The use of thyroid hormones as an effective adjunct treatment for affective disorders has been studied over the past three decades and has been confirmed repeatedly. Interaction of the thyroid and monoamine neurotransmitter systems has been suggested as a potential underlying mechanism of action. While catecholamine and thyroid interrelationships have been reviewed in detail, the serotonin system has been relatively neglected. Thus, the goal of this article is to review the literature on the relationships between thyroid hormones and the brain serotonin (5-HT) system, limited to studies in adult humans and adult animals. In humans, neuroendocrine challenge studies in hypothyroid patients have shown a reduced 5-HT responsiveness that is reversible with thyroid replacement therapy. In adult animals with experimentally-induced hypothyroid states, increased 5-HT turnover in the brainstem is consistently reported while decreased cortical 5-HT concentrations and 5-HT2A receptor density are less frequently observed. In the majority of studies, the effects of thyroid hormone administration in animals with experimentally-induced hypothyroid states include an increase in cortical 5-HT concentrations and a desensitization of autoinhibitory 5-HT1A receptors in the raphe area, resulting in disinhibition of cortical and hippocampal 5-HT release. Furthermore, there is some indication that thyroid hormones may increase cortical 5-HT2 receptor sensitivity. In conclusion, there is robust evidence, particularly from animal studies, that the thyroid economy has a modulating impact on the brain serotonin system. Thus it is postulated that one mechanism, among others, through which exogenous thyroid hormones may exert their modulatory effects in affective illness is via an increase in serotonergic neurotransmission, specifically by reducing the sensitivity of 5-HT1A autoreceptors in the raphe area, and by increasing 5-HT2 receptor sensitivity.
The thyroid system and mood modulation in affective illness
Disorders of the thyroid gland are frequently associated with severe mental disturbances.1,2 This intimate association between the thyroid system and behavior has been the impetus for exploring the effects of thyroid hormones in modulating affective illness, and the role of the hypothalamic-pituitary thyroid (HPT) axis in the pathophysiology of mood disorders.3 Thyroid hormones (TH) have a profound influence on behavior and mood, and appear to be capable of modulating the phenotypic expression of major affective illness.3,4,5,6 Thyroid supplementation is now widely accepted as an effective treatment option for patients with affective disorders.7,8,9
Actions of thyroid hormones in the adult brain
It is well established that thyroid hormones are essential for both the development and maturation of the human brain, affecting such diverse events as neuronal processing and integration, glial cell proliferation, myelination, and the synthesis of key enzymes required for neurotransmitter synthesis.10,11 Thyroid deficiency during the perinatal period results in irreversible brain damage and mental retardation. However, despite this accepted body of knowledge and in disregard of the clinical and therapeutic observations in association with affective illness, the action of thyroid hormones in CNS function in adults has not been widely acknowledged by general endocrinologists. This lack of interest seems to have originated in the 1950s and 1960s, when early physiological studies suggested that oxygen consumption in the mature human brain did not change with changing thyroid status.12,13,14
Thus, in contrast to our understanding of thyroid hormone's critically important role in the development of the CNS, until recently, little has been known about the function and effects of thyroid hormones in the mature mammalian brain.15 However, with improved methods in basic research the action of thyroid hormones in the mature brain has become a subject of greater interest.16 There are several lines of evidence suggesting that thyroid hormones affect mature brain function. First, thyroid hormone receptors are prevalent in the mature brain. Nuclear receptors for T3, the thyroid hormone with the highest biological activity, are widely distributed in adult rat brain with higher densities of nuclear T3 receptors in phylogenetically younger brain regions—in the amygdala and hippocampus—and lower densities in the brain stem and cerebellum.17,18 A second line of evidence pertains to brain thyroid hormone metabolism. The process of 5-deiodination by which both thyroid hormones, T4 and T3, are metabolized to inactive iodothyronines has been demonstrated to be different in the adult brain from that in peripheral tissues. Specifically, the type D2 and type D3 deiodinases catalyze these metabolic processes in spatially distinct patterns in the central nervous system and appear to be segregated into specific cell types.19 D2 is expressed primarily in the brain and anterior pituitary gland where it metabolizes T4 to the active thyroid hormone form, T3. The activity of D2 in distinct regions of the brain varies widely, with the highest levels found in cortical areas and lesser activity in the midbrain, pons, hypothalamus and brainstem.20 In rat brain D2 is expressed in neurons, in particular in the nerve terminals, but also in astrocytes.21 Third, thyroid hormones have been detected in relatively high (nanomolar) concentrations in cortical tissue.22 In contrast to peripheral tissue where T4 concentrations usually far exceed those of T3, in the brain T4 and T3 concentrations are in an equimolar range.
Monoamines and mood
Over the past two decades it has become apparent that the monoamines, specifically norepinephrine and serotonin play a major role in mood modulation.23,24,25,26,27 These long track systems which begin in the brainstem and extend through the midbrain into the limbic system and cortex modulate the activity of many of the brain regions related to emotion and memory. The interdependence of these long tracks—including the dopamine system—with thyroid hormone metabolism has become better understood as our technology has improved.
The catecholaminergic system was initially investigated largely because of the known physiological association between sympathetic activity and thyroid hormones.26 Thyroid hormones appear to play an important role in regulating central noradrenergic (NA) function and it has been suggested that thyroid dysfunction may be linked with abnormalities in central NA neurotransmission.27 Evidence for a thyroid–NA interaction derives largely from immunohistochemical mapping studies demonstrating that T3 is concentrated in both nuclei and projection sites of central NA systems.28 Recent evidence that T3 is also delivered from the locus coeruleus to its NA targets via anterograde axonal transport indicates that T3 may function as a cotransmitter with norepinephrine in the adrenergic nervous system.29
However, the neuropharmacological effects and functional pathways underlying the therapeutic effects of thyroid hormones in patients with affective disorders are still unclear. One of the most intuitive hypotheses postulates the existence of a brain thyroid hormone deficiency in affective illness. Thyroid hormone therapy can then be considered a replacement therapy with a possible mechanism of action being its pharmacological effects on monoamine neurotransmitter systems, eg, by increasing β-adrenergic receptor activity and thus promoting the action of catecholamines at central receptor sites.27
The CNS serotonin system
As with the noradrenergic and dopaminergic systems, the bulk of the CNS serotonergic nerve terminals originate in the neuronal cell bodies of the brainstem raphe nuclei and project, both rostrally and caudally, to neuroanatomically discrete areas throughout the brain but with extensive innervation of the cerebral cortex and the limbic system.30 Although the serotonin system has been given prominence in recent deliberation regarding mood modulation, particularly since the advent of drugs that specifically interfere with serotonin neuronal reuptake systems, there has been little investigation of the relationship of this system to the thyroid system. This paper analyzes the existing literature pertaining to this relationship and explores areas which may be fruitful for further study.
The brain serotonin system and its role in depression
Basic and clinical research of the past three decades has yielded compelling evidence that the serotonergic system is intimately involved in the pathogenesis of depression.23,25,31,32 Changes in serotonergic neurotransmission have been repeatedly associated with the therapeutic response to antidepressant and mood stabilizing medication.23,33 Almost all currently employed treatments for depression, including the tricyclic antidepressants, the SSRIs, the MAO inhibitors, lithium and ECT, directly or indirectly augment serotonergic neurotransmission.34 Another line of evidence derives from the tryptophan-depletion paradigm, a procedure that lowers central serotonin levels, and which produces a rapid relapse of SSRI-responsive depression.33,35 Other support comes from studies demonstrating lowered levels of 5-hydroxyindoleacetic acid (5-HIAA), a metabolite of 5-HT whose levels reflect central serotonin activity, in the CSF in unmedicated depressed patients.25 In brain imaging studies, clinical depression was associated with reduced serotonin transporter availability.36,37
Objectives and elements of this review
This article explores the hypothesis that the mood modulating activity of thyroid hormones may be mediated in part by interaction with the brain serotonin system, specifically by enhancing cortical serotonergic neurotransmission. Because of the specific organization of the brain serotonin system, an analysis of the literature has been divided into anatomical areas: specifically the brainstem, the midbrain, the limbic system, and the cortex; those studies that were not specified in regard to brain area, as were many of the early studies, are referred to as ‘whole brain studies’. The other important element running through these analyses is the technical advance which has occurred over the 25 years that are the subject of our review. Specifically, many early studies were based upon crude analyses of levels of serotonin and its metabolites in homogenized brain tissue of animals that had been previously exposed to hypo- or hyperthyroid states. As technology advanced and our chemical dissection of thyroid hormone system metabolism gained specificity, more sophisticated studies emerged that have improved our understanding of turnover and receptor activity. In the 1990s, ligand studies and studies of transporter systems began to complement the earlier studies, and most recently, microdialysis techniques have provided new insights by measuring levels of serotonin in vivo. We have tried to reflect this technical advance in the analysis of the papers that are reviewed.
Methods of literature research
An attempt was made to identify all reports studying the interaction between thyroid hormones and the brain serotonergic system both in animals and in humans, but with a focus on studies in the adult brain. A computer-aided search of the National Library of Medicine MEDLINE database for 1966 to August 2000 using the subject headings ‘thyroid hormones’, ‘serotonin’, ‘brain’ and ‘affective disorders’ was performed, supplemented by the bibliographies of reports identified.
Results of the review
Effects of experimentally-induced hypothyroid states on brain serotonin system in animals
Historical perspective: studies in neonatal animals
Stimulated by the essential role of thyroid hormones in brain development, the effects of hypothyroidism on serotonergic neurotransmission were originally studied in neonatal rats. In these studies, 5-hydroxytryptamine (5-HT, serotonin) and 5-HIAA, the main 5-HT metabolite, were found to be significantly elevated and the serotonin precursor 5-hydroxytryptophan (5-HTP) to be decreased compared to euthyroid controls indicating an increased serotonin turnover rate in the neonatal period.38 Other data have demonstrated that neonatal hyperthyroidism induced by daily application of T3 also resulted in an increased turnover of 5-HT.39
Measurements of 5-HT and its metabolites in adult hypothyroid animals
In the adult rat brain, hypothyroidism generally induced lesser changes in the serotonergic system compared to the studies in neonatal animals. Thirteen studies were identified that measured the effects of experimentally-induced hypothyroidism on the serotonergic system. The methods and results of these studies are shown in Table 1. One early study measured brainstem 5-HT concentrations and did not find significant differences compared to euthyroid animals.40 Later, using more sensitive assay techniques, five studies measured 5-HT and 5-HIAA concentration or the 5-HIAA/5-HT ratio as an indicator of the serotonin turnover and reported increased 5-HT metabolites in the brainstem41,42,43,44,45 (notice: one study calculated the inverse ratio, 5-HT/5-HIAA41). Reduced 5-HT concentrations in the cortex,46,47 and reduced concentrations of the serotonin precursor 5-HTP were reported in the whole brain48 of hypothyroid adult rats. These findings of increased 5-HT turnover in the brainstem and decreased levels of 5-HT and its precursors in the cortex/whole brain are in accordance with the hypothesis that increased brainstem 5-HT turnover might activate raphe 5-HT1A autoreceptors and subsequently decrease serotonin release in the cortical projection areas.23
Receptor studies: changes in 5-HT1A and 5-HT2 receptors in the adult hypothyroid brain
Among the many 5-HT receptor subtypes with different regional distributions throughout the CNS, it is the 5-HT1A and 5-HT2 receptor densities that have been most studied in experimentally-induced hypothyroid animals. The 5-HT1A receptor subtype, predominantly located on the cell bodies and dendrites of the serotonergic neurons in the raphe nuclei, functions as a control point of activity for these neurons. In contrast, the postsynaptic entities of 5-HT neurotransmission consist of several subtypes of 5-HT2 receptors located in distinct projection areas of the 5-HT neurons.
In experimentally-induced hypothyroid states the 5-HT1A (presynaptic) receptor density in the brainstem and midbrain was not altered49,50,51,52 (Table 1). Studies on the density of 5-HT1A (postsynaptic) receptors outside the brainstem yielded contradictory results. An increase in cortical and hippocampal 5-HT1A (postsynaptic) receptors was observed by Tejani-Butt et al50 but not by Hong et al49 and Kulikov et al,51 who found no significant differences compared to euthyroid adult rodents.
An early study by Mason et al52 found a decrease in 5-HT2 receptor density in the striatum but not in the cortex of hypothyroid adult rats. However, when 5-HT2A receptors were assessed selectively in severely hypothyroid rats, a significant cortical reduction was recently reported by Kulikov et al.51
Hence in summary, several lines of evidence indicate that an experimentally-induced hypothyroid state in adult rodents is associated with an increased 5-HT turnover rate in the brainstem, but not with a change in 5-HT1A autoreceptor density in the raphe area. There is also some evidence that hypothyroid states result in a decrease in cortical 5-HT serotonin concentrations and 5-HT2A receptor density.
Effects of thyroid hormone application on brain serotonin system in animals
Fourteen studies were located that measured the acute effects of thyroid hormone (T3 and/or T4) on 5-HT and/or L-tryptophan, 5-hydroxytryptophan (5-HTP), and 5-HIAA concentrations in adult rodent brain (for methods and results see Table 2).
Acute effects of thyroid hormones on levels of 5-HT and its metabolites in the brainstem and midbrain
Rastogi and Singhal53 observed an increase in the 5-HT precursor L-tryptophan (L-TP), and Heal and Smith54 found an increase in both 5-HT and 5-HIAA in the midbrain after T3 application in euthyroid animals. In contrast, Henley et al42,45 examined animals after thyroidectomy, which resulted in an elevated serotonin turnover rate; in these animals, T3 replacement resulted in a significant decrease in the 5-HIAA/5-HT ratio in the brainstem. In the two studies of thyroid replacement after thyroidectomy, T3 replacement for longer than 3.5 days reduced the 5-HT turnover in caudal brainstem to completely normal values. The activity of tryptophan hydroxylase (TPH), the rate-limiting enzyme in the synthesis of 5-HT, was found unaltered in the midbrain after T3 application.51,53
Acute effects of thyroid hormones on levels of 5-HT and its metabolites in cortex and whole brain
More consistent than the effects of ‘micro-dissection’ reported above were the results of studies that measured the effects of thyroid hormone on levels of 5-HT and its metabolites in the cortex or in the whole brain. Thyroid hormone application to euthyroid rodents increased cortical or whole brain 5-HT, 5-HTP and 5-HIAA concentrations in 10 studies.41,46,47,48,54,55,56,57,58,59 These results indicating increased cortical 5-HT turnover were consistent despite changing technologies over a 25-year period, and may be considered robust. In only one study did the whole brain 5-HT level not increase after thyroid hormone administration.60
Chronic effects of thyroid hormones on levels of 5-HT and its metabolites in cortex and whole brain
Fewer studies assessed the effects of a single vs multiple T3 or T4 application on cortical serotonergic neurotransmission in euthyroid rodents54,55,58,59 (Table 2). In three of four studies, increases in cortical or whole brain 5-HT, 5-HTP and 5-HIAA contents were observed only after repeated (chronic) thyroid hormone application.54,55,59
Thus, the increased concentration of 5-HT and its precursors and metabolites in the cortex or whole brain that were observed in the majority of studies were more pronounced after repeated (chronic) thyroid hormone application. Similar studies investigating the effects on 5-HT levels in the brainstem and midbrain are less consistent (Table 2).
Effects of thyroid hormones on 5-HT1A receptor density and sensitivity
Three autoradiographic studies have reported that thyroid hormone application induces no significant reduction in raphe and midbrain 5-HT1A receptor density.49,50,51 However, a recent study by Gur et al59 indicated a loss of autoinhibitory 5-HT1A receptor sensitivity mediated by T3 (Table 2). Gur et al,59 for the first time in the study of the 5-HT-thyroid interaction, used an in vivo microdialysis technique that allows the measurement of 5-HT concentrations in the brain with a high degree of accuracy in the living animal. In this study, the decrease in hippocampal and cortical serotonin release that follows the application of a 5-HT1A agonist via the activation of inhibitory autoreceptors was significantly reduced by T3 alone, or T3 combined with clomipramine administration in euthyroid rats.59
Effects of thyroid hormones on 5-HT2 receptor density and sensitivity
The database concerning changes in 5-HT2 receptor density after thyroid hormone application reveals contradictions. Mason et al52 observed an increase in 5-HT2 receptor density in the striatum, hippocampus and cortex of thyroidectomized rats only after long-term application of a relatively high dose of either T3 (250–1000 μg kg−1 for 7–10 days) or T4 (250–500 μg kg−1 for 7–10 days). Kulikov et al51 showed that T4 application in thyroidectomized animals returned cortical 5-HT2A receptor densities to normal levels, irrespective of whether a replacement or high T4 dose was applied (15 μg kg−1 vs 200 μg kg−1 T4 for 21 days each). Lower doses and shorter duration of T3 application yielded different results: Sandrini et al58 found no significant effect on 5-HT2 receptor density in the hippocampus and a decrease in cortical 5-HT2 receptor density after application of T3 in euthyroid rats (100 μg kg−1 for 3–7 days). In the study of Heal and Smith54 the same T3 dose applied to euthyroid rats (100 μg kg−1 for 10 days) also decreased cortical 5-HT2 receptor density. A reduction in prefrontal 5-HT2A receptors was observed after coadministration of T3 and the antidepressant desipramine.61
A thyroid hormone-induced change in receptor sensitivity was observed for cortical 5-HT2 receptor function in adult euthyroid rats. Heal and Smith54 observed an increase in 5-HT2 receptor sensitivity after short-term T3 application, and Atterwill60 reported similar findings after both short- and long-term T3 application (Table 2). Under stress conditions, on the other hand, administration of high doses of T4 (350 μg kg−1 for 7 days) resulted in a blunting of the immobilization stress-induced activation of hypothalamic 5-HT2 receptors.62
Effects of low vs high doses of thyroid hormones on the 5-HT system
Some studies compared the effects of a low (replacement) vs a high dose of thyroid hormone on 5-HT receptors.50,51,52 The effects on the serotonergic system were more pronounced with higher doses of thyroid hormone in two out of three studies.50,52 Cortical and hippocampal 5-HT2 receptors.52 and hippocampal and hypothalamic (postsynaptic) 5-HT1A receptor density50 were significantly increased after administration of higher doses of T3 or T4. However, excess serum thyroid hormone in thyroidectomized rats, achieved by administration of high doses of T4, did not produce any changes in cortical 5-HT2A receptors when compared to thyroidectomized animals with normalized thyroid hormone levels.51
In summary, 5-HT receptor studies in adult euthyroid rodents indicate that thyroid hormone application may desensitize presynaptic 5-HT1A raphe autoreceptors, and thus increase cortical serotonin release, an effect similar to that described after addition of the 5-HT1A receptor antagonist pindolol to an ongoing SSRI treatment.63 The receptor studies also indicate that thyroid hormone application may increase cortical 5-HT2 receptor sensitivity. This increase in 5-HT2 receptor function does not seem to be linear, as stress-induced activation of hypothalamic 5-HT2 receptors was blunted in hyperthyroid rats.62 Cortical 5-HT2 receptor densities were only increased after prolonged treatment with relatively high doses of thyroid hormone in thyroidectomized rats. In contrast, standard doses of T3 in euthyroid rats resulted in a decrease in the number of cortical 5-HT2 receptors.
Clinical studies of the thyroid–serotonin interaction
The serotonin system in hypothyroid patients and effects of thyroid hormone replacement
In three studies, parameters of the serotonergic system were examined in hypothyroid patients (Table 3). Sjöberg et al64 measured 5-HT, L-TP and 5-HIAA concentrations in the CSF of seven hypothyroid patients before and after T4 replacement. A significant decrease in the serotonin precursor L-TP after T4 treatment was found which may indicate increased conversion to 5-HT. However, no significant increase in CSF 5-HT or 5-HIAA concentrations after T4 replacement was found.
Several studies in an effort to evaluate functional components of the serotonergic system in humans have examined the neuroendocrine responses to d-fenfluramine (D-FEN). D-fenfluramine stimulates the serotonergic projecting pathways from the dorsal raphe nuclei to the paraventricular nucleus of the central hypothalamus and seems to release cortisol via activation of 5-HT1A or 5-HT2 receptors.65,66 Two such challenge studies found a significantly decreased D-FEN-induced cortisol response in hypothyroid patients65,67 (Table 3), which normalized with T4 replacement.67 This enhancement of central 5-HT2 receptor activity after T4 application in previously hypothyroid patients67 is in agreement with the findings in animal studies of increased 5-HT2 receptor sensitivity after T3 application.54,60
The serotonin system in hyperthyroid patients
One study examined peripheral 5-HT concentrations and the activity of the metabolizing enzyme monoamine oxidase (MAO) before and after treatment in 45 hyperthyroid patients and compared the activity to that present in healthy, euthyroid controls. Serotonin blood levels were found to be increased, and MAO activity decreased, in the hyperthyroid state68 (Table 3). After 3 months of treatment with carbimazole and the associated decline of plasma T3 and T4 concentrations towards normal levels, MAO activity increased and plasma serotonin concentrations decreased, however, not to within the range of the normal control subjects.68 These findings suggest altered serotonin metabolism during the hyperthyroid state.
Serotonin-HPT system interaction in patients with major depression
The interaction of the 5-HT system and thyroid axis function was investigated in patients with major depression using the D-FEN stimulation test69 (Table 3). Patients with HPT system abnormalities (as indicated by a blunted TSH response to the TRH stimulation test suggesting ‘hyperactivity’ of the HPT system) had hormonal D-FEN responses comparable to those of healthy controls, while patients without HPT abnormalities showed reduced hormonal responses to D-FEN compared to controls. The authors suggested that the blunted TSH response to TRH stimulation found in a subgroup of depressed patients might be a compensatory mechanism for diminished central 5-HT activity.69
Implications for thyroid hormone modulation of mood disorder
Does the information reviewed here of the interaction of thyroid hormones with serotonergic neurotransmission, have relevance for our understanding of the mood modulating effects of thyroid hormones in the clinical setting, and can it promote our understanding of the pathophysiology and treatment of mood disorders?
The molecular mechanisms underlying the efficacy of thyroid hormone treatment in patients with mood disorders, and in patients with primary hypothyroidism who have comorbid depression, are not known. From the few studies in humans with thyroid dysfunction, there is some evidence from the neuroendocrine challenge studies that hypothyroid status is associated with a reduced 5-HT responsiveness. Furthermore, this appears to be reversible with thyroid replacement therapy.65,67 However, given the small number of studies in humans definitive conclusions cannot be drawn. Not only is the number of studies limited but the sample sizes in the studies were small and the methods employed to assess central 5-HT function varied considerably. It is also questionable whether the peripheral blood and CSF content of 5-HT and its metabolites provide an index of brain serotonergic neurotransmission,70 while neuroendocrine challenge studies provide only an indirect way of ‘probing’ central 5-HT function.71
In contrast, results from studies in animals provide strong evidence that thyroid status has a considerable impact in serotonergic neurotransmission in the adult brain. Experimentally-induced hypothyroid states result in an increase in 5-HT turnover in the brainstem. Increased 5-HT turnover in hypothyroid states may lead to an increase in raphe 5-HT1A autoreceptor activity and a decrease in cortical 5-HT concentrations (Figure 1). This observation indicates that in the raphe area increased serotonin turnover may activate inhibitory autoreceptors on the serotonergic cell bodies and thus, reduce serotonin turnover in the cortical and subcortical projection areas of these serotonergic neurons. The value of direct measurements of 5-HT and its precursors/metabolites from homogenized brain tissues is limited. However, in more recent receptor studies, it was found that hypothyroid states result in a decrease in cortical 5-HT2A receptor density, an observation that reinforces the postulate that hypothyroidism is associated with a reduced cortical serotonergic neurotransmission.
Thyroid hormone application may increase cortical serotonergic neurotransmission via two independent mechanisms: (1) by reducing the sensitivity of 5-HT1A autoreceptors in the raphe area, thus disinhibiting cortical and hippocampal serotonin release; and (2) by increasing cortical 5-HT2 receptor sensitivity, a potentially independent way of increasing 5-HT neurotransmission (Figure 1). These latter two potential mechanisms for thyroid hormone modulation of serotonin transmission warrant further elaboration. With respect to the first mechanism, it is important to note that in animal studies it has been demonstrated that an acute blockade of serotonin transporters by SSRIs increases raphe serotonin concentrations immediately.72,73 However, application of SSRIs also activates presynaptic 5-HT1A autoreceptors located on serotonergic cell bodies in the raphe area and may thus inhibit serotonin release in the cortical projection areas.23 Subsequently, an increase in frontal serotonin release is only found after prolonged SSRI application,74 when increased synaptic serotonin concentrations in the brainstem induce a down-regulation of 5-HT1A autoreceptors, or after a drug-induced blockade of 5-HT1A receptors.75 This mechanism has been postulated to be responsible for the delayed antidepressive effects of SSRIs.23 In accordance with this hypothesis, a blockade of 5-HT1A receptors would facilitate the antidepressive action of SSRIs in patients with major depression.63,76 A similar mechanism involving a desensitization of presynaptic 5-HT1A autoreceptors could be involved in the efficacy of thyroid hormones to accelerate and augment antidepressant agents. This could also explain why T3 augmentation treatment in patients with depression usually takes effect in the first 2 weeks after initiation of treatment. Thus, potential similarities exist in the putative mechanism of action of the 5-HT1A receptor antagonist pindolol and of thyroid hormones, and that both pindolol and T3 are found to speed recovery from depression.7,63,76 With respect to the potential second mechanism, it should be noted that reduced 5-HT2 receptor sensitivity has been observed in most,77,78,79,80 although not all studies of patients with major depression.81 Subsequently, treatment with clomipramine or SSRIs increased 5-HT2 receptor sensitivity.82,83 Thus, thyroid hormone-induced increases in 5-HT2 receptor sensitivity might potentiate the effects of antidepressant drugs on the 5-HT2 receptors, as has been demonstrated in studies with animals54,60 and humans.65,67 However, a hypothesis of 5-HT2 receptor-mediated antidepressive effects of thyroid hormones faces some limitations. First, the serotonin receptor subtypes perturbed in the neuroendocrine challenge studies in humans are unknown. In these clinical studies, the in vivo serotonin receptor sensitivity is indirectly assessed by measuring cortisol or prolactin release after serotonergic challenge with various drugs (5-hydroxytryptophan, fenfluramine or meta-chlorophenylpiperazine).82,83,84,85 The observed effect might be mediated via various contributions of both 5-HT2C and 5-HT1A receptor stimulation.66,86,87 Second, autoradiographic and brain imaging studies, measuring not the sensitivity but the density of 5-HT2 receptors among patients with major depression, observed significant decreases in frontocortical 5-HT2 receptor availability after antidepressive drug treatment.88,89 Of course, increases in receptor sensitivity may be accompanied by decreases in receptor number to avoid overstimulation of the monoamine neurotransmitter system and this may be the explanation. Such a hypothesis would be supported by the observation in animal studies of Heal and Smith,54 Sandrini et al58 and Watanabe61 that cortical 5-HT2 receptor density was reduced after thyroid hormone application, a procedure which has been shown to increase the sensitivity of this receptor subtype.54,60,65,67
Post-receptor and molecular actions of thyroid hormones
While not the primary focus of this review, other sites of thyroid hormone action that are important for understanding thyroid hormone effects on brain function, include post-receptor, transcriptional, and gene regulatory mechanisms. A series of studies indicate that these signaling pathways, downstream from receptors, are also influenced by changes in thyroid status. In rats, hypothyroidism induced a significant up-regulation of G-protein complexes in synaptosomal membranes from different brain regions.90 Conversely, in studies of euthyroid animals, treatment with T3 decreased the abundance of the alpha-subunits of Gi in synaptosomal membranes of the cerebral cortex.91 Impaired signal transduction via adenylate cyclase and inositol phosphatase has also been demonstrated in the adult brain of hypothyroid rats. Hypothyroid rats also showed enhanced inhibition of adenylate cyclase in synaptosomal membranes by GTP,92 and decreased formation of inositol phosphate in response to the muscarinic cholinergic agonist carbachol.93 Thus, it appears that thyroid hormones exert an important influence on the activity and synthesis of G-proteins and the receptor/G-coupling systems that serve the monoamine receptor system. Thus thyroid hormone deficiency leads to an impairment in adenylate cyclase activity and phosphoinositide-based signaling pathways involved in transcriptional activities in the adult CNS.16,90,91
The molecular action of thyroid hormone is mediated through specific nuclear TH receptors (TRs) α and β (β1, β2), functioning as ligand-dependent transcription factors that increase or decrease the expression of target genes.94 Although the two genes that encode the related TRα and TRβ are differentially expressed, the two receptors usually coexist in the same cell type. The relative contribution of these two TR genes encoding for TRα and TRβ in mediating a particular T3 response is poorly understood because of a lack of in vivo functional information. Knock-out mouse models lacking a particular TR isoform have been generated to explore the relative contribution of each of the TR isoforms to the TH-mediated regulation of various biological processes in different tissues. However, the animals tested to date showed little overt behavioral or neuroanatomical abnormalities compared with animals rendered hypothyroid by thyroidectomy95,96,97,98 suggesting that other TR forms may compensate or substitute for lacking or defective receptors in these knock-out mouse models. In contrast, a TR knock-in mouse model with a T3 binding mutation in the TRβ locus resulted in severe neuroanatomical and behavioral dysfunction (eg, abnormal hippocampal gene expression of brain-derived neurotrophic factor (BDNF), myelin basic protein (MBP), and tyrosine protein kinase receptor B (TrkB), learning deficiency, and cerebellar dysfunction) indicating a specific and deleterious action of unliganded TR in the brain.99 Recent studies have also indicated that the adult brain has various genetic loci that are responsive to thyroid hormones.100 Among the most extensively studied loci is RT3/neurogranin, a brain-specific gene encoding a protein kinase C substrate that binds calmodulin and is located in dendritic spines and forebrain neurons;101 in these studies, adult-onset hypothyroidism led to a decrease of RC3/neurogranin, an effect that was reversible with T4 treatment.102 Thyroid hormone also modulates glucose transport processes across the blood–brain barrier (BBB)103 and in astrocytes,104 and may alter the expression of glucose transporter one (GLUT-1) gene, the principal isoform responsible for glucose transport across the BBB.105 Furthermore, the effects of thyroid hormones on CNS gene expression have been demonstrated for various other neuroactive peptides, eg, TRH,106 corticotropin-releasing hormone (CRH),107 brain-derived neurotrophic factor, nerve growth factor and neurotrophin 3,108,109 angiotensinogen,110 and several structural brain-specific genes (eg, myelin-associated glycoprotein, Pcp-2, microtubule-associated proteins).111 Of particular relevance is a recently reported interaction between thyroid and serotonin systems, indicating synergistic effects of T3 and 5-HT1A receptors on hippocampal brain-derived neurotrophic factor (BDNF) expression. T3 administration prior to treatment with a 5-HT1A agonist caused a downregulation of hippocampal BDNF mRNA expression in adult rats.111 These molecular studies clearly indicate that thyroid hormones actively regulate a broad spectrum of genes in the adult brain although the behavioral significance of such activity is unknown.
In our review we found evidence, particularly from results in animal studies, to support the hypothesis that thyroid status impacts the serotonin system in the adult brain, and that increasing thyroid hormone levels increase serotonin neurotransmission. Given the important role of the serotonin system in the pathogenesis of depression we speculate that the serotonin system may be involved in the mood modulating effects of thyroid hormones among patients with affective disorders. This hypothesis would explain why thyroid hormones are most effective in patients with affective disorders when administered as an adjunctive treatment to antidepressants and/or mood stabilizers that perturb the serotonin system. This is also supported by evidence that thyroid hormones alone appear to have limited clinical use in affective illness.4,5
It must be emphasized however that this interaction with the serotonin system is probably only one of the mechanisms through which thyroid hormones may have modulatory effects in mood disorders. Thyroid hormones interact with a broad range of neurotransmitter systems thought to be involved in the regulation of mood including post-receptor and signal transducing processes, as well as gene regulatory mechanisms.
We thank Georg Juckel, MD, and Faustino Lopez-Rodriguez, PhD, MD, for comments on the manuscript. This work has been supported by a grant from the Deutsche Forschungsgemeinschaft to MB (Ba 1504/3–1).