High-calorie diet results in reversible obesity-related glomerulopathy in adult zebrafish regardless of dietary fat

Abstract

Obesity is a risk factor for the development of kidney disease. The role of diet in this association remains undetermined, in part due to practical limitations in studying nutrition in humans. In particular, the relative importance of calorie excess versus dietary macronutrient content is poorly understood. For example, it is unknown if calorie restriction modulates obesity-related kidney pathology. To study the effects of diet-induced obesity in a novel animal model, we treated zebrafish for 8 wk with diets varied in both calorie and fat content. Kidneys were evaluated by light and electron microscopy. We evaluated glomerular filtration barrier function using a dextran permeability assay. We assessed the effect of diet on podocyte sensitivity to injury using an inducible podocyte injury model. We then tested the effect of calorie restriction on the defects caused by diet-induced obesity. Fish fed a high-calorie diet developed glomerulomegaly (mean: 1,211 vs. 1,010 µm² in controls, P = 0.007), lower podocyte density, foot process effacement, glomerular basement membrane thickening, tubular enlargement (mean: 1,038 vs. 717 µm² in controls, P < 0.0001), and ectopic lipid deposition. Glomerular filtration barrier dysfunction and increased susceptibility to podocyte injury were observed with high-calorie feeding regardless of dietary fat content. These pathological changes resolved with 4 wk of calorie restriction. Our findings suggest that calorie excess rather than dietary fat drives obesity-related kidney dysfunction and that inadequate podocyte proliferation in response to glomerular enlargement may cause podocyte dysfunction. We also demonstrate the value of zebrafish as a novel model for studying diet in obesity-related kidney disease.

NEW & NOTEWORTHY Obesity is a risk factor for kidney disease. The role of diet in this association is difficult to study in humans. In this study, zebrafish fed a high-calorie diet, regardless of fat macronutrient composition, developed glomerulomegaly, foot process effacement, and filtration barrier dysfunction, recapitulating the changes seen in humans with obesity. Calorie restriction reversed the changes. This work suggests that macronutrient composition may be less important than total calories in the development of obesity-related kidney disease.

INTRODUCTION

Obesity is a rapidly growing global health threat affecting nearly 2 billion people and is an independent risk factor for chronic kidney disease (CKD) and its progression (1–5). The pathology known as obesity-related glomerulopathy (ORG) was first described in 1974 (6). ORG encompasses a constellation of clinical and pathological findings in patients with obesity, including the indolent onset of subnephrotic- or nephrotic-range proteinuria, glomerulomegaly, and mild foot process effacement (2). Although isolated ORG is uncommon, obesity accelerates kidney disease progression due to other causes, such as diabetes and IgA nephropathy, and increases the risk of acute kidney injury (3, 7–9).

Clarifying the role of diet in initiating and maintaining kidney injury is an important step toward understanding the pathogenic link between obesity and obesity-related kidney disease. Although diet indirectly impacts kidney function by inducing obesity and metabolic dysfunction (10), experimental evidence suggests that dietary macronutrient (protein, fat, and carbohydrate) composition also directly impacts kidney function. In animal models, a high-fat diet causes lipotoxicity in the kidney (11–14), and longitudinal cohort data have associated high dietary fat intake with incident albuminuria and estimated glomerular filtration rate decline (15). At the same time, high protein intake has for decades been experimentally demonstrated to cause hyperfiltration (16, 17) and is associated with a greater risk of incident CKD in humans (18). Ambiguity also surrounds the use of hypocaloric diets in patients with early CKD. Although weight loss reduces hyperfiltration and proteinuria (19, 20), concerns about malnutrition and higher mortality risk have traditionally dissuaded clinicians from recommending weight loss via calorie restriction as a means of slowing CKD progression. Thus, substantial uncertainty about the “optimal” dietary macronutrient profile and caloric intake level for people with or at risk for kidney disease remains. These critical knowledge gaps demand new approaches to evaluating the role of diet in kidney injury.

Animal models represent one such opportunity. Some animal models of ORG rely on the induction of metabolic dysfunction by irreversible means, such as genetic manipulation. Leptin-deficient ob/ob mice, a model of human insulin resistance and obesity, have been shown to develop proteinuria in early life, which progresses to glomerulomegaly and eventually to nodular glomerulosclerosis (21, 22). Interestingly, podocyte density is decreased in ob/ob mice compared with wild-type controls, although the mechanisms leading to this observation remain under investigation. Similarly, db/db leptin receptor-deficient mice develop features of early diabetic nephropathy, but do so over a much longer timeframe than ob/ob mice (23).

Fewer animal models have investigated the role of diet composition on the development of kidney disease (2). Diet-induced models of obesity have multiple potential advantages for the study of ORG. They most closely mimic the causes of obesity in patients, have the potential to be reversible, and facilitate the study of specific metabolic pathways. For example, the addition of 30% sucrose to the drinking water of Wistar rats resulted in increased abdominal adiposity, decreased creatinine clearance, and proteinuria. Although kidney histology was not examined, sucrose-fed rats had higher kidney free-fatty acid content, leading to lipotoxicity, increased reactive oxygen species generation, and decreased autophagy (24). Similarly, mice fed a high-fat diet develop pathology consistent with ORG, and this pathology is accompanied by changes in mitochondrial structure in multiple cell types including podocytes and proximal tubular cells (25). Using a small molecule to protect mitochondrial function abrogates these changes, suggesting roles for lipid metabolism and energy balance in the pathogenesis of ORG (26). Despite these important insights, rodent models remain limited by the fact that satiety limits ad libitum consumption of standard chow, necessitating energy-dense diets such as the high-fat and sucrose-enriched diets described earlier.

Zebrafish (Danio rerio) are an attractive nonrodent model for exploring the causes and consequences of ORG. The adult zebrafish kidney (mesonephros) comprises several hundred nephrons that are remarkably similar to those of higher-order animals, including having all glomerular components (27). Zebrafish exhibit genetic conservation, with 71% of human genes represented by a zebrafish ortholog (28). Zebrafish also have the advantages of facile genetic manipulation, high fecundity, amenability to drug screening, and ease of dietary manipulation. Only one study has investigated the effect of diet-induced obesity in fish (29). Medaka (Oryzias latipes) fed a high-fat diet develop obesity with kidney pathology consistent with ORG, including glomerulomegaly and capillary dilation.

We thus sought to answer three central questions: 1) does diet-induced obesity cause ORG in zebrafish; 2) if so, is high dietary fat necessary to cause ORG; and 3) does calorie restriction reverse ORG? Our major findings indicate that even a short period of high calorie feeding results in an ORG-like phenotype in zebrafish that is independent of dietary fat content. We also found that a high-calorie diet sensitized the fish to podocyte injury. Finally, we demonstrated that calorie restriction following high calorie overfeeding can reverse obesity-induced kidney defects. These results provide insights into the mechanisms of ORG and establish the zebrafish as a novel animal model for its study.

METHODS

Animal Husbandry and Morphometry

Adult (>90 days postfertilization) zebrafish (D. rerio) were maintained at 28.5°C on a 14:10-h light-dark cycle at a density of no more than 5 fish/L on a water circulation system. The AB wild-type strain was used for all experiments unless otherwise specified. Double transgenic lfabp::VDBP-GFP and pod::NTR-mCherry fish (30) were generated by Dr. Weibin Zhou and obtained from Dr. Erica Davis and Dr. Nico Katsanis (Duke University). All experimental procedures were reviewed and approved by the University of North Carolina Chapel Hill Animal Care and Use Committee. Animals were housed in an Association for Assessment and Accreditation of Laboratory Animal Care-accredited facility in compliance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals as detailed on protocols.io (dx.doi.org/10.17504/protocols.io.bg3jjykn). Fish were anesthetized with tricaine before weight and standard length (i.e., snout to caudal peduncle) were obtained. Body mass index was calculated as grams per centimeter squared (31). Kidney area was measured following fixation in 4% paraformaldehyde (32).

Feeding Regimens

Zebrafish were fed one of four diets for 8 wk (n = 12–15 fish/group). The normal-calorie, normal-fat diet consisted of a single feeding of Gemma Micro 300 (GM300, Skretting) daily, providing 59% protein and 14% fat by weight, at the manufacturer’s recommended amount (3% of the total fish weight per tank). The high-calorie diet was the same as the normal diet, but daily feeding amounts were 10 times the amount by weight. High-fat diets consisted of GM300 mixed with dried chicken egg yolk (Sigma-Aldrich) at a ratio of 1:9 GM300-egg yolk, providing 21% protein and 29.3% fat by weight. This high-fat feed was provided daily at an amount to provide either a similar caloric intake (termed the high-fat diet) or at 10 times this caloric amount (high-calorie, high-fat diet; Table 1). Fish were weighed every 4 wk, and feeding was adjusted to maintain caloric intake with respect to fish weight.

Table 1.

Composition of control and intervention diets

Diet	Calories, kcal/g fish wt	Fat, %	Protein, %
Control (normal diet) and maintenance diet	0.15	14	59
High-fat diet	0.11	29	21
High-calorie diet	1.5	14	59
High-calorie, high-fat diet	1.10	29	21
Calorie-restricted diet	0.05	14	59

Calorie Restriction

A separate series of zebrafish was fed either the normal diet or high-calorie diet for 8 wk. After 8 wk, normal diet-fed fish were maintained on the normal-diet, whereas high-calorie diet-fed fish were divided into the following three groups for 4 wk: continuation of high-calorie diet, reduction in calories to manufacturer recommendations (maintenance diet), or to a calorie restriction 1/3 of the maintenance diet (n = 6 fish/group).

Blood Glucose Measurement

Fasting blood glucose was measured after 8 wk of feeding, 24 h after the last meal in live, anesthetized fish using blood sampled from the dorsal aorta (33) using a FreeStyle Lite blood glucose monitor and test strips (Abbott) (34).

Histology and Electron Microscopy

Zebrafish were fixed and embedded, and 5-µm sections were stained with hematoxylin and eosin according to the standard protocol of the Histology Research Core of the University of North Carolina at Chapel Hill. A minimum of 60 glomeruli and 100 proximal tubules were imaged from each group, and cross-sectional areas were measured. Nucleated cells in each glomerular section were counted, and cell density was calculated as the number of hematoxylin-positive nuclei divided by the cross-sectional area of the glomerulus.

Transmission electron microscopy was carried out on zebrafish kidneys (n = 3 kidneys/group) following fixation in electron microscopy (EM) fixative (Electron Microscopy Sciences). Samples were stained with osmium tetroxide and counterstained with aqueous saturated uranyl acetate. Ultrathin sections (70 nm) were obtained using a Leica UC6 Ultratome and imaged using a JEM-1400 plus (JEOL) at 120 kV with image acquisition on a Gatan digital camera.

To evaluate foot process effacement, the ratio of total width of podocyte foot process cytoplasm to total glomerular basement membrane (GBM) width in each of five images per group was calculated. GBM thickness was determined as the distance between the foot process membrane and endothelial cell membrane at each of 15 points identified by overlaying each image with a grid.

Quantification of Glomerular Cell Density

To further assess changes in glomerular cell density, we used the dual transgenic fish line mentioned earlier, which includes a podocin promoter-driven mCherry-NTR marker labeling podocytes. Transgenic fish (n = 3 fish/group) were fed either the normal diet or high-calorie diet for 8 wk before euthanasia. The kidneys were fixed and stained with DAPI at 1:1,000. We created a z-stack of seven glomeruli from each fish using the Zeiss LSM700 laser scanning confocal and Zeiss Plan Apochromat ×63/1.4 oil objective and a slice thickness of 2 µm. We then counted the total number of nuclei and total number of podocytes. We calculated the mean Feret’s diameter of podocyte nuclei and all cell nuclei from each section. We used these measurements, along with section thickness, to calculate corrected total cell density and corrected podocyte density using previously derived equations (35, 36).

ImageJ (National Institutes of Health) was used for all image analyses. All image measurements were conducted by graders blinded to the identity of the specimen.

Dextran Permeability Assay

The zebrafish mesonephric filtration barrier is similar in form and function to the metanephric barrier in mammals. When injected, dextrans up to 70 kDa pass into the urinary space; larger dextrans are retained within the vasculature (37, 38).

To measure effective size exclusion by the glomerular filtration barrier, 10 µL of a 1:1 mixture of Texas red-conjugated 70-kDa dextran and FITC-conjugated 500-kDa dextran (10 mg/mL, deionized water, Invitrogen) were injected intraperitoneally into zebrafish from each group (39). We confirmed uptake from the peritoneum by measuring retinal fluorescence using wide-field microscopy 4 h after injection (data not shown). Fish were recovered for 24 h before euthanasia and fixation in 4% paraformaldehyde. The kidney was dissected and mounted in Aqua-Poly/Mount (Polysciences). Slides were imaged on a Zeiss LSM700 laser scanning confocal using the Zeiss Plan Apochromat ×20/0.8 objective and Zen software (Carl Zeiss Microscopy). We performed colocalization analysis using the Coloc 2 plugin for ImageJ, using manually selected regions of interest including tubules and Spearman’s rank correlation coefficient for colocalization.

Induction of Podocyte Injury

Calorie restriction has previously been shown to protect from acute kidney injury (40). To assess how calorie excess may predispose podocytes to injury, we used a previously developed double transgenic fish line expressing podocyte-specific bacterial nitroreductase (NTR) as well as a vitamin D-binding protein-green fluorescent protein (VDBP-GFP) construct (30). When exposed to metronidazole (MTZ), podocytes in this fish line reduce the drug to a cytotoxin leading to podocyte-specific cell death, which can be visualized as the 79.6-kDa construct is filtered into the tubules and endocytosed. Previous work has shown that a low dose of MTZ (2 mM) does not induce proximal tubule GFP uptake 72 h later (30).

In a separate experiment, after being fed with the aforementioned diet regimens for 8 wk, transgenic fish were exposed to 2 mM MTZ in 0.1% DMSO for 24 h. Fish were recovered for 72 h before euthanasia. Dissection and whole kidney imaging were performed as described earlier.

Statistical Methods

GraphPad Prism was used for statistical analyses. Data are expressed as means ± SD. Welch ANOVA was used for comparisons of continuous variables among more than two groups with Dunnett’s T3 multiple comparisons test used for post hoc comparison between groups. P values of <0.05 were considered statistically significant.

RESULTS

Effect of Diet on Body Morphometry and Blood Glucose

We assigned wild-type adult fish to one of the following four diets: normal diet; high-fat diet; high-calorie diet; or high-calorie, high-fat diet (see methods). After 8 wk, fish in the high-calorie (mean weight: 364 mg) and high-calorie, high-fat (mean weight: 506 mg) diet-fed groups were significantly heavier than normal diet-fed controls (mean weight: 269 mg, P = 0.008 for the high-calorie diet and 0.05 for the high-calorie, high-fat diet; Fig. 1A). High calorie feeding resulted in longer fish, but only high calorie with high fat feeding resulted in a higher body mass index compared with controls (mean body mass index: 0.065 g/cm² for the high-calorie, high-fat diet vs. 0.05 g/cm² for the normal diet, P = 0.03; Fig. 1, B and C). Regardless of caloric or fat content, there were no significant differences in 24-h fasting glucose between groups (Fig. 1D). The cross-sectional area of the kidney was greater in the high-calorie diet-fed group compared with the normal diet-fed group (14.7 vs. 10.4 mm², P = 0.01). High-fat, normal-calorie diet-fed fish had smaller kidneys compared with controls (7.2 mm², P = 0.01; Fig. 1, E–G). There was no difference in kidney area normalized to body weight between any experimental group and controls.

Figure 1.

High calorie feeding caused higher body mass without fasting hyperglycemia. A: zebrafish fed high-calorie diets (high-calorie diet and high-calorie, high-fat diet) were heavier than those fed a normal diet after 8 wk. B: standard length increased over the course of the experiment in all groups except for the normal diet-fed group and was significantly longer in the high-calorie diet-fed group (vs. the normal diet-fed group) at 8 wk. C: body mass index (BMI) was significantly higher in the high-calorie, high-fat diet-fed group than in the normal diet-fed group at 8 wk. D: fasting glucose was not significantly different between the diet-fed groups. E: kidney area as measured following fixation and dissection. High-calorie diet-fed fish had significantly larger kidneys, and high-fat diet-fed fish had significantly smaller kidneys. There was no difference in relative kidney size when kidney size was normalized to total fish weight. F: lateral views of zebrafish in each group revealed increased abdominal girth compared with head size in the high-calorie diet-fed and high-calorie, high-fat diet-fed groups. G: ventral view of the kidney at the time of dissection. The kidney is outlined in black, adherent to the dorsal abdominal wall with all other abdominal organs removed. Representative images are shown, and all images were taken from the same distance with equal magnification. Scale bars = 1 cm. n = 11–15 fish/group for A–G. Data are presented as means ± SD. Statistical tests were as follows: Welch ANOVA with Dunnett’s multiple comparisons test. *P < 0.05 and **P < 0.01. HC, high-calorie diet; HCHF, high-calorie, high-fat diet; HF, high-fat diet; ND, normal diet.

High-Calorie Diets Induce Glomerulomegaly and Foot Process Effacement

Fish fed high-calorie and high-calorie, high-fat diets developed marked glomerulomegaly compared with those fed the normal diet (mean glomerular area for the high-calorie diet: 1,211 µm², high-calorie, high-fat diet: 1,229 µm², and normal diet: 1,020 µm², P = 0.001; Fig. 2, A and B). Fish fed the high-fat diet without increased caloric content did not have glomerulomegaly (Fig. 2, A and B). The absolute number of cells per glomerulus was higher in high-fat, high-calorie, and high-calorie, high-fat diet-fed groups (36.6, 41.4, and 38.5 cells/glomerulus compared with 32.5 cells/glomerulus with the normal diet, P < 0.05 for all comparisons; Fig. 2C), but there were no differences between groups in the number of nucleated cells per area of glomerulus (data not shown).

Figure 2.

High-calorie diet led to glomerulomegaly and decreased podocyte density. A and B: hematoxylin and eosin-stained paraffin kidney sections revealed marked glomerulomegaly in zebrafish fed a high-calorie diet, without evidence of glomerular sclerosis. Kidneys from the high-calorie and high-calorie, high-fat diet-fed groups had significantly larger glomeruli than those in the normal diet-fed group. C: as evaluated by light microscopy, the high-calorie diet-fed group had a higher total cell number in each glomerulus compared with normal diet-fed controls. D: confocal microscopy of kidneys from transgenic fish also showed no difference in total cell density between fish fed a normal diet and fish fed a high-calorie diet. E: in contrast, glomeruli from fish fed a high-calorie diet had lower podocyte density compared with fish fed a normal diet. Scale bars = 25 µm. In B and C, at least 60 glomeruli from 3–4 fish/group were measured on three sections per fish. In D and E, n = 21 glomeruli per group. Data are presented as means ± SD. Statistical tests were as follows: Welch ANOVA with Dunnett’s multiple comparisons test. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. HC, high-calorie diet; HCHF, high-calorie, high-fat diet; HF, high-fat diet; ND, normal diet.

To further evaluate this finding, we quantified cell density and podocyte density within individual glomeruli using previously described methods (36). In accordance with findings by light microscopy, confocal microscopy revealed no difference in the density of total nuclei between transgenic fish fed a normal diet or fish fed a high-calorie diet (Fig. 2D). However, the density of podocytes was significantly lower in fish fed the high-calorie diet compared with those fed a normal diet (784 vs. 1,133 podocytes/10⁶ µm³ of glomerular volume, P < 0.0001; Fig. 2E).

Electron microscopy of kidneys from fish fed the high-calorie diet revealed foot process effacement (Fig. 3A, black arrows). This resulted in a significantly higher percentage of GBM length covered by foot process cytoplasm in the high-calorie diet-fed group and a trend toward greater foot process coverage in the high-calorie, high-fat diet-fed group (84.0% for the high-calorie diet vs. 74% for the normal diet, P = 0.006; Fig. 3B). Patients with obesity often have thicker GBMs (2, 41). Consistent with this, zebrafish exposed to either high-calorie diet had GBM thickening compared with normal diet-fed control fish (Fig. 3C).

Figure 3.

Electron microscopy reveals foot process effacement and glomerular basement membrane (GBM) thickening in the setting of high-calorie diets. A: transmission electron micrographs of representative glomeruli from each diet-fed group revealed foot process effacement in both groups fed higher-calorie diets. Arrows indicate areas of foot process effacement. B: the percentage of the GBM covered by intact podocyte foot processes increased in the high-calorie diet-fed group, suggesting loss of the slit diaphragm area. C: glomeruli from high-calorie and high-calorie, high-fat diet-fed fish had increased GBM thickness. Scale bars = 1 µm. In B, n = 5 electron micrographs from each group. In C, GBM thickness was measured at 15 randomly assigned points in each of five images. Data are presented as means ± SD. Statistical tests were as follows: Welch ANOVA with Dunnett’s multiple comparisons test. **P < 0.01 and ****P < 0.0001. HC, high-calorie diet; HCHF, high-calorie, high-fat diet; HF, high-fat diet; ND, normal diet.

High-Calorie and High-Fat Diets Alter Proximal Tubule Size and Cell Organization

Tubular changes in ORG include enlargement of tubular epithelial cells and accumulation of ectopic lipid-containing vesicles (2, 42). Fish fed a normal diet exhibited proximal tubules with an eosinophilic cytoplasm, regular cell spacing, and a well-defined brush border. However, high-fat and high-calorie feeding, alone or in combination, resulted in proximal tubules with a pale, foamy cytoplasm and irregular cell spacing, with diminished and irregular brush borders and widened lumens (Fig. 4A). Proximal tubules were significantly larger in high-calorie and high-fat diet-fed groups compared with normal diet-fed controls, with the effect most prominent in high-calorie diet-fed fish (high-calorie diet: 1,038 µm² and normal diet: 717 µm², P < 0.0001; Fig. 4B).

Figure 4.

High-calorie diets caused proximal tubule enlargement and ectopic lipid accumulation. A and B: hematoxylin and eosin-stained sections from fish fed high-fat and high-calorie diets (high-fat, high-calorie, and high-calorie, high-fat diets) revealed pale cytoplasm with enlargement of the proximal tubule and proximal tubular lumen. The cross-sectional area of the proximal tubules in the high-fat and high-calorie diet-fed fish were significantly larger than those fed a control diet. C: fish fed either high-calorie diet developed ectopic lipid accumulation and brush border abnormalities. Representative electron micrographs of the apical membrane revealed a reduced and abnormal brush border (arrowheads) and copious lipid-containing vesicles (arrows). Scale bars = 25 µm in A and 1 µm in C. In B, at least 100 tubular cross-sections were measured from each group. Data are presented as means ± SD. Statistical tests were as follows: Welch ANOVA with Dunnett’s multiple comparisons test. **P < 0.01 and ****P < 0.0001. HC, high-calorie diet; HCHF, high-calorie, high-fat diet; HF, high-fat diet; ND, normal diet.

Electron microscopy revealed differences in tubular morphology and organization. The apical border of proximal tubular epithelial cells in normal diet-fed fish demonstrated typical brush border morphology with small endocytic vesicles. In high-fat diet-fed proximal tubules, the brush borders were intact, but the cytoplasm had scattered lipid-containing vesicles. In both high-calorie and high-calorie, high-fat diet-fed groups, the brush border was depleted (Fig. 4C, arrowheads), the cytoplasm was filled with lipid-containing vesicles (Fig. 4C, arrows) that displaced other organelles to the basal portion of the cell, and mitochondrial morphology was abnormal (Fig. 5).

Figure 5.

High-calorie and high-calorie, high-fat diets led to ectopic lipid deposition and displacement of organelles. Transmission electron microscopy of proximal tubules from fish fed the normal-calorie diets (normal diet and high-fat diets) showed evenly spaced cells with elongated mitochondria distributed throughout the cell and few endocytic vesicles. In contrast, proximal tubule cells in the high-calorie diets (high-calorie and high-calorie, high-fat diets) had bulging apical membranes and cytoplasm filled with lipid droplets, displacing the nucleus and mitochondria to the basal portion of the cell. Scale bars = 5 µm. HC, high-calorie diet; HCHF, high-calorie, high-fat diet; HF, high-fat diet; ND, normal diet.

High-Calorie Diet Perturbs Glomerular Filtration Barrier Function

To assess the effects of diet on the glomerular filtration barrier, we injected zebrafish in each group with a mixture of Texas red-conjugated 70-kDa dextran and fluorescein-conjugated 500-kDa dextran. In unaffected glomeruli, 70-kDa dextran is filtered and accumulates in proximal tubule cells, whereas larger 500-kDa dextran remains intravascular. Indeed, 24 h after intraperitoneal injection, no 500-kDa dextran was seen in the proximal tubules of normal diet-fed fish, whereas the smaller 70-kDa signal was present (Fig. 6A). In high-calorie and high-calorie, high-fat diet-fed fish, however, proximal tubular cells revealed fluorescent signals from both dextrans, indicating filtration barrier dysfunction (Fig. 6A). Colocalization analysis confirmed the presence of the larger dextran in proximal tubular epithelial cells of high-calorie and high-calorie, high-fat diet-fed fish (Fig. 6A).

Figure 6.

High-calorie diets led to glomerular filtration barrier dysfunction and susceptibility to injury. A: high-calorie diet feeding resulted in filtration barrier dysfunction. In normal diet-fed fish, no signal was present from 500-kDa dextran uptake in proximal tubules, whereas 70-kDa dextran was filtered and reabsorbed. In high-calorie and high-calorie, high-fat diet-fed fish, the signal from the larger dextran colocalized with the smaller, indicating failure of size exclusion at the glomerulus. Colocalization was demonstrated by Spearman’s ranked correlation coefficient, with 1 = absolute colocalization and 0 = no colocalization. Scatterplots demonstrate the relative intensities of red (x-axis) versus green (y-axis) pixels. B: schematic representation of dietary preconditioning to metronidazole (MTZ) injury. Zebrafish were fed one of four diets for 8 wk prior to treatment with 2 mM MTZ for 24 h and were then allowed to recover for 72 h before euthanasia. C: proximal tubule green fluorescent protein (GFP) fluorescence (arrows) was present after low-dose MTZ injury in fish preconditioned with a high-calorie diet (high-calorie or high-calorie, high-fat diets) but not in those fed a normal diet (arrowheads indicate normal tubules). The high-fat diet-fed group is not shown, as effects were equivalent to normal diet. No fish in the normal diet- or high-fat diet-fed groups had observable GFP in the proximal tubules, while 20% and 40% of those in the high-calorie and high-calorie, high-fat diet-fed groups did, respectively. In C, 5 fish/group were evaluated. Data are presented as percentages of fish with GFP in the proximal tubules. Scale bars = 20 µm. HC, high-calorie diet; HCHF, high-calorie, high-fat diet; HF, high-fat diet; ND, normal diet.

High-Calorie Diet Predisposes Podocytes to Injury

Obesity can promote cytotoxic injury in mammalian podocytes (43, 44). To test this in zebrafish, we took advantage of dual transgenic fish expressing both a VDBP-GFP construct and podocyte-specific NTR; when exposed to MTZ, podocytes convert MTZ to a cytotoxin, leading to podocyte cell death (30). This disrupts the glomerular filtration barrier, resulting in leakage of VDBP-GFP into proximal tubules, where it is endocytosed and can be observed in affected nephrons.

Transgenic fish were preconditioned with each of four diets for 8 wk before 2 mM MTZ exposure (Fig. 6B). This dose of MTZ does not result in measurable filtration dysfunction after 72 h in normally fed fish (30). However, at 72-h post-MTZ treatment, high-calorie and high-calorie, high-fat diet-fed fish showed evidence of barrier dysfunction as manifested by GFP accumulation in the proximal tubules, whereas fish on a normal-calorie diet (normal diet or high-fat diet) did not (Fig. 6C). Transgenic fish not treated with MTZ did not have GFP accumulation in the tubules when fed either normal or high-calorie diets.

Calorie Restriction Ameliorates the Effects of High Calorie Feeding

We hypothesized that the effects of short-term overfeeding on the zebrafish kidney would be reversible with cessation of the high-calorie diet. Wild-type zebrafish were fed normal or high-calorie (normal fat content) diets for 8 wk. At that point, fish fed the high-calorie diet were randomly assigned to continue the high-calorie diet or to a diet of “maintenance” caloric intake or a calorie-restricted diet, consisting of 30% of normal calorie provision (Fig. 7A). No variation of dietary fat content was undertaken in this experiment. As in the prior series, high-calorie feeding caused significant weight gain compared with normal diet feeding, whereas 4 wk of decreased calorie intake at the maintenance or restriction level resulted in slight weight loss (Fig. 7B).

Figure 7.

Calorie restriction ameliorated the pathological and functional changes caused by high-calorie diets. A: schematic of overfeeding and calorie restriction. Zebrafish were fed a high-calorie diet for 8 wk and then either continued on the high-calorie diet or transitioned to maintenance calories or calorie restriction for 4 wk and compared with fish fed a control diet throughout the 12 wk. B: high-calorie diet induced significant weight gain. Maintenance and calorie restriction diets resulted in slight weight loss, whereas continued high calorie feeding resulted in continued weight gain. C: glomerulomegaly receded with 4 wk of either maintenance or calorie restriction diet, with glomerular size similar to that of control diet-fed fish. High-calorie diet-fed fish continued to exhibit glomerulomegaly. At least 70 glomeruli from 3–4 fish/group were measured on hematoxylin and eosin (H&E)-stained sections. D: proximal tubular enlargement also returned to normal after 4 wk of maintenance or calorie restriction diet. At least 100 proximal tubules from 3−4 fish/group were measured on H&E-stained slides. E: filtration barrier function was restored after 4 wk of either maintenance or calorie restriction diet. High-calorie diet feeding resulted in the presence of 500-kDa dextran in the proximal tubules and glomerulomegaly at 12 wk. Reducing calories to maintenance or calorie-restricted levels eliminated dysfunctional filtration and resulted in normalization of glomerular size. Scale bars = 20 µm. In B, n = 6 fish/group. In C, >60 glomeruli per group were measured per group. In D, >100 tubular cross sections per group were measured. Data are presented as means ± SD. Statistical tests were as follows: Welch ANOVA with Dunnett’s multiple comparisons test. *P < 0.05 and **P < 0.01. CR, calorie-restrict diet; HC, high-calorie diet; MTN, maintenance diet; ND, normal diet.

Reduced-calorie diets resulted in a return to normal glomerular and tubular size, with no significant difference observed between the two types of reduced-calorie diets (maintenance or calorie restriction) and normal diet (Fig. 7, C and D). Continued high calorie feeding maintained significant glomerulomegaly and cross-sectional tubular area compared with the normal diet (Fig. 7, C−E).

We also assessed filtration barrier function using the previously described dextran filtration assay (see Fig. 6A). In fish fed the high-calorie diet for 12 wk, both 70-kDa dextran (red) and 500-kDa dextran (green) were present in proximal tubular epithelial cells, suggesting dysfunction of the ability of the glomerular barrier to exclude the larger dextran from filtration. Only the smaller dextran was present in fish transitioned back to low-calorie diets, a pattern similar to that in fish fed a normal diet throughout the experiment (Fig. 7E).

DISCUSSION

Contributions of diet and obesity to the development and progression of CKD are difficult to study in humans as effects of lifestyle accrue slowly. Developing a clear understanding of how dietary factors modulate kidney pathology is critical to providing accurate, preventive dietary guidance to patients. Although feeding trials in humans are challenging, animal models can provide valuable insights under highly controlled conditions in short timeframes.

Our data demonstrate that short-term high-calorie feeding results in pathological changes in the zebrafish kidney, drawing a direct line from excess calorie intake to the development of ORG. The findings of glomerulomegaly, foot process effacement, and GBM thickening are consistent with the glomerular changes observed in biopsy samples from humans with ORG (45). Although the underlying cause of the glomerulomegaly remains to be determined, we observed greater cell numbers in the glomeruli of high-calorie diet-fed fish, similar to findings in mammalian models of ORG (46, 47), suggesting increased cell proliferation. In mammalian glomeruli, mesangial and endothelial cells are more capable of proliferation than podocytes, which results in podocyte stress and injury during glomerular enlargement (48, 49). Obese leptin-deficient rats develop glomerulomegaly and decreased podocyte density even before the onset of overt diabetes, and podocyte density is directly correlated with albuminuria (50, 51). In another model of hypertrophic stress induced by dominant negative expression of AA-4E-BP1 in rats, failure of the podocytes to respond to glomerular hypertrophy led directly to glomerulosclerosis and kidney failure (51).

In contrast, zebrafish podocytes are capable of regeneration and proliferation in response to acute insult (52). Obesity diminishes the regenerative capacity of other zebrafish tissues, but the effects of high calorie feeding on podocyte regeneration are unknown (53, 54). We found that fish fed a high-calorie diet have lower podocyte density, supporting inadequate podocyte proliferation in response to glomerular expansion (and the resulting demand for each podocyte to cover increased capillary surface area) as a mechanism potentially underlying filtration barrier dysfunction in overfed zebrafish. Why zebrafish podocytes fail to regenerate or expand in the face of high calorie feeding is not known, but lipotoxicity related to increased lipid endocytosis in podocytes may be one mechanism connecting diet to podocyte damage and death. Downregulation of the phosphatase and tensin homolog in mouse podocytes leads to increased endocytosis of lipoproteins and a phenotype consistent with ORG, and podocytes from patients with ORG also demonstrated enhanced lipid endocytosis (55). Investigating lipotoxicity and other mechanisms that may cause zebrafish podocytes not to regenerate will yield important insights into how podocytes maintain homeostasis under metabolic and mechanical stress.

The pathological glomerular changes we described have physiological relevance, both within the model and in the broader context of ORG. Although the degree of foot process effacement in humans is not always associated with proteinuria, ORG commonly results in proteinuria (56). Electron microscopy in the zebrafish model revealed segmental foot process effacement, and size-selective permeability was impaired in fish fed a high-calorie diet, as evidenced by the loss of 500-kDa dextran across the filtration barrier. These findings indicate a correlation between histopathological changes and function in this model.

In addition to changes in the glomeruli, abnormalities of the proximal tubules mimicked those observed in human ORG. Greater tubular diameter has been documented in patients with obesity and proteinuria. Both tubular epithelial cells and the tubular lumen are enlarged in ORG, putatively due to both increased urinary flow present and increased metabolic load on proximal tubule endothelial cells, along with the absorption of fatty acids (57). Electron microscopy of overfed fish revealed vacuolated proximal tubule endothelial cells, consistent with the ectopic lipid accumulation observed in other models of kidney lipotoxicity (42, 58). A previously established mouse model of high-fat diet-induced kidney dysfunction demonstrated that these vacuoles contain both neutral lipids (in the form of cholesterol esters) and phospholipids (58), likely resulting from dysfunction of the lysosomal system in the proximal tubule. Fish fed a high-calorie diet also had abnormal mitochondrial morphology in the proximal tubule. A recent transcriptomic study has suggested that the zebrafish proximal tubule relies on oxidative phosphorylation to meet its energy requirements, even from early in development (59). Both high-fat and high-calorie diets have been shown to disrupt mitochondrial oxidative phosphorylation, and this may be an important mechanism by which a high-calorie diet induces tubular injury and fibrosis (14, 60, 61).

The changes we observed in obese zebrafish occurred without regard to diet composition suggesting that excess calories, not fat, drive ORG in this system. Further investigation of the role of diet composition using our model, including variations in dietary protein and fat sources, will clarify this finding and could have important implications for dietary counseling in CKD. We also found that calorie restriction results in the resolution of both filtration barrier dysfunction and pathological changes to the glomeruli and tubules induced by a high-calorie diet. This finding harmonizes with the observation that weight loss via a hypocaloric diet results in improvements in both hyperfiltration and proteinuria in humans (62, 63). Our model is the first to show that these functional improvements are accompanied by the reversal of pathological changes in the kidney.

In addition to a higher risk of incident CKD, patients with obesity experience a higher risk of acute kidney injury than patients with normal weight. Preconditioning with high calorie feeding before subnephrotic insult predisposed zebrafish podocytes to cytotoxic injury, leading to filtration barrier dysfunction. This is in keeping with existing data showing that podocyte lesions are common in obesity and may predispose the kidney to injury (13, 64, 65).

Obesity leads to acute and chronic structural kidney changes, which underpin the development and progression of CKD. Our observations help elucidate the pathogenic mechanisms that interconnect diet, obesity, and pathological changes. We demonstrated maladaptive changes in both the glomerulus and proximal tubules as early waypoints in the course of ORG that can be induced with high calorie intake and reversed with calorie restriction. Future work should focus on how diet-induced obesity impacts nephron hyperplasia and podocyte stress in our ORG animal model. In the interim, our findings suggest that further investigation into calorie restriction in animal models and patients with early CKD is warranted and may uncover effective methods for ORG prevention and treatment.

GRANTS

This work was supported by University of North Carolina (UNC) at Chapel Hill Renal Epidemiology Training Grant 5T32DK007750-21 and UNC Kidney Center Endowment Funds. J.E.F. is supported by National Institutes of Health Grant R01HL152034.

DISCLOSURES

J.C.J. has received speaking honoraria from the Japanese Society of Nephrology and Korean Society of Nephrology; research funding from Alexion unrelated to this project; and paid consulting fees unrelated to this project from BioCryst Pharmaceuticals, ChemoCentryx, Protalix Biotherapeutics, and Sangamo Therapeutics. J.E.F. has received speaking honoraria from the American Society of Nephrology and multiple universities as well as investigator-initiated research funding unrelated to this project from the Renal Research Institute, a subsidiary of Fresenius Kidney Care, North America. She serves on the medical advisory board to NxStage Medical, now owned by Fresenius Kidney Care, North America, and a scientific advisory board and Data and Safety Monitoring Committee for the National Institute of Diabetes and Digestive and Kidney Diseases. She has received consulting fees from Fresenius Kidney Care, North America, and AstraZeneca. None of the other authors has any conflicts of interest, financial or otherwise, to disclose. The results presented in this paper have not been published previously in whole or part, except in abstract format.

AUTHOR CONTRIBUTIONS

E.M.Z., R.J.F., and J.S.P. conceived and designed research; E.M.Z. performed experiments; E.M.Z., J.C.J., and J.S.P. analyzed data; E.M.Z., J.C.J., J.E.F., R.J.F., and J.S.P. interpreted results of experiments; E.M.Z. and J.C.J. prepared figures; E.M.Z. drafted manuscript; E.M.Z., J.C.J., J.E.F., R.J.F., and J.S.P. edited and revised manuscript; E.M.Z., J.C.J., J.E.F., R.J.F., and J.S.P. approved final version of manuscript.