Gut microbial trimethylamine N-oxide generation promotes risk of atrial fibrillation via muscarinic receptor-mediated autonomic dysfunction

Figure 1. Plasma TMAO, choline, and betaine independently associate with prevalent AF. (A) The relationship between tertiles (T) of increasing plasma concentrations of TMAO, choline, and betaine and the proportion of prevalent AF in GeneBank subjects. The P value represents a Pearson’s χ2 test. (B) Forest plots of the OR of prevalent AF and tertiles of increasing concentrations of plasma TMAO, choline, and betaine with no adjustment (white circles), adjustment for patient characteristics and comorbidities (model 1: age, sex, diabetes mellitus, CVD, systolic blood pressure, current smoker, BMI, and high-sensitivity C-reactive protein; gray circles), and after full model adjustment (model 1 and eGFR; black circles). The circles represent the OR and the horizontal bars extend from the lower limit to the upper limit of the 95% CI. (C) Cubic spline plots of plasma concentrations of TMAO, choline, and betaine and the OR of prevalent AF. The solid red line represents the estimated OR from unadjusted logistic regression analysis at increasing values of the metabolite, and the dashed blue lines represent the 95% CI. **P < 0.01.

Figure 2. Association of baseline TMAO tertiles with odds of AF. Forest plots show OR for AF with increasing TMAO levels in the indicated subgroups. Open circles represent OR for tertile 3 (T3) versus tertile 1 (T1); horizontal lines indicate 95% CIs. P interaction (P) values were obtained from likelihood ratio tests comparing logistic models with and without the interaction between TMAO tertiles and each subgroup factor. Units for continuous variables: age (years), BMI (kg/m2), HDL cholesterol (HDLc; mg/dL), LDL cholesterol (LDLc; mg/dL), triglycerides (TG; mg/dL), high-sensitivity C-reactive protein (hs-CRP; mg/L), and eGFR (mL/min/1.73 m2). DM, diabetes mellitus.

Figure 3. Kaplan-Meier plots and HRs of the incidence of AF in CREM-IbΔC-X mice on TMAO diet and plasma TMA concentrations. (A) The incidence of paroxysmal and persistent AF in all CREM-IbΔC-X mice on a chow or TMAO-supplemented diet, as represented by Kaplan-Meier plots and HRs. The circles represent the HR, and the horizontal bars extend from the lower limit to the upper limit of the 95% CI of the estimate of the HR. HRs between respective dietary groups were calculated using a Cox proportional hazard model. (B) Plasma concentrations of TMAs in all CREM-IbΔC-X mice used in the TMAO mouse ECG study. A Mann-Whitney test was used to assess significant differences between groups. Data represent mean ± SD. (C) Trials of AF percent induction in C57BL/6J mice treated with TMAO and chow diets. (D) TMAO levels and AF percent induction correlation in all C57BL/6J mice. Pearson’s correlation coefficient measuring the relationship between TMAO levels and AF percentage of AF. P values were determined using a linear regression model. *P < 0.05, **P < 0.01, ****P < 0.0001.

Figure 4. Kaplan-Meier plots and HRs of the incidence of paroxysmal AF and persistent AF in CREM-IbΔC-X mice on respective diets. (A and B) The incidence of paroxysmal (A) and persistent AF (B) in all CREM-IbΔC-X mice, as represented by a Kaplan-Meier plots and HRs, on a chow diet, a choline-supplemented diet, a chow + IMC-supplemented diet, and a choline + IMC-supplemented diet. The circles represent the HR, and the horizontal bars extend from the lower limit to the upper limit of the 95% CI of the estimate of the HR. HRs between respective dietary groups were calculated using a Cox proportional hazard model. (C) Average plasma concentrations of TMAO, choline, and betaine in all CREM-IbΔC-X mice used in the ECG study. Data represent mean ± SD. A Kruskal-Wallis test with Dunn’s post hoc analysis was used to determine the significant differences among the medians of 4 independent diet groups, and multiple paired comparison adjustments were performed. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Figure 5. HRs and Kaplan-Meier plot of the incidence of paroxysmal and persistent AF. (A) The incidence of paroxysmal and persistent AF and average plasma TMAO, choline, and betaine concentrations in male and female CREM-IbΔC-X mice. (B) Kaplan-Meier plot and HRs of the duration of paroxysmal AF in male and female CREM-IbΔC-X mice. In A and B, the circles represent the HR, and the horizontal bars extend from the lower limit to the upper limit of the 95% CI of the estimate of the HR. HRs between respective dietary groups were calculated using a Cox proportional hazard model. Each HR represents the risk over the entire analyte concentration range. *P < 0.05, **P < 0.01.

Figure 6. IMC concentrations in plasma and cecum samples and assessment of CutC/D activity. (A) Quantification of IMC in cecum and plasma samples of CREM-IbΔC-X mice on respective diets. (B) Quantification of the conversion of d9-TMA to d9-TMAO in liver samples of mice on respective diets. (C and D) Quantification of the conversion of d9-choline to d9-TMA in cecum of CREM-IbΔC-X mice under aerobic (C) and anaerobic (D) conditions on respective diets. Replicates represent 4 independent samples from each mouse. (E and F) α-Diversity plot (E; Shannon’s index) and principal component analysis (F) of β-diversity plot of CREM-IbΔC-X mice on respective diets. Permutational ANOVA (Permanova) was used to compare the β-diversity distributions of dietary groups. (G) CutC/D mRNA counts in cecum samples from all, male, and female CREM-IbΔC-X mice utilized in the mouse ECG study. In A–E and G, data represent mean ± SD. A Kruskal-Wallis test with Dunn’s post hoc analysis was used to determine the significant difference among the medians of 4 independent diet groups, and multiple paired comparison adjustments were performed. CCA, canonical correspondence analysis. *P < 0.05, **P < 0.01, ***P < 0.001,…

Figure 7. CutC levels and correlation with gut microbiome and differential abundance of gut microbial species with the time of onset of paroxysmal and persistent AF. (A) Gut microbiota taxa abundance and its association with mouse plasma TMA and TMAO and CutC/D mRNA counts in linked cecal samples of male and female CREM-IbΔC-X mice on respective diets. The circles represent the mean, and the bars represent SEM with regression and P values determined using the linear regression model. *P < 0.05, **P < 0.01. (B and C) Volcano plot showing the association of differential abundance of gut microbial species with the time of onset of paroxysmal (B) and persistent (C) AF. In the plot, the x axis shows the coefficient (effect size and direction), and the y axis shows the –log10(q value). Each circle represents a specific microbial species, colored on a yellow-to-red gradient indicating q values, with yellow being more significant. The dashed dark red line at y ≈ 1.30 marks the statistical significance threshold (q value = 0.05). Circles above this line and labeled are statistically significant (q < 0.05). Circles in the upper corners of the plot, far from 0 on the x axis and high on the y axis,…

Figure 8. Choline diet increases the LA size during diastolic and systolic function. (A) LA diastolic size in male and female CREM-IbΔC-X mice at 8 weeks of age on respective diets. (B) LA systolic size in male and female CREM-IbΔC-X mice at 8 weeks of age on respective diets. (C) Echocardiography of LA size in female and male CREM-IbΔC-X mice at 8 weeks of age on respective diets. In A and B, data represent mean ± SD. A Kruskal-Wallis test with Dunn’s post hoc analysis was used to determine the significant difference among the medians of 4 independent diet groups, and multiple paired comparison adjustments were performed. *P < 0.05, **P < 0.01, ****P < 0.00001.

Figure 9. Analysis of CV and APD in CREM-IbΔC-X mouse hearts using ex vivo optical mapping on chow versus choline-supplemented diets. (A) Schematic of the optical mapping experimental setup used for whole-heart atrial action potential measurements (see text for details). (B) Representative examples of action potentials (left) measured from the LA and posterior wall for chow (black) and choline (red), and atrial activation maps (posterior view) during 130 ms pacing (asterisk) with local CV vectors superimposed for chow (middle) and choline (right) RA, right atria. (C–G) Grouped data of CV (C), APD80 (D), APD80 maximum (E), APD 80 minimum (F), and wavelength (G) at a pacing cycle length of 130 ms. All P values shown were determined using 2-tailed Student’s t test to compare the means of 2 independent groups, after confirmation of normality using Shapiro-Wilk and Kolmogorov-Smirnov tests. *P < 0.05, **P < 0.01.

Figure 10. TMAO modulates autonomic signaling. (A) TMA/TMAO induced inhibition of M2R (clone 1) signaling in HEK293 cells. Unpaired 2-tailed Welch’s t test was used to obtain the fold change P value by comparing the mean EC50 values of carbachol (Car) and TMA/TMAO (n = 6 independent experiments). (B) Heart rate in CREM-IbΔC-X mice treated with choline and chow diet. Data represent mean ± SD. A Mann-Whitney test was used to assess significant differences between groups. *P < 0.05, **P < 0.01. (C) Schematic demonstrating that IMC inhibits TMA/ TMAO formation, AF induction, and progression. Choline diet alters gut microbial TMA and its metabolite formation. TMAO formation plays a key role in autonomic dysfunction via inhibition of M2R and the resulting increased sympathetic tone and susceptibility to AF. IMC that inhibits formation of TMA/TMAO can reduce AF induction and progression.