Progression of diabetic kidney disease in T2DN rats

Abstract

Diabetic kidney disease (DKD) is one of the leading pathological causes of decreased renal function and progression to end-stage kidney failure. To explore and characterize age-related changes in DKD and associated glomerular damage, we used a rat model of type 2 diabetic nephropathy (T2DN) at 12 wk and older than 48 wk. We compared their disease progression with control nondiabetic Wistar and diabetic Goto-Kakizaki (GK) rats. During the early stages of DKD, T2DN and GK animals revealed significant increases in blood glucose and kidney-to-body weight ratio. Both diabetic groups had significantly altered renin-angiotensin-aldosterone system function. Thereafter, during the later stages of disease progression, T2DN rats demonstrated a remarkable increase in renal damage compared with GK and Wistar rats, as indicated by renal hypertrophy, polyuria accompanied by a decrease in urine osmolarity, high cholesterol, a significant prevalence of medullary protein casts, and severe forms of glomerular injury. Urinary nephrin shedding indicated loss of the glomerular slit diaphragm, which also correlates with the dramatic elevation in albuminuria and loss of podocin staining in aged T2DN rats. Furthermore, we used scanning ion microscopy topographical analyses to detect and quantify the pathological remodeling in podocyte foot projections of isolated glomeruli from T2DN animals. In summary, T2DN rats developed renal and physiological abnormalities similar to clinical observations in human patients with DKD, including progressive glomerular damage and a significant decrease in renin-angiotensin-aldosterone system plasma levels, indicating these rats are an excellent model for studying the progression of renal damage in type 2 DKD.

INTRODUCTION

The diabetic kidney disease (DKD) animal models used in research did not reproduce most or all of the lesions of human diabetic nephropathy (DN), and the severity and similarity to the DN phenotype observed in humans are the main reasons for skepticism when applying knowledge from basic biology to address critical medical needs. The initial characterization of the type 2 diabetic nephropathy (T2DN) phenotype by Nobrega et al. (26) is missing the important evidence required for the correct classification of advanced DN in T2DN rats. Here, using novel innovative approaches, we provide a detail characterization of the DN phenotype in this strain.

DKD is a common complication of diabetes, which frequently leads to end-stage kidney disease and an increase in the risks for cardiovascular morbidity and mortality. The changes to renal structure and function initiated by hyperglycemia and subsequent decline in kidney functions eventually result in DKD. The number of patients suffering from diabetes is expected to reach ~35 million by 2035, and roughly 40% of these cases will develop chronic kidney disease and progress to DKD (23, 30). Of those diagnosed with diabetes, ~90−95% have type 2 diabetes. As of 2014, >20% of these individuals had developed DKD (10, 50). Treatments based on inhibition of the renin-angiotensin system (RAS) reduce the rates of vascular dysfunction but seldom affect progression to kidney disease (38). Multiple clinical trials using a combination of angiotensin-converting enzyme and angiotensin receptor blocker dual therapy have proven that, for most patients with type 2 diabetes, such treatments are not effective and often lead to an increased risk of hyperkalemia and acute kidney injury (12, 28, 39). In contrast, the combination of RAS and Na⁺-glucose cotransporter-2 (SGLT2) inhibitors has demonstrated a profound renoprotective effect, but further clinical investigation is still required because of the increased risk of ketoacidosis, hypoglycemia, hyperkalemia, and osmotic diuretic effects, all of which could potentially lead to volume depletion, bone fractures, and urinary tract infections during this treatment (4, 24, 31, 36, 54, 55).

Hallmarks of DKD include a decline in glomerular filtration rate (GFR), albuminuria, changes to renal morphology (thickening of the glomerular basement membrane, mesangial expansion, interstitial fibrosis, glomerular hypertrophy, glomerulosclerosis, podocyte foot process effacement, and arterial hyalinosis), and the comorbidity of hypertension (1, 8, 40, 51). The irreversible damage of glomeruli in DKD striving to maintain whole kidney filtration and reabsorption within the normal range often results in hyperfiltration (49). This effect is accompanied by renal hypertrophy and polyuria followed by increased albuminuria and decreased GFR. Dyslipidemia is one of the major risk factors for cardiovascular disease in the incidence and progression of DKD (25), which can cause a nephrotic syndrome or renal insufficiency and is therefore strongly associated with glomerular damage (52). High cholesterol and triglyceride plasma levels were observed in type 2 diabetes and have been demonstrated to be an independent risk factor for cardiorenal disease progression in humans (53).

Preclinical studies on animal models of DKD are used to predict new targets for therapy, but most have failed in subsequent randomized clinical trials or have shown no renoprotective effects (2). Therefore, animal models that more closely mimic the characteristics of the diabetic type 2 phenotype in humans are required for reliable research studies. T2DN rats were previously created by the introgression of the mitochondria and some passenger loci from the Fawn hooded hypertensive rat into the background of the Goto-Kakizaki (GK; type 2 diabetic) rat. Nobrega et al. (26) characterized this strain as having the defining characteristics of DKD mentioned above, such as proteinuria and focal segmental glomerulosclerosis. The histological changes exhibited by T2DN rats were noted to be similar to patients diagnosed with this disease.

Most interestingly, they found that, as the animals aged, the severity of the DKD phenotype increased (26). Despite the aforementioned publication, this model has largely been ignored, likely because of the lack of a thoroughly described phenotype demonstrating the similarities between T2DN rats and hallmarks of human DN pathology.

The primary goal of the present study was to reassess the phenotype of the T2DN rat using modern techniques with an emphasis on the development of glomerular and podocyte injury and to determine the potential of the T2DN rat to be used as a type 2 DKD model for the assessment of potential therapeutic agents. We performed detailed analyses of equilibrium renin-angiotensin-aldosterone system (RAAS) components and a comprehensive histological evaluation of renal and glomerular damage and applied a novel nonoptical topographical imaging (SICM) approach to reveal the podocyte foot process effacement during the development of DKD.

MATERIALS AND METHODS

Experimental protocol and animals.

The animal use and welfare adhered to the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals following protocols reviewed and approved by the Medical College of Wisconsin Institutional Animal Care and Use Committee. Animals at different ages, 8-, 12- (young adult), or >48 (aged)-wk-old male Wistar (nondiabetic), GK (type 2 diabetic), and T2DN rats, were used for experiments. Wistar rats were purchased from Charles River Laboratories. GK and T2DN rats had been inbred for multiple generations at the Medical College of Wisconsin. Animals were fed a normal salt diet (no. 5001, LabDiet, Purina) with water and food provided ad libitum. Dahl salt-sensitive rats used in the SICM experiments obtained from the Medical College of Wisconsin colony were fed a normal salt chow. Rats were anesthetized, and their kidneys were flushed with PBS via aortic catheterization (3 ml·min⁻¹·kidney⁻¹ until blanched). For each rat, one kidney was used for glomeruli isolation and Western blot analysis, and the other kidney was used for immunohistochemistry analyses.

Electrolyte measurements and albuminuria assay.

Blood glucose levels were measured at 12 wk of age using a tail prick and a glucometer. For urine collection, rats were placed in metabolic cages (no. 40615, Laboratory Products) for a 24-h urine collection. These urine samples were used to determine electrolytes, microalbumin, creatinine, and nephrin levels. Whole blood and urine electrolytes and creatinine were measured with a blood gas and electrolyte analyzer (ABL system 800 Flex, Radiometer, Copenhagen, Denmark) (29). Kidney function was determined by measuring albuminuria using a fluorescent assay (Albumin Blue 580 dye, Molecular Probes, Eugene, OR) read by a fluorescent plate reader (FL600, Bio-Tek, Winooski, VT). Nephrin levels in urine were assessed using Western blot analysis (ab58968, Abcam) as previously described (45).

Histological staining and analysis of kidney injury.

Rat kidneys were cleared of blood, formalin fixed, paraffin embedded, sectioned, and mounted on slides as previously described (29). Slides were stained with Masson’s trichrome stain. The localization of glomeruli in Wistar, GK, and T2DN rats was performed by a novel and robust application of convolutional neural nets (9). Briefly, color deconvolution was performed using the solved color-basis vectors. Intensity images were automatically created for the “red” and “blue” color components. Histogram equalization was then performed as previously described (9) on the extracted red stains over only tissue-related pixels with the background removed through masking. A glomerular injury score was assessed using semiquantitative morphometric analysis based on a scale of 0–4 as previously described (29, 35).

Spatial resolution of glomerular damage was shown on the generated heatmaps and based on blind observers’ assigned scores for representative examples within groups. Specifically, heatmaps were generated using a cubic interpolation over the area of the kidney with the human-assigned scores acting as reference points that were anchored at the center of each glomeruli location. The colors of the heatmap were then scaled to between 0 and 4 to show spatial trends of injury. Finally, a cumulative distribution plot was generated (OriginPro 9.0) using glomerular injury scores obtained by semiquantitative morphometric analysis, and the probability for a corresponding score interval was calculated. Protein cast analysis was performed using a color thresholding method involving Metamorph software (Molecular Devices, Sunnyvale, CA). Fibrosis was assessed using color deconvolution in the Fiji image application (ImageJ 1.51u, NIH).

For podocin (Nphs2) staining, freshly isolated glomeruli were allowed to adhere to microscopy coverslips double coated with poly-l-lysine. These adhered glomeruli were then fixed with chilled 4% paraformaldehyde in PBS with 1 mM CaCl₂ and 2 mM MgCl₂ for 20 min and then gently washed three times with ice-cold PBS. Next, the glomeruli were probed with a podocin primary antibody in 2% BSA-0.1% Triton X-100-PBS overnight at 4°C (NPHS2, 1:500, ab50339, Abcam). The following day, the glomeruli were washed three times with cold PBS and incubated with Alexa fluorophore-labeled secondary antibody (1:500, ThermoFisher, Pittsburgh, PA) in 2% BSA-PBS at room temperature in the dark. After three PBS washes, the glomeruli were incubated with 0.5 μg/mL Hoescht nuclear stain in PBS for 10 min at room temperature in the dark. After five final washes with PBS, the tissue was preserved and coverslip mounted with Fluoromount-G (SouthernBiotech, Birmingham, AL). Z-stack images with 2 µm z-steps were captured on a confocal Nikon A1R inverted microscope using a Plan Apo ×40 DIC M N2 objective with 0.95 numerical aperture controlled by Nikon Elements AR software (Nikon, Tokyo, Japan). Postimage processing was performed with Fiji image software.

Scanning ion conductance microscopy.

Podocyte morphology was imaged in freshly isolated glomeruli using super-resolution hopping probe ion conductance microscopy, an advanced version of SICM. The identification of glomerular structures and SICM three-dimensional topographical imaging was carried out using the custom-modified SICM scanner ICNano (ICAPPIC), as previously described (27). Briefly, the sample was manually positioned in the x-y directions under an optical microscope, and the scanning pipette was positioned in the z-direction with a piezoelectric actuator. Fine-tipped scanning nanopipettes were pulled from borosilicate glass (outer diameter/internal diameter: 1/0.5 mm) with the horizontal laser puller P-97 (Sutter Instruments, Novato, CA). The pipette resistance was in the range of 80–100 MΩ corresponding to an estimated tip diameter of 90–120 nm. Nanopipettes were held in voltage-clamp mode with an Axopatch 200B patch-clamp amplifier (Axon Instruments). The amplifier headstage was mounted on the z-scanning head. The amplifier output signal was monitored by the SICM electronics, which simultaneously controlled sample and pipette positioning. The scan system was mounted on a Nikon TE2000-U inverted microscope (Nikon Instruments). Raw SICM data were processed using SICMImageViewer microscopy analysis software (ICAPPIC).

After constructing a three-dimensional topographical image of identified glomerular structures, their x-y-z dimensions were determined (changes in z-axes displayed as a pseudocolor). The three-dimensional profiles of the secondary processes of podocytes were built using longitudinal z-axis line-scan sections on the glomerular vessels imaged with SICM. The line scan was defined on a selected interval of [0, 2.5] µm and represented as a differentiable function [f(z)]. The total variation (V) of each corresponding line scan was calculated according to the following formula:

$$V\left(f\right)={\displaystyle {\int }_0^{2.5}\left|f^{\prime }\left(z\right)\right|}\text{d}z$$

Quantification of the RAAS.

Analysis and quantification of ANG I, ANG II, ANG-(1–7), ANG-(1–5), ANG-(2–8), ANG-(3–8), ANG-(2–10), ANG-(2–7), ANG-(1–9), ANG-(3–7), and aldosterone were carried out according to quantifications of steady-state levels in equilibrated heparin plasma samples by Attoquant Diagnostics (Vienna, Austria) according to the company’s protocol.

Briefly, angiotensin peptide levels and aldosterone were measured in conditioned lithium-heparin plasma at 37°C. Stable isotope-labeled internal standards for each metabolite were added to stabilized plasma samples at a concentration of 200 pg/mL and subjected to LC-MS/MS-based angiotensin quantification by Attoquant Diagnostics, as previously described (3, 7, 16).

Statistical analysis.

Data are presented as means ± SE. In the box chart diagram, the ends of the box are ±SE, and the median is marked by a vertical line inside the box. The whiskers are ±SD. Data were tested for normality (Shapiro-Wilk) and equal variance (Levene’s homogeneity test). Statistical analysis consisted of one-way ANOVA or other if indicated (SigmaPlot 12.5 or OriginPro 9.0).

To reduce the number of false discoveries, we declared a statistical significance when P < 0.01. In addition, when an ANOVA test was significant (P < 0.01), Holm-Sidak’s multiple-comparisons adjustment was applied to all pairwise P values. For each ANOVA test, the number of groups is denoted by k.

RESULTS

T2DN rats display typical characteristics of DKD.

Despite the high degree of genetic similarity between T2DN and GK rats (97%), diabetes starts earlier and progresses more severely in T2DN rats compared with GK rats (26). The GK rat, a spontaneously diabetic rat that was developed through selective breeding of Wistar rats with high glucose levels, displayed mild or no DN symptoms (15).

First, we confirmed the development of diabetes and major characteristics of DKD in T2DN compared with GK (diabetic control) and Wistar (nondiabetic control) male rats. Rats develop rapidly during infancy and become completely sexually mature at 8 wk. Two main age groups of animals were examined in our experiments: young adult (12 wk old) and aged (over 48 wk old) groups, which are roughly equivalent to ~20 and >40 human years (43). Figure 1A shows that, at 12 wk of age, both diabetic strains display mild hyperglycemia, which further progresses with age and becomes more pronounced in T2DN rats. The reduction in overall body weight in diabetic animals likely driven by insulin resistance and impairment of cellular glucose reabsorption has a profound effect in both age groups, as shown in Fig. 1B. Adaptations to increase renal mass primarily reflect tubular growth to restore renal function toward normal (13). Figure 1C shows pathological renal growth in diabetic strains with significantly elevated index in aged T2DN rats compared with both Wistar and GK rats. This damage precedes the development of glomerular and tubular damage in DN. In addition, young (12 wk) T2DN rats showed renal glycosuria (up to 4 times increase in the urine) and significant blood Na⁺ deficits, which can be associated with polyuria and osmotic diuresis without the presence of acidosis (Table 1).

Fig. 1.

Development of diabetes in type 2 diabetic nephropathy (T2DN), Goto-Kakizaki (GK), and nondiabetic Wistar rats. A: blood glucose (Glu) levels (random, nonfasting) in 12-wk-old (top; ANOVA, P < 0.025, k = 3 groups) and >48-wk-old (bottom, *P < 0.002, k = 3) rats. B: total body weight (TBW) of age-matched 12- or 48-wk-old rats (*P < 0.001 vs. Wistar rats for both groups, k = 3). C: normalized two kidney-to-body weight (2K/BW) ratio for 12-wk-old (ANOVA, *P < 0.001 vs. Wistar rats, k = 3) or 48-wk-old (ANOVA, *P < 0.001, k = 3) rats.

Table 1.

Blood electrolytes analysis of 12-wk-old Wistar, GK, and T2DN rats

	Wistar	GK	T2DN
pH	7.31 ± 0.02	7.37 ± 0.03	7.44 ± 0.01
Calcium, mmol/L	1.3 ± 0.1	1.2 ± 0.1	1.1 ± 0.1
Sodium, mmol/L	141.1 ± 0.7	138.6 ± 0.9	136.4 ± 1.1*
Potassium, mmol/L	3.8 ± 0.1	4.0 ± 0.1	3.6 ± 0.2
Chloride, mmol/L	104.4 ± 1.3	101.1 ± 0.8	100.0 ± 1.4

Data are presented as means ± SE (ANOVA, k = 3 groups); N ≥ 8 rats for each group. GK, Goto-Kakizaki; T2DN, type 2 diabetic nephropathy. Wistar and GK groups were not significantly different. In addition, when an ANOVA test was significant (P < 0.01), a Holm-Sidak multiple-comparisons adjustment was applied to all pairwise P values.

^*P < 0.01, T2DN compared with Wistar rats.

Diabetic T2DN and GK rats have significant alterations in components of the RAAS.

Glucose and blood pressure control by RAAS inhibitors helps to reduce the risk of DKD but does not entirely prevent it. Together with SGLT2 inhibitors (25a), dual therapy is currently recommended to reach the maximum renoprotective effect and treat complications associated with DKD.

Therefore, we were interested in testing components of the RAAS in GK and T2DN rats compared with control Wistar rats. Equilibrium levels of RAAS hormones, including ANG I, ANG II, and aldosterone, and alternative RAAS components were simultaneously quantified in plasma samples of 12-wk-old Wistar, GK, and T2DN rats using an innovative mass spectrometry (LC-MS/MS)-based approach (33). Figures 2 and 3 show the results from this analysis. Both diabetic strains had significant decreases in most RAAS metabolites and plasma renin activity (PRA), as presented by the sum of ANG I and ANG II. T2DN rats exhibited low renin activity with normal aldosterone levels (Fig. 3B). Interestingly, aldosterone levels were decreased in GK rats. Angiotensin-converting enzyme activity (calculated by normalizing ANG II to ANG I levels) did not differ between groups.

Fig. 2.

Angiotensin peptide plasma expression levels in diabetic rats. A−D: equilibrium levels of ANG I-(1–10) (A), ANG II-(1–8) (B), ANG III-(2–8) and ANG-(1–7) (C), and ANG IV-(3–8) and ANG-(1–5) (D) in 12-wk-old Wistar, Goto-Kakizaki (GK), and type 2 diabetic nephropathy (T2DN) rats. P values are shown for each graph (determined by one-way ANOVA, k = 3 groups). Each dot represents one rat. ACE, angiotensin-converting enzyme; ACE2, angiotensin-converting enzyme 2; APA, aminopeptidase A; APN, aminopeptidase N. *Statistically significant differences are shown.

Fig. 3.

Aldosterone levels and angiotensin-converting enzyme (ACE) and renin activity in diabetic rats. A: summary graph of plasma aldosterone levels in 12-wk-old Wistar, Goto-Kakizaki (GK), and type 2 diabetic nephropathy (T2DN) rats. B: ratio of aldosterone and ANG II (AA2 ratio) is representative of changes in primary aldosteronism. C: ratio of ANG II to ANG I demonstrating ACE activity. D: sum of ANG I and ANG II representing circulating plasma renin activity (PRA). P values are shown for each graph (ANOVA, k = 3 groups). Each dot represents one rat. *Statistically significant differences are shown.

Progression of DKD and kidney injury in T2DN rats.

Abnormal renal growth precedes the development of glomerular hypertrophy that increases in glomerular size and represents both cellular hypertrophy and hyperplasia (13). Figure 4A shows representative images of medullary and cortical areas of both diabetic GK and T2DN rats at ×20 and ×40 magnification, respectively. The renal lesion was characterized by hyaline (protein) casts, glomerular hypertrophy, and glomerulosclerosis, which were highly present in the renal tissue of T2DN compared with diabetic control GK rats. Figure 4, B and C, shows summarized analyses of medullary casts and whole kidney fibrosis, indicating that T2DN rats had increased renal damage at 48 wk (late stage) compared with both T2DN rats at 12 wk old and age-matched GK rats. Histological changes of the tubular epithelium in T2DN include acute tubular necrosis with flattened and dilated epithelium presenting along with glomerular lesions. A high presence of mononuclear cells in the interstitium was also observed.

Fig. 4.

Progression of kidney injury in type 2 diabetic nephropathy (T2DN), Goto-Kakizaki (GK), and nondiabetic Wistar rats. A: representative images of kidney tissue stained with Masson’s trichrome (×20 magnification and expanded areas at ×40 magnification). Note tubulointerstitial disease present along with diffuse mesangial sclerosis and diffuse thickening of the capillary walls in the T2DN glomerulus. Scale bars = 100 and 50 µm, respectively. B and C: summary graphs of the medullary protein cast area (percent total kidney area; B) and cortex fibrosis (C) for GK and T2DN rats at 12 and 48 wk (ANOVA, *P < 0.002, k = 6 groups). D: diuresis in T2DN rats at 8, 12, and >48 wk (repeated-measures ANOVA, *P < 0.01). E: albuminuria (urinary albumin normalized to creatinine) in 8-, 12-, and >48-wk-old T2DN and >48-wk-old Wistar and GK rats (*P < 0.01, ANOVA, k = 10). F: blood cholesterol in 12- and 48-wk-old rats (ANOVA, *P < 0.01, T2DN vs. GK or Wistar rats, k = 15). G: plasma blood urea nitrogen (BUN) in 12- and 48-wk-old rats (ANOVA, P < 0.021 for 48-wk-old T2DN vs. GK rats, k = 6). H and I: creatinine clearance [creatinine urine concentration × urine flow)/creatinine serum concentration] (H) and urine osmolarity (I) in >48-wk-old rats (ANOVA, *P < 0.01, T2DN vs. GK or Wistar rats, k = 3). For all data sets, urine samples were collected for 24 h.

Excess levels of sugar in the blood are commonly associated with polyuria. T2DN rats clearly develop polyuria as young adults (12 wk), and it continued in aged animals (48 wk; Fig. 4D), likely causing glomerular hyperfiltration progressing to persistent albuminuria associated with glomerular and proximal tubule injury. Significant increases in urine albumin were observed in T2DN rats after 48 wk (Fig. 4E); however, age-matched Wistar or GK rats did not develop albuminuria, as was previously reported (51). Dyslipidemia in type 2 diabetes is characterized by high triglyceride levels and decreased HDL-cholesterol (42). Cholesterol metabolism was not significantly changed in nondiabetic and GK controls, indicating high cholesterol synthesis specifically in the aged T2DN population (Fig. 4F). The elevated blood urea nitrogen levels (Fig. 4G) and substantial decrease in creatinine clearance (24-h urine collection; Fig. 4H) among the aged T2DN group demonstrated the progressive decline in kidney function and development of DKD. Similar to patients with impaired renal function, aged T2DN animals could not concentrate urine. As a result of these diabetic complications, urine osmolarity in aged T2DN animals fell and reached values that were comparable to those in plasma (Fig. 4I).

Evaluation of podocyte and glomerular damage in T2DN rats.

Glomerular hypertrophy is linked with the development of sclerotic lesions in human diabetes. For a detailed evaluation of glomerular damage in T2DN rats, we performed a novel and robust application of convolutional neural nets for the localization of glomeruli in healthy and damaged trichrome-stained whole renal sections (9). Figure 5A shows representative qualitative illustrations (heatmaps) based on glomerular score damage within whole kidney sections from 48-wk-old GK and T2DN rats. The red color indicates areas around one or a number of localized glomeruli with high damage scores (range of 3–4), showing the presence of sclerotic lesions in T2DN cortexes compared with GK diabetic controls. Qualitative and statistical comparisons of glomerular injury in the young adult and aged groups are shown in Fig. 5, B–D (note the number of glomeruli count for each score, with up to 1,200 glomeruli/group analyzed). Cumulative probability distribution analysis of the glomerular injury score (Fig. 5D) revealed that there was a slight shift for young adult T2DN rats to have a higher glomerular damage. In aged rats, ~1% of the GK rat glomeruli had a score of 4, whereas 27% of T2DN rat glomeruli had the most severe damage score.

Fig. 5.

Glomerular damage in Wistar, Goto-Kakizaki (GK), and type 2 diabetic nephropathy (T2DN) rats. A: representative section of Masson’s trichrome-stained kidneys from >48-wk-old GK and T2DN rats. The glomerular damage map (pseudocolor) shown on the right was built from the Mason’s trichrome-stained image using the automatic localized algorithm. The red color on pseudocolor mapping refers to high damage levels, and low or no damage is shown in blue. B–D: glomerular injury score (0–4, where 0 = no damage) assessed by semiquantitative morphometric analysis for 12-wk-old (B) and >48-wk-old (C) Wistar, GK, and T2DN rats. Numbers of glomeruli per score are shown on the y-axes. The percentage of glomeruli within the selected score range defined from cumulative distribution is shown. D: cumulative probability distributions of the obtained glomerular scores in 12-wk-old (top) and >48-wk-old (bottom) Wistar, GK, and T2DN rats shown in B and C, respectively. The plot indicates the probability of a glomerulus having a score within the selected range. The example of how to estimate the glomerular damage percentage on cumulative function was shown for the T2DN (>48 wk) group (red arrows). A Kolmogorov–Smirnov test was used to identify significant differences between the groups [at the level P < 0.001, GK and T2DN (>48 wk) distributions were significantly different].

Altered expression in the key components of the glomerular filtration slit, such as podocin and nephrin, is widely used a marker to identify proteinuria and nephrotic syndrome in experimental models of diabetes and in various human proteinuric pathologies. Nephrinuria, or a high amount of soluble nephrin in the urine, shows positive correlation with severity of the podocyte injury. Western blot analysis revealed elevated levels of nephrin in urine samples of aged T2DN rats (Fig. 6A), in line with the temporal development of glomerular injuries described above. Immunofluorescent labeling of freshly isolated T2DN glomeruli for podocin further demonstrated the dramatic loss of this key protein in focal or segmental areas compared with healthy Wistar controls (Fig. 6B).

Fig. 6.

Podocyte damage in type 2 diabetic nephropathy (T2DN) rats. A: Western blot analysis of nephrin shedding, a marker of podocyte damage, analyzed from urine (24-h collection) of young and aged Wistar and T2DN rats. The summary graph shows changes in urinary nephrin normalized to creatinine levels (N ≥ 3, ANOVA, k = 6 groups). B: podocin expression in aged (>48-wk-old) Wistar and T2DN glomeruli. Representative images of freshly isolated glomeruli labeled with the podocyte marker podocin (green) are shown. Scale bars = 50 µm. The two-dimensional image is a single section of a z-stack image sequence showing surface podocin and cell nuclei (blue). The three-dimensional projection is a summary of fluorescence intensity from the entire z-stack sequence overlaid with the transmitted light (TL) image projection. The arrows indicate a focal region with almost no podocin in T2DN glomeruli.

To evaluate structural changes in glomerular morphology and podocyte foot processes, we applied the novel nonoptical microscopy approach of SICM. This approach allows for the evaluation of structural changes in cells of interest with a resolution close to electron microscopy. Here, we analyzed podocyte foot processes from freshly isolated glomeruli from aged T2DN and Dahl salt-sensitive (nondiabetic, low-salt) rats. Figure 7A shows three-dimensional topographical images (z-axis represented by changes in color) of a glomerular filtration barrier obtained by SICM. The super-resolution scans were performed in a preselected area of the glomeruli under different magnifications and revealed interdigitating foot processes that completely enwrap the glomerular capillaries and established tight control of glomerular filtration in healthy nondiabetic control animals. In contrast, topographical images of a glomerular slit diaphragm in isolated T2DN glomeruli (48 wk old) exhibited a complete absence of interdigitated foot processes on the selected region of the glomerular tuft. The effacement of the epithelial cell foot processes and loss of the normal barrier are commonly analyzed by electron microscopy techniques. Similarly, to characterize the three-dimensional architecture of the glomerular filtration barrier obtained by SICM, we built a z-axis line scan profile along with podocyte foot processes and calculated variations over the total length of line on the selected x,y interval (Fig. 7B). Increased variation in the z-axis is indicative of the changes in height between different foot processes; therefore, the decreased variation in T2DN glomeruli is a quantitative assessment of foot process effacement. The obtained topographical z-axis profile example and statistical summary of three-dimensional architecture of the glomerular filtration barrier revealed significant foot process loss leading to the observed nephrinuria and albuminuria in T2DN animals (Fig. 7C).

Fig. 7.

Assessment of the structural changes in podocyte foot processes by scanning ion conductance microscopy (SICM). A: three-dimensional topographic images of nondiabetic (Dahl salt-sensitive) rat and type 2 diabetic nephropathy (T2DN) rat (>48-wk-old) glomeruli obtained by SICM. The super-resolution scans were performed in a selected area of the glomeruli under different magnifications and reveal interdigitated foot processes that enwrap the glomerular capillaries. The scan of a glomerulus isolated from a T2DN rat completely lacks the interdigitated foot processes on the selected region of the glomerular tuft. B: example of quantification of the z-axis profile to evaluate form and variation of the secondary podocyte processes. Total variation (total length of an irregular arc segment curve through the z-axis line scan) represents the variability/density of podocyte foot processes obtained from SICM topography images and was defined by an integral of the first derivative of the z-axis curve on the selected interval. C: line scan z-axis profile examples for the three-dimensional topographical images shown in A. The summary graph revealed significant changes in total z-axes profile variations and indicated severe foot process losses in T2DN animals (N ≥ 3 glomeruli, n = 7 line scans, *P < 0.001, ANOVA).

DISCUSSION

The present study provides a detailed characterization of the spontaneous development of advanced diabetic nephropathy in the T2DN model. Here, using a comparison with Wistar and GK rats, we described changes in the RAAS during the development of diabetes and characterized the severity of glomerular damage in T2DN rats. Overall, we concluded that significant injury is typically present in up to 27% of glomeruli of adult T2DN kidneys. The observed glomeruli lesions include severe mesangial expansion, the occasional presence of nodular sclerosis, significant effacement of the epithelial cell foot processes, and loss of the filtration barrier integrity accompanied by a reduction in podocin expression and nephrinuria. According to the pathological classification of DN in human kidney biopsy specimens, a phenotype was observed in aged T2DN rats that was comparable to glomerular classes IIb or III (47). Additionally, the presence of tubular and vascular lesions, shown as tubular atrophy, interstitial fibrosis, and diffuse thickening of the glomerular capillary walls, was described. The present validation was carried out using a number of novel approaches and will provide strong background for the use of the T2DN model in diabetic research.

Patients with DKD often demonstrate insulin resistance, lipid abnormalities, albuminuria, cardiovascular complications, hypercoagulation, and other metabolic syndrome-related pathologies. Additional genetic stressors have been incorporated in rat models commonly used for DKD studies to promote the development of renal lesions observed in humans. Zucker diabetic fatty or Wistar fatty rats usually have hyperinsulinemia, elevated cholesterol, and renal abnormalities before the development of hyperglycemia (6, 22). T2DN rats develop most of the human complications listed above later in life, together or after the substantial increase in blood glucose levels. Significant renal abnormalities, including the characteristic histological changes in glomeruli and tubulointerstitial lesions, are clearly detectable in T2DN rats after 40 wk, similar to the Zucker diabetic fatty strain (17).

The progression of DKD in Otsuka Long-Evans Tokushima fatty rats, an established model for the investigation of advanced DN (18), has a similar phenotype to that observed in T2DN. Both strains are spontaneously diabetic rats with signs of early polyuria, decreased body weight, and abnormal glycosuria at some stage of the disease and renal hypertrophy. Some important characteristics of DKD in Otsuka Long-Evans Tokushima fatty and T2DN rats, as well as KK-Ay mice (48), are the presence of moderate to severe mesangial matrix expansion with mesangial cell proliferation, diffuse glomerulosclerosis, nodular glomerular lesion, and decreased podocyte numbers (19). For example, one of the most widely used models of type 2 diabetes, endothelial nitric oxide synthase^−/− db/db (14, 57), does not develop nodular mesangial sclerosis or severe tubulointerstitial fibrosis, which are features of advanced DN in humans (19).

Because the T2DN rat model replicates the normal progression and metabolic characteristics of human type 2 diabetes, it is also suitable for pharmacological screening. Williams et al. (56) used T2DN male rats for the chronic administration of combined therapy of selective angiotensin-converting enzyme and metalloprotease inhibitors to successfully reduce the degree of glomerulosclerosis and delay the progression of DKD. In another study (21), use of the SGLT2 inhibitor luseogliflozin alone or in combination with lisinopril to control hyperglycemia and GFR in T2DN rats resulted in the reduction of blood pressure, degree of glomerular injury, and tubular necrosis. As expected, control of hyperglycemia with insulin had no effect on the progression of renal disease in T2DN rats, thus indicating an insulin resistance syndrome, a hallmark of prediabetes and a type 2 diabetes pathological condition. T2DN rats were also used to test chronic endothelin A receptor blockade with atrasentan. It was shown that this treatment produces advantageous changes in renal hemodynamics that slow the progression of renal disease and also reduces renal histopathology in the absence of reducing arterial pressure and proteinuria (46). These results indicate that the T2DN model may be suitable for investigating pharmacological targets to treat and prevent advanced DN.

To further test the development of DKD and support the use of the T2DN model, we applied several novel technical approaches. First, we used a comprehensive analysis of RAAS components in diabetic models. The relationship between the RAAS and diabetes has been studied for many years, and the general assumption is that patients with DN have low levels of RAAS components in the blood, masking and perhaps reflecting an activated intrarenal renin system (34, 37). Here, we reported that main equilibrium components of the RAAS are lowered in the plasma of both T2DN and GK young rats compared with nondiabetic Wistar rats. The PRA test, which is routinely used in clinical practice to measure PRA enzyme responsible for the body's regulation of blood pressure, thirst, and urine output, showed significantly low levels in diabetic strains (Fig. 3D) similar to reported observations in patients with DN (34). Further use of the T2DN model and LC-MS/MS-based quantification of RAAS components for the comparison between the blood and renal interstitium, in sex differences, under the decline in GFR and high K⁺ intake, and during the different stages of DKD will provide important knowledge for assessing the therapeutic effectiveness of interrupting the renin system in DN.

The term “DKD” is characterized as the presence of renal insufficiency and albuminuria and increases the probability for a diagnosis of DN, which is determined in patients with diabetes via histological confirmation of renal involvement due to diabetes (11, 47). Processing of kidney biopsy samples requires highly qualified assessments to differentiate the presence of specific glomerular, vascular, and tubulointerstitial lesions to detect and classify DN. Here, we applied semiquantitative morphometric analysis on diabetic kidney histological slides to precisely describe glomerular injury and construct representative heatmaps, which allowed easy and quick evaluations of glomerular pathology. Nearly all T2DN rats exhibited diffuse global glomerulosclerosis with nodule formation and arteriolar hyalinosis, which is reflected by a significant shift in glomerular damage score compared with GK rats at 48 wk of age (Fig. 5C).

It is important to note that a statistical comparison of means from data that have a non-normal distribution, like a glomerular score distribution, is widely used in research papers and may provide incorrect results. We have circumvented this potential mathematical error by generating cumulative probability distributions to compare glomerular damage scores. Tubulointerstitial pathogenic processes, including tubular atrophy and interstitial inflammation, are also highly present in T2DN animals. Dramatic differences between GK and T2DN rat histological analyses support previous reports and indicate that GK rats are relatively resistant to the development of DN.

Our data indicate that the GK rat provides a suitable control for the T2DN rat since it exhibits diabetes without overt DN. Mean arterial pressure was not significantly different between T2DN and GK rats at 3, 6, or 18 mo (20). Therefore, the contribution of elevated blood pressure is minimal. However, multiple metabolic syndrome-related factors leading to diabetes dyslipidemia, glycosuria, and albuminuria are present in the T2DN strain (Fig. 4). Interestingly, specific sequence analysis of mitochondrial DNA showed that T2DN rats carry mitochondria from the Fawn hooded hypertensive rat, whereas the GK rat carries mitochondria from the Wistar rat (41). Impaired respiratory complex activity and energy depletion without mitochondrial uncoupling were identified in T2DN rats (44). Therefore, the occurrence of mitochondrial uncoupling as a mechanism for low ATP synthetic capacity and mitochondrial damage might play a critical role in the development of DKD and require further investigation.

The diabetic milieu has profound effects on glomerular health, affecting glomerular cell morphology, structures, and function. In the present study, we specifically addressed glomerular and podocyte injury during the development of DKD. Loss of key molecular components of podocyte foot processes and slit diaphragm structure, as podocin and nephrin, is considered a major contributor to DKD complications. The elevated urinary nephrin and loss of glomerular podocin staining in T2DN rats (Fig. 6A) were consistent with downregulation of nephrin expression and loss of the electron-dense structure of the slit diaphragm observed in patients with diabetes (5). Podocytes maintain the integrity of the glomerular filtration barrier, with a distinct three-dimensional structure of interdigitating protrusions that covers the capillaries.

This complicated architecture is difficult to visualize and quantify using conventional optical or scanning electron microscopy methods. Here, we applied a novel super-resolution SICM approach to visualize and quantify changes in podocyte foot process morphology during the development of DN in T2DN rats (Fig. 7). Our data indicated severe podocyte injury and foot process effacement in the T2DN strain, including several extreme cases with completely bare endothelial cells forming the wall of glomerular capillaries (Fig. 7A, bottom). Moreover, these three-dimensional topographical images allowed for the quantification of foot process size and density (both parameters represented as z-profile variation), which we believe in the future may be useful for research and clinical evaluations of podocyte pathology in biopsy samples.

In summary, the progression of renal disease in T2DN rats closely parallels that of human DKD. The combination of GK and T2DN stains represents a useful model to study the development of diabetes in resistant or advanced DN cases, respectively. Therefore, several novel approaches described here for analysis of the pathological process in glomerulus and podocyte will be useful for the understanding of the progression of DKD and renal injury in diabetes.

GRANTS

This work was supported by Department of Veteran Affairs Grant I01 BX004024 (to A. Staruschenko), National Institutes of Health Grants R35-HL-135749 (to A. Staruschenko), P01-HL-116264 (to A. Staruschenko and A. W. Cowley), DK-020595 (Pilot & Feasibility project to O. Palygin), R00-DK-105160 to (D. V. Ilatovskaya), and T32-HL-134643 (CVC A.O. Smith Fellowship to C. A. Klemens), a Michael Keelan, Jr., MD, CVC Research Foundation Grant (to O. Palygin), American Diabetes Association Grant 1-15-BS-172 (to A. Staruschenko), and an American Physiological Society Research Career Enhancement Award (to A. Staruschenko).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

O.P., J.L., and A.S. conceived and designed research; O.P., D.S., V.L., R.B., M.F., C.A.K., O.S., and D.V.I. performed experiments; O.P., D.S., V.L., R.B., M.F., C.A.K., O.S., J.D.B., and D.V.I. analyzed data; O.P., D.S., A.W., J.L., and A.S. interpreted results of experiments; O.P., D.S., R.B., M.F., D.V.I., and A.S. prepared figures; O.P., D.S., and A.S. drafted manuscript; O.P., D.S., V.L., R.B., M.F., C.A.K., O.S., J.D.B., A.W., J.L., D.V.I., and A.S. approved final version of manuscript.