Gut microbiota-derived indole-3-propionic acid alleviates diabetic kidney disease through its mitochondrial protective effect via reducing ubiquitination mediated-degradation of SIRT1 (original) (raw)

Graphical abstract

The proposed model for the beneficial effects of indole-3-propionic acid (IPA) on diabetic kidney disease (DKD) involves the mitigation of mitochondrial dysfunction in glomerular endothelial cells (GECs) through the following mechanism: IPA increases SIRT1 levels by inhibiting its phosphorylation-mediated ubiquitination degradation. Elevated SIRT1 levels subsequently promote the deacetylation and nuclear translocation of PGC-1α, leading to increased expression of SOD2 and mtTFA. These actions enhance mitochondrial biosynthesis, reduce oxidative stress, ultimately resulting in decreased proteinuria, and attenuated progression of DKD.

graphic file with name ga1.jpg

Keywords: Indole-3-propionic acid, SIRT1, Oxidative stress, Mitochondria, Diabetic kidney disease, Glomerular endothelial cells

Highlights

Abstract

Introduction

Gut microbes and their metabolites play crucial roles in the pathogenesis of diabetic kidney disease (DKD). However, which one and how specific gut-derived metabolites affect the progression of DKD remain largely unknown.

Objectives

This study aimed to investigate the potential roles of indole-3-propionic acid (IPA), a microbial metabolite of tryptophan, in DKD.

Methods

Metagenomic sequencing was performed to analyze the microbiome structure in DKD. Metabolomics screening and validation were conducted to identify characteristic metabolites associated with DKD. The protective effect of IPA on DKD glomerular endothelial cells (GECs) was assessed through in vivo and in vitro experiments. Further validation via western blot, immunoprecipitation, gene knockout, and site-directed mutation elucidated the mechanism of IPA on mitochondrial injury.

Results

Alterations in gut microbial community structure and dysregulated tryptophan metabolism were evident in DKD mice. Serum IPA levels were significantly reduced in DKD patients and correlated with fasting blood glucose, HbA1c, urine albumin-to-creatinine ratio (UACR), and estimated glomerular filtration rate (eGFR). IPA supplementation ameliorated albuminuria, bolstered the integrity of the glomerular filtration barrier, and mitigated mitochondrial impairments in GECs. Mechanistically, IPA hindered SIRT1 phosphorylation-mediated ubiquitin–proteasome degradation, restoring SIRT1′s role in promoting PGC-1α deacetylation and nuclear translocation, thereby upregulating genes associated with mitochondrial biosynthesis and antioxidant defense.

Conclusion

Our findings underscore the potential of the microbial metabolite IPA to attenuate DKD progression, offering novel insights and potential therapeutic strategies for its management.

Introduction

Diabetic kidney disease (DKD) affects around 40% of individuals with diabetes, making it the primary contributor to chronic kidney disease (CKD) [1]. It substantially elevates the risk of cardiovascular events, healthcare costs, and mortality [2], [3]. DKD poses a significant global health challenge. Despite its complexity, the pathogenesis of DKD remains poorly understood, and existing preventive and therapeutic approaches offer limited efficacy in reducing morbidity and mortality.

The gut microbiome, often considered the second genome of the human body [4], profoundly influences host physiology and is recognized as the body's largest endocrine organ. Microbiota-derived biologically active molecules directly target host cells and receptors or indirectly regulate human responses through interactions with other microorganisms and their metabolites [5]. Indole-3-propanoic acid (IPA), an exclusive metabolite generated by enteric bacteria from dietary tryptophan, is absorbed by the gut epithelium and released into the circulation system [6]. Recent epidemiological investigations have linked IPA to metabolic disorders. Circulating IPA levels were inversely correlated with type 2 diabetes mellitus (T2DM) [7], [8], [9] and the incidence of advanced atherosclerosis [10]. IPA significantly decreased in obesity [11], liver fibrosis [12], CKD [13], and in patients with a rapid decline in estimated glomerular filtration rate (eGFR) [13]. However, the molecular mechanism behind these collective phenomena remains largely elusive. Previous studies have shown that IPA possess, like melatonin, a high resonance stable heterocyclic aromatic ring structure, can efficiently scavenge free radicals without generating reactive and pro-oxidative intermediate compounds. So, IPA is considered as an ideal powerful antioxidant and can effectively promote the repair of mitochondrial structural and functional damage [16], [17]. In cardiomyocytes, IPA acts as a mitochondrial modulator, enhancing maximal mitochondrial respiration and improving myocardial contraction. In an ex vivo heart failure mouse model with abnormal mitochondrial energy metabolism, IPA improved cardiac function [18]. In diabetic animal models, exogenous IPA supplementation improved cognitive impairment and neuropathy by increasing mitochondrial DNA content, promoting mitochondrial biogenesis, and alleviating oxidative damage [19], [20]. Therefore, targeting mitochondrial dysfunction appears to be the focus of IPA in metabolic diseases.

The kidney is a highly metabolic organ, and its function is closely linked to mitochondrial homeostasis. Mitochondrial dysfunction, including impaired mitochondrial biogenesis, abnormal mitochondrial dynamics, and disorders in mitochondrial energy metabolism, contributes to the advancement and exacerbation of DKD [21]. Mitochondrial injury induces the overproduction of reactive oxygen species (ROS), leading to cellular oxidative damage [22], [23]. The increased ROS and oxidative stress, in turn, accelerate abnormalities in mitochondrial structure and function [24], [25]. These two aspects form a vicious circle, ultimately aggravates the progression of DKD [26]. Therefore, strategies that promote mitochondrial biogenesis, reduce oxidative stress in kidney cells, and target mitochondrial damage are of great significance in alleviating DKD.

However, the role of IPA in DKD remains uncertain. In this study, we first screened and analyzed the link between IPA and the pathogenesis of DKD in both patients and mouse models. We further explored the effect of IPA on structural and functional kidney impairments and investigated the molecular mechanisms underlying its renal protective effect. Our findings highlight the potential of the microbial metabolite IPA to attenuate DKD progression, providing new insights and potential therapeutic strategies for its management.

Materials and methods

Study population

Between March and June 2021, we conducted a cross-sectional study involving 28 patients with T2DM, including 14 with DKD, randomly selected from the Department of Endocrinology and Metabolism at the Affiliated Hospital of Southwest Medical University. Controls (n = 9) were age- and sex-matched participants undergoing physical examinations at the Health Examination Center during the same period. Inclusion criteria were as follows: (1) age 18–60 years, (2) T2DM diagnosis according to the 1999 World Health Organization (WHO) criteria [27], and (3) DKD diagnosis following the 2020 American Diabetes Association (ADA) guidelines [28]. Exclusion criteria included: (1) pregnant and lactating women, (2) other kidney diseases, (3) acute diabetes complications, (4) blood system diseases or infectious diseases affecting blood routine, (5) malignant tumors, (6) autoimmune diseases, (7) serious gastrointestinal diseases, and (8) use of nephrotoxic drugs, immunosuppressants, hormones, antibiotics, or probiotics in the past six months. All serum samples were processed within 2 h of fasting blood collection and frozen at −80°C after quenching by liquid nitrogen. For histological and immunohistochemical studies, three DKD patients (aged 56–65 years, two males and one female) from the Department of Pathology and three nondiabetic patients from the Department of Urology at the Affiliated Hospital of Southwest Medical University were recruited.

Animal experiments

In this study, we conducted three stages of animal experiments. Male C57BL/6J mice (5-weeks-old) were purchased from Teng Xin, China, and SIRT1 gene knockout mice were generated with the assistance of Cyagen, China. All mice were housed in a specific-pathogen-free environment with a 12-hour light–dark cycle, maintaining humidity and temperature at 55 ± 5% and 22 ± 1°C, respectively. The mice were fed either a standard diet (Keao Xieli, China) or a high-fat diet (HFD) (60% fat; TrophicDiet, China) for 26 weeks. In the 12th week, the HFD group received streptozotocin (STZ, Solarbio, China) injections (50 mg/kg, d1-d5, i.p.) to establish the DKD model. Within the following two weeks, mice with blood glucose >16.7 mmol/L were induced successfully, representing the establishment of the diabetes model [29], [30]. By the 26th week, a notable increase in the urine albumin-to-creatinine ratio (UACR), a key indicator for diagnosing, staging, and predicting DKD progression in clinical practice [31], confirmed the successful construction of the DKD model [29], [30].

During the first stage of the animal experiment (Fig. S1), 5 DKD mice and 6 controls were sacrificed at the 26th week, and serum samples were collected for untargeted metabolomics analysis. Fecal samples from 5 DKD mice and 4 controls (the fecal samples from the other two controls were insufficient) were collected for metagenomic sequencing. The second stage aimed to investigate the therapeutic effect of IPA (20 mg/kg/day; cat. 220027, Sigma, USA) on DKD mice. Four groups were established for a 12-week treatment: NC+vehicle (n = 7), NC+IPA (n = 7), DKD+vehicle (n = 8), and DKD+IPA (n = 8). The third stage primarily involved validating the effects of IPA in SIRT1 heterozygous mice (SIRT1 homozygotes lethal). Six groups were established: WT (n = 6), WT DKD (n = 6), WT DKD+IPA (n = 6), SIRT1+/- (n = 6), SIRT1+/- DKD (n = 6), SIRT1+/- DKD+IPA (n = 6).

Blood glucose content and body weights were monitored every 2 weeks. Serum and urine samples were collected before sacrificing the mice, and kidney tissues were harvested immediately afterward. Notably, serum was separated and subpackaged within 2 h. All serum, urine, and fecal samples, along with some renal cortices, were frozen at –80°C after quenching with liquid nitrogen. The remaining kidney tissues were fixed in 4% paraformaldehyde (Servicebio, China) for 24 h, dehydrated, embedded in paraffin wax, and cut into 4 μm sections mounted on positively charged slides for subsequent experiments.

Ethics statement

The animal study strictly adhered to the guidelines of the National Institutes of Health (NIH) and was approved by the Animal Ethics Committees of Southwest Medical University. The human study has been approved by the Ethics Committee of the Affiliated Hospital of Southwest Medical University (KY2021086) and was registered online (Clinical trial register no. ChiCTR2100048381). The human study adhered to the Declaration of Helsinki, and all participants provided written informed consent.

The isolation and identification of primary glomerular endothelial cells (GECs)

The separation procedures involved differential sieving to extract glomeruli and the magnetic Dynabeads method for selecting GECs, as previously described with some modifications [32], [33], [34]. (1) Reagents and materials: heparin (cat. H8060-1) and collagenase IV (cat. C8160) were purchased from Solarbio, China. Fetal bovine serum (FBS), endothelial cell medium (ECM) (cat. 1001) and 0.2% gelatin solution (cat. 0423) were acquired from Sciencell, USA. DMEM medium, Dynabeads™ sheep anti-rat IgG (cat.11035), magnet (DynaMag™-15, cat. 12301D) and magnetic grates (DynaMag™-Spin, cat. 12320D) were obtained from Invitrogen, USA. Platelet endothelial cell adhesion molecule-1 (PECAM, CD31) (ab7388) and IgG H&L (Alexa Fluor® 647) (ab150167) were purchased from Abcam, USA. HBSS (cat. SH30030.02) and benzonase nuclease (cat. E1014-5KU) were acquired from HyClone and Merck respectively. Other chemicals and regents used in this part were of analytical grade. (2) Male C57BL/6J mice (1-week-old, 5 ± 2 g) were heparinized (0.3–0.4 U/g heparin, i.p.) and then euthanized (1 g/kg pentobarbital sodium). After full lavage (perfusing with HBSS through the heart) of the blood circulation system, the kidneys were collected and their medulla were removed as completely as possible. The harvested renal cortices were minced into tiny particles and then digested (1 mg/mL collagenase IV, 7 U/mL heparin and 250 U/mL benzonase nuclease dissolved in HBSS) at 37°C for 10 min with gentle agitation. The digestion was terminated by adding equal volume of DMEM containing 20% FBS. Next, with the washing of cold HBSS, the digested mixture was sequentially passed through 188-μm, 100-μm and 50-μm stainless steel cell strainer. Finally, the suspension mixture retained on the top of 50-μm strainer were transferred into a new centrifuge tube and centrifuged at 200 g and 4°C for 5 min. After discarding supernatant, the highly-purified glomeruli were deposited at the bottom of the centrifuge tube. Notably, kidneys and mixture were kept at 4°C except for the digestion at 37°C . (3) The isolation and culture of GECs involved transferring the harvested glomerular mixture to dishes precoated with 0.2% gelatin solution and then culturing in ECM at 37°C and 5% CO2 for 5–7 days without shaking. The cells were collected when they grew to about 80%, and anti-CD31 antibody-conjugated Dynabeads were constructed following the manufacturer’s instructions to select CD31+ cells from them. Ultimately, the harvested CD31+ cells were transferred to dishes precoated with 0.2% gelatin solution and then cultured in ECM at 37°C and 5% CO2. (4) The identification of GECs (Fig. S2A) utilized immunofluorescence analysis. When the cell purity reached 80%, they were used for research purposes. All primary GECs used in this study were within four generations. (5) The treatment of GECs included exposing them to different concentrations of glucose (10, 20, 30, 40, 50, 60 mM) and IPA (0, 5, 50, 100, 500, 1000, 2500, 5000, 10,000 μM) respectively, and stimulating with 30 mM glucose for different times (0, 24, 48, 72, 96 h) to test cytotoxicity. The results of the CCK-8 toxicity test are shown in Fig. S2B. Finally, cells were pretreated with 100 μM and 500 μM IPA for 2 h respectively in the following experiments, followed by 30 mM glucose stimulation for an additional 48 h.

Metagenomic sequencing

Fecal samples were collected and stored at –80°C. DNA was extracted using the FastPure Stool DNA Isolation Kit (MJYH, Shanghai, China). The quality and concentration of the DNA were verified using a SynergyHTX microplate reader and a NanoDrop 2000 spectrophotometer. DNA libraries were then constructed and sequenced on a NovaSeq 6000 platform with a 150 bp paired-end sequencing strategy by Berry Genomics (Beijing, China). The raw sequencing data were quality-checked and filtered on the Majorbio Cloud Platform (https://www.majorbio.com), and assembled using MEGAHIT [35] (version 1.1.2, https://github.com/voutcn/megahit). Gene prediction was performed using Prodigal [36] (version2.6.3, https://github.com/hyattpd/Prodigal), and non-redundant gene catalogs were created with CD-HIT [37] (version 4.7, https://weizhongli-lab.org/cd-hit/). Taxonomic and functional annotations were conducted using DIAMOND [38] (version 2.0.13, https://ab.inf.uni-tuebingen.de/software/diamond/) against various databases including NCBI NR and KEGG.

Untargeted metabolomics analyses

Untargeted metabolomics was employed to elucidate the serum metabolic alterations associated with DKD. For sample preparation, 100 μL of serum was mixed with 400 μL of a cold methanol/acetonitrile solution (1:1, v/v) to precipitate proteins. After centrifugation at 14,000 g and 4°C for 15 min, the supernatant was collected, followed by vacuum freeze-drying. The dried samples were re-dissolved in 100 μL acetonitrile/water (1:1, v/v) solvent.

The analyses were performed using an ultra-high-performance liquid chromatography (UHPLC) system (1290 Infinity LC, Agilent Technologies) coupled to a quadrupole time-of-flight mass spectrometer (AB Sciex TripleTOF 6600) at Applied Protein Technology Co., Ltd (Shanghai, China). A 2.1 × 100 mm ACQUIY UPLC BEH 1.7 µm column (Waters, Ireland) was used for separation. Electrospray ionization (ESI) was conducted in both positive and negative modes. The gradient elution procedure was as follows: 85% B for 1 min, 65% B for 11 min, 40% B for 4 min, 85% B for 0.1 min, with a flow rate of 400 L per minute.

Raw MS data (wiff.scan files) were converted to MzXML files using ProteoWizard MSConvert before being loaded into the publicly accessible XCMS software. For isotope and adduct annotation, the R program CAMERA was utilized. Metabolite identification involved comparing the accuracy m/z value (<10 ppm) and MS/MS spectra with an in-house database established using available authentic standards.

The R package (ropls) was employed for pareto-scaled principal component analysis (PCA) and orthogonal partial least-squares discriminant analysis (OPLS-DA). Student’s _t_-test was conducted to investigate the significance of differences, and VIP-value > 1 and _P_-value < 0.05 were used to screen for significantly changed metabolites. Pearson correlation analysis was applied to determine associations between variables.

Analysis of tryptophan metabolites

Tryptophan metabolites in serum were determined by UHPLC-multiple reaction monitoring (MRM)-MS/MS platform at Biotree Ltd. (Shanghai, China). Metabolites were extracted using an extraction solution (acetonitrile: methanol = 1:1, containing isotopically-labeled internal standard mixture). Standard solutions were prepared. UHPLC separation was performed using an ExionLC System (SCIEX) equipped with a Waters Acquity UPLC HSS T3 column (100 × 2.1 mm, 1.8 μm, Waters). Mass spectrometry was conducted using a SCIEX 6500 QTRAP + triple quadrupole mass spectrometer equipped with an IonDrive Turbo V ESI interface. MRM data acquisition, and quantitative analysis of target compounds were carried out with SCIEX Analyst Work Station software (version 1.6.3) and SCIEX MultiQuant software (version 3.0.3).

Histopathological analysis

Hematoxylin and eosin (H&E) staining, periodic acid-schiff (PAS) glycogen staining, and Masson’s trichrome staining were employed to assess histological changes. Staining procedures followed standard methods in routine pathology. Finally, the stained sections were observed using an optical upright microscope (Leica, Germany).

Transmission Electron Microscopy (TEM)

To observe ultrastructural changes in the glomerulus and GECs mitochondria, renal cortical lamellas and GECs were pre-fixed in 3% glutaraldehyde for 12 h and then post-fixed in 3% osmium tetroxide for 90 min at 4°C. Subsequently, after dehydration in ascending concentrations of acetone and embedding in epoxy resin, the tissues were cut into ultrathin sections (50 nm) for staining (uranium acetate and lead citrate, 15–20 min at room temperature). The Jem-1400plus TEM was utilized for observing and photographing. ImageJ (National Institutes of Health, Bethesda, MD) was employed to measure and quantify the number and size of mitochondria, as previously described [39].

Measurement of UACR, ROS, malondialdehyde (MDA), mitochondrial membrane potential and ATP

The UACR of mice was measured and calculated following the manufacturer’s procedures outlined in the kit (cat. AD3429MO, Andygene, USA). ROS content was assessed using the DCFH-DA fluorescent probe according to the ROS assay kit (cat. S0033M-1, Beyotime, China) protocol. MDA levels were probed in accordance with the instructions of the MDA assay kit (cat. A003-1-2, JianCheng, Nanjing, China). Mitochondrial Membrane Potential was assessed by JC-1 staining per the JC-1 assay kit (cat. m8650, Solarbio, China) operating manual. The concentration of ATP was assessed using the ATP assay kit (cat. S0026, Beyotime, China).

Western blotting and immunoprecipitation

Western blotting was performed following standard procedures. Briefly, total proteins from the mice renal cortex and GECs were extracted using RIPA buffer (Beyotime, China). Nuclear proteins were extracted according to the instructions of the nucleoprotein extraction kit (Sangon Biotech, China). The protein concentration in cell lysates was measured using the Bicinchoninic Acid Kit (BCA, Beyotime, China) to calculate the loading amount. Protein samples were separated by 10% SDS-PAGE and transferred onto a PVDF membrane (Millipore, USA). The membranes were blocked with 5% BSA (Solarbio, China) for 2 h at room temperature and then incubated with primary antibodies overnight at 4°C. Subsequently, the membranes were washed with TBST buffer and incubated with secondary antibodies at room temperature for 1 h. The bands were visualized on an imaging system with an ECL luminescence reagent. Cytosolic proteins were normalized to Tubulin, β-actin, or Histone H3.

For immunoprecipitation of native proteins, the Dynabeads™ Protein G Immunoprecipitation kit (cat. 10007D, Invitrogen, USA) was used for subsequent analysis by western immunoblot. In brief, 3 µg primary antibody was incubated with 200 μL cell lysate (at least 1 mg/mL) overnight at 4°C with rotation to construct the Ab-Ag compound. Then, add 50 μL Dynabeads™ Protein G to construct Dynabeads-Ab-Ag compound and incubate with rotation for 3 h at 4°C. After elution and denaturation, the immunoprecipitated proteins were separated by SDS-PAGE and measured by western blot.

The antibodies used were as follows: SIRT1 (cat. 8469/ cat. 9475, CST, USA), phospho-SIRT1 (Ser47) (cat. 2314, CST, USA), PGC-1α (cat. 518025, Santa Cruz, USA), SOD2 (cat. 13141, CST, USA), mtTFA (cat. 166965, Santa Cruz, USA), ubiquitin (cat. 3936, CST, USA; cat. 154650, Absin, China), MDM2 (cat. 965, Santa Cruz, USA), Smurf2 (cat. 12024, CST, USA), CUL4 (cat. 377188, Santa Cruz, USA), COP1 (cat. 166799, Santa Cruz, USA), IgG (cat. 131368, Abcam, USA), Tubulin (cat. 3163, Bioworlde, USA), β-actin (cat. 3700, CST, USA), Histone H3 (cat. 3190, Bioworlde, USA), HRP-labeled goat anti-mouse/ rabbit IgG(H + L) (cat. 0216/cat. 0208, Beyotime, China), HRP-labeled mouse anti-rabbit IgG (Conformation Specific) (cat. 5127, CST, USA), and HRP-labeled rabbit anti-mouse IgG (Light Chain Specific) (cat. 58802, CST, USA).

Immunohistochemistry and immunofluorescence staining

For paraffin sections, samples underwent dewaxing and rehydration, followed by antigen retrieval and blocking with 10% serum (cat. 16210072, Gibco, USA). Regarding cell samples, GECs were cultured and treated in six-well chamber slides. The cells were fixed with 4% paraformaldehyde, permeabilized with 0.2% Triton X-100, and subsequently blocked with 10% serum. Sections or cells were then incubated with primary antibodies, including SIRT1, p-SIRT1, PGC-1α, SOD2, mtTFA, and CD31 (cat. 281583, Abcam, USA) at 4°C overnight. Subsequently, they underwent sequential incubation with corresponding CY3- or FITC-conjugated secondary antibodies and 4,6-Diamidino-2-phenylindole (DAPI, Abcam, UK) for immunofluorescence staining, or with streptavidin-horseradish peroxidase complex, diaminobenzidine, and hematoxylin for immunohistochemistry. Finally, a fluorescence microscope (Leica, Germany) was used for scanning and photographing.

Plasmid and transfection

The expression plasmid for mouse SIRT1 and the site-directed mutagenesis plasmid for mouse SIRT1 were procured from HANBIO, China. In brief, the SIRT1 Ala-46 mutant was generated by substituting Ser-46 with alanine, and the SIRT1 Asp-46 mutant was created by substituting Ser-46 with aspartate. Transient transfection was performed using LipoFiterTM3.0 Liposomal Transfection Reagent (cat. HB-TRLF3-1000, HANBIO, China), following the manufacturer’s instructions.

Real-time quantitative reverse transcription PCR (RT-qPCR)

The RT-qPCR analysis was conducted following previously described standard methods [40]. Briefly, total RNAs from renal cortex or cell samples were extracted using TRIzol reagent (Invitrogen, USA). The concentration and purity of RNA were assessed with a spectrophotometer (OD260/280 and OD260/230 ratios). Reverse transcription of RNAs to cDNA was performed using the ReverTra Ace qPCR RT Master Mix (FSQ-201, TOYOBO). The resulting cDNA was subjected to RT-qPCR with gene-specific primers, in the presence of the Applied Biosystems™ SYBR™ Green PCR kit (Thermo Fisher Scientific, USA), utilizing the Analytikjena qTOWER 3G real-time PCR system (JENA, Germany). β-actin served as the endogenous control, and all samples were analyzed in triplicates. The relative quantity of mRNA was expressed as 2^-ΔΔCT. Primers were synthesized by Sangon Biotech, Shanghai, China, with the following sequences used in this study: GCTGACGACTTCGACGACG (Forward) and TCGGTCAACAGGAGGTTGTCT (Reverse) for mouse SIRT1; TATGGAGTGACATAGAGTGTGCT (Forward) and CCACTTCAATCCACCCAGAAAG (Reverse) for mouse PGC-1α; CAGACCTGCCTTACGACTATGG (Forward) and CTCGGTGGCGTTGAGATTGTT (Reverse) for mouse SOD2; ATTCCGAAGTGTTTTTCCAGCA (Forward) and TCTGAAAGTTTTGCATCTGGGT (Reverse) for mouse mtTFA; ACCTCTATGCCAACACAGTG (Forward) and GGACTCATCGTACTCCTGCT (Reverse) for mouse β-actin.

Statistical analyses

For normally distributed values, the data are presented as means ± standard deviation (SD), while categorical variables are expressed as ratios. The normality of continuous variables was assessed using Shapiro-Wilk’s test. Statistical analysis included unpaired two-tailed Student’s _t_-test (for two groups) or one-way ANOVA (for more than two groups) for normally distributed values, Kruskal–Wallis H test for nonparametric values, and the chi-squared test for categorical variables. Pearson correlation analysis was employed to examine associations between two variables. Statistical analyses were performed using IBM SPSS Statistics 26 software. Statistical significance was defined as *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Results

Altered gut microbial community structure and metabolic pathways in DKD mice

Metagenomic sequencing was performed to elucidate the microbial community structure in DKD mice and control mice. Detailed information on the relative abundance of detected gut microbial taxa is provided in Table S1. PCA analysis, principal coordinates analysis (PCoA), and non-metric multidimensional scaling (NMDS) of gut microbiome data from DKD and control mice indicated significant differences in microbial composition between the two groups (Fig. 1A). Although not statistically significant, the alpha and beta diversity of the microbiota in DKD mice were lower than those in the control mice (Fig. 1B). At the phylum level, community bar plot analysis (Fig. 1C) and Circos plot (Fig. 1D) highlighted notable differences in microbial community structure, particularly the dysregulated Firmicutes/Bacteroidetes ratio, a well-documented feature in diabetes [41], [42]. Additionally, the top 50 most abundant phyla exhibited significant differences between DKD mice and controls, as shown by the heatmap (Fig. 1E). These data underscore the distinct alterations in gut microbiota composition in the DKD state.

Fig. 1.

Fig. 1

Altered gut microbial community structure and metabolic pathways in DKD mice. (A) Multivariate statistical analyses (PCA, PCoA, and NMDS) of gut microbiome from DKD and control mice. These analyses were used to assess differences in microbial composition. PCA shows separation along the first principal component (PC1), PCoA illustrates clustering patterns based on Bray-Curtis distances, and NMDS visualizes community dissimilarities with stress values indicating goodness of fit. (B) Alpha and beta diversity analyses of the gut microbiota in DKD and control mice. Alpha diversity metrics (e.g., Chao1 index) were used to evaluate species richness, while beta diversity was assessed using Bray-Curtis dissimilarity indices. Statistical significance was determined using the Wilcoxon rank-sum test. (C, D) Community bar plot and Circos plot analyses at the phylum level, displaying the relative abundance and distribution of bacterial phyla in individual samples from DKD and control mice. The bar plot shows the proportional differences in microbial community structure between groups, while the Circos plot highlights the relative abundance and distribution of various phyla within the gut microbiota of both groups. (E) Heatmap of the top 50 most abundant phyla in DKD and control mice, showing relative abundance across individual samples. Clustering analysis visualizes similarities and differences in microbial composition between the groups. (F) KEGG pathway analysis comparing the metabolic functions of gut microbiota in DKD and control mice, with statistical significance assessed using the Wilcoxon rank-sum test. (G) Multivariate statistical analyses (PCA, PCoA, and NMDS) of the tryptophan metabolism pathway (KEGG ko00380) in gut microbiota from DKD and control mice. (H) Pearson correlation analysis showing the top 50 bacterial species significantly associated with serum IPA levels. (I) Redundancy analysis illustrating the relationship between gut microbiota composition and three explanatory variables: blood glucose levels, UACR, and IPA levels, with statistical significance determined using permutation tests. * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001.

Focusing on the functional aspect of these gut microbiota, KEGG pathway analysis revealed significant differences in the metabolism of carbohydrates, lipids, and amino acids between DKD and control mice (Fig. 1F). Notably, the tryptophan metabolism pathway (KEGG ko00380) showed significant differences (P = 0.0373). Multivariate statistical analyses, including PCA, PCoA, and NMDS, corroborated the distinct distribution patterns of bacteria related to this pathway between DKD mice and controls (Fig. 1G). These results suggest aberrant tryptophan metabolism in the gut under DKD condition.

Microbial tryptophan metabolism has gained growing attention in recent years. We reviewed and summarized the bacteria [6], [43], [44], [45], [46] reported to have direct or strong associations with the production of tryptophan gut metabolites—indole derivatives—and constructed a heatmap based on this information (Fig. S3). The results showed significantly different clustering patterns of indole-derivative-associated bacterial communities between DKD mice and controls, suggesting abnormal tryptophan-indole metabolism in the gut under DKD condition. As part of this complex gut metabolism, IPA has also garnered increasing attention. An increasing number of bacteria associated with IPA production have been identified, including Clostridium botulinum, Clostridium caloritolerans, Clostridium sporogenes, Peptostreptococcus russellii, and Peptostreptococcus anaerobius [6], [43], [44]. Our observations indicate that many species are correlated with IPA levels, as detailed by Pearson correlation analysis in Table S2. Fig. 1H displays the top 50 species significantly associated with IPA levels, including the known IPA-producing bacterium Clostridium sporogenes.

Furthermore, redundancy analysis was applied to demonstrate the relationship between gut microbiota composition and three explanatory variables: blood glucose levels, UACR, and IPA levels (Fig. 1I). The results showed that higher blood glucose and UACR levels significantly impact the gut microbiota composition of DKD mice, while higher IPA levels are more closely linked to the gut microbiota composition of control mice. This suggests a close association between a healthy gut microbiota metabolic state and IPA levels.

These findings collectively highlight the substantial alterations in gut microbial community structure and metabolic pathways in DKD mice, with particular emphasis on dysregulated tryptophan metabolism and its potential impact on IPA levels.

Dysregulation of tryptophan metabolism in DKD mice

Untargeted metabolomics analyses were performed to identify differential metabolites in the serum of DKD mice and controls. A total of 579 and 563 metabolites were detected in positive and negative ion modes, respectively (Fig. 2A). Multivariate statistical analyses, including PCA, OPLS-DA, and PLS-DA (Fig. 2B and C), as well as the heat map (Fig. 2D and E), revealed a distinct clustering pattern within each group and noticeable discrimination between DKD mice and the controls. In the negative ion model, 42 metabolites were significantly downregulated, and 81 metabolites were upregulated in DKD mice. In the positive ion model, 27 metabolites were downregulated, and 15 were upregulated (Fig. 2F and G). Detailed information on these differential metabolites is available in Table S3.

Fig. 2.

Fig. 2

Fig. 2

Dysregulation of tryptophan metabolism in DKD mice. (A) Schematic representation of the untargeted metabolomics workflow for serum samples from DKD mice and controls. (B, C) Multivariate statistical analyses (PCA, OPLS-DA, and PLS-DA) showing distinct clustering within each group and clear discrimination between DKD mice and controls in both ion modes. (D, E) Heat maps and (F, G) metabolic profiles of differential metabolites (OPLS-DA: VIP-value > 1, _P_-value < 0.05), with 123 metabolites significantly changed in the negative ion mode and 42 in the positive ion mode compared to controls. (H) Principal metabolic pathways of dietary tryptophan in the body. (I) Heatmaps of all identified tryptophan metabolites, with each row representing a different metabolite and each column a sample. The color intensity indicates the expression level, with red indicating upregulation and blue indicating downregulation. (J) Chord diagram showing correlation coefficients (r) between all differential metabolites and UACR. (K) Scatter plot displaying the correlation coefficients (r) and significance levels (_P_-value) for various indole derivatives with UACR. Each dot represents an indole derivative, with the dot size indicating the correlation strength (larger dots represent stronger correlations) and color intensity representing the _P_-value (darker colors indicate more significant correlations). * P< 0.05, ** P< 0.01, *** P< 0.001, **** P< 0.0001. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Six primary dysregulated metabolic pathways were identified in DKD mice: amino acid metabolism, lipid metabolism, carbohydrate metabolism, cofactors and vitamins metabolism, nucleotide metabolism, and energy metabolism (Fig. S4). The markedly altered metabolites fell into several categories, including glycerophospholipids, fatty acyls, carboxylic acids and derivatives, organooxygen compounds, benzene and substituted derivatives, prenol lipids, steroids and steroid derivatives, and indole derivatives (Fig. S5).

Indole derivatives are small molecules produced by the metabolism of dietary tryptophan by intestinal bacteria [6]. Tryptophan is metabolized through three main pathways in the body: a small portion for protein synthesis and serotonin and tryptamine production, over 90% is catabolized via the kynurenine pathway, and the remainder is converted into indole derivatives by gut microorganisms (Fig. 2H) [47]. These indole derivatives have garnered significant attention in recent years due to their role in signal transduction, regulation of intestinal homeostasis, metabolism, and inflammation [45]. Cluster analysis showed that these indole derivatives were altered in the serum of DKD mice (Fig. 2I, Table S4). Furthermore, Pearson correlation analysis revealed a close association between indole derivatives and UACR, as demonstrated by the chord diagram (Fig. 2J) and scatter plot (Fig. 2K, Table S5). Fig. 2K shows the correlation coefficients (r) and significance levels (P) for various indole derivatives with UACR. Significant negative correlations were observed for IPA, indole acrylic acid, methyl indole-3-acetate, and 5-methoxydimethyltryptamine, while positive correlations were observed for indole and arachidonoylserotonin. These data revealed a dysregulation of intestinal tryptophan metabolism in DKD mice, which may be linked to the decline in renal function.

Impaired serum IPA levels in DKD patients correlate significantly with fasting blood glucose, HbA1c, UACR, and eGFR

To investigate the role of the tryptophan metabolism pathway and indole derivatives in DKD, a cross-sectional study analyzed tryptophan metabolite levels in 37 human serum samples, including 9 healthy subjects, 14 individuals with T2DM, and 14 DKD patients. Renal pathological staining indicated that the DKD participants included in this study were highly representative of typical DKD patients, as they adhered to identical inclusion and exclusion criteria. Histomorphometric analysis (Fig. 3A) of H&E-stained sections revealed hyaline-like changes in glomerular capillary lumina in DKD patients, accompanied by significant inflammation and eosinophil infiltration in the renal interstitium. Masson staining indicated collagenous matrix deposition in glomeruli and tubules, while PAS glycogen staining showing thickening of the glomerular basement membrane, augmented mesangial matrix in glomeruli, and the presence of Kimmelstiel-Wilson nodules—typical pathological features of DKD.

Fig. 3.

Fig. 3

Impaired serum IPA levels in DKD patients correlate significantly with fasting blood glucose, HbA1c, UACR, and eGFR. (A) Representative images of H&E staining, Masson staining, and PAS staining showing the pathological renal structure and deposited collagenous matrix in DKD patients. Scale bar: 200 μm and 50 μm, respectively. (B) Pearson correlation analyses display the correlation coefficients (r) and significance levels (_P_-value) for tryptophan metabolites with fasting blood glucose, HbA1c, eGFR, and UACR. Each dot represents an indole derivative, with the size of the dot indicating the strength of the correlation (larger dots represent stronger correlations) and the color intensity representing the _P_-value (darker colors indicate more significant correlations). (C) Serum IPA concentrations in DKD (n = 14), T2DM (n = 14), and healthy subjects (n = 9). Pearson correlation analyses assessed the associations of circulating IPA levels with (D) fasting blood glucose, (E) HbA1c, (F) UACR, and (G) eGFR. Data are presented as mean ± SD; * P< 0.05, ** P< 0.01, *** P< 0.001, **** P< 0.0001.

In the cross-sectional study, participants across the three groups were matched for age, sex, body weight, and BMI (details in Table S6). Using the UHPLC-MRM-MS/TS platform, we identified 30 compounds in the tryptophan metabolic pathway (Table S7). Pearson correlation analysis revealed significant associations between tryptophan metabolites—particularly indole derivatives such as indole, indole-3-acetic acid, and IPA—and fasting blood glucose, HbA1C, eGFR, and UACR (Fig. 3B, Table S8). Notably, a significant decrease in serum levels of IPA was observed in DKD patients based on targeted validation (Fig. 3C), consistent with the untargeted metabolomics screening results in DKD mice (Fig. 2I). Consequently, we focused our subsequent work on IPA. The serum concentrations of IPA in healthy controls, T2DM, and DKD patients were 300.33 ± 33.71 nmol/L, 248.31 ± 33.54 nmol/L, and 184.88 ± 49.80 nmol/L (Fig. 3C), respectively. Additionally, IPA showed a negative correlation with fasting blood glucose (r = −0.6414, P < 0.0001, Fig. 3D), HbA1C (r = −0.4398, P = 0.0064, Fig. 3E), and UACR (r = −0.7887, P < 0.0001, Fig. 3G), while demonstrating a positive correlation with eGFR (r = 0.7038, P < 0.0001, Fig. 3F). These results collectively suggest that serum IPA could serve as a biomarker for DKD progression.

IPA mitigated albuminuria and enhanced glomerular filtration barrier integrity in DKD mice

In vivo experiments were conducted to investigate the effects of IPA on DKD (Fig. 4A). In previous studies, 10, 20, and 40 mg/kg IPA were administered to mice via gavage, with no observed side effects [43], [48]. Additionally, supplementation with 27.3 mg/kg/day of IPA for 7 weeks improved insulin resistance [49], and administering 20 mg/kg/day of IPA for 8 weeks effectively alleviated hepatic steatosis in Sprague-Dawley rats [50]. Therefore, we administered IPA at 20 mg/kg/day in our research. Reduced IPA levels were confirmed in DKD mice using UHPLC-MRM-MS/MS, and oral administration of IPA at 20 mg/kg/day for 12 weeks significantly increased the serum IPA concentration in DKD mice (Fig. 4B). While IPA administration moderately lowered random blood glucose concentrations (Fig. 4C), it had no impact on body weight (Fig. 4D). Importantly, IPA supplementation significantly reduced UACR levels in DKD mice (64.01 ± 16.92 mg/g vs. 41.15 ± 11.67 mg/g, P < 0.01, Fig. 4E).

Fig. 4.

Fig. 4

IPA alleviates albuminuria and improves the integrity of the glomerular filtration barrier in DKD mice. (A) Experimental protocol outlining the animal study. (B) Serum IPA levels after 12 weeks of intragastric IPA administration. Changes in (C) random blood glucose and (D) body weight across different groups. (E) UACR levels of mice in different groups after 12 weeks of intragastric IPA administration. Renal morphology assessed by (F) H&E staining and (G) Masson staining showing glomerular and cortical interstitial changes. Scale bar, 200 μm. (H) TEM images illustrating ultrastructural alterations in the glomerular filtration barrier. Scale bar, 2 μm. Data are presented as mean ± SD; * P< 0.05, ** P< 0.01, *** P< 0.001, **** P< 0.0001.

Histological analysis of kidney sections, as demonstrated by H&E staining (Fig. 4F), revealed a disrupted renal structure, interstitial and tubular edema, glomerular shrinkage, and perivascular inflammatory cell infiltration in DKD mice. Masson staining (Fig. 4F) demonstrated an increased area of renal fibrosis and glomerulosclerosis in DKD mice. TEM observation (Fig. 4G) unveiled impairment of the glomerular filtration barrier in DKD mice, characterized by diffuse thickening of the glomerular basement membrane, widespread proliferation of endothelial cells, and the disappearance or extensive fusion of endothelial fenestrae and foot processes. Following 12 weeks of IPA treatment, kidney structure was restored, renal fibrosis was reduced, and the integrity of the glomerular filtration barrier improved. These histological analyses confirmed the beneficial effects of IPA administration in alleviating renal pathological damage in DKD.

IPA attenuated mitochondrial morphology, structure, and functional impairments in GECs of DKD mice

The TEM images of renal cortex (Fig. 4G) illustrated the improved structure of GECs under IPA administration. As a crucial component of the glomerular filtration barrier, GECs play a key role in reducing abnormal albumin excretion [51]. However, being highly differentiated cells, GECs have limited proliferation and regenerative capacity, relying heavily on mitochondrial efficiency [34]. Hyperglycemia-induced excessive ROS production leading to mitochondrial dysfunction is a significant factor in GEC injury in DKD patients [52], [53], [54]. Therefore, we investigated whether exogenous IPA supplementation could protect the mitochondrial structure and function of GECs in DKD mice.

As depicted in Fig. 5A, the mitochondrial structure in the normal control was intact, characterized by clearly visible inner and outer bilayer membranes, densely stacked cristae membranes, and an evenly distributed matrix. In contrast, mitochondria in DKD mice appeared disorganized and swollen, showing shortened, broken, blurred, or reduced cristae and a transparent, vesicular matrix. IPA treatment effectively mitigated the observed damage to mitochondrial structures in GECs. Quantitative image analysis revealed a decrease in the number of mitochondria (Fig. 5B) and an increase in volume (Fig. 5C) of GECs in DKD mice. Additionally, the renal cortex of DKD mice exhibited a reduction in ATP content (Fig. 5D) and an accumulation of MDA (Fig. 5E), While IPA administration reversed these outcomes (Fig. 5B–E).

Fig. 5.

Fig. 5

IPA ameliorates alterations in mitochondrial morphology, structure, and function in GECs of DKD mice. (A) Representative TEM images revealing mitochondrial morphological and structural changes in GECs, with blue arrows indicating GEC mitochondria. Scale bar, 1 μm. Quantitative analysis of (B) mitochondrial density and (C) volume based on electron microscope images of GEC mitochondria. (D) ATP content and (E) MDA levels in the kidney cortex of mice. In GECs stimulated with 30 mM glucose, (F) DCHF-DA staining indicated changes in ROS levels, and (G) the fluorescent probe JC-1 demonstrated alterations in mitochondrial membrane potential. Scale bar, 200 μm. Quantification of (H) ATP levels and (I) MDA contents under different interventions in GECs. Data are presented as mean ± SD; * P< 0.05, ** P< 0.01, *** P< 0.001, **** P< 0.0001. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Subsequently, we established an in vitro DKD model of high glucose injury to explore the role of IPA on primary GECs mitochondria by assessing ROS, mitochondrial membrane potential, ATP, and MDA contents. ROS (DCHF-DA staining) in cells displayed green fluorescence under a fluorescence microscope. After 48 h of 30 mM glucose stimulation, the green fluorescence in GECs increased (Fig. 5F). Treatment with 100 μM and 500 μM IPA significantly reduced the fluorescence intensity. JC-1, a fluorescent probe for detecting mitochondrial membrane potential, shifts from red to green fluorescence under a microscope, signaling a decrease in mitochondrial membrane potential. In Fig. 5G, GECs exposed to 30 mM glucose for 48 h exhibited reduced red fluorescence and increased green fluorescence, indicating a decline in mitochondrial membrane potential. Furthermore, increasing IPA supplementation led to a gradual rise in red fluorescence intensity and a decrease in green fluorescence intensity. Intracellular ATP content significantly dropped (Fig. 5H) and intracellular MDA accumulated (Fig. 5I) after 48 h of 30 mM glucose stimulation compared to the control group. However, IPA treatment reversed the dysregulation of ATP and MDA levels (Fig. 5H and I). Results from in vivo and in vitro experiments indicated that IPA effectively alleviated mitochondrial damage in GECs induced by high glucose.

IPA enhanced mitochondrial biosynthesis and antioxidant defense via PGC-1α deacetylation and nuclear translocation

The progression of DKD is associated with abnormalities in mitochondrial biogenesis and oxidative injury [55], [56]. Peroxisome proliferator-activated receptor-γ co-activator-1α (PGC-1α) is a transcriptional coactivator that translocates to the nucleus upon activation, where it binds with various transcription factors, including mitochondrial transcription factor A (mtTFA), a crucial regulator of mitochondrial biogenesis [57], and superoxide dismutase 2 (SOD2), a key antioxidant enzyme [58]. This interaction regulates essential physiological processes such as mitochondrial biogenesis and oxidative stress [59], [60]. Therefore, PGC-1α is considered the hub for mitochondrial regulation. Given the recognized benefits of IPA on mitochondrial function and the observed protective effects of IPA on mitochondria in this experiment, we investigated its impact on the PGC-1α-SOD2/mtTFA signaling pathway.

Western blotting (Fig. 6A, C, D, and F), immunofluorescence (Fig. S6A, B, D, and E), and RT-qPCR (Fig. S6G and H) analyses revealed a significant decrease in the levels of PGC-1α, SOD2 and mtTFA in the renal cortex of DKD mice and in GECs exposed to 30 mM glucose for 48 h compared to controls. Administration of IPA increased the expression of SOD2 and mtTFA but did not alter the overall content of PGC-1α (Fig. 6A, C, D, and F; Fig. S6A, B, D, and E; Fig. S6G and H). Nucleoprotein analysis (Fig. 6B, C, E, and F) and immunofluorescence images (Fig. 6G and H) demonstrated that IPA facilitated the nuclear translocation of PGC-1α. Previous studies have demonstrated that the cellular localization and activity of PGC-1α are determined by its acetylation/deacetylation ratio [61]. To elucidate the molecular mechanisms underlying the nuclear translocation of PGC-1α following IPA treatment, we measured the deacetylation levels of PGC-1α. Immunoprecipitation analyses (Fig. 6I and J) revealed that exposure to 30 mM glucose increased the acetylation of PGC-1α in GECs, while IPA intervention reversed this modification.

Fig. 6.

Fig. 6

Fig. 6

IPA promotes mitochondrial biogenesis and antioxidant defense by facilitating PGC-1α deacetylation and nuclear translocation. Protein levels of PGC-1α (in whole cell lysate, WCL), SOD2, mtTFA, and PGC-1α (in nucleus) in (A–C) the renal cortex tissue of mice and in (D–F) GECs. Representative images of PGC-1α immunofluorescence in (G) kidney sections and (H) GECs. Scale bar, 50 μm. (I, J) Immunoprecipitation of Ac-lysine with PGC-1α in GECs. Endogenous PGC-1α was immunoprecipitated using anti-PGC-1α, and the immunoprecipitates were analyzed with anti-Ac-lysine. (K, L) Protein levels of PGC-1α, SOD2, and mtTFA in GECs with the intervention of ZLN005, a PGC-1α agonist, and SR-18292, a PGC-1α inhibitor. (M) ROS levels, (N) mitochondrial membrane potential, (O) ATP levels, and (P) MDA contents of GECs with the intervention of SR-18292. Scale bar, 200 μm. Data are presented as mean ± SD; * P< 0.05, ** P< 0.01, *** P< 0.001, **** P< 0.0001.

Next, we investigated the pivotal role of PGC-1α in mitochondrial protection using ZLN005, a PGC-1α agonist, and SR-18292, a PGC-1α inhibitor. As depicted in Fig. 6K–P, the intervention with 10 μM SR-18292 reversed the regulatory effect of IPA on the expression of PGC-1α, SOD2, mtTFA, ROS, JC-1, ATP, and MDA.

The data in this section suggested that IPA promoted the deacetylation and nuclear translocation of PGC-1α in GECs of DKD mice. Subsequently, it enhanced the expression of SOD2 and mtTFA, mediating mitochondrial biogenesis and mitigating oxidative damage.

SIRT1 mediated IPA's renal protection via PGC-1α-SOD2/mtTFA in GECs of DKD

Silent mating type information regulation 2 homolog 1 (Sirtuin 1, SIRT1), a member of the mammalian histone deacetylases Sirtuin family, plays a pivotal role in metabolism and renal health by deacetylating various proteins, including PGC-1α [62]. We boldly hypothesized that SIRT1 maybe a major component in the process of IPA regulating PGC-1α signaling pathway.

Western blotting (Fig. 7A–D) and immunofluorescence (Fig. S6C and F) analyses revealed that IPA administration rescued the reduction of SIRT1 protein levels in the kidney of DKD mice and in GECs stimulated by 30 mM glucose. The intervention with 5 μM with EX 527, an inhibitor of SIRT1, annulled the nuclear accumulation of PGC-1α promoted by IPA (Fig. 7E and G), and reversed the up-regulation effects of IPA on SOD2 and mtTFA (Fig. 7F and H).

Fig. 7.

Fig. 7

Fig. 7

IPA upregulates SIRT1 expression and mediates mitochondrial biogenesis and antioxidant defense through the SIRT1-regulated PGC-1α-SOD2/mtTFA signaling pathway. Protein levels of SIRT1 in the (A, C) renal cortex tissue of mice and in (B, D) GECs. (E–H) Protein levels of PGC-1α (in the nucleus), SIRT1, PGC-1α (in whole cell lysate, WCL), SOD2, and mtTFA in GECs with the intervention of SRT 1720, an agonist of SIRT1, and EX 527, an inhibitor of SIRT1. To evaluate the impact of SIRT1 deficiency on IPA's protective effect, SIRT1 heterozygous (SIRT1+/-) mice were utilized in (I-N), and (O-V) GECs were derived from wild type (WT) and SIRT1+/- mice. (I) UACR levels in mice from various groups after 12 weeks of intragastric IPA administration. (J–L) Protein levels of PGC-1α (in the nucleus), SIRT1, SOD2, and mtTFA, and (M) the ATP and (N) MDA content in the renal cortex tissue of mice in different groups. (O-R) Protein levels of PGC-1α (in the nucleus), SIRT1, SOD2, and mtTFA in GECs. (S) The ATP levels, (T) MDA content, (U) ROS levels (V), and mitochondrial membrane potential in GECs. Scale bar, 200 μm. (W) Representative images of immunohistochemistry analysis of SIRT1, PGC-1α, SOD2, and mtTFA on kidney biopsy specimens from both healthy subjects and DKD patients. Scale bar, 250 μm. Data are presented as mean ± SD; * P< 0.05, ** P< 0.01, *** P< 0.001, **** P< 0.0001.

To assess whether SIRT1 deficiency could diminish the protective effect of IPA, we conducted studies using SIRT1 heterozygous mice, as homozygotes exhibit embryonic and perinatal lethality [63]. Genetic ablation of one allele (SIRT1+/-) significantly reduced the expression of SIRT1 mRNA and protein in the kidney cortex tissue compared to wild-type (SIRT1+/+) mice (Fig. S7A–F). IPA administration in SIRT1-deficient DKD mice failed to prevent the elevation of UACR (Fig. 7I), did not rescue the downregulation of PGC-1α-SOD2/mtTFA signaling (Fig. 7J–L), and had no impact on ATP and MDA levels in kidney cortex tissue (Fig. 7M and N). In vitro experiments using GECs from SIRT1+/- mice also indicated the loss of protective effect of IPA in SIRT1-deficient DKD mice after exposure to high glucose (Fig. 7O–V).

Subsequently, we conducted immunohistochemistry analysis on kidney biopsy specimens from three healthy subjects and three DKD patients. As depicted in Fig. 7W, immunohistochemistry staining indicated a decrease in the expression of SIRT1, SOD2, and mtTFA, along with reduced nuclear translocation of PGC-1α in the kidneys, particularly in the glomeruli. This underscored the significant impairment of the SIRT1/PGC-1α signaling pathway in patients with DKD.

The data in this section indicates that SIRT1 mediated the promotion of mitochondrial biosynthesis and antioxidant defense by IPA through the PGC-1α-SOD2/mtTFA signaling pathway in GECs of DKD.

IPA upregulated SIRT1 levels by inhibiting its ubiquitin–proteasome degradation

The administration of IPA increased SIRT1 protein content, as indicated by the data above, but did not alter SIRT1 mRNA levels, as determined by RT-qPCR analysis (Fig. 8A and B). To exclude the influence of protein synthesis on SIRT1 levels, GECs were pretreated with the protein synthesis inhibitor cycloheximide (CHX). As depicted in Fig. 8C and D, CHX decreased SIRT1 levels but did not nullify the up-regulation of SIRT1 protein abundance induced by IPA administration. This suggests that inhibiting SIRT1 degradation, rather than promoting protein synthesis, contributes to the up-regulation of SIRT1 levels in response to IPA treatment.

Fig. 8.

Fig. 8

Fig. 8

IPA elevated SIRT1 levels by preventing its ubiquitin–proteasome degradation. (A, B) SIRT1 mRNA expression in the kidney cortex tissue of mice and in GECs. (C, D) SIRT1 protein expression under the intervention of the protein synthesis inhibitor cycloheximide (CHX). (E, F) SIRT1 protein expression with the intervention of MG132, a proteasome inhibitor, and Bafilomycin A1, an autophagy inhibitor. (G, I, J) Polyubiquitin and SIRT1 protein expression in whole cell lysates (WCL) of GECs. (H, K) Immunoprecipitation of polyubiquitin with SIRT1 in GECs. Endogenous SIRT1 was immunoprecipitated using anti-SIRT1, and the immunoprecipitates were analyzed with anti-ubiquitin. (L, N) Protein expression of SIRT1, MDM2, Smurf2, CUL4, and COP1 in WCL of GECs. (M, O) Immunoprecipitation of MDM2, Smurf2, CUL4, and COP1 with SIRT1 in GECs. Endogenous SIRT1 was immunoprecipitated using anti-SIRT1, and the immunoprecipitates were analyzed with anti-MDM1, anti-Smurf2, anti-CUL4, and anti-COP1. Data are presented as mean ± SD; * P< 0.05, ** P< 0.01, *** P < 0.001, **** P < 0.0001.

The ubiquitin–proteasome system (UPS) and the autophagy-lysosome pathway are the two major intracellular protein degradation mechanisms [64]. As shown in Fig. 8E and F, the intervention of MG132, a proteasome inhibitor, but not Bafilomycin A1, an autophagy inhibitor, rescued the reduction of SIRT1 protein levels after exposure to high glucose. These data do not support autophagy-lysosome-mediated protein degradation in regulating SIRT1 levels in GECs exposed to high glucose but rather seem to favor the involvement of the UPS in SIRT1 degradation. To address this issue, we examined the polyubiquitination levels of SIRT1 through immunoprecipitation (Fig. 8G–K). The results revealed mild polyubiquitination of SIRT1 in GECs under basal conditions and an upregulation of polyubiquitination upon exposure to high glucose. However, IPA mitigated this increase induced by high glucose in GECs (Fig. 8H and K).

The UPS is dedicated to degrading intracellular proteins to modulate cellular homeostasis and organismal functions. It involves a series of enzymatic reactions, catalyzed by ubiquitin-activating enzyme E1, ubiquitin-conjugating enzyme E2, and ubiquitin ligase E3, leading to the covalent attachment of a single ubiquitin molecule or ubiquitin chain to substrate proteins [65]. The final step, where E3 transfers ubiquitin from E2 to the substrate, determines the specificity of ubiquitination modification [66]. To clarify the proteasomal degradation of SIRT1 in our study, we reviewed previously published literature [67], [68], [69], [70] and used the UbiBrowser software (https://ubibrowser.ncpsb.org/ubibrowser) to predict the E3 ligase of SIRT1 and found several candidates including murine double minute 2 (MDM2), SMAD ubiquitin regulatory factor 2 (SMURF2), cullin 4B (CUL4)-ring complex, and constitutively photomorphogenic 1 (COP1). According to the immunoprecipitation analysis presented in Fig. 8L–O, SIRT1 emerged as a target substrate for the four aforementioned E3 ligases in GECs. Both MDM2 and SMURF2 exhibited an augmented interaction with SIRT1 upon exposure to high glucose stimulation. Notably, treatment with IPA attenuated the interaction between MDM2 and SIRT1, suggesting that IPA inhibited the proteasomal degradation of SIRT1.

IPA inhibited the phosphorylation induced ubiquitination and proteasome-dependent degradation of SIRT1

The various post-translational modifications of proteins in organisms often complement each other, working in concert to facilitate intracellular signal transduction [72]. Typically, substrate phosphorylation is necessary for the recognition by E3 ligases and subsequent degradation [73]. Previous evidence has demonstrated that Ser-46 phosphorylation (Ser-47 in human SIRT1) leads to the reduction of mouse SIRT1 protein through ubiquitination-mediated protein degradation. In our study, we assessed SIRT1 phosphorylation using a phospho-specific antibody designed against phospho-serine 47 of human SIRT1. With this antibody, we observed mild phosphorylation of SIRT1 in mouse kidney cortex tissue (Fig. 9A and C) and in GECs (Fig. 9B and D). This modification was upregulated in DKD mice or in GECs exposed to high glucose, and the administration of IPA inhibited this phosphorylation (Fig. 9A–D).

Fig. 9.

Fig. 9

IPA suppresses phosphorylation-induced ubiquitination and proteasome-dependent degradation of SIRT1. (A–D) Expression of pSIRT1 (Ser46) protein in the kidney cortex tissue of mice and in GECs. (E) SIRT1 mutational plasmid constructed by substituting Ser46 with alanine (Ala-46, A46) and aspartate (Asp-46, D46). (F–J) Protein expression of polyubiquitin, pSIRT1 (Ser46), and SIRT1 in whole cell lysates (WCL) of GECs. Immunoprecipitation of polyubiquitin with SIRT1 in GECs. Endogenous SIRT1 was immunoprecipitated using anti-SIRT1, and the immunoprecipitates were analyzed with anti-ubiquitin. Data are presented as mean ± SD; * P < 0.05, ** P< 0.01, *** P< 0.001, **** P< 0.0001.

The association between substrate phosphorylation and protein reduction suggests that SIRT1 protein may undergo degradation in response to Ser-46 phosphorylation. To explore this possibility, we generated a SIRT1 mutant by replacing Ser46 with alanine (Ala-46, A46) and aspartate (Asp-46, D46) (Fig. 9E). Immunoprecipitation analysis (Fig. 9F) demonstrated that the transfection of SIRT1-A46 plasmid significantly decreased SIRT1 phosphorylation (Fig. 9H) and down-regulated the polyubiquitylation modification of SIRT1 (Fig. 9J), while transfection with SIRT1-D46 plasmid markedly increased SIRT1 phosphorylation (Fig. 9H) and up-regulated the polyubiquitylation modification of SIRT1 (Fig. 9J). Importantly, the site-directed mutation of SIRT1 abolished the regulatory effect of IPA on SIRT1 phosphorylation (Fig. 9H), SIRT1 expression (Fig. 9I), and SIRT1 polyubiquitylation (Fig. 9J).

The data in this section suggest that SIRT1 phosphorylation, critical for ubiquitylation degradation, is modulated by IPA. Mutational analysis confirms the regulatory impact on SIRT1 phosphorylation, expression, and ubiquitylation, highlighting the intricate interplay between IPA and SIRT1 degradation pathways.

Discussion

The bidirectional crosstalk between gut microbiota and kidneys is increasingly recognized. A pivotal mechanism involves the production of small molecules by gut microbes, capable of exerting effects at or beyond the host gut barrier [75]. In our study, metagenomic sequencing revealed significant alterations in the gut microbiota composition of DKD mice, partially reflected in changes in the microbial communities and functions related to intestinal tryptophan metabolism. Untargeted serum metabolomics analysis further revealed dysregulated intestinal tryptophan metabolism in DKD mice. Decreased levels of IPA were confirmed in DKD patients, showing a significant negative correlation with diabetes-related renal injury. Moreover, IPA treatment in DKD mice decreased albuminuria and improved the integrity of the glomerular filtration barrier. Mechanistically, IPA inhibited SIRT1 phosphorylation-mediated ubiquitination degradation, restoring SIRT1′s role in promoting PGC-1α deacetylation and nuclear translocation, thereby increasing the expression of genes related to mitochondrial biosynthesis and antioxidant defense in GECs. Our results provide evidence of mitochondrial protection by IPA in GECs of DKD mice and provide new insight for DKD management.

Metabolic abnormalities in DKD have been extensively reported [76]. Initial research focused on short-chain fatty acids [77], branched-chain amino acids [78], and carnitine-derived metabolites [79]. Advances in metabolomics have facilitated the identification of increasingly crucial intestinal metabolites, hastening the discovery of potential biomarkers to enhance the diagnosis and prognosis of kidney diseases [80]. Our untargeted metabolomics study on mouse serum provided further evidence of metabolic abnormalities in DKD, underscoring the importance of focusing on metabolic changes in this condition.

The metabolism of tryptophan, a precursor of various bioactive compounds, has gained increasing attention due to its importance in health [81]. Previous studies have revealed dysregulated tryptophan metabolism in DKD [82], [83]. Consistently, our metagenomics analysis showed significant alterations in the gut microbiota composition and metabolic functions under DKD condition, including bacterial communities directly or strongly associated with the production of tryptophan indole derivatives [6], [43], [44], [45], [46]. This suggests abnormal tryptophan-indole metabolism in the gut under DKD condition. Furthermore, our untargeted metabolomics study observed significant tryptophan metabolic abnormalities in the serum of DKD mice, particularly in indole derivatives produced by gut microbiota, which were significantly associated with renal function decline. This prompted us to further investigate this phenomenon in a human cohort. Contrary to our findings in mice, no significant changes in the levels of indole acrylic acid and methyl indole-3-acetate were observed in the human cohort, despite their significant decrease in DKD mice and associations with renal function decline. Instead, indole-3-acetic acid, which was not prominent in the untargeted screening, was found to be elevated in DKD patients and strongly correlated with fasting blood glucose, UACR, and eGFR. IPA, as expected, maintained consistent results across both screening and validation sets, showing significantly reduced levels in the serum of both DKD mice and patients. This finding may partially explain the reduced abundance of IPA-metabolizing bacteria, such as Clostridium sporogenes, observed in the fecal samples of DKD mice in our metagenomic analysis. Additionally, it corroborates the redundancy analysis observation that higher IPA levels are more closely linked to the gut microbiota composition of control mice. Moreover, IPA was strongly correlated with glycometabolic abnormalities and renal function decline. Thus, it is an attractive hypothesis that serum IPA may serve as a biomarker for DKD, warranting further exploration into its potential beneficial effects on DKD.

Previous studies have indicated that higher serum level of IPA is associated with a reduced risk of developing T2DM and correlates with increased insulin secretion and sensitivity [7]. In a recent clinical survey involving 9,180 participants from diverse racial/ethnic backgrounds, IPA content in the blood demonstrated a negative correlation with T2DM risk [8]. Consistent with these findings, our study revealed significantly lower serum IPA concentrations in T2DM patients compared to healthy individuals, displaying a negative association with fasting blood glucose and HbA1c, indicating a close relationship between IPA and T2DM severity. Additionally, clinical investigation revealed a notably lower serum IPA level in CKD patients compared to healthy subjects, especially those with rapid renal function decline [13], suggesting IPA as a potential indicator for CKD progression. However, this study [13], excluding participants with diabetes, left the relationship between IPA and CKD caused by T2DM unclear. In our present study, we observed a reduced circulating level of IPA in DKD patients compared to healthy and T2DM subjects. IPA levels was positively correlated with eGFR and significantly negatively correlated with UACR. Excitingly, animal experiments provide compelling evidence for the renal protective effects of IPA. IPA intervention significantly decreased albuminuria in DKD mice and effectively ameliorated renal fibrosis and morphological abnormalities, including disrupted renal cortex structure, diffuse interstitial and tubular edema, inflammatory infiltration, and damage to the glomerular filtration barrier. Conclusively, our findings underscore a close association between serum IPA levels and renal function decline in patients with DKD and a renal protective effect of IPA in DKD mice.

In early research on screening indole compounds as neuroprotective agents, IPA stands out as an ideal antioxidant due to its high efficiency in scavenging free radicals without generating pro-oxidative intermediates [14], [15]. Given the pivotal role of mitochondria in energy metabolism, oxidative phosphorylation, and oxidative stress in eukaryotes [84], the antioxidant properties of IPA prompted inquiries into its potential mitochondrial protection. In Alzheimer's disease mice, IPA effectively mitigated mitochondrial damage in the hippocampus, striatum, and cortex by enhancing mitochondrial respiratory rate, membrane potential, and ATP content, while reducing ROS levels [16]. Recent studies identified IPA as a mitochondrial modulator in cardiomyocytes, influencing maximal mitochondrial respiration and myocardial contractility in an ex vivo heart failure mouse model [18]. In DKD mice, IPA intervention improved cognitive impairment by increasing mitochondrial DNA levels in the brain [20]. These studies suggest that mitochondria may be a target of IPA. In the current study, what is particularly noteworthy is the protective effect of IPA on GECs' mitochondrial morphology, structure, and function. As expected, IPA administration alleviated morphological abnormalities in GECs' mitochondria in DKD mice, with increased mitochondrial quantity and ATP content. In a high glucose-induced DKD cell model, IPA treatment reduced ROS and MDA content in GECs, improving mitochondrial membrane potential and ATP levels. Our findings underscore the role of IPA in mitochondrial protection.

Due to transcriptional activation of genes in mitochondrial biogenesis and antioxidant enzyme, PGC1-α is known as a central regulator of mitochondrial structure and function [85]. Strategies to enhance PGC-1α activity and its mediated mitochondrial protection offer substantial benefits in kidney disease [86]. Previous study has shown that activating PGC-1α improves oxidative stress and apoptosis in GECs induced by high glucose exposure [87]. A recent study reported that IPA intervention alleviated neuropathy in STZ-induced diabetic rats and high glucose-stimulated mouse brain neuroma cells, increasing ATP content, reducing MDA, and improving mitochondrial structure and function [19]. This protective effect of IPA on mitochondrial function was mediated by upregulated expression of PGC-1α [19]. However, our research revealed that IPA protected against mitochondrial injury induced by high glucose and upregulated the expression of SOD2 and mtTFA by promoting PGC-1α nuclear translocation, without affecting the protein levels of PGC-1α in GECs. It has been shown that the acetylation/deacetylation ratio of PGC-1α determines its cellular location and function [60], [88]. Current evidence suggests that deacetylation of PGC-1α facilitates its translocation into the nucleus, where it regulates downstream mitochondrial biogenesis and antioxidant genes [60], [89], [90], [91]. In our cellular experiment, exposure to 30 mM glucose increased the acetylation levels of PGC-1α, while IPA intervention significantly promoted its deacetylation and nuclear translocation, subsequently enhancing the expression of key genes involved in mitochondrial biogenesis and antioxidant defense.

SIRT1, a known protein deacetylase, has been reported to activate PGC-1α transcriptional activity by interacting with specific lysine residues, leading to the deacetylation of PGC-1α. This interaction facilitates nuclear translocation and the transcriptional regulation of downstream target genes [61], [88], [92]. Numerous studies have highlighted the significance of SIRT1 in DKD progression. Overexpressing SIRT1 in proximal tubules of DKD mice successfully reduced albuminuria production [93]. Additionally, both podocyte-specific SIRT1 overexpression and treatment with a SIRT1 agonist attenuated diabetes-induced podocyte loss and reduced oxidative stress in glomeruli [94]. Immunohistochemistry analysis in our study of kidney biopsy specimens from healthy subjects and DKD patients revealed substantial impairment of the SIRT1/PGC-1α signaling pathway in patients with DKD. Administration of IPA reversed the decrease in SIRT1 protein levels in the kidneys of DKD mice and in GECs stimulated by high glucose. Treatment with the SIRT1 inhibitor, EX 527, effectively counteracted the effect of IPA on the PGC-1α-SOD2/mtTFA signaling pathway. Moreover, SIRT1 deficiency impeded the suppressive effect of IPA on elevated UACR in DKD mice and hindered IPA's regulatory effect on the PGC-1α signaling pathway. These results emphasize the crucial role of SIRT1 in regulating IPA-mediated mitochondrial protection through the PGC-1α-SOD2/mtTFA signaling pathway.

Interestingly, our findings revealed that IPA selectively increased the protein levels of SIRT1 without affecting SIRT1 mRNA levels. We hypothesized that IPA regulated SIRT1 protein levels by inhibiting its degradation rather than increasing its synthesis. This speculation was subsequently validated through the administration of the protein synthesis inhibitor CHX. These results indicate that the protein stability of SIRT1 should be enhanced by IPA intervention, leading to decreased degradation and increased protein levels. In eukaryotic cells, protein degradation primarily relies on the ubiquitin–proteasome system and autophagy-lysosome system [64]. In our study, intervention with the autophagy inhibitor Bafilomycin A1 did not affect the reduction in SIRT1 expression induced by 30 mM glucose in GECs, while the proteasome inhibitor MG132 successfully increased SIRT1 levels. This suggests that proteasomal degradation, rather than autophagic lysosomal degradation, was involved in the decreased SIRT1 protein levels induced by high glucose in GECs.

Post-translational modifications play a critical role in regulating the diverse functions of proteins within cells [95]. A large-scale proteomic study revealed the ubiquitination of SIRT1 [96]. Similarly, our investigation revealed that IPA effectively suppressed glucose-induced polyubiquitination of SIRT1 in GECs exposed to high glucose concentrations. The process of protein ubiquitination is intricately controlled by the E1-E2-E3 cascade, where E3 ligases dictate substrate recognition and specificity [66]. Through literature review [67], [68], [69], [70] and analysis using the UbiBrowser software [71], we identified several E3 ligases responsible for SIRT1 ubiquitination, including MDM2, SMURF2, CUL4, and COP1. Previous research indicates that SIRT1 is ubiquitinated by the E3 ligase MDM2 in response to oxidative stress-induced cell senescence [67], ubiquitinated by the E3 ligase SMURF2 to inhibit cell proliferation and tumor formation [68], ubiquitinated by the E3 ligase CUL4 to promote cancer cell autophagy [69], and ubiquitinated by the E3 ligase COP1 to exacerbate lipid toxicity [70]. Our study corroborates SIRT1′s status as a substrate for ubiquitination by these E3 ligases, particularly highlighting the regulatory roles of MDM2 and SMURF2. Furthermore, IPA intervention effectively attenuated MDM2-mediated enhancement of SIRT1 polyubiquitination, thereby upregulating SIRT1 protein expression and preventing its proteasomal degradation.

The interplay of diverse post-translational modifications on a specific protein enhances signal processing specificity and combinatorial logic [97]. Cross-talk between phosphorylation and ubiquitination has been documented, with phosphorylation activating E3 ubiquitin ligase activity [98], [99]. Prior research has shown that SIRT1 phosphorylation at Ser 46 (mouse) triggers its ubiquitination and proteasome-dependent degradation [74]. In a recent study, elevated SIRT1 phosphorylation at Ser 46 was linked to mitochondrial dynamic changes and dysfunction in podocytes during the progression of DKD mouse podocyte injury [100]. Inhibiting SIRT1 at Ser 46 mitigated podocyte damage and mitochondrial abnormalities induced by high glucose. In our study, increased phosphorylation of SIRT1 (Ser 46) was observed in mouse kidney cortex tissue and GECs exposed to 30 mM glucose, and IPA administration inhibited this phosphorylation. Mutation of Ser-46 to alanine prevented phosphorylation, ubiquitination, and degradation of SIRT1, while mutation to aspartate increased these modifications. However, site-directed mutation hindered IPA's regulatory effect on SIRT1 levels and its modifications. Our study revealed that IPA closely orchestrated SIRT1 phosphorylation, ubiquitination, and proteasomal degradation, thereby influencing SIRT1 protein expression.

Conclusion

In summary, our study revealed a significant decrease in IPA, a tryptophan-derived metabolite, in both DKD patients and mice, showing a negative correlation with decreased renal function. Mechanistically, IPA elevated SIRT1 levels by inhibiting SIRT1 phosphorylation-mediated ubiquitylation degradation. This, in turn, activated the SIRT1/PGC-1α signaling pathway, promoting mitochondrial biosynthesis and alleviating oxidative stress in GECs. However, there are several limitations in this current study. First, tryptophan is the precursor for IPA production. a tryptophan supplementation group should be included to explore the relationship between tryptophan availability and IPA synthesis in the context of DKD. Second, this study did not initially set different concentration gradients for intervention. Further studies are needed to identify the minimal effective dose and the safe dose.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgement

This study received financial support from the Natural Science Foundation of China (Grant Nos. U22A20286 and 82300911), the Sichuan Science and Technology Program (Grant No. 2023YFS0471), the collaborative project between Sichuan Province and Luzhou-Southwest Medical University (Grant Nos. 2022YFS0617 and 2021LZXNYD-D09), and the Office of Science, Technology, and Talent Work of Luzhou (Grant Nos. 2020LZXNYDP02 and 2021LZXNYD-G01).

Footnotes

Contributor Information

Betty Yuen-Kwan Law, Email: yklaw@must.edu.mo.

Yang Long, Email: longyang0217@swmu.edu.cn.

Yong Xu, Email: xywyll@swmu.edu.cn.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Supplementary Data 1

Supplementary Data 2

References

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Data 1

Supplementary Data 2