Central serotonergic systems have long been appreciated to have key roles in the regulation of feeding and body weight (for review, see Lam and Heisler 2007; Nelson and Gehlert 2006; Heisler et al. 2003). Studies of mutant mice with functional alterations of 5-HT2CR activity have elucidated some of the mechanisms underlying this important regulation.
In the hypothalamus, 5-HT2CRs are prominently expressed on POMC-containing neurons of the ventrolateral arcuate nucleus (Wright et al. 1995; Pasqualetti et al. 1999; Heisler et al. 2002), and serotonin binding to these receptors is thought to contribute to neuronal activation (Qiu et al. 2007). POMC neurons in the arcuate are also activated by leptin signaling through the leptin receptor (Cheung et al. 1997; Elias et al. 1999; Cowley et al. 2001) and serotonin signaling through the 5-HT1B receptor (Ho et al. 2007; Nonogaki et al. 2007). POMC neurons in turn signal the ventromedial hypothalamus (VMH) via mel-anocortin neurotransmission; binding of melanocortin to the melanocortin 4 receptor (MC4R) inhibits expression of feeding behavior (Cowley et al. 1999; Adage et al. 2001). This anorexigenic pathway is in opposition to signaling through anorexigenic pathway consisting of neuropeptide Y (NPY)/agouti-related protein (AgRP) neurons (which are inhibited by leptin signaling, see Yokosuka et al. 1998) projecting to VMH via NPY neurotransmission (which stimulates feeding behavior) and agouti-related protein (AgRP) neurotransmission (which is a functional antagonist of the MC4R, and thus blocks the inhibition of feeding behavior, see Morton and Schwartz 2001). This "canonical" circuit is depicted in Fig. 4.2.
Multiple studies have examined the effects of 5-HT2CR loss on overall mouse weight, and consensus results suggest that while 5-HT2CR loss does not initially influence body weight of newborn mouse pups, weanlings, or young adults (<3 months old), constitutive 5-HT2CR /Y mice develop a progressive middle-aged onset obesity. This obese phenotype becomes more exaggerated when comparing body-weight trajectories of mice fed a high fat diet (Nonogaki et al. 1998). Of note, this obesity phenotype occurs despite observing only modest hyperphagia (particularly during the light cycle, when the animals are expected to be inactive) in young constitutive 5-HT2CR /Y mice (Goulding et al. 2008).
As depicted in the Fig. 4.2, current models of ingestive behavior neurophysiol-ogy suggest that 5-HT2CR related feeding regulation occurs through stimulation of POMC neurons, which in turn excite MC4Rs and inhibit expression of feeding behaviors. Supporting this hypothesis, recent studies have demonstrated decreased expression of both POMC and a nucleobindin 2 (NUCB2, a satiety-related molecule) in the hypothalamus of constitutive 5-HT2CR /Y mice compared with wild types (Nonogaki et al. 2008). Additionally, stimulation of 5-HT2CRs using mCPP leads to dose-dependent increases in hypothalamic NUCB2 expression that are not
Fig. 4.2 (a) Integration of leptinergic and serotonergic signaling at the hypothalamic arcuate nucleus regulates both orexigenic and anorexigenic neuronal pathways controlling feeding behavior. (b) Loss of 5-HT2CR function leads to decreased signaling within the pro-opiomelanocortin (POMC) neuronal population, decreased activation of anorexigenic signaling within the hypothalamus, and an imbalance between activities of the orixigenic and anorexigenic pathways favoring increased food consumption. 5-HT2CR indicates serotonin2C receptor, 5-HT1BR serotonin1B receptor; MC4R melanocortin 4 receptor, Y1 neuropeptide Y receptor 1, POMC pro-opiomelano-cortin, NYP neuropeptide Y, AgRP agouti-related protein, CRH cortisol-releasing hormone, CART cocaine- and/or amphetamine-regulated transcript, MCH melanin-concentrating hormone
Fig. 4.2 (a) Integration of leptinergic and serotonergic signaling at the hypothalamic arcuate nucleus regulates both orexigenic and anorexigenic neuronal pathways controlling feeding behavior. (b) Loss of 5-HT2CR function leads to decreased signaling within the pro-opiomelanocortin (POMC) neuronal population, decreased activation of anorexigenic signaling within the hypothalamus, and an imbalance between activities of the orixigenic and anorexigenic pathways favoring increased food consumption. 5-HT2CR indicates serotonin2C receptor, 5-HT1BR serotonin1B receptor; MC4R melanocortin 4 receptor, Y1 neuropeptide Y receptor 1, POMC pro-opiomelano-cortin, NYP neuropeptide Y, AgRP agouti-related protein, CRH cortisol-releasing hormone, CART cocaine- and/or amphetamine-regulated transcript, MCH melanin-concentrating hormone observed in constitutive 5-HT2CR -/Y mice. These results suggest that serotonergic neurotransmission onto these ventrolateral arcuate POMC neurons is a major factor driving the expression of both an anorexigenic-related transcription factor and the melanocortin transcript.
The functional status of this above-described pathway appears to be necessary for expression of SSRI-evoked anorexia. For example, fluvoxamine does not mediate its usual anorectic effect in mice with intact melanocortin signaling but constitutively lacking functional 5-HT2CRs (Nonogaki et al. 2009). However, fluvoxamine again has anorectic action (potentially through 5-HT1B receptor signaling) in mice with a combined homozygous deletion of 5-HT2CRs and a hemizygous mutation of b-endorphin. Thus, SSRI-evoked anorexia appears to require intact signaling at both the 5-HT2CR/ POMC neuron and MC4R/arcuate neuron. Mice with constitutive 5-HT2CR knockout also show a similar resistance to the anorectic effects of d-fenfluramine, a potent stimulator of serotonin release and inhibitor of serotonin reuptake (Vickers et al. 1999). This finding was seen in analyses of behavioral satiety sequences, where it was noted that intermediate doses of d-fenfluramine had a much-blunted effect of suppressing feeding in constitutive 5-HT2CR /Y mice compared with its effect in wild types.
Double mutant mice constitutively lacking both 5-HT2CRs and the adipocyte hormone leptin (htr2c~/Y, ob-/ob-) provide particularly convincing evidence that signal integration at the POMC neuron plays a vital role in control of ingestive behaviors. Young double mutant mice had many hallmarks of significant diabetes, including hyperphagia and marked polydipsia, obesity, hyperglycemia, aberrant glucose tolerance tests, glucosuria, and hyperinsulinemia (Wade et al. 2008). Double mutant mice also had markedly elevated serum concentrations of the antiinsulin counterregulatory hormones corticosterone and glucagon. While there were no phenotypic differences appreciated in the above outcomes for young mice carrying a single constitutive 5-HT2CR deletion, addition of the 5-HT2CR mutation on top of the ob-/ob- genetic background markedly increased the size of each of these deficits well above the baseline ob-/ob- phenotype.
Similarly, mice lacking 5-HT2CRs demonstrate a complex interplay regarding how 5-HT2CRs and 5-HT1BRs regulate food intake. Wild-type mice receiving the 5-HT1BR agonist CP-94,253 demonstrate a modest decrease in proportion of time spent feeding and performing other behaviors as evaluated by behavioral satiety sequences (over a 5-min bin resolution) (Dalton et al. 2006). However, performing this same treatment on 5-HT2CR mutant mice led to a dramatic decrease in feeding and other activities. These studies suggested that 5-HT1BR mediated hypophagia occurs independently of functioning 5-HT2CRs. Since 5-HT2CR mutation does not lead to significant changes in expression of other serotonin receptors (Lopez-Gimenez et al. 2002), these results also suggest that persistent loss of 5-HT2CR function effects cellular responses to 5-HT1BR stimulation. Furthermore, contrary to expectation, middle-aged mice carrying double mutations for both the 5-HT2CR and 5-HT1BR are less heavy than 5-HT2CR-mutant, 5-HT1BR intact littermates (Abdallah, personal communication).
Interestingly, htr2c transcripts are one of the relatively few gene products that undergo posttranscriptional editing, an unusual RNA regulatory process where individual transcript nucleotides are enzymatically modified (usually to inositol, a process that changes their complementary binding to tRNAs and thus their peptide sequence). This process occurs through the action of multiple enzymes, predominantly through adenosine deaminase acting on RNA-1 (ADAR1, which selectively edits adenosines on residues 157, 159, and 161, see Hartner et al. 2004), as well as ADAR2 and ADAR3. Multiple, different posttranscriptional 5-HT2CR variants have been described, with the main difference being the overall level of constitutive signal transduction activity in the absence of bound serotonin. The INI variant results from an unedited htr2c transcript, while the VGV variant results from a fully edited htr2c transcript. INI variant 5-HT2CRs demonstrate the greatest constitutive activity and native ligand affinity; conversely, VGV variant 5-HT2CRs demonstrate the least constitutive activity and weakest native ligand affinity (Herrick-Davis and Niswender 1999). Recent studies using transgenic mice expressing solely the INI or VGV 5-HT2CR variants demonstrate that how htr2cs undergo posttranscriptional editing has a dramatic effect on mouse phenotype and feeding regulation. While mice solely expressing the INI 5-HT2CR variant grew normally compared with wild-type littermates, mice expressing the VGV 5-HT2CR variants demonstrated dramatically decreased fat mass, marked hyperphagia, increased energy expenditure and sympathetic tone, and increased overall 5-HT2CR expression (Kawahara et al. 2008). Of note, the VGV phenotype was not suppressed in mice that expressed only the VGV 5-HT2CR variant and a homozygous MCR/ MC4R- mutation. This data suggests that 5-HT2CR mediated control of food ingestion is mediated through both MC4R-dependent and -independent pathways.
It is intriguing to note that many of the differences in feeding behavior observed between constitutive 5-HT2CR -/Y mice and their wild-type counterparts can be fully reversed by selective reintroduction of 5-HT2CR expression into hypothalamic POMC neurons (Xu et al. 2008). In this study, a 5-HT2CR construct was generated containing a transcription blocker flanked by loxP sites between exons 3 and 4 (loxTB 5-HT2CR, as per authors). This construct was substituted for the native 5-HT2CR allele, and mice containing this construct were bred to transgenic mice expressing Cre recombinase under control of the POMC promoter (POMC-Cre). In loxTB 5-HT2CR/POMC-Cre mice, expression of Cre recombinase is restricted to POMC containing neurons. In these POMC neurons, Cre recombinase expression leads to excision of the transcription blocker site, thus restoring intact 5-HT2CR expression in these cells. Dramatically, restoring 5-HT2CR expression to POMC neurons alone normalized many of the above-described deficits. For example, loxTB 5-HT2CR/POMC-Cre mice do not display middle-aged obesity, hyperadipos-ity, or hyperleptinemia. Additionally (and unexpectedly), both the daily and dark cycle locomotor hyperactivity seen in mice constitutively lacking 5-HT2CR function is not observed in loxTB 5-HT2CR/POMC-Cre mice.
Was this article helpful?