Gut microbiota modulates distal symmetric polyneuropathy in diabetic patients

The gut microbiota from patients with DSPN induced more severe peripheral neuropathy in db/db mice

We collected fecal samples from the DM patients with DSPN (DSPN, n = 27), and compared their gut microbiota with that in DM patients without DSPN (DM, n = 30) and subjects with normal glucose level (NG, n = 29) from our previous dataset (Deng et al., 2022; Fang et al., 2021) (Table S1). Based on the sequencing data of V3-V4 regions of 16S rRNA genes, the richness of gut microbiota in patients with DSPN was significantly higher than that in DM and NG groups (Figure S1A). Principal coordinate analysis (PCoA) based on Jaccard distance showed that the overall structure of gut microbiota in DSPN was different from that in DM and NG groups (Figure S1B).

Next, to evaluate whether the dysbiosis of gut microbiota contributes to the development of peripheral neuropathy, we transplanted the fecal microbiota from DSPN and NG group to the genetically diabetic mice respectively. After treatment with a cocktail of antibiotics for 2 weeks, the db/db mice at 13-weeks old were gavaged with the fecal microbiota either from DSPN (M-DSPN, n = 6) or NG group (M-NG, n = 6). The structure of the gut microbiota in the M-DSPN mice was significantly different from that in M-NG mice (P = 0.004 by PerMANOVA test with 999 permutations), and the gut microbiota of the recipient mice was more similar to that of the corresponding human donors (Figure S2). We evaluated the severity of peripheral neuropathy in the mice at the end of third week after FMT. Compared to the M-NG group, the M-DSPN mice showed higher mechanical and thermal sensitivity, which were reflected by the level of 50% threshold and thermal latency, and lower MNCV of sciatic nerve (Figure 1A). The M-DSPN mice also had significant lower intraepidermal nerve fiber density (IENFD) of the posterior plantar skin by immunohistochemistry analysis, which was used to assess the neuropathy in the small fiber (Figure 1B). In the peripheral nervous systems, neurofilament 200 (NF200) was a special marker for A-fibers, myelin basic protein (MBP) was associated with the formation and maintenance of myelin, and brain derived neurotrophic factor (BDNF) was required for survival, proliferation, migration and differentiation of neurons and neurogliocytes (Fan et al., 2020; Zhang et al., 2019) (McGregor and English, 2018). Previous studies have found that NF200, MBP and BDNF decreased significantly in dorsal root ganglion (DRG) and sciatic nerve when peripheral nerve injured by diabetes (Fan et al., 2020; Zhang et al., 2019).The levels of these three proteins were significantly lower in M-DSPN mice than M-NG group both in DRG and sciatic nerve (Figure 1C and D). The above results suggest that the gut microbiota from patients with DSPN aggravates peripheral neuropathy in db/db mice.

• Download figure
• Open in new tab

Figure 1 Gut microbiota from patients with DSPN induced more severe peripheral neuropathy than that from the subjects with normal glucose level in db/db mice.

(A) Mechanical sensitivity, thermal sensitivity and motor nerve conduction velocity (MNCV) of sciatic nerve. Immunohistochemical staining of the posterior plantar skin and integrated optical density (IOD) analysis for (B) Intraepidermal nerve fibre density (IENFD, marked by PGP9.5). Immunohistochemical staining of (C) dorsal root ganglion and (D) sciatic nerve and IOD analysis for Neurofilament 200 (NF200), myelin basic protein (MBP) and Brain derived neurotrophic factor (BDNF). (E) Immunohistochemical staining of clone and IOD analysis for tight junction proteins ZO-1, Claudin-1 and Claudin-4. (F) Plasma levels of LBP, IL-6 and TNF-α. Data are presented as the mean ± s.e.m. The comparison between the two groups was tested by Student’s t-test (two-tailed). * P < 0.05, ** P < 0.01 and *** P < 0.001. M-DSPN indicates db/db mice received fecal microbiota from patients with DSPN, M-NG indicates db/db mice received fecal microbiota from the subjects with normal glucose level. For A, M-DSPN group, n = 6; M-NG group, n = 6; For B and F, M-DSPN group, n = 4; M-NG group, n = 4; For C, M-DSPN group, n = 3; M-NG group, n = 3; For D and E, M-DSPN group, n = 3; M-NG group, n = 4. For B and C, scale bars indicate 20 μm; For D and F, scale bars indicate 50 μm.

We also assessed the effects of gut microbiota from patients with DSPN on the gut barrier function. The immunofluorescence staining showed that the expression of the tight junction proteins ZO-1, Claudin-1 and Claudin-4 in the colonic mucosal biopsy specimens were significantly lower in M-DSPN mice than that in M-NG group (Figure 1E), indicating worse gut barrier function in M-DSPN mice. Then we tested the serum level of lipopolysaccharide (LPS)-binding protein (LBP), which can bind to antigens such as endotoxin produced by bacteria and represent a surrogate biomarker that links bacterial antigen load in the blood and the host inflammatory response. Notably, the level of LBP was significantly higher in the M-DSPN mice than that in M-NG group (Figure 1F). We assessed the inflammatory status of the mice. Indeed, the serum levels of TNF-α and IL-6, which are systemic inflammation biomarkers that have been associated with DSPN progression (Feldman et al., 2017; Herder et al., 2017), were higher in the M-DSPN mice than that in M-NG group (Figure 1F). These results suggest that gut barrier dysfunction, higher antigen load and severer systemic inflammation may contribute to the aggravation of peripheral neuropathy by the gut microbiota from patients with DSPN.

FMT alleviated the severity of peripheral neuropathy in patients with DSPN

As the dysbiosis of gut microbiota contributed to the development of peripheral neuropathy, we would like to find out whether improvement of gut microbiota can alleviate DSPN in patients. Thirty-seven DSPN patients were randomized into either FMT or placebo group in a 2:1 ratio. Before enrolment into this trial, all patients showed no improvement in their DSPN symptoms after at least 84 days of conventional treatments (lifestyle modification, glucose control and drug intervention, medical details in Table S2). After the run-in period, these patients were transplanted with transplants made from fecal microbiota of the healthy donors or placebo and then were followed up for 84 days. Finally, 22 patients in the FMT group and 10 patients in the placebo group completed this RCT study (Figure S3). At baseline (0D), demographic and anthropometric characteristics and glycemic control were similar between the two groups (Table S3). In particular, there were no significant differences in the level of Toronto Clinical Scoring System (TCSS (Arumugam et al., 2016)) score and the sensory and motor nerve functional status of DSPN (Figure 2 and Figure S4). At 84 days after transplantation (84D), glucose and lipid metabolism, blood pressure and body weight remained unchanged in both groups (Table S3). The insulin requirement was similar between the two groups (Table S3) and no serious adverse events were documented during the trial (Table S4).

• Download figure
• Open in new tab

Figure 2 FMT alleviated the severity of peripheral neuropathy, anxiety, depression, and improved sleep and life quality in patients with DSPN.

(A) Toronto clinical scoring system (TCSS) score. (B) Visual analogue scale (VAS) score. (C) Hamilton anxiety scale (HAMA) score; (D) Hamilton depression rating scale (HAMD) score. (E) Pittsburgh sleep quality index (PSQI) score. (F) World health organization’s quality of life (WHOQOL)-BREF score. 0D indicates baseline, and 84D indicates 84 days after FMT. In the box plots, the line in the middle of the box is plotted at the median, and the inferior and superior limits of the box correspond to the 25^th and 75^th percentiles, respectively. The whiskers correspond to the 10^th and 90^th percentiles, and outliers are denoted. Mann-Whitney U test was used to analyze differences between FMT and Placebo. * P <0.05. Wilcoxon matched-pairs signed rank test was used to analyze each pairwise comparison 0D and 84D within each group. ^# P<0.05 and ^## P<0.01. The bars represent the mean change from the baseline value per group, with the corresponding s.e.m.. Mann-Whitney U test was used to analyze differences in the changes between the FMT and placebo groups (intergroup changes). ^†P < 0.05 and ^††P < 0.01. FMT group, n = 22; Placebo group, n = 10.

TCSS score is used to evaluate the neuropathic symptoms and signs of patients, as the standard to assess the severity of DSPN (Arumugam et al., 2016). We used the change of TCSS score at 84D compared with baseline as the primary outcome and the results showed that the reduction of TCSS score was significantly greater in the FMT group than that in the placebo group (Figure 2A). Moreover, the levels of TCSS at 84D for the patients subjected to FMT were significantly lower than baseline, but no changes were observed in the placebo group, and the levels of TCSS in the FMT group were significantly lower than that in the placebo group at 84D (Figure 2A). We also evaluated neuropathic pain of the patients with visual analogue scale (VAS (Petersen et al., 2021)) score. Compared to baseline, the VAS scores also decreased significantly after FMT, but not in placebo group. The levels of VAS were significantly lower than that in the placebo group at 84D (Figure 2B). Notably, fifteen patients in the FMT group, who suffered from moderate/severe neuropathic pain (VAS score ⩾4) at baseline and eight of them (53.3%) had more than 50% pain relief at 84D as compared with baseline, which was considered a good outcome (Petersen et al., 2021). While only 14.29% patients in placebo group showed more than 50% pain relief at 84D, based on the VAS, which was significantly lower than that in FMT group. These results indicate that FMT induced stable relief of the symptoms and signs, particularly the neuropathic pain, in the patients with DSPN.

As the neuropathic symptoms severely induce anxiety, depression, sleep disorders and reduce the quality of life in DSPN patients (Gylfadottir et al., 2020), we assessed the effects of FMT on anxiety, depression, sleep quality and overall quality of life by using the Hamilton Anxiety Scale (HAMA (Zhao et al., 2021)), Hamilton Depression Rating Scale (HAMD (Zhao et al., 2021)), Pittsburgh Sleep Quality Index (PSQI (Buysse et al., 1989)) and the Brief table of the World Health Organization’s Quality of Life (WHOQOL-BREF (Saxena et al., 2001)). After 84 days of transplantation, HAMA and HAMD score showed that the anxiety and depression status were significantly improved in the FMT group but not in the placebo group (Figure 2C and D). PSQI assessment showed that sleep quality was significantly improved in patients subjected to FMT at 84D but not in those that received the placebo (Figure 2E). The WHOQOL-BREF score significantly increased in both FMT and placebo groups at 84D, suggesting improved quality of life for all the patients (Figure 2F).

We used electrophysiological measurements (nerve conduction velocity (NCV) and current perception threshold (CPT)) to objectively assess whether FMT improve peripheral neuropathy (Dyck et al., 1997). At 84 days after transplantation, the sensory nerve conduction velocities (SNCVs) of sural nerves and ulnar nerves were significantly increased only in patients of the FMT group and the changes of SNCVs were significantly greater in the FMT group than in the placebo group (Figure S4A-C). The motor nerve conduction velocities (MNCVs) of the distal median nerves, posterior tibial nerves and common peroneal nerve were significantly improved in patients of the FMT group but not in placebo (Figure S4D-G). CPT can be used to visualize the sensitivity of different types of never fibers to electrical stimulation, including the thick myelinated (Aβ) fibers (2000Hz), thin myelinated (Aδ) fibers (250Hz), and unmyelinated (C) fibers (5Hz) (Lv et al., 2015). For superficial and deep peroneal nerve, the CPT level at 5 Hz was significantly decreased in FMT group after 84 days, and significantly lower in FMT group than that in placebo group, indicating that the hypoesthesia of C-fibers in legs of FMT group improved (Figure S4H-J). For distal median nerve, the CPT levels at 2000Hz, 250Hz and 5Hz all decreased significantly in FMT group but not in placebo after 84 days, and the changes at 250Hz and 5Hz were significantly greater in the FMT group than in the placebo group, suggesting that FMT may influence on Aβ, Aδ and C nerve fibers of distal upper limbs (Figure S4K-M). This suggests that introduction of gut microbiota from healthy donors improved the electrophysiological functions of peripheral nerves in patients with DSPN.

After finishing the RCT study, to further verify the alleviation of neuropathic symptoms and signs in patients with DSPN by FMT, we transplanted the transplants made from fecal microbiota of the healthy donors to all of the ten patients in placebo group (post-RCT study). Compared to the baseline or the end of the RCT study, these patients showed significantly improvement of the neuropathic symptoms and signs, anxiety status, sleep quality, overall quality of life, and electrophysiological functions of peripheral nerves after 84 days of transplantation with the transplants from healthy donors (Figure S5 and S6).

FMT improved gut barrier integrity and systemic inflammatory status in patients with DSPN

As we had found that the dysbiosis gut microbiota from the patients with DSPN made damage of gut barrier in the aforementioned db/db mice study, in order to find out whether FMT, the gut microbiota – targeted treatment, could improve gut barrier function and decrease chronic inflammation of the patients, we collected biopsy specimens of colon from part of our patients at baseline and 84 days after transplantation in RCT study. The immunofluorescence staining showed that the expression of the tight junction proteins ZO-1 and Claudin-1 in the colonic mucosa of biopsy specimens were significantly higher in FMT than that in placebo group at 84D (Figure 3A and B). In addition, Claudin-4 was promoted in specimens from patients in both FMT and placebo group, but the changes were significantly greater in the FMT group (Figure 3C).

• Download figure
• Open in new tab

Figure 3 FMT improved gut barrier integrity and systemic inflammatory status in patients with DSPN.

Representative immunofluorescence staining of the tight junction proteins (A) ZO-1, (B) Claudin-1 and (C) Claudin-4 in clone. Plasma levels of (D) LBP, (E) IL-6 and (F) TNF-α. Scale bars indicate 20 μm. 0D indicates baseline, 84D indicates 84 days after FMT. In the box plots, the line in the middle of the box is plotted at the median, and the inferior and superior limits of the box correspond to the 25^th and 75^th percentiles, respectively. The whiskers correspond to the 10^th and 90^th percentiles, and outliers are denoted. Student’s t-test (two-tailed) (A, B and C) and Mann-Whitney U test (D, E and F) were used to analyze differences between FMT and Placebo. * P < 0.05. Paired t-test (two-tailed) (A, B and C) and Wilcoxon matched-pairs signed rank test (D, E and F) was used to analyze each pairwise comparison 0D and 84D within each group. ^# P < 0.05 and ^## P < 0.01. The bars represent the mean change from baseline per group, with the corresponding s.e.m. Student’s t-test (two-tailed) (A, B and C) and Mann-Whitney U test (D, E and F) were used to analyze differences in the changes between the FMT and placebo groups (intergroup changes). ^† P < 0.05 and ^†† P < 0.01. For A and C, FMT group, n = 6; placebo group, n = 5. For B, FMT group, n = 4; placebo group, n = 5. For D, E and F, FMT group, n = 22; placebo group, n = 10.

Because the improvement of the gut barrier integrity would decrease the serum antigen load from microbiota, we tested the serum level of LBP, and found that the level of LBP was significantly reduced in the FMT group but not in the placebo group (Figure 3D). Then we assessed the inflammatory status of patients with DSPN after FMT. Indeed, the serum levels of IL-6 and TNF-α decreased in the FMT group at 84D but did not change in the placebo group (Figure 3E and F). This suggests that FMT improved the gut barrier function, reduced serum antigen load and aforementioned alleviation of chronic inflammation, which may contribute to alleviate DSPN in patients.

FMT induced overall structural changes in the gut microbiota in patients with DSPN

To further explore the role of gut microbiota in the beneficial effect on DSPN alleviation via FMT, we did shotgun metagenomic sequencing of all the transplants made from gut microbiota of the donors, and the fecal samples from all the patients in the RCT study and post-RCT study before transplantation (0D), 3 days (3D), 28 days (28D), 56 days (56D), and 84 days (84D) after transplantation (Table S5). To characterize the strain-level changes in the gut microbiota in patients following FMT, we de novo assembled 1,572 high-quality draft genomes from the metagenomic dataset (details in Materials and Methods, Table S6 and S7). After integrating the genomes from Human Gastrointestinal Bacteria Genome Collection (HGG) (Forster et al., 2019) to improve the metagenomic analysis, 1,999 non-redundant high-quality metagenome-assembled genomes (HQMAGs) were obtained and used in the further analysis (Table S8). Transplantation with placebo showed no significant alteration in the gut microbiota in the patients during the RCT study (Figure S7). Then samples from the FMT group in RCT study and sample from the post-RCT study together with transplants from healthy donors were analyzed to reveal the alteration in gut microbiota during FMT. To determine probable microbial origins in longitudinal samples after FMT, unique HQMAGs detectable in pre-FMT patient (0D) samples and FMT transplants were compared (Figure 4A). The relative abundance of HQMAGs only detected in the FMT transplants only accounted for 48.1% of the HQMAGs shared between the 0D samples of patients and corresponding transplants. Transplant-derived HQMAGs accounted for a considerable proportion of the population from day 3 to day 84 after FMT (3D to 84D). The distance between the patient’s gut microbiota and that of corresponding transplants was significantly decreased after FMT, although the gut microbiota structures in the patients were still closer to each individual’s own baseline composition than to the transplants (Figure 4B). There was a significant increase in richness of the gut microbiota at 28D after FMT, but the diversity measured by Shannon index was not altered (Figure 4C). Subject adjusted principal coordinate analysis (aPCoA) of Jaccard distances at the HQMAGs level demonstrated that the gut microbiome of patients significantly shifted after 3 days of FMT (3D). The gut microbiome further significantly changed after 28 days (28D) and remained relative stable afterward (56D-84D) (Figure 4D). These results suggest that FMT induced early and significant alterations in the gut microbiota of DSPN patients, which were concomitant with significant and stable improvement of neuropathic symptoms and signs.

• Download figure
• Open in new tab

Figure 4 FMT induced overall structural changes in the gut microbiota in patients with DSPN.

(A) Strains from samples were classified by origin (unique to patient at baseline, unique to transplant, present in both patient baseline and transplant, or absent in both patient baseline and transplant). (B) Jensen-Shannon distance (JSD) of the gut microbiotas. JSD vs. donor: JSD between a sample at the particular time point and that of the transplant; JSD vs. T0: JSD between a sample at the particular time point and the baseline sample. Data are presented as the mean ± s.e.m. Wilcoxon matched -pair signed-rank tests were used to analyze each pairwise comparison. *P < 0.05 and **P < 0.01. (C) Richness (number of strains) and diversity (Shannon index) of the gut microbiota. Data are presented as the mean ± s.e.m. Wilcoxon matched-pair signed-rank tests were used to analyze each pairwise comparison. *P < 0.05 and **P < 0.01. (D) Subject adjusted principal coordinate analysis (aPCoA) of Jaccard distances at the strain level. The lower triangular heat map shows the marginal PerMANOVA test based on Jaccard distance (strata in subject, 999 permutations, *P < 0.05 and **P < 0.01.). The box plots show the changes in the gut microbiota from different time points on PC1 or PC2 (the line in the middle of the box is plotted at the median, the inferior and superior limits of the box correspond to the 25th and 75th percentiles, the whiskers correspond to the 10^th and 90^th percentiles, and outliers are denoted). Wilcoxon matched-pair signed-rank tests were used to analyze each pairwise comparison. *P < 0.05, **P < 0.01 and ***P < 0.001. The samples from FMT group in RCT study and samples from Placebo group in the post-RCT study of transplantation with transplants from healthy donors were analyzed together. 0D indicates before transplantation (n = 32), 3D (n = 31), 28D (n = 29), 56D (n = 26) and 84D (n = 32) indicate 3 days, 28 days, 56 days and 84 days after transplantation with transplants from healthy donors, respectively.

The gut microbial genomes correlated with TCSS score were organized in two competing guilds

Next, we identified that TCSS scores were significantly correlated with 54 HQMAGs (Figure 5A and Table S9), among which 21 and 33 HQMAGs had negative and positive correlations, respectively, by using linear mixed effect model via MaAsLin2 (Mallick et al., 2021). Most of the HQMAGs (19 of 21) negatively correlated with TCSS score were belonged to Firmicutes, and many of them are potentially beneficial bacteria that can regulate mucosal barrier function to reduce gut permeability and/or modulate the host immune response, possibly via production of butyrate (Cani et al., 2022; Derrien et al., 2022; Gomes et al., 2018; Li et al., 2019; Maioli et al., 2021; Waters and Ley, 2019), such as Fecalibacterium prausnitzii, Agathobacter rectalis, Agathobaculum butyriciproducens and Anaerobutyricum hallii. Out of the 33 HQMAGs positively correlated with TCSS score, 26 were belonged to Bacteroidetes, in especial 17 from Bacteroides uniformis. As bacteria in the gut ecosystem interact with each other and work as coherent functional groups (a.k.a “guilds”) (Wu et al., 2021), we applied co-abundance analysis on these 54 HQMAGs to explore the interactions between them and to find potential guilds structure. The 54 HQMAGs were organized into two guilds-the 21 HQMAGs negatively correlated with TCSS score were positively interconnected with each other and formed as Guild 1, which was true for the 33 HQMAGs positively correlated with TCSS score as Guild 2 (Figure 5B). There were only negative correlations between the two guilds, suggesting a potentially competitive relationship between the two guilds. We found that the abundance of Guild 1 was almost equal to Guild 2 in the gut microbiota of transplants from healthy donors, but the abundance of Guild 2 was significantly higher than Guild 1 in the patients with DSPN at baseline (Figure 5C). Compared to their transplants, the gut microbiota of patients with DSPN contained significantly much more Guild 2 and slightly fewer Guild 1 (Figure 5C). Notably, FMT significantly decreased the abundance of Guild 2 and mildly increased the abundance of Guild 1 from 3D, and there was no significant difference between Guild 1 and 2 from 28D to 84D (Figure 5C).

• Download figure
• Open in new tab

Figure 5 The gut microbial genomes correlated with TCSS score were organized in two competing guilds.

(A) The high-quality metagenome-assembled genomes (HQMAGs) had significantly correlation with TCSS score. The orange bar indicates negative correlation and purple bar indicates positive correlation. (B) Co-abundance network of the HQMAGs reflects two competing guilds. The co-abundance correlation between the HQMAGs were calculated using repeated measures correlation. All significant correlations with BH-adjusted P < 0.05 were included. Edges between nodes represent correlations. Red and blue colors indicate positive and negative correlations, respectively. Node color indicates Guild 1 (orange) and Guild 2 (purple), respectively. (C) The abundance of Guild 1 and Guild 2 in transplants and fecal samples from patients. The samples from FMT group in RCT study and samples from Placebo group in the post-RCT study of transplantation with transplants from healthy donors were analyzed together. 0D indicates before transplantation (n = 32), 3D (n = 31), 28D (n = 29), 56D (n = 26) and 84D (n = 32) indicate 3 days, 28 days, 56 days, and 84 days after transplantation with transplants from healthy donors, respectively. Data are presented as the mean ± s.e.m. Wilcoxon matched-pair signed-rank tests (two-sided) were used to compare differences of abundance between Guild 1 and 2 at the same timepoint, ***P < 0.001. Mann-Whitney tests were used to compare transplants and recipients, ^# P < 0.05 and ^## P < 0.01. (D) Functional genes in the HQMAGs significantly correlated with TCSS score. The encode key enzymes in the propionate, butyrate and indole production, lipid A biosynthesis pathways and antibiotic resistant genes are shown in the right part of the cycle. The HQMAGs that were significantly corelated with TCSS score and whose genomes contain these genes are shown in the left part of the cycle. The linkage between strains and genes indicates that the genome of the strain contained the gene. The orange line indicates that the HQMAGs negatively corelated with TCSS score, whereas the purple line indicates that the HQMAGs positively corelated with TCSS score pst: propionyl-CoA:succinate-CoA transferase; pct: propionate CoA transferase; but: butyryl–coenzyme A (butyryl-CoA):acetate CoA transferase; tnaA: tryptophanase; arnT: 4-amino-4-deoxy-L-arabinose transferase; lpxA: UDP-N-acetylglucosamine acyltransferase; lpxB: lipid A disaccharide synthase; lpxCf (lpxC-fabZ): bifunctional enzyme LpxC/FabZ; lpxD: UDP-3-O-(3-hydroxymyristoyl)glucosamine N-acyltransferase; lpxH: UDP-2,3-diacylglucosamine hydrolase; lpxK: tetraacyldisaccharide 4’-kinase; lpxL: lipid A biosynthesis lauroyltransferase; ARGs: antibiotic resistant genes.

Then we studied the genomic features of the two guilds to understand the potential functions underlying their associations with the alleviation of DSPN (Table S9). The SCFAs, derived from gut bacterial fermentation, have been considered as the most important bacterial metabolites affecting gut permeability and regulate intestinal and systemic inflammation in humans (Koh et al., 2016). We used the abundances of genes that encode key enzymes to indicate changes in the SCFAs genetically producing capacity (including acetic acid, propionate, and butyric acid) production pathways (Claesson et al., 2012). Most of the 54 HQMAGs carried fhs gene involved in the acetic acid biosynthesis pathway, regardless of the Guilds they belonged to (Table S9). For genes encoding the key enzyme propionyl-CoA:succinate-CoA transferase and propionate CoA transferase (pst and pct) in the propionate synthetic pathway, which has been reported to be negatively associated with glycemic control (Sanna et al., 2019; Tirosh et al., 2019), significantly much more HQMAGs in Guild 2 than that in Guild 1 harbored pst or pct (pst: 14.29 in Guild 1 vs 78.79% in Guild2, Fisher’s exact test P = 4.01×10 ^-6 ; pct: 23.81% in Guild 1 vs 78.79% in Guild 2, Fisher’s exact test P = 1.59×10 ^-4) (Figure 5D and Table S9). Conversely, only the HQMAGs in Guild 1 had but and buk genes, which are the predominate terminal genes for the butyrate biosynthetic pathways in gut microbiota (Figure 5D and Table S9). Moreover, we also assessed the change in endotoxin (LPS) biosynthesis, which is the major antigen from the gut microbiome. Lipid A, the amphipathic LPS glycolipid moiety, stimulates the immune system by tightly binding to Toll-like receptor 4 (Zhang et al., 2021). We showed that most of the HQMAGs in Guild 2 (78.79%) carried the genes related to Lipid A biosynthesis, but only two HQMAG in Guild 1 had these genes (Figure 5D and Table S9). In terms of antibiotic resistance genes (ARGs), 8 HQMAG s in Guild 1 encoded 14 ARGs and 29 genomes in Guild 2 encoded 112 ARGs (Figure 5D and Table S9). Taken together, FMT induced increase of Guild 1 that had the beneficial capacity for butyrate production and decrease of Guild 2 that could produce more antigen load, which contribute to the reduced low-grade, systemic and chronic inflammation.

Matched enterotype between FMT transplants and recipients linked to better improvement of DSPN

In the FMT studies, it is always an interesting question whether the resemblance between transplants of gut microbiota from the healthy donors and/or the original gut microbiota in the recipients could be linked to the therapeutic efficacy. Here, we clustered all the samples, including 27 transplants and 150 samples from patients during FMT, into two enterotypes (C1 and C2) based on Jaccard distance of their microbiota at HQMAGs-level (Figure 6A). There were 19 transplants and 17 original gut microbiota in the recipients (samples of the patients at 0D) classed to C1, and 8 transplants and 15 original gut microbiota in the recipients classed to C2 (Figure 6B). Firstly, in order to test whether the specific transplant enterotype can affect therapeutic efficacy of FMT, we subjected the patients in two subgroups based on enterotypes of the transplants they received. TCSS scores in these two subgroups both showed a significant reduction at 84D after FMT, but there was no significant difference between the two subgroups (Figure S8). Next, we resubjected the patients into transplant-recipient matched or unmatched groups based on whether their original gut microbiota and transplants belonged to the same enterotype. Notably, the abundance of Guild 1 that negatively correlated with TCSS score, was significantly higher in the transplant-recipient matched than that in unmatched group from 56D to 84D after FMT (Figure 6C). Meanwhile, the abundance of Guild 2 that positively correlated with TCSS score, only significantly decreased in transplant-recipient matched group at 3D and kept on lower level from 28D to 84D after FMT, which was significantly lower than that in unmatched group (Figure 6C). Moreover, transplant-recipient matched group had significantly lower TCSS score than unmatched group at 84D (Figure 6D). These results suggest that matched enterotype between transplants and recipients in FMT may introduce better alteration in gut microbiota of the patients and induce better improvement of the symptoms and signs of DSPN.

• Download figure
• Open in new tab

Figure 6 Matched enterotype between FMT transplants and recipients linked to better improvement of DSPN.

(A) Classification of two enterotypes among all samples based on Jaccard distance and complete linkage. The samples included all the transplants, FMT group in RCT study and Placebo group in the post-RCT study of transplantation with fecal microbiota of healthy donors. Calinski-Harabasz (CH) index was used to assess the optimal number of clusters. (B) The enterotype of baseline gut microbiota of recipients paired with their transplants. (C) The abundance of Guild 1 and Guild 2 in matched and unmatched groups. The samples from FMT group in RCT study and samples from Placebo group in the following trial of transplantation with transplants from healthy donors were analyzed together. Matched group indicates the baseline gut microbiota of recipients and their transplants belong to the same enterotype (n = 24). unmatched group indicates the baseline gut microbiota of recipients and their transplants belong to the different enterotypes (n = 8). 0D indicates before transplantation, 3D, 28D, 56D and 84D indicate 3 days, 28 days, 56 days and 84 days after transplantation with transplants from healthy donors, respectively. Data are presented as the mean ± s.e.m. Mann-Whitney U test was used to analyze differences of abundance of Guild 1 or 2 between matched and unmatched groups, *P < 0.05. Friedman test followed by Nemenyi post-hoc test was used to compare the different timepoints within the same group, and values of each group with same letters are not significantly different (P > 0.05). (D) Changes of TCSS score in matched and unmatched groups. In the box plots, the line in the middle of the box is plotted at the median, and the inferior and superior limits of the box correspond to the 25^th and 75^th percentiles, respectively. The whiskers correspond to the 10^th and 90^th percentiles, and outliers are denoted. Mann-Whitney U test was used to analyze differences between matched and unmatched groups. * P < 0.05. Wilcoxon matched-pairs signed rank test was used to analyze each pairwise comparison 0D and 84D within each group. ^# P < 0.05 and ^## P < 0.01. The bars represent the mean change from the baseline value per group, with the corresponding s.e.m. Mann-Whitney U test was used to analyze differences in the changes between the matched and unmatched groups (intergroup changes).