Metaorganismal choline metabolism shapes olfactory perception

Abstract

Microbes living in the intestine can regulate key signaling processes in the central nervous system that directly impact brain health. This gut-brain signaling axis is partially mediated by microbe-host-dependent immune regulation, gut-innervating neuronal communication, and endocrine-like small molecule metabolites that originate from bacteria to ultimately cross the blood-brain barrier. Given the mounting evidence of gut-brain crosstalk, a new therapeutic approach of "psychobiotics" has emerged, whereby strategies designed to primarily modify the gut microbiome have been shown to improve mental health or slow neurodegenerative diseases. Diet is one of the most powerful determinants of gut microbiome community structure, and dietary habits are associated with brain health and disease. Recently, the metaorganismal (i.e., diet-microbe-host) trimethylamine N-oxide (TMAO) pathway has been linked to the development of several brain diseases including Alzheimer's, Parkinson's, and ischemic stroke. However, it is poorly understood how metaorganismal TMAO production influences brain function under normal physiological conditions. To address this, here we have reduced TMAO levels by inhibiting gut microbe-driven choline conversion to trimethylamine (TMA), and then performed comprehensive behavioral phenotyping in mice. Unexpectedly, we find that TMAO is particularly enriched in the murine olfactory bulb, and when TMAO production is blunted at the level of bacterial choline TMA lyase (CutC/D), olfactory perception is altered. Taken together, our studies demonstrate a previously underappreciated role for the TMAO pathway in olfactory-related behaviors.

Keywords: Nutrition; gut microbiome; metabolism; olfaction.

Figure 1

Trimethylamine N-Oxide (TMAO) pathway metabolites are differentially abundant across the gut-liver-brain axis. Tissue levels of TMAO pathway co-metabolites were quantified using stable isotope dilution liquid chromatography-tandem mass spectrometry (LC-MS/MS) in the gut, liver, and distinct brain regions. Tissues were harvested from male C57Bl/6J mice including intestinal segments including duodenum (SI-1), jejunum (SI-2), ileum (SI-3), cecum, liver, and different brain regions including cortex, subcortical region, cerebellum, brainstem, and olfactory bulb. Metabolites quantified included (A) choline, (B) L-carnitine, (C) betaine, (D) γ-butyrobetaine, (E) trimethylamine (TMA), and (F) trimethylamine N-oxide (TMAO). Data are presented as mean ± SEM from n = 4 mice.

Figure 2

Regional Differences in Fmo1-5 expression in the mouse brain. Brain regions (Brain stem, Cerebellum, Cortex, Olfactory Bulb, and Subcortical Region) were harvested from male C57Bl/6J, and expression of Fmo1-5 transcripts was quantified using qRT-PCR. Data are presented as mean + SEM from n = 5 mice; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001.

Figure 3

Gut microbe-targeted choline trimethylamine lyase inhibition alters innate behaviors including olfactory perception. Wild-type female C57BL/6J mice were treated with or without the gut microbial CutC/D inhibitor iodomethylcholine (IMC) in the drinking water and then subjected to a series of behavioral phenotyping tests to broadly assess alterations in innate behaviors. A–D, Plasma levels of trimethylamine N-oxide (TMAO) pathway co-metabolites (A) choline, (B) L-carnitine, (C) trimethylamine (TMA), and (D) TMAO were quantified via LC-MS/MS. Innate behavioral phenotyping tests included (E and F) grooming test, (G) olfactory cookie test, (H) food odor discrimination tests, (I) novel mouse odor discrimination test, (J and K) nest building test, (L) hotplate sensitivity test, (M) startle test, and (N) forepaw grip strength test. Data are presented as mean ± SEM; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; n = 8/group for metabolomics, n = 20 control, n = 20 IMC-treated for grooming test, all other tests had n = 19 control, n = 20 IMC-treated.

Figure 4

Gut microbe-targeted choline trimethylamine lyase inhibition does not significantly alter cognitive, anxiety-like, or depression-related behaviors. Wild-type female C57BL/6J mice were treated with or without the gut microbial CutC/D inhibitor iodomethylcholine (IMC) in the drinking water, and then subjected to a series behavioral phenotyping tests to broadly assess cognition, anxiety, and depression including (A) Y-maze test, (B–D) Standard Morris water maze, (E) Probe trial in Morris water maze, (F) Elevated plus maze, (G–I) Open field test, (J) Tail suspension test, and (K) Forced swim test. Data are presented as mean ± SEM; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; n = 20 control, n = 20 IMC-treated for the forced swim test, all other tests had n = 19 control, n = 20 IMC-treated.

Figure 5

Gut microbe-targeted choline trimethylamine lyase inhibition selectively alters some, but not all, social interaction-related behaviors. Wild-type female C57BL/6J mice were treated with or without the gut microbial CutC/D inhibitor iodomethylcholine (IMC) in the drinking water and then subjected to a series of behavioral phenotyping tests to assess social interactions including (A) 3-chamber social interaction test, (B) Social preference test, (C) Social novelty test, and (D–G) Social interaction with a juvenile test. Data are presented as mean ± SEM; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗∗p < 0.0001; n = 20 control, n = 20 treated for 3-chamber test, all other tests n = 19 control, n = 20 IMC-treated.

Figure 6

TMA lyase inhibition alters olfactory TMAO, but not TMA and consistently shows defects in olfactory behavior. Female wild type were treated with or without the gut microbial CutC/D inhibitor iodomethylcholine (IMC) in the drinking water, and then subjected to plasma and tissue metabolomics as well as the olfactory cookie test and a broadened olfactory discrimination test compared to that in Figure 2. A–D, Plasma and (E–H) olfactory levels of trimethylamine N-oxide (TMAO) pathway co-metabolites (A and E) choline, (B and F) L-carnitine, (C and G) trimethylamine (TMA), and (D and H) TMAO were quantified via LC-MS/MS. Latency to find the cookie was significantly increased in FMO3^TG mice (I). Attractant (J) and deterrent (K) odor discrimination testing was performed. Data are presented as mean ± SEM and represent n = 20 for the plasma metabolites and behavioral testing (panels A–D, I–K), n = 4 for tissue metabolomics (E–H).