Shotgun metagenomic sequencing identified distinct gut microbial signatures and metabolic pathways associated with coronary artery disease, including enrichment of pro-inflammatory bacterial taxa and depletion of short-chain fatty acid-producing bacteria, according to a study published in mSystems.
Researchers analyzed fecal samples from 14 patients with coronary artery disease and 28 propensity score-matched healthy controls using metagenome-assembled genome (MAG) reconstruction to provide strain-resolved insights into microbial contributions to cardiovascular disease pathophysiology.
Taxonomic Alterations
Differential abundance analysis identified 15 bacterial species with significantly different relative abundances between patients with coronary artery disease (CAD) and controls. Seven species were enriched in patients with CAD; the most significantly enriched were CAG-303 sp000437755 (coefficient 3.262) and AM51-8 sp003478275 (coefficient 2.724).
Members of the Lachnospiraceae family, previously associated with trimethylamine-N-oxide production, were significantly enriched in patients with CAD. Conversely, eight bacterial species showed depletion in patients with CAD, including Slackia isoflavoniconvertens (coefficient −2.017) and Faecalibacterium prausnitzii (coefficient −2.473).
Metabolic Pathway Enrichment
Functional pathway analysis identified 10 MetaCyc metabolic pathways that differed significantly between groups. Seven metabolic pathways were enriched in patients with CAD, including the urea cycle, L-citrulline biosynthesis, glycolysis III (from glucose), and CDP-diacylglycerol biosynthesis. Three pathways—L-isoleucine biosynthesis III and the superpathway of branched-chain amino acid biosynthesis—were depleted in patients with CAD.
In the urea cycle pathway, Alistipes finegoldii and Flavonifractor plautii were primary contributors, whereas Alistipes onderdonkii was detected exclusively in patients with CAD. For L-citrulline biosynthesis, Alistipes, Escherichia, and Klebsiella were predominant contributors, with Escherichia especially enriched in patients with CAD.
Analysis of gut metabolic modules identified seven modules enriched in patients with CAD, including three amino acid degradation-associated modules (aspartate, serine, and arginine degradation) and four carbohydrate degradation-associated modules (galactose, lactose, xylose, and pectin degradation).
Metabolite Signatures and Predictive Modeling
Three metabolites differed significantly between patients with CAD and controls. Inosine was elevated in patients with CAD (coefficient 0.267), whereas C18:0e MAG (coefficient −0.074) and α-muricholate (coefficient −0.407) were depleted.
Random forest classification models demonstrated the predictive potential of gut microbiota features for CAD. A model based on bacterial species achieved a mean area under the curve of 79%, with AM51-8 sp003478275, F. prausnitzii, and Faecalibacillus intestinalis as top-ranking predictors. The metabolites-based model yielded a mean AUC of 78%.
Integrating both bacterial taxa and metabolites improved predictive performance to a mean AUC of 89%, with inosine showing the highest feature importance, followed by α-muricholate, F. prausnitzii, C18:0e MAG, and AM51-8 sp003478275.
MAG Reconstruction and Analysis
MAG reconstruction yielded 31 bins from coassemblies, with 17 originating from patients with CAD and 14 from controls. Individual assemblies resulted in 326 bacterial MAGs, which after dereplication at 95% average nucleotide identity yielded 144 distinct species-level MAGs.
Community-scale metabolic profiling revealed that MAGs from patients with CAD exhibited more than 1.5-fold higher metabolic weight scores for N₂ fixation, CO oxidation, carbon fixation, iron oxidation, nitrous oxide reduction, and sulfite reduction, with Sutterella wadsworthensis, Phascolarctobacterium succinatutens, and Clostridium sp001916075 as predominant contributors.
Control-derived MAGs showed higher metabolic weight scores for aromatic degradation, formaldehyde oxidation, acetate oxidation, nitrate reduction, nitrite ammonification, iron reduction, and thiosulfate disproportionation, with major contributors including Megamonas funiformis, Prevotella copri, S. wadsworthensis, and Bilophila wadsworthia.
Strain-Level Functional Heterogeneity
Strain-level comparative genomic analysis revealed distinctive functional profiles between CAD-derived and control-derived MAGs. Control-derived Akkermansia muciniphila MAGs carried functional modules for arabinan and xyloglycan degradation compared with CAD-derived A. muciniphila MAGs.
The CAD-derived M. funiformis MAG displayed carriage of metabolic modules related to methanogenesis, mercury reduction, acetate conversion, and lipid metabolism compared with control-derived MAGs.
Regarding trimethylamine-related genes, the cutC gene was identified in nine MAGs, including six Anaerostipes hadrus MAGs from both patients with CAD and controls. The cntA/yeaW gene was identified in 10 M. funiformis MAGs from both groups. The non-trimethylamine-producing gene mtxB was identified from Phocaeicola plebeius_A and F. prausnitzii MAGs exclusively from controls.
Study Characteristics and Limitations
The study analyzed participants with a mean age of 53 years (range 40-70) and mean body mass index of 25 (range 19-31). Shotgun metagenomic sequencing yielded an average of nearly 46 million raw reads per participant, with nearly 40 million remaining following host DNA removal.
The relatively small sample size of 42 participants and predominantly male cohort (39 males, 3 females) may limit the generalizability of findings and prevent assessment of sex-specific microbiome differences in CAD. The cross-sectional design precludes causal inference regarding the relationship between gut microbiome alterations and CAD development.
The authors declared having no competing interests.
Source: mSystems
