Genetic Models of Serotonin (5-HT) Depletion: What do They Tell Us About the Developmental Role of 5-HT?

Authors


Abstract

A large number of hyposerotonergic genetic models have been generated over the past few years. Serotonin (5-HT) depletion has been obtained via targeting of genes involved in 5-HT synthesis (Tph1 and Tph2), specification and determination of the 5-HT phenotype during development (GATA3, Pet1, and Lmx1b), and 5-HT storage or clearance (Vmat2 and SERT). Here we review these various models from a developmental perspective, beginning with a description of the sources of 5-HT during development. We then summarize the neurological and behavioral alterations that have been observed in the genetic hyposerotonergic models. Although these models appear to have normal brain development and do not exhibit any gross morphological defects, problems in somatic growth and physiological functions have been observed. Abnormal adult behavior is also seen, although whether it results from depletion of 5-HT during development or functional 5-HT deficiencies in adult life remains unclear. Evidence from these hyposerotonergic models suggests that the developing brain may not need 5-HT for the establishment of general organization and structure. However, central 5-HT appears to be necessary for postnatal body growth, maturation of respiratory and vegetative control, and possibly for the development of normal adult behavior. Anat Rec, 2011. © 2011 Wiley-Liss, Inc.

INTRODUCTION

Although serotonin (5-HT) is best known as a mood-stabilizing neurotransmitter in the adult brain, it has a variety of important functions during development. Indeed, 5-HT modulates a number of developmental processes, including cell proliferation, migration, and differentiation, in peripheral tissues and in the brain (Azmitia,2001; Buznikov et al.,2001; Gaspar et al.,2003). Because of the importance of 5-HT during development, a link between 5-HT dysfunction and neurological developmental disorders, such as autism, has been suggested (Whitaker-Azmitia,2001). Furthermore, it is possible that variations in 5-HT levels have subtle effects on the wiring of the brain during development, leading to an increased risk for various mental disorders, including depression and anxiety in adulthood (Gaspar et al.,2003; Murphy and Lesch,2008).

Genetic models in the mouse can provide valuable insight into the developmental role of 5-HT. Early genetic models demonstrated that increasing 5-HT transmission during development by reducing clearance of 5-HT has long-term consequences on brain organization. These models had either a loss of function of the monoamine oxidase A (MAOA) gene, which encodes the enzyme that degrades 5-HT, or a loss of function of the 5-HT transporter (SERT) gene, which is required to clear 5-HT from the extracellular space. These models provided clear evidence that 5-HT modulates activity-dependent refinement of neural circuits. MAOA KO and SERT-KO mice had altered development of the barrel cortex (Cases et al.,1996; Persico et al.,2001; Salichon et al.,2001) and refinement of the retinal projections in central targets (Upton et al.,1999,2002). Moreover, these models demonstrated that abnormal adult behavior could result from changes in serotonin levels during development. More specifically, the MAOA −/− mice showed increased aggressivity (Cases et al.,1995; Mejia et al., 2002) and the SERT −/− mice had a pro-depressive phenotype (Lira et al.,2003, Alexandre et al.,2006) and increased anxiety (Ansorge et al.,2004) that could be linked to the increased brain levels of serotonin during critical developmental periods. Genetic models in which 5-HT receptors were selectively invalidated further demonstrated the developmental importance of 5-HT. Notably, loss of function of the 5-HT1A receptor altered developmental programming of the hippocampal neural circuits, which resulted in increased anxiety levels (Gross et al.,2002). Additionally, loss of function of the 5-HT2B receptor caused early developmental abnormalities in cardiovascular development (Nebigil et al.,2000).

Although these genetic models suggest that 5-HT dysfunction affects neural circuit development and subsequent adult behavior, the evidence is much less clear in genetic models with decreased 5-HT levels in the brain. Although there is ample evidence for effects of 5-HT depletion on brain development, this research primarily relies on pharmacological approaches (reviewed in this issue by Mary Blue). In contrast, in most of the recently described genetic models of central 5-HT depletion, generated over the last 6 years, the brain developed normally. How should these conflicting results be interpreted?

In this article, we will review the existing literature on genetic hyposerotonergic models. We will begin with a description of the various sources of 5-HT during development. Then we will present the main strategies that have been used to obtain hyposerotonergic mice and summarize the neurological and behavioral alterations that have been observed in these models. We will end by providing some speculative remarks on the diversity of phenotypes induced by altered serotonin transmission at different periods in life.

5-HT Sources During Development

Brain development can be broadly divided into two phases: during the first phase, the general structure and organization of the brain is established, and, during the second phase, refinement of neural circuits occurs in an activity-dependent manner (Fig. 1). Genetically programming 5-HT depletion during development requires identification of the sources of 5-HT at these different developmental time-points. In the adult brain, all 5-HT is produced by the raphe neurons in the brainstem. In contrast, there are multiple sources of 5-HT in the developing brain. The first and most important source is the mother: maternal 5-HT is actively transported through the placental brush border cells via SERT (Cool et al.,1990; Carrasco et al.,2000). 5-HT immunostaining is present in the placenta and the ectoplacental cone (Yavarone et al.,1993). Raphe neuron differentiation and production of 5-HT starts by E10.5 in mice (Hendricks et al.,1999) and by the fifth gestational week in humans (Sundstrom et al.,1993; Verney et al.,2002; Paterson and Darnall,2009). However, in cases where the embryo has deficient 5-HT production, maternal 5-HT can likely compensate during the embryonic phase of development. In addition to the raphe, several other tissues produce large amounts of 5-HT by E14, namely the pineal gland, where 5-HT is a precursor of melatonin, and the enterochromaffin and myenteric cells in the gut (Cote et al.,2007). Other tissues, such as the pancreatic beta cells (Paulmann et al.,2009), the parafollicular cells of the thyroid, the ovarian cumulus cells (Dube and Amireault,2007), the dorsal root ganglia, and the taste buds (Huang et al.,2005) produce smaller amounts of the amine. Collectively, these tissue sources outside of the brain produce ∼95% of 5-HT. During development, 5-HT produced in these different body parts could reach the brain through the immature blood brain barrier, and act to compensate deficiencies in 5-HT production by raphe neurons.

Figure 1.

Sources of serotonin during embryonic and postnatal development. Maternal 5-HT is actively transported through the placenta via SERT. Raphe neuron differentiation and production of 5-HT starts by E10.5 in mice. The pineal gland and the enterochromaffin and myenteric cells in the gut produce large amounts of 5-HT by E14. Other tissues, such as the pancreatic beta cells, the parafollicular cells of the thyroid, the ovarian cumulus cells, the dorsal root ganglia, and the taste buds produce smaller amounts of the amine. Synthesis of 5-HT in the brain involves Tph2, whereas synthesis in other parts of the body, except for the myenteric cells, involves Tph1. In addition to these sites of 5-HT production, there are a number of cells that act as 5-HT reservoirs during development. Uptake of 5-HT during early development is first observed in non-neuronal tissues such as the heart, the liver, the cranial mesenchyme, and the notochord. Later, 5-HT accumulation is detected in several categories of neurons, both in the brain and in neural crest derivatives. All these cells and neurons express SERT and the vesicular monoamine transporter (Vmat1 or Vmat2).

In addition to these sites of 5-HT production, there are a number of cells that act as 5-HT reservoirs during development. These are sites where 5-HT is captured and stored. This mechanism allows the tissues to buffer 5-HT levels by modulating 5-HT uptake and release. Uptake of 5-HT during early development is first observed in non-neuronal tissues such as the heart, the liver, the cranial mesenchyme, and the notochord (Lauder and Zimmerman,1988; Shuey et al.,1992; Pavone et al.,2008). Later, 5-HT accumulation is detected in several categories of neurons, both in the brain and in neural crest derivatives (Hansson et al.,1999; Narboux-Neme et al.,2008). All these cells and neurons express SERT and the vesicular monoamine transporter (Vmat1 or Vmat2), which permits them to take up 5-HT and then concentrate it in intracellular organelles, where it is protected from degradation.

Because the blood brain barrier is not fully functional before postnatal day 12 (Ribatti et al.,2006), 5-HT from the periphery can easily enter the brain and vice versa during the embryonic and early postnatal periods. Thus, when targeting peripheral or central 5-HT production, it is important to be aware of the alternative sources of 5-HT which could compensate for deficiencies in 5-HT levels in other areas.

Targeting 5-HT Synthesis: Tryptophan Hydroxylase Knockout Mice

The most straightforward way to obtain 5-HT depletion is to interfere with 5-HT synthesis by targeting the rate-limiting synthetic enzyme. Tryptophan hydroxylase (Tph), the key enzyme in 5-HT metabolism, converts tryptophan to 5-hydroxytryptophan. There are two isoforms of Tph, Tph1 and Tph2, which are encoded by different genes on different chromosomes. Tph2 is more abundant in the brain, although it is also expressed in myenteric cells in the gut (Cote et al.,2007) (Fig. 1). Tph1 is expressed in the pineal gland, the enterochromaffin cells (Cote et al.,2003), the beta pancreatic cells, and the other peripheral tissues which produce 5-HT (Fig. 1). Expression of Tph1 has also been reported in the raphe neurons during a late developmental phase, peaking at P21 (Nakamura et al.,2006), although this was not confirmed in a more recent study (Gutknecht et al.,2009). Studies of the kinetics of both enzymes have shown that Tph1 has a higher affinity for tryptophan and a higher enzymatic activity than Tph2 (McKinney et al.,2004; Nakamura et al.,2006).

The Tph1 gene was cloned first and was invalidated via gene recombination approaches by two different teams. The Tph1 −/− mice showed a 90% reduction of circulating 5-HT in the blood and in the gut, but had normal levels of 5-HT in the brain (Cote et al.,2003; Walther and Bader,2003). The Tph1 −/− mice appeared to develop normally when they were born from heterozygote mothers. However, Tph1 −/− mothers gave birth to embryos which were malformed and had stunted growth at E10 and E12, irrespective of the embryonic genotype (Cote et al.,2007; Fligny et al.,2008). In embryos collected from Tph1 −/− mothers, brain malformations were noted in the hindbrain and in the pallium, with reduced cell proliferation in the ventricular zone. Thus, maternal 5-HT is required for early embryonic development, before the development of the raphe.

The Tph1 −/− mice have been instrumental to the characterization of the numerous non-neuronal roles of 5-HT. These include the role of 5-HT in cardiovascular function (Cote et al.,2007; Fligny et al.,2008), platelet function (Walther et al.,2003), and in the secretion of insulin by beta cells in the pancreas (Paulmann et al.,2009). This last observation led to the unexpected discovery that 5-HT can act via serotonylation of GTPases to control exocytosis. In addition, Tph1 was found to play a role in liver regeneration (Lesurtel et al.,2006), in mammary gland development and involution (Matsuda et al.,2004), and in the regulation of bone formation (Yadav et al.,2009).

Because the Tph2 gene was identified later as a separate gene from Tph1, Tph2 −/− mouse models have been generated only over the last 2 years (Gutknecht et al.,2008; Savelieva et al.,2008; Alenina et al.,2009; Migliarini et al.,2009). In the Tph2 −/− mice, levels of 5-HT in the brain are reduced by 90%, whereas peripheral levels remain unchanged (Savelieva et al.,2008; Alenina et al.,2009). Despite the severe depletion of brain 5-HT, the mice are viable and are born in normal Mendelian ratios, although two groups noticed increased lethality during the first 4 postnatal weeks (Alenina et al.,2009; Migliarini et al.,2009). Tph2 −/− mice had severe growth defects, as compared with their wildtype littermates (Alenina et al.,2009; Migliarini et al.,2009). Growth defects were sustained during the first few postnatal weeks, with a recovery by 4 months of age. Additionally, Tph2 −/− mice had impaired thermoregulation, altered sleep patterns, and decreased heart-rate and blood pressure (Alenina et al.,2009). Somewhat surprisingly, no gross morphological defects were found in the brain, although one group reported abnormal organization of the serotonergic fibers (Migliarini et al.,2009).

Tph2 −/− mice showed some subtle behavioral abnormalities as adults. One group found that male Tph2 −/− mice buried significantly more marbles in the marble-burying test, a test for anxiety (Savelieva et al.,2008). In two tests for depression, the forced swim test and the tail suspension test, Tph2 −/− mice showed opposing nonsignificant trends, with the forced swim test suggesting a slight anti-depressive effect and the tail suspension test suggesting a slight pro-depressive effect (Savelieva et al.,2008). Interestingly, Tph2 −/− mothers showed altered nurturing behavior, neglecting and often cannibalizing pups (Alenina et al.,2009; Migliarini et al.,2009).

A rare hypomorphic variant of the Tph2 gene has been identified in some human patients with major depression. A knock-in of this variant (R439H) in mice led to an 80% depletion of 5-HT in the brain (Beaulieu et al.,2008). These mice showed pro-depressive behavior and higher anxiety levels, as demonstrated by increased immobility in the tail suspension test and increased latency to cross toward the light in dark-light emergence tests. Male mice were also more aggressive and attacked nonaggressive males more frequently.

A double knock-out of both Tph1 and Tph2 has also been obtained (Savelieva et al.,2008). In these mice, brain 5-HT levels were similar to those found in Tph2 −/− mice, and peripheral 5-HT levels were also reduced. The mice were generally smaller, but this was only significant in male mice. The behavioral phenotype was similar to that found in Tph2 −/− mice, with more marble burying, reduced immobility in the forced swim test, and increased immobility in the tail suspension test. However, the authors noted that a decrease in body fat in double knockout mice could have influenced the results in the forced swimming test.

Targeting the Development of 5-HT Neurons in the Raphe

Another genetic approach for 5-HT depletion is targeting the transcription factors that regulate the region-specific generation and differentiation of 5-HT neurons. In recent years, a number of factors have been identified as playing a role in 5-HT cell specification. These factors can act sequentially or in parallel. A first set of transcription factors, Nkx2.2, Nkx2.9, and Ascl1, are required to generate 5-HT precursors in the raphe between embryonic days 8 and 10, in mice. A second set of transcription factors, GATA2, GATA3, Insm1, Fox A2, Lmx1b, and Pet1, are required for 5-HT subtype selection and 5-HT neuron terminal differentiation between E10 and E12 (Cordes,2005; Jacob et al.,2007; Jacob et al.,2009). Other transcription factors, such as Phox2B and Hoxb1, are involved in repressing the 5-HT phenotype of the precursors in specific brainstem regions (Pattyn et al.,2004). A number of these transcription factors, particularly those implicated in the generation of the 5-HT precursors or early differentiation of the neurons, have multiple targets and pleiotropic effects. Thus, knockout of these genes frequently leads to early lethality, and conditional deletions or chimeras are required to analyze the postnatal phenotypes. We will discuss here only those genetic models that gave rise to viable offspring.

GATA2 and GATA3 are zinc-finger transcription factors, a family of genes known to play important roles in cell-type specific gene expression. The GATA2 and GATA3 genes are expressed in the midbrain at the level of the raphe. These genes have complementary roles in the differentiation of the raphe cell identity. GATA2 governs the differentiation of the 5-HT and GABA raphe neurons in the rostral raphe (Craven et al.,2004; Kala et al.,2009), whereas GATA3 is involved in the differentiation of the caudal raphe 5-HT neurons (van Doorninck et al.,1999). GATA3 −/− mice die by E10 and have severely abnormal development (Pandolfi et al.,1995). Thus, the role of GATA3 in serotonergic neurons was analyzed in chimeric mice, in which only a fraction of the cells are GATA3 −/− (van Doorninck et al.,1999). Analysis of these chimeras showed that the mutant cells in the caudal raphe lost the 5-HT phenotype and had disturbed cytoarchitectural organization, whereas neurons in the rostral raphe developed normally. Impaired locomotor performance on a rotating rod was only seen in chimeric mice with affected caudal raphe neurons.

Pet1 is an ETS transcription factor, and its human homologue, fifth Ewing variant (FEV), is expressed exclusively in raphe 5-HT neurons (Hendricks et al.,1999; Maurer et al.,2004; Lillesaar et al.,2009) and in 5-HT intestinal cells (Wang et al.,2010). Pet1 is a terminal differentiation factor that acts in the final step of the transcriptional cascade to establish the final identity of 5-HT neurons. In Pet1 −/− mice, the raphe neurons and 5-HT precursors are generated in normal numbers. However, 70%–80% of these neurons do not reach the mature serotonergic phenotype and remain in the raphe in an arrested state of differentiation (Hendricks et al.,2003). No clear developmental defects were noted in the cerebral cortex of the Pet1 −/− mice (Stankovski et al.,2007), and the barrel cortex developed normally (our unpublished observations). However, recent reports suggest that 5-HT neurons have an altered migration pattern (Krueger and Deneris,2008). Furthermore, several behavioral alterations have been noted in Pet1 −/− mice, such as severely impaired maternal behavior (Lerch-Haner et al.,2008), increased aggression, and increased anxiety in the elevated plus maze (Hendricks et al.,2003). Altered respiratory control was also noted during early postnatal life, as indicated by irregularities in breathing and increased respiratory pauses (Erickson et al.,2007). Interestingly, genetic variants of FEV in humans have been associated with sudden infant death syndrome, suggesting that there is a link between the development of the serotonin neurons in the brainstem and early postnatal respiratory control (Rand et al.,2007; Broadbelt et al.,2009).

Lmx1b is a LIM homeodomain-containing gene that has multiple functions in the development of many tissues, including the kidney, limbs, and central nervous system. Lmx1b acts upstream of Pet1 and also coordinates with Pet1 to maintain the serotonergic phenotype. When the Lmx1b gene is invalidated, the 5-HT precursors in the raphe are generated but do not acquire a serotonergic phenotype (Ding et al.,2003). Because homozygous Lmx1b −/− mice have an embryonic lethal phenotype (Chen et al.,1998), a conditional knock-out of Lmx1b has been obtained by crossing Lmx1B floxed mice with Pet1-Cre mice (Zhao et al.,2006). These mice initially have normal generation of 5-HT neurons in the raphe (E11), but by E12-13, 5-HT neuron-specific markers, such as SERT and Tph2, are downregulated. Thus, in adults, only a few 5-HT positive neurons are found in the raphe, and no 5-HT fibers are detected. Whether the Lmx1b −/− raphe neurons transdifferentiate into another phenotype remains unclear. No structural abnormalities of the brain or general body were reported in these mice, but several functional disorders have been noted during the postnatal period. Hodges et al. (2009) observed that the conditional Lmx1b −/− mice had irregular respiration with frequent apneic episodes between P2 and P14; this phenotype normalized between P14 and P28. These transient respiratory problems are probably responsible for the increased mortality seen during the early postnatal period and were suggested to cause the transitory growth defects. 5-HT2A receptor agonists were noted to partially reverse the respiratory alterations, suggesting that these are essentially functional rather than developmentally hard-wired abnormalities.

The conditional Lmx1b −/− mice also exhibit some behavioral abnormalities as adults (Dai et al.,2008). They have impaired retrieval of spatial memory, as tested in the Morris water maze. They also have reduced levels of anxiety in the elevated plus maze and the novelty-suppressed feeding test. Finally, these mice have faster acquisition and enhancement of fear memory, a phenotype which is partially reversed after administration of a 5-HT1A receptor agonist. Again, the reversal seen after treatment with 5-HT receptor agonists suggests that this particular behavioral phenotype may not be a developmental effect.

Targeting the Serotonin Clearance or Release Mechanisms

Increasing 5-HT catabolism or clearance are two additional strategies, which have been used to deplete brain 5-HT. Preventing 5-HT storage in release vesicles increases 5-HT degradation, whereas overexpressing SERT increases 5-HT clearance.

Vmat2 is required for the storage of monoamines in synaptic vesicles in both the central and peripheral nervous system. When 5-HT is not stored into vesicles, it is catabolized to 5-hydroxyindole acetic acid (5-HIAA) by mitochondrial enzymes, the monoamine oxidases. Vmat2 inhibition by reserpine or knockout of the Vmat2 gene results in a profound depletion of 5-HT and the other monoamines that rely upon Vmat2 for vesicular storage (dopamine, noradrenaline, and histamine). Vmat2 −/− mice are not viable beyond the first postnatal week and show major growth retardation due to feeding abnormalities (Fon et al.,1997; Wang et al.,1997). No gross abnormalities in brain organization or in the layout of the barrel cortex in the primary somatosensory cortex were found (Persico et al.,2001; Alvarez et al.,2002). However, more careful histological studies showed a twofold increase in developmental cell death in the superficial layers of the cingulate cortex, which was correlated with a reduced expression of the anti-apoptotic factor Bclxl (Stankovski et al.,2007). Furthermore, maturation of the upper cortical layers (II–IV) was altered in the Vmat2 −/− mice. Specifically, barrels did not form in layer IV, despite the normal organization and layout of the thalamocortical axons (Alvarez et al.,2002). Because Vmat2 −/− mice were severely atrophic and did not survive beyond P5, it was not possible to determine whether these changes reflected a developmental delay or a permanent change in the cortical architecture.

We recently obtained a conditional knockout of the Vmat2 gene under the control of SERT, which has allowed us to examine the consequences of blocking vesicular release specifically in serotonergic neurons (Narboux-Neme et al.,2009). In this genetic model, Vmat2 was ablated in the raphe neurons, as well as in the neurons and cell types that transiently accumulate serotonin during development, such as the thalamocortical axons and the neurons of the limbic cortices (Narboux-Neme et al.,2008). This model showed a protracted delay in the maturation of the upper cortical layers similar to the phenotype observed in the Vmat2 −/− mice. However, neurons in layer IV did form normal barrels by P7, indicating that the defect in layer IV neuron organization previously seen in the full Vmat2 −/− mice reflected a developmental delay, rather than a permanent change. A major retardation of general somatic growth was also noted (Fig. 2); this was most severe during the second postnatal week and tended to normalize after weaning (Narboux-Neme et al.,2009). Interestingly, the somatic growth retardation and the delayed cortical maturation appeared to be dissociated in time, suggesting that they result from perturbations in independent pathways. Somatic growth abnormalities could be corrected by the administration of clorgyline, an inhibitor of monoamine oxidase. Thus, the deficiency in somatic growth results from defective serotonin release and can be reversed by nonvesicular release of 5-HT.

Figure 2.

Growth defects in Vmat2 conditional −/− mice. Conditional Vmat2 −/− mouse (red arrow) and its control littermate at P16. Knockout mice display severe growth retardation that normalize after weaning.

We have only just begun to explore the behavioral phenotype of the conditional Vmat2 −/− mice. Thus far, we have found that the mice have reduced spontaneous locomotor abilities, antidepressive responses in the tail suspension test, and reduced anxiety levels in the hyponeophagia test (Mongeau and Diaz, unpublished results).

In humans, a common variant of SERT has increased expression, which increases 5-HT uptake and is associated with decreased anxiety (Canli and Lesch,2007). To examine this genotype in mice, Jennings et al. (2006) produced a mouse strain in which SERT was over-expressed in the raphe, producing a hyposerotonergic phenotype (50% reduction). These mice exhibited decreased anxiety-related behavior, spending more time in the open arm of the elevated plus maze and showing a shortened latency to first enter the open arm. Animals also showed reduced levels of anxiety in the hyponeophagia test. Inhibition of SERT reversed this phenotype, suggesting that the decrease in anxiety was not a developmental effect. Furthermore, it is unclear whether sites that transiently express serotonin during development are affected in this model.

Consistencies and Inconsistencies Between Different Hyposerotonergic Models

The main phenotypes that have been observed in the different hyposerotonergic models are listed in Table 1. The models have consistently demonstrated that central 5-HT depletion results in impaired postnatal growth and altered thermoregulation and breathing, which could explain the increased lethality during early postnatal life. Interestingly, three of the mouse models (Tph2, Pet1, and conditional Vmat2) have impaired maternal behavior. Also, three mouse models (Tph2, Pet1, and Tph2-R439H) show increased male aggression. However, other behavioral effects have been reported less consistently. Different tests for depression showed both pro-depressive and anti-depressive trends in the Tph2 knockout. Pro-depressive effects were seen in Tph2-R439H knock-in mice, whereas anti-depressive effects have been seen in Vmat2 conditional knockout mice. In tests for anxiety, certain models (Pet1, Tph2-R439H) have increased anxiety levels, whereas other models (Lmx1b, SERT, and Vmat2) have reduced anxiety levels. These paradoxical results may illustrate the complexity of serotonergic systems. The 15 different serotonin receptor sub-types and the varied and shifting sources of 5-HT during development create a complex system in which slight alterations in the method of 5-HT depletion or in the timing of the depletion result in surprisingly different phenotypes.

Table 1. Comparison of the major phenotypes reported in the hyposerotoninergic mouse models
Model5-HT levelsVegetative effectsSomatic growthMaternal behaviorAggressivityAnxietyDepressionReferences
TPH190% Reduction of peripheral 5-HT, normal brain levels NormalNormalNormalNormalNormalCote et al.,2003; Walther and Bader,2003
TPH290% Reduction of brain 5-HTImpaired thermoregulation, decreased heart rate/blood pressure, respiratory problemsRetardation of growthImpairedIncreasedNDUnclearGutknecht et al.,2008; Savelieva et al.,2008; Alenina et al.,2009; Migliarini et al.,2009
TPH2 (R439H)80% Depletion of brain 5-HTNormalNormalNDIncreasedIncreasedIncreasedBeaulieu et al.,2008
Pet185%–90% Decrease in brain 5-HTRespiratory problemsNDImpairedIncreasedIncreasedNDHendricks et al.,2003; Erickson et al.,2007; Lerch-Haner et al.,2008
Lmx1b99% Decrease in brain 5-HTImpaired thermoregulation, respiratory problemsRetardation of growthNDNDReducedNDZhao et al.,2006; Dai et al.,2008; Hodges et al.,2009
VMAT2 conditional95% Decrease in brain 5-HTRespiratory problemsRetardation of growthImpairedNDReducedReducedNarboux-Neme et al.,2009
Sert50% Reduction in brain 5-HTNormalNormalNDNDReducedNDJennings et al.,2006

The somewhat blurred picture of the functional consequences of reducing 5-HT transmission contrasts with the more consistent reports on the effects of increasing brain levels of 5-HT during development. However, a similar paradoxical pattern emerges when the effects of elevated 5-HT levels are studied in adult and developing brains. Inhibition of SERT function during the early postnatal period, by genetic (SERT-knockout mice) or pharmacological approaches (administration of antidepressants) caused an increase in anxiety related behaviors, REM sleep, and pro-depressive responses (rev. Homberg et al.,2010). In adults, however, pharmacological inhibition of SERT function has antidepressant and anxiolytic effects. Indeed, the most frequently prescribed drugs for depression and anxiety are selective serotonin reuptake inhibitors, which target SERT. The behavioral changes in young animals are the result of overactivation of 5-HT receptors during critical developmental periods. More specifically, increased 5-HT1A receptor signaling during the first postnatal month was implicated in the pro-depressive effects and the modification of sleep patterns of SERT knockout mice (Alexandre et al.,2006). In the SERT and MAOA knockout mouse models, the brain wiring defects in the somatosensory and visual systems are mediated by overactivation of a different 5-HT receptor, 5-HT1B, whose overactivation during the the first 2 weeks of life disrupts refinement of sensory maps (Salichon et al.,2001). Thus, when working with hyperserotonergic models and hyposerotnergic models, it is important to consider the many different 5-HT receptor subtypes, as well as the effects of age and development.

As we try to gain a more precise understanding of serotonergic function, the behavioral phenotype of hyposerotonergic mice needs to be better characterized. More studies are required to tease apart the effects of developmental depletion of 5-HT from the effects of functional 5-HT deficiencies during adult life. Finally, understanding the roles of the specific receptor subtypes in the development of different phenotypes will lead to better explanations of the often paradoxical results found in both hypo- and hyper-serotonergic models.

CONCLUDING REMARKS

The hyposerotonergic genetic models that have been generated over the last few years are finally allowing us to analyze the diversity of the physiological functions of 5-HT, complementing previous pharmacological and genetic studies. An important and unexpected outcome of these models was a clearer identification of the highly diversified sources of 5-HT in the body. This has led to the proposal that there are many different serotonin sub-systems, which serve specific functions in the organism, rather than a global serotonin system. These studies have, in particular, alerted researchers in neuroscience to the importance of 5-HT in a wide range of essential bodily functions, which could, in turn, impact brain function.

The finding that there are no major abnormalities in brain development in the hyposerotonergic models was especially surprising. As previously discussed, there are many different sources of serotonin during development, and these different sources could compensate when only one serotonin source is targeted. Even so, serotonin levels in the brain were found to be severely reduced in most of the models discussed, suggesting that the brain can indeed develop more or less normally in the absence of serotonin. These results, paired with studies of hyperserotonergic models, as discussed above, suggest that the developing brain may actually need to be protected from high levels of circulating serotonin when the brain blood barrier is still immature. The transient ability of certain neurons to take up and store serotonin during development could serve as a clearance pathway, acting to reduce extracellular levels of serotonin.

However, mice with depleted levels of serotonin in the brain have major defects in body growth, even when peripheral levels of serotonin are normal. Thus, it is possible that serotonin, as not required for overall brain development, is required for normal body growth, perhaps through regulation of the endocrine system via the hypothalamus (Yadav et al.,2009). Respiration and other physiological functions are also impaired in hyposerotonergic animals, suggesting that serotonin is required for proper functioning of respiratory motor neurons during early postnatal life.

Interestingly, despite apparently normal brain development, hyposerotonergic mice have behavioral abnormalities as adults, showing increased aggressivity, impaired maternal behavior, and changes in anxiety levels and depressive behavior. These abnormalities suggest that subtle miswiring in the brain may occur as a result of serotonin depletion. Moving forward, the behavioral phenotype of these hyposertonergic models will need to be clarified and the organization and connectivity of the brain should be examined on a more detailed level. These mice could provide important insights into the mechanisms underlying a number of human mental disorders, which manifest themselves in adulthood but are believed to originate from developmental abnormalities.

Ancillary