Medicinal slices form of PLR were purchased from Guangzhou Weida Company. Standard substance of puerarin (purity ≥99.0%), daidzin (purity ≥98.0%) and daidzein (purity ≥98.0%) were purchased from Chengdu Must Bio-Technology Co., Ltd. PLR decoction preparation was conducted as previously described (14,20,21). In brief, a certain quality of PLR was received and extracted in boiling distilled water at a ratio of 1:10 (w/v) for 1 h; the filtrate was collected and the residue was boiled with distilled water (1:6, w/v) for another 1 h. The aforementioned filtrates were concentrated and the content of puerarin, daidzin and daidzein was quantified by using ultra-performance liquid chromatography (specific method is listed in Table SI and Fig. S1). Briefly, the column was an Agilent ZORBAX SB-C18 (3.0×100 mm, 1.8-Micron). The mobile phase comprised methanol and water with a flow rate of 0.3 ml/min. Changes in the proportion of the mobile phase are shown in Table SI. Puerarin, daidzin and daidzein were detected at a UV wavelength of 250, 270 and 260 nm, respectively. Testosterone (240 nm) acted as the internal standard. The retention times of puerarin, daidzin and daidzein were 3.82, 5.19 and 9.55 min, respectively. Organic phase methanol was obtained from Tianjin Damao Chemical Reagent Factory. The respective concentrations of these three compounds were 10.9, 17.5 and 0.3 mg/ml.

All animal experimental procedures were conducted in strict accordance with the National Institute of Health guidelines and were approved (approval no. L-2019216) by the Institutional Animal Ethical Care Committee of Southern Medical University Experimental Animal Center (Guangzhou, China) (22). Specific pathogen-free male C57BL/6 mice (6–8 week, weighing 20–26 g) were purchased from SPF Biotechnology Co., Ltd. The total number purchased was 55. All mice were housed under standard conditions of temperature (22±1°C) and humidity (50±5%) with a 12/12-h light-dark cycle and free access to food and water. After 1 week of acclimatization, HS-feeding mice were administered drinking water containing 2% (w/v) NaCl for 8 weeks as previously described (11) (HS, n=6), The amount of drinking water provided minus the amount of water remaining indicated the amount of water consumed during the day. while PLR-intervention mice received Pueraria decoction (containing 18, 36 and 72 mg/kg Puerarin, respectively) by oral gavage once a day along with 8 weeks HS feeding (PLR-L, PLR-M and PLR-H, n=6 per group). Control mice received normal drinking water during the experiment (CON, n=10). There were no differences in baseline covariates among groups. The feces and urine were collected every week for further examination. After an 8-week period of modeling, cervical dislocation was performed under complete anesthesia with an intraperitoneal injection of 1.5% sodium pentobarbital at a concentration of 40 mg/kg. Death of mice was verified by cardiac arrest, a reduction in body temperature, and a lack of response to strong stimuli. Sampling operation time was ~10–15 min per mouse. Serum, kidney, liver, spleen, colon and cecal contents samples were harvested in a sterile manner for further analysis. Calculation of organ index was performed using organ weight/body weight ratio. After separation of the colon, the length was measured from the junction of the cecum and colon to the anal canal, which was recorded as the length of the colon. Blood collection (0.5–0.8 ml) was performed by open laparotomy through the abdominal aorta with a fine needle after euthanasia under complete anesthesia.

For blood pressure measurement, systolic blood pressure (SBP) and mean blood pressure (MBP) were measured every week via non-invasive tail cuff method, using a BP-2010A instrument (Beijing Softron Biotechnology Co., Ltd.). All measurements were operated between 8 to 12 am. At least six continuous stable results of each mouse were obtained and the average value was calculated as the final result.

The mice were weighed at regular times each day, their bedding was changed and their health was assessed. Humane endpoints were as follows: The body weight of mice being 20% lower than before the experiment, mice being unable to feed or drink, or not responding to gentle stimulation. In the aforementioned circumstances, mice would be considered unsuitable for further experimentation and would be euthanized by neck dissection after complete anesthesia using the anesthesia method and dosage as aforementioned. In the present study, none of the mice reached these humane endpoints.

FMT experiment was performed based on the modified method previously described (11,23). In brief, feces from the donor mice (HS and PLR-H group mice) were collected in aseptic centrifugal tubes respectively and resuspended in PBS at 125 mg/ml, the mixtures were centrifuged at 1,000 × g for 1 min at 37°C and the supernatants were collected and saved in separate 1.5 ml tubes for subsequently microbiota transplantation. Before FMT experiment, the acceptor male C57BL/6 mice (6–8 weeks) were given antibiotics (vancomycin, 100 mg/kg; neomycin sulfate 200 mg/kg; metronidazole 200 mg/kg; and ampicillin 200 mg/kg) intragastrically once a day for 1 w to deplete the gut microbiota. All the mice then received drinking water containing 2% (w/v) NaCl for 8 weeks as aforementioned. A volume of 150 µl of the fecal microbiota solution was simultaneously orally gavaged to mice once a day in the first 2 weeks, every other day in the following 2 weeks and twice each week in the last 4 weeks. The mice that received fecal microbiota solution from HS mice or PLR-H mice were referred to as HS-FMT group or PLR-FMT group (n=8 per group). Control mice received normal drinking water during the experiment (CON, n=5). The total number of mice purchased in this part of the experiment was 21. Blood pressures were measured as aforementioned every week. All the mice had free access to food and water, and the feces and urine were collected before sacrifice. Mice were sacrificed under anesthesia and then blood, kidney, liver, spleen, colon and cecal contents samples were harvested in a sterile manner for further analysis.

Feces were collected in sterilized 1.5 ml tubes and frozen at −80°C before DNA extraction. Microbial DNA from fecal samples were extracted as previously described (24,25). V3-V4 region of 16S rDNA was amplified by PCR using the following primers: 341F, 5′-CCTACGGGNGGCWGCAG-3′ and 806R, 5′-GGACTACHVGGGTATCTAAT-3′. Amplicons were purified using the AxyPrep DNA Gel Extraction Kit (cat. no. AP-GX-250; Axygen; Corning, Inc.). Equimolar pooling of purified amplicons was performed and paired end sequenced (PE250) was conducted on an Illumina platform according to the standard protocols. Alpha diversity were calculated in QIIME (26) (version 1.9.1). The abundance statistics of each taxonomy was visualized using Krona (27) (version 2.6). Principal coordinate analysis (PCoA) and non-metric multidimensional scaling (NMDS) of weighted unifrac distances were generated in R project Vegan package (version 2.5.3). Biomarker features in each group were screened by LEfSe software (28) (version 1.0). Microbial dysbiosis index (MDI) was calculated as follows: MDI=log₁₀[(total abundance in genera increased in disease group)/(total abundance in genera decreased in disease group)]. The Kyoto Encylopedia of Genes and Genomes (KEGG) pathway analysis of the operational taxonomic units was inferred using Tax4Fun (29) (version 1.0) and was generated using Omicsmart, a dynamic real-time interactive online platform for data analysis (http://www.omicsmart.com). Spearman correlation coefficient between environmental factors and species was calculated in R project psych package (version 1.8.4) then generated using the Wekemo Bioincloud (https://www.bioincloud.tech). The calculated P-value was conducted through false discovery rate (FDR) correction, taking FDR ≤0.05 as a threshold.

Kidney samples were collected in sterilized 1.5-ml tubes and frozen at −80°C before DNA extraction. Total RNA was extracted using the Eastep™ Super Total RNA Extraction Kit (cat. no. LS1040; Promega Corporation) according to the manufacturer's instructions. After total RNA was extracted, eukaryotic mRNA was enriched by Oligo(dT) beads. Then fragments were transcribed into cDNA by using NEBNext Ultra RNA Library Prep Kit for Illumina (cat. no. 7530; New England Biolabs, Inc.). The ligation reaction was purified, and PCR amplified. The resulting cDNA library was sequenced using Illumina Novaseq6000 by Gene Denovo Biotechnology Co. Differential expression analysis of RNAs was performed by DESeq2 (30) software between two different groups. The transcripts with the parameter of FDR below 0.05 and absolute fold change ≥2 were considered differentially expressed transcripts. Gene set enrichment analysis (GSEA) was performed using software GSEA (31) and MSigDB (31) to identify whether a set of genes in specific KEGG pathway shows significant differences in two groups.

Total RNA was extracted from kidney and colon tissue using the Animal Total RNA Isolation Kit (cat. no. RE-03011/03014; Foregene Co., Ltd,) according to the manufacturer's instructions. A reverse transcription enzyme, HiScript III RT SuperMix for qPCR (+gDNA wiper) (Vazyme Biotech Co., Ltd.), was applied to obtain cDNA (temperature and duration: 50°C for 15 min; 85°C for 5 sec). The quantitative PCR reactions (initial denaturation: 95°C for 30 sec; 40 cycles of amplification at 95°C for 10 sec and 60°C for 30 sec; followed by melting curve analysis at 95°C for 15 sec, 60°C for 1 min and 95°C for 15 sec) were performed on the LightCycler480 (Roche Diagnostics) using a ChamQ SYBR qPCR Master Mix (Vazyme Biotech Co., Ltd.). Relative quantification of target genes were calculated using the 2^−ΔΔCq method (32). GAPDH was used as a control gene. All the target gene primer sequences are listed in Table SII.

The kidney and colon tissue protein were extracted with a commercial RIPA lysis buffer (cat. no. PC101; Epizyme Biomedical Technology Co., Ltd.) supplemented with 1% cocktail protease inhibitor (cat. no. MB2678; Dalian Meilun Biology Technology Co., Ltd.). Protein samples were quantified using a BCA Protein Assay Kit (cat. no. ZJ102; Epizyme Biomedical Technology Co., Ltd.) and 40 µg protein was loaded per lane. Samples were separated by SDS-PAGE on 10% gels and transferred to a PVDF membrane. After blocking with 5% non-fat milk for 1 h at 37°C, samples were incubated overnight at 4°C with the following primary antibodies: Rabbit anti-Occludin (1:2,000; cat. no. ab216327; Abcam), rabbit anti-ZO-1 (1:2,000; cat. no. ab216880; Abcam), rabbit anti-Claudin-1 (1:2,000; cat. no. YT0942; ImmunoWay Biotechnology Company), mouse anti-TNF-α (1:1,000; cat. no. YM3477; ImmunoWay Biotechnology Company), rabbit anti-IL-1β (1:1,000; cat. no. YT5201; ImmunoWay Biotechnology Company), rabbit anti-β-catenin (1:5,000; cat. no. 51067-2-AP; Proteintech Group, Inc.) and mouse anti-β-actin (1:5,000; cat. no. RM2001; RayBiotech, Inc.). Subsequently, they were incubated with horseradish peroxidase-conjugated goat anti-mouse/rabbit secondary antibody at 37°C for 1 h (1:5,000; cat. no. LF101/LF102; Epizyme Biomedical Technology Co., Ltd.). Membranes were visualized with a Femto Light Chemiluminescence Kit (cat. no. SQ201; Epizyme Biomedical Technology Co., Ltd.) and imaged with a gel image processing system (cat. no. 92-15313-00; FluorChem R System; Bio-Techne). The band intensity was assessed using ImageJ software (version 1.53K; National Institutes of Health).

Serum creatinine was determined by using a Creatinine Assay kit (sarcosine oxidase) (cat. no. C011-2-1; Nanjing Jiancheng Bioengineering Institute). The urinary protein, serum lipopolysaccharide (LPS) and serum TNF-α level were detected manually with commercial ELISA Kits (cat. nos. MM-44286M2, MM-0634M2 and MM-0132M2; Shanghai Meimian Biotechnology, Co., Ltd.) according to the manufacturer's protocols. Serum and urine K⁺, Ca²⁺ and Na⁺ concentrations were determined on an automatic biomedical analyzer (Roche Diagnostics).

FITC Dextran 4-KD (FD-4; Sigma-Adrich; Merck KGaA) was used to detect intestinal permeability in vivo. FD-4 was orally administered to mice at a dose of 0.6 mg/kg 4 h before serum was harvested. The FD-4 level in serum was detected in the dark, based on a standard curve and spectro-fluorometrically with an excitation wavelength of 485 nm and an emission wavelength of 530 nm in a microplate fluorescence reader (Infinite^® M1000 PRO; Tecan Group, Ltd.).

L-012 (FUJIFILM Wako Pure Chemical Corporation) was used to confirm intestinal inflammation in vivo (33). Mice were anesthetized with 3–4% isoflurane induction by inhalation and maintained with 1–1.5% isoflurane and injected with 100 µl of a 20-mmol L-012 solution. After 1 min, IVIS Spectrum CT system was used to acquire bioluminescent images. For post-acquisition analysis, the Living Image software was used to quantify the bioluminescent signals at standardized regions of interest (ROIs) defined on the abdomen.

The left kidney and colon tissue of mice were collected and fixed in 4% paraformaldehyde for 3 days at room temperature. The samples were then embedded in paraffin, sectioned at 4-µm thickness and stained with hematoxylin and eosin (H&E) (hematoxylin for 4 min and eosin for 20 sec) or Masson's trichrome (MASSON) (Weigert's iron hematoxylin stain for 8 min, ponceau S for 5 min and aniline blue for 2 min) at room temperature. The kidney tissue histological damage was assessed according to a previously described scoring method. Briefly, the extent of damage was assessed based on interstitial inflammatory cell infiltration, tubular pattern of renal tubules and loss of brush border. The lesion score was quantified as follows: score 0=0%; score 1=1–10%; score 2=11–25%; score 3=26–75%; score 4=76–100% (34). The kidney fibrosis area was selected randomly from at least 5 points of cortical fields and quantified using ImageJ software (version 1.53K; National Institutes of Health). Immunohistochemistry was performed using the primary antibodies for Occludin, ZO-1 and Claudin-1. Quantification of the average optical density was performed by automated image analysis in five randomly chosen fields of each sample (magnification, ×200). All images were scanned by a NanoZoomer Digital slide scanner and captured at ×200 or ×400 magnification with an NDP. View2 Plus Image viewing software (Hamamatsu Photonics K.K.) was used.

Colon paraffin section samples were generated as aforementioned. TBS solution containing 0.3% Triton X-100 (cat. no. 9002-93-1; Beijing Solarbio Science & Technology Co., Ltd.), 0.25% protein free rapid blocking buffer (cat. no. PS108P; Epizyme Biomedical Technology Co., Ltd.) and 5% goat serum (cat. no. BL210A; Biosharp Life Sciences) was used to block samples for 1 h at room temperature. The samples were then incubated at 4°C with rabbit anti-β-catenin antibody (1:50; cat. no. 51067-2-AP; Proteintech Group, Inc.) overnight. Subsequently, they were incubated with Alexa Fluor 488-labeled goat anti-rabbit secondary antibody (1:500; cat. no. A0423; Beyotime Institute of Biotechnology) at 37°C for 1 h. Then the slices were washed and stained with DAPI (0.5 µg/ml) for 10 min. Images were captured with fluorescent microscopy (Olympus Corporation). Quantification of the fluorescence intensity was calculated by automated image analysis in five randomly chosen fields of each sample (magnification, ×200).

Results were represented as the mean ± standard error of the mean (SEM) and performed using GraphPad Prism software (version 8.0) (GraphPad Software; Dotmatics). When comparing the difference in means between two samples, the Mann-Whitney test was used for non-normal distributions; when comparing the difference in means between multiple samples, the one-way ANOVA was used if the data were normally distributed and the variance was homogeneous, and the Kruskal-Wallis test was used for non-normal distributions. Comparisons between groups involving the same time point and between groups at different time points within the same group were made using two-way repeated-measures ANOVA. Dunn's post hoc test was used after Kruskal-Wallis test, and Bonferroni's post hoc test or Tukey's HSD post hoc test was used after ANOVA. Adonis analysis and anosim analyses were used for PCoA. P<0.05 was considered to indicate a statistically significant difference.

There were no differences in the baseline weight and blood pressure among the groups (Fig. S2A-C); however, after consuming the same amount of sodium, there was a significant difference in SBP among the groups (Fig. 1A). Compared with the CON group, the SBP of the HS group showed significant elevation from the 5th week. By contrast, the SBP of the PLR-H group maintained no difference from the CON group during the whole experiment. SBP in the PLR-L and PLR-M groups were lower than the HS group; however, this reduction was not significant at the end of the experiment (Fig. 1A). Additively, changes in MBP were not significantly different among the groups (Fig. S2D). These results suggested that high-dose PLR supplementation could effectively reduce SBP in CKD. Besides, the weight change was also recorded; HS group gained weight more slowly and exhibited significant different in weight compared with those in the CON group from the 4th week onwards. PLR intervention could maintain a normal rate of weight gain to some extent (Fig. S2E).

In the 8th week, serum creatinine, urine protein and kidney histologic stains were used to assess the extent of kidney damage as well as the efficiency of PLR on kidney protection. Compared with the CON group, the HS group revealed significantly aggravated serum creatinine and urine protein levels (Fig. 1B and C); however, these biochemical parameters were significantly improved in the PLR-M and PLR-H groups. Moreover, the macroscopic images of the kidneys of the PLR-M and PLR-H groups exhibited markedly less hyperemia and swelling compared with those of the HS and PLR-L groups (Fig. 1D). Consistent with this appearance, the kidney coefficient and the histopathological injury observed by H&E staining and the proportion of fibrosis observed by MASSON staining were significantly lower in the PLR-M and PLR-H groups than those in the HS group (Fig. 1E and F). Such histological coefficient differences were not indicated in the liver and spleen among the groups (Fig. S2F and G), however. Notably, middle-and high-dose PLR supplementation significantly alleviated CKD. PLR-H more efficiently reduced the degree of fibrosis and renal H&E scores compared with PLR-M. Taken together, the high PLR dose was considered as the most effective for CKD protection and was evaluated in the following experiment.

Compared with the CON group, the HS mice demonstrated significantly upregulated mRNA levels of Col1a1, Col3a1, TIMP-1 and Fibronectin (Fig. 1G), which indicated fibrosis in the kidneys of the HS group. High-dose PLR supplementation significantly downregulated these fibrosis-related gene expression levels. In addition, the mRNA levels of Kim-1 and Ngal, which indicated renal tubular injury, were significantly higher in the HS group than in the PLR-H group. Considering that the amount of salt intake could influence kidney injury, the total water intake was compared and it was found that the HS and PLR (all doses) groups consumed similar amounts of salt in the whole experiment (Fig. S2H). Additionally, both HS intervention and administation of high-dose PLR did not alter serum sodium loading, but they did increase urinary sodium excretion, as measured by urine and serum Na⁺ (Fig. 1H and I). Serum concentrations of the other major ions also had no differences among the groups (Fig. S2I and J).

To investigate the mechanism by which high-dose PLR exerted its renoprotective effects, RNA-seq was performed to assess the gene expression profile of the kidneys in the HS group vs. the PLR-H group. Compared with the HS group, administration of high-dose PLR altered the expression of 532 genes in the kidneys, with 53 transcripts being significantly upregulated and 479 being downregulated (Fig. S2K). GSEA based on the Gene Ontology (GO) and KEGG databases were performed on the RNA-seq data and revealed that a high PLR dose substantially decreased the expression of genes related to the canonical Wnt signaling pathway (Fig. 1K and L). Furthermore, it was found that HS intake induced a significant increase in the mRNA levels of IL-6, TNF-α, CCL2 and CCL3 in the kidney (Fig. 1J), and this effect was reversed in the PLR-H group. To further verify whether the protective effect of high-dose PLR is related to the Wnt/β-catenin pathway, the mRNA levels of molecules in this pathway were next analyzed. As presented in Fig. 1M, HS intake significantly increased the Wnt3 and Wnt4 mRNA levels. Furthermore, the protein level of β-catenin was measured (Fig. 1N and O), the expression level of which was significantly increased in the HS mice compared with the CON group, demonstrating that HS intake could activate the canonical Wnt signaling pathway. While high-dose PLR administration could significantly downregulate the mRNA and protein level of β-catenin, it also downregulated the mRNA levels of Wnt3 and Wnt4. In addition, TNF-α, which was able to induce β-catenin activation, was also found to have an elevated protein level in the kidney tissue in the HS group and was reduced by high-dose PLR administration. Overall, it was found that high-dose PLR could exert nephroprotective effects by reducing the activation of the Wnt/β-catenin pathway.

Maternal intestinal damage, including inflammation and concomitant increased gut permeability, are involved in numerous extraintestinal diseases including CKD (35). Hence, the intestinal alterations were examined, as shown in Fig. 2A. Compared with the HS group, colon shortening was relieved to certain extent in the PLR-H group, although no statistical difference was observed. To further assess the inflammation level, pro-inflammatory cytokines in the colonic tissues were measured (Fig. 2C). The mRNA levels of IL-6, CXCL1 and CCL2 in the HS group were significantly higher, whereas the mRNA level of inflammatory protective factor IL-10 was reduced compared with those in the CON group. However, high-dose PLR supplementation significantly inhibited the expression of pro-inflammatory factors. In Fig. 2B, an L-012 chemiluminescent probe was used to assess intestinal reactive oxygen species (ROS) in vivo (33). The images indicated that chronic HS intervention markedly elevated the ROS concentration in the colon of mice, with a significantly increased total flux in photons, and this elevation was significantly dampened by high-dose PLR administration.

To further demonstrate the role of PLR administration in intestinal protection, intestinal tight junction markers in the colon were measured. Compared with the CON group, the HS group revealed significantly downregulated mRNA levels of tight junction proteins (Fig. 2I); whereas the serum TNF-α level, LPS level and FD-4 permeability were elevated (Fig. 2D-F). Moreover, the HS mice presented significantly deteriorative histopathological injury in the colon as observed by H&E staining, as well as decreased mRNA and protein levels of ZO-1, Occludin and Claudin-1 observed by immunohistochemical staining and western blotting (Fig. 2G-J). However, all the aforementioned damages induced by HS were largely attenuated by high-dose PLR administration. Collectively, the aforementioned results demonstrated that therapeutic high-dose PLR administration alleviated intestinal inflammation and reinforced the gut barrier.

A pathophysiological colon change is always associated with enteric dysbiosis (36,37). The impact of high-dose PLR on the gut microbiota composition and equilibrium was further examined via 16S rDNA sequencing. The relative abundance of Firmicutes as well as the Firmicutes/Bacteroidetes ratio in the cecal content of HS mice were lower than those in the cecal content of CON mice, whereas the Bacteroidetes level was significantly higher in the HS group (Fig. S3A). High-dose PLR intervention reversed the dominant position of Bacteroidetes and upregulated the Firmicutes/Bacteroidetes ratio. In addition, the bacterial composition in the cecal content samples from the three groups in terms of bacterial phyla and genera were significantly different (Figs. 3A and S3B). At the phylum level, Verrucomicrobia was the predominant phylum in the PLR-H group compared with that of the HS group. At the genus level, the fecal microbiota was dominated by Lachnospiraceae_NK4A136_group, Akkermansia and Lactobacillus. Besides, different abundances of fecal bacterial taxa in the three groups were identified by LEfSe analysis (Fig. 3B), which indicated that seven bacterial genera including short-chain fatty acid (SCFA)-producing bacteria, such as Akkermansia and Bifidobacterium, were enriched by PLR-H, while the other eight taxa, including Rikenella, Prevotellaceae_UCG-001 and Lachnoclostridium were enriched in the HS group (LDA score >3). Moreover, indicator species analysis showed a statistically significant higher relative abundance of Faecalibaculum and Akkermansia at the genus level in the PLR-H group than in the HS group (Fig. 3C). It was also found that Lactobacillus and Bifidobacterium showed an increasing trend after high-dose PLR administration compared with the HS group. Additionally, the MDI revealed an increased value in HS mice (Fig. 3D), which was reversed by high-dose PLR intervention, indicating a decrease in the level of intestinal flora disruption.

In the present study, the alpha diversity was impacted by neither HS nor PLR-H administration (Fig. S3C). However, NMDS and PCoA demonstrated that the overall structure of the gut microbiota was significantly different among the three groups (Fig. 3E and F), indicating that the gut microbiota structure in HS mice was markedly influenced and shifted closer towards that of the CON group by PLR intervention.

To further the present investigation, Spearman correlation analysis was performed (Figs. 3G and S3D). The relative abundances of Akkermansia, Lactobacillus and Bifidobacterium were negatively correlated with kidney injury-related parameters and the Wnt/β-catenin signaling pathway transcripts, while positively correlated with tight junction proteins in the colon. The relative abundance of these beneficial bacteria was also negatively correlated with inflammatory factors not only in the colon but also in the kidney. By contrast, the relative abundance of dominant bacteria (such as Rikenella, Prevotellaceae_UCG-001 and Lachnoclostridium) in the HS group had a strong positive correlation with injury severity in the colon and kidney.

Tax4Fun prediction analysis was then used to annotate the 16S rDNA data with metabolic pathways from the KEGG database (Fig. 3H). The relative abundance of the metabolism-associated signal transduction pathway showed a significant downregulation of the Wnt signaling pathway in the PLR-H group compared with that of the HS group, which was consistent with the GSEA results obtained from RNA-seq about the protective role of high-dose PLR in kidney injury. In addition, the vascular endothelial growth factor (VEGF) signaling pathway, which was proven by previous studies to promote fibrosis in CKD (38), was lower in the PLR-H group than that in the HS group, suggesting that fibrotic signaling pathways were suppressed by high-dose PLR supplementation.

Collectively, PLR administration could rectify gut microbiota dysbiosis and enhance the relative abundances of beneficial bacteria in mice, which were significantly negatively correlated with CKD severity.

Next, to confirm whether the protective effect of PLR against CKD was dependent on the gut microbiome, the gut microbiota derived from mice treated with PLR-H were transplanted to CKD mice (Fig. 4A). Baseline weight and blood pressure had no differences among the groups (Fig. S4A-C). Compared with the HS-FMT group, SBP, serum creatinine and urine protein were significantly reduced in the PLR-FMT group (Fig. 4B-D). As presented in Fig. 4E and F, the PLR-FMT group demonstrated slight macroscopic injury and a lower kidney coefficient than that of the HS-FMT group; however, no statistical differences were exhibited in the liver and spleen coefficient among the three groups (Fig. S4D). Additionally, the kidney histopathological injury and fibrosis proportion were also reduced in the PLR-FMT group compared with those of the HS-FMT group (Fig. 4G). Furthermore, the mRNA levels of the inflammatory factors and chemokines, along with the mRNA levels of renal tubular-related injury and interstitial fibrosis, were also decreased in the PLR-FMT group compared with those of the HS-FMT group (Figs. 4H, I and S4E).

In addition, the activation of the Wnt/β-catenin pathway in the kidney was also explored. As shown in Fig. 4J, the mRNA levels of Wnt1, Wnt3, Wnt4 and β-catenin of the PLR-FMT group were lower than those in the HS-FMT group. Moreover, compared with the HS-FMT group, the protein levels of β-catenin and TNF-α were significantly reduced in the PLR-FMT group (Fig. 4K-M). Taken together, these results suggested that the gut microbiota remodeled by PLR could protect kidney function, the effect of which is probably associated with the downregulation of the Wnt/β-catenin pathway.

Afterwards, it was investigated whether the gut microbiota derived from the PLR group was effective in alleviating the intestine inflammatory response and protecting intestinal barrier function. Colonic length was significantly reduced in the HS-FMT group compared with the CON group; however, this reduction was significantly mitigated by the intervention of PLR-derived gut microbiota and was restored to a length that was not significantly different from that of the CON group (Fig. 5A). Decreased mRNA levels of IL-6, CXCL1 and CCL2, as well as increased IL-10 mRNA expression in the colon were observed in the PLR-FMT group (Fig. 5C) compared with those of the HS-FMT group, which indicated a lower colonic inflammatory environment generated by PLR-derived gut microbiota. Furthermore, the bioluminescence imaging system exhibited images of less ROS in the colonic tissues of the PLR-FMT group than that of the HS-FMT group (Fig. 5B). Furthermore, H&E staining of the PLR-FMT colon revealed less abscission and epithelial cell necrosis, and less inflammatory cell infiltration compared with that of the HS-FMT group. As presented in Fig. 5G-J, the mRNA and protein levels of ZO-1, Occludin and Claudin-1 in the colon were significantly higher; while the serum TNF-α level, LPS level and FD-4 permeability were significantly lower in the PLR-FMT group than those in the HS-FMT group (Fig. 5D-F). These results suggested that the HS-derived gut microbiota triggered gut permeability by deteriorating intestinal barrier function, while the microbiota remodeled by PLR protected against colonic inflammation and intestinal barrier disruption, which reduced translocation of bacterial products and inflammatory cytokines into the serum. In addition, the aforementioned results indicated that the gut microbiota reestablished by PLR could reduce renal tissue fibrosis and colonic epithelial barrier damage.

To investigate the gut microbiota characteristic that exerted renoprotective effects in the FMT intervention, the microbial compositions from the HS-FMT and PLR-FMT groups were further analyzed. The main composition in either the phylum or genus level was similar among the groups (Figs. 6A and S5A). Notably, LEfSe and indicator species analysis indicated that Lactobacillus and Ruminococcaceae_UCG_014 at the genus level were significantly enriched by PLR-FMT (LDA score >3.6, Fig. 6B and C), while higher relative abundances of Akkermansia and Bifidobacterium were also found in the PLR-FMT group. As shown in Fig. S6A, the main composition in the genus level was similar between the PLR-H and the PLR-FMT groups. Indicator species analysis revealed 111 shared genera species in the PLR-H and PLR-FMT groups (Fig. S6B). There were no significant differences in the relative abundance of major genera. The relative abundance of Akkermansia and Alistipes in the PLR-FMT group was significantly lower than that of the PLR-H group, but the relative abundance of Lactobacillus and Lachnoclostridium was significantly higher than that of the PLR-H group (Fig. S6C). In addition, the increased MDI value in the HS-FMT mice was reversed by PLR-FMT intervention (Fig. 6D). Although alpha diversity was not impacted by either HS-FMT or PLR-FMT administration (Fig. S5B), PCoA analysis and NMDS showed that the gut microbiota structure was completely clustered (Fig. 6E and F). Furthermore, Spearman correlation analysis revealed that Lactobacillus, Akkermansia and Ruminococcaceae_UCG_014, which were enriched in the PLR-FMT group, were significantly positively correlated with the gut barrier-related tight junction protein, while negatively correlated with damage-related parameters and the Wnt signaling pathway-related genes in kidney (Fig. 6G). Finally, different gut microbial metabolic functions were explored via Tax4Fun prediction analysis from the KEGG data (Fig. 6H). The VEGF and Wnt signaling pathways were significantly downregulated in PLR-FMT mice compared with the HS-FMT group.

Overall, the aforementioned results indicated that FMT intervention could alter the structure of the gut microbiota, and the supplementary microbiota remodeled by high-dose PLR exerted a similar renoprotective effect as the PLR-H group. Further, Lactobacillus, Akkermansia and Bifidobacterium might be the dominating beneficial bacteria that induced CKD alleviation in both the PLR-H and PLR-FMT groups.