Interactions between Roseburia intestinalis and diet modulate atherogenesis in a murine model

Abstract

Humans with metabolic and inflammatory diseases frequently harbour lower levels of butyrate-producing bacteria in their gut. However, it is not known whether variation in the levels of these organisms is causally linked with disease development and whether diet modifies the impact of these bacteria on health. Here we show that a prominent gut-associated butyrate-producing bacterial genus (Roseburia) is inversely correlated with atherosclerotic lesion development in a genetically diverse mouse population. We use germ-free apolipoprotein E-deficient mice colonized with synthetic microbial communities that differ in their capacity to generate butyrate to demonstrate that Roseburia intestinalis interacts with dietary plant polysaccharides to: impact gene expression in the intestine, directing metabolism away from glycolysis and toward fatty acid utilization; lower systemic inflammation; and ameliorate atherosclerosis. Furthermore, intestinal administration of butyrate reduces endotoxaemia and atherosclerosis development. Together, our results illustrate how modifiable diet-by-microbiota interactions impact cardiovascular disease, and suggest that interventions aimed at increasing the representation of butyrate-producing bacteria may provide protection against atherosclerosis.

Figure 1.. Colonization with R. intestinalis increases cecal levels of short chain fatty acids (SCFAs) in mice fed a high plant polysaccharide (HPP) diet.

a, Experimental design. b, COPRO-Seq (community profiling by sequencing) analysis of fecal samples from gnotobiotic Apolipoprotein E knock-out mice colonized with the “core” community (n=5) or the “core plus R. intestinalis” community (n=5). The bar charts show the abundance of each species in each community. c,d, Levels of SCFAs in cecum (c, n=11 in the “core” group and n=10 in the “core plus R. intestinalis” group) and plasma (d, n=5 per group) in the “core” group and the “core plus R. intestinalis” colonized mice assessed by gas chromatography-mass spectrometry. Cecal SCFA levels are indicated by micromoles per gram wet weight (wt). Symbols represent values for individual animals, with averages (horizontal line) plus standard errors of the means indicated. Significance was calculated by an unpaired two-tailed Student’s t test.

Figure 2.. Colonization with R. intestinalis inhibits the development of atherosclerosis in mice fed a HPP diet.

a,e,f, Representative photographs of Oil Red O staining (a) and quantitative analysis of plaque area (e) and Oil Red O (ORO) positive area (f) in the aortic sinus (n=11 in the “core” group and n=10 in the “core plus R. intestinalis” group). b-d,g-i, Representative sections and quantitative analyses of MOMA-2 positive macrophages (b,g), collagen (c,h), and α-actin⁺ smooth muscle (α-SMA) cells (d,i) in the aortic sinus (n=6 per group). Box areas are enlarged below (a-d). Symbols represent values for individual animals, with averages (horizontal line) plus standard errors of the means indicated. Significance was calculated by an unpaired two-tailed Student’s t test. j, Gene expression of Tnf-α, Il-6, Il-1β, Mcp1, Icam1, and Vcam1 in the aorta from mice colonized with the core” community (n=6) or the “core plus R. intestinalis” community (n=5). The data are expressed as box-and-whisker plots where boxes represent median values and interquartile ranges and whiskers represent minimum and maximum values. Significance was calculated by an unpaired two-tailed Student’s t test, followed by the Benjamini-Hochberg procedure to control the false discovery rate. Gene expression of Tnf-α and Vcam1 showed statistical significance (q value of<0.10).

Figure 3.. R. intestinalis affects histone post-translation modifications (PTMs) in the colon.

Relative abundance of histone PTMs on H3 and H3.3. a-e, K4 methylation on H3 (a), K18 and K23 acetylation and methylation on H3 (b), K27 and K36 methylation on H3 (c) and H3.3 (d), and K79 methylation on H3 (e) in the proximal colon. Values are reported as a fold change in “core plus R. intestinalis” colon relative to “core” control (n=4 in the “core” group and n=5 in the “core plus R. intestinalis” group) for both diets. Significance was calculated by a two-tailed Welch’s T-test as follows: *, P value of <0.05; **, P value of <0.01; ***, P value of <0.001. Actual P values are found in Supplementary Table 8. HPP and LPP indicate high plant polysaccharide and low plant polysaccharide, and ac, me, and un indicate acetylation, methylation, and unmodified, respectively. For example, H3:K4me3 indicates the trimethylation of lysine 4 of H3.

Figure 4.. Colonization with R. intestinalis regulates energy metabolism and improves intestinal barrier function.

a,b, Gene ontology (GO) term enrichments with Fisher’s Exact Tests for genes upregulated (a) and downregulated (b) from RNA-seq analysis in colon samples of HPP diet-fed mice colonized with the “core plus R. intestinalis” community (n=5) relative to the “core” community (n=3). c, Fluorescein isothiocyanate (FITC) -dextran assay in HPP diet-fed mice colonized with the “core” (n=4) or the “core plus R. intestinalis” (n=5). d, Plasma LPS levels (EU/ml) measured in mice colonized with the core” community (n=7) or the “core plus R. intestinalis” community (n=7). Significance was calculated by an unpaired two-tailed Student’s t test and the data are expressed as box-and-whisker plots where boxes represent median values and interquartile ranges and whiskers represent minimum and maximum values (c,d). e, Untargeted metabolome analysis of plasma (n=6 per group). Relative mass spectrometry scaled intensities of long chain fatty acids (LCFAs), β-hydroxybutyrate (β-OHB), and eicosanoids detected at significantly different levels between mice colonized with “core” vs “core plus R. intestinalis” community (all indicated differences significant at P<0.05). Significance was calculated by two-tailed Welch’s T-tests, followed by the Benjamini-Hochberg procedure to control the false discovery rate, and the data are expressed as box-and-whisker plots where boxes represent median values and interquartile ranges and whiskers represent minimum and maximum values. 12-HETE and 15-KETE indicate 12-hydroxyeicosatetraenoic acid and 15-ketoeicosatetraenoic acid, respectively.

Figure 5.. Tributyrin supplementation reduces the development of atherosclerosis in mice colonized with non-butyrate producers.

a, Experimental design. b, COPRO-Seq (community profiling by sequencing) analysis of fecal samples from gnotobiotic ApoE^−/− mice fed with the HPP diet (n=8) or the 6% tributyrin (TB)-supplemented diet (n=9). The bar charts show the abundance of each species in the community. c,e,f, Representative photographs of Oil Red O staining (c) and quantitative analysis of plaque area (e) and Oil Red O (ORO) positive area (f) in the aortic sinus (n=8 in the HPP diet group and n=9 in the tributyrin group). d,g, Representative sections (d) and quantitative analysis (g) of MOMA-2 positive macrophages in the aortic sinus (n=8 in the HPP diet group and n=9 in the tributyrin group). Box areas are enlarged below (c,d). h, Fluorescein isothiocyanate (FITC) -dextran assay in mice fed the HPP diet (n=5) or the 6% tributyrin-supplemented diet (n=5). i, Plasma LPS levels (EU/ml) measured in mice fed the HPP diet (n=8) or the 6% tributyrin-supplemented diet (n=9). The data are expressed as individual dots with mean ± SEM (e-g) or box-and whisker plots where boxes represent median values and interquartile ranges and whiskers represent minimum and maximum values (h,i).