MS

  liquid chromatography–mass spectrometry

 
• LFQ

  label-free quantification

 
• MR

  mineralocorticoid receptor

 
• NCC

  sodium chloride co-transporter

 
• NTA

  nanoparticle tracking analysis

 
• PA

  primary aldosteronism

 
• PBS

  phosphate-buffered saline

 
• PRA

  plasma renin activity

 
• ROC

  receiver operator characteristic

 
• TEM

  transmission electron microscopy

 
• uEV

  urinary extracellular vesicle

Acknowledgments

Proteomics and bioinformatic support were provided by the Department of Molecular Biology at the Radboud Institute of Molecular Life Sciences, and the Cambridge Baker Systems Genomics Initiative at Baker Heart & Diabetes Institute. We appreciate the support of Jorge Perez-Lopez in technical procedures.

Financial Support: This study was supported partially by Grants Agencia Nacional de Investigación y Desarrollo - Fondo Nacional de Desarrollo Científico y Tecnológico (ANID-FONDECYT) 1160695, 1190250, 1190419, 1212006, and 3200646; Comision Nacional de Ciencia e Investigacion - Fondo de Equipamiento Científico y Tecnológico (CONICYT-FONDEQUIP) EQM150023; Agencia Nacional de Investigación y Desarrollo–Millennium Science Initiative Program- IMII P09/016-F, ICN09_016; CORFO-Biomedical Research Consortium 13CTI-21526-P1; Sociedad Chilena de Endocrinologia y Diabetes 2019-09; and Centro Traslacional de Endocrinologia-UC. E.B. holds a fellowship from Comision Nacional de Ciencia e Investigacion de Chile (21171092), CONICYT Pasantía Doctoral, DIDEMUC support grant, and Faculty of Medicine fellowship from Pontificia Universidad Católica de Chile. J.P.R. is a fellow of the Radboud Excellence Initiative (Radboud University, Nijmegen, the Netherlands).

Author Contributions: E.B. and C.C. worked on experimental design. C.C., A.T-C., A.V., and C.F. enrolled patients. E.B., J.P.R., and C.C. performed the experiments. E.B., J.P.R., and C.C.M. analyzed the data. E.B., J.P.R., C.C., R.B., J.H., and M.J.Y. wrote the manuscript.

Additional Information

Disclosures: The authors have nothing to disclose.

Data Availability

Some or all data generated or analyzed during this study are included in this published article or in the data repositories listed in References.