Once daily administration of the SGLT2 inhibitor, empagliflozin, attenuates markers of renal fibrosis without improving albuminuria in diabetic db/db mice

Abstract

Blood glucose control is the primary strategy to prevent complications in diabetes. At the onset of kidney disease, therapies that inhibit components of the renin angiotensin system (RAS) are also indicated, but these approaches are not wholly effective. Here, we show that once daily administration of the novel glucose lowering agent, empagliflozin, an SGLT2 inhibitor which targets the kidney to block glucose reabsorption, has the potential to improve kidney disease in type 2 diabetes. In male db/db mice, a 10-week treatment with empagliflozin attenuated the diabetes-induced upregulation of profibrotic gene markers, fibronectin and transforming-growth-factor-beta. Other molecular (collagen IV and connective tissue growth factor) and histological (tubulointerstitial total collagen and glomerular collagen IV accumulation) benefits were seen upon dual therapy with metformin. Albuminuria, urinary markers of tubule damage (kidney injury molecule-1, KIM-1 and neutrophil gelatinase-associated lipocalin, NGAL), kidney growth, and glomerulosclerosis, however, were not improved with empagliflozin or metformin, and plasma and intra-renal renin activity was enhanced with empagliflozin. In this model, blood glucose lowering with empagliflozin attenuated some molecular and histological markers of fibrosis but, as per treatment with metformin, did not provide complete renoprotection. Further research to refine the treatment regimen in type 2 diabetes and nephropathy is warranted.

Figure 1. Circulating glucose levels.

(a,b) Fasted blood glucose at baseline and three days after treatment start, (c,d) fasted plasma glucose at baseline and treatment end, and (e) glycated hemoglobin at treatment end in db/m (open) and db/db (grey) mice. Circles (ο) vehicle-treated; squares (◽) empagliflozin-treated; triangles (▵) metformin-treated; and diamonds (◊) empagliflozin + metformin co-treated. Data are means ± SEM (n = 5–11). (a,c,e): *P < 0.05 vs db/m vehicle, ^†P < 0.05 vs db/db vehicle, ^δP < 0.05 vs db/db metformin within a time point by one-way ANOVA and Tukey’s post hoc. Comparisons by Student’s unpaired t-test: significance denoted by solid lines and trends denoted by dashed lines. (b,d): ^#P < 0.05 delta change (▵) from baseline by Student paired t-test. N.B. (a,b): Some mice exceeded the upper limit of the glucometer, in which case the recorded blood glucose value was 33.3 mmol/L. Differences from baseline may therefore be underestimated.

Figure 2. Oral glucose tolerance.

(a) Plasma glucose concentrations over time, (b) area under glucose curve; AUCglucose, (c) plasma insulin concentrations over time, (d) area under insulin curve; AUCinsulin, and (e) insulinogenic index; AUCinsulin:glucose 0–30 mins in response to an oral glucose bolus (2 g/kg body weight) and (f) fasted plasma glucose-to-insulin ratio (t = 0 mins) in db/m (open) and db/db (grey) mice. Circles (ο) vehicle-treated; squares (◽) empagliflozin-treated; triangles (▵) metformin-treated; and diamonds (◊) empagliflozin + metformin co-treated. Data are means ± SEM (n = 6–11). *P < 0.05 vs db/m vehicle, ^†P < 0.05 vs db/db vehicle, ^δP < 0.05 vs db/db metformin, ^ΨP < 0.05 vs all other db/db groups by one-way ANOVA and Tukey’s post hoc. Comparisons by Student’s unpaired t-test: significance denoted by solid lines and trends denoted by dashed lines.

Figure 3. Filtered glucose load, glucosuria, and cortical glucose content.

Predicted filtered glucose load determined by (a) fasted plasma glucose and (b) GFR on dual y-axes figures, (c) urinary glucose excretion, and (d) cytosolic glucose levels within kidney cortices in db/m (open) and db/db (grey) mice. Circles (ο) vehicle-treated; squares (◽) empagliflozin-treated; triangles (▵) metformin-treated; and diamonds (◊) empagliflozin + metformin co-treated. Data are (a,b) means ± SEM or (c,d) individual mice with means ± SEM (n = 5–11). *P < 0.05 vs db/m vehicle, ^†P < 0.05 vs db/db vehicle, ^δP < 0.05 vs db/db metformin by one-way ANOVA and Tukey’s post hoc. Comparisons by Student’s unpaired t-test: significance denoted by solid lines and trends denoted by dashed lines.

Figure 4. mRNA and protein expression of glucose transporters in kidney cortices.

(a–c) Real-time qPCR for Slc5a1 (encoding SGLT1), Slc5a2 (encoding SGLT2), and Slc2a2 (encoding GLUT2), and (d,e) Western immunoblot for total cell membrane SGLT2 expression in db/m (open) and db/db (grey) mice. Data are means ± SEM (n = 4–6). *P < 0.05 vs db/m vehicle, ^δP < 0.05 vs db/db metformin by one-way ANOVA and Tukey’s post hoc. Comparisons by Student’s unpaired t-test: significance denoted by solid lines and trends denoted by dashed lines. Blot has been cropped to improve clarity; see Supplementary Fig. S3 for full length blots.

Figure 5. mRNA expression of gluconeogenic enzymes in kidney cortices.

(a–c) Real-time qPCR for Pck1 (encoding phosphoenolpyruvate carboxykinase 1; PEPCK), Fbp1 (encoding fructose bisphosphatase 1), and G6pc (encoding glucose-6-phosphatase) in db/m (open) and db/db (grey) mice. Data are means ± SEM (n = 4–6). *P < 0.05 vs db/m vehicle by one-way ANOVA and Tukey’s post hoc. Comparisons by Student’s unpaired t-test: significance denoted by solid lines and trends denoted by dashed lines.

Figure 6. Kidney function.

(a–c) Albuminuria, (d) urinary concentrations of kidney injury molecule 1 (KIM-1), (e) neutrophil gelatinase-associated lipocalin (NGAL), and (f) plasma cystatin C levels in db/m (open) and db/db (grey) mice. Circles (ο) vehicle-treated; squares (◽) empagliflozin-treated; triangles (▵) metformin-treated; and diamonds (◊) empagliflozin + metformin co-treated. Weeks refer to treatment duration. Data are individual mice with means ± SEM (n = 5–10). *P < 0.05 vs db/m vehicle, ^δP < 0.05 vs db/db metformin by one-way ANOVA and Tukey’s post hoc. Comparisons by Student’s unpaired t-test: significance denoted by solid lines and trends denoted by dashed lines. Headings refer to duration of treatment.

Figure 7. Glomerulosclerosis and mesangial expansion.

(a,b) Periodic acid-Schiff (PAS) stain for glomerulosclerosis, and (c,d) collagen IV and (e,f) fibronectin immunostaining in glomeruli in db/m (open) and db/db (grey) mice. Data are means ± SEM (n = 5–9). *P < 0.05 vs db/m vehicle by one-way ANOVA and Tukey’s post hoc. Comparisons by Student’s unpaired t-test: significance denoted by solid lines and trends denoted by dashed lines.

Figure 8. Tubulointerstitial fibrosis.

(a,b) Masson’s trichrome, (c,d) Sirius Red, and (e,f) collagen IV and (g,h) fibronectin immunostaining in tubulointerstitium in db/m (open) and db/db (grey) mice. Data are means ± SEM (n = 5–10). *P < 0.05 vs db/m vehicle by one-way ANOVA and Tukey’s post hoc. Comparisons by Student’s unpaired t-test: significance denoted by solid lines and trends denoted by dashed lines.

Figure 9. Renal cortical expression of profibrotic genes and macrophage marker.

Real-time qPCR for (a) ColIVα1 (encoding collagen type IVα1), (b) Fn1 (encoding fibronectin), (c) Ctgf (encoding connective tissue growth factor, (d) Tgf β1 (encoding transforming growth factor β1), and (e) Cd14 (encoding CD14) in db/m (open) and db/db (grey) mice. Data are means ± SEM (n = 4–6). *P < 0.05 vs db/m vehicle, ^†P < 0.05 vs db/db vehicle by one-way ANOVA and Tukey’s post hoc. Comparisons by Student’s unpaired t-test: significance denoted by solid lines and trends denoted by dashed lines.

Figure 10. Circulating and intra-renal RAS.

(a) Plasma renin activity, and renal cortical (b) renin activity and (c) angiotensin II content in db/m (open) and db/db (grey) mice. Data are (a,b) means ± SEM and (c) median ± IQR with min and max values (n = 5–10). *P < 0.05 vs db/m vehicle, ^†P < 0.05 vs db/db vehicle, ^δP < 0.05 vs db/db metformin by one-way ANOVA and Tukey’s post hoc. (c) Analyzed by Kruskal-Wallis one-way ANOVA and no significant differences observed. Comparisons by Student’s unpaired t-test: significance denoted by solid lines and trends denoted by dashed lines.