Parabacteroides distasonis ameliorates insulin resistance via activation of intestinal GPR109a

Gut microbiota plays a key role in insulin resistance (IR). Here we perform a case-control study of Chinese adults (ChiCTR2200065715) and identify that Parabacteroides distasonis is inversely correlated with IR. Treatment with P. distasonis improves IR, strengthens intestinal integrity, and reduces systemic inflammation in mice. We further demonstrate that P. distasonis-derived nicotinic acid (NA) is a vital bioactive molecule that fortifies intestinal barrier function via activating intestinal G-protein-coupled receptor 109a (GPR109a), leading to ameliorating IR. We also conduct a bioactive dietary fiber screening to induce P. distasonis growth. Dendrobium officinale polysaccharide (DOP) shows favorable growth-promoting effects on P. distasonis and protects against IR in mice simultaneously. Finally, the reduced P. distasonis and NA levels were also validated in another human type 2 diabetes mellitus cohort. These findings reveal the unique mechanisms of P. distasonis on IR and provide viable strategies for the treatment and prevention of IR by bioactive dietary fiber.


Supplemental Figures and Legends
Source data can be found in Source Data file.and TC of (g), Kruskal-Wallis test for (f) and (i).Source data can be found in Source Data file.(%).i-j OGTT (i) and AUC (j).k-l ITT (k) and AUC (l).m Fasting insulin level.n HOMA-IR.o Serum TNF-α.p Serum IL-1β.q Serum LPS.r Relative mRNA expression of genes related to intestinal permeability in colon.s Relative mRNA expression of Gpr109a in colon.t Relative mRNA expression of genes related to inflammation in liver.Data are presented as the mean ± SEM.Statistical analysis was performed using One-way ANOVA with Tukey's post hoc test for (b), (f), (h), (i), (k), (l), (m), (n), (o), (p), (q), Zo1 of (r), (s), and (t), One-way ANOVA with Dunnett's T3 post hoc test for (c), (j), Claudin1, and Occludin of (r), Kruskal-Wallis test for (d), (g) and Muc2 of (r).*, P < 0.05; **, P < 0.01.Source data can be found in Source Data file.Statistical analysis was performed using two-tailed unpaired t-test for data marked with * , two-tailed Mann-Whitney test for data marked with § , and two-tailed Chi-Squared test for data marked with † .€ Some patients take more than one antidiabetic drug.

Fig. 1
Gut microbiota profiling in patients with T2DM a-b Relative abundance of bacteria at the phylum (a) and genus (b) level.c Gut microbial co-occurrence network analysis indicated Parabacteroides and Faecalibacterium are core genus (average relative abundance > 0.1%) in gut microbiota of humans (T2DM and HC groups).The red line indicates Spearman's rank (two-tailed Spearman's rank test) correlation coefficient > 0.30 and P < 0.05; the blue line indicates Spearman's rank correlation coefficient < -0.30 and P < 0.05.T2DM (type 2 diabetes mellitus); HC (health control).Source data can be found in Source Data file.

7 A F 1 2 LSupplementary Fig. 3
b a c te ro id e s R o s e b u ri a D o re a C o p ro c o c c u s O s c il lo s p ir a [R u m in o c o c c u s ] B a c te ro id e s U n c la s s if ie d _ L a c h n o s p ir a c e a e B il o p h il a R u m in o c o c c u s D e s u lf o v ib ri o U n c la s s if ie d _ H e li c o b a c te ra c e a e U n c la s s if ie d _ S 2 4a c to c o c c u s U n c la s s if ie d _ R u m in o c o c c a c e a e A ll o b a c u lu m L a c to b a c il lu s [P re v o te ll a ] Effects of DOP on HFD-induced IR mice After an 8-week HFD treatment, mice were given PBS (HFD group) or DOP (DOP group) for 5 weeks, and the control group (Ctrl) was fed with a chow diet and given an equivalent volume of PBS.(a-g) n = 8 mice per group, (h-o) n = 6 mice per group.a Food intake.b Liver weight.c Epididymal fat weight.d Epididymal fat/body weight (%).e Serum free fatty acids.f Serum triglyceride.g Serum cholesterol.h Representative images of the appearance of liver, liver sections after H&E staining, oil red O staining, and F4/80 immunohistologic staining.Scale bars, 50 μm.i Relative mRNA expression of genes related to inflammation in liver.j-k Relative abundance of bacteria at the phylum (j) and genus (k) level in HFD and DOP groups.l-n α-diversity indexes in HFD and DOP group as indicated by the Shannon, ACE, and Chao1 indexes.The box plots showing the minima, maxima, centre, bounds of box and whiskers and percentile.o Principal coordinate analysis (PCoA) of the gut microbiota between the HFD and DOP groups by weighted UniFrac distance (ANOSIM test).p Taxonomic cladogram generated from LEfSe analysis.Each circle's size is proportional to the taxon's abundance.q Spearman correlations (two-tailed Spearman's rank test) between the top 20 genera and IR phenotypes.The color represents positive (red) or negative (blue) correlations and FDRs are denoted.*, FDR < 0.05; **, FDR < 0.01.DOP (Dendrobium officinale polysaccharide); HFD (High-fat diet).Data (a-g, i) are presented as the mean ± SEM.Statistical analysis was performed using One-way ANOVA with Tukey's post hoc test for (e), (f), and (i), One-way ANOVA with Dunnett's T3 post hoc test for (c) and (g), Kruskal-Wallis test for (a) and (b).Source data can be found in Source Data file.Supplementary Fig. 4 DOP ameliorates IR in a gut microbiota-dependent manner Mice were treated with antibiotic cocktails (Abx) for 7 days to deplete gut microbiota, followed by DOP administration.n = 8 mice per group.a Experimental scheme for b to o. b Body weight curve.c Body weight change (%).d Food intake.e-f OGTT (e) and AUC (f).g-h ITT (g) and AUC (h).i Serum insulin.j HOMA-IR.k Serum LPS.l Serum free fatty acids.m Liver weight.n Epididymal fat weight.o Epididymal fat/body weight (%).DOP (Dendrobium officinale polysaccharide).Data are presented as the mean ± SEM.

Supplementary Fig. 7
Nicotinic acid is a bioactive metabolite of P. distasonis a The MS/MS spectra of NAM standard and NAM detected in P. distasonis-incubated samples.b Extracted ion chromatograms of NAM from cultured P. distasonis compared to the Vehicle.c The level of NAM in the culture supernatant of P. distasonis at different time points, n = 3 biologically independent samples per group.d The change of fecal NAM after gavage with P. distasonis.n = 12 mice per group.e The intensity Bacteroides fragilis Daily gavage with P. distasonis NSP007 Daily gavage with Bacteroides thetaiotaomicron Daily gavage with Klebsiella pneumoniae Daily gavage with Escherichia coli Daily gavage with Clostridium difficile the culture supernatant of 44 P. distasonis after 24 h of incubation.n = 3 biologically independent samples per group.f Colonic NAM level in Vehicle, LPD, and KPD group of mice.n = 8 mice per group.g Colonic NAM level in HFD and DOP group of mice.n = 8 mice per group.h Fecal NAM level in HC (n = 30) and T2DM (n = 30) group of humans.i Spearman correlations (two-tailed Spearman's rank test) between the colonic NA level and the severity of IR in mice.j Spearman correlations (two-tailed Spearman's rank test) between the fecal NA level and the severity of IR in humans.k The intensity of NA in the culture supernatant of different strains after 24 h of incubation.n = 3 biologically independent samples per group.l-n The changes of fecal NA before and after colonization by different strains.Mice were treated with bacteria strains (2×10 8 CFU/mice/d) for 7 days.n = 6 mice per group.(l) Experimental scheme for m to n, (m) Relative abundance of different bacteria strains in feces before and after treatment.(n) The change of fecal NA level before and after different bacteria strains treatment.DOP (Dendrobium officinale polysaccharide); HFD (High-fat diet); LPD (P.distasonis NSP007); KPD (heat-killed P. distasonis NSP007); T2DM (type 2 diabetes mellitus); HC (health control).Data are presented as the mean ± SEM.Source data can be found in Source Data file.for (c), (e), (i), and Claudin1, Muc2, and Occludin of (j), One-way ANOVA with Dunnett's T3 post hoc test for Zo1 of (j), Kruskal-Wallis test for (b).Source data can be found in Source Data file.Supplementary Fig.9 P. distasonis ameliorates IR in a GPR109a activation-dependent manner Mice were fed an HFD for 8 weeks, then treated with PBS, P. distasonis, MPN, and P. distasonis + MPN for a further five weeks.(b-q) n = 8 mice per group, (r-t) n = 6 mice per group.a Experimental scheme for b to t. b Relative abundance of P. distasonis in feces accessed by qPCR.c Body weight curve.d Body weight change (%).e Food intake.f Liver weight.g Epididymal fat weight.h Epididymal fat/body weight

2 Effect of DOP, IN, and BG on the composition of gut microbiota a
The changes of total carbohydrates content of medium (DOP, IN, and BG as the sole carbon) during in vitro fermentation with fecal microbiota from T2DM patients.n = 30 per group.b Relative abundance of bacteria at the phylum level in different groups.c-e α-diversity indexes in different groups as indicated by the Shannon, ACE, and Chao1 indexes.The box plots showing the minima, maxima, centre, bounds of box and whiskers and percentile.n = 30 per group.f Principal coordinate analysis (PCoA) of the bacteria in different groups by weighted UniFrac distance (ANOSIM test).n = 30 per group.g Taxonomic cladogram generated from LEfSe analysis.Each circle's size is proportional to the taxon's abundance.h LDA score represents the taxonomic data with significant differences between IN, BG, and Vehicle groups.Only LDA scores > 2 are shown.Green indicates enriched taxa in the IN group; red indicates enriched taxa in the BG group; blue indicates enriched taxa in the Vehicle group.i Gut microbial co-occurrence network analysis