The impact of autoimmune thyroid disease on cognitive and psychiatric disorders: focus on clinical, pre-clinical and molecular studies

Abstract

Autoimmune thyroid disease (AITD) is the most prevalent organ-specific autoimmune condition, encompassing Graves’ disease (typically linked with hyperthyroidism) and Hashimoto’s thyroiditis (generally associated with hypothyroidism). The growing body of evidence suggests that AITD can interfere with brain function. Here, we review the impact of AITD on cognition, mood, and psychiatric disorders by analysing data from clinical, animal, ex vivo and in vitro studies to reveal the molecular mechanisms by which AITD affects brain function. Most reports indicate a stronger association between cognitive impairments and hyperthyroidism (including AITD-related) than hypothyroidism. Both hypothyroidism and hyperthyroidism are linked with a higher risk of depression. At least some of those effects can be mediated by altered concentrations of T3 (3,3′,5-triiodo-L-thyronine), which regulates gene expression in the brain microenvironment, affecting neurogenesis, angiogenesis, neurotransmitter release, and synaptic transmission. Diminished TSH (thyrotropin) signalling may also impair learning and memory by inhibiting the Wnt5a-β-catenin pathway. Thyroid disorders may also contribute to neurodegeneration by T3-mediated attenuation of amyloid-β elimination or TRH-induced formation of neurofibrillary tangles. Surprisingly, most clinical studies do not specify the immune origin of hypothyroidism or hyperthyroidism, therefore further studies involving large, well-characterised patient cohorts are needed to clarify the relationships between AITD and cognitive impairments and psychiatric disorders. Furthermore, data on the effect of anti-thyroid antibodies on brain function are scarce and inconclusive. Given the association between hyperthyroidism and an increased risk of dementia, cognitive impairment and mood disorders, adequate treatment and careful monitoring of AITD patients are essential to prevent the induction of exogenous hyperthyroidism.

Introduction

This review examines the influence of autoimmune thyroid diseases (AITDs), specifically Graves’ disease (GD) and Hashimoto’s thyroiditis (HT), on the development and progression of cognitive and psychiatric disorders. Furthermore, it synthesises the current understanding of the molecular and cellular mechanisms underlying the associations between AITD and neuropsychiatric disturbances.

AITD is triggered by an autoimmune attack directed against proteins expressed by thyroid cells and involved in thyroid hormone (TH) synthesis: the thyrotropin receptor (TSHR), thyroid peroxidase (TPO), and thyroglobulin (Tg). Other autoantigens play a smaller role in AITD pathology (1). The autoimmune assault leads to the dysfunction of the thyroid gland, with two key AITD manifestations: hypothyroidism (TH deficiency) or TH excess (hyperthyroidism) (1). Both hypo- and hyperthyroidism can substantially influence brain function, thereby affecting cognition and mood.

Cognitive impairment and psychiatric disorders in AITD-related thyroid dysfunctions

The impact of autoimmune thyroid dysfunctions on broadly defined cognitive impairments, encompassing deficits in mental functions such as memory, attention, reasoning, perception, language, and executive functions, as well as on psychiatric disorders such as depression, anxiety, and manic psychosis, has been extensively studied in numerous clinical trials involving diverse patient groups. Systematic analyses, meta-analyses, population-based studies, and observational research highlight the associations between specific thyroid dysfunctions and certain types of cognitive and psychiatric disorders (2, 3). Awareness of these associations is critical for: i) diagnosing and managing cognitive and psychiatric disorders, recognising thyroid dysfunction as a modifiable risk factor; ii) closely monitoring AITD patients for an increased risk of psychiatric and cognitive disorders; and iii) avoiding overtreatment of patients with subclinical hypothyroidism in the context of HT to prevent the induction of exogenous hyperthyroidism and its associated risk of cognitive impairments.

The influence of AITD on neurological disorders is highly complex and may stem from TH excess or deficiency, altered TSH levels, or elevated titres of thyroid antibodies such as anti-TPO, anti-Tg, and anti-TSHR, characteristic for HT and GD, respectively. In most of the clinical studies summarised below, it was not possible to distinguish whether the potential impact of thyroid dysfunction on cognitive impairment and psychiatric disorders was attributable to TH deficiency or excess, abnormal TSH levels outside the reference range, or immune system activation and the presence of thyroid antibodies. In studies where the autoimmune origin of thyroid dysfunction was considered during patient selection, this has been explicitly noted (Supplementary Tables S1 and S2 (see section on Supplementary materials given at the end of the article)). The potential mechanisms through which TH, TSH, and thyroid-specific antibodies affect the nervous system are analysed in subsequent sections of this review.

Cognitive impairment and dementia

A meta-analysis of 15 longitudinal and cross-sectional studies involving more than 300,000 patients showed that both overt and subclinical hyperthyroidism increase the risk of dementia, with odds ratios of 1.14 and 1.56, respectively. No such association was found for either overt or subclinical hypothyroidism (4). Similarly, a meta-analysis of 11 cohort and case–control studies (24,000 participants) indicated that elevated fT4 levels and reduced TSH levels correlate with an increased risk of dementia, while no such correlations were found for reduced fT4 levels and elevated TSH (5). Moreover, a long-term observational study lasting 9 years and involving 60,000 patients over 65 years of age showed that low TSH, regardless of its cause (endogenous or exogenous thyrotoxicosis), is associated with an increased risk of cognitive impairment (6). In addition, Agrawal et al. reported a higher prevalence of hyperthyroidism in patients with dementia of both Alzheimer’s disease and vascular origin compared to patients without cognitive impairment (7).

Clearly, weaker associations were found between cognitive impairment and hypothyroidism. Admittedly, an analysis of a Danish registry of 100,000 patients showed a 1.22-fold higher risk of dementia in individuals with hypothyroidism and a 12% increase in this risk for every 6 months of persistently elevated TSH (8). A smaller Mexican study (1,750 participants) also found an association between overt hypothyroidism and mild cognitive impairment (9). However, meta-analyses involving a total of more than one million individuals did not confirm an association between either overt or subclinical hypothyroidism and dementia (4, 5, 10).

Numerous studies have focused on patients with only subclinical hypothyroidism and hyperthyroidism, as these are the most common forms of thyroid dysfunction in the population. A meta-analysis of 11 prospective studies involving 16,000 patients showed that subclinical hyperthyroidism was associated with a higher risk of dementia (HR = 1.67) than subclinical hypothyroidism (HR = 1.14) (11). Similar results were found in a cohort study involving 2,500 participants aged 70–79 years, where low TSH correlated with a higher risk of dementia and cognitive decline (12). In contrast, an association between TSH remaining in the upper normal range (5.5–12.0 mIU/L) and dementia was observed only in participants under 75 years of age (13).

The results of the studies cited above (see Supplementary Table S1 for more details) suggest a stronger association between cognitive impairment and hyperthyroidism, both overt and subclinical, than hypothyroidism. However, a large cross-sectional study involving 74,000 elderly participants did not confirm an association between either overt or subclinical hypothyroidism and hyperthyroidism, raising questions about the relevance of screening for thyroid dysfunction in the context of cognitive decline in older adults (14).

Research on the role of autoimmunity in thyroid disease in the development of cognitive impairment has also yielded inconclusive results. In a cohort study involving 12,000 patients, hyperthyroidism increased the risk of dementia, but no association was found between this risk and the presence of anti-TPO antibodies (15). Similarly, an analysis of over 50,000 patients with hyperthyroidism did not show differences in dementia risk between patients with elevated anti-TSHR antibody titres in hyperthyroidism due to Graves’ disease and those with non-autoimmune hyperthyroidism resulting from toxic multinodular goitre without positive antibody titres (16). In another study, patients with Hashimoto’s disease in the euthyroid state had lower cognitive test scores than those treated with T4 for other reasons (17). Conversely, higher baseline anti-TPO titres in euthyroid patients with steroid-responsive encephalopathy and associated autoimmune thyroiditis correlated with better cognitive function (18). However, larger population-based studies did not show an association between anti-TPO antibodies and cognitive decline (19, 20) (Supplementary Table S1).

Extensive clinical studies have not provided conclusive evidence of the impact of thyroid disorders on cognitive function and dementia development, nor a clear association between elevated anti-TPO or anti-TSHR antibody titres and cognitive impairment. Discrepancies in findings may stem from differences in study protocols and methods used to assess thyroid function and cognitive impairment (2). However, most studies suggest a stronger association between the risk of cognitive impairment and hyperthyroidism than hypothyroidism. Therefore, hyperthyroidism, even in its subclinical form, should be treated not only to reduce the risk of coronary artery disease (in accordance with current clinical guidelines) (21) but also to lower the risk of dementia or cognitive decline. In contrast, excessive treatment of subclinical hypothyroidism, particularly in elderly individuals, is not recommended, as it may lead to exogenous hyperthyroidism and increase the risk of dementia (6).

Psychiatric disorders: depression, anxiety, mental fatigue

A meta-analysis of 20 case-control and population-based studies involving more than 36,000 patients with hypothyroidism (overt or subclinical) showed a strong association with psychiatric disorders such as depression (OR = 3.31) and anxiety (OR = 2.32). However, these results may be overstated due to the inclusion of case–control studies that involved patients with more severe symptoms in the meta-analysis, as well as the inclusion of cases of subclinical depression in the analyses (22, 23). Indeed, a meta-analysis involving almost 350,000 participants, which excluded case–control studies, found weaker associations of overt (OR = 1.77) and subclinical (OR = 1.13) hypothyroidism with the occurrence of depression. Importantly, the presence of anti-TPO antibodies was not associated with an increased risk of depression, suggesting that low TH levels may play a greater role in the development of psychiatric disorders than autoimmunity per se (24) (Supplementary Table S2).

An association between hyperthyroidism and psychiatric disorders has also been demonstrated, including an increased risk of psychiatric hospitalisation and the use of psychotropic medication. However, this applied only to patients with hyperthyroidism due to Graves’ disease and not to those with toxic nodular goitre. This suggests that in hyperthyroidism, autoimmunity – rather than excess hormones – may contribute to psychiatric disorders (25), in contrast to hypothyroidism, where elevated anti-TPO antibody titres were not associated with a higher risk of depression (24). In addition, a study of 65 premenopausal women with hyperthyroidism due to Graves’ disease showed high rates of depression, mental fatigue, and anxiety during the active phase of the disease, with significant improvement after achieving a euthyroid state following treatment. However, mental fatigue and depression persisted in some patients despite achieving euthyroidism, particularly in younger individuals with pre-existing psychiatric disorders or mild ocular symptoms (26, 27).

Less conclusive results were obtained in studies investigating the associations between psychiatric disorders and subclinical thyroid dysfunction (see Supplementary Table S2 for more details). A meta-analysis of 15 studies involving 12,000 patients showed a 2.35-fold increased risk of depression in subclinical hypothyroidism, particularly in younger patients under 60 years of age. In addition, L-thyroxine treatment did not reduce the risk of depression (28). Consistent with these findings, another meta-analysis involving 6,000 patients with subclinical hypothyroidism showed an increased risk of depression, but only in patients under 60 years of age. In these patients, treatment with L-thyroxine also did not reduce the risk of depression (29). In contrast, a large prospective study involving more than 220,000 participants found no significant association between subclinical hypothyroidism and the onset of depressive symptoms during a 2-year follow-up (30).

Similarly, a smaller observational study involving more than 20,000 patients, but with a longer follow-up period of more than 8 years, found no association between subclinical hypothyroidism or subclinical hyperthyroidism and the occurrence of depressive symptoms (31). In addition, another prospective study with a 3-year follow-up found no changes in depression scale scores in patients over 70 years of age with subclinical hypothyroidism. Interestingly, these changes were significant in patients with subclinical hyperthyroidism, suggesting that it may be associated with an increased risk of depression in an older population (32).

In summary, evidence suggests a higher risk of depression in patients with overt hypothyroidism and hyperthyroidism, while the impact of subclinical thyroid dysfunction on psychiatric disorders remains inconclusive. However, certain subgroups appear more vulnerable: younger patients, particularly women, in subclinical hypothyroidism, and older adults in subclinical hyperthyroidism. Notably, current findings do not support routine L-thyroxine supplementation for subclinical hypothyroidism to prevent or treat depression. In older patients, its use requires careful consideration due to the risk of inducing exogenous hyperthyroidism, which may further elevate the risk of depression.

Establishing a true link between specific types of thyroid dysfunction and psychiatric disorders requires the precise design of studies in specific patient groups and the adoption of consistent criteria for assessing psychiatric disorders, including conditions such as moderate depression, clinical depression, or mental fatigue. In addition to precisely designed clinical trials, it is also necessary to understand the molecular mechanisms underlying the increased risk of cognitive and psychiatric disorders in thyroid disease.

AITD-related cognitive and mood impairment in animal models

The observational clinical studies provide data on the associations between thyroid disorders and cognitive impairments in humans. The causative links between thyroid disease and cognition have been verified by experimental studies using animal models that also help in the understanding of the impact of AITD on cognitive impairment.

Cognitive impairment and dementia: animal models

Several animal models provided evidence of the impact of hyperthyroidism on cognitive abilities (Supplementary Table S3). Hyperthyroidism induced by the supplementation of T4/T3 in rodents impaired learning ability and spatial memory (33, 34, 35), as well as recognition memory (35). The impairment affected both short-term and long-term spatial memory (33). Similar cognitive impairment, accompanied by reduced neurogenesis, was observed in a mouse model of TSHR deletion in hippocampal neurons (TSH^CKO mice), which is characterised by a condition similar to subclinical hyperthyroidism only within the brain, with preserved normal peripheral hormone levels (36). An interesting study was conducted by Lou et al. (37), in which female mice with Alzheimer’s disease (APP/PS1 mice) were induced to develop Graves’ disease by administration of Ad-TSHR289 adenovirus. Behavioural tests indicated that autoimmune hyperthyroidism exacerbates cognitive impairment and increases amyloid-β deposits in the hippocampus, hallmarks of Alzheimer’s disease.

Numerous experimental studies also indicate that hypothyroidism leads to cognitive deficits. Learning impairments and difficulties in spatial memory were observed in rodents with TH deficiency induced by the administration of substances inhibiting hormone synthesis, propylthiouracil (PTU), or methimazole (MMI) (38, 39, 40, 41, 42). However, Salazar et al. (43) showed that the impairment only affected the learning stage without affecting memory. Animals with overt hypothyroidism also exhibited decreased recognition memory and impaired emotional memory (39, 40, 41, 44). Further assessment of spontaneous locomotor activity remained unchanged, indicating that the observed impairment in the behavioural tests is due to impaired cognitive function and memory rather than general motor activity. Interestingly, hypothyroidism may induce age-dependent cognitive impairment. Young female Kunming mice show deficits in spatial, recognition, and emotional memory. In contrast, middle-aged mice show only moderate impairment in spatial and recognition memory, whereas old mice performed similarly in all tests to age-matched wild-type controls (45).

Subclinical hypothyroidism may also cause cognitive impairment in animal models, although these deficits are less severe than those observed in overt hypothyroidism (46). It was observed that cognitive impairments positively correlated with TSH levels, whereas exploratory activity negatively correlated with TSH levels. Another study compared the effects of TH deficiency and excess on cognitive functions (47). In female rats with both overt and subclinical hyperthyroidism, as well as overt hypothyroidism, a decrease in learning and spatial memory was noted compared to the control group. In contrast, no significant differences were observed in the subclinical hypothyroidism group. Normal cognitive function in hypothyroid animals can be restored by supplementation with either T3 (48) or T4 (49). However, Xu et al. (44) observed only a partial reversal of cognitive dysfunction after T4 supplementation and restoration of euthyroidism, which confirms the clinical findings described above.

Animal models allow for the investigation of the effects of anti-Tg and anti-TPO antibodies on cognitive function in AITD, independent of TH and TSH levels. In two mouse HT models, both in NOD females immunised with Tg (50) and in CBA/J females immunised with α-enolase (51), learning and spatial memory deficits were observed, accompanied by elevated anti-Tg and anti-TPO antibody levels and euthyroidism (Supplementary Table S3).

In summary, experimental animal studies indicate the causative associations between hyperthyroidism and impaired cognition, confirming clinical observations of human patients. Moreover, in contrast to the largely inconclusive human studies on hypothyroidism, animal models clearly demonstrate that TH deficiency also exerts adverse effects on cognitive function. The reason behind these differences between animal studies and patient observations may stem from the more controlled experimental conditions in rodent models, which allow for the induction of profound and prolonged hypothyroidism. In humans, such observations are not feasible due to ethical considerations and the necessity for the prompt initiation of L-thyroxine supplementation. Moreover, based on various animal models, it can be cautiously concluded that adverse brain outcomes associated with AITD may result not only from an excess or deficiency of TH but also from altered TSH levels and immune system activation.

Psychiatric disorders in animal models

Experimental studies do not definitively confirm an association between hyperthyroidism and psychiatric disorders (Supplementary Table S4). Importantly, these observations were made only in animal models of hyperthyroidism induced by TH supplementation. Elevated TH and reduced TSH concentrations did not result in social deficits or impairments in social recognition – T3-treated male mice displayed a natural social instinct and preferred the company of other animals over solitude (35). In addition, hyperthyroidism was not associated with animal depression and did not lead to loss in the ability to experience pleasure from drinking sucrose water, anhedonia. It does, however, appear that elevated TH levels moderately increase anxiety (35, 52), but the results of the studies are inconclusive (35). Interestingly, Yu et al. (52) reported a bidirectional effect of thyroid dysfunction on anxiety- and depression-like behaviours in rats with hyperthyroidism induced by T4 administration and hypothyroidism induced by I¹³¹. TH excess exacerbated anxiety- and depressive-like behaviours, while, conversely, TH deficiency exerted anti-anxiety and antidepressant effects.

Intriguing observations have been made regarding the relationship between hypothyroid status and psychiatric disorders in experimental settings. TH deficiency induced by MMI is not associated with psychiatric disorders, including depressive and anxiety-like behaviours, in either females (48, 53, 54) or males (49, 55). However, it is noteworthy that in one study, a reduction in spontaneous rearing behaviour was observed, which may be interpreted as a depression-like state (55). In contrast, different results were observed in female NOD mice immunised with Tg (56). The activation of the immune system during euthyroidism led to elevated anxiety- and depression-like behaviours.

Experimental studies suggest that changes in hormonal levels can exert differential effects on psychiatric disorders. TH excess may exacerbate anxiety- and depression-like behaviours, although findings are not entirely consistent. Conversely, low TH levels do not substantially influence psychiatric disorders. Pre-clinical data imply that the observed increase in anxiety- and depression-like behaviours noted in patients with HT is more likely attributable to immune system activation rather than TH imbalances.

Cellular and molecular background of AITD-induced cognitive aberrances

The key molecular mechanisms by which AITD may possibly affect brain and cognitive function include those triggered directly by hormonal alterations (T3/T4, TSH, or TRH) and the presence of anti-thyroid antibodies. Analysis of animal models and in vitro studies revealed multiple TH-regulated genes involved in brain function and cognition (Table 1). However, in practice, it is challenging to separate the effects induced by TSH or TH due to hormonal interconnections induced by negative feedback regulation of the hypothalamus–pituitary–thyroid axis. In laboratory practice, selective data on the brain/neuronal TH or TSH actions can be obtained from in vitro/ex vivo studies, which usually are devoid of the physiological context provided by the brain microenvironment. Furthermore, it is difficult to differentiate between the molecular effects of TH concentration changes (reflecting the hyperthyroidism or hypothyroidism) resulting from autoimmune vs non-autoimmune thyroid disease. Therefore, it can only be assumed that the molecular mechanisms by which increased/decreased T3/T4 concentrations affect brain functioning are similar in autoimmune versus non-autoimmune thyroid disease.

Table 1.

TH-regulated genes and proteins related to brain function and cognition.

Gene/protein	Function	Ref
APP/amyloid precursor protein	Source of β-amyloid	(57)
BDNF/Brain-derived neurotrophic factor	Involved in neuronal differentiation and synaptic plasticity	(58, 59)
DAB1/DAB adaptor protein 1	Involved in neuronal migration	(60)
ELAVL4 (HuD)/ELAV like RNA binding protein 4	Regulator of neuroserpin mRNA stability	(61)
FN1 (fibronectin)	A component of extracellular matrix	(62)
GAP43/neuromodulin	Involved in synaptic plasticity and memory processes	(63)
GFAP/glial fibrillary acidic protein	Astrocyte maturation marker	(62)
IL1-R	Regulation of hippocampal microglia autophagy	(64)
Laminin	ECM component; astrocyte adhesion	(65, 66)
NR3C1 (GR)/Glucocorticoid receptor	Involved in spatial learning and memory	(63, 67)
NRGN (RC3)/neurogranin	Involved in the LTP in the hippocampus	(63, 68, 69)
RELN/reelin	ECM component; involved in neuronal migration	(61)
SERPINI1/neuroserpin	Inhibitor of tissue plasminogen activator (tPA)	(70)
SLC1A2/GLT-1	Sodium-dependent, high-affinity amino acid transporter that mediates the uptake of L-glutamate	(71)
SLC1A3/GLASt	Sodium-dependent, high-affinity amino acid transporter that mediates the uptake of L-glutamate	(71)
SYN1/synapsin I	Axonogenesis and synaptogenesis	(58)
S100B/S100beta	Astrocyte maturation marker	(62)

TH and TR actions in the brain

TH affect key cellular processes by regulating gene expression via two types of mechanisms, involving genomic or non-genomic processes. In the genomic mechanisms, T3 acts as a ligand of nuclear receptors (TRs), TRα, and TRβ that regulate transcription of target genes. Non-genomic mechanisms are activated by T3 and other iodothyronines that trigger rapid cellular effects without affecting gene expression. Intracellular T3 concentrations are independent of plasma TH levels due to the activity of iodothyronine deiodinases (Box 1). The key role in the regulation of neuronal gene expression is played by T3 taken up via the MCT8 transporter from glial cells, where it is locally synthesised due to intracellular deiodination catalysed by DIO2 (72). The genes involved in TH signalling are widely expressed throughout the brain, supporting efficient transport, local synthesis, as well as TH-mediated regulation of gene expression (Table 2).

Box 1. The key mechanisms and proteins involved in TH signalling pathways.

TH genomic mechanisms: T3 acts as a ligand of the nuclear receptors (TRs), TRα, and TRβ, encoded by THRA and THRB genes, respectively. In the brain, the TRs are mainly represented by TRα1. TRs bind thyroid hormone response element (TRE) sequences in target genes, and recruit coregulatory proteins, including histone acetylases and deacetylases, resulting in changes in chromatin condensation, thereby activating or inactivating gene expression. TRs can regulate gene expression in both ligand-dependent and ligand-independent manner. Unliganded TRs bind transcriptional co-repressors such as NCOR1 or NCOR2 (aka SMRT), leading to the decreased expression of target genes (73).

TH non-genomic mechanisms: various ligands (T4, T3, rT3) bind to different target proteins (e.g. plasma membrane integrin αVβ3 receptors, cytoplasmic TRβ isoforms, truncated TRα isoforms), inducing rapid changes in the localisation and/or activity of target proteins and triggering intracellular signalling involving MAPK or PI3K (73).

Iodothyronine deiodinases (DIOs): enzymes that catalyse deiodination of iodothyronines (74).

• DIO1: catalyses deiodination of rT3 (the preferable substrate), although it also can deiodinate T4, resulting in T2 and T3, respectively.

• DIO2 catalyses deiodination of T4, thereby contributing to the intracellular pool of T3.

• DIO3: catalyses deiodination of T3 and rT3, resulting in the synthesis of T2, thereby decreasing intracellular availability of T3.

TH transport through the blood barrier: T4/T3 transported in the blood enters the brain via specific plasma membrane transporters MCT8 and Oatp1c1 (Table 2).

Table 2.

The expression of genes related to the TH signalling pathway in the human brain.

Gene	Brain area/cell type	Reference
TSHR	The limbic system including the amygdala, cingulate gyrus, frontal cortex, hippocampus, hypothalamus, and thalamus	(75)
THRA	Detected in radial glial cells, neuronal progenitor cells, early differentiating neurons, astrocyte precursors, and choroid plexus cells, with a strong dominance of the THRA2 isoform over THRA1	(76)
THRB	Low expression overall but significant in the hippocampus, hypothalamus, and cortex. It is found in GABAergic neurons and glial cells	(77)
DIO1	Not expressed in human brain	(78)
DIO2	High expression in astrocytes and tanycytes, especially in the hypothalamus and hippocampus	(77)
DIO3	High expression in the hippocampus, cortex, and hypothalamus. Found in neurons and glial cells	(79)
MCT8	Widely expressed across many human brain regions. In the developing cortex, found in neuroepithelial cells, radial glia, intermediate progenitors, and neurons of cortical layers. In adults, expressed in striatal spiny neurons, basal ganglia interneurons, motor thalamus projection neurons, pyramidal neurons, GABAergic interneurons, and endothelial cells of the blood–brain barrier. Also present in the ventricular/subventricular zones and choroid plexus	(80, 81)
OATP1C1	In contrast to MCT8, predominantly expressed in neurons (e.g. pyramidal neurons and interneurons), OATP1C1 is more highly expressed in endothelial cells at the blood–brain barrier and in the choroid plexus	(81, 82)

T3 and the brain microenvironment

The brain microenvironment (BME) includes various cell types, such as neurons, glial cells, endothelial and immune cells, as well as non-cellular components including the extracellular matrix (ECM). The interactions between these cells are vital for maintaining cognitive functions. The specific effects of TH actions in the brain depend on the type of the targeted cell as well as the presence of microenvironment components. T3 often affects neuronal function indirectly, by modifying the functioning of the surrounding BME. The examples of such T3-BME interconnections are provided below.

T3 and neurogenesis

The hippocampal dentate gyrus, where neurogenesis continues through the lifespan, is intensively targeted by T3, as illustrated by a high density of expressed TRs. Consequently, TH are crucial for adult brain hippocampal neurogenesis, survival of progenitor cells, proliferation, maturation, growth, and migration of neurons. To some extent, these effects are mediated by T3-dependent oxygen supply. In neuroblastoma cells cultured under hypoxic conditions, T3 increases the expression of HIF-2α (hypoxia-inducible factor 2α), as well as the genes crucial for hypoxia response, including VEGF, Enolase, and c-Jun (83). In line with these data, T3 induces expression of proangiogenic growth factors Fgf2 and VEGF, as well as proliferation and tubulogenesis of rat brain-derived endothelial cells (84).

T3-treatment of astrocytes supports the generation of neurites in neurons, as shown by co-culture experiments (66). Mechanistically, T3 stimulates astrocytes to produce EGF, which contributes to neuron proliferation, neuritogenesis, and migration (85, 86, 87) (Fig. 1). Remarkably, EGF secreted by astrocytes in response to T3 treatment acts in two modes. By acting in a paracrine mode, it targets the neurons to induce their proliferation. On the other hand, the T3-induced EGF acts in an autocrine mode and stimulates astrocytes to produce laminin and fibronectin fibrils in the extracellular space, which facilitates the outgrowth of neuronal neurites (87). T3 also affects the functioning of cerebellar astrocytes by stimulating their growth, proliferation and differentiation, as well as adhesion and spreading. Some of these effects are mediated by T3-induced astrocyte autocrine activity. Specifically, T3 stimulates the secretion of FGF2, which in turn acts in an autocrine manner, stimulating the production of laminin and fibronectin, thereby regulating cellular adhesion (66) (Fig. 1). Intriguingly, the studies on TH-mediated regulation of laminin in astrocytes provided conflicting results. In the earliest study, Farwell et al. demonstrated that T4 (but not T3) increases extracellular deposition of laminin without changing its expression (65). In contrast, Trentin et al. found that T3 increased laminin expression (66), while Dezonne et al. reported that treatment of astrocytes with TH-depleted serum did not change laminin content in astrocytes (62). These differences were explained by variable experimental approaches in the three studies. The in vitro analysis of cultured mouse cerebral cortex astrocytes demonstrated that TH depletion resulted in decreased synthesis of GFAP (glial fibrillary acidic protein) and S100b (astrocyte maturation markers), as well as fibronectin (62). Altogether, these data indicate that T3 not only regulates the functioning of astrocytes but also supports their neuron–protective activity and proper transmission of glutamate-induced signalling.

Figure 1.

T3-induced effects on astrocytes and neurons. 1) T3 upregulates the expression of glutamate transporters GLASt and GLT1 on astrocytes, thereby increasing glutamate uptake and clearance from the extracellular space, minimising glutamate-induced toxicity, and increasing the survival of astrocytes and neurons. 2) T3 stimulates the secretion of FGF2, which acts on astrocytes in an autocrine manner, stimulating the secretion of laminin and fibronectin, thereby regulating adhesion. 3) T3 induces secretion of EGF by astrocytes, triggering its autocrine actions, which stimulate secretion of laminin and fibronectin into the extracellular space. This, in turn, facilitates outgrowth of neurites in neurons. 4) In response to T3, astrocytes secrete EGF, which acts in a paracrine manner to induce neuronal proliferation and migration. Red arrows: increase of a process (e.g. of expression/secretion/signalling); blue arrows: decrease of a process.

T3 and the release of neurotransmitters

Methimazole-induced hypothyroidism in rats decreases the synthesis and release of glutamate, the key hippocampal neurotransmitter involved in learning and memory processes. These effects result from reduced glutaminase activity in the hippocampus (88). On the other hand, in vitro analyses demonstrated that T3 contributes to the increased glutamate clearance from the extracellular space in the cerebellum (71). The latter, traditionally linked with motor coordination and balance, was recently connected with regulation of cognition, as well as emotional, and reward processes. Mechanistically, T3 increases the expression of glutamate receptors GLASt and GLT-1, which mediate glutamate clearance from the extracellular space (Fig. 1). This, in turn, leads to increased glutamate uptake in cerebellar astrocytes (71). By mediating glutamate clearance, T3 protects the astrocytes from cell death induced by the excess of extracellular glutamate. Indirectly, these T3 actions contribute to the viability of neurons, which are protected against glutamate toxicity when co-cultured with T3-treated astrocytes (71) (Fig. 1). The role of T3 in cerebellar functioning is underscored by studies in mice expressing mutated TRα1 devoid of efficient T3 binding. In these mice, unliganded TRα1 receptor activates pathways that lead to the delay of cerebellum development, contributing to locomotor dysfunction, as well as reduced GABAergic signalling, manifested with extreme anxiety, reduced cognition, memory and locomotor dysfunctions (89). To some extent, these murine phenotypes recapitulate features presented by patients with RTHα (resistance to TH due to THRA mutations), such as cerebellar involution, motor incoordination or reduced perceptual reasoning (90). Interestingly, about 70% of RTH children suffer from ADHD (91), while mice with mutated TRβ have phenotypes resembling ADHD (92), which further supports the crucial role of T3/TR-mediated control of gene expression in the regulation of cognition, mood, and emotions.

T3 and the synaptic transmission

Hypothyroidism affects learning and memory by impairing synaptic transmission, in particular, long-term potentiation (LTP), an activity-dependent enhancement of synaptic transmission. Hypothyroidism impairs both the early phase (E-LTP) and the late phase LTP (L-LTP). At the molecular level, E-LTP is controlled by CAMKII (calcium/calmodulin-dependent protein kinase II) phosphorylation, while L-LTP relies on CREB (cyclic-AMP response element-binding protein), which is phosphorylated by several kinases such as PKA (protein kinase A), MAPK (mitogen-activated kinase), and CAMKIV (calcium/calmodulin-dependent protein kinase type IV). In the animal model, hypothyroidism decreases the hippocampal expression of phosphorylated CAMKII, total CAMKII, neurogranin, and calmodulin. These changes are largely normalised by T4 (93). On the other hand, hypothyroidism reduces the expression of ACI (adenyl cyclase I), CaMKIV, and phosphorylated CREB, a crucial regulator in L-LTP and long-term memory. These effects are probably not mediated by TSH-cAMP-PKA-CREB signalling, since increased TSH levels in hypothyroidism should promote the activation of CREB-dependent genes and facilitate L-LTP. However, it needs to be verified if lower expression of LTP genes under hypothyroidism results from T3/TR-dependent decreased transcription.

T3 and the microglia

T3 alterations may contribute to cognitive impairments by affecting the functioning of microglia. The latter are the primary brain immune cells involved, among others, in cognitive abnormalities (94, 95). Microglial migration and phagocytosis are stimulated by T3, which is taken up through TH transporters and binds to TRα, triggering signalling which leads to changes in cell morphology (e.g. membrane ruffling). T3-induced migration involves complex signalling pathways, including NOS, Gi/o-protein, PI3K, MAPK/ERK, GABA_A and GABA_B receptors, Na+/K+-ATPase, NCX-1, and SK channels. In contrast, microglial phagocytosis does not depend on GABAA and GABAB receptors (96). T3-dependent regulation of microglia functioning highly depends on their microenvironment and T3 induces migration of brainstem microglia only in the presence of neurons. In contrast, microglia derived from the cortex respond to T3 regardless of the presence or absence of neurons (97).

Studies in rats demonstrate that hypothyroidism-induced cognitive impairment may be mediated by autophagy. MMI-induced hypothyroidism in rats leads to activation of microglia, production of IL-1a and IL-1b, and IL-1R1 expression in hippocampal neurons. This in turn triggers the downregulation of mTOR-ULK-p62 pathway and upregulation of Beclin-1, ATG5-12, ATG7, and LC3BII, synonymous with autophagy activation, with the following induction of apoptosis and loss of hippocampal neurons. All those changes culminate in impaired learning memory that could be reversed by treatment with T4, autophagy inhibitor, or IL1-R antagonist (64). However, that study did not address the question of whether the observed changes result from a direct action of T3/T4 or TSH on target hippocampal cells.

AITD-related T3 effects on neurodegeneration

Experimental studies suggest that both hypo- and hyperthyroidism may contribute to the development of AD. Induction of GD-related hyperthyroidism in the murine AD model resulted in increased neuronal Aβ deposition and inflammation, culminating in RIPK3/MLKL-induced necroptosis and neuronal loss (37). This was associated with a change of hypothalamus microglia polarisation, with M1 as a major microglia type, and inhibition of M2 polarisation. These effects resulted from direct T3 actions on microglia. Specifically, T3 decreased the elimination of Ab deposits from microglia in a RIPK3-dependent manner, while T3 treatment of microglia exposed to Aβ1-42 enhanced TNFa, NOS, and IL-1b production, synonymous with proinflammatory response. At the same time, the levels of Arginase 1, IL-4, and CD206 were reduced by T3, indicating attenuation of anti-inflammatory mechanisms. These data indicated that GD hyperthyroidism accompanying AD results in T3-induced microglia necroptosis and inhibition of Aβ phagocytosis, contributing to the progression of cognitive impairment (37).

It was suggested that the AD-promoting effects of hypothyroidism could be mediated by the T3-TTR axis (98). Plasma TTR levels are decreased in AD patients, and TTR attenuates neuronal Aβ accumulation and AD progression. It was postulated that T3 induces TTR concentrations in serum, and thus T3 decrease in hypothyroidism may lead to decreased TTR expression and abrogation of its protective effects against AD. However, T3-dependent regulation of TTR was demonstrated by only one study performed in fish (99). On the other hand, T3 and TR seem to have protective action against β-amyloid depositions. Specifically, TRα expression is reduced in the hippocampi of AD patients, while T3 attenuates the expression and processing of APP in a manner dependent on TRs that target and bind TRE sequences in the APP gene. Furthermore, double TRα/TRβ knock-out in mice increases APP expression (72). Hypothyroidism in rats leads to cognitive deficits (spatial learning, memory), associated with decreased volumes of brain and hippocampus, upregulation of Aβ levels and Tau phosphorylation, along with increased production of inflammatory cytokines. Those deficits can be ameliorated by T3 administration (72). Interestingly, AD is associated with hippocampal hypothyroidism and reduced expression of DIO2, as demonstrated by the murine AD model as well as the analysis of human AD brain organoids. This local TH deficiency induces microglia phagocytosis, leading to impaired immune microglial responses, which facilitates Aβ generation and leads to memory impairment (100).

Untreated hypothyroidism results in alterations in grey matter cerebellar morphology, possibly contributing to cognitive impairment and mood dysregulation (101). In thyroidectomised rats, these effects are mediated by increased hypothalamic endoplasmic reticulum stress (ERE) (102). ERE results from the accumulation of misfolded proteins that trigger the unfolded protein response, relying on the attenuation of general translation and increased expression of proteins facilitating elimination of the unfolded/misfolded proteins by the endoplasmic reticulum-associated protein degradation (ERAD) mechanism. Unsuccessful ERE alleviation leads to apoptosis activation and cell death. Hypothyroidism induces ERE of hippocampal cells, with the following induction of expression of the proapoptotic protein GADD153, ER-membrane-bound caspase 12, as well as the effector caspase 3, suggestive of activation of apoptosis in hippocampal neurons (102).

TSH actions in the brain

TSH binds and activates its plasma membrane-bound receptor (TSHR), which belongs to the GPCR (G-protein coupled receptors) family. Structurally, TSHR consists of seven transmembrane domains, a large extracellular domain, and a small intracellular domain. TSH binding to the TSHR triggers conformational changes that facilitate the binding of the effector molecules (G proteins, arrestins), leading to the activation of the signalling pathways involving secondary messenger molecules such as cAMP, diacylglycerol (DAG), or inositol 1,4,5-trisphosphate (IP3). Apart from TSH, TSHR can also be modulated by thyrostimulin and anti-TSHR antibodies (103).

TSHR is expressed in various brain regions, including the hippocampus and cortex (Table 2). TSH triggers TSHR-cAMP-dependent signalling in neurons and affects their functioning, including neurite outgrowth (104). Tshr deficiency in mice impairs their spatial and working memory. These effects are mediated by the changed expression of genes involved in learning and memory (including Bdnf, Egr1, and Fos), leading to the alterations of the density and structure of excitatory synapses. In line with these data, clinical studies demonstrated that low TSH levels can be linked with an increased risk of cognitive impairment (105).

Deficiency of TSH-TSHR signalling may also be associated with ADHD (attention-deficit/hyperactivity disorder) (91). TSHR KO mice are hyperactive, impulsive, aggressive, and reluctant to socialise, while their short-term memory and object recognition are impaired, which resembles the symptoms observed in ADHD patients. Some of those deficits are ameliorated by treatment with methylphenidate (MPH), the key first-line treatment of ADHD. The phenotypic effects of TSHR deficiency are probably mediated by NAT (noradrenaline transporter), which enables reuptake of norepinephrine into presynaptic nerve terminals and regulates noradrenaline homoeostasis. NAT expression is increased in TSHR KO mice, resulting in decreased noradrenaline turnover in the frontal cortex. MPH treatment inhibits NAT, thereby counteracting the expression changes induced by TSHR deficiency. Furthermore, the phenotypic effects of Tshr KO are recapitulated by treatment of mice with supraphysiological T3 concentrations, which results in decreased pituitary TSH. This indicates that the cognitive and behavioural changes result from diminished TSH-TSHR signalling, not merely the absence of TSHR. The postulated mechanisms mediating TSHR-induced NAT expression include regulation by AP-1, CREB, and NF-kB, which are activated by TSHR signalling, although these mechanisms require experimental verification (91). A more recent study revealed that TSH-TSHR-mediated regulation of learning and memory involves Wnt/β-catenin pathway. Selective deletion of Tshr in hippocampal neurons decreases the expression of Wnt5a, leading to the inactivation of β-catenin signalling pathway. These changes result in a reduced number of neural stem cells in the dentate gyrus and impaired learning and memory (36).

TRH effects on the brain

Although there is no direct evidence of causal links between AITD-related TRH changes and cognitive impairments, several studies have shown that TRH itself can affect cognition. One of the key AD hallmarks is the formation of neurofibrillary tangles and neuropil threads by phosphorylated tau protein (106). TRH-induced signalling leads to the decreased phosphorylation of tau protein in hippocampal neurons. These effects are mediated by two independent pathways involving Src and PKA kinases. Src activation leads to the inhibition of GSK3β, the key kinase responsible for tau phosphorylation. In contrast, PKA can lead to the attenuation of tau phosphorylation (107).

Anti-thyroid antibodies: impact on the brain function

The studies on the influence of anti-thyroid antibodies on the CNS are limited. TSHR expressed in limbic neurons may be targeted by autoantibodies, leading to gradual TSHR loss, resulting in diminished TSH-TSHR signalling in neurons and mood dysregulation (108). This finding is supported by decreased TSHR expression in bipolar limbic neurons compared with normal human brains (109). Interestingly, expression of Tg is similarly decreased in the brains of patients with bipolar disease (109). However, the latter study did not confirm whether bipolar patients suffered from AITD; therefore, the causal links between AITD and TSHR/Tg loss in the brain remain to be confirmed.

In contrast, observations from a mouse model show that GD is associated with an increased TSHR expression in vascular endothelial cells of the medulla oblongata. Furthermore, the same study demonstrated that anti-TSHR antibodies stimulate TSHR expression and cAMP production in endothelial cells in vitro (110). However, the exact mechanisms by which anti-TSHR antibodies stimulate TSHR expression remain unknown.

A mouse HT model revealed the presence of anti-TPO and anti-Tg antibodies in the frontal cortex (56). Interestingly, one study reported the ability of anti-TPO antibodies to bind cerebellar cells expressing glial fibrillary acidic protein (106). Such binding was observed for monoclonal anti-TPO antibodies or antibodies derived from the CSF of patients diagnosed with Hashimoto’s encephalopathy, a controversial disease entity, described as ‘corticosteroid-responsive neurological syndrome occurring in patients with AITD’ (106, 111). No binding was observed for anti-TPO antibodies derived from patients with Hashimoto’s thyroiditis (106); therefore, it needs to be verified if antibodies present in AITD patients may indeed affect CNS cells and influence the functioning of the brain.

Conclusion and perspectives

Clinical and pre-clinical studies suggest that AITD can lead to impairments of cognition, memory, or mood. Most reports indicate that hyperthyroidism is associated with the risk of cognitive impairments, while reports on hypothyroidism provide variable data. Furthermore, clinical studies clearly indicate the associations between hyperthyroidism (both overt and subclinical) and increased risk of dementia. Levothyroxine is the fourth top prescribed drug in the U.S. and the third in the UK (112, 113), which exemplifies the potential impact of thyroid disease on public mental health. Despite multiple reports that evaluated the links between thyroid disease and brain function, the studies focussing on the relationship between AITD and cognition are still limited. The vast majority of clinical studies analysed the associations between cognition and hypo-/hyperthyroidism, without deeper evaluation of its autoimmune nature. Therefore, comprehensive clinical studies involving large, well-characterised cohorts of patients are needed to clarify the relationships between AITD and cognitive impairments and psychiatric disorders. Such studies should be carefully designed to ensure patients are thoroughly characterised not only by TH/TSH concentrations, but also by age, gender, BMI, the presence of anti-thyroid antibodies, and co-morbidities.

Due to manuscript length limitations, we were unable to discuss many detailed mechanisms that may also influence cognitive dysfunction and psychiatric disorders associated with AITD. The most crucial omitted topics include the potential impact of obesity and microbiota. Hypothyroidism is often associated with weight gain, while fat tissue is an important source of inflammatory cytokines that can affect brain function. The gut microbiota is altered in AITD patients (114), which can contribute to cognitive and mood impairments through the gut–microbiota–brain axis. Obesity alters gut microbiota (115), which further adds to the complexity of potential interrelations between AITD, gut microbiota, and brain function.

Data from clinical, as well as in vivo, ex vivo, and in vitro studies suggest that changes in concentrations of T3 and TSH, and possibly also of TRH, as well as the presence of anti-thyroid antibodies, can trigger signalling pathways in different brain regions, resulting in cognitive impairments. These effects highly depend on the brain microenvironment context, as illustrated by T3 effects on microglia migration. This indicates that basic research aimed at understanding the molecular mechanisms underlying cognitive disorders in AITD should be based on multicellular (co-culture) models and/or organoids. The implementation of novel high-throughput molecular techniques, such as single-cell and spatial transcriptomics, may help in deciphering the complex relationship between AITD and brain function.

Declaration of interest

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the work reported.

Funding

The study was supported by Centre of Postgraduate Medical Education grant 501-1-025-01-24.