SUN-310 URINARY EXTRACELLULAR VESICLES REFLECT THE PATHOLOGY OF DIABETIC NEPHROPATHY

Abstract

Diabetic nephropathy (DN) involves the injury of podocytes, tubular epithelium, and endothelium. These cells release substantial extracellular vesicles (EVs) into urine, which are supposed to increase when their parent cells encounter stressed conditions. We therefore hypothesize that various populations of urinary EVs could reflect the pathological features of DN, thus might help differentiate patients with DN from those with non-DN proteinuria. Altogether 78 patients with nephrotic proteinuria underwent renal biopsy were enrolled, including 42 with DN, 7 with DN superimposed on membranous nephropathy (MN), and 29 cases of proteinuric controls with MN or minimal change disease (MCD), where 10 cases were diabetic and the other 19 were nondiabetic. Urinary EVs were detected by flowcytometry and microvesicles (MVs, 0.2-1μm EVs positive for Annexin-V or cell specific markers) and >1μm EVs were analyzed. Podocyte (nephrin or podocin positive), proximal tubular cell (AQP1 or megalin positive), and endothelial cell (CD31 or CD144 positive)-derived MVs and >1μm EVs were numbered respectively. The correlations between urinary EVs and pathological features were analyzed. Multivariant binary logistic regression model was used to figure out optimal combinations of EVs populations discriminating DN patients from proteinuric controls. Almost all EVs from podocytes, proximal tubular and endothelial cells were significantly increased in DN patients compared to proteinuric controls. There was no significant difference in urinary EVs between patients with DN and DN superimposed on MN, or between proteinuric controls with diabetics and those with not. DN patients from class Ⅲ had greater number of endothelial cell (CD31+, p=0.024) and proximal tubular cell (AQP1+/megalin+, p=0.022) -derived MVs compared to those from class Ⅱ. EVs from these cells also correlated with the degree of mesangial expansion and vascular injury, reflecting the severity of DN classic glomerular and vascular impairments. In tubulointerstitial lesions, proximal tubular cell-derived EVs correlated with the severity of acute kidney injury, whereas did not correlate with the severity of chronic tubulointerstitial injury or interstitial inflammation in DN patients. Correlation analysis among EVs revealed that the numbers of >1μm EVs from endothelial cells correlated closely with those of MVs from podocytes (r range 0.695-0.783) and proximal tubular cells (r range 0.664-0.805), the number of endothelial MVs (CD144+) correlated with those of proximal tubular MVs (r range 0.614-0.619). In binary logistic regression model, the combination of megalin positive >1μm EVs(with every 10-fold increase: OR=3.74, 95%CI 1.12-12.53; p=0.033), podocin and nephrin positive MVs (with every 10-fold increase: OR=4.03, 95%CI 1.63-9.95; p=0.003), and Annexin-V positive MVs (with every 10-fold increase: OR=0.30, 95%CI 0.10-0.85; p=0.024) best discriminated DN patients from proteinuric controls with the specificity of 89.29% and sensitivity of 67.35%. Urinary EVs might be an intriguing noninvasive approach to discriminate DN from non-DN proteinuric glomerulopathy, which could be particularly meaningful in differential diangosis of diabetic patients with overt proteinuria, whereas additional validation is needed. Further researches are also needed to unveil mechanisms underlying the changes of urinary EVs in DN.