Trimethylamine-N-Oxide Promotes Age-Related Vascular Oxidative Stress and Endothelial Dysfunction in Mice and Healthy Humans

Abstract

Age-related vascular endothelial dysfunction is a major antecedent to cardiovascular diseases. We investigated whether increased circulating levels of the gut microbiome-generated metabolite trimethylamine-N-oxide induces endothelial dysfunction with aging. In healthy humans, plasma trimethylamine-N-oxide was higher in middle-aged/older (64±7 years) versus young (22±2 years) adults (6.5±0.7 versus 1.6±0.2 µmol/L) and inversely related to brachial artery flow-mediated dilation (r2=0.29, P<0.00001). In young mice, 6 months of dietary supplementation with trimethylamine-N-oxide induced an aging-like impairment in carotid artery endothelium-dependent dilation to acetylcholine versus control feeding (peak dilation: 79±3% versus 95±3%, P<0.01). This impairment was accompanied by increased vascular nitrotyrosine, a marker of oxidative stress, and reversed by the superoxide dismutase mimetic 4-hydroxy-2,2,6,6-tetramethylpiperidin-1-oxyl. Trimethylamine-N-oxide supplementation also reduced activation of endothelial nitric oxide synthase and impaired nitric oxide-mediated dilation, as assessed with the nitric oxide synthase inhibitor L-NAME (NG-nitro-L-arginine methyl ester). Acute incubation of carotid arteries with trimethylamine-N-oxide recapitulated these events. Next, treatment with 3,3-dimethyl-1-butanol for 8 to 10 weeks to suppress trimethylamine-N-oxide selectively improved endothelium-dependent dilation in old mice to young levels (peak: 90±2%) by normalizing vascular superoxide production, restoring nitric oxide-mediated dilation, and ameliorating superoxide-related suppression of endothelium-dependent dilation. Lastly, among healthy middle-aged/older adults, higher plasma trimethylamine-N-oxide was associated with greater nitrotyrosine abundance in biopsied endothelial cells, and infusion of the antioxidant ascorbic acid restored flow-mediated dilation to young levels, indicating tonic oxidative stress-related suppression of endothelial function with higher circulating trimethylamine-N-oxide. Using multiple experimental approaches in mice and humans, we demonstrate a clear role of trimethylamine-N-oxide in promoting age-related endothelial dysfunction via oxidative stress, which may have implications for prevention of cardiovascular diseases.

Keywords: acetylcholine; aging; brachial artery; microbiota nitric oxide; superoxide.

Conflict of interest statement

CONFLICTS OF INTEREST / DISCLOSURES

None.

Figures

Figure 1.. Plasma concentrations of trimethylamine N-oxide (TMAO) are inversely related to conduit artery endothelial function in healthy adults varying in age.

Plasma TMAO (A) and endothelial function (B; assessed by brachial artery flow-mediated dilation [FMDBA]) in young (N=19) and middle-aged to older (MA/O; N=98) adults. Data are mean ± S.E.M. *p<0.05 vs. young adults (unpaired t-test). C) The unadjusted relation between plasma TMAO (log transformed due to skewness) and FMDBA. D) The adjusted partial residual scatterplot between the Ln TMAO and FMDBA residuals calculated from the linear model adjusted for age, sex, and cardiovascular risk factors (see text), with 95% confidence intervals for the fitted values (grey shaded area). This plot illustrates the strength of the independent effect of TMAO on FMDBA.

Figure 2.. Chronic TMAO supplementation induces aging-like endothelial dysfunction via reduced NO bioavailability and increased superoxide-driven oxidative stress in young mice.

In arteries from young adult mice (12 mo.) supplemented without (Control) or with 0.12% trimethylamine N-oxide (TMAO) for 6 months: A) Carotid artery endothelium-dependent dilation (EDD) to increasing doses of acetylcholine (ACh) in the absence and presence of the nitric oxide (NO) synthase inhibitor NG-nitro-L-arginine methyl ester (L-NAME); B) Peak carotid artery endothelium-independent dilation to increasing doses of the NO donor sodium nitroprusside; C) NO-mediated dilation, as assessed by the difference in carotid artery peak EDD in the absence vs. presence of L-NAME; (D) Protein abundance of phosphorylated endothelial nitric oxide synthase (p-eNOS) at Ser1177, assessed by both Western blotting in aorta lysates (left panel; normalized to GAPDH with representative Western blot images generated from WES electropherograms shown below) and immunofluorescence in aorta rings (right panel; representative images show below); and (E) abundance of nitrotyrosine (sum of high and low molecular weight residues normalized to GAPDH with representative Western blot images generated from WES electropherograms shown to the right); and F) Carotid artery peak EDD to ACh in the absence and presence of the superoxide dismutase mimetic 4-hydroxy-2,2,6,6-tetramethylpiperidin-1-oxyl (TEMPOL). All data are mean ± S.E.M. N=6–10/group. *p<0.05 TMAO vs. Control (panel A: 2-way mixed design ANOVA with Tukey’s post hoc; panels C-F: unpaired t-test). †p<0.05 vs. ACh alone in TMAO-treated mice (paired t-test).

Figure 3.. Acute TMAO incubation directly induces endothelial dysfunction in isolated arteries.

A) Endothelium-dependent dilation (EDD) to increasing doses of acetylcholine (ACh) before (ACh alone) and after 60 min pre-incubation with 25μM, 50μM or 100μM trimethylamine N-oxide (TMAO). Statistics are pairwise comparisons of each concentration of TMAO on mean EDD (averaged across doses of ACh; i.e., main TMAO concentration effect). *p<0.05 vs. ACh alone. †p<0.05 vs. 25μM TMAO. #p<0.05 vs. 50μM TMAO. N=11. B) Peak EDD derived from the dose response curves presented in (A). C) Endothelium-independent dilation to sodium nitroprusside (SNP) in paired carotid arteries from the same mice pre-incubated without (SNP alone) or with 100μM TMAO. N=7. D-E) Peak EDD to ACh, either alone or following pre-incubation with 100μM TMAO and/or with the nitric oxide synthase inhibitor NG-nitro-L-arginine methyl ester (L-NAME) (D) or the superoxide dismutase mimetic TEMPOL (E). Both N=4. For B-D, *p<0.05 vs. ACh alone (RM ANOVA with Tukey’s post-hoc). ‡p<0.05 vs. TMAO. All data are mean ± S.E.M.

Figure 4.. DMB treatment reverses age-related impairments in endothelial function.

In young (5-6 mo.) and old (26-27 mo.) mice supplemented without (YC, OC) or with (YDMB, ODMB) 1% 3,3-dimethyl-1-butanol (DMB) for 8-10 weeks (unless otherwise noted, young: N=8-11/group; old: N=10-14/group): A) Plasma TMAO concentrations; B) (Left panel) Carotid artery endothelium-dependent dilation (EDD) to increasing doses of acetylcholine (ACh) in the absence and presence of the nitric oxide (NO) synthase inhibitor NG-nitro-L-arginine methyl ester (L-NAME), and (right panel) carotid artery endothelium-independent dilation to increasing doses of the NO donor sodium nitroprusside; C) NO-mediated dilation, as assessed by the difference in peak EDD in the absence vs. presence of L-NAME; D) Abundance of phosphorylated (Ser1177) endothelial NO synthase (p-eNOS) in aorta rings, with representative images shown to the right; E) Superoxide production, measured in 1mm aorta rings by electron paramagnetic resonance spectroscopy, with representative tracings above; and F) Peak carotid artery EDD to ACh in the absence and presence of the superoxide dismutase mimetic 4-hydroxy-2,2,6,6-tetramethylpiperidin-1-oxyl (TEMPOL) in old mice (N=6-14/group). All data are mean ± S.E.M. Statistics are pairwise comparisons across groups and/or drug condition (1-way [panels A, C-E] or 2-way mixed design [panels B and F] ANOVA with Tukey’s post hoc). *p<0.05 OC vs. YC. †p<0.05 OC vs. ODMB. ‡p=0.07 OC vs. ODMB.

Figure 5.. Endothelial oxidative stress and its suppression of vascular function with aging are associated with higher circulating trimethylamine N-oxide (TMAO).

A) Brachial artery flow-mediated dilation (FMDBA) following 20-min intravenous infusion with saline or the antioxidant ascorbic acid in young adults (18-27 years; N=14) and middle-aged to older (MA/O; 50-79 years) adults with either lower (<4.4μM; N=11) or higher (>4.5μM; N=16) plasma TMAO. Statistics are pairwise comparisons (Sidak’s multiple comparison test) assessed following 2-way mixed design ANOVA (group x drug: p<0.01). *p<0.05 vs. young within drug condition. †p<0.05 saline vs. ascorbic acid within group. B) The change in FMDBA from saline control in response to ascorbic acid (AA) infusion, i.e., oxidative stress-mediated suppression of FMDBA. Statistics are pairwise comparisons (Tukey’s post-hoc test) assessed following 1-way ANOVA (main effect: p<0.01). *p<0.05 vs. young. ‡p<0.05 vs. MA/O adults with lower TMAO. C-D) Abundance of nitrotyrosine in biopsied endothelial cells from MA/O adults is higher in individuals with higher vs. lower plasma TMAO levels (C; N=11-12/group; representative fluorescent images shown on the right) and is related to higher plasma TMAO levels (D; r2 for the unadjusted model is shown; TMAO values are log-transformed due to skewness).