Empagliflozin alleviates podocytopathy and enhances glomerular nephrin expression in db/db diabetic mice

Abstract

BACKGROUND

Modern guidelines recommend sodium-glucose cotransporter-2 (SGLT2) inhibitors as the preferred antihyperglycemic agents for patients with type 2 diabetes and chronic kidney disease. However, the mechanisms underlying the renal protective effect of SGLT2 inhibitors are not fully understood.

AIM

To estimate the effect of the SGLT2 inhibitor, empagliflozin (EMPA), on the structure of podocytes and nephrin expression in glomeruli in db/db diabetic mice.

METHODS

We treated 8-wk-old male db/db mice with EMPA (10 mg/kg/d) or vehicle for 8 wk. Age-matched male db/+ mice were included as non-diabetic controls. Parameters of body composition, glycemic and lipid control, and plasma concentrations of leptin, insulin and glucagon were assessed. We evaluated renal hypertrophy as kidney weight adjusted to lean mass, renal function as plasma levels of creatinine, and albuminuria as the urinary albumin-to-creatinine ratio (UACR). Renal structures were studied by light and transmission electron microscopy with a focus on mesangial volume and podocyte structure, respectively. Glomerular nephrin and transforming growth factor beta (TGF-β) were assessed by immunohistochemistry.

RESULTS

Severe obesity and hyperglycemia developed in db/db mice prior to the start of the experiment; increased plasma concentrations of fructosamine, glycated albumin, cholesterol, leptin, and insulin, and elevated UACR were detected. Mesangial expansion, glomerular basement membrane thickening, and increased area of TGF-β staining in glomeruli were revealed in vehicle-treated mice. Podocytopathy was manifested by effacement of foot processes; nephrin-positive areas in glomeruli were reduced. EMPA decreased the levels of glucose, fructosamine and glycated albumin, UACR, kidney hypertrophy, mesangial expansion, glomerular basement membrane thickening, and glomerular TGF-β staining, alleviated podocytopathy and restored glomerular staining of nephrin.

CONCLUSION

These data indicate that EMPA attenuates podocytopathy in experimental diabetic kidney disease. The anti-albuminuric effect of EMPA could be attributed to mitigation of podocyte injury and enhancement of nephrin expression.

Key Words: Diabetes; Chronic kidney disease; Albuminuria; Podocyte; Sodium-glucose transporter 2 inhibitors; Empagliflozin

Core Tip: In the present study, we assessed the influence of the sodium-glucose cotransporter-2 (SGLT2) inhibitor empagliflozin (EMPA) on glomerular structure, with a focus on podocytes, and glomerular staining of nephrin in db/db mice, a model of type 2 diabetic nephropathy. We treated 8-wk-old mice with EMPA (10 mg/kg/d) or vehicle for 8 wk. The results demonstrated that EMPA attenuates podocytopathy and enhances glomerular nephrin staining. These effects were accompanied by mitigation of renal hypertrophy and glomerular transforming growth factor beta expression and a decrease in albuminuria. These results contribute to the understanding of the renal protective effect of SGLT2 inhibitors in diabetes.

INTRODUCTION

Kidney disease is a significant contributor to global morbidity and mortality. In 2017, 697.5 million cases of chronic kidney disease (CKD) were registered; the global prevalence was 9.1%. CKD is considered to be the cause of 35.8 million disability-adjusted life-years, a third of them are associated with diabetes[1]. Since 2000, an increase in the incidence of diabetes-associated end-stage renal disease (ESRD) has been reported in many countries and populations around the globe, including Australia, Bosnia and Herzegovina, Malaysia, Mexico, the Philippines, the Republic of Korea, Russia, Scotland, Singapore, and Taiwan[2]. In the United States, 746557 cases of ESRD were observed in 2017[3]. In patients with diabetes the CKD stage is associated with lower scores in health-related quality of life[4]. Consequently, the development of new approaches for renal protection remains a priority challenge.

The inhibitors of sodium-glucose cotransporter-2 (SGLT2) have opened up new prospects for the prevention of diabetes-related CKD. In randomized clinical trials, SGLT2 inhibitors [empagliflozin (EMPA), canagliflozin and dapagliflozin specifically] ameliorated albuminuria and prevented the decline in renal function in participants with type 2 diabetes (T2D)[5-10]. In particular, in the EMPA-REG OUTCOME trial, EMPA treatment reduced the risk of macroalbuminuria, the decline in renal function, the induction of renal replacement therapy and renal death[5,10]. SGLT2 inhibitors are recommended as preferred add-on antihyperglycemic agents for patients with T2D and CKD by the American Diabetes Association and the European Association for the Study of Diabetes[11,12].

The mechanisms underlyingthe renal protective effect of SGLT2 inhibitors are not fully understood. Among other effects, diminishing intraglomerular pressure and hyperfiltration, suppression of inflammation and fibrogenic pathways, protection against ischemic kidney damage, elevation of glucagon-like peptide-1 and glucagon levels, have been widely discussed[13-15]. Improving tubular metabolism and oxy-genation is considered another mechanism of the protective effect[16]. In glomeruli, podocytes may be the cornerstone of the effect of these drugs. Recently it was revealed that the SGLT2 molecule is expressed in podocytes and is upregulated by the albumin load both in vivo and in vitro[17].

Diabetic kidney disease (DKD) is characterized by podocyte damage depicted by the effacement of foot processes (FPs) and disruption of the slit diaphragm, resulting in elevated permeability for albumin[15]. The effacement of FPs is considered to be a result of actin cytoskeleton disruption. Normally, the cortical actin network binds with proteins of the slit diaphragm, such as nephrin and podocin. Nephrin is considered an adhesion protein and is expressed by mature podocytes[18]. However, the intracellular domain of nephrin acts as a signaling molecule[15]. Mutation-caused abnormalities of the nephrin gene result in congenital nephrotic syndrome of the Finnish type[18]. The increased urinary excretion of nephrin is associated with the albuminuric pattern of CKD in patients with T2D[19]. Since the slit diaphragm provides the most selective barrier for albumin[20], it can be assumed that the anti-albuminuric action of SGLT2 inhibitors could be mediated by an effect on podocytes. Indeed, in the model of fructose- and streptozotocin-induced diabetes in rats, dapagliflozin enhanced glomerular nephrin expression and mitigated histological signs of diabetic ne-phropathy[21].

The aim of this study was to estimate the influence of the SGLT2 inhibitor EMPA on glomerular structural changes, with a special focus on podocytes, and glomerular staining of nephrin in db/db mice, a model of T2D.

MATERIALS AND METHODS

Animals

The experiment was carried out on SPF male db/db mice (BKS.Cg-Dock7m^+/+Lepr^db/J, stock No. 000642, The Jackson Laboratory, United States). These animals are characterized by knock-down of the leptin receptor, which causes diabetes due to impaired satiety, elevated food consumption, and insulin resistance in hepatic, adipose and muscle tissue[22-24]. The db/db mice usually become polyphagic and obese at 4 wk of age and exhibit elevated blood glucose between the fourth and eighth wk. These mice demonstrate shifts in the production of pancreatic and gastrointestinal hormones, and adipokines, which are similar to those in human T2D[25].

Ethical approval

The study was approved by the Animal Ethics Committee of the IC&G SB RAS (Permission 21, April 1, 2014).

Design of the experiment

The experiment was a comparative placebo-controlled study in three parallel groups (Figure 1). At 8 wk of age, db/db diabetic mice were randomized to either EMPA or vehicle (EMPA group, Placebo group). Age-matched heterozygous db/+ mice were included as non-diabetic controls.

Figure 1 Design of the experiment . The db/db diabetic mice were randomized into vehicle and empagliflozin groups. Heterozygous non-diabetic db/+ mice were included as controls. The studied drug or vehicle was administered by gavage once per day from 8 to 16 wk of age. Procedure 1 at wk 8: Weight measurement, body composition assessment, blood and urine sampling. Procedure 2 at wk 12: Weight measurement, body composition assessment, blood sampling. Procedure 3 at wk 16: Weight measurement, body composition assessment, blood and urine sampling, sacrifice and kidney sampling. EMPA: Empagliflozin.

EMPA (Boehringer Ingelheim, Germany) was given at a daily dose of 10 mg/kg, in accordance with previous investigations[26,27]. The agent was suspended in 200 µL of saline and administered intragastrically. The db/db mice, randomized to vehicle, received 200 µL of saline intragastrically. The duration of the experiment was 8 wk.

Experimental conditions and procedures

Mice were housed in OptiMICE cages (Animal Care Systems, United States), 2-3 animals/cage. Food (Ssniff, Germany) and water was provided ad libitum. A 24h cycle of light and dark, temperature of 23±2°C and relative humidity of 45%±10% were maintained.

Body weight and parameters of body composition were estimated every 4 wk. A magnetic resonance imaging (MRI) Body Composition Analyzer (Echo Medical Systems, United States) was used for this procedure. We sampled 150-200 μL of heparinized blood from the retro-orbital sinus with an interval of 4 wk. In addition, urine samples were collected in Petri dishes, which were placed under the animal until spontaneous urination occurred. The obtained samples were frozen (–80 ºC).

At 16 wk of age (8^th wk of the experiment), the animals were decapitated following administration of isoflurane and kidney tissue samples were obtained. Renal structure was estimated by light microscopy, immunohistochemistry (IHC) and transmission electronic microscopy (TEM).

Laboratory investigations

The plasma levels of glucose, total cholesterol, low-density lipoprotein (LDL) and high-density lipoprotein (HDL) cholesterol, triglycerides, creatinine, and uric acid were measured using an AU480 Chemical Analyzer (Beckman Coulter, United States). The plasma concentrations of glycated proteins (fructosamine and glycated albumin) were assessed by ELISA (Mouse FTA ELISA kit and Mouse GAL ELISA kit, MyBioSource Inc., United States). The levels of leptin, insulin and glucagon were assessed by a bead array assay (BioRad, United States). Urinary concentrations of albumin and creatinine were determined by ELISA (Mouse Albumin ELISA kit and Mouse Creatinine kit, Cristal Chem, United States). Albuminuria was expressed as the urinary albumin-to-creatinine ratio (UACR).

Renal histology and IHC

Mayer`s hematoxylin and eosin (H&E) stained 5-µm paraffin-embedded renal tissue sections were observed by light microscopy. For IHC, formalin-fixed, paraffin-embedded 5-µm sections were prepared. Primary antibody to nephrin (1:100, ab58968, Abcam, United Kingdom) or transforming growth factor beta (TGF-β) (1:100, ab66043, Abcam, United Kingdom) were used with the corresponding secondary antibodies (1/2000; Goat Anti-Rabbit IgG H&L+HRP antibody, ab205718, Abcam, United Kingdom). PBS-Tween 20 solution was used for the secondary antibody control.

A morphometric analysis was performed using the Image J platform (https://imagej.net, United States). At least ten glomeruli were estimated on each H&E and IHC slide. The cells under lattice sites of a 5-μm grid were counted and fractional mesangial and capillary volume, volumetric density (V_V) of nephrin-positive and TGF-β positive areas in glomeruli were assessed.

Glomerular ultrastructure assessment

For TEM, paraformaldehyde-OsO₄–fixed samples of renal cortex were embedded in Epon-812. The uranyl acetate and lead citrate contrasted thin sections (70-100 nm) were investigated using a JEM-1400 microscope (JEOL, Japan). In the ultrastructural images, the GBM width, the mean width and numerical density (N_A) of podocyte FPs were estimated. The GBM was assessedat magnification ×100000. Magnification ×40000 was used for the FP width, whereas the measurement of N_A was performed at magnification ×15000.

Statistical processing

Statistical analysis was carried out using STATISTICA 12.0 software (Dell, United States). The non-parametric Mann-Whitney U-test, the ANOVA-χ² test, and Wilcoxon signed rank test were used to assess the differences between two independent groups, three and two dependent groups, respectively. The non-parametric Spearman rank test was used to determine the correlations between variables. We considered the differences as significant at a Pvalue below 0.05.

RESULTS

Weight and body composition

At the age of 8 wk, i.e. at the start of experiment, all diabetic db/db mice showed a substantial elevation in body weight and fat mass when compared to non-diabetic controls (Table 1). These shifts in the placebo group remained stable throughout the experiment. A slight reduction in body weight was observed in control animals (P = 0.01). In contrast, in EMPA-treated mice, both body weight and fat mass demonstrated a further increase (P = 0.008 and P = 0.03, respectively).

Table 1 Body weight and fat mass in db/db and db/+ mice.

Parameter	Age in wk	Group
		Non-diabetic db/+ mice (n = 9)	db/db diabetic mice
			Vehicle (n = 10)	EMPA (n = 9)
Body weight (g)	8	26.8 (25.0-31.1)	37.1 (32.1–41.6)b	37.4 (28.5–41.5)b
	12	26.2 (23.5–28.2)	37.5 (35.8–40.5)b	46.2 (38.8–53.0)b,c
	16	24.8 (21.6–27.7)e	37.8 (34.3–40.4)b	52.6 (42.5–62.2)b,c,f
Fat mass (g)	8	3.1 (1.9–4.1)	18.3 (16.2–22.5)a	18.4 (13.5–20.6)b
	12	2.7 (1.9–3.3)	19.5 (17.0–22.7)b	25.3 (21.7–30.9)b,c
	16	2.7 (2.1–4.5)	19.6 (17.4–22.7)b	28.0 (22.2–35.3)a,d,e
Fat mass (%)	8	11.4 (8.5–16.1)	50.6 (46.6–55.3)b	49.3 (46.8–50.0)b
	12	10.7 (8.1–12.4)	52.5 (45.2–57.0)b	55.5 (51.5–58.3)b
	16	11.7 (7.6–13.2)	55.0 (42.5–62.8)b	53.4 (49.1–56.7)b

^aP< 0.01 vs non–diabetic control (db/+ mice, Mann-Whitney U-test).

^bP< 0.001 vs non–diabetic control (db/+ mice, Mann-Whitney U-test).

^cP< 0.01 vs placebo (Mann-Whitney U-test).

^dP< 0.001 vs placebo (Mann-Whitney U-test).

^eP< 0.01 vs wk 8 and 12 (ANOVA-χ² test).

^fP < 0.001 vs wk 8 and 12 (ANOVA-χ² test). Data are presented as medians (min–max values). EMPA: Empagliflozin.

Plasma biochemical parameters

High glucose levels were observed in db/db mice at the start of the experiment (P < 0.000001, Table 2). Both fructosamine and glycated albumin concentrations were increased also (P < 0.000001 and P = 0.00001, respectively). The vehicle-treated db/db mice demonstrated a further elevation of these parameters during the experiment (P = 0.005). In the EMPA group, all glycemic parameters improved significantly (all P = 0.02), although normal values were not reached.

Table 2 Plasma biochemistry in db/+ and db/db mice.

Parameter	Age in wk	Group
		Non-diabetic db/+ mice (n = 9)	Diabetic db/db mice
			Vehicle (n = 10)	EMPA (n = 9)
Glucose (mmoL/L)	8	9.50 (5.10–11.4)	28.5 (16.8–40.1)b	23.1 (15.3–35.4)b
	12	9.20 (6.10–9.50)	30.4 (22.7–48.0)b	16.3 (13.6–20.7)b,d,e
	16	9.50 (8.50–12.2)	32.7 (22.5–53.1)b	16.1 (9.9–23.6)b,d,e
Fructosamine (µmoL/L)	8	237 (217–249)	456 (424–511)b	480 (425–579)b
	16	239 (222–296)	622 (524–672)b,f	468 (341–491)b,c,e
Glycated albumin (µmoL/L)	8	107 (103–127)	227 (206–239)b	235 (217–261)b
	16	117 (109–133)	283 (252–349)b,f	210 (166–245)b,c,e
Total cholesterol (mmoL/L)	8	2.76 (1.80–3.96)	2.72 (1.48–4.86)	2.76 (1.60–4.05)
	16	2.37 (1.98–2.97)	3.57 (1.68–4.86)e	4.05 (2.10–5.22)e
HDL-cholesterol (mmoL/L)	8	1.17 (0.50–1.62)	1.20 (0.51–2.01)	1.40 (0.93–1.89)
	16	1.26 (0.45–1.80)	1.50 (0.48–2.19)	2.31 (1.05–2.88)e
LDL-cholesterol (mmoL/L)	8	1.32 (0.81–2.10)	1.02 (0.84–2.72)	1.07 (0.84–2.60)
	16	0.81(0.75–1.11)f	0.81 (0.72–1.14)	1.05 (0.60–1.41)
Triglycerides (mmoL/L)	8	1.74 (0.90–2.70)	1.14 (0.51–2.30)	1.25 (0.87–2.30)
	16	1.35 (0.81–2.58)	1.79 (1.17–3.54)e	1.77 (1.29–3.87)e
Creatinine (µmoL/L)	8	67.0 (47.0–78.0)	76.4 (70.2–84.2)	70.5 (54.0–82.2)
	12	68.6 (48.5–74.5)	76.5 (64.5–100.5)a	74.7 (46.2–85.8)a
	16	61.8 (54.0–71.7)	75.6 (63.9–89.4)a	81.3 (55.8–91.2)a
Uric acid (µmoL/L)	8	273 (104–420)	173 (143–341)	281 (165–507)
	16	235 (159–271)	170 (143–187)	204 (179-242)

^aP < 0.01 vs non-diabetic control (db/+ mice, Mann-Whitney U-test).

^bP < 0.001 vs non-diabetic control (db/+ mice, Mann-Whitney U-test).

^cP < 0.01 vs placebo db/db group (Mann-Whitney U-test).

^dP < 0.001 vs placebo db/db group (Mann-Whitney U-test).

^eP < 0.05 vs placebo db/db group (Mann-Whitney U-test).

^fP < 0.01 compared with wk 8 (Wilcoxon test). Data are presented as medians (min–max values). EMPA: Empagliflozin.

There was an increase in plasma total cholesterol levels throughout the experiment in mice treated with placebo (+31.3%, P = 0.04) or EMPA (+46.7%, P = 0.03). In the EMPA group, a marked elevation of HDL-cholesterol was observed (+65%, P = 0.04). However, we recorded no significant changes in the plasma LDL-cholesterol concentrations. There was an increase in the levels of plasma triglycerides in both diabetic groups throughout the experiment (placebo: +57%, P = 0.048; EMPA: +41.6%, P = 0.046).

Hormones in plasma

Diabetic mice had extremely high levels of leptin (Table 3). In the vehicle group, the median level of this hormone exceeded the control by 29.4-fold at wk 8 (P = 0.00005) and by 23.7-fold at wk 16 (P = 0.00002). In EMPA-treated mice a further increase in concentrations of leptin was observed at wk 16 (+46.0%, P = 0.04). A marked increase in insulin levels was revealed in vehicle-treated db/db mice at the start (P = 0.0003) and at the end of the experiment (P = 0.04). Glucagon levels did not significantly change. Treatment with EMPA did not significantly change the levels of insulin and glucagon.

Table 3 Plasma levels of leptin and insulin in db/+ and db/db mice.

Hormone	Age in wk	Group
		Non-diabetic db/+ mice (n = 9)	Diabetic db/db mice
			Vehicle (n = 10)	EMPA (n = 9)
Leptin (ng/mL)	8	3.3 (1.4–6.54)	97.1 (53.2–114.4)b	93.4 (80.8–133.1)b
	16	3.8 (1.5–6.3)	90.0 (21.2–151.4)b	136.4 (53.0–171.2)b,c,d
Insulin (ng/mL)	8	5.6 (3.4–18.0)	25.6 (11.5–45.2)b	21.2 (10.8–40.8)b
	16	10.0 (2.1–22.2)	22.1 (9.8–33.2)a	20.8 (6.0–56.7)a
Glucagon (ng/mL)	8	370 (250–2670)	660 (190–2760)	450 (290–2180)
	16	370 (150–2120)	605 (260–2960)	390 (240–860)

^aP < 0.01 vs non-diabetic control (db/+ mice, Mann-Whitney U-test).

^bP < 0.001 vs non-diabetic control (db/+ mice, Mann-Whitney U-test).

^cP < 0.05 vs placebo (Mann-Whitney U-test).

^dP < 0.05 vs wk 8 (Wilcoxon test). Data are presented as medians (min–max values). EMPA: Empagliflozin.

Renal tests

Diabetic db/db mice had slightly increased plasma creatinine levels at wk 12 and 16 compared to db/+ heterozygotes (P <0.01, Table 2). No deviations were observed in creatinine levels at wk 16 when compared to wk 8 in all the experimental groups.

An elevation in the UACR was noted in db/db mice at 8 wk of age (Figure 2). In the vehicle group, a further increase in the UACR was observed after 8 wk. In contrast, EMPA treatment significantly decreased the UACR.

Figure 2 Urinary albumin-to-creatinine ratio at week 0 and week 8 of the experiment in non-diabetic db/+ mice and diabetic db/db mice, treated with vehicle or empagliflozin. The data are presented as medians, lower and upper quartiles. ^aP < 0.001 vs non-diabetic control (db/+ mice); ^bP < 0.001 vs placebo(Mann-Whitney U-test); ^cP < 0.01 vs week 8 (Wilcoxon test). EMPA: Empagliflozin.

Mesangial volume and GBM width

We observed glomerular hypertrophy, mesangial expansion and GBM thickening, which matched structural signs of diabetic nephropathy, in the vehicle-treated diabetic mice. Table 4 demonstrates the differences in mesangial fractional volume and GBM width between the groups. A significant reduction in mesangial volume and GBM width was revealed in the EMPA group (P = 0.0008 and P = 0.02). At the end of the experiment, these parameters were similar in the control and EMPA groups. There was no difference in the volume of glomerular capillaries between control and diabetic animals.

Table 4 Renal structural parameters in db/+ and db/db mice.

Parameter	Group
	Non-diabetic db/+ mice (n = 9)	Diabetic db/db mice
		Vehicle (n = 10)	EMPA (n = 9)
Mesangium, fractional volume (%)	14.4 (9.8–18.5)	38.6 (34.5–42.7)a	25.5 (20.4–35.8)d
Capillaries, fractional volume (%)	31.3 (28.9–38.7)	34.6 (30.9–36.1)	33.3 (29.5–37.0)
GBM, mean width (nm)	135 (116–157)	163 (135–203)a	139 (116–225)b
Podocyte FPs, mean width (nm)	220 (191–242)	372 (299–426)a	203 (162–317)c
Podocyte FPs (NA, nm–1)	3.39 (3.00–3.79)	2.73 (1.92–3.51)a	3.31 (2.32–4.18)b

^aP < 0.01 vs non-diabetic control (db/+ mice, Mann-Whitney U-test).

^bP < 0.05 vs non-diabetic control (db/+ mice, Mann-Whitney U-test).

^cP < 0.01 vs non-diabetic control (db/+ mice, Mann-Whitney U-test).

^dP < 0.001 vs placebo (Mann-Whitney U-test). Data are presented as medians (min–max values). GBM: Glomerular basement membrane; FPs: Foot processes; EMPA: Empagliflozin.

Podocyte structural changes

Diabetic db/db mice showed effacement of FPs as a manifestation of podocytopathy (Figure 3). In vehicle-treated animals, the mean width of FPs was 1.7-fold greaterand N_A of FPs was 1.5-fold lower compared with controls (Table 4). In the group of actively treated animals, the width and N_A of podocyte FPs were similar to those in db/+ controls.

Figure 3 Podocyte foot processes. A: Non-diabetic db/+ mice; B: Vehicle-treated db/db diabetic mice demonstrate effacement of podocyte foot processes(FPs); C: Empagliflozin restores the structure of FPs in db/db mice. Transmission electron microscopy: ×100000.

Glomerular staining for nephrin and TGF-β

A reduction in the percentage of glomerular nephrin-positive areas was observed in vehicle-treated mice (Figures 4 and 6). In the EMPA group, staining for glomerular nephrin was markedly increased.

Figure 4 Glomerular staining for nephrin. A:Non-diabetic db/+ mice; B: Vehicle-treated db/db mice; C: Empagliflozin-treated db/db mice; A-C: Immunohistochemistry with anti-nephrin antibody; D: Negative control.

Figure 5 Glomerular staining for TGF-β. A:Non-diabetic db/+ mice; B: Vehicle-treated db/db mice; C: Empagliflozin-treated db/db mice; A-C: Immunohistochemistry with anti-TGF-β antibody; D: Negative control.

Figure 6 Volumetric density of nephrin-positive areas and TGF-β-positive areas in glomeruli of nondiabetic db+ and diabetic EMPA-treated and vehicle-treated db/db mice. A: Nephrin-positive areas; B: TGF-β-positive areas. ^aP < 0.001 vs non-diabetic db/+ mice; ^bP < 0.01; ^cP < 0.001 vs vehicle treated db/db mice (Wilcoxon test). Vv: Volumetric density; TGF-β: Transforming growth factor β; EMPA: Empagliflozin.

The mice treated with vehicle demonstrated larger TGF-β positive areas in the glomeruli than non-diabetic mice (P=0.00001, Figures 5 and 6). EMPA decreased TGF-β staining (P = 0.0004).

The relationships between laboratory parameters and renal structural changes

Fructosamine and glycated albumin demonstrated positive correlations with mesangial volume, mean width of GBM and FPs, and TGF-β staining; the correlations with N_A of podocyte FPs and nephrin-positive areas were negative (Table 5). Plasma creatinine correlated positively with mesangial volume, and demonstrated a negative correlation with nephrin-positive areas in glomeruli. The UACR levels were positively correlated with mesangial volume, GBM width, width of podocyte FPs and TGF-β-staining. The UACR was negatively correlated with nephrin staining.

Table 5 The relationships of glycemic control indicators, plasma creatinine and urinary albumin-to-creatinine ratio with renal structural parameters.

Parameter	Fructosamine	Glycated albumin	Plasma creatinine	UACR
Mesangial volume, VV	r = 0.69	r = 0.67	r = 0.57	r =0.81
	P = 0.0002	P = 0.0004	P = 0.004	P = 0.000002
Capillary volume, VV	r = 0.13	r = 0.11	r = 0.30	r = 0.05
	P > 0.05	P > 0.05	P > 0.05	P > 0.05
GBM width	r = 0.69	r = 0.69	r = 0.21	r = 0.53
	P = 0.0008	P = 0.007	P > 0.05	P = 0.02
Podocyte FPs, width	r = 0.62	r = 0.63	r = –0.15	r = 0.50
	P = 0.003	P = 0.003	P > 0.05	P = 0.03
Podocyte FPs, NA	r = –0.57	r = -0.58	r = 0.04	r = –0.25
	P = 0.008	P = 0.007	P > 0.05	P > 0.05
Nephrin-positive areas in glomeruli, VV	r = –0.68	r = –0.65	r = –0.51	r = –0.76
	P = 0.006	P = 0.009	P = 0.04	P = 0.001
TGF-β positive areas in glomeruli, VV	r = 0.65	r = 0.65	r = 0.09	r = 0.67
	P = 0.003	P = 0.002	P >0.05	P = 0.001

UACR: Urinary albumin-to-creatinine ratio; GBM: Glomerular basement membrane; FPs: Foot processes; Vv: Volumetric density; TGF-β: Transforming growth factor β. Spearman’s rank correlation coefficients are shown.

DISCUSSION

In this study, we tested the hypothesis that the anti-albuminuric activity of EMPA could be mediated by its effect on podocyte integrity. We chose db/db mice as a model of T2D. In these animals, we documented obesity as an increase in the fat mass verified by MRI. The observed hyperglycemia, hypercholesterolemia, hypertriglyceridemia, hyperinsulinemia and hyperleptinemia characterized the metabolic profile of T2D. The structural signs of DKD in db/db mice were accompanied by a marked increase in urinary albumin excretion and a slight increase in plasma creatinine. These findings are in accordance with data from prior investigations which demonstrated that structural signs of DKD in db/db mice appeared at 8 wk and were retained at 12 and 16 wk of age[28], whereas a reduction in creatinine clearance occurred later, at the age of 15 wk[29]. In our study, we used male db/db mice in the experiment; in a previous study, EMPA also demonstrated anti-albuminuric and anti-fibrotic effects in the kidneys of female db/db mice[30].

In our model of T2D, EMPA prevented the elevation of albuminuria and preserved podocyte morphology. Specifically, EMPA alleviated the effacement of FPs, a principal sign of podocytopathy. It was previously shown that ipragliflozin, another SGLT2 inhibitor, preserved podocyte integrity in db/db diabetic mice[31]. Dapagliflozin also limited podocyte injury in a protein-overload model of proteinuric nondiabetic nephropathy[17]. In uninephrectomized db/db mice, treatment with metformin, ramipril, and EMPA increased a density of podocyte filtration slit [32]. In these studies, as well as in our work, the alleviation of podocytopathy was accompanied by a decrease in the protein excretion.

As the podocyte FPs form the most refined layer of the glomerular filter, changes in their molecular and structural organization could lead to increased permeability to albumin. In this study, we found a reduction in glomerular nephrin-positive staining in db/db mice. The data are in agreement with those from previous studies which demonstrated down-regulation of glomerular nephrin expression in patients with T2D[33,34]. Traditionally, nephrin is considered a structural component of the slit diaphragm. At present, the cytoplasmic domain of the nephrin molecule is presumed to be involved in the development of glomerular diseases[35]. Nephrin regulates podocyte cytoskeleton organization, and nephrin down-regulation may alter the actin structure, leading to the effacement of FPs and slit diaphragm breakdown[15]. In our study, an enlargement of nephrin-positive areas under EMPA treatment was associated with a reduction in albuminuria. These results provided further support for the notion that the anti-albuminuric effect of EMPA could be mediated, at least partially, via enhancement of nephrin expression.

We observed a reduction in glomerular TGF-β-positive areas in EMPA-treated mice. Overexpression of TGF-β is considered to be a cornerstone among molecular mechanisms of diabetic glomerulosclerosis. In addition, TGF-β1 enhances the permeability of the glomerular barrier and decreases tubular reabsorption of albumin[36]. It was found that the TGF-β signaling pathway impairs nephrin expression in podocytes[37,38]. Therefore, a decrease in TGF-β expression could also contribute to the anti-albuminuric effect of EMPA.

It is still a matter of debate whether SGLT2 inhibitors exert their renal protective activity in a glucose-independent manner. In our study, the severity of podocytopathy correlated with hyperglycemia. In EMPA-treated mice, we observed a reduction in plasma fructosamine and glycated albumin levels, regarded as intermediate-term indicators of glycemic control[39], accompanied by a decrease in plasma glucose. Although normoglycemia was not achieved under EMPA treatment, the podocyte morphology improved significantly. The results give further support to the notion that the non-glycemic effects of SGLT2 inhibitors could be important for their protective effect on the kidneys. Further evidence of this assumption was obtained recently in the DAPA-CKD randomized clinical trial. In this trial, dapagliflozin demonstrated a protective effect on renal function in both diabetic and non-diabetic CKD[40].

The changes in the body weight should be taken into account when interpreting the effects of SGLT2 inhibitors. In our study, EMPA contributed to the additional increase in body weight and fat mass, while the opposite pattern was observed in clinical studies[41]. The increase in the body weight was observed under administration of EMPA[26,42], dapagliflozin[43] and ipragliflozin[44] in experiments similar to ours. However, no increase in food consumption was documented during EMPA treatment[26,42]. The mechanism of EMPA-induced increase in rodent body weight requires further research. Interestingly, we found a marked increase in plasma cholesterol levels in the EMPA group, but the change was attributed to HDL-cholesterol elevation. Thus, it is unlikely that the shifts in the fat mass or lipid metabolism could contribute to the renal protective activity of EMPA.

In this study, we demonstrated that EMPA restores the structure of podocyte FPs in diabetes. It was postulated that the changes in FP morphology may be the result of podocyte cytoskeletal reconfiguration[45]. In cultured podocytes, the SGLT2 inhibitor dapagliflozin limited cytoskeletal remodeling induced by albumin load[17]. Thus, podocytes could be considered direct targets of SGLT2 inhibitors. We have shown recently that EMPA reactivated autophagic flux in the podocytes of db/db mice[46]; the promotion of autophagy could be an explanation of the renal protective effect of SGLT2 inhibitors[46,47].

CONCLUSION

The results of this study indicate that EMPA attenuates podocytopathy in db/db mice, a model of T2D nephropathy. The anti-albuminuric effect of EMPA could be attributed to the mitigation of podocyte injury and enhanced nephrin expression.

ARTICLE HIGHLIGHTS

Research background

Sodium-glucose cotransporter-2 (SGLT2) inhibitors have opened up new prospects for the prevention of chronic kidney disease (CKD). Modern guidelines recommend SGLT2 inhibitors as the preferred antihyperglycemic agents for patients with type 2 diabetes (T2D) and CKD. In the empagliflozin (EMPA)-REG OUTCOME trial, the SGLT2 inhibitor EMPA prevented increased albuminuria and progression of renal function decline in T2D patients.

Research motivation

The exact mechanisms of the renal effect of SGLT2 inhibitors are not fully understood. Since diabetic kidney disease is accompanied by podocyte injury and this injury results in albuminuria, it can be speculated that the anti-albuminuric action of SGLT2 inhibitors could be mediated via podocytes. The influence of SGLT2 inhibitors on podocyte structure and function remains to be clarified.

Research objectives

In this study, we estimated the effect of EMPA on glomerular structural changes, with a special focus on podocytes, and glomerular nephrin staining, in db/db mice, a model of T2D. Thereby, we assessed the hypothesis that the anti-albuminuric effect of EMPA is linked with the preservation of podocyte integrity in diabetic kidney disease.

Research methods

We treated 8-wk-old db/db male mice with EMPA (10 mg/kg/d) or vehicle for 8 wk. Heterozygous db/+ mice were included as non-diabetic controls. Body weight and body composition were assessed every 4 wk. The plasma levels of glucose, glycated proteins, lipids, creatinine, leptin, insulin, glucagon, and albuminuria were monitored. Renal structure was studied by light and transmission electron microscopy. Glomerular nephrin and transforming growth factor beta (TGF-β) were assessed by immunohistochemistry. The fractional mesangial and capillary volume, glomerular basement membrane and podocyte foot processes width, numerical density of podocyte foot processes, as well as volumetric density of nephrin-positive and TGF-β-positive areas in glomeruli were quantified.

Research results

Throughout the experiment, diabetic mice showed a dramatic elevation in the body weight and fat mass. The observed increase in serum levels of glucose, glycated proteins, cholesterol, triglycerides, leptin and insulin characterized the metabolic profile of T2D. In db/db mice, EMPA mitigated renal hypertrophy, decreased mesangial fractional volume, the width of glomerular basement membrane, and glomerular TGF-β staining. EMPA-treated db/db mice demonstrated fewer signs of podocyte foot process effacement and restored glomerular staining of nephrin. These effects were correlated with a reduction of albuminuria. The improvement in podocyte integrity was observed even though normoglycemia was not achieved.

Research conclusions

This is the first study to quantitatively describe the effects of EMPA on podocyte structure and glomerular staining of nephrin in a model of T2D. The data indicate that EMPA attenuates podocytopathy in experimental diabetic kidney disease. The anti-albuminuric effect of EMPA could be attributed to mitigation of podocyte injury and enhancement of nephrin expression. The protective effect of EMPA on the kidneys is realized even in condition of suboptimal glycemic control.

Research perspectives

Uncovering the molecular pathways which are important for the effect of SGLT2 inhibitors on podocyte homeostasis is a challenge for further research.

ACKNOWLEDGEMENTS

Light and electron microscopy was carried out at the Joint Access Center for Microscopy of Biological Objects of IC&G SB RAS (Novosibirsk, Russia). We are sincerely thankful to Irina Ischenko, Maria Borisova, Sergey Bayborodin, Taisiya Aleshina, and Victor Rosin for their facilitation with animal procedures, tissue processing, and microscopy.