The 5-hydroxytryptamine (serotonin) receptor 2C (HTR2C) is one of a family of receptors that is involved in coordinating the intracellular responses to serotonin in the mammalian nervous system 1 2 . This particular protein couples the binding of ligand to the surface receptor domain with the mobilization of intracellular calcium via G-protein coupling and activation of protein kinase pathways. Although the best understood activities of HTR2C are those elicited via activation of phospholipase C (PLC), there is evidence that other signalling cascades such as phospholipase D and A2 pathways can be stimulated 3 .

The full-length HTR2C receptor is a seven transmembrane domain protein, and has a phosphorylation site in the cytoplasmic tail. It is a target for drug agonists and antagonists used to treat a variety of physiological and psychiatric conditions; the structures of the receptor-ligand complexes have been computationally predicted based on successes with other G-protein coupled receptors such as the β2 adrenergic receptor and D2 dopamine receptor 4 . The exon encoding the second intracellular loop, which interacts with G-proteins, is subject to a complex series of mRNA editing events representing 24 potential isoforms of the protein 5 6 . The potential contribution of each amino acid change to receptor activity in vivo has been tested by studies using PLC activation as a reporter 3 7 8 . An additional level of protein regulation occurs through a cryptic internal splice site in the third coding exon which results in premature termination of the open reading frame. In rodents, both the splice junction usage and the mRNA editing are thought to be regulated via non-coding RNAs. In particular the snoRNAs of the Snord115 locus, syntenic with the Prader-Willi Syndrome locus on human chromosome 15 9 , is implicated in mouse models 10 11 . However, the role of the syntenic locus in humans has been controversial, and more recent analyses have revealed mir-22 as a regulator of not only HTR2C mRNA expression, but also of other candidate genes for panic disorder 12 . Unfortunately, this latter study does not report which isoform(s) of the transcripts are affected. Alternative experimental evidence suggests that the intramolecular dsRNA loops formed by the full length HTR2C RNA transcript are sufficient to promote editing via the ADAR proteins, and that inhibition of the secondary structures through the introduction of an RNA binding peptide opens the possibility of novel, editing-directed therapies in the future 13 . It is proposed that this intricate series of transcriptional modifications allows fine-tuning of the serotonergic and dopaminergic pathways in response to neuronal cell types, stress and circulating neurotransmitter concentrations. All act to down-regulate the constitutive activity of the unedited, full-length form of the protein. Furthermore, detailed sequencing analysis of the extent of editing in the mouse brain confirms that it varies according to the stage of development from the embryo through to the adult 14 : thus, the regulatory mechanisms are highly responsive both spatially and temporally.

Advances in our understanding of the regulation of expression of the HTR2C receptor through transcriptional control provide additional insights into the physiological roles of the protein. HTR2C is central to a number of homeostatic and behavioural responses. These include control of circadian rhythms, leptin regulation of appetite, energy expenditure and bone mass, anxiety, depression, and the control of the release of other hormones such as pro-opiomelanocortin and prolactin from the pituitary gland 15 16 . These serotonin responses are elicited mainly via those HTR2C receptors present in the limbic system and the hypothalamus. In contrast, using a variety of agonists and antagonists both HTR2C and HTR2A receptors in the prefrontal cortex of the rat brain have been shown to be required for locomotor activity 17 .

Overall, HTR2C is an attractive candidate to study in animal models of behavioural responses, particularly aggressive behaviours, and has been investigated at the genetic and functional levels in a range of species including rat, mouse, dog and chicken 18 19 20 21 22 23 . In human neurodevelopmental disorders such as attention deficit hyperactivity disorder (ADHD) 24 and psychiatric disorders such as puerperal psychosis 25 there is a modest association with HTR2C specific haplotypes, but many genetic association studies in humans are inconclusive or contradictory. More detailed recent analyses looking at the levels of gene expression, protein activity and the potential isoform composition in different brain subregions with relevance to genotypes and behaviours support a role for editing, not just polymorphism, at the HTR2C locus, which could explain why genetic association studies have been inconclusive 26 27 28 29 30 . The argument for a central role for HTR2C receptor in psychoses, anxiety disorders and stress based on the biological and functional evidence remains strong, and has been reviewed by Drago and Serretti 31 .

Our previous studies have identified some modest changes in HTR2C transcriptional levels in the hypothalamic regions of sows showing infanticidal aggressive behaviours when compared with their non-aggressive counterparts 32 . In a parallel study to identify QTL contributing to the behaviour, we identified at least two peaks on the porcine X chromosome, one of which lay on the long arm 33 . This locus on the long arm of the X chromosome has been independently confirmed in a porcine resource population created from European White Duroc crossed with Chinese Erhualian 34 . Of those genes already implicated in human mood disorders and rodent behaviours, HTR2C is both a functional and positional candidate for further investigation, as it maps to porcine Xq22-23.

In this paper we describe the analysis of the porcine gene at both the genomic and cDNA levels. The open reading frame of the receptor has been scanned for polymorphisms, and the mRNA editing of the transcript investigated in the hypothalamic samples from sows showing maternal infanticide and their matched controls. Overall, we found evidence for increased mRNA expression but not for significant differences in the editing profiles in the hypothalamus of infanticidal animals.

This paper describes the structure of the porcine HTR2C transcripts as determined using RT-PCR with cDNA prepared from a pool of mRNA isolated from brain tissue. The full-length open reading frame sequence is most analogous to human transcript ENST00000276198 (reference sequence NM_000868). Although exon 3 of human transcript ENST00000371951 is conserved at the genomic level, we assume that this isoform and its orthologues have limited expression in humans and in other mammals, and may be confined to specific neurons in the brain. Two major splice sites define the 3' end of exon 6: both are canonical splice junctions. The first splice junction generates the short form of the protein, which leads to premature truncation of the polypeptide, the loss of transmembrane domains 6 and 7 and the intracellular region thought to be essential for G protein binding. The position of the splice site is analogous to that in human and rodent genes: however, in each species the truncation of the polypeptide occurs at a unique position within the final coding exon. The second splice junction generates the full length transcript, but the complex post-transcriptional editing mechanism generates a range of different isoforms that modify the ability of the receptor-ligand complex to initiate and perpetuate phospholipase dependent intracellular signalling pathways.

Only one of the four polymorphisms in the transcript was identified in the protein coding region in this study (S275N). This amino acid lies in the linker region between transmembrane domains 5 and 6, at a site not previously implicated in functional changes to HTR2C proteins. Any impact of this change on downstream signalling pathways is unknown.

For each of the two breed backgrounds in this study, three of the four pairs show upregulation of HTR2C gene expression in infanticidal animals. This is consistent with previously published results 32 using array probe data and qRT-PCR primers from the final coding exon, where a trend for increased expression of HTR2C in infancticidal animals was noted. In this paper, we have looked at both the relative quantity of the longer and shorter isoforms, and the potential impact of editing on protein expression between the two groups. Our results support increased mRNA expression in the hypothalamic samples from the infanticidal animals. It is interesting that there is a more significant difference between the two behavioural states for the relative quantity of the longer isoform since this is potentially the most active variant of the receptor, which can subsequently be modified through the action of the editing process. However, there is no statistically significant difference in the qualitative data we have obtained for the degree of mRNA editing.

One obvious limitation to this study is the small number of available samples from animals of the same genetic background. However, the implication of the three-way ANOVA is that breed may also be relevant in determining the overall expression of the receptor, and the observed breed specific differences in the incidence of abnormal maternal behaviour support such a complex, multifactorial model.

Although qualitative changes have been described in a number of other studies relating to both human and rodent populations with presumed genetic or drug-induced behavioural differences, it should also be noted that most of these have analysed different regions of the brain from that described in this paper. The distribution of HTR2C isoforms is remarkably diverse and dependent upon the neuroanatomical region under investigation 14 . In general, the evidence for editing patterns influencing behaviour has been replicated in studies of material collected from the (pre)frontal cortex in both humans 30 and rat 17 18 20 . All of these papers support a role for the gene in anxiety related behaviours, including some forms of aggression. In contrast, using whole brain mRNA from a murine model of Lesch-Nyhan Disease 19 , Bertelli et al. showed that although the level of the HTR2C transcript is elevated, the pattern of editing per site is not affected, which is entirely consistent with our observations here. Again, this raises the issue of whether significant changes are confined to specific sub-sets of neurones within the brain, and that the explicit role of editing can only be fully understood with more precisely dissected samples.

To date, the only positive correlation between behaviour and serotonin receptors within the hypothalamus is with HTR2A. This receptor is de-sensitized preferentially following prolonged glucocorticoid exposure 36 . This implicates the HTR2A receptor in the dampened HPA axis response following prolonged stress, and specifically the action of HTR2A bearing neurones within the paraventrical nucleus (PVN). However, the HTR2C receptor did not seem to be directly involved in this model.

Although HTR2C has been implicated in genetic studies of human behavioural abnormalities, many of these have focussed on using well described polymorphic markers, either singly or as haplotypes. Results are contradictory, and currently the role of this specific receptor remains inconclusive. At the DNA level, the locus is complex, since the intronic regions of the gene also harbour loci for non-coding micro-RNA (miRNA) species. Such RNA molecules are known to be expressed at a high level in the brain, and regulate the expression of downstream targets. In addition HTR2C has itself been identified as a potential target for miRNAs encoded elsewhere in the genome. With such molecular and regulatory complexity, clarification of the evidence will require further experimental work before the HTR2C region of the X-chromosome can be excluded from the list of candidate loci involved in the control of behaviours. Firstly, the contributions of HTR2C gene expression in different neuroanatomical regions of the porcine brain need to be analysed more precisely, and, secondly, we need more understanding of how the associated miRNAs regulate expression patterns of both the HTR2C locus, and other brain expressed genes.

This study supports elevated levels of gene expression in infanticidal animals, but has not yet provided direct evidence for the involvement of HTR2C editing in the hypothalamus. Further work to investigate the distribution of the respective splice variants in cross-sections representing more detailed subregions of the brain, combined with a fuller understanding of the biological functions of the encoded miRNAs within the gene are still required. Such in-depth analysis may help to explain why there are recurrent data that implicate this locus in behavioural phenotypes in the absence of consistent genetic correlates.

The authors declare that they have no competing interests.

CRQ and CAS provided the project ideas, design, management and interpretation of results. CRQ and MB collected samples, and prepared cDNA for the experimental sections, while AM and FJ carried out PCR experiments, subcloning, sequencing and virtual translations of the products. All authors read and approved the final manuscript.