Specific pathogen-free male C57BL/6J mice (6–8 weeks old, 17–21 g) were obtained from SPF Biotechnology Co., Ltd. (Suzhou, China; Permit No. SYXK [SU] 2021-0025) and maintained under standardized conditions (24 $\pm$ 2 °C, 12-h light/dark cycle) with ad libitum water access. Notably, the mice underwent a one-week acclimatisation period prior to the commencement of the study. All procedures were conducted in accordance with the Guidelines for the Care and Use of Animals established by the Chinese Animal Management Committee, as approved by the Animal Experiment Ethics Committee of the Affiliated Hospital of Nanjing University of Chinese Medicine (Ethics No. AEWC-20230531-309).

After a week of acclimatisation, mice were divided into two cohorts: the normal control (NC) group (n = 8, housed 4 per cage) received a standard diet, while the remaining 32 animals were subjected to a 12-week high-fat diet. Diabetes was subsequently induced via intraperitoneal administration of streptozotocin (S0130, Sigma-Aldrich, St. Louis, MO, USA; 40 mg/kg/day $\times$ 4 days). After the injections, the mice were continued on the high-fat diet. Mice with fasting blood glucose (FBG) levels $\ge$16.7 mmol/L at 10 weeks after streptozotocin injection were considered to have diabetes. Thereafter, diabetic mice were randomized into four groups (n = 8/group, housed 4 per cage): normal protein (NP, 20% protein), CR (30% restriction), low-protein (LP, 13% protein), and LPCR (13% protein + 30% restriction). All diets (Jiangsu Xietong Pharmaceutical Bioengineer Co., Ltd.) were formulated as detailed in Supplementary Table 1. Daily food intake was measured, and FBG and body weight were recorded weekly, with dietary intervention lasting for a total of five weeks (Fig. 1).

Fig. 1.

Dietary paradigms and study design. (A) Dietary composition (green: carbohydrate, brown: fat, blue: protein). Standard diet (STD) provided 70% of calories from carbohydrates, 10% from fat, and 20% from protein. High-fat diet (HFD) provided 60% of calories from fat, 20% from carbohydrates, and 20% from protein. Caloric restriction (CR) consisted of a 30% reduction in total calories from the STD (i.e., 70% of STD calories). Low-protein diet (LPD) provided 13% of calories from protein, 74% from carbohydrates, and 13% from fat. Low-protein calorie-restricted (LPCR) diet combined the LPD with a 30% caloric restriction (i.e., 70% of LPD calories). (B) Study design. Eight mice assigned to normal control (NC) group were maintained on a STD. A mouse model of diabetes was established by feeding the mice a 12-week HFD, followed by multiple intraperitoneal injections of streptozotocin (STZ), and an additional 10 weeks of HFD. The diabetic mice were randomized into four groups: a normal protein (NP) group fed STD, and three groups receiving dietary interventions of CR, low-protein (LP), and LPCR (n = 8 mice/group). After 5 weeks of dietary intervention, blood, urine, fecal samples, and kidney tissues were collected for relevant analyses.

After a five-week period of dietary intervention, random urine samples were obtained from the mice using individual metabolic cages, followed by the determination of urinary albumin (LA128107H, Nanjing Jin Yibai Biological Technology Co., Ltd, Nanjing, Jiangsu, China) and creatinine (C011-2-1, Nanjing Jiancheng Bioengineering Institute, Nanjing, Jiangsu, China) levels using enzyme-linked immunosorbent assay (ELISA). Retro-orbital blood samples were centrifuged (4 °C, 1000 g for 15 min) to extract serum. Serum creatinine, triglycerides (TG) (A110-1-1, Nanjing Jiancheng Bioengineering Institute, Nanjing, Jiangsu, China), total cholesterol (T-CHO) (A111-1-1, Nanjing Jiancheng Bioengineering Institute, Nanjing, Jiangsu, China), low-density lipoprotein cholesterol (LDL-C) (A113-1-1, Nanjing Jiancheng Bioengineering Institute, Nanjing, Jiangsu, China), high-density lipoprotein cholesterol (HDL-C) (A112-1-1, Nanjing Jiancheng Bioengineering Institute, Nanjing, Jiangsu, China), glutathione (GSH) (A006-2-1, Nanjing Jiancheng Bioengineering Institute, Nanjing, Jiangsu, China), superoxide dismutase (SOD) (A011-3-2, Nanjing Jiancheng Bioengineering Institute, Nanjing, Jiangsu, China), trimethylamine oxide (TMAO) (MM-1122M1, Elabscience Biotechnology Co., Ltd, Wuhan, Hubei, China), and tumor necrosis factor-$\alpha$ (TNF-$\alpha$) (H052-1-2, Nanjing Jiancheng Bioengineering Institute, Nanjing, Jiangsu, China) levels were determined using ELISA kits.

At the end of the experiment, mice were anaesthetised with pentobarbital sodium (30 mg/mL, 100 mg/kg, i.p.) (CAS: 57-33-0, Department of Pharmacy, Affiliated Hospital of Integrated Chinese and Western Medicine, Nanjing University of Chinese Medicine). All mice were humanely euthanized at the conclusion of the experiment using the method of cervical dislocation. Intestinal feces were collected, rapidly frozen in liquid nitrogen, and transported in dry ice to Majorbio Bio-Pharm Technology Co., Ltd. for gut microbiota analysis. Kidney samples were divided into two portions, either fixed in 4% paraformaldehyde (G1101, Servicebio, Wuhan, Hubei, China) for histopathological examination or stored at –80 °C for western blotting.

Paraffin-embedded renal tissue sections (4 µm thickness) were prepared and subjected to staining with haematoxylin and eosin, periodic acid-Schiff, and Masson’s trichrome for structural evaluation. For Masson’s trichrome staining, the extent of collagen fiber deposition (blue-stained areas) was quantitatively analyzed using ImageJ software (National Institutes of Health, Bethesda, MD, USA). For immunohistochemical assay, the sections were probed with rabbit polyclonal antibodies targeting fibronectin (Cat No.15613-1-AP, proteintech, Wuhan, Hubei, China) (1:200) and collagen I (Cat No.14695-1-AP, proteintech, Wuhan, Hubei, China) (1:200) overnight at 4 °C, followed by incubation with horseradish peroxidase-conjugated goat anti-rabbit secondary antibody (Cat No. SA00001-2, proteintech, Wuhan, Hubei, China) at room temperature for 30 min. The expression levels of fibronectin and collagen I were analysed quantitatively using ImageJ software.

Microbial genomic DNA was isolated from fecal samples using the PF Mag-Bind Stool DNA Kit (M5635-02, Omega Bio-tek, Norcross, GA, USA). DNA purity and concentration were assessed using a Nanodrop 2000 spectrophotometer (Thermo Scientific, Waltham, MA, USA). Qualified DNA samples were subjected to amplification of the 16S rRNA gene. PCR was performed with custom universal primers: forward primers 27F (5^′-AGRGTTYGATYMTGGCTCAG-3^′) and reverse primer 1492R (5^′-RGYTACCTTGTTACGACTT-3^′). The optimized thermal cycling protocol included initial denaturation at 95 °C for 180 s, followed by 27 cycles of denaturation (95 °C, 30 s), annealing at 60 °C (30 s), extension (72 °C, 45 s), and a final extension at 72 °C for 600 s. Subsequently, a PacBio library was constructed and sequenced. PacBio data were analysed using SMRTLink 11.0 (Pacific Biosciences, Menlo Park, CA, USA), resulting in at least three complete passes with 99% sequence accuracy for high-fidelity sequences. High-fidelity sequences were processed using UPARSE 11 (Drive5, Sonoma, CA, USA) to classify the operational taxonomic units at 97% similarity.

The Gut Microbiome Health Index (GMHI) was calculated to quantitatively assess the microbial community health potential, the formula was constructed based on the theoretical framework proposed by Gupta et al. [15]. Permutational multivariate analysis of variance (PERMANOVA; 999 permutations, Bray-Curtis distance) was performed to test the significance of differences in gut microbiota composition among the dietary intervention groups. Principal coordinate analysis (PCoA) was performed to derive key coordinates and display sample variations in a multidimensional space, and visualizations were generated using the R software (Version 3.3.1, R Development Core Team, Auckland, New Zealand).

Proteins were extracted frozen kidney tissues, mixed with western blotting loading buffer, separated using SDS-PAGE (P0012A, Beyotime, Shanghai, China). After electrophoretic separation, samples were transferred onto PVDF membranes. The membranes were blocked with 5% skim milk for 2 h at room temperature, followed by overnight incubation at 4 °C with primary antibodies targeting apoptosis-associated speck-like protein containing a CARD (ASC) (Cat No.67824, CST, Boston, MA, USA), NOD-like receptor family pyrin domain containing 3 (NLRP3) (Cat No.15101, CST, Boston, MA, USA), interleukin-1$\beta$ (IL-1$\beta$) (Cat No.31202, CST, Boston, MA, USA), and $\beta$-actin (Cat No.3700, CST, Boston, MA, USA) (all diluted at 1:3000). Subsequently, membranes were washed and incubated with species-matched horseradish peroxidase-conjugated secondary antibodies (anti-rabbit: Cat No.7074, CST, Boston, MA, USA; anti-mouse: Cat No.7076, CST, Boston, MA, USA) at 4 °C for 48 h. Protein bands were visualized using a chemiluminescent detection system, and quantitative densitometry was performed using the ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Data analysis was conducted using GraphPad Prism 8.0.2 (GraphPad Software, Inc., San Diego, CA, USA). Normally distributed variables are presented as mean $\pm$ SD. Normality was assessed using the Shapiro-Wilk test. For homogeneity of variances, either Bartlett’s or Brown-Forsythe test was employed depending on the normality test outcome. For parametric comparisons across multiple groups, one-way ANOVA followed by Tukey’s post-hoc test was applied. Nonparametric comparisons between the two groups were conducted using the Mann–Whitney U test, followed by false discovery rate correction for multiple comparisons. A p-value 0.05 was considered statistically significant.

Table 1 summarises the changes in body weight and FBG levels across the different groups before and after the dietary intervention. At the experimental endpoint, diabetic mice across the intervention groups exhibited markedly reduced body weights relative to the NC group (all p 0.01). Although the body weight in the LPCR group displayed a downward trend compared with the NP group, this difference was not statistically significant. All three dietary intervention groups experienced a reduction in FBG levels compared with that at baseline, whereas the NP group showed an increase. Notably, FBG levels were significantly lower in the LP and LPCR groups than in the NP group at the end of the intervention (both p 0.01).

After five weeks of dietary intervention, the NP group exhibited marked elevations in serum TG, T-CHO, and LDL-C (all p 0.01) alongside a reduction in HDL-C (p = 0.023) relative to the NC group. However, LP and LPCR diets significantly improved the levels of all four lipid parameters compared to the NP group (all p 0.01), with the LPCR intervention displaying superior efficacy in modulating LDL-C and HDL-C profiles (Fig. 2).

Fig. 2.

LPCR diet improves lipid metabolism in DKD mice. (A) Mean TG level, (B) mean HDL-C level, (C) mean T-CHO level, and (D) mean LDL-C level in each group after dietary intervention. Values are expressed as mean $\pm$ SD, n = 6. *p 0.05, **p 0.01 vs. NC group; ^#⁢#p 0.01 vs. NP group. LPCR, low-protein calorie-restricted; DKD, diabetic kidney disease; TG, triglycerides; HDL-C, high-density lipoprotein cholesterol; T-CHO, total cholesterol; LDL-C, low-density lipoprotein cholesterol; NC, normal control; NP, normal protein; CR, caloric restriction; LP, low-protein.

To examine the effects of LPCR on kidney function, we measured serum creatinine concentrations and urinary albumin-to-creatinine ratios after five weeks of dietary intervention. The NP group exhibited significantly elevated serum creatinine levels and urinary albumin-to-creatinine ratios relative to the NC group (both p 0.01). In contrast, the three dietary interventions showed substantial reductions in both serum creatinine levels and urinary albumin-to-creatinine ratios compared to NP group measurements (all p 0.01; Fig. 3).

Fig. 3.

LPCR diet improves renal function in DKD mice. (A) Mean serum creatinine level and (B) mean urinary albumin-to-creatinine ratio in each group. Values are expressed as mean $\pm$ SD, n = 6. **p 0.01 vs. NC group; ^#⁢#p 0.01 vs. NP group. LPCR, low-protein calorie-restricted; DKD, diabetic kidney disease; NC, normal control; NP, normal protein; CR, caloric restriction; LP, low-protein.

Haematoxylin and eosin staining revealed that mice in the NP group exhibited pathological changes characteristic of DKD, including irregular glomerular morphology, cytoplasmic vacuolation of renal tubular epithelial cells, and chronic inflammatory cell infiltration into the renal interstitium. However, the three dietary interventions ameliorated these pathological changes to varying degrees, with the LPCR group demonstrating the most significant effect (Fig. 4A).

Fig. 4.

LPCR diet alleviates partial pathological abnormalities of the kidneys in DKD mice. (A) Representative photomicrographs of Haematoxylin and eosin (HE)-stained kidney sections in each group after dietary intervention. Yellow arrows: irregular glomerular morphology; green arrow: cytoplasmic vacuolation in renal tubular epithelial cell, black arrows: chronic inflammatory cell infiltration in the renal interstitium. Scale bar: 100 µm; original magnification: 200$\times$. (B) Representative photomicrographs of Periodic acid-Schiff (PAS)-stained kidney sections in each group after dietary intervention. Black arrows: mesangial matrix widening in glomeruli; green arrows: renal tubular basement membrane thickening. Magnification = 200$\times$. Scale bar: 100 µm. (C) Representative photomicrographs of Masson’s trichrome-stained kidney sections in each group after dietary intervention. Blue-stained areas indicate collagen fiber deposition. Magnification = 200$\times$. Scale bar: 100 µm. (D) Quantitative analysis of the mean collagen fiber area in each group after dietary intervention. Values are expressed as mean $\pm$ SD, n = 3. **p 0.01 vs. NC group; ^#⁢#p 0.01 vs. NP group. LPCR, low-protein calorie-restricted; DKD, diabetic kidney disease; NC, normal control; NP, normal protein; CR, caloric restriction; LP, low-protein.

Periodic acid-Schiff staining indicated enlarged areas of glycogen deposition in the kidneys of mice in the NP group, along with widening of the mesangial matrix in the glomeruli and thickening of the tubular basement membrane. However, treatment with CR, LP, and LPCR diets alleviated these pathological alterations to varying degrees, with the LPCR group showing the most significant improvement (Fig. 4B).

Masson’s trichrome staining revealed marked collagen accumulation in renal tissues of NP group mice (p 0.01), indicating pronounced renal fibrosis. However, CR, LP, and LPCR dietary interventions significantly attenuated collagen deposition (all p 0.01; Fig. 4C,D). Additionally, immunohistochemical analysis revealed significantly elevated glomerular fibronectin (p 0.01) and renal interstitial collagen I (p 0.01) in the NP group versus NC group. In contrast, the LPCR intervention significantly suppressed the expression of these fibrosis-associated proteins (both p 0.01; Fig. 5).

Fig. 5.

LPCR diet alleviates renal fibrosis in DKD mice. (A,B) Representative immunohistochemical images of fibronectin (A) and collagen I (B) in each group after dietary intervention. Scale bar: 100 µm; original magnification: 200$\times$. (C,D) Quantitative analysis of the mean positive area of fibronectin (C) and collagen I (D) in each group after dietary intervention. Values are expressed as mean $\pm$ SD, n = 3. **p 0.01 vs. NC group; ^#⁢#p 0.01 vs. NP group. LPCR, low-protein calorie-restricted; DKD, diabetic kidney disease; NC, normal control; NP, normal protein; CR, caloric restriction; LP, low-protein.

$\beta$-diversity analysis was performed to evaluate gut microbiota compositional similarities across different groups. Notably, the NP group displayed a clear separation from the NC group, indicating significant alterations in microbial diversity. PCoA further revealed that the LPCR group clustered most closely to the NC group. PERMANOVA confirmed the statistically significant differences between dietary interventions (p 0.01; Fig. 6A).

Fig. 6.

LPCR diet modulates gut microbiota in DKD mice. (A) Principal coordinate analysis (PCoA) of each group. (B) Comparison of Gut Microbiome Health Index (GMHI) among different groups. (C,D) Mean Firmicutes abundance (C) and mean Firmicutes/Bacteroidetes ratio (D) in each group after dietary intervention. (E,F) Comparison of gut microbiota at the genus level (E) and species level (F) between NP and LPCR groups. (G) Mean serum trimethylamine oxide (TMAO) level in each group after dietary intervention. Values are expressed as mean $\pm$ SD, n = 6–8. *p 0.05, **p 0.01 vs. NC group; ^#p 0.05, ^#⁢#p 0.01 vs. NP group. LPCR, low-protein calorie-restricted; DKD, diabetic kidney disease; NC, normal control; NP, normal protein; CR, caloric restriction; LP, low-protein.

The GMHI, a metric quantifying health-associated microbial signatures via stool metagenomic profiling, showed significant impairment in the NP group compared to the NC group (p 0.01). Notably, the LPCR diet led to a significant improvement in GMHI compared with that in the NP group (p 0.01; Fig. 6B).

To delineate how dietary interventions counteract gut microbiota dysbiosis in DKD mice, we analysed the taxonomic shifts across the phylum, genus, and species levels. At the phylum level, Firmicutes abundance (p 0.01) and Firmicutes/Bacteroidetes (F/B) ratio (p = 0.017) were higher in the NP group than those in the NC group. Notably, treatment with the LPCR diet markedly decreased Firmicutes abundance (p 0.01) and F/B ratio (p = 0.02) in mice with DKD (Fig. 6C,D). Genus- and species-level analyses further demonstrated significantly increased abundance of Bacteroides (p 0.01) and Bacteroides acidifaciens (p 0.01), respectively, in the LPCR group versus NP group (Fig. 6E,F).

Given the critical role of TMAO, a gut microbiota-derived metabolite, in renal pathophysiology, we assessed dietary modulation of this uremic toxin. Serum TMAO levels were significantly elevated in the NP group relative to NC group (p 0.01). Conversely, CR, LP, and LPCR dietary interventions demonstrated significant reductions in serum TMAO levels relative to NP group measurements (p = 0.019, p 0.01, and p 0.01, respectively), with the LPCR diet exerting the most pronounced inhibitory effect (Fig. 6G).

Oxidative stress and systemic inflammation were identified as key mechanisms underlying DKD progression. Compared to the NC group, the NP group exhibited notably reduced serum GSH and SOD levels, along with a marked increase in TNF-$\alpha$ levels (all p 0.01). However, treatment with LPCR diet for five weeks reversed these trends, demonstrating a significant upregulation of GSH and SOD alongside suppression of TNF-$\alpha$ relative to the NP group (all p 0.01; Fig. 7A–C).

Fig. 7.

LPCR diet provides antioxidant and anti-inflammatory benefits in DKD mice. (A–C) Mean GSH level (A), mean SOD level (B), and mean TNF-$\alpha$ level (C) in each group after dietary intervention. (D) Representative Western blot images of NLRP3 inflammasome-related protein in the kidney tissue of DKD mice after dietary intervention. (E–G) Western blot densitometric analysis of ACS (E), NLRP3 (F), and IL-1$\beta$ (G) in each group after dietary intervention. Values are expressed as mean $\pm$ SD, n = 3–6. *p 0.05, **p 0.01 vs. NC group; ^#⁢#p 0.01 vs. NP group. LPCR, low-protein calorie-restricted; DKD, diabetic kidney disease; GSH, glutathione; SOD, superoxide dismutase; TNF-$\alpha$, tumor necrosis factor-$\alpha$; NLRP3, NOD-like receptor family pyrin domain containing 3; ASC, apoptosis-associated speck-like protein containing a CARD; IL-1$\beta$, interleukin-1$\beta$; NC, normal control; NP, normal protein; CR, caloric restriction; LP, low-protein.

Activation of NLRP3 inflammasomes served as a driving factor in immune-inflammatory dysregulation. ASC, NLRP3, and IL-1$\beta$ expression levels in kidney tissue were markedly higher in the NP group than in the NC group (all p 0.01). Conversely, dietary interventions with CR, LP, and LPCR markedly attenuated the expression of these inflammatory markers (all p 0.01), with the LPCR group showing the most pronounced decrease in NLRP3 expression (Fig. 7D–G).