Changes of gut microbiota in diabetic nephropathy and its effect on the progression of kidney injury

We aimed to illustrate gut microbiota and short chain fatty acid (SCFA) levels in diabetic nephropathy (DN) patients, and investigate the mechanism of sodium butyrate in diabetic mellitus (DM) rats. Gut microbiota and serum SCFA levels were measured by 16S rDNA and GC-MS. After being built by streptozotocin (DM rats), the DM rats were administered 300 mg/kg sodium butyrate for 12 weeks (DM + BU rats). Gut microbiota, serum and fecal butyrate level were measured. RT-PCR, WB and transmission electron microscopy were performed to explore LC3mRNA or LC3B protein expression, and autophagosomes in kidney tissues. AMPK/mTOR protein expression in renal tissue were also measured. The gut microbial dysbiosis was found in DM and DN groups, and some SCFAs-producing bacteria were decreased in DN group. The serum butyrate concentrations were lower in SCFA-DN group compared with SCFA-HC group and SCFA-DM group in the other cohort. Serum butyrate level was positively correlated with eGFR. Sodium butyrate increased serum and fecal butyrate levels, and improved the enlargement of glomerular area and fibronectin and collagen IV expressions in renal tissues in DM + BU rats. The LC3 mRNA, LC3BII/I ratio and number of autophagosomes were increased in renal tissue of DM + BU rats. Higher p-AMPK/AMPK ratio and lower p-mTOR/ mTOR ratio were shown in renal tissue of DM + BU rats compared with DM rats. We found the decrease in SCFAs-producing bacteria and low SCFAs concentrations in DN patients. Oral butyrate supplementation may improve kidney injury in DM rats, possibly by increasing autophagy via activating AMPK/mTOR pathway.


Introduction
Diabetic nephropathy (DN) is a common co-morbidity of diabetic mellitus (DM) and the major cause of end stage renal disease (ESRD) in the world. The prevalence of DM was reported to 11.2% recently in China [1]. About 35% of all diabetic patients may develop DN [2]. Intensive glycemic control and the use of sodium-glucose cotransporter-2 (SGLT2) inhibitors and renin-angiotensin-aldosterone system (RAAS) blockade may delay the progression of DN [3,4]. However, it is necessary to study the mechanism and risk factors of DN progression and find more treatment strategies to delay disease progression to ESRD.
Recently, an intimate connection between gut and kidney has been proposed [5]. The alteration in construction of the gut microbiome in DM participants was obvious [6]. It was reported that patients with type 2 diabetes had a moderate These authors contributed equally: Kedan Cai, Yanhong Ma degree of gut microbial dysbiosis characterized by decrease in the abundance of some universal butyrate-producing bacteria and increase of various opportunistic pathogens [6].
The short-chain fatty acids (SCFAs), mainly including acetate, butyrate and propionate were generated by gut microbiota [7]. In chronic kidney disease (CKD) patients, serum and fecal SCFAs levels were significantly decreased, which may contribute to disease progression [8]. SCFAs can diffuse into bloodstream through intestinal mucosa [9]. In an ischemia/reperfusion model, SCFAs inhibited apoptosis to increase cell proliferation, and then improved the renal function [10]. Autophagy is a highly conserved cellular process to maintain intracellular homeostasis [11]. When cells encounter starvation or augmented energy demands, AMP-activated protein kinase (AMPK) (a key energy sensor) activates and downregulates mammalian target of rapamycin (mTOR) pathway, which may activate autophagy to maintain homeostasis [12]. In this study, we aimed to identify the changes of the intestinal flora and the serum SCFAs concentrations in DN patients. Then, we investigated the effects and mechanism of sodium butyrate (the most biological effective SCFA) on the progression of DN in streptozotocin-induced diabetic rats [13]. Our findings may provide further understanding of gut-kidney axis and a promising strategy in DN treatment.

Methods
The research protocols were conformed to the provisions of the Declaration of Helsinki [14]. DN was diagnosed clinically based on the following criteria as persistent albuminuria >300 mg/24 h in at least two of three consecutive 24-hour urine collections, presence of retinopathy, and no evidence of other kidney diseases [15]. The medical history, blood pressure, and biochemical results including assessment of fasting glucose and lipids, renal function and urinary albumin creatinine ratio (UACR) were collected. Estimated glomerular filtration rate (eGFR) was calculated by CKD-EPI formula.
For DN patients and DM patients, the exclusion criteria included (1) patients complicated with other kidney diseases, active infection, severe liver dysfunction, cancer and history or existence of gastro-intestinal or systemic diseases known to affect gut bacterial composition; (2) under immunosuppressive therapy; (3) after kidney allograft transplantation or reaching ESRD before renal biopsy; (4) receiving antibiotics or probiotics within 2 months before collecting samples; (5) with incomplete baseline data. The normal controls were from health examination centers in the same hospital, with no hypertension, diabetes, dyslipidemia, metabolic syndrome, cancer, abnormal liver or kidney function, or history of gastro-intestinal or systemic diseases known to affect gut bacterial composition. Fresh fecal samples were collected at hospital in the morning and kept frozen at −80°C until usage. The gene sequencing of gut microbiota was performed using 16S ribosomal DNA sequencing technology. DNA extraction was conducted using PowerSoil® DNA Isolation Kit (MO BIO). DNA concentration (1 ng/μL) and purity were monitored on 1% agarose gels. The V3-V4 region of the bacterial 16S ribosomal DNA genes were amplified via PCR barcode-indexed primers, using KAPA HiFi™ HotStart ReadyMix (2X). The amplified PCR products were purified with Beckman DNA Clean Beads and quantified by the Qubit 2.0 fluorometer (Invitrogen, Carlsbad, CA, USA). The enriched library was loaded in Ion 530™ Chip and sequenced on an Ion S5TM platform (Thermo Fisher Scientific, Waltham, MA) and 600 bp single-end reads were generated, supplied by LC-Bio Technology Co., Ltd (Hang Zhou, China).
Sequencing analysis was performed by Uparse software (Uparse V8.1.1861, http://drive5.com/uparse/). Sequences with ≥97% similarity were assigned to the same operational taxonomic units (OTUs). The subsequent analysis of αdiversity and β-diversity were performed base on this output normalized data. The α-diversity metrics determines species richness and evenness within bacterial populations, including ace and Chao1 (microbial richness), and Shannon index and Simpson index (microbial diversity). All these indices in our samples were calculated with QIIME (Version 1.9.1). The β-diversity determines the shared diversity between bacterial populations. Non-Metric Multi-Dimensional Scaling (NMDS) was carried out using R programming language (Version 3.2.2) and Vegan package. The Linear Discriminant Analysis Effect Size (LEfSe) algorithm was used to identify the genomic features of taxa differing in abundance among three groups. A size-effect threshold of 2.0 on logarithmic LDA score was used for discriminative functional biomarkers.
Participants and measurement of SCFAs using gas chromatograph-mass spectrometry (GC-MS) We enrolled a separate cohort of 27 T2DM patients without renal injury (SCFA-DM group) and 100 DN patients (SCFA-DN group) who received a renal biopsy in our center. The diagnosis of T2 DM or DN was according to the criteria proposed by the American Diabetes Association in 2017 [14] or the criteria by the Renal Pathology Society in 2010 [16], respectively. The control group included 30 healthy controls from the same hospital (SCFA-HC group). The exclusion criteria for all the participants were the same as those listed above.
Blood samples were collected during fasting, centrifuged at 3500 rpm for 5 min after collection, then stored in aliquots at −80°C until usage. Each serum sample was sufficiently mixed with 50 μL 50% sulfuric acid and 200 μL ether, centrifuged at 12,000 rpm for 20 min at 4°C, then stood for 30 min at 4°C. The supernatant ether layer was filtered through anhydrous sodium sulfate and the solution was transferred to a glass vial for GC-MS analysis. The analysis was performed using the GC-MS 7890A-5975C (Agilent Technology, USA). A FFAP capillary column (30 m × 0.25 mm × 0.25 μm) was used for chromatographic separation, and helium (1 mL/min) was used as the carrier gas. These were conducted in Metabo-Profile Biotechnology (Shanghai) Co.,Ltd.

Animal experiments
Sixteen male Sprague-Dawley (SD) rats aged eight weeks and weighed around 180-200 g were purchased from Shanghai Super--B&K laboratory animal Corp.Ltd and maintained at Zhejiang Chinese Medical University Laboratory Animal Research Center with a SPF facility at (22 ± 2°C), 40-70% humidity, with a 12/12 h light/dark cycle and air exchange, and free access to food and water ad libitum. All animal procedures were approved by the Institutional Ethics Committee on Animal Use of the Zhejiang Chinese Medical University (NO.11346). After 2 weeks of adaptive feeding, rats were randomly divided to receive modeling of diabetes or control. Diabetes was induced by single injection of streptozotocin (55 mg/kg) (n = 12). Blood glucose levels were measured one week after the injection to confirm the establishment of diabetic models. The control group was given the same volume of citric acid buffer (NC, n = 4). The diabetic rats were randomly divided to receive sodium butyrate (303410, Sigma, USA) (DM + BU, n = 6) or equal volume of distilled water (DM, n = 6) after three weeks of establishment of diabetic model. Rats in DM + BU group were administered 300 mg/kg sodium butyrate by oral gavage once a day for 12 weeks [17].
The blood glucose levels were determined by a J&J OneTouch UltraVue Blood Glucose Meter every week. The concentration of blood urea nitrogen and 24 h urinary protein was measured every four weeks. The blood urea nitrogen levels were measured by an automatic analyzer (Hitachi Tokyo, Japan). The rats were kept alone in metabolic cages to collect their 24 h urine samples, of which 24 h urinary protein concentration was quantitatively analyzed by an automatic analyzer (Hitachi Tokyo, Japan). All rats were sacrificed after 12 weeks of sodium butyrate or control intervention, with blood, urine, feces in cecum and colon, and renal tissues collected. Kidney weights were measured at the time of sacrifice.
Fresh fecal samples were collected at the time of sacrifice, then frozen until usage at −80°C. The gene sequencing of gut microbiota was performed using 16S ribosomal DNA (16S rDNA) sequencing technology described above. Each colonic fecal sample with 100 mg was mixed with 0.9 mL ultrapure water, crushed with a tissue grinder and then centrifuged at 12,000 rpm for 20 min at 4°C. The detailed method of measurement of serum and fecal SCFAs were same as described above.
LC3 mRNA was measured by real time PCR. Primers for LC3 was from Sangon Biotech (Shanghai, CN) Co., Ltd. The primer sequences were listed in Supplementary Table  3. The LC3B, Phospho-AMPK(Thr172), AMPK, Phospho-mTOR (Ser2448) and mTOR protein from the renal cortex were measured by western blot. The detailed information of primary antibodies was shown in Supplementary Table 4. Protein expression was observed using Image J.
The morphology of renal tissue stained with hematoxylin and eosin was visualized by the microscopy (Olympus VS120, Japan). The glomerular area was measured by Image-Pro Plus 6.0 software. Renal section was stained with specific primary antibody of fibronectin and Collagen-IV in the immunohistochemical staining. Protein expression was observed using Image J. Transmission electron microscopic (TEM) examination was performed to observe autophagy using a HITACHI 7650 electron microscope (HITACHI, Japan).

Statistical analysis
All statistical analyses were performed with SPSS v19.0 software or Prism 7.0 GraphPad Software. The results were expressed as means ± standard deviation for normally distributed continuous variables, median values (interquartile ranges) for non-normally distributed continuous variables and frequencies and percentages for categorical variables. The significance of differences between three groups were analyzed by one-way analysis of variance with Student-Newman-Keuls test or least post-tests when data were normally distributed. Comparisons of non-normally distributed continuous variables were performed using the Kruskal-Wallis test. For categorical variables, the chisquare test was used. Correlations between serum SCFAs and clinical data were analyzed by Pearson's R coefficient using psych package 1.9.12, and visualized by heatmap in corrplot package 0.84. A P value < 0.05 was considered statistically significant.

Results
The gut microbial diversity and microbial composition in DN Baseline clinical and biochemical characteristics of all participants in DM, DN and HC groups are shown in Supplementary Table 1. There was no significant difference of age among the three groups. Calculated by chao1 and ace indexes, the fecal α-microbial richness in DN group was higher than that in DM group (p = 0.0042, 0.0049 respectively), similar to HC group (p = 0.78, 0.80 respectively) (Fig. 1a, b). Whereas the evenness of gut microbiome shown by the community diversity calculated by Shannon and Simpson indexes, was similar in DN group and DM group (p = 0.50, 0.79 respectively) (Fig. 1c, d).
To identify specific bacterial taxa associated with DN, we compared fecal microbiome using LEfSe. The fecal microbial structure and their predominant bacteria showed the significant differences in taxa among the DM, DN and HC groups by a cladogram representative (all p < 0.05, Fig. 1e) and LDA score (p < 0.05, Fig. 1f). The LEfSe showed that the relative abundances of Proteobacteria, δ-proteobacteria and γprobacteria, Moraxellaceae and its Acinetobacter, Pseudomonadales, Erysipelatoclostridium and Hungatella of Firmicutes, Desulfovibrionales and its Desulfovibrionaceae, Desulfovibrio in DN group were higher than those in HC group and DM group. The relative abundances of Clostridia and Clostridiales, Firmicutes, Ruminococcaceae and its Ruminococcaceae_UCG_013, Ruminococcaceae_UCG_014, Ruminococcaceae_UCG_003, Lachnospiraceae and its Lachnospira, Lachnospiraceae_NK4A136_group, Butyricicoccus and Eubacterium were lower in DN group than those in HC group. As we know, Ruminococcaceae, Butyricicoccus and Lachnospiraceae were the SCFA-producing bacteria.  Supplementary Table 2. The major SCFAs identified in the serum samples included acetate, propionate and butyrate. The serum concentrations of acetate, butyrate and total SCFAs were lower in SCFA-DN group than those in SCFA-HC group and SCFA-DM group ( Table 1). The correlation analysis showed serum butyrate level was positively correlated with eGFR level (r = 0.15, p = 0.04) and negatively correlated with urinary albumin creatinine ratio (UACR) level (r = −0.20, p = 0.02) (Fig. 2). Also, there were positive correlations between serum levels of total SCFAs, acetate, propionate and eGFR level (r = 0.35, p < 0.05; r = 0.35, p < 0.05; r = 0.17, p = 0.03; respectively) ( Fig. 2).

Treatment with sodium butyrate improved the renal injury in DM rats
The blood glucose levels in DM rats (n = 6) and DM + BU rats(n = 6) were significantly increased compared with the controls (n = 4), confirming that the streptozotocin-induced diabetic model was successfully established. The blood glucose level decreased in DM + BU rats compared with DM rats after treatment with sodium butyrate at eight weeks and 12 weeks without any adverse sides (Fig. 3a). The ratio of kidney-to-body weight, serum nitrogen level and 24 h urine protein in DM rats was significantly higher than those in controls (Fig. 3b-d). After treatment with sodium butyrate, the DM + BU rats deceased in the levels of nitrogen and ratio of kidney-to-body weight compared with DM rats (Fig. 3b, c). However, there was no decrease in 24 h urine protein in DM + BU rats compared with DM rats (Fig. 3d). The glomerular area in DM rats was bigger than the control rats. After the treatment of sodium butyrate, the glomerular area was smaller in DM + BU rats than DM rats (Fig. 3e). The expression of fibronectin and Collagen IV by immunohistochemistry in renal tissue showed that the glomerular expressions of fibronectin and Collagen IV increased in DM rats compared with control rats. But these changes were inhibited by treatment of sodium butyrate in DM + BU rats (Fig. 3e, f). These results indicated sodium butyrate may be a new therapy for alleviating DN progression.

Change of gut microbiota after treatment of sodium butyrate in DM rats
A total of 670 OTUs were obtained at a 97% homology cutoff. To display microbiome space among samples, the β-diversity calculated by NMDS showed the distribution of fecal microbial community among the three groups (Stress = 0.10) (Supplementary Fig. 1a). The results of sequence analysis in the gut microbiota of each group showed that the relative abundance of Ruminococcaceae_ UCG-014 in the cecum of DM + BU group was higher than that of DM group, while similar to that of NC group (Supplementary Fig. 1b), demonstrating treatment of sodium butyrate may restore the normal gut microbiota in DM rats.
The GC-MS analysis showed that the serum and fecal levels of butyrate significantly decreased in the DM rats compared with normal rats. Sodium butyrate intervention in DM + BU rats increased the serum and fecal butyrate levels ( Supplementary Fig. 1c, d). There was positive correlation between serum butyrate and fecal butyrate levels (r = 0.636, p = 0.008) (Supplementary Fig. 1e).

Sodium butyrate induces autophagy via activating AMPK/mTOR pathway in DM rats
LC3-B formation is recognized as a marker of the existence of autophagosomes in animal experiments [18]. RT-PCR analysis showed that LC3 mRNA level was lower in renal tissue of DM rats than the control rats; after the treatment of sodium butyrate, LC3 mRNA level was increased in renal tissue of DM + BU rats (Fig. 4A). The protein levels of autophagy-associated protein, LC3B II and LC3BI showed LC3B II/I ratio was up-regulated significantly in DM + BU rats compared with DM rats (Fig. 4B). TEM showed more autophagosomes in renal tissue of DM + BU rats compared with DM rates and control rates (Fig. 4C). These results demonstrated that the autophagy level increased in DM rats after the treatment of sodium butyrate.
AMPK is a trimeric protein complex composed of α, β and γ subunits [19], which downregulates mTOR and induces autophagy [20]. The protein levels of p-AMPK, AMPK, p-mTOR and mTOR showed significant higher p-AMPK/AMPK ratio and lower p-mTOR/ mTOR ratio in renal tissue of DM + BU rats compared with DM rats (Fig. 4B). We speculated that sodium butyrate may induce autophagy by activating AMPK/mTOR signal pathway.

Discussion
In the present study, we showed an obvious change in the richness of gut microbiota between DM patients and DN patients, and the altered bacterial community may play an important role in the progression of DN. Tao et al. also showed the similar shift of the richness of gut microbiota in DN patients [21]. SCFAs such as acetate, propionate, and butyrate are the major microbial products and important energetic and signaling molecules. It was reported that serum SCFA levels were lower in CKD patients than healthy controls [8]. In this study, the concentration of  serum acetate, propionate, butyrate and total SCFAs were lower in DN group than those in NC group and DM group. The serum butyrate level was positively correlated with eGFR level, but negatively correlated with UACR level. It suggested that SCFAs may be a new therapy to delay the progression of DN. The reduced level of serum butyrate was positively correlated with the decreased level of fecal butyrate in DM rats in our study. Yamamura et al. demonstrated that fecal SCFAs level was likely to be positively associated with serum SCFAs level, and the relative abundance of several SCFAs-producing bacteria were positively associated with fecal SCFAs even after adjusting for age and sex [22][23][24]. We speculated that the disordered gut microbiome and decreased certain SCFAproducing bacteria may result in the reduced serum SCFAs level in DN patients. Tao et al. showed that at the genus level SCFAs-producing bacteria Prevotella_9 decreased in DN group [21]. In our study, the results revealed that DN patients had apparent dysbiosis in gut microbiota, especially the decreased relative abundance of Ruminococcaceae, Butyricicoccus and Lachnospiraceae, which were SCFAs-producing bacteria of the human gut in Eastern China. In addition, the relative abundance of butyrate-producing bacteria, Ruminococcaceae_UCG-014, was decreased obviously in DM rats compared with normal rats in our study.
In order to demonstrate the effects of SCFAs on DN, we selected sodium butyrate, one of the most biologically effective SCFAs from microbial fermentation in the gut, to treat DM rats. Although probiotic supplements can increase the butyrate level, they occasionally cause detrimental metabolic activities or produce host deleterious metabolites, and cause inappropriate immune responses and systemic infections [25]. Herein, oral butyrate supplement was used to increase the serum butyrate level. After treatment with sodium butyrate, the DM rats showed reduced kidney damage in our study. Furthermore, sodium butyrate improved gut dysbiosis, promoting expansion of the genera Ruminococcaceae_UCG-014. Other reported studies also observed improvements in kidney injury in other models upon the butyrate treatment [26,27]. Li et al. revealed fiber promoted expansion of Prevotella and Bifidobacterium, which increased fecal and systemic SCFA concentrations and reduced expression of genes encoding inflammatory cytokines, chemokines, and fibrosis-promoting proteins in diabetic kidneys [28]. The fecal microbial transplantation was shown to increase the level of fecal butyrate in DN mice, and also improved renal pathological injury [29]. The In vitro studies also showed sodium butyrate played protective effects on podocytes [28], tubular cells [30], endothelial cells [31] and mesangial cells under high glucose status.
Autophagy is a highly conserved cellular process to maintain intracellular homeostasis. Accumulated evidences have shown that activation of autophagy can prevent the progression of DN [11]. In this study, our results showed that LC3 mRNA level and LC3BII/I ratio in DM group were significantly lower than those in control group, and there was little autophagosomes in renal tissue, which indicated that autophagy was inhibited in DM rats. This finding is consistent with the other study [32]. Then, upon the treatment with sodium butyrate, LC3 mRNA level and LC3BII/I ratio, as well as the number of autophagosomes in DM + BU rats increased compared with DM rats. So, we firstly showed that sodium butyrate could trigger autophagy in kidneys of DM rats. Recently, it was reported that SCFAs could regulate autophagy in acute kidney injury by promoting the expression of Atg 7 in renal tubular epithelial cells to inhibit renal fibrosis and improve renal function [10]. In other diseases, sodium butyrate has been found to induce Atg 5-dependent autophagy in mouse neuroendocrine STC-1 cells, which was mainly characterized by increasing expression of LC3-II and decreasing expression of p62 [33]. Other studies also found that sodium butyrate promotes autophagy to inhibit tumor growth in colorectal cells [34,35].
We further investigated the mechanism of sodium butyrate in regulating autophagy in DM rats. The AMPK/mTOR pathway is the classic pathway to regulate autophagy. Butyrate was reported to activate AMPK phosphorylation to improve renal injury in CKD animals [36]. Wang et al. proved that the AMPK/mTOR pathway mediated sodium butyrate induced autophagy in human bladder cancer cells [37]. Our study found that sodium butyrate dramatically increased p-AMPK/ AMPK, with AMPKα phosphorylation at Thr172, and downregulated p-mTOR/mTOR, with mTOR phosphorylation at Ser2448. These findings confirmed that the AMPK/mTOR pathway may mediate sodium butyrate induced autophagy in kidneys of DM rats.
Also, we observed that oral sodium butyrate supplement in DM rats decreased blood glucose level significantly after 8 weeks and 12 weeks. It was in coincided with previous studies [38,39]. Gao et al. showed that there was a significant and positive effect of sodium butyrate on glucose metabolites and insulin resistance [38]. In a randomized double-blinded placebo-controlled trial, combination of butyrate and inulin significantly reduced fasting blood sugar and waist to hip ratio [40]. A study also demonstrated that butyrate-producing bacteria Clostridium butyricum CGMCC0313.1 (CB0313.1) improved diabetic markers of fasting glucose, glucose tolerance and insulin tolerance [41]. Butyrate-mediated glucagon like peptide-1 (GLP-1) secretion may be a major factor in the improvement of insulin sensitivity and diabetes [42]. GLP-1 has been demonstrated to decrease hepatic gluconeogenesis and promote insulin secretion [43]. The increased glucose transporter-4 (GLUT-4) level induced by butyrate in adipose tissue may be considered as one of the other factors to improve glucose metabolites [38]. GLUT4, expressed in skeletal muscle and adipose tissue, is a member of glucose transporter protein family to uptake the glucose in tissues [44,45]. Taken together, this may be an extra effect for controlling hyperglycemia to alleviate the progression of DN.
There were several limitations in our study. Firstly, the direct correlation between serum SCFAs and gut microbiota would not be analyzed in the participants, as they were not from the same cohort. Secondly, the lifestyles or diets were not analyzed due to a lack of records. So, the association between gut microbiota composition and lifestyles or diets needs to be explored in further studies. Thirdly, gene knockout mice, germ-free mice and fecal microbiota transplantation (FMT) were not used in the animal study. Fourthly, the blot for beta-actin as a control for normalization of protein levels in western blot was not performed when comparing the ratio of p-AMPK/AMPK, p-mTOR/mTOR and LC3B II/I. Herein further studies on the underlying mechanism and interactions are needed. Fig. 5 The study of intestinal micro-ecology in DN and its effect on the progression of the nephropathy. It revealed that there was the gut microbial dysbiosis in DN.Meanwhile, the serum total SCFAs, acetate, propionate and butyrate concentrations were significantly decreased in DN. The supplement of sodium butyrate might delay DN progression via AMPK/mTOR pathway to activate autophagy In conclusion, we demonstrated the changes in composition of the gut microbiota, especially the decrease in SCFAsproducing bacteria in DN patients. In addition, we showed that serum concentrations of SCFAs were low in DN patients, and oral butyrate supplementation may improve kidney injury in DM rats, possibly by increasing autophagy via activating AMPK/mTOR pathway (Fig. 5). These findings provided further understanding of gut-kidney axis in diabetic nephropathy and may provide a new promising strategy.

Data availability
The datasets generated during the current study are available from the corresponding author on reasonable request