Kidney tea was collected from Yunnan Province, China, and then identified by Prof. Binghong Fei. Kidney tea (600 g) was soaked in 6,000 mL of water for 1 h, followed by decoction for 30 min and filtration. This process was repeated twice, and the combined extract was concentrated to 1 g/mL using a rotary evaporator and then stored in a refrigerator at −20°C for subsequent experiments.

The phytochemical profiles of kidney tea were carried out with ultrahigh-performance liquid chromatography (UHPLC)-mass spectrometry (MS). Thermo UHPLC vanquish (Thermo Fisher Scientific) was used for chromatographic separation with an ACQUITY UPLC HSS T3 column (100 mm × 2.1 mm, 1.8 μm) at 35°C. Water containing 0.1% formic acid and acetonitrile solution containing 0.1% formic acid were served as mobile phases A and B, respectively. The gradient program of mobile phase B was as follows: 0–17.0 min, 5%–98%; 17.0–17.2 min, 98%–5%; 17.2–20.0 min, 5%. The injection volume was 2 μL with a flow rate of 0.3 mL/min.

Thermo Q-Exactive HFX (Thermo Fisher Scientific) was used for MS analysis, and the electrospray ionization (ESI) source was used for detection in positive and negative ion modes. The parameters of ESI source were set as follows: the spray voltage was 3.8 kV (+) and 3.0 kV (−), the source temperature was 320°C, and the range of MS scan was 90–1,300 m/z.

The animal experiment was approved by the Life Science Ethics Committee of the First Affiliated Hospital of Zhengzhou University (No. 2024-KY-0015-002). 30 male C57BLKS/J db/db and 6 male C57BLKS/J db/m mice (7-week-old) were obtained from Nanjing Junke Bioengineering Co., Ltd. (Nanjing, China). db/db mouse exhibits leptin receptor deficiency and often serves as a common animal model for spontaneous type 2 diabetes (Suriano et al., 2021). All mice were maintained in an SPF-grade environment with a temperature of 22°C ± 2°C, a relative humidity of 50% ± 5%, and a 12 h light/dark cycle. All animals had free access to standard feed and water with 1 week of acclimatization. Subsequently, 30 db/db mice were randomly allocated into five groups (n = 6), namely, the model group (Mod), the kidney tea low-dose group (KTL), the kidney tea medium-dose group (KTM), the kidney tea high-dose group (KTH), and the dapagliflozin group (DAPA). Moreover, 6 db/m mice were used as the control group (Con). Of these, the KTL, KTM and KTH groups were given 2, 4 and 8 g·kg⁻¹·d⁻¹ kidney tea extract, respectively. The DAPA group was given 1.3 mg·kg⁻¹·d⁻¹ dapagliflozin. After 1 week of adaptive feeding, the mice in each intervention group received the corresponding drug, and the mice in the Con and Mod groups received distilled water. All mice were gavaged with a volume of 10 mL·kg⁻¹ and administered once daily for 8 weeks.

The blood from the tail vein was collected after an 8-hour fast at 0, 2, 4, 6, and 8 weeks, and the fasting blood glucose (FBG) levels were measured by a glucometer (Bayer Healthcare LLC, NY, United States). The urine was collected at week 8 by metabolic cages. Ultimately, the mice were anaesthetized with pentobarbital sodium, the eyeballs were removed for blood collection, and serum was obtained by centrifugation. After the mice were executed, the kidneys and contents from the ileocecal junction were collected.

Serum tumor necrosis factor-α (TNF-α), interleukin-6 (IL-6), and IL-10 levels were detected with ELISA kits purchased from Jiangsu Enzyme Labeling Biotechnology Co., Ltd. (Yancheng, China). Blood urea nitrogen (BUN) and serum creatinine (Scr) levels were detected by an automatic biochemical analyzer (Rayto Life and Analytical Sciences Co., Ltd., Shenzhen, China). Urine N-acetyl-β-D-glucosaminidase (NAG), albumin, and neutrophil gelatinase-associated lipocalin (NGAL) levels were detected by assay kits purchased from Nanjing Jiancheng (Nanjing, China) to monitor renal injury.

The kidney tissues of appropriate size were fixed in 4% paraformaldehyde, and the fixed tissues were embedded in paraffin. The samples were sectioned and stained with hematoxylin and eosin (H&E), periodic acid-Schiff (PAS) and Masson, and subsequently observed under light microscope for renal histopathological changes.

Total microbial DNA was extracted from the contents of the ileocecal junction of mice by a Magnetic Soil And Stool DNA Kit (Tiangen Biochemical Technology Co., Ltd., Beijing, China). Amplification of 16S rDNA gene (V3-V4 regions) was performed with the specific primers 341F (5′-CCTACGGGNGGCWGCAG-3′) and 806R (5′-GGACTACHVGGGTATCTAAT-3′) with barcodes. After electrophoresis and purification, the purified amplicons were sequenced by Illumina NovaSeq 6000 platform (Illumina, San Diego, CA, United States).

The FASTQ format data were processed through QIIME 2 pipeline. Taxonomic assignment of amplicon sequence variants (ASVs) was performed based on the SILVA v138 16S database. The composition of species in the samples was analyzed based on the results of species annotation. Principal coordinates analysis (PCoA) was used to assess the beta diversity, and linear discriminant analysis effect size (LEfSe) analysis was performed to further screen significantly different species within different groups.

The serum processing method and instrument details in the untargeted metabolomics analysis are reported in our previous literature (Zhu et al., 2023). The mobile phase was a mixture of solvent A (acetonitrile) and solvent B (water containing 25 mM ammonium acetate and 25 mM ammonium hydroxide) at a flow rate of 0.5 mL/min. The gradient program was as follows: 0–0.5 min, 5% B; 0.5–7.0 min, 5%–35% B; 7.0–8.0 min, 35%–60% B; 8.0–9.0 min, 60% B; 9.0–9.1 min, 60%–5% B; 9.1–12.0 min, 5% B.

Orthogonal partial least squares discriminant analysis (OPLS-DA) and principal component analysis (PCA) were performed to evaluate the overall distribution trend among groups. Variable importance in the projection (VIP) value > 1 and p < 0.05 (Student’s t-test) were utilized to identify significantly differential metabolites, and Kyoto Encyclopedia of Genes and Genomes (KEGG) was used for pathway enrichment analysis.

A 1 mm³ volume of renal tissue was fixed overnight in 2.5% glutaraldehyde. After dehydration, embedding, curing and sectioning, ultrathin sections were stained with 2% uranyl acetate and lead citrate. Subsequently, cellular ultrastructural changes were observed under a transmission electron microscope (Hitachi HT7700, Tokyo, Japan).

The kidney tissues were homogenized in PBS and used for the determination of ferroptosis-related markers. The activity of superoxide dismutase (SOD) was measured by SOD activity assay kits (Elabscience Biotechnology Co., Ltd., Wuhan, China), and the levels of 4-hydroxynonenal (4-HNE) and malondialdehyde (MDA) were measured via MDA and 4-HNE assay kits purchased from Nanjing Jiancheng (Nanjing, China), respectively.

0Kidney tissues were washed with PBS, and TRIzol reagent was used to extract the total RNA. The above RNA was reverse transcribed by Supermo III M-MLV reverse transcriptase to obtain the corresponding cDNA. PCR amplification was carried out using 2 × Power Taq PCR MasterMix and corresponding primers. The amount of PCR products was calculated using the fluorescence intensity of SYBR Green I, and the results were processed using the 2^−ΔΔCT method. These reagents were purchased from Beijing BioTeke Biotechnology Co., Ltd. (Beijing, China). Primer sequences were shown in Supplementary Table S1.

Kidney tissues were added to RIPA buffer containing 1% PMSF, which were incubated at 4°C for 30 min. The proteins were obtained by centrifugation at low temperature. After separation, the protein samples were transferred to polyvinylidene difluoride membranes. Subsequently, the membranes were sealed with 5% skimmed milk and incubated with primary antibody overnight. After washing with TBST buffer solution, the membranes were incubated with secondary antibody. The primary antibodies were as follows: anti-GPX4 (Affinity Biosciences, Changzhou, China), anti-ACSL4 (Proteintech, Wuhan, China), anti-NCOA4 (Santa Cruz, Dallas, Texas, United States), anti-FTH1 (BOSTER Biological Technology Co., Ltd., Wuhan, China) and anti-β-actin (Proteintech, Wuhan, China).

GraphPad Prism 9.0 was used for statistical analysis. Data were presented as mean ± SD. One-way ANOVA was used to analyze the significance, followed by Dunnett’s test. p < 0.05 was considered statistical significant. Moreover, the potential relationships between gut microbiota, serum metabolites and host phenotypes were performed via Spearman correlation analysis.

Based peak intensity (BPI) chromatograms were shown in Supplementary Figure S1. Polyphenols are the main active ingredients of kidney tea (Chiang et al., 2023), and such compounds are considered to have antioxidant and nephroprotective effects (Jiang et al., 2012; Oršolić et al., 2021). Therefore, we identified the polyphenols in kidney tea. 12 polyphenols with relative percentages higher than 0.5% were identified by qualitative analysis (Table 1). Among them, the relative percentage of rosmarinic acid was the highest, accounting for 6.40%. The total content of 12 identified polyphenols in kidney tea reached 19.76%.

The animal experimental design was described in Supplementary Figure S2. First, we evaluated the therapeutic effects of kidney tea on DN mice via measuring FBG. Compared to normal mice, DN mice exhibited significant hyperglycemia, whereas kidney tea treatment reduced FBG levels to some extent (Figure 1A). Chronic inflammation is closely related to the pathogenesis of DN (Zhang et al., 2024). Therefore, the levels of the inflammatory factors IL-6, IL-10 and TNF-α were measured. The results found that TNF-α and IL-6 levels were significantly increased, while anti-inflammatory cytokine IL-10 levels were significantly decreased in DN mice (Figures 1B–D). High and medium doses of kidney tea could obviously reverse the levels of these cytokines. Moreover, renal injury-related markers including Scr, BUN and urinary albumin were measured. Ferroptosis has been reported to occur in renal tubular epithelial cells in diabetic nephropathy (Jin and Chen, 2022). Therefore, we also measured NGAL and NAG levels. The levels of Scr, BUN, urinary albumin, NGAL and NAG in the Mod group were significantly higher than those in the Con group, which could be significantly reduced by kidney tea treatment (Figures 1E–I).

Next, the effects of kidney tea on renal morphological changes in DN mice were determined by H&E, PAS and Masson staining (Figures 2A–C). By H&E staining, shrunken glomeruli, interstitial inflammatory infiltrate, and massive degeneration of tubular epithelium were observed in DN mice (Figure 2A). PAS and Masson staining showed thickened glomerular basement membrane and increased collagen fiber proteins in DN mice, respectively (Figures 2B,C). Notably, these pathological changes were ameliorated after kidney tea treatment. Taken together, these results suggested that kidney tea improved renal function and injury in DN mice.

Considering that DN mice treated with high-dose kidney tea showed the best renal protection, the KTH group was used for subsequent gut microbiota and metabolomics analysis. We first investigated the composition of the gut microbiota after kidney tea treatment by 16S rDNA gene sequencing. PCoA analysis found an obvious separation between the clusters of the Con, Mod and KTH groups (Figure 3A), indicating that kidney tea caused significant changes in the composition of the gut microbiota. At the phylum level, Bacteroidota, Desulfobacterota, Firmicutes, and Actinobacteriota were the dominant species among all of the groups (Figure 3B). Subsequently, we examined the structural alterations at the genus level (Figure 3C). The results revealed that the relative abundance of Muribaculaceae, Lachnoclostridium, Akkermansia, Corynebacterium and Prevotellaceae_UCG-001 was increased, while the relative abundance of Blautia, Alloprevotella and Lachnospiraceae_NK4A136_group was decreased in DN mice compared to normal mice (Figures 3D–K). Significantly, kidney tea administration could reverse the dysbiosis of these bacterial taxa to some extent.

Furthermore, LEfSe analysis with an LDA score ≥ 3 was used to identify the characteristic bacterial taxa of each group (Figure 3L). Our results indicated that the significantly different microbiota in the Con, Mod and KTH groups were 9, 15 and 36, respectively. Of these, f_Muribaculaceae, c_Verrucomicrobiae and g_Akkermansia were the dominant microbiota in the Mod group. Importantly, high-dose kidney tea could enrich the relative abundance of o_Bacteroidales, g_Bacteroides, f_Rikenellaceae, g_Rikenellaceae_RC9_gut_group, and f_Tannerellaceae.

In addition, we used untargeted serum metabolomics to investigate the effects of kidney tea on host metabolites. The PCA plots showed the distinct separation of each group under positive and negative ion modes (Figures 4A, B). Meanwhile, VIP value > 1 and p < 0.05 were used as screening criteria for significantly differential metabolites. A total of 105 differential metabolites were identified in the Mod group compared with the Con group, indicating that the metabolite profile changed significantly during the development of DN. Compared with the Mod group, kidney tea intervention downregulated 41 metabolites and upregulated 31 metabolites. The heatmap showed the 50 most differentiated serum metabolites, implying that kidney tea exerted a profound effect on host metabolism (Figure 4C). Therefore, we further investigated the function of the differential metabolites between the two groups with pathway enrichment analysis. The results found that kidney tea mainly participated in the regulation of metabolic pathways such as ferroptosis, arginine biosynthesis, and mammalian target of rapamycin (mTOR) signaling pathway (Figure 4D). Among them, 14 differential metabolites were from the top 11 metabolic pathways (Supplementary Table S2). Notably, the relative abundance of serotonin, deoxyinosine, saccharin and 1-palmitoyl-2-oleoyl-sn-glycerol was increased, while the relative abundance of isopentenyl pyrophosphate and N-alpha-acetyl-l-ornithine was decreased in DN mice compared with normal mice (Figures 4E–J). Kidney tea intervention reversed the alterations of serum metabolites caused by DN.

Based on the results of untargeted metabolomics, we further investigated the roles of kidney tea in the regulation of ferroptosis pathway. Electron microscope analysis revealed destruction of mitochondrial cristae in renal tissues of DN mice, while high and medium doses of kidney tea significantly attenuated mitochondrial damage (Figure 5A). SOD is a key intracellular antioxidant system, whereas MDA and 4-HNE are indicators of lipid peroxidation. Therefore, the relationship between ferroptosis and DN was further explored by measuring SOD, MDA and 4-HNE levels. The Mod group had lower activity of SOD as well as higher levels of MDA and 4-HNE than the Con group (Figures 5B–D). These findings suggested the presence of ferroptosis in the kidneys of DN mice. Of note, these indicators were ameliorated after kidney tea intervention. Next, we determined the expression of the ferroptosis-related proteins acyl-CoA synthetase long-chain 4 (ACSL4), glutathione peroxidase 4 (GPX4), ferritin heavy chain 1 (FTH1), and nuclear receptor coactivator 4 (NCOA4) in renal tissues. The mRNA expression of NCOA4 and ACSL4 was significantly increased in DN mice compared to normal mice, while kidney tea could decreased their expression (Figures 5E,G). The mRNA expression of FTH1 and GPX4 was significantly decreased in DN mice, which was reversed by kidney tea administration (Figures 5F,H). Western blot was further utilized to detect the expression of these proteins, and the results were consistent with qRT-PCR (Figures 5I–M). In conclusion, these data indicated that ferroptosis contributed to DN development, and that kidney tea played a therapeutic role in DN by improving ferroptosis in the kidneys.

Spearman correlation analysis was conducted to determine the potential relationships between differential bacterial genera, altered metabolites, and host phenotypes (Figure 6). Our results found that Corynebacterium and Akkermansia were positively correlated with serotonin but negatively correlated with isopentenyl pyrophosphate and N-alpha-acetyl-l-ornithine. Meanwhile, Bacteroides was significantly and negatively correlated with serotonin. In addition, Prevotellaceae_UCG-001, Corynebacterium, Akkermansia, and Muribaculaceae were positively correlated with renal injury indicators (Scr, BUN, NAG and NGAL), pro-inflammatory cytokines (IL-6 and TNF-α) and ferroptosis markers (MDA and 4-HNE), but negatively correlated with SOD and IL-10. On the contrary, Blautia and Bacteroides had an opposite relationships with these phenotypes. Altogether, these data supported that some key gut microbiota might be related to host metabolites and DN development.