Dietary Interventions Improve Diabetic Kidney Disease, but not Peripheral Neuropathy, in a db/db Mouse Model of Type 2 Diabetes

Abstract

Patients with type 2 diabetes often develop the microvascular complications diabetic kidney disease (DKD) and diabetic peripheral neuropathy (DPN), which decrease quality of life and increase mortality. Unfortunately, treatment options for DKD and DPN are limited. Lifestyle interventions, such as changes to diet, have been proposed as non-pharmacological treatment options for preventing or improving DKD and DPN. However, there are no reported studies simultaneously evaluating the therapeutic efficacy of varying dietary interventions in a type 2 diabetes mouse model of both DKD and DPN. Therefore, we compared the efficacy of a 12-week regimen of three dietary interventions, low carbohydrate, caloric restriction, and alternate day fasting, for preventing complications in a db/db type 2 diabetes mouse model by performing metabolic, DKD, and DPN phenotyping. All three dietary interventions promoted weight loss, ameliorated glycemic status, and improved DKD, but did not impact percent fat mass and DPN. Multiple regression analysis identified a negative correlation between fat mass and motor nerve conduction velocity. Collectively, our data indicate that these three dietary interventions improved weight and glycemic status and alleviated DKD but not DPN. Moreover, diets that decrease fat mass may be a promising non-pharmacological approach to improve DPN in type 2 diabetes given the negative correlation between fat mass and motor nerve conduction velocity.

Graphical Abstract

Low carbohydrate, caloric restriction, and alternate day fasting dietary interventions promoted weight loss, ameliorated glycemic status, and improved diabetic kidney disease (DKD) in a db/db type 2 diabetes mouse model. In contrast, dietary interventions did not impact fat mass or diabetic peripheral neuropathy (DPN). Fat mass negatively correlated with motor nerve conduction velocity, suggesting that diets that decrease fat mass may be promising non-pharmacological approaches to improve DPN. Image created in BioRender.

INTRODUCTION

Diabetes is a global public health concern that impacts 537 million adults worldwide and is predicted to increase in parallel with the growing obesity epidemic.¹ Type 2 diabetes (T2D), which accounts for over 90% of cases, is characterized by hyperglycemia, dyslipidemia, and insulin resistance.^1,2 T2D is associated with common and severe complications that reduce patients’ quality of life and increase mortality. Of these complications, diabetic kidney disease (DKD) and diabetic peripheral neuropathy (DPN) are the most common.^3,4 DKD affects 25–40% of T2D patients and is characterized by increased albuminuria and a subsequent reduction in glomerular filtration rate.⁵ DKD is one of the leading causes of end-stage renal disease and a key risk factor for cardiovascular events.^4,6 DPN affects around 50% of individuals with T2D and is a chronic loss of the peripheral nerves that progresses in a distal-to-proximal fashion.^3,7 Loss of sensation in the extremities increases susceptibility to ulcer development and foot infection and can ultimately lead to lower-limb amputation.³ Unfortunately, despite decades of research, there are currently no effective treatment options for DKD or DPN.^8,9

Optimal glycemic control and blood pressure regulation can slow DKD progression.^10,11 However, optimal glucose control is difficult to achieve and DKD risk persists even in well-controlled patients.^12,13 For DPN, large intervention studies indicate that intensive glucose control in T2D patients has little effect on disease development and progression.^14,15 Together, these results suggest that available treatment options are not sufficient to mitigate DKD and DPN onset and progression.

Current recommendations from the American Diabetes Association encourage weight reduction through dietary interventions and exercise programs for the prevention and management of DKD and DPN in T2D.¹⁶ These guidelines are supported by clinical studies that examined the impact of dietary interventions, such as low carbohydrate Mediterranean (LC), caloric restriction (CR), and alternate day fasting (ADF), on DKD and DPN in T2D.^17–19 A LC diet limits simple sugars and starchy foods and exerts cytoprotective effects following renal and neuronal injury, potentially due to reduced chronic glycemia.^17,20,21 CR diet is characterized by decreased caloric intake and reverses insulin resistance, decreases DKD incidence, and ameliorates DPN in diabetic mouse models.^18,22,23 Intermittent energy restriction, which consists of alternate periods of restricted caloric intake or intervals of fasting, is gaining popularity as an alternative to CR and can be achieved by ADF.¹⁹ Evidence from rodent models suggests that intermittent energy restriction is renoprotective,²⁴ supports peripheral nerve health, and delays degeneration.²⁵ However, to our knowledge, there is no single study that characterized and compared the effect of all three of these dietary interventions on DKD and DPN in one T2D mouse model.

Mouse models of T2D that consistently develop DKD and DPN are invaluable for investigating disease pathogenesis.^26–28 The db/db mouse model, which has a leptin receptor mutation, develops obesity, hyperglycemia, and insulin resistance, and is considered one of the best characterized T2D mouse models.^26,27,29 These mice produce a robust DKD phenotype with albuminuria and histopathological changes, such as glomerular basement membrane thickening, podocyte loss, and moderate mesangial expansion.^27,29 Additionally, db/db mice consistently display DPN as evidenced by hypoalgesia, impaired motor and sensory nerve conduction velocities (NCVs), and decreased intraepidermal nerve fiber densities (IENFDs).^30,31

The goal of the current study was to evaluate and compare the effects of varying dietary interventions on DKD and DPN onset and progression in the db/db T2D mouse model. To do so, we compared metabolic, DKD, and DPN phenotypes between db/db mice fed a standard (SD), LC, CR, or ADF diet. We report that all three dietary interventions ameliorated body weight and glycemic status, but not body composition, compared to db/db mice fed SD. Additionally, all three dietary interventions improved the DKD, but not the DPN phenotype. Interestingly, percent fat mass negatively correlated with motor NCV. Overall, our results suggest that dietary interventions that improve glycemic status may suffice to prevent DKD onset and progression, but that ameliorating DPN may additionally require improved percent fat mass.

MATERIALS AND METHODS

Animals

Nine-week old male db/+ (non-diabetic control) or db/db (T2D) mice on a C57BLKS background (BKS.Cg-Dock^7m +/+ Lepr^db/J, Stock No: 000642; RRID:IMSR JAX:000642) were purchased from Jackson Laboratory (Bar Harbor, ME, USA) and allowed a 1-week acclimation period before the start of the study. Mice were housed in a pathogen-free environment (20±2 °C, 12:12 h light:dark cycle) and cared for by the Unit for Laboratory Animal Medicine at the University of Michigan. Terminal metabolic, DKD, and DPN phenotyping were performed after 12 weeks when mice were 22 weeks of age. Experimental protocols were approved by the University of Michigan’s Institutional Animal Care and Use Committee and were in compliance with guidelines by the Diabetes Complications Consortium (http://www.diacomp.org), US National Research Council’s Guide for the Care and Use of Laboratory Animals, the US Public Health Service’s Policy on Humane Care and Use of Laboratory Animals, and the Guide for the Care and Use of Laboratory Animals.

Dietary interventions

The study consisted of five experimental groups (n=9–10 mice/group; Fig 1). Nine-week-old mice were allowed a 1-week acclimation period before random cage assignment and dietary intervention. We found that on average, db/+ mice consumed 3 g chow/day while db/db mice consumed 6 g chow/day. In order to assess and compare the effects of different dietary interventions on DKD and DPN, experimental groups included: (1) db/+ mice fed ad libitum a SD (3 g chow/day; RD Open Source Diet D11112201; New Brunswick, NJ, USA); (2) db/db mice fed ad libitum a SD (6 g chow/day); (3) db/db mice fed a LC diet (1.5 g chow/day; RD Open Source Diet D19021107);⁹ (4) db/db mice fed a CR diet (1.5 g chow/day; RD Open Source Diet D11112201); and (5) db/db mice fed an ADF diet (3 g SD/1.5 g LC/day; RD Open Source Diet D11112201 and D19021107 every other day).³² Full dietary composition is included in Supplementary Table 1.

Figure 1. Study design and dietary paradigms.

A) Dietary intervention study design. Starting at 10 weeks, db/db mice were fed a standard (SD) ad libitum, low carbohydrate (LC), caloric restriction (CR), or alternate day fasting (ADF) diet (n = 9–10/group). After 12 weeks on diet, terminal metabolic, diabetic kidney disease (DKD), and diabetic peripheral neuropathy (DPN) phenotyping were performed. B) Dietary composition. SD consisted of the RD Open Source Diet D11112201. Db/+ mice fed SD ad libitum ate, on average, 3 g chow/day and db/db mice fed SD ad libitum ate, on average, 6 g chow/day. LC diet contained more calories from protein and fat. CR diet was restricted to 50% of the SD. Mice were fed 1.5 g LC or CR chow/day. ADF diet alternated SD (3 g chow/day) and LC diet (1.5 g chow/day) every other day.

Metabolic phenotyping

Body weight and fasting blood glucose (FBG) were measured at 22 weeks, the experimental endpoint. FBG was measured using an AlphaTrak Glucometer (Abbott Laboratories, Abbott Park, IL, USA). Terminal hemoglobin A1c (HbA1c) levels were quantified via ELISA (Mouse Hemoglobin A1C Assay Kit, cat #80310; CrystalChem, Elk Grove Village, IL, USA) as per the manufacturer’s instructions. The Mouse Metabolic Phenotyping Center (MMPC) at the University of Cincinnati Medical Center (Cincinnati, OH, USA) measured terminal serum lipid profiles. Plasma insulin levels and body compositions were assessed at study termination by the Michigan MMPC at the University of Michigan Animal Phenotyping Core (Ann Arbor, MI, USA). Plasma interleukin 6 (IL-6), tumor necrosis factor α (TNF-α), and monocyte chemoattractant protein-1 (MCP-1) were measured using customized mouse adipokine plex panels (#22171; Multiplexing Eve technologies, Calgary, AB Canada; discoveryassay.com). Body composition (fat mass, lean mass, and fluid content) was measured by nuclear magnetic resonance (NMR; EchoMRI, 4in1-900, Houston, TX, USA).

DKD phenotyping

DKD phenotyping was performed at 22 weeks of age according to criteria established by the Diabetes Complications Consortium (www.diacomp.org) and our published protocols.^33–35 Urine was collected for 24 h following a 3-day stay in metabolic cages (Hatteras Instruments, Cary, NC, USA) and analyzed to measure polyuria, creatinine, and albumin levels (Albuwell M and Companion Creatinine systems, Exocell, Philadelphia, PA, USA). The albumin-to-creatinine ratio (ACR), an early measure of impaired kidney filtration, was quantified using creatinine and albumin levels.⁵

For kidney histopathology, the left kidneys were harvested after systemic phosphate buffered saline perfusion. Kidneys were weighed and fixed overnight in 2% paraformaldehyde in phosphate buffered saline. Mesangial index and glomerular area, which reflect mesangial expansion and glomerular hypertrophy, were quantified.^33–36 Briefly, kidneys were embedded in paraffin, sectioned (3 μm), and stained with periodic acid-Schiff reagent. Fifteen glomerular tufts per animal were randomly selected for microscopic imaging and analyzed. Mesangial index was quantified by calculating the percentage of periodic acid-Schiff-positive area relative to the total glomerular area using MetaMorph (v.7.7.0.8, Molecular Devices, San Jose, CA, USA).

DPN phenotyping

DPN phenotyping was performed at 22 weeks according to criteria established by the Diabetes Complications Consortium (www.diacomp.org) and our published protocols.^33,37 Large nerve fiber function was assessed using sural sensory and sciatic motor NCVs. Mice were anesthetized with isoflurane and their body temperature was maintained at 34 °C using a heating lamp. Sural sensory NCVs were measured by inserting stainless steel needle electrodes (Natus Medical, Pleasanton, CA, USA) at the dorsum of the foot, and recording following an antidromic supramaximal electrode stimulation of the ankle. Sensory NCVs were then quantified by dividing the distance between the recording and stimulating electrodes by the sensory nerve action potential take-off latency. Sciatic motor NCVs were measured by recording at the dorsum of the foot and stimulating first at the ankle and then at the sciatic notch in an orthodromic supramaximal manner. Sciatic NCVs were calculated by dividing the distance between the two stimulation sites by the difference between the two onset latencies.

IENFDs were measured using published protocols.^38,39 Briefly, plantar surface hind paw footpads were removed, fixed in Zamboni (cat# 1459A, Newcomer Supply, Middleton, WI, USA), washed, embedded, and cryo-preserved for sectioning. For immunohistochemistry, 30 μm sections were labeled with PGP9.5 (cat# 14730–1-ap, RRID:AB_2210497; Proteintech, Rosemont, IL, USA) and three z-series images were taken using a confocal microscope (Leica SP5, 20×1.2 water-immersion objective, 1024×1024 pixel resolution; Leica Microsystems, Wetzlar, Germany). MetaMorph software (v.7.7.0.8, Molecular Devices, San Jose, CA, USA) determined the number of fibers per mm of epidermis.

Statistical analysis

Data analysis and statistical tests were performed using Prism v.9 (GraphPad, San Diego, CA, USA). Data points are expressed as the mean ± standard error of the mean. Data sets were compared using one-way ANOVA with a Tukey’s post-hoc for multiple comparisons. Multiple regression analysis was conducted using R Statistical Software.⁴⁰ Multiple regression analyses were performed by assessing the correlation between percent fat mass and DPN factors while adjusting for treatments (db/+ Ctrl, db/db, db/db LC, db/db CR, and db/db ADF). Significance was assigned when p < 0.05 and results represented as least square mean ± the standard error of the mean.

RESULTS

Dietary interventions improve glycemic status, but not body composition

To compare the effects of dietary intervention on the metabolic phenotype, and kidney and nerve functions in T2D, db/+ mice were placed on a SD and db/db T2D mice were placed on either a SD, LC, CR, or ADF diet. Terminal metabolic, DKD, and DPN phenotyping was performed at 22 weeks (Fig 1). T2D db/db mice fed a SD consistently developed a diabetic phenotype, with increased body weight, impaired glycemic status, and dyslipidemia (Fig 2A–C, E). Dietary interventions significantly reduced body weight relative to db/db mice (Fig 2A). FBG levels were significantly decreased following LC and CR interventions compared to db/db mice on SD but remained high with ADF intervention (Fig 2B). All three dietary intervention paradigms significantly decreased HbA1c levels (Fig 2C). Plasma insulin levels were higher in db/db mice on SD compared to db/+ mice, but not significantly (p = 0.1, Fig 2D); however, plasma insulin levels were higher in CR and ADF versus db/+ mice. The db/db mice on SD developed lipid abnormalities with significant increases in cholesterol levels compared to db/+ animals (Fig 2E). All three dietary interventions decreased cholesterol levels, which was only significant with LC diet (Fig 2E). While plasma triglycerides were not altered in db/db relative to db/+ mice, these were significantly elevated following an LC intervention versus controls (Fig 2F). We found no statistical differences in plasma cytokine (IL-6 and TNF-α) and chemokine (MCP-1) levels across experimental conditions (Supplemental Fig 1). In terms of body composition, db/db mice on a SD displayed a significant increase in percent fat mass and decrease in percent lean mass and fluid content compared to db/+ mice (Fig 2G). Interestingly, dietary interventions did not change the body composition pattern of db/db mice (Fig 2G). Together, these findings indicate that the db/db mouse model develops robust metabolic dysfunction, consistent with essential features of the human diabetic phenotype. Overall, dietary interventions improved body weight and hyperglycemia, but not body composition, in diabetic mice.

Figure 2. Dietary intervention corrects terminal metabolic phenotype in the db/db mouse model of type 2 diabetes.

A) Body weight, B) fasting blood glucose, C) HbA1c, D) insulin, E) cholesterol, F) triglycerides, and G) body composition were measured at 22 weeks in control (db/+), T2D (db/db), and T2D with dietary intervention (db/db low carbohydrate (LC), caloric restriction (CR), and alternate day fasting (ADF)) mice. n = 9–10 mice/group; *p < 0.05, **p < 0.01, ****p < 0.0001, one-way ANOVA followed by Tukey’s post hoc of db/+ versus db/db mice with or without interventions; ^##p < 0.01, ^###p < 0.001, ^####p < 0.0001, one-way ANOVA followed by Tukey’s post hoc of db/db versus db/db with interventions. Data are presented as the mean ± s.e.m.

Dietary interventions improve DKD

To investigate the impact of dietary intervention on DKD onset and progression, polyuria, urinary ACR, kidney weight, mesangial index, and glomerular area were measured at study termination. As anticipated, db/db mice on a SD displayed compromised renal function, with polyuria and increased urinary ACR relative to db/+ mice (Fig 3A, B). LC and CR dietary interventions normalized both polyuria and ACR (Fig 3A, B). An ADF regimen reduced polyuria compared to db/db mice (Fig 3A) but was unable to restore urinary ACR (Fig 3B). Additionally, db/db mice fed a SD developed kidney hypertrophy, as indicated by a significant increase in kidney weight, kidney weight to body weight ratio, and glomerular area compared to db/+ mice (Fig 3C, E), which was normalized following all dietary interventions (Fig 3C, E). However, the increase in mesangial index, a measure of mesangial expansion, observed in db/db mice on a SD was unaffected by dietary interventions (Fig 3D). Overall, dietary interventions improved DKD phenotypes in diabetic mice.

Figure 3. Dietary intervention corrects terminal diabetic kidney disease (DKD) phenotype in the db/db mouse model of type 2 diabetes (T2D).

A) Polyuria, B) albumin-to-creatinine ratio (ACR), C) left kidney weight (LKW), D) LKW/body weight (BW) ratio, E) mesangial index, and F) glomerular area were measured at 22 weeks in control (db/+), T2D (db/db), and T2D with dietary intervention (db/db low carbohydrate (LC), caloric restriction (CR), and alternate day fasting (ADF)) mice. n = 9–10 mice/group; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, one-way ANOVA followed by Tukey’s post hoc of db/+ versus db/db mice with or without interventions; ^#p < 0.05, ^##p < 0.01, ^###p < 0.001, ^####p < 0.0001, one-way ANOVA followed by Tukey’s post hoc of db/db versus db/db with interventions. Data are presented as the mean ± s.e.m.

Dietary interventions do not improve DPN

Next, we performed DPN phenotyping to determine if dietary interventions improve nerve function and reduce nerve loss. As anticipated, db/db mice fed a SD displayed significantly decreased sensory and motor NCVs by 22 weeks compared to db/+ mice on a SD (Fig 4A, B). This decrease in large nerve fiber function was not improved by any of the three dietary interventions (Fig 4A, B). In fact, db/db mice fed an ADF diet had significantly delayed sensory NCV compared to db/db mice on a SD (Fig 4A). We also measured changes in IENFDs in the footpad of the mouse hind paw to quantify small nerve fiber loss. IENFDs were significantly reduced in db/db mice fed a SD compared to db/+ mice on a SD (Fig 4C). Again, dietary interventions failed to improve IENFDs in db/db mice (Fig 4C). In summary, while LC, CR, and ADF dietary interventions improve glycemic status, they do not appear to improve DPN in T2D mice.

Figure 4. Dietary intervention does not correct terminal diabetic peripheral neuropathy (DPN) phenotype in the db/db mouse model of type 2 diabetes (T2D).

A) Sensory nerve conduction velocity (NCV), B) motor NCV, and C) intraepidermal nerve fiber densities (IENFDs) were measured at 22 weeks in control (db/+), T2D (db/db), and T2D with dietary intervention (db/db low carbohydrate (LC), caloric restriction (CR), and alternate day fasting (ADF)) mice. n = 9–10 mice/group; **p < 0.01, ***p < 0.001, ****p < 0.0001, one-way ANOVA followed by Tukey’s post hoc of db/+ versus db/db mice with or without interventions; ^##p < 0.01, one-way ANOVA followed by Tukey’s post hoc of db/db versus db/db with interventions. Data are presented as the mean ± s.e.m.

Motor nerve conduction function correlates with fat mass

Since dietary interventions did not improve body composition or DPN phenotype, we assessed whether DPN severity correlated with fat mass (Fig 5). Multiple regression analysis indicated that accounting for percent fat mass within experimental conditions had a significant association with motor NCV, but not sensory NCV or IENFD (Table 1). Specifically, increase percent fat mass correlated with decrease motor NCV (Fig 5A). These results suggest that preventing DPN development and progression may require a dietary intervention that improves percent fat mass.

Figure 5. Multiple regression analysis of diabetic peripheral neuropathy parameters with percent fat mass.

Relationship between (A) motor nerve conduction velocities (NCVs), (B) sensory NCV, and (C) intraepidermal nerve fiber densities (IENFDs) with percent body fat as determined by multiple regression analysis for control (db/+), T2D (db/db), and T2D with dietary intervention (db/db low carbohydrate (LC), caloric restriction (CR), and alternate day feeding (ADF)) mice. n = 9–10 mice/group for NCVs and n = 4–7 mice/group for IENFDs Shading represents the spread of the data.

Table 1.

Multiple regression analysis of diabetic peripheral neuropathy parameters with percent fat mass.

Parameter	Source (adjusted for group)	p-value
Motor NCV (m/s)	% Fat	0.017
Sensory NCV (m/s)	% Fat	0.26
IENFD (fibers/mm)	% Fat	0.97

IENFD, intraepidermal nerve fiber density; NCV, nerve conduction velocity.

DISCUSSION

DKD and DPN are common complications of T2D with limited treatment options. Multiple studies suggest that lifestyle interventions, including changes in diet, are promising non-pharmacological options to improve DKD and DPN.^9,41 This is the first study to evaluate the simultaneous effects of LC, CR, or ADF diets on DKD and DPN in the well-established db/db T2D mouse model. Mouse diets were designed to mimic the composition of human diets and metabolic, DKD, and DPN phenotyping were performed to assess complication severity. We report that all three dietary interventions significantly improved the metabolic profile of db/db mice by lowering body weight and hyperglycemia. These metabolic improvements were accompanied with reduced renal damage and improved measures of DKD. However, dietary interventions failed to improve body composition or DPN in db/db mice. Interestingly, we identified a negative correlation between percent body fat and motor NCV, suggesting that dietary interventions that improve body fat may be an effective treatment option for DPN.

In line with our current findings, both animal models^37,42 and patient studies^43–45 have previously shown that dietary interventions that reduce body weight also improve metabolic disorders, including T2D and its complications. Moreover, we found that all three dietary interventions ameliorated glycemic status, consistent with other reports in db/db mice^46,47 and in patients with T2D.^48,49 In this study, we also showed an improvement in circulating cholesterol levels, but not plasma insulin, triglycerides, or body composition following interventions. Indeed, plasma cholesterol was significantly reduced in the LC group, whereas only trending decreases were observed following CR and ADF. These findings are in line with previous clinical studies, showing a superior effect of LC in reducing cholesterol, compared to other dietary interventions, such as a lower-fat diet.^50–53 In the current study, dietary interventions failed to normalize percent body fat and lean mass, similar to previous reports in ob/ob and db/db mice.⁵⁴ Yet, these animal findings are conflicting with clinical studies, showing reduced fat mass following dietary interventions in adults with obesity and T2D.^55–57 Discrepancies between human data and the mouse data presented here may be related to leptin deficiency in ob/ob mice and leptin resistance in db/db animals. In fact, Sloan et al. showed that caloric restriction alone was not sufficient to alter body composition.⁵⁴ Instead, intraperitoneal injection of leptin for 1 week was necessary to reduce adipose mass in caloric-restricted ob/ob mice.⁵⁴

While plasma insulin levels were not significantly increased at 22 weeks of age in db/db mice fed a SD compared to db/+ animals, we did observe an increase in insulin levels following dietary intervention. We previously reported elevated plasma insulin levels in db/db mice at 16 weeks.⁵⁸ However, by 22 weeks these animals experience pancreatic failure and the inability to produce high levels of insulin.⁵⁹ These results agree with clinical and pre-clinical evidence that hyperinsulinemia associates with T2D in early- and mid-stage, but not after pancreatic failure in late-stage disease.^60–63 An increase in plasma insulin levels following dietary intervention was also observed by Nonaka et al. in a db/db late-stage diabetes mouse model following a CR diet, despite improved fasting blood glucose and glucose tolerance.⁴⁷ Increased blood insulin levels in late-stage disease may be due to reduced insulin degradation as levels of insulin degrading enzyme were significantly reduced in the liver following a 3-week CR diet.⁴⁷ A separate study with db/db mice found that β-cells and GLUT2 expression were preserved following 22 weeks on a carbohydrate-free diet but decreased after a standard chow or a high-fat diet.⁶⁴ Future studies assessing glucose tolerance, β-cell function and proliferation, as well as peripheral glucose uptake will highlight the effect of dietary interventions on insulin resistance and secretion in the db/db mouse.

Interventions that improve glycemic control in T2D are known to help manage DKD.^36,65 Here, reduced hyperglycemia and weight loss following dietary interventions was similarly associated with renoprotection, as indicated by improved functional measures (polyuria and ACR) and histopathological markers (left kidney weight to body weight ratio and glomerular area). An ADF diet, however, failed to reduce ACR in db/db mice. None of the dietary interventions restored mesangial index. Improved DKD phenotype following dietary interventions is in line with previous data from T2D rodent models.^66,67 However, failure of ADF to improve ACR conflicts with a study in which STZ-induced type 1 diabetic Sprague Dawley rats were provided access to food every day (ad libitum) or every other day (intermittent fasting). After 8 weeks, diabetic rats on an intermittent fasting diet displayed decreased plasma creatinine and increased plasma albumin compared to diabetic rats fed ad libitum.²⁴ Similarly, while dietary interventions failed to improve mesangial index in the current study, previous work in diabetic rat models showed that CR and LC diets prevent diabetes-induced glomerular expansion.^27,66 These discrepancies may be due to differences in rodent model, disease stage at initiation of dietary intervention,⁶⁶ and exposure to dietary interventions before or after diabetes induction.⁶⁸ In the clinic, a low calorie diet improves renal function in obese diabetic patients even before weight loss, suggesting a direct impact of diet on renal function.^69,70 Overall, our data suggest that dietary interventions rescue the main phenotypic markers of DKD. Our results also raise the possibility that dietary intervention at earlier disease states may result in better outcomes.

Here, we found that 12-week LC, CR, or ADF dietary interventions did not improve large- and small-fiber deficits in the db/db model. Of note, multiple regression analysis indicated that percent body fat negatively correlates with motor NCVs. Failure of dietary interventions to improve DPN is in disagreement with emerging pre-clinical and clinical research which suggests that dietary interventions, including CR and dietary reversal from a high-fat to a SD, are effective in reducing DPN severity in the presence or absence of T2D.^{18,37,43,71–74} These discrepancies may be due to the increased fat mass and elevated circulating triglyceride levels (specifically following an LC diet) that were retained in mice in the current study despite normalization of body weight and glycemia.⁵⁴ However, the observed association between percent body fat and motor NCVs is in line with human clinical data which show that fat mass and fat percent are higher in individuals with DPN than in those without DPN.⁷⁵ Additionally, both obesity and dyslipidemia are independent metabolic drivers of DPN.^76–79 Therefore, it appears that improvement of glycemic status and weight is not sufficient to improve the DPN phenotype. Instead, diets that improve percent fat mass may be more effective at improving nerve function than those tested in the current study.^80–82 Thus, additional investigation exploring the effect of dietary intervention and type of intervention on DPN is required.

Our study does have limitations. First, this study used the genetic db/db mouse model of T2D. As noted above, discrepancies between human and mouse data may be due to leptin resistance in db/db animals. Carefully designed studies in non-genetic diet-induced models could overcome the confounding effects of genetic manipulation and provide additional insight. Second, dietary interventions were not matched for macronutrient content, which may have influenced the findings. Third, failure to improve DPN may be due to the timing or type of intervention. As db/db animals display DPN symptoms as early as 6 weeks of age³¹, it is possible that DPN had already progressed past the point at which dietary intervention would have been beneficial. Dietary interventions that promote unsaturated healthy fats should also be explored in the db/db model. We have previously carried out lipidomics analysis in the sciatic nerve and kidneys of db/db mice and found that changes in lipid metabolism were mostly tissue specific.⁸³ We particularly identified nerve triglycerides and diglycerides as important drivers of DPN.⁸³ Assessing how interventions impact kidney and nerve lipidome might explain the differential effects and would therefore be a valuable avenue for future research.

In summary, this is the first comparative study of the effect of dietary interventions on DKD and DPN in a db/db T2D mouse model. We conclude that LC, CR, and ADF interventions improve body weight and glycemic status of the db/db T2D mouse. Furthermore, we report that dietary interventions improve both functional and histopathological aspects of DKD, suggesting that strict and specific dietary control could provide a non-pharmacologic treatment option for DKD. Interestingly, dietary interventions tested here did not improve body composition or DPN phenotype. While glycemic control alone was not sufficient to ameliorate DPN, the negative correlation between percent body fat and motor NCVs suggests that diets that improve fat profiles may improve nerve function in T2D. Finally, this model provides the basis for the mechanistic studies required to investigate the role of specific molecular pathways in the protective effects of dietary interventions.

Nonstandard Abbreviations:

ACR

albumin-to-creatinine ratio

ADF

alternate day fasting

CR

caloric restriction

DKD

diabetic kidney disease

DPN

diabetic peripheral neuropathy

HbA1c

hemoglobin A1c

IENFDs

intraepidermal nerve fiber densities

LC

low carbohydrate

NCVs

nerve conduction velocities

SD

standard diet

T2D

type 2 diabetes

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author.