The study was approved by the Institutional Ethics Committee of Sanjay Gandhi Postgraduate Institute of Medical Sciences (IEC Code 2018-139-EMP-106) and Pondicherry Institute of Medical Sciences (PIMS IEC No. RC 18/105). Participant recruitment and blood analysis were described previously (28). Briefly, individuals aged 18–60 years without or with diabetes and/or hypertension and reporting to have no CKD were enrolled after written informed consent. Fasting blood and second morning urine were collected.

Urine was pre-processed, and 10 ml was used to isolate exosomes. The exosomes were isolated using the ultracentrifugation method as described by us previously (7, 22, 29, 30) and stored at −80°C until further analysis. Exosome characterization was done using nanoparticle tracking analysis (NTA), flow cytometry, and immunoblotting. Urinary exosomal (uE) samples were used to estimate uE-WT1 by sandwich ELISA (EH1321, FineTest) as per the manufacturer's protocol. An aliquot of whole urine was used for the estimation of the following markers using solid-phase sandwich ELISA (Human Lipocalin-2/NGAL DuoSet kits, R&D Systems), albumin (DY1455), KIM-1 (DY1750B), and NGAL (DY1757). Other parameters, including urine creatinine, serum creatinine, triglycerides, cholesterol, and blood urea nitrogen (BUN), were estimated using an autoanalyzer (XL-640, Erba Mannheim, Germany). For the analysis, urine KIM-1, NGAL, and uE-WT1 were normalized with urine creatinine.

The Institutional Animal Ethics Committee (IAEC) of Sanjay Gandhi Postgraduate Institute of Medical Sciences (Registration No. 57/PO/ReBi/SL/99/CPCEA) approved the study. Fourteen male Wistar rats (R. norvegicus) weighing between 230 and 280 g and aged 90–100 days were obtained from the Indian Institute of Toxicology Research, Lucknow, India. The animals were acclimatized and housed as per our routine procedure (30). The standard chow diet and water were freely accessible to all the animals under study. Intraperitoneal administration of streptozotocin (STZ) (50 mg/kg body weight of rat) dissolved in 0.1 M citric acid buffer with pH 4.5 was used to induce diabetes after 18 h of fasting. Since there can be early mortality due to excessive insulin secretion by damaged beta cells, rats were provided with sucrose (15 g/L) supplemented with drinking water for 48 h. Diabetes in rats was confirmed at a blood glucose level of 11 mmol/L (1 mmol = 18 mg glucose). To monitor blood glucose using a glucometer (Optimum Exceed, Abbott Diabetes Care Inc., Alameda, CA, United States), blood from the tail vein was collected after 48 h of STZ injection. Urine samples were collected at baseline (before STZ injection) and at the 6th and 9th weeks after STZ-induced diabetes using metabolic cages (Lab Products, United States). Rats were euthanized, and the kidneys were harvested after the 9th week to confirm kidney injury by histopathology using periodic acid-Schiff (PAS) staining. Kidney tissue sections from vehicle-treated rats were used as controls.

Nanoparticle tracking analysis (NTA using NanoSightNS300, Malvern Instruments, United Kingdom) was performed to analyze the size distribution and concentration of urinary exosomes. Briefly, exosomes were diluted in 1 × PBS and loaded into a sample chamber. A green laser was used to detect the nanoparticles, and five videos of 60 s duration were taken. Data were analyzed using the NTA 3.2 software (Malvern Instruments), which tracks the individual particles and calculates their size and velocity to generate a size distribution profile and determine the nanoparticle concentration.

Kidney tissues were fixed in 4% paraformaldehyde. Thereafter, tissues were embedded in paraffin blocks and 3 μm slices were cut from embedded blocks for staining. Periodic acid-Schiff (PAS) staining was done following the manufacturer's protocol (Sigma, St. Louis, MO, United States) as described previously (29). Stained sections were visualized at 400 × by using an Olympus IX73 light microscope.

For Western blotting of urinary exosomes, paired urine samples from rats collected at baseline (before STZ-injection), at the 6th and 9th weeks after STZ-injection, were used (n = 3 rats/time point). Western blotting of exosomal protein was performed as described by us previously (22). Briefly, exosomal protein samples were solubilized in Laemmli sample buffer. For each sample, an equal volume of solubilized protein was loaded onto a 10% polyacrylamide gel. Separated proteins were transferred onto nitrocellulose membranes and blocked with 5% non-fat dry milk for 1 h. Membranes were then incubated with a primary antibody against the WT1 protein (Abcam, MA, United States) overnight at 4°C. Thereafter, the incubation membranes were washed and then incubated with horseradish peroxidase-conjugated anti-rabbit secondary antibody (1:5,000). The antibody-antigen reactions were visualized using chemiluminescence (GE Healthcare, NJ, United States). Densitometry analysis of the uE-WT1 protein band was done. Total uE protein loaded on the gel was used for normalization.

For exosome characterization, a similar protocol was performed using antibodies against CD81 (Abcam, cat no. ab23505).

Exosomes were analyzed for the presence of surface marker CD63 by flow cytometry using magnetic coated CD63-Dynabeads (Thermo Fisher Scientific, Waltham, MA, United States, Cat No. 10606D) as per the manufacturer's instructions. Isolated exosomes were resuspended in 1 × PBS and bound to magnetic-coated CD63-Dynabeads (Thermo Fisher Scientific, Waltham, MA, United States) overnight at 4°C. The following day, the Dynabeads-bound exosomes were incubated with an anti-CD63 mouse antibody (Abcam, Cat No. ab59479) overnight at 4°C along with an appropriate isotype control. After incubation, the beads were washed two times with 1 × PBS and stained with Alexa Flour 488 anti-mouse secondary antibody (detection antibody) for 1 h at 4°C and analyzed by flow cytometry (Beckman Coulter DxFLEX Flow Cytometer).

Categorical variables were represented as frequency (percentage), while continuous variables were represented as median (interquartile range) or mean ± standard deviation. uE-WT1 was divided into four groups using the first, second, and third quartiles. The Kruskal-Wallis H test was used to test the significance of urinary and serum markers by ordering ACR categories and uE-WT1 quarters, followed by the Mann-Whitney U test with Bonferroni correction as the post-hoc test. Furthermore, Spearman correlation (rho) and stepwise multivariable linear regression were performed to test the linear association between variables and to determine which of the several urinary markers (uE-WT1, KIM-1, and NGAL) best determined the differences in ACR levels. The Spearman rank correlation was performed to test the correlation between variables. To compare the levels of ACR and uE-WT1 in rats before and after STZ injection (at the 9th week), a Wilcoxon signed rank test was used.

The model with age, gender, and BMI as the only explanatory variables was considered the base model. Gender was converted to a dummy variable as required for linear regression. A stepwise variable selection approach using multivariable linear regression was used to determine the performance of urinary markers in explaining variability in ACR. The influential points were identified through regression diagnostics, and sensitivity analysis was performed. The coefficient of determination (R²) and Akaike information criterion (AIC) were used to evaluate model performance.

A p < 0.05 was considered to be statistically significant. All analysis was performed using Stata 16 (StataCorp) and GraphPad Prism 8. The power of the study was calculated to be more than 80%.

NTA analysis showed that the maximum number of extracellular vesicles in our preparation had a particle size of 162 nm, with the second-highest concentration of 226 nm size (Figure 1A). Characterization by immunoblotting revealed a specific band for CD81 (exosome-specific marker protein) in our samples (Figure 1B). The percentage of exosome-bead complexes with positive staining for CD 63, another exosome-specific marker protein, was quantitated in our urinary exosome preparation using a flow cytometer. Figure 1C shows the results obtained in samples of the determined exosome-bead complex with 93.55% CD63 positive staining and 90.45% CD63 positive staining (Figure 1C).

We enrolled 627 participants reporting to have no CKD. However, 43 (6.86%) of them were found to have compromised kidney function (ACR≥30 mg/g or eGFR <60 ml/min/1.73 m²) after laboratory analysis. Individuals with kidney disease were removed from our study cohort; however, their characteristics were summarized in Supplementary Table 1. A total of 584 participants were involved in the study, including 312 (53.42%) men with an average [median (IQR)] age of 44.00 (36.00–51.00) years. In total, 335 (57.36%) of individuals had diabetes and/or hypertension (Table 1).

To further check the association of uE-WT1 with kidney injury markers, the individuals with normal kidney function were stratified by uE-WT1 quartiles. The individuals in the highest uE-WT1 quarter showed elevated urinary KIM1 and NGAL, relative to lower quarters (Tables 2A, B). There were fewer men than women in the highest quarter relative to the gender ratio in the lower quarter (p < 0.01). However, urinary NGAL remained significantly elevated in individuals in the highest uE-WT1 quarter, even after age and gender stratification. The median age of the individuals between the quarters ranged between 42 and 46 years. The urine ACR was also statistically different among the four groups; it was highest (but well below the diagnostic cutoff) in the upper quarter (Table 2A). The four groups had similar eGFR levels and lipid profiles (Table 2A).

KIM-1, NGAL, and uE-WT1 showed a positive association with ACR. The level of renal injury markers in the ACR category (11–30 mg/g), a.k.a. high-normal albuminuria, had significantly higher levels of KIM-1, NGAL, and uE-WT1 when compared to those with ACR below 1 mg/g or 1–10 mg/g (Tables 3A, B).

Correlation analysis showed a moderate but significant correlation between ACR and uE-WT1, (rho = 0.47 vs. rho = 0.25). Stepwise multivariable linear regression analysis was performed to determine the ability of uE-WT1, NGAL, and KIM-1 as determinants of high-normal ACR levels. The known urinary markers (KIM-1 and NGAL) were found to perform well in explaining the variability in ACR within the normal range (p < 0.01) (Supplementary Table 2). Base model with (1) KIM-1 could explain 5.9% of the variability in ACR, (2) NGAL could explain 8.70% of the variability in ACR, and (3) uE-WT1 could explain 24.50% of the variability in ACR. The performance of the regression model slightly improved after combining KIM-1 with uE-WT1 (Radj2 = 0.273, AIC: 1,218.73) or NGAL with uE-WT1 (Radj2 = 0.258, AIC: 1,237.26). After removing influential points, the model performance improved for uE-WT1 and KIM-1 (Radj2 = 0.273 vs. 0.350, AIC = 1,218.73 vs. 1,067.23) as well as for NGAL and uE-WT1 (Radj2 = 0.258 vs. 0.280, AIC = 1,237.26 vs. 1,031.68) (Figure 2, Supplementary Table 3).

We next tested if induction of kidney injury may cause a rise in uE-WT1 in diabetic rats. Wistar rats were made diabetic and then followed for 9 weeks to monitor diabetes-induced kidney injury. Rats were made diabetic by streptozotocin injections (i.p.). Analysis of uE-WT1 levels in the paired uE samples from these rats showed a significant rise at the 9th week of STZ injection, relative to baseline (Figure 3A). Furthermore, a substantial rise in uACR was also evident in these rats in the 9th week after STZ injections, relative to baseline (Figure 3B). Moreover, uACR strongly correlated with uE-WT1 protein levels in diabetic rats in the 9th week after STZ injection (Figure 3C). Histopathological changes were also confirmed in the kidney tissues by PAS staining in the 9th week after STZ injection (Figure 3D), similar to what has been reported by us previously (29). PAS staining clearly indicates considerable deposition of polysaccharides (collagen, glycogen) in rats after the 9th week of STZ injection compared to the kidney tissue section of age-matched vehicle-treated rats.

Moreover, the change in uE-WT1 protein levels was also confirmed by immunoblotting of the paired uE samples collected at baseline and the 6th and 9th weeks from diabetic rats (n = 3/time point). WT-1 protein bands were detected at the 6th and 9th weeks after STZ injection, while no band was detected in the baseline uE samples (Figure 3E). The densitometric analysis confirmed a significant rise in band density at 9th-week uE samples relative to baseline (Figures 3E, F).