Mineralocorticoid receptor antagonism in diabetes reduces albuminuria by preserving the glomerular endothelial glycocalyx

Abstract

The glomerular endothelial glycocalyx (GEnGlx) forms the first part of the glomerular filtration barrier. Previously, we showed that mineralocorticoid receptor (MR) activation caused GEnGlx damage and albuminuria. In this study, we investigated whether MR antagonism could limit albuminuria in diabetes and studied the site of action. Streptozotocin-induced diabetic Wistar rats developed albuminuria, increased glomerular albumin permeability (Ps'alb), and increased glomerular matrix metalloproteinase (MMP) activity with corresponding GEnGlx loss. MR antagonism prevented albuminuria progression, restored Ps'alb, preserved GEnGlx, and reduced MMP activity. Enzymatic degradation of the GEnGlx negated the benefits of MR antagonism, confirming their dependence on GEnGlx integrity. Exposing human glomerular endothelial cells (GEnC) to diabetic conditions in vitro increased MMPs and caused glycocalyx damage. Amelioration of these effects confirmed a direct effect of MR antagonism on GEnC. To confirm relevance to human disease, we used a potentially novel confocal imaging method to show loss of GEnGlx in renal biopsy specimens from patients with diabetic nephropathy (DN). In addition, patients with DN randomized to receive an MR antagonist had reduced urinary MMP2 activity and albuminuria compared with placebo and baseline levels. Taken together, our work suggests that MR antagonists reduce MMP activity and thereby preserve GEnGlx, resulting in reduced glomerular permeability and albuminuria in diabetes.

Trial registration: ClinicalTrials.gov NCT00381134.

Keywords: Chronic kidney disease; Diabetes; Endocrinology; Glycobiology; Nephrology.

Figure 1. Development of albuminuria and increased glomerular permeability in early diabetic nephropathy is ameliorated by MR antagonism.

(A) Schematic overview of STZ-induced diabetic model and spironolactone (spiro) treatment protocol for male Wistar rats. An injection of STZ was given at week 0. Four weeks after STZ injection, spiro (an MR inhibitor) was given for 28 days, and rats were culled at week 8 after STZ injection. Rats were randomized to receive STZ and spiro. (B) Treatment with spiro for 28 days reduced the fold change in urinary albumin/creatinine ratio (uACR) from initiation of treatment, week 4 to week 8 (control, n = 10; diabetes, n = 12; diabetes-spiro, n = 13). Data were log transformed and presented as log₂ (fold change). (C) Representative images of an isolated glomerulus stained with R18 and Alexa Fluor 488–BSA (AF488-BSA). Magnification, 20×. (D) Glomerular albumin permeability (Ps’_alb) was measured at week 8 (control, n = 7 rats [32 glomeruli]; diabetes, n = 8 [36 glomeruli]; diabetes-spiro, n = 5 [21 glomeruli]). In B and D, 1-way ANOVA was used for statistical analysis, followed by Tukey’s multiple comparisons. Each dot, triangle, and square on the graph represents a rat. Data are expressed as mean ± SEM. **P < 0.01; ***P < 0.001.

Figure 2. Fluorescence profile peak-to-peak measurements confirm that glomerular endothelial glycocalyx damage is prevented by MR antagonism and correlates strongly with glomerular albumin permeability.

Rats were perfused with Ringer solution, and the left kidney was removed for lectin staining. (A) Representative images show glomerular capillaries labeled red (R18) and the luminal glomerular endothelial glycocalyx (GEnGlx) labeled green with Marasmium oreades agglutinin (MOA) or wheat germ agglutinin (WGA). Scale bars: 20 μm and 5 μm. ROI, region of interest for fluorescence profile peak-to-peak (P-P)measurement. (B) Representative relative intensity peaks of R18 (red) and MOA (green) profiles showing P-P assessment of the GEnGlx; Gaussian curves (dashed lines) were fit to the raw intensity data of each plot for P-P measurements. (C and D) Quantification at week 8 after STZ of GEnGlx MOA labeling P-P (control, n = 7; diabetes, n = 6; diabetes-spironolactone [diabetes-spiro], n = 7) and functional association with the rate of glomerular albumin leakage (Ps’_alb) (n = 9). (E and F) Quantification at week 8 after STZ of GEnGlx WGA labeling P-P (control, n = 5; diabetes, n = 6; diabetes-spiro, n = 5) and functional association with the rate of glomerular Ps’_alb (n = 16). In C and E, 1-way ANOVA was used for statistical analysis, followed by Tukey’s multiple comparisons. Each dot, triangle, and square on the graph represents a rat. Data are expressed as mean ± SEM. **P < 0.01; ***P < 0.001.

Figure 3. The effect of MR antagonism in preventing the diabetes-induced increase in glomerular permeability is dependent on the glomerular endothelial glycocalyx.

(A) Schematic overview of enzymatic degradation of the glomerular endothelial glycocalyx (GEnGlx) with hyaluronidase on spironolactone-treated (spiro-treated) male Wistar rats. An injection of STZ was given at week 0. Four weeks after STZ injection, spiro (an MR inhibitor) was given for 28 days, and rats were given hyaluronidase (200 units) at week 8 after STZ via tail vein injection 1 hour before being culled for tissue collection. Rats were randomized to receive hyaluronidase. (B and C) Quantification at week 8 after STZ of GEnGlx WGA labeling peak-to-peak (diabetes-spiro, n = 5; diabetes-spiro-enzyme, n = 5) and GEnGlx MOA labeling peak-to-peak (diabetes-spiro, n = 5; diabetes-spiro-enzyme, n = 5) confirmed enzyme degradation of GEnGlx. In B and C, unpaired t test was used for statistical analysis. (D) Albuminuria levels returned to those expected in vehicle treated diabetic rats. Urinary albumin/creatinine ratio (uACR) was determined from the same rats (n = 5) at week 4 after STZ (diabetes before spiro 4wk), week 8 after treatment with spiro (diabetes-spiro 8wk preenzyme), and week 8 after hyaluronidase (diabetes-spiro 8wk post-enzyme). The connecting line (gray) represents the same rat for each data point. Repeated-measures 1-way ANOVA was used for statistical analysis, followed by Tukey’s multiple comparisons. Each triangle or square on the graph represents a rat. Data are expressed as mean ± SEM. *P < 0.05; ***P < 0.001.

Figure 4. Increased matrix metalloproteinase activity in early diabetic nephropathy is ameliorated by MR antagonism.

(A and D) Matrix metalloproteinase (MMP) activities were measured for the sheddases MMP2 and MMP9. Systemic circulation of plasma active MMP2 (A) and plasma active MMP9 (D) (control, n = 10; diabetes, n = 7; diabetes-spironolactone [diabetes-spiro], n = 13) were determined. (B and E) Localized activity of glomerular active MMP2 (B) and glomerular active MMP9 (E) (control, n = 8; diabetes, n = 7; diabetes-spiro, n = 8) were determined from isolated glomeruli and normalized to glomerular total protein. (C and F) Urine active MMP2 (C) and urine active MMP9 (F) (control, n = 13; diabetes, n =11; diabetes-spiro, n = 15) were determined and normalized to urine creatinine. Each dot, triangle, and square represents a rat. The fold change relative to control was calculated to enable pooling of results from different experiments. One-way ANOVA was used for statistical analysis, followed by Tukey’s multiple comparisons. Data are expressed as mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 5. Exposing human GEnC to diabetic conditions resulted in MMP upregulation and GEnGlx damage, effects ameliorated by MR antagonism.

Human conditionally immortalized glomerular endothelial cells (CiGEnC) were maintained in the presence of glucose, insulin, TNF-α, and IL-6 to mimic a diabetic environment. (A, C, E, and G) Representative images of CiGEnC stained with MMP2, MMP9, MMP14, or WGA lectin (an endothelial glycocalyx label) shown for control, diabetes, and diabetes-spironolactone (diabetes-spiro) samples. DAPI, nuclear label; R18, endothelial membrane label. Scale bars: 25 μm. (B, D, F, and H) Fluorescence intensity was quantified in CiGEnC for MMP2 (control, n = 35; diabetes, n = 34; diabetes-spiro, n = 33), MMP9 (control, n = 23; diabetes, n = 24; diabetes-spiro, n = 19), MMP14 (control, n = 18; diabetes, n = 31; diabetes-spiro, n = 14), and WGA lectin (control, n = 30; diabetes, n = 41; diabetes-spiro, n = 43). In A and B, 1-way ANOVA was used for statistical analysis, followed by Tukey’s multiple comparisons. In C and D, Kruskal-Wallis test was used for statistical analysis. Data are expressed as mean ± SEM. **P < 0.01; ***P < 0.001.

Figure 6. The glomerular endothelial glycocalyx is damaged in human diabetic nephropathy.

(A) Representative images show glomerular capillaries labeled red (R18) and the luminal glomerular endothelial glycocalyx (GEnGlx) labeled green with Ulex europaeus agglutinin I (UEA-I) in renal biopsies from thin basement membrane nephropathy (TBMN) and diabetic nephropathy (DN) patients from Bristol, United Kingdom, and from histologically normal controls and patients with DN from Bari, Italy. Scale bar: 40 μm and 5 μm. (B) Quantification of GEnGlx UEA labeling peak-to-peak (Bristol: TBMN, n = 8; diabetes, n = 12; Bari: control, n = 7; diabetes, n = 7) confirms that GEnGlx damage may contribute to the disease phenotype seen in human DN. Unpaired t test was used for statistical analysis. (C) The number of samples, glomeruli, and capillaries used to analyze each group. Each dot or square on the graph represents a patient. Data are expressed as mean ± SEM. ***P <0.001.

Figure 7. Matrix metalloproteinase activity in human diabetes is reduced by MR antagonism.

Matrix metalloproteinase (MMP) activities were measured for the sheddases MMP2 and MMP9. (A and B) Urine active MMP2 and urine active MMP9 were determined and normalized to urine creatinine (placebo, n = 20; spironolactone, n = 15). Data were log transformed and presented as log₂ (fold change). Mean fold change in placebo and spironolactone groups were presented for urine active MMP2 and urine active MMP9. Normalized changes for each individual taking placebo/spironolactone are also displayed to illustrate the variability in individuals’ progression with time/response. Paired t test was used for statistical analysis between baseline and week-48 data. Data are expressed as mean ± SEM. *P < 0.05; **P < 0.01. Individual data for urine active MMP fold change from baseline to week 48 were presented for placebo and spironolactone groups. Each dot represents an individual patient.