The Diverse Metabolic Roles of Peripheral Serotonin

Abstract

Serotonin (5-hydroxytryptamine or 5-HT) is a multifunctional bioamine with important signaling roles in a range of physiological pathways. Almost all of the 5-HT in our bodies is synthesized in specialized enteroendocrine cells within the gastrointestinal (GI) mucosa called enterochromaffin (EC) cells. These cells provide all of our circulating 5-HT. We have long appreciated the important contributions of 5-HT within the gut, including its role in modulating GI motility. However, evidence of the physiological and clinical significance of gut-derived 5-HT outside of the gut has recently emerged, implicating 5-HT in regulation of glucose homeostasis, lipid metabolism, bone density, and diseases associated with metabolic syndrome, such as obesity and type 2 diabetes. Although a new picture has developed in the last decade regarding the various metabolic roles of peripheral serotonin, so too has our understanding of the physiology of EC cells. Given that they are scattered throughout the lining of the GI tract within the epithelial cell layer, these cells are typically difficult to study. Advances in isolation procedures now allow the study of pure EC-cell cultures and single cells, enabling studies of EC-cell physiology to occur. EC cells are sensory cells that are capable of integrating cues from ingested nutrients, the enteric nervous system, and the gut microbiome. Thus, levels of peripheral 5-HT can be modulated by a multitude of factors, resulting in both local and systemic effects for the regulation of a raft of physiological pathways related to metabolism and obesity.

Enteroendocrine cells collectively constitute the largest endocrine tissue in the body. These cells are scattered within the gastrointestinal (GI) epithelium, and make up about 1% of all cells lining the GI tract. They consist of an array of different cell types, each containing specific hormone markers that are actively secreted in response to various physiological stimuli, including nutrients and gut distension. Enterochromaffin (EC) cells within the GI mucosa are specialized enteroendocrine cells that synthesize and secrete between 90% and 95% of total body serotonin (5-hydroxytryptamine or 5-HT) (1). They are a considerable source of this important multifunctional bioamine and constitute about half of all enteroendocrine cells. The remaining pool of 5-HT is synthesized predominantly by serotonergic neurons within the central nervous system (CNS) (2–6), with much smaller amounts produced by pancreatic islets (7), mammary glands (8–11) and adipose tissue (12). Plasma 5-HT does not cross the blood-brain barrier (13), and indeed, central and peripheral sources of 5-HT may have opposing roles in energy homeostasis (14).

The synthesis of 5-HT requires the rate-limiting enzyme tryptophan hydroxylase (TPH), which exists as two isoforms. TPH1 is predominantly localized in gut EC cells as well as other nonneuronal sources, whereas TPH2 is found largely in neurons of the CNS (2). A small proportion of enteric neurons (∼1%) also contain 5-HT (15). EC-cell–derived 5-HT is a pleiotropic bioamine, with broad functions, including roles in platelet aggregation (16), GI motility (17), regulation of bone density (18), liver regeneration (19), and inflammation (20). Recent studies demonstrate that 5-HT also regulates glucose homeostasis, hepatic gluconeogenesis, mobilization of hepatic free fatty acids, and the browning of white adipose tissue (WAT) (12, 21, 22). Such effects have direct implications for metabolic disorders such as type 2 diabetes (T2D) and obesity, in which energy homeostasis is significantly perturbed. In addition, EC cells may act as a signaling nexus between the gut microbiome and host; gut dysbiosis and altered bacterial signaling in obese individuals is identified as a pathway through which gut 5-HT may affect metabolic control (23, 24).

In this review, we consider the physiological role that 5-HT plays in regulating metabolism via actions on an array of target tissues. In addition, we discuss the involvement of altered 5-HT homeostasis in clinically important metabolic diseases, and examine the evidence for EC cells as critical sensory cells within the gut.

5-HT Regulates Blood Glucose and Obesity Through Effects on Hepatocyte and Adipocyte Function

Gut-derived 5-HT is a regulator of metabolism (25) through interactions with key metabolic target tissues (Table 1), and altered gut-derived 5-HT is related to both T2D and obesity. Plasma 5-HT and blood glucose levels are positively correlated in humans (26), fish (27), sheep (28), and rodents (21). Circulating 5-HT is increased in individuals with T2D, as evidenced through the measurement of urinary levels of the major 5-HT metabolite, 5-hydroxyindoleacetic acid (26). These levels are doubled in patients with T2D, and positively correlate to glycated hemoglobin. Gain-of-function polymorphisms in TPH1 were associated with body mass index (BMI) and waist circumference, both of which are measures of obesity, in a genome-wide association study of 8842 nondiabetic individuals (29). Increased intestinal and plasma 5-HT levels are observed in mice with diet-induced obesity (30, 31). However, the absence of 5-HT, through genetic or pharmacological blockade of peripheral TPH, protects against the development of metabolic syndrome in mice on a high-fat diet (12, 22).

Table 1.

Summary of Tissue Targets of Peripheral 5-HT Involved in Metabolism and Effect of 5-HT on These Target Tissues

Tissue	Effect of 5-HT on Target Tissue
GI tract	Modulation of intestinal motility via 5-HT3 and 5-HT4 receptors
	Increased luminal bicarbonate and electrolyte secretion
	Increased fat absorption by increase of bile acid turnover
Pancreas	Serotonylation of exocytosis proteins
	Inhibition of insulin secretion via 5-HT1D receptors
	Increased GSIS insulin secretion via 5-HT2 and 5-HT3 receptors
	Increased β-cell mass during pregnancy via 5-HT2B and 5-HT3 receptors
Liver	Increased bile acid synthesis and secretion
	Increased gluconeogenesis via 5-HT2B and 5-HT2C receptors
	Decreased glycogen synthesis
	Decreased glucose uptake
WAT	Increased lipolysis via 5-HT2B receptors
	Decreased adiponectin secretion via 5-HT2A
	Inhibition of WAT “browning”
BAT	Decreased thermogenesis via 5-HT3A receptors
Bone	Decreased bone turnover and renewal via 5-HT1B and 5-HT2A receptors and uptake via SERT
	Decreased osteocalcin synthesis and secretion

Abbreviations: BAT, brown adipose tissue; GSIS, glucosestimulated insulin secretion.

One of the major regulators of plasma glucose is the liver, and hepatic gluconeogenesis is the main contributor to plasma glucose levels during periods of fasting (32). Fasting-induced increases in blood glucose are mediated, in part, by the activation of hepatocyte 5-HT_2B receptors. Binding of plasma 5-HT to 5-HT_2B receptors enhances the activity of two key, rate-limiting enzymes in gluconeogenesis, glucose-6-phosphatase and fructose 1,6-bisphosphatase (22). Enhanced activity of these key enzymes occurs at a transcriptional level through increased cyclic adenosine monophosphate (cAMP) downstream of 5-HT_2B receptor stimulation, subsequent activation of cAMP-dependent protein kinase A, and increased activity of the transcription factor, cAMP response element binding factor (33). Although 5-HT increases glucose production by hepatocytes, this increase is limited by the presence of endogenous substrates, particularly glycerol that is generated from lipolysis, and glycogen that is stored within hepatocytes.

As 5-HT increases hepatic gluconeogenesis, it simultaneously decreases glycogen synthesis and reduces GLUT2-mediated uptake of glucose (22, 34). In addition, both circulating 5-HT and hepatic 5-HT_2B receptor expression are increased in mice under fasting conditions (22). This fasting-induced increase in plasma 5-HT is driven by reduced glucose availability, leading to higher EC-cell TPH1 expression and 5-HT synthesis (35). Exaggerated hepatic glucose production is the major determinant of glycemic control, and its contribution to plasma glucose levels is significantly augmented in T2D (32). As such, augmented gut 5-HT may contribute to the development and progression of metabolic diseases through increased hepatic glucose output.

WAT is an energy store that is also regulated by gut-derived 5-HT (Fig. 1). 5-HT is capable of mobilizing free fatty acids and glycerol from adipocytes. This occurs via 5-HT_2B receptor-mediated phosphorylation, and consequent activation, of intracellular hormone-sensitive lipase (HSL), the rate-limiting enzyme involved in lipolysis (22). HSL is stimulated through the 5-HT_2B receptor either indirectly by an increase in cAMP and downstream activation of cAMP-dependent protein kinase A, or directly by phosphorylated perilipin (36). However, it is currently unknown by which of these pathways 5-HT regulates HSL. The 5-HT–induced lipolysis of stored triacylglycerol results in increased plasma levels of free fatty acids and glycerol. Glycerol can then be used to further fuel hepatic gluconeogenesis or can be converted, along with free fatty acids, to acetyl-CoA by β oxidation in the liver for the synthesis of ketone bodies. Experimentally raising plasma 5-HT levels in mice significantly increases plasma glycerol and free fatty acids, an effect which is blunted in mice lacking 5-HT_2B receptors in adipose tissue (22).

Figure 1.

Actions of 5-HT on adipocytes. Peripheral 5-HT stimulates lipolysis within white adipocytes, therefore increasing circulating free fatty acids and glycerol. This occurs via activation of 5-HT_2B receptors, which in turn results in activation of HSL by two possible mechanisms: (1) HSL translocates to the surface of lipid droplets and is activated by direct phosphorylation by cAMP-dependent protein kinase A, or (2) HSL is activated by phosphorylated (Ⓟ) perilipin on the surface of lipid droplets. Whether 5-HT activates HSL by either or both of these mechanisms is currently unknown. 5-HT also inhibits the release of the metabolically active hormone adiponectin via activation of 5-HT_2A receptors. GABA, γ-aminobutyric acid; PACAP, pituitary adenylate cyclase-activating polypeptide.

Serotonin hinders thermogenic capacity and energy catabolism in interscapular brown adipose tissue (iBAT), the primary tissue responsible for heat production (thermogenesis) in mice (12). Exogenous 5-HT significantly attenuates the thermogenic potential of the β-adrenergic receptor agonist, isoproterenol. 5-HT reduces cAMP levels in iBAT, lowers the activation of HSL, and reduces expression of uncoupling protein 1 (UCP1), the mitochondrial protein responsible for thermogenesis) (12). In contrast, mice treated with the peripherally restricted TPH inhibitor LP533401 display increased thermogenic capacity as a result of increased UCP1 expression in iBAT, and importantly, resistance to the glucose intolerance and insulin resistance that normally accompanies a high-fat diet (12).

In addition to direct effects on iBAT, 5-HT also inhibits the “browning” of WAT, the process through which WAT is converted to a brown adipose tissue (BAT)-like phenotype, resulting in so-called “beige” adipocytes. Beige adipocytes have a higher expression of key thermogenic genes encoding for UCP1 and Ppc1α compared with WAT, and consume more energy through the production of heat (37). 5-HT actions on thermogenesis are partially mediated via 5-HT_3A receptors in BAT (38). In addition, the ablation of Tph1 specifically in adipose tissue confers resistance to high-fat diet–induced obesity in mice. These mice display reduced adipocyte mass, lower weight gain, and improved glucose tolerance and insulin sensitivity under such conditions, suggesting 5-HT synthesis within WAT itself has important metabolic consequences (38).

Serotonin can suppress adipose tissue release of adiponectin (Fig. 1), the adipokine capable of attenuating hepatic gluconeogenesis and increasing insulin sensitivity (39–41). Circulating adiponectin levels are lower in rodent models of obesity and patients with T2D (42–47) and negatively correlate with BMI (46, 48). Adiponectin acts in a paracrine manner on adipocytes to decrease lipid accumulation within WAT, to promote thermogenesis in BAT, and to induce the browning of WAT through increased UCP1 expression (49). Activation of 5-HT_2A receptors on mesenteric adipose tissue suppresses adiponectin release, whereas increased expression of these receptors in genetically obese mice markedly reduces adiponectin levels (50). Moreover, knockdown of 5-HT_2A receptor expression, or pharmacological inhibition of its signaling, results in an increase in adiponectin expression in a differentiated adipocyte cell line (50). Thus, peripheral 5-HT contributes to metabolic dysfunction not only through increases in hepatic gluconeogenesis and fasting blood glucose levels, but also by acting on adipocytes to mobilize free fatty acids and glycerol. In addition, peripheral 5-HT augments obesity through suppression of the “browning” of white fat, reduces UCP1 in BAT, and suppresses secretion and expression of metabolically beneficial adiponectin.

The underlying mechanisms behind the increase in plasma 5-HT associated with metabolic dysfunction are unclear. One possible means is leptin, a satiety factor that is primarily synthesized in WAT but is also released by the stomach in response to a meal. Gastric and circulating leptin levels are markedly increased in obese individuals and in mice fed a high-fat diet, and hyperleptinemia is a feature of obesity (51, 52). Mice with an intestine-specific deletion of the leptin receptor isoform B exhibit a marked decrease in duodenal mucosa 5-HT content (53). Increased expression of gastric leptin, gut TPH1, and increased EC-cell numbers are all observed in mice fed a high-fat diet, and, importantly, precede the onset of obesity and hyperleptinemia (53). Such findings indicate that these factors may play a causative role in the development of obesity and the metabolic changes associated with obesity.

Insulin Secretion and Pancreatic β-Cell Function

Insulin-secreting β cells in the pancreas are capable of 5-HT uptake and contain 5-HT receptors; however, the primary relevant source of 5-HT signaling in islets appears to be intrinsically derived. Pan inhibitors of TPH (and thus, total-body 5-HT) have no relevant effect on plasma insulin during fed or fasted states in mice (22). Similarly, plasma insulin remains unchanged in Tph1 knockout mice (22). β cells express both TPH1 and TPH2 and are thus able to synthesize 5-HT (54). This 5-HT is stored in β-cell secretory granules and is cosecreted with insulin (55–59). 5-HT confers intracellular effects by modifying proteins involved in exocytosis in a process termed serotonylation. This term refers to the process by which transglutaminases covalently couple 5-HT during insulin exocytosis to two key players in insulin secretion, the small guanosine triphosphatases Rab3a and Rab27a, resulting in the augmentation of glucose-stimulated insulin secretion (GSIS) (60).

The most well-characterized roles of 5-HT in pancreatic islets occur through 5-HT receptor signaling. Human islets express a number of 5-HT receptors (61), in addition to SERT, supporting functional roles for 5-HT–receptor stimulation and uptake in β cells. 5-HT–receptor activation has diverse effects on insulin secretion. The binding of 5-HT to 5-HT_1D receptors inhibits insulin secretion in healthy human islets and in β-cell model cell lines (60, 61). This is proposed to create a negative-feedback loop for both basal secretion and GSIS (56–59, 62). In contrast, activation of 5-HT₂ and 5-HT₃ receptors augments the respiratory capacity and membrane excitability of pancreatic islet cells to exert an incretinlike effect by increasing insulin secretion from β cells, while suppressing glucagon release from α cells through activation of 5-HT_1F receptors (61–64). Thus, the effect of 5-HT on insulin secretion is likely dictated by the number and type of 5-HT receptors present within individual β cells.

The diverse roles of 5-HT in pancreatic islets are also illustrated under specific conditions of metabolic change. During pregnancy, the maternal β-cell mass expands in response to increased insulin resistance. Increased β-cell proliferation drives the expansion of β-cell numbers that underpins this compensatory hyperinsulinemia. The increase in β-cell mass is due to prolactin-dependent induction of β-cell TPH1 expression, which leads to increased 5-HT production and the autocrine/paracrine activation of intrinsic 5-HT_2B and 5-HT₃ receptors to enhance cell proliferation (7, 54, 63). The increase in β-cell mass and insulin production plays a critical role in reducing the incidence of gestational diabetes. Serotonin can also augment β-cell function during periods of hyperglycemia, when there is an increased demand for insulin secretion. Mice with β-cell–specific knockout of Tph1 have a maladaptive response to a high-fat diet and display lower GSIS than control mice. This effect is driven by reduced 5-HT availability and can be rescued by exogenous administration of 5-HT (65). The expression of a number of 5-HT receptors is also known to be altered in human islets from individuals with T2D and in islets of mice fed a high-fat diet (61, 62, 65). Such changes are purported to contribute to defective insulin and glucagon secretion in human T2D. However, the mechanisms that drive altered 5-HT signaling in β cells during T2D remain unclear.

Osteoskeletal Regulation

The skeletal system is regulated by a variety of hormonal signals and through sympathetic outflow from the CNS (18, 66). Bone renewal depends on a balance between bone reabsorption by osteoclasts and bone synthesis by osteoblasts. Alterations to this balance considerably affect bone structure. Examples of this include premature ovarian failure and menopause-associated osteoporosis, both of which are associated with a marked loss of bone density and strength (67). The relationship between 5-HT and bone structure was first identified by the decreased bone density that occurs in patients who are administered selective serotonin reuptake inhibitors for the treatment of depression (68, 69). A number of studies have shown that TPH1 is expressed in bone, but at orders of magnitude lower than within the GI tract (70, 71). Gut-derived 5-HT is now clearly established as capable of inhibiting bone turnover and renewal by inhibiting osteoblast and osteoclast differentiation and proliferation via actions at 5-HT_1B and 5-HT_2A receptors, and uptake via SERT (66, 71–75). Correspondingly, inhibition of gut-derived 5-HT has anabolic effects on bone formation in an ovariectomized mouse model of osteoporosis, acting as both a preventative and therapeutic treatment of bone loss (72). Such findings demonstrate that gut-derived 5-HT is a powerful determinant of bone homeostasis.

Bone plays an important role in whole-body metabolism through the release of osteocalcin from osteoblasts. Osteocalcin is a powerful driver of β-cell proliferation and insulin production (76). It also promotes adipocyte thermogenesis and the secretion of adiponectin (76, 77), both of which have positive metabolic effects. Circulating osteocalcin levels are inversely correlated with fasting plasma glucose, glycated hemoglobin, visceral fat mass, and BMI in healthy humans (78) and in individuals with T2D (78–84). In longitudinal studies in humans both before and after the development of metabolic syndrome, plasma osteocalcin levels were found to be an independent risk factor for developing T2D (14, 76, 77, 85, 86). Administration of osteocalcin can also prevent the onset of obesity, impaired glucose tolerance, and insulin insensitivity in mice fed a high-fat diet (76). Inhibition of gut-derived 5-HT in ovariectomized mice markedly increased osteoblast numbers, bone formation, and serum osteocalcin (72). Considering that plasma 5-HT levels increase in T2D (26), gut-derived 5-HT may promote the development of metabolic disease secondary to its inhibition of osteoblast function by decreasing circulating levels of osteocalcin.

EC Cells Are GI Luminal Sensors That Regulate GI Function

EC cells act as important luminal sensors within the intestinal tract by responding to a wide array of luminal and circulatory cues, including nutrient and immune modulators, as well as neurochemical and mechanical stimuli (Fig. 2). Isolating and studying primary EC cells has proven difficult, because EC cells are dispersed throughout the GI tract epithelial cell layer. Although methods for the purification and culture of primary human and rodent EC cells have recently been developed (35, 82, 87–89), the majority of studies of EC-cell function have been undertaken in BON cells, a human carcinoid EC-cell line. Both BON cells and primary EC cells are capable of sensing increases in glucose availability (35, 88, 90). An acute increase in luminal glucose levels also triggers 5-HT release in intact mouse colonic tissue preparations (35) and primary EC cells in culture (88). However, the concentrations of glucose required to trigger gut 5-HT release are far in excess of those observed in the circulation. Such findings clearly illustrate that increases in ingested, but not plasma, glucose levels are sensed by EC cells, evidently through glucose-sensing mechanisms that are located on the apical plasma membrane. This detection of glucose levels occurs as a result of membrane depolarization and subsequent extracellular Ca²⁺ influx through voltage-gated l-type Ca²⁺ channels (35, 87). Despite the fact that almost all glucose absorption occurs over the length of the small intestine, colonic EC cells sense and respond to luminal concentrations of glucose (35, 88, 89). This indicates that, although glucose is absent in the colon under normal physiological conditions, colonic EC cells may be equipped to sense and respond to the presence of glucose under disease states, such as diarrhea or fast transit. A lack of nutrient availability during fasting also stimulates the synthesis and release of 5-HT from EC cells in mice due to increased transcription of Tph1 (22). Similarly, exposure of primary colonic EC cells to low glucose, reflecting in vitro fasting conditions, also induces TPH1 expression and increases 5-HT synthesis and release (35).

Figure 2.

EC cells (shown in purple) within the mucosal lining of the GI tract act as sensory cells for a wide array of nutrient cues within the GI lumen. They also sense neuromodulatory agents from extrinsic and intrinsic afferent nerve fibers, and are responsive to mechanical stimuli, such as intestinal muscle contractions.

EC cells are sensitive to both acute increases and chronic reductions in glucose availability, although the mechanisms governing such responses remain unclear. Glucose entry to EC cells through the Na⁺-dependent glucose cotransporter 1 has been suggested as a potential pathway for glucose-stimulated 5-HT secretion in BON cells (90). Isolated duodenal and colonic EC cells from mice show region-specific expression of several GLUT hexose transporters, which are particularly tuned to sensing glucose and fructose (89) and differentially respond to the presence of these sugars in culture. The sweet taste receptor, T1R2, which forms a broadly tuned G-protein–coupled receptor with T1R3 to detect all known sweet tastants, including sucrose and artificial sweeteners, is expressed in a subset of 5-HT–containing cells within the human duodenal mucosa (91). The artificial sweetener sucralose increases 5-HT release from human EC cells (92), whereas sucrose increases 5-HT release from primary mouse duodenal EC cells (88). Sweet taste–triggered 5-HT release may be important in patients with T2D, who have defective regulation of intestinal sweet taste receptors and exaggerated postprandial glucose absorption (91). The reduced ability of artificial sweeteners to undergo absorption and metabolism has resulted in the increased consumption of these compounds in the population. Given this, sweet taste–triggered 5-HT release may represent a risk to normal blood glucose control, particularly for patients with T2D.

EC cells express a range of G protein–coupled receptors that are tuned to other nutrients, such as tastants, amino acids, and free fatty acids (89, 93). Both primary EC cells and BON cells express the olfactory receptor OR1G1, and olfactory cues from spices and odorants, such as thymol, have the capacity to trigger 5-HT release (92, 94). Pungent compounds, including allyl isothiocyanate and cinnamaldehyde, activate the TRPA1 channel in EC cells to stimulate 5-HT secretion and gut contraction (95). The detection of these types of nutrients in the intestinal lumen is largely restricted to the small intestine; expression of receptors for olfactory and pungent agents is enriched in this region of the GI tract and absent in the colon (96). There is limited knowledge regarding the concentrations of odorants and pungent compounds in the intestinal lumen following oral consumption. However, the rapid absorption and metabolism of thymol (97) indicates limited intestinal exposure to these types of compounds.

A number of receptors for sensing short-chain fatty acids (SCFAs), medium-chain fatty acids, and long-chain fatty acids are expressed in human and mouse EC cells (89, 98), and human colonic EC cells respond to amino acids and some fatty acids, including lauric acid (98). SCFAs such as butyric, acetic, and propionic acids, in particular, occur naturally in the diet, but are also produced de novo by bacteria residing in the gut. The production of luminal SCFAs and secondary bile acids, such as deoxycholate, by gut microbiota, are likely activators of host EC cells, given that both increase 5-HT release in a variety of experimental settings (99–102). SCFAs can act in a chronic fashion on EC cells by increasing TPH1 expression and 5-HT synthesis (24). The interaction of SCFAs with free fatty acid receptor (FFAR) 2 may mediate this response (93); however, clear evidence for this in primary EC cells has not yet been provided.

EC cells respond to neuromodulatory agents (Fig. 2); acetylcholine, γ-aminobutyric acid, and pituitary adenylate cyclase–activating peptide are all capable of increasing EC-cell 5-HT release in vitro (87, 92, 103–106). This suggests that EC cells may also be responsive to neural signals that are derived from the CNS and/or enteric nervous system (ENS). Indeed, adrenergic, cholinergic, and peptidergic inputs from the ENS terminate in close proximity to EC cells (106), and direct stimulation of vagal inputs to the duodenal mucosa significantly increases 5-HT levels in the portal vein of cats (107). Correspondingly, β-adrenergic receptor blockade reduces 5-HT release (106, 108). Although it is plausible that transmitter release from mucosal afferents stimulates 5-HT release, results from studies may be confounded by contraction-induced EC-cell stimulation. EC cells are mechanically sensitive, likely through the mechanosensitive ion channel, Piezo2 (109). 5-HT release correlates with intestinal motility (motor complexes) both in vivo and ex vivo (110). In addition, there is strong evidence that stimulation of EC cells can, in turn, activate the nerve endings of intrinsic primary afferent neurons with cell bodies in myenteric ganglia (106, 111). As such, EC cells have been a focus in efforts to understand the regulatory feedback mechanisms that underlie intestinal motility and transit. Interestingly, although 5-HT is synthesized in some enteric neurons (112), there is limited evidence that 5-HT is a neurotransmitter under normal conditions in mouse colon (113, 114), and rarely (115) or never in the guinea-pig colon (116, 117), human colon (118), and rat colon (119). There is, however, evidence that during inflammation of the gut wall, 5-HT may contribute to fast synaptic transmission (120).

EC-cell–derived 5-HT has long been known to have modulatory effects on GI motility (Fig. 3). This occurs via activation of 5-HT₃ and 5-HT₄ receptors on vagal afferent nerve terminals that innervate the intestinal mucosa (111, 121). Morphological studies demonstrate that myenteric Dogiel type 2 neurons project into the mucosa with nerve endings in close apposition to EC cells (108), and 5-HT potently activates these nerve endings (111). The release of 5-HT from EC-cell vesicles is finely tuned, given that the small amounts of 5-HT released per vesicle (82, 87, 122), compared with other endocrine cells (123–125), is sufficient to activate nearby 5-HT receptors (87). However, release of 5-HT from EC cells is not explicitly required for neurogenic contractions of the gut wall to occur (126, 127), because propagating neurogenic contractions are maintained in vitro in the absence of the intestinal mucosa (17, 128). Furthermore, in vivo intestinal transit is maintained when TPH1 is inhibited pharmacologically (72).

Figure 3.

The effects of 5-HT on GI tract absorption and motility. Release of 5-HT in response to chemical and physical cues increases bile acid turnover, thereby increasing fat absorption, in addition to increasing bicarbonate and electrolyte secretion into the GI lumen. Modulation of intestinal motility occurs via activation of 5-HT₃ and 5-HT₄ receptors on afferent nerve terminals from the ENS.

Although it is not required for the initiation of contractile propagation, gut 5-HT is nevertheless a potent activator of gut motility and gastric emptying, and has established metabolic effects through such pathways. Luminal 5-HT is required for intestinal nutrient and water absorption, and influences bicarbonate and electrolyte secretion into the lumen (Fig. 3) (129–131). Through stimulation of 5-HT₃ receptors on intestinal vagal afferent nerve endings that innervate the pancreas, EC-cell–derived 5-HT indirectly increases postprandial pancreatic enzyme secretion in a manner synergistic with cholecystokinin (CCK) (96). This, in turn, increases digestion and absorption of luminal nutrient contents. Intestinal fat absorption is also mediated by the release of liver-synthesized bile salts and bile acids from the gall bladder, which can be reabsorbed in the intestinal lumen. Gut 5-HT increases bile acid turnover by inducing the expression of the apical sodium-dependent bile salt transporter, in addition to increasing bile acid synthesis and secretion by the liver and gall bladder (21). Thus, gut 5-HT indirectly increases intestinal fat absorption, and may indirectly contribute to obesity via these effects on bile acids.

Hormonal signaling between EC cells and other enteroendocrine cells exists. A portion of EC cells express receptors for glucagonlike peptide 2 (GLP-2) (132), a hormone secreted by intestinal L cells in response to food intake. As such, 5-HT may act as a mediator of several effects that are attributed to both 5-HT and GLP-2, including inhibition of gastric emptying and nutrient-stimulated gastric acid secretion (133–136). In addition, GLP-2–stimulated 5-HT secretion may mediate the reported effects of GLP-2 on bone (a tissue that lacks the GLP-2 receptor), which include reductions in bone reabsorption and bone mineral density (137). In contrast, 5-HT appears to have a stimulatory effect on the secretion of GLP-1 from L cells in ileal tissue segments (138). This effect is also seen in STC-1 cells, an immortalized secretin tumor L-cell model, and is blocked by the addition of a broad-spectrum 5-HT–receptor antagonist (138). Through activation of 5-HT₂ and 5-HT₃ receptors that innervate the submucosal plexus, 5-HT indirectly evokes intestinal CCK secretion by increasing active CCK-releasing peptide (139). Together, CCK and 5-HT act to stimulate postprandial pancreatic enzyme secretion, and contribute to the aforementioned effects on bile acid regulation (140–144).

The expression of enteroendocrine hormones is not classically restricted to a single cell type, given that coexpression of hormones has been highlighted along the length of the gut. Cells containing 5-HT coexpress a combination of other hormones, including substance P, secretin, and CCK (96, 145, 146). Establishing the contributions of these subpopulations of cells to GI and peripheral physiology is complex. Individually, however, these hormones act as potent stimuli within the ENS to alter GI motility and function (147–149), and act synergistically to aid in nutrient digestion and absorption (140, 142–144).

5-HT, the Gut Microbiome, and Metabolism

The gut is host to a densely populated and diverse commensal microbial ecosystem. This microbial community, together with its genetic material and traits that are encoded (collectively, the gut microbiome), performs a range of essential functions, perhaps the most important of which is the digestion of complex carbohydrates and plant-derived polysaccharides (150). The influence of the intestinal microbiome on host metabolism and fat storage is well established (37, 151–155). Intervention studies involving the transplantation of fecal microbiota from lean to obese individuals, from obese individuals to germ-free (GF) mice, or to mice that have reduced microbial abundance and diversity following antibiotic treatment, have clearly demonstrated a causal role of gut microbiota in the dysregulation of host glucose and lipid metabolism (154, 156–158). In particular, depletion of the intestinal microbiota protects mice against adverse changes in glucose tolerance and insulin sensitivity in response to genetic and diet-induced obesity, and increases thermogenesis in BAT and “browning” of WAT (37). These observations have led to intense interest in understanding the mechanistic relationship between gut microbiota and syndromes associated with metabolic dysfunction, such as obesity and T2D.

Serum, plasma, colonic, and fecal concentrations of 5-HT are substantially reduced in GF and antibiotic-treated mice compared with conventionally raised controls (99, 159–161). These decreases in 5-HT result from reduced gut 5-HT biosynthesis, which is supported by associated decreases in TPH1 expression and EC-cell density, particularly in the colon (99, 161). Furthermore, the introduction of an intestinal microbiome in GF mice, for example, through the transplantation of human fecal matter, increases synthesis of EC-cell 5-HT, a shift that can be attenuated by TPH inhibition (24).

Although the intestinal microbiota can synthesize and secrete 5-HT (162), bacterial production of 5-HT is far exceeded by biosynthesis in EC cells. Rather, the resident microbiota of the healthy gut provides ongoing signals to host mucosal EC cells to maintain gut 5-HT content and plasma 5-HT levels, through stimulation of TPH1 expression. Regulation of EC-cell 5-HT biosynthesis by the gut microbiota appears to be maintained by the production of metabolites (Fig. 4), including SCFAs and secondary bile acids (24, 99), particularly by resident spore-forming bacteria (99, 163, 164).

Figure 4.

Ramifications of the interactions between the gut microbiome and EC cells. Gut microbiota (specifically, spore-forming bacteria) play a central role in regulating EC-cell 5-HT biosynthesis and secretion through the release of secondary bile salts such as deoxycholic acid and SCFAs, particularly acetate, butyrate, and propionate. Microbial and dietary SCFAs stimulate the free fatty acid (FFA) receptors, FFAR2 and FFAR3, resulting in an increase in the expression of TPH1 and, consequently, increased 5-HT biosynthesis and secretion. Secondary bile acids activate the G-protein–coupled bile acid receptor, TGR5, to stimulate the secretion of 5-HT. Increased plasma 5-HT can have detrimental consequences in the context of metabolic disease through an overall decrease in insulin sensitivity and glucose tolerance.

Given the central role that the gut microbiome plays in regulating 5-HT biosynthesis, it is unsurprising that diet-driven changes in the composition of the intestinal microbiota, as well as the associated changes in available substrates for bacterial growth, have a marked impact on gut and serum 5-HT levels. However, functional redundancy between bacterial species, as well as the ability of certain taxa to produce metabolites that stimulate 5-HT in response to varied dietary substrates, means that microbial regulation of 5-HT biosynthesis is complex. The impact of diet on microbiota-mediated 5-HT biosynthesis can be illustrated by examining the effects of increases in either plant-derived complex carbohydrates or saturated fats, dietary components that are broadly associated with beneficial and detrimental health implications, respectively. An increased intake of nonabsorbable dietary fiber results in a selective stimulus for bacterial species able to use it for growth, notably members of the genus Clostridia. These bacteria ferment polysaccharides in the colon to produce SCFAs, which, in turn, stimulate EC-cell free fatty acid receptors to trigger 5-HT biosynthesis (Fig. 4) (particularly acetate and propionate, via FFAR2) (93, 100), as well as intestinal gluconeogenesis and insulin-stimulated glucose uptake (particularly propionate and butyrate, via FFAR3) (165). In humans, an increased dietary intake of saturated fats is also selective for Clostridia, as well as members of the phylum Bacteroidetes, but in this case it results from increased levels of bile acid in the gut lumen. The activities of the bile-tolerant Bacteroidetes and Clostridia result in the deconjugation of bile acids and the conversion of the secondary metabolites to deoxycholic acid, respectively. The levels of deoxycholic acid that are produced can stimulate EC-cell 5-HT biosynthesis (24, 99) via the G-protein–coupled bile acid receptor TGR5 (101) (Fig. 4). In parallel, high-fat diet–driven increases in Bacteroidetes and Clostridia, and corresponding decreases in the relative abundance of Bifidobacterium and Lactobacillus, are also likely to result in an alteration of both the absolute and relative levels of microbiota-derived SCFAs. These similar effects of dietary fiber and saturated fats to increase gut 5-HT are in contrast to their opposing metabolic effects. This is just one example of the complexity of the systems physiology within the context of the nutrient/gut hormone/metabolism axis. Additionally, these changes in metabolites that stimulate 5-HT biosynthesis take place in a wider context of altered bacteria-mediated energy harvest. An additional layer of complexity exists because of the potential effects of the gut microbiome on the nervous system (166), and interactions between other enteroendocrine cells. Understanding the complex interactions between diet, gut microbiota, EC cells, the nervous system, and host metabolism in more detail may be of key clinical relevance.

Concluding Remarks

There is growing evidence that gut-derived 5-HT provides a signaling nexus between diet, gut microbiome, and metabolism. Peripheral serotonin contributes to overall energy homeostasis through effects on a number of important metabolic sites, including the gut, pancreas, liver, fat, and bone. Plasma 5-HT increases in diabetes and obesity, and factors such as ingested nutrients and the gut microbiome affect gut 5-HT secretion and synthesis.

Acute stimulation of EC cells in response to luminal nutrients such as glucose has the capacity to increase the release of 5-HT. In this setting, increased circulating 5-HT may aid in nutrient digestion and absorption through increases in pancreatic enzyme secretion and bile acid turnover, while modulating intestinal motility to allow for maximum energy harvest. Concurrently, endogenous β-cell 5-HT contributes to GSIS to help maintain postprandial blood glucose levels. Cross-talk between EC cells and other enteroendocrine cells is also likely to contribute to the finely tuned metabolic consequences following ingestion of certain nutrients. In contrast, alterations in energy availability during more chronic situations, such as fasting, and through interactions with the gut microbiome, increase the biosynthesis of 5-HT at the level of increased Tph1 transcription, and through an increase in EC-cell density. In response to chronic exposure to such conditions, circulating 5-HT is likely to act on longer-term processes to increase blood glucose levels, and therefore energy availability, through increased hepatic gluconeogenesis and lipolysis from adipose tissue. In conjunction, 5-HT conserves energy by decreasing BAT thermogenesis and WAT browning, while also decreasing bone turnover, therefore reducing other prometabolic hormones such as osteocalcin and adiponectin. The chronic consumption of a high-fat or high-sugar diet may therefore shift the nature in which EC cells respond to their environment, resulting in a change from physiological outcomes associated with acute nutrient exposure to those consistent with chronic stimulation. Although substantial advances have been made in understanding the role of 5-HT in metabolism and metabolic disorders, considerable challenges remain.

EC cells are tuned to a wide range of luminal stimuli, and release 5-HT in response to dietary glucose and fatty acids; however, the precise mechanisms underlying this have yet to be elucidated. In the face of increasing exposure to sugars, artificial sweeteners, and fats in Western diets, and causal links to the development of obesity, insulin resistance, and T2D, the relationship between these dietary nutrients and 5-HT release warrants deeper investigation. Our understanding of connections between the gut microbiome and metabolism are also gaining momentum, with evidence now in strong support of endogenous peripheral 5-HT as a key mediator. Further knowledge of the nature of EC-cell/microbiome communication is required to shed light on the complex host-microbiota interactions that exist within the gut. In addition to direct effects on key metabolic tissues, 5-HT may further exacerbate metabolic disorders by limiting hormones with beneficial metabolic roles, such as osteocalcin and adiponectin; however, these mechanisms await further clarification. The key question of what drives gut 5-HT synthesis in humans, and how this relates to human metabolic diseases, remains an important and clinically relevant question to be answered.

Abbreviations:

 
• 5-HT

  5-hydroxytryptamine

 
• BAT

  brown adipose tissue

 
• BMI

  body mass index

 
• cAMP

  cyclic adenosine monophosphate

 
• CCK

  cholecystokinin

 
• CNS

  central nervous system

 
• EC

  enterochromaffin

 
• ENS

  enteric nervous system

 
• FFAR

  free fatty acid receptor

 
• GI

  gastrointestinal

 
• GLP-2

  glucagonlike peptide 2

 
• GSIS

  glucose-stimulated insulin secretion

 
• HSL

  hormone-sensitive lipase

 
• iBAT

  interscapular brown adipose tissue

 
• SCFA

  short-chain fatty acid

 
• T2D

  type 2 diabetes

 
• TPH

  tryptophan hydroxylase

 
• UCP1

  uncoupling protein 1 expression

 
• WAT

  white adipose tissue.

Acknowledgments

Disclosure Summary: The authors have nothing to disclose.

Author notes

Address all correspondence and requests for reprints to: Damien Keating, PhD, Discipline of Human Physiology and Centre for Neuroscience, Flinders University of South Australia, Adelaide 5001, Australia. E-mail: [email protected].