Serotonin Receptor Expression in Human Prefrontal Cortex: Balancing Excitation and Inhibition across Postnatal Development

Serotonin and its receptors (HTRs) play critical roles in brain development and in the regulation of cognition, mood, and anxiety. HTRs are highly expressed in human prefrontal cortex and exert control over prefrontal excitability. The serotonin system is a key treatment target for several psychiatric disorders; however, the effectiveness of these drugs varies according to age. Despite strong evidence for developmental changes in prefrontal Htrs of rodents, the developmental regulation of HTR expression in human prefrontal cortex has not been examined. Using postmortem human prefrontal brain tissue from across postnatal life, we investigated the expression of key serotonin receptors with distinct inhibitory (HTR1A, HTR5A) and excitatory (HTR2A, HTR2C, HTR4, HTR6) effects on cortical neurons, including two receptors which appear to be expressed to a greater degree in inhibitory interneurons of cerebral cortex (HTR2C, HTR6). We found distinct developmental patterns of expression for each of these six HTRs, with profound changes in expression occurring early in postnatal development and also into adulthood. However, a collective look at these HTRs in terms of their likely neurophysiological effects and major cellular localization leads to a model that suggests developmental changes in expression of these individual HTRs may not perturb an overall balance between inhibitory and excitatory effects. Examining and understanding the healthy balance is critical to appreciate how abnormal expression of an individual HTR may create a window of vulnerability for the emergence of psychiatric illness.


Introduction
In the adult, modulation of the prefrontal cortex by the neuromodulator serotonin is critical for emotional regulation and resilience to stress [1,2,3,4]. The prefrontal cortex is a late maturing brain region [5,6,7,8,9] with an extensive interrelationship with the serotonin system [10,11,12]. Although the serotonergic innervation of the prefrontal cortex matures early in primate postnatal brain development [13], its effects on the developmental neurophysiology and growth of the human prefrontal cortex are much less well understood. Medicines that target the serotonin system are used to treat symptoms of anxiety or depression but appear to be less effective during childhood and have an increased risk of adverse effects when administered to children as compared to adults [14,15]. In fact, recent work suggests that exposure to serotonergic medicines in development increases behaviours suggestive of greater anxiety and sensitivity to stress [16,17].
Serotonin has been implicated as a trophic factor in brain development [18,19,20,21], and its effects show significant developmental changes [22,23]. It is thought that these developmental changes in the functional and pharmacological effects of serotonin are due to changes in the expression of individual postsynaptic receptors [22,24]. In humans, developmental changes in the expression of serotonin receptors (HTRs) would alter the functional effects of serotonin and serotonergic medicines, yet developmental changes in HTRs have not been systematically examined in human prefrontal cortex.
Postsynaptic serotonin receptors control how the prefrontal cortex responds to serotonin at baseline and to the increased release of serotonin during stress [1]. The most well-studied receptors, HTR1A and HTR2A, are highly expressed in excitatory neurons of prefrontal cortex [25,26,27,28]. Under baseline conditions, endogenous serotonin release triggers an Htr1Amediated inhibition of prefrontal cortex [11,29,30] with Htr2Amediated effects potentially recruited at higher levels of serotonin release [25,30]. Behaviourally, these two serotonin receptors in adult prefrontal cortex appear to have opposite effects on anxiety and mood, with high levels of the inhibitory HTR1 being anxiolytic [31,32], the excitatory Htr2A being required for normal anxiety levels [33] and higher HTR2A levels being associated with mood disturbance [34,35]. Work in rodents suggests that there is a striking developmental change in the functional balance of Htr1A and Htr2A expression [22,23]; this relationship has not yet been examined in the developing human prefrontal cortex.
Beyond the well-studied HTR1A and HTR2A, the serotonin receptor family is large, consisting of 14 members in total. Little is known about the human developmental trajectories of all the serotonin receptors expressed in prefrontal cortex. In this work, we examine HTR1A and HTR2A together with a selection of the other receptors, including HTR2C, HTR4, HTR5A, and HTR6. The excitatory Htr2C receptor appears to be more strongly expressed in cortical interneurons [36,37,38] and plays a critical role within prefrontal cortex in controlling impulsivity [39,40]. The inhibitory Htr5A receptor is expressed in a significant portion of cortical pyramidal neurons [41,42,43,44] and may influence anxiety levels under stress [45]. The excitatory Htr4 and Htr6 are thought to be expressed respectively in pyramidal neurons [22,46,47] and interneurons [48,49] and are both important in cognition [50,51].
We examine the developmental changes in the expression of these serotonin receptors in human prefrontal cortex from infancy to adulthood. Since serotonin receptors comprise a large family with diverse localization and functions, we have chosen six receptors with a variety of expression patterns and neurophysiological coupling. We have also made particular effort to relate developmental changes in the HTRs expression to developmental changes in inhibitory interneuron markers in order to determine if changes in serotonin neurotransmission may be synchronized with the maturation of interneuron subtypes. This human developmental information is important in order to appreciate vulnerable time periods in the postnatal prefrontal cortex and to gain insight into potential mechanisms underlying changes in the effects of serotonin and serotonergic medicines.

Human Postmortem Brain Samples
Human postmortem tissue from the dorsolateral prefrontal cortex was obtained from the NICHD Brain and Tissue Bank for Developmental Disorders at the University of Maryland, Baltimore, MD, USA (contract HHSN275200900011C, Ref. No. N01-HD-9-0011). Written consent was obtained from individuals or their next of kin before tissue donation. Samples were obtained from 59 individuals who ranged in age from six weeks to 49 years and were grouped into seven developmental periods: neonates (n = 8), infants (n = 13), toddlers (n = 7), school age (n = 7), teenagers (n = 7), young adults (n = 9) and adults (n = 8). Demographic details and sample characteristics are summarised in Table 1 with full details available in Table S1. Sample preparation for mRNA expression analyses have been described previously [52]. This study was carried out in accordance with the latest version of the Declaration of Helsinki after specific approval by the University of NSW Human Research Ethics Committee (HREC # 07261).

Gene Expression Normalization
In order to reduce the variability inherent in measuring mRNA from this number of individuals the data was normalized through the use of housekeeping genes. We chose four -b-glucuronidase (GUSB), hydroxymethylbilanesynthase (HMBS), peptidylprolylisomerase A (PPIA), and ubiquitin C (UBC) -which did not change through development (ANOVA, GUSB; F(6,61) = 0.5 , p = 0.8, HMBS; F(6,61) = 0.5, p = 0.7, PPIA; F(6,60) = 0.4, p = 0.9, UBC; F(6,61) = 0.08, p = 0.7). The stability of the expression of these genes was assessed through the use of the geNorm program which calculates the average pair-wise variation of a particular gene with all other control genes yielding an M-Value with lower number indicating greater stability (GUSB: 0.871, HMBS: 0.811, PPIA: 0.876, UBC: 0.923) [54]. We have avoided ribosomal mRNAs, which are expressed at very high levels, and instead chose two high expression and two medium expression genes. The chosen housekeepers are involved in the varied biological processes of the cell following the criteria previously established [54]. The housekeeping genes were averaged through the use of a geometric mean which was then used to normalize the HTR gene expression data.

Statistical Analysis
Population outliers were identified for each transcript measured through qPCR using the Grubb's test (p,0.05), resulting in the removal of the following individuals: one toddler from HTR2C, one young adult and one adult from HTR4, one young adult from HTR5A, and two infants and an adult from HTR6. All data were normally distributed (Lilliefors) and showed homogeneity of variance (Levene's, p.0.20). Statistical tests were performed using PASW statistics (version 18 for MacOSX, SPSS, Chicago, Il, USA) and included one-way analysis of variance (ANOVA) with a post hoc Fisher LSD to determine significant changes in gene expression between groups. Using normalized qPCR data, Pearson correlations were performed between the genes of interest and the brain sample characteristics (pH, RIN, PMI) with significant correlations being co-varied in an Analysis of Covariance (ANCOVA). For Figure 1b and to calculate the relative balance between HTRs with excitatory and inhibitory contributions, the measurements for HTR5A were multiplied by 10 to correct for the difference in cDNA dilution used in detection. Only individuals who had qPCR measurements from all six HTRs were included in the ANOVA to assess for changes in the expression ratio of excitatory and inhibitory receptors (neonate = 8, infant = 11, toddler = 6, school age = 7, teenage = 7, young adult = 6, adult = 6). Identifying significant linkages between the mRNA expression of the HTRs expressed by interneurons and other interneuron markers was done using a stepwise regression analysis and the interneuron marker mRNA expression data from Fung et al. [7]. This method allowed us to include only interneuron markers whose developmental expression patterns are significantly associated with developmental changes in individual HTR mRNA expression.

Results
To investigate developmental changes in the expression of a subset of important serotonin receptors in human prefrontal cortex, we performed qPCR on a large sample of postmortem brains from infancy to adulthood. The HTRs we targeted include the two major serotonin receptors expressed in the prefrontal cortex, HTR1A and HTR2A, as well as HTR2C, HTR4, HTR5A, HTR6. As illustrated in Table 2, these receptors include representatives of all three classes of G-protein coupling (Ga i , Ga q , and Ga s ), which have distinct effects on cellular excitability. However, to appreciate the potential net overall effect of each receptor on the excitability of the prefrontal cortex, this receptor coupling must be viewed in the context of cellular localization. As illustrated in Table 3, some HTRs are predominantly expressed by excitatory pyramidal neurons whereas others are mainly expressed in inhibitory interneurons. Lastly, we looked at possible confounds to HTR mRNA expression levels arising from our sample characteristics.
HTR1A and HTR5A: Ga i -coupled receptors Differences were observed with qPCR in the expression in one of the two inhibitory serotonin receptors in prefrontal cortex, as illustrated in Figure 1. We find that the mRNA for HTR1A does not change significantly across the age range evaluated (F(6,53) = 0.6, p = 0.8). This developmental pattern contrasts with that for HTR5A mRNA expression which remains consistent across postnatal development, but then climbs sharply in adulthood (ANOVA: F(6,53) = 2.9, p,0.02, Fisher LSD: p,0.02 adult vs. infant, neonate, school age, young adult), as illustrated in Figure 1b.

HTR2A and HTR2C: Ga q -coupled receptors
There is strong developmental regulation of HTR2A mRNA in the human prefrontal cortex (F(6,53) = 5.3, p,0.001). As illustrated in Figure 2a, this receptor shows a low level of expression in infancy, increases during the infant years, plateaus in the toddler to teenager period, and then declines to adult levels.
The similarly-coupled HTR2C appears to have opposing behavioural effects to the HTR2A [39,40] potentially as a result of different cellular expression [36,37,38,46], as illustrated in Table 3. Here, we find that in human prefrontal cortex HTR2C mRNA shows gradual developmental up-regulation (F(6,52) = 4.7, p = 0.001). Figure 2b illustrates that prefrontal HTR2C expression reaches a highest expression during young adulthood.
In contrast to the anatomical location of HTR4, HTR6 has been previously suggested to have strong expression in inhibitory interneurons [48,49]. Here, we find that there is a significant developmental regulation (F(6,50) = 4.9, p = 0.001), With the greatest HTR6 mRNA expression in the toddler age group and a significant down-regulation after childhood (Figure 3b).

Relationship with inhibitory interneuron marker mRNA
Since HTR2C and HTR6 are thought to be more strongly expressed in cortical interneurons, we examined whether their expression was related to developmental changes in common interneuron markers, such as parvalbumin (PV), calbindin (CB), somatostatin (SST) and cholecystokinin (CCK). As hypothesized, developmental changes in HTR2C and HTR6 mRNA were significantly associated with individual interneuronal markers, but these two HTRs appear related to distinct interneuron markers. Developmental changes in HTR6 mRNA were significantly matched with developmental changes in CB mRNA

Demographic effects other than age
None of the mRNAs encoding serotonin receptors varied significantly with gender. Caucasians did have a significantly increased expression of HTR2A mRNA compared to African Americans (t = 22.7,df = 57, p = 0.01). There was no universal demographic correlate in our study with these six HTRs.

Discussion
We found large developmental changes in the expression of several individual serotonin receptors in human prefrontal cortex. While the mRNAs for two Ga i -coupled inhibitory HTRs have a relatively constant expression level across the life span, the mRNAs encoding two excitatory Ga q -coupled and two Ga s -coupled HTRs increase to higher levels during childhood and teenage years before declining to adult levels. However, some of these excitatory HTRs may be expressed predominantly in inhibitory interneurons, and thus may act to inhibit the prefrontal cortex. In support of this hypothesis, we found significant relationships between the developmental expression of certain HTRs and developmental expression of specific interneuron markers. Overall, a collective look at six prefrontal HTRs in terms of their neurophysiological effects and cellular localization suggests the possibility that there Table 3. Major and minor cortical cellular distribution of the analyzed serotonin receptors.

Model for the developmental balance of the expression of excitatory and inhibitory HTRs
We found strong developmental regulation of the individual HTR mRNAs, in particular the Ga q -coupled and Ga s -coupled HTRs. By considering the possible excitatory or inhibitory effect of our measured HTRs, as well as their proposed cellular locations, we produced the model illustrated in Figure 4. We calculated the proposed interneuron contributions of the HTR receptors (HTR2A + HTR2C + HTR6 -HTR1A) as well as the pyramidal cell contributions (HTR2A + HTR4 -HTR1A -HTR5A). By taking into account that excitation of inhibitory cells will result in an inhibition of pyramidal cell firing, we were able to propose a net effect (dotted line in Figure 4) on the prefrontal cortex. This overall effect was seen to be balanced throughout the course of prefrontal cortical development. Overall, serotonin signalling may have an increased saliency in the first 5-10 years of life since the expression of most HTR mRNAs increase during this developmental period. In order to reduce the complexity of the system, the model makes several assumptions. It only includes 6 of the 14 known HTR receptors and does not take into account variations in receptor densities and possible interactions of second messenger systems. However, the model provides a useful framework for evaluating the possible net effects of serotonin over the course of prefrontal cortical development.
Developmental expression of Ga i -coupled HTRs in human prefrontal cortex HTR1A: Perspective and clinical relevance. Our finding of relatively constant HTR1A expression across human development is consistent with the results of a previous developmental investigation in human postmortem tissue [57]. HTR1A has been a focus of myriad human postmortem and imaging studies in adults. Work examining HTR1A binding in prefrontal cortex suggests that such Ga i -coupled receptors are protective against anxiety in controls [32] and are reduced in people with social anxiety [31]. Interestingly, there appears to be an increase in HTR1A expression and binding in schizophrenia [26,58,59,60].
The importance of maintaining constant expression of HTR1A in prefrontal cortex during development should not be underestimated, since conditional loss of the forebrain Htr1A receptor during development leads to life-long abnormal anxiety behaviour [61,62]. Of note, HTR1A appears to be regulated by early life experience in rodents and primates [41,63]. Human HTR1A is highly expressed in superficial layers of prefrontal cortex [64]. Analysis of Htr1A expression and function in superficial prefrontal cortex of rodent suggests that healthy animals have relatively stable levels of this key inhibitory receptor [41]. However, a careful examination of the in situ data [41], in the context of previous neurophysiological examination [22,23], suggests that Htr1A may have varying developmental mRNA expression patterns in the superficial and deep cortical layers. An apparent lack of change in HTR1A mRNA expression in the development of the prefrontal cortex by qPCR analysis may obscure possible opposite developmental changes in the superficial and deeper cortical layers [22,41]. Despite a link to PV-containing interneurons [30,55,56,65], HTR1A mRNA levels did not display a relationship through regression with any mRNA marker of inhibitory neurons.
HTR5A: Perspective and clinical relevance. The other Ga i -coupled receptor we examined, HTR5A, is less fully understood. We found its expression levels remain relatively constant throughout development but it shows an unexpected and fairly large increase in expression in adulthood. Stimulating HTR5A may be underappreciated as a clinical target to protect the brain under stressful conditions. A genetic deletion study in mouse suggests that knocking out Htr5A produces abnormal behavioural effects on stressful tests and inappropriate responses to novel situations [45]. Interestingly, a close family member (Htr5B) has recently been shown to be upregulated by the developmental stress of social isolation after weaning [66].
Developmental expression of Ga q -coupled HTRs in human prefrontal cortex HTR2A: Perspective and clinical relevance. The HTR2A receptor is the second most widely expressed serotonin receptor in the prefrontal cortex after HTR1A and is also predominantly expressed in pyramidal neurons [25,26,27,28,67]. In rodents, the excitatory effects of cortical Htr2A are thought to promote anxiety [33] and impulsivity [40]. In adult humans, HTR2A binding is elevated in prefrontal cortex of subjects with mood disorders [35] and also has been correlated with personality risk factors for affective disorders, such as neuroticism [68,69]. HTR2A antagonism contributes to the efficacy of medicines used to treat schizophrenia to the extent that it may be a defining feature of the atypical antipsychotic medicines [70]. However, it is controversial whether there are changes in HTR2A expression or binding in schizophrenia [26,59,71,72,73]. Yet, medicines that antagonize HTR2A improve the symptoms of anxiety and depressive disorders in patients who did not improve on other treatments [74,75,76,77]. In rodents, genetic deletion of cortical Htr2A diminished anxiety levels [33], and pharmacological modulation of Htr2A within the prefrontal cortex has bidirectional effects on impulsivity with stimulation increasing and antagonism decreasing impulsivity [40]. Rodent work has also Figure 4. Working model to illustrate the balanced developmental expression of HTRs with excitatory effects on prefrontal cortex and those with inhibitory effects. The complex cellular localization of the HTRs and their different excitatory or inhibitory effects makes it difficult to conceptualize how the developmental changes in their expression would affect the prefrontal cortex. Several assumptions were required in order to formulate a working model that attempted to synthesize the potential consequences of these developmental changes. The mRNA expression of each serotonin receptor was normalized to its individual peak expression which was given the value of one. Receptors on which serotonin has an excitatory effect were assigned a positive contribution value while those receptors on which serotonin has an inhibitory effect were assigned a negative value. The net HTR contribution to interneuron activity predicted by this model is plotted in black and was calculated as follows: (HTR2A + HTR2C + HTR6 -HTR1A). The net HTR contribution to pyramidal neuron activity predicted by this model is plotted in gray and was calculated as follows: (HTR2A + HTR4 -HTR1A -HTR5A). The overall effect of serotonin receptors on the output from the prefrontal cortex was based on the predominant localization of each receptor. This net prefrontal effect is illustrated by the dashed line which takes into account that excitation on inhibitory interneurons would likely result in inhibition of the firing of the pyramidal cells responsible for output. Note this working model suggests a lack of a large developmental change in the predicted overall net effect of HTRs on the prefrontal cortex, despite an overall trend towards increased expression of the excitatory serotonin receptors within the first decade of life. However, the 5HTR subtypes responsible for maintaining this balance appear to change. doi:10.1371/journal.pone.0022799.g004 shown strong neurophysiological effects of this receptor in deep layers of cortex in the first two weeks of postnatal development [22,23] and reduced effects in adulthood [78], likely due to the competing inhibitory effects of Htr1A which is highly co-expressed in cortical pyramidal neurons [25,28]. Recent work suggests that early stress exposure may increase the functional effects of Htr2A in adulthood with negligible changes in receptor expression, binding, or sensitivity [78], suggesting potential elevation of downstream signalling.
HTR2A mRNA also shows a remarkably strong positive relationship with the interneuron markers PV and CB, both of which show a protracted increase during the first decade of human life [7]. Of note, the up-regulation of HTR2A, PV, and CB mRNA mirrors the increase in gamma oscillation power which is found during development [79]. PV and CB are found primarily in inhibitory neurons which are electrophysiologically classified as fast-spiking and which synapse onto the cell bodies of excitatory neurons [80]. Htr2A has been suggested to modulate cortical gamma oscillations [9,30] at least in part through its effects on fast spiking interneurons [30,56].
HTR2C: Perspective and clinical relevance. HTR2C is thought to have effects opposite to those of HTR2A within prefrontal cortex. This difference may be due to the fact that HTR2C is expressed in inhibitory interneurons [36,37,38] and not in pyramidal neurons [46], while HTR2A is predominantly expressed in pyramidal neurons with a minority of its expression in interneurons [28]. Here, we show that the developmental changes in HTR2C are linked to developmental changes in the interneuronal marker, CCK. This relationship suggests that as CCK interneurons mature they may increase HTR2C expression and become more responsive to serotonin. Work in rodents suggests that greater prefrontal HTR2C tone would act to decrease impulsivity, since impulsivity is increased either by genetic deletion of Htr2C [39] or by its pharmacological antagonism within prefrontal cortex [40].
Developmental expression of Ga s -coupled HTRs in human prefrontal cortex HTR4: Perspective and clinical relevance. The HTR4 receptor is found in prefrontal cortical pyramidal neurons [24,46,47] and appears to have excitatory effects within prefrontal cortex [22,24]. There is evidence that HTR4 agonists improve cognitive functioning in rodent and primate models [81,82]. Our findings are broadly consistent with rodent studies suggesting that there is gradual up-regulation of this receptor during development [83]. The surface expression and function of HTR4 are thought to be regulated by the S100 family member p11 [84], which has been implicated in mood disorders and suicide [85]. The HTR4 receptor gene also has a polymorphism which has been linked to schizophrenia in a Japanese population [86].
HTR6: Perspective and clinical relevance. Htr6 appears to be more strongly expressed in inhibitory interneurons [48,49] than in pyramidal neurons [46]. We found that developmental changes in HTR6 expression follow an inverted-U shaped pattern, with an initial increase followed by a decrease. This expression profile mirrors that for CB expression in our cohort, which corresponds to the proposed interneuron localization of HTR6. Stimulation of this receptor has been suggested to have anxiolytic [49,87] and anti-depressant effects [88]. Yet HTR6 antagonism is an unusual feature of the atypical antipsychotic clozapine [89]. Indeed, pharmacologically antagonizing HTR6 acts to enhance cognition [51]. In our cohort, the expression of HTR6 peaks at the toddler age group before decreasing into adulthood, thus drugs acting on this receptor may produce exaggerated effects if administered in childhood.
A balance seemingly tipped towards inhibition and potential mechanisms of dysregulation Quantification of the normalized expression of two serotonin receptors likely to have net excitatory effects on the prefrontal cortex (HTR2A and HTR4) based on the available evidence, and four likely to have net inhibitory effects (HTR1A, HTR2C, HTR5A, HTR6), suggests that the effect of serotonin through these receptors may remain balanced in the inhibitory range across development and in adulthood. This prediction of a net inhibitory cortical response to serotonin is supported by in vivo studies of the cortical effects of raphe stimulation in adult rodents [11,25,29]. The overall effect of serotonin on the cortex is complicated by the localization of a number of these receptors on inhibitory interneurons (HTR2C, HTR6, and to a lesser extent HTR1A, HTR2A). This is shown by the widespread relationship between developmental changes in GA-BAergic interneuron marker mRNA expression and those of HTRs expression. In particular, there is a statistical relationship demonstrated through linear regression for HTR2A and HTR6 mRNA with the interneuron marker CB, in addition to a relationship between HTR2C, and the interneuron marker CCK.
The hypothesized balance between excitation and inhibition for this group of receptors may be disrupted during development when the expression or function of individual HTRs are changed by polymorphisms, early life stress, or developmental exposure to serotonergic medicines. For example, the developmental trajectory of HTR2A suggests that the period of increased HTR2A mRNA expression from childhood through the teenage years may be a time when the brain would be particularly vulnerable to increased developmental Htr2A expression resulting from early maternal deprivation [78,90]. If unchecked by the highly coexpressed inhibitory Htr1A [25,28], or by the excitatory Htr2C expressed in inhibitory neurons [35,36,37,38,39], such changes would be hypothesized to lead to greater impulsivity [39,40] particularly under stressful conditions which would increase serotonin release [1,11,91]. Changes to the expression or function of HTRs have been shown to occur as a result of genetic polymorphisms [92,93,94], early life experience [41], and developmental pharmaceutical exposure [15,95]. Disrupting a normal balance in the effects of serotonin on the prefrontal cortex during a vulnerable developmental period may result in abnormal emotional regulation, cognition, and behaviour in the short term. Furthermore, the trophic effects of serotonin on maturing neurons [18,19,20,21], suggests that dysregulation of an individual HTR may have lasting repercussions for adult prefrontal cortical structure and function.

Supporting Information
Table S1 Demographic details of cases used in this study. Abbreviations: m, male; f, female; PMI, post-mortem interval defined as interval between death and freezing of the brain; AA, African American; C, Caucasian; RIN, RNA integrity number; SIDS, sudden infant death syndrome; MVA, motor vehicle accident; SVCS, Superior Vena Cava Stenosis; ASCVD, Arteriosclerotic Cardiovascular Disease; HASCVD, Hypertensive Arteriosclerotic Cardiovascular Disease. (DOCX)