Fecal Microbiome Composition Does Not Predict Diet-Induced TMAO Production in Healthy Adults

Abstract

Background Trimethylamine-N-oxide (TMAO) is a small molecule derived from the metabolism of dietary nutrients by gut microbes and contributes to cardiovascular disease. Plasma TMAO increases following consumption of red meat. This metabolic change is thought to be partly because of the expansion of gut microbes able to use nutrients abundant in red meat. Methods and Results We used data from a randomized crossover study to estimate the degree to which TMAO can be estimated from fecal microbial composition. Healthy participants received a series of 3 diets that differed in protein source (red meat, white meat, and non-meat), and fecal, plasma, and urine samples were collected following 4 weeks of exposure to each diet. TMAO was quantitated in plasma and urine, while shotgun metagenomic sequencing was performed on fecal DNA. While the cai gene cluster was weakly correlated with plasma TMAO (rho=0.17, P=0.0007), elastic net models of TMAO were not improved by abundances of bacterial genes known to contribute to TMAO synthesis. A global analysis of all taxonomic groups, genes, and gene families found no meaningful predictors of TMAO. We postulated that abundances of known genes related to TMAO production do not predict bacterial metabolism, and we measured choline- and carnitine-trimethylamine lyase activity during fecal culture. Trimethylamine lyase genes were only weakly correlated with the activity of the enzymes they encode. Conclusions Fecal microbiome composition does not predict systemic TMAO because, in this case, gene copy number does not predict bacterial metabolic activity. Registration URL: https://www.clinicaltrials.gov; Unique identifier: NCT01427855.

Keywords: fecal microbiome; metagenomics; trimethylamine N‐oxide; trimethylamine lyase.

Figures

Figure 1. Overview of analyses.

A, Healthy adults were provided non‐meat, white meat, or red meat diets for 4‐week periods and donated plasma, urine, and stool samples during clinic visits. Trimethylamine N‐oxide (TMAO) and related metabolites were measured in plasma and urine samples. A subset of urine samples (n=38) was taken after oral challenge with isotope labeled choline and carnitine. Isotope‐labeled TMAO was measured in these urine samples. Stool was used for both anoxic cultures to determine the activities of enzymes related to trimethylamine production and metagenomics sequencing. Metagenomic sequence data were used to compute metagenomic operational taxonomic units, as well as global gene and gene family abundances. Low abundance trimethylamine‐related genes were quantified using a gene cluster‐base d approach. B, Plasma and urine TMAO were not correlated with any stool data set, and stool trimethylamine ‐related enzyme activities were not correlated with related genes or with TMAO production during the oral challenge. C, Elastic net models trained with trimethylamine‐related gene abundances did not predict TMAO or stool enzyme activities. CV indicates cardiovascular; FMO, flavin mono‐oxygenases; GBB, gamma‐butyrobetaine; HPLC/MS/MS, high performance liquid chromatography with tandem mass spectrometry; MI, myocardial infarction; mOTUs, metagenomic operational taxonomic units; and TMAO, trimethylamine‐N‐oxide.

Figure 2. Synthesis of trimethylamine N‐oxide requires dietary substrate as well as host and microbial enzymes.

Nutrients including choline and carnitine, which are abundant in red meat, are converted to trimethylamine via multiple synthetic pathways. The copy number of related genes is altered by diet and may contribute to trimethylamine N‐oxide production by host enzymes. caiA/B/C, crotonobetainyl‐CoA dehydrogenase; cntA/B, carnitine oxygenase/reductase; cutC/D, choline utilization gene cluster; FMOs, flavin monooxygenases; GBB, gamma‐butyrobetaine; grdH, betaine reductase complex component B subunit beta; MI, myocardial infarction; TMA, trimethylamine; TMAO, trimethylamine‐N‐oxide; torA, trimethylamine‐N‐oxide reductase (cytochrome c); and yeaX/W, carnitine monooxygenase subunit YeaX.

Figure 3. cutC predicts trimethylamine N‐oxide in those assigned to non‐meat before red meat.

Plasma trimethylamine N‐oxide is higher during red meat consumption vs a meat‐free diet regardless of diet order. Although cutC is not differentially abundant in red meat vs meat‐free, cutC abundance is correlated with plasma trimethylamine N‐oxide among those assigned to non‐meat before red meat. P values represent Wilcox tests. RPKM indicates reads per kilobase per million reads; TMAO, trimethylamine‐N‐oxide.

Figure 4. Abundance of genes related to trimethylamine synthesis do not improve models of trimethylamine N‐oxide (TMAO).

Elastic net was used to train and test linear models of plasma TMAO, as well as percent changes in TMAO from non‐meat to red meat diets. Models were tested with 100 randomly selected training/testing groups, using adjusted R2 and root‐mean‐squared error as performance metrics. The base model, using data summarized in Table 1, predicted 41% of the variance in TMAO, and the addition of trimethylamine‐related gene abundances did not improve model performance. TMAO fractional excretion rate outperformed the base model in predicting changes in plasma TMAO, in terms of root‐mean‐squared error (P<0.001, Wilcox test). FER, fractional excretion rate; RMSE indicates root‐mean‐squared error; and TMAO, trimethylamine‐N‐oxide.

Figure 5. Fecal trimethylamine‐lyase activities are not correlated with trimethylamine N‐oxide (TMAO) production in host.

In a single visit, 13 participants randomly selected from the APPROACH cohort donated a fecal sample and participated in a heavy‐isotope‐labeled choline and carnitine challenge (38 visits total). A, Participants were dosed orally with 250mg d6‐choline and d3‐carnitine. Labeled TMAO and synthetic intermediates were quantitated in 24‐hour urine collection. B, Paired fecal samples were cultured with 200 µmol/L d6‐choline and d9‐carnitine and labeled TMAO and synthetic intermediates were quantitated after 18‐ and 36‐hour incubations. Analytes not detected in >30% of samples were excluded. C, The heatmap shows Spearman correlations between fecal enzyme activities and the indicated isotopologue in 24‐hour urine collection. TMAO produced from carnitine in the host was not highly correlated with carnitine‐derived trimethylamine in fecal cultures. Urine choline‐derived TMAO was negatively correlated with choline‐trimethylamine‐lyase activity in feces (ρ=−0.46, P=0.003). Cr, creatinine; RMSE indicates root‐mean‐squared error; and TMAO, trimethylamine‐N‐oxide. * Adjusted P<0.05. P values adjusted with the method of Benjamani and Hochberg.

Figure 6. cutC abundance marginally outperforms clinical data in predicting fecal choline‐trimethylamine lyase activity.

Fecal samples were cultured with 200 µmol/L d6‐choline and labeled trimethylamine and synthetic intermediates were quantitated after 18‐hour incubation. Elastic net models were trained to predict 18‐hour d6‐choline with 70% of samples and tested on the remaining 30%, and testing was iterated over 100 different training/testing sets. Although the cutC models had less error than models trained with clinical data (Table 1), cutC models did not result in higher correlations between predicted and actual d6‐choline concentration. RMSE indicates root‐mean‐squared error.