A new role for microbiota? Dulling the thrust of serotonin and 5-HT3 signaling cascade.

THE IMPORTANCE OF SEROTONIN in gastrointestinal biology is often overlooked given the intense interest of investigators focused on neurotransmitter (mal)functions of the central nervous system. Regardless, the gastrointestinal tract is the major site of serotonin (5-hydroxytryptamine, 5-HT) production (95%) and its crucial role as a paracrine mediator and neurotransmitter in regulating gastrointestinal function has been well established (2). Its functions range from modulating visceral sensation, neuronal protection and regeneration to stimulating secretion, motility, and proinflammatory responses. Serotonin elicits its varied effects by the coordinated action of multiple receptor subtypes exhibiting cell-specific localization and signaling cascades as well as the regulation of the Na/Cl-dependent serotonin transporter, SERT. Furthermore, serotonin’s interplay with the gut microbiome and its role as a key intermediary in the gut-brain axis are gaining prominence. These multiplex roles have led serotonin to earn the moniker, the “sword and shield of the gut,” where Dr. Gershon ascribes the sword function to mucosal serotonin and the shield function to neuronal serotonin (3). In this issue of the Americal Journal of Physiology-Gastrointestinal Liver Physiology, the work of Bhattarai and colleagues (1) adds a twist to this, suggesting that the sword function could be manipulated by the microbiome. With a view to gain an insight into the human microbiota, the authors compared serotonin-sensitive ion secretion in age-matched Swiss Webster Germ Free (GF) mice with those colonized with human gut microbiota (HM). They focused on the proximal colon and measured ion transport by transepithelial shortcircuit current (Isc), an index of Cl secretion in this segment of the colon. While there were no differences in baseline transepithelial resistance, Isc or cAMP-stimulated secretion, the GF mice had a significantly greater secretory response to serotonin than their HM counterparts, suggesting that the presence of gut microbiota specifically tempers the secretory response to 5-HT. The prevalent 5-HT receptors in the intestine are 5-HT3, the cys loop family, ligand-gated ion channel, and 5-HT4, a G protein-coupled receptor. Although not specifically defined in the mouse colon, 5-HT3 is known to exist as a heteropentamer of 5-HT3A and 5-HT3B isoforms. There was an increase in colonic 5-HT3B, but not 5-HT4, transcript in GF versus HM animals. This was paralleled by an increase only in the 50-kDa, but not 40-kDa, 5-HT3B isoform of the protein. The 5-HT3-specific antagonist, ondansetron, inhibited serotonin-stimulated current, whereas the 5-HT3-specific agonist, 2-methyl-5 hydroxytryptamine HCl, also stimulated Isc in GF but not HM mice. The serotonin response in GF persisted in the presence of tetrodotoxin, suggesting a nonneuronal component of regulation. In contrast, the 5-HT4 antagonist, GR118308, decreased Isc both in GF and HM, even in the presence of tetrodotoxin, suggesting that the GPCR signaling cascade has a role in the nonmicrobial-dependent response to serotonin. An intriguing observation is that both human microbiota (HM) and conventional mouse microbiota (CR) can overcome the 5-HT3mediated response seen in GF animals, suggesting this was a common effect of commensal microbiota regardless of host species. It still begs the question as to whether the alpha diversity of the HM mice is similar to that of CR. Clearly all these point to a role of the microbiota in tempering the serotonin response, but how is the signal conveyed? Suggestive hints are provided that need further proof. First, as a proof of concept, the authors use 5-HT3A-GFP (not 5-HT3B) mice to demonstrate the presence of these receptors in nerve fibers as well as enterchromaffin cells of the proximal colon. Second, in colonoids derived from GF mice, acetate at a single concentration, but not butyrate, decreased the levels of 5-HT3B mRNA, hinting that the signal may be via short chain fatty acids, in particular, acetate. As with many good studies, the findings raise more questions: What is the diversity and composition of the protective commensals? While treating all mice with samples from one human allows for a well-controlled study, it begs the question of what effects -diversity will have on the response? What is the trigger? Is the microbiome profile altered in disease states, and does it use the same or different triggers? There is some hint to the trigger in nondiseased states, in that the short chain fatty acid, acetate, attenuates the 5-HT3 mRNA levels in GF-derived colonoids. It is interesting that acetate is effective at a very narrow concentration range and perhaps evokes, suggestions of a niche effect in the intact animal. Here it is helpful to set it in the larger context of what other changes are seen in the serotonin signaling scenario. In a 2015 paper by Reigstad et al. (5), this investigative group reported that there was no difference in the contractility seen in GF mice or those exposed to humanized microbiota (HM) or conventional mouse microbiota (CR). The HM and CR mice, however, exhibit an increase in serotonin production, which was related to an increase in its rate-limiting synthetic enzyme, tryptophan hydroxylase (Tph), but there were no changes in 5-HT4, SERT, or monoamine oxidase. Again in a proof of Address for reprint requests and other correspondence: E. B. Chang, Univ. of Chicago, 900 E. 57th St., Knapp Center for Biomedical Discovey, Chicago, IL 60637 (e-mail: [email protected]). Am J Physiol Gastrointest Liver Physiol 313: G14–G15, 2017; First published May 24, 2017; doi:10.1152/ajpgi.00166.2017.