Seven reports were included in this classification. Six studies contained data about TRH expression in the Hyp or limbic areas and behavioral tests such as the elevated plus-maze (EPM), the defensive burying test (DBT), the open field test (OFT), and the novelty-induced hypophagia test (NIHT), to analyze anxiety-like behavior in rodents (rats or mice). One report describes TRH content changes during the estrous cycle of female rats, given that estradiol is a neuroprotective hormone and its fluctuations have been associated with anxiety disorders. It shows that proestrus promotes a decrease in TRH content in Hyp, Amy, NAc and Hipp [29]. Three reports that use growth hormone-releasing hormone (GHRH) knockout (KO) mice [30], and Wistar or Wistar Kyoto (WKY) rats (strain with increased depression- and anxiety-like behaviors) [6, 31] find an inverse correlation between TRH mRNA expression either in the Hyp or in the Amy with anxiety-like behaviors. GHRH KO mice exhibit reduced anxiety in the OFT and the EPM and increased TRH mRNA expression in the Hyp [30]; whereas when rats are subjected to the DBT, increases in TRH content and decreases on its mRNA levels after 30 or 60 min of performing the test are observed in Amy of Wistar and WKY rats [6, 31]. These results support the anxiolytic role of TRH. The other three reports fail to find a clear association between changes in the TRH system and anxiety. In two different studies that use KO mice for TRH-R1 and TRH-R2 [32, 33], anxiety-like behavior is increased in TRH-R1 KO mice when performing the EPM and decreased in TRH-R2 KO female mice after subjecting them to NIHT. In a model of anorexia in rats [34], TRH mRNA expression increases in the PVN, but anxiety-like behavior increases in such female rats subjected to the DBT.

We included three studies where TRH was intracerebroventricularly (i.c.v.) or intraperitoneally (i.p.) administered, and the DBT and the EPM test anxiety-like behavior. One of those studies was also included in classification 1 [6]. According to these reports, there is controversy regarding TRH’s effects on anxiety. I.c.v. TRH administration to Wistar rats reduces anxiety in the DBT and decreases corticosterone serum levels [6].

On the other hand, in another report where TRH is i.c.v. Injected to Holtzman rats shows no effects on anxiety-like behavior analyzed with the EPM [35]. Likewise, an i.p. administration of TRH does not affect anxiety in rats subjected to the EPM [36]. More efforts are needed to discriminate TRH effects of different limbic regions on anxiety behavior.

Seven studies were included where the administration of different drugs or hormones alters TRH expression or content in several brain regions. Two studies evaluate TRH changes after administering anxiolytic drugs ketamine or prazosin, glutamatergic or adrenergic neurotransmission inhibitors, respectively. The i.p. injection of ketamine induces an increase in the content of TRH in the Hyp, NAc, PFC and Amy of Sprague-Dawley rats [37], and the administration of prazosin enhances TRH content in the Amy, NAc and striatum of Sprague-Dawley and Lewis rats, a strain with high-extreme behavioral responses [38]. One report evaluates the inhibition of TRH binding to its receptor by benzodiazepines (anxiolytic drugs) and finds that in Amy, its receptor (not identified as its subtype at the time) is more sensitive to the inhibitory actions of these drugs [39]. This evidence suggests that TRH could be mediating the effects of anxiolytic drugs. Two articles report changes in the early stages of life (gestation or neonatal) on the TRH system and anxiety-like behavior displayed by rodents. Neonatal serotonin depletion by administering para-chlorophenylalanine (pCPA) induces a reduction in anxiety in the OFT, but TRH mRNA expression in the PVN does not change [40]. On the other hand, oral administration of thyroxine (T${}_4$) to WKY dams provokes an increase in TRH content in the Amy of adult male offspring. Still, it increases their anxiety-like behavior in the DBT [41].

The administration of drugs that directly affect TRH transcription has also been evaluated. The role of TRH from the NAc in anxiety has been supported by the reduced anxiety-like behavior in the DBT, which is observed along with an upregulation of TRH expression after an intra-accumbal administration of an inhibitor of the enzyme phosphodiesterase 7 (PDE7B) in ad libitum fed rats. This approach maintains elevated phosphorylated CREB (pCREB) levels, enhancing TRH transcription. When an antisense oligonucleotide (ODN) is injected before administrating the PDE7B inhibitor, a decrease in accumbal TRH mRNA levels develops, and anxiety returns to the level of a vehicle-treated group. It is noteworthy that the i.p. injection of PDE7B inhibitor affects feeding and anxiety depending on the dose administered: while a lower concentration (0.2 mg) of PDE7 inhibitor increases the expression of TRH in the PVN and decreases feeding (without affecting anxiety-like behavior), a higher dose enhanced accumbal TRH mRNA and induces anxiolysis, with no changes in food intake [42]. Lastly, i.p., corticosterone administration induces TRH content in Hipp and Amy’s striatum [43]. Overall, we can conclude that TRH from limbic regions, especially from Amy and NAc, regulates mood and modulates anxiolytic drugs’ actions.

Only one report found that TRH content in the cerebrospinal fluid (CSF) was correlated to anxiety in diagnosed patients [44]. Although no statistical differences are found, a weak trend of a lower CSF TRH content in anxious patients is observed. This could be accounted for the small number of patients included in the report (45 patients). The lack of research of TRH association to anxiety in clinical studies shows that this field could be exploited further.

We found sixteen articles that analyze the participation of the peptide in the display of depression-like behaviors by using various models. Some of the most recent articles show that TRH may be a factor that facilitates depression-like behaviors in rats since TRH mRNA content and that of its receptor TRH-R1, which are both highly expressed in the cortico-limbic pathway, result upregulated in the basolateral Amy (BLA) of animals with chronic stress (2 h of restraint for 14 days). This TRH upregulation is observed along with high scores of depressive parameters when performing the forced-swim test (FST) or tail-suspension test (TST); in an elegant experiment, researchers show that siRNAs injection directed against TRH or TRH-R1 in BLA decrease depression-like conducts [45].

Also, TRH content and expression increase in the BLA of chronically stressed animals, which also present an up-regulation of interleukin-18 (pro-depressant factor) in this area and elevated immobility time when performing the FST and the TST. Interestingly, interleukin-18 KO mice and rodents with interleukin-18 silenced expression show low depressive parameters and fail to present the elevated TRH mRNA levels in BLA of wild-type stressed animals. Therefore, the amygdalar TRHergic pathway seems to be associated with high levels of depression and downstream of interleukin-18 signaling inducing depression-like behaviors in chronically stressed animals [46].

In WKY rats that develop a hypothyroideal condition and high immobility scores in the FST, the prepro-TRH protein levels of Amy increase in gestational T${}_4$-treated offspring (an anti-depressant agent). However, their immobility scores do not change, and results are not supportive of an association between higher amygdalar expression of TRH and depression-like behavior [41]. Therefore, TRH anti-depressant effects are not always supported

Another approach is using acute or chronic corticosterone administration to rats, given its involvement in displaying depression in humans and animals. A single i.p. injection of a high dose of corticosterone elevates TRH content in the Hipp, Amy, entorhinal cortex and striatum of rats after 4 or 8 h of being injected [43]. In contrast, male rats ingesting the steroid in drinking water for 20 days, a condition that induces depression-like behaviors, show decreased TRH content in the Amy or PFC, depending on low or high corticosterone dose, respectively [43, 47].

GHRH KO mice is a useful model to analyze anti-depressant roles of peptides as TRH since those animals present less immobility time in the FST and TST along with high hypothalamic TRH mRNA levels [30]. Also, KO animals for TRH-R1 [32] or TRH-R2 [33] support the role of limbic TRH in mood regulation since both transgenic mice display depressive-like behaviors.

Female rats in proestrus and estrus present low depression parameters due to the characteristic high estrogen serum levels observed in those phases, which are implicated with anti-depressant actions. In those cycle phases, females show decreased content of TRH in various regions of the limbic system, as if the peptide is being released and thus implicated in estrogen-induced depression regulation [29].

The utilization of electroconvulsive shocks in rats as anti-depressant treatment results in elevated TRH mRNA levels in the granular cells of the hippocampal dentate gyrus analyzed by microarrays [48]; in addition, with the same approach, others find high peptide levels in limbic regions, including the Amy and the pyriform cortex besides the Hipp [49, 50].

Olfactory bulbectomy (OBX) is considered a model of depression and is commonly used in rats to evaluate the efficacy of different anti-depressant drugs. Controversies are found regarding TRH changes in these animals: OBX increases hypothalamic TRH content, and when treated with sertraline, a serotonin reuptake inhibitor, such change is reversed, and their locomotor activity is reduced [51]. In contrast, other studies that also use OBX rats find no changes in brain TRHmRNA and no association with depressive-like behaviors [52, 53].

We only found one in vitro report showing that lithium (anti-depressant) treatment in cultured GH3 cells can interfere with the TRH-induced increase in inositol 1,4,5 triphosphate (IP3), which argues against the anti-depressant effect of the peptide [54].

Only two articles use a different approach to elucidate the role of TRH in mood regulation, which is the injection of the peptide or some of its analogs, allowing the observation of depressive-behavioral changes in rodents. By injecting the TRH analog taltirelin in TRH-R1, TRH-R2 or both KO mice and evaluating their immobility time in the TST, the results point out to TRH-R1 activation as needed for TRH effects as anti-depressant [55].

Choi et al. (2015) [45] (already mentioned in the 3.2.1. section) show that the injection of TRH or taltirelin in the BLA induces depression-like behaviors in animals performing the FST, TST, U-field sociability and U-field tests.

An i.p. injection of TRH in rats induces an increased release but not the turnover of noradrenaline in hypothalamic cultured tissue, which supports the anti-depressant role of TRH [56].

We found seven articles that investigated the effect of administering different drugs known as anti-depressant factors such as ketamine, prazosin, glycogen synthase kinase 3$\beta$ (GSK-3$\beta$) and escitalopram. They find that TRH content is enhanced in brain regions involved in depression regulation evaluated by high-performance liquid chromatography (HPLC), which supports its participation in the effects of the NT pathways altered by those anti-depressants. Ketamine (an NMDA receptor antagonist) increases TRH in Amy, Hyp, PFC and NAc [37]; prazosin ($\alpha$-adrenergic antagonist) augments TRH content also in Amy, striatum and entorhinal cortex [38]; GSK-3$\beta$ elevates TRH concentration in Amy, striatum and piriform cortex among others [57], whereas escitalopram, a serotonin reuptake inhibitor, increases TRH levels in NAc and striatum [58].

TH administration, which also participates in displaying depression symptoms, induces an increment in brain TRH [59]. As TH can only affect TRH transcription specifically in the hypophysiotopic neurons of the PVN, it is interesting that by stimulating a non-genomic mechanism, TRH levels also change in Amy, pyriform cortex and striatum. With acute and chronic valproate administration or after its withdrawal in rats, TRH content increases in the pyriform cortex, Amy and Hipp, the expression levels of TRH receptors in NAc and PFC, and the receptor binding in Amy and PFC [60]. Acute or chronic consumption of lithium in the diet also changes limbic TRH content, decreasing it in the entorhinal cortex but increasing peptide levels in the pyriform cortex and striatum when lithium withdrawal is induced [61].

We selected ten clinical studies, and some suggest that TRH ameliorates depressive symptoms. Postmortem analyses showed decreases in TRH mRNA levels in the PVN when patients had been chronically treated with corticosterone, which, if elevated in serum in a sustained manner, maybe underlying the development of depression [62]. A similarly reduced gene expression of TRH is observed in PVN of patients diagnosed with major depression [63]. Others have not found changes in limbic TRH since postmortem studies of depressed patients show similar peptide levels in the Amy of males and females and similar to those of the control group [64].

When the analysis is performed in the CSF of depressed untreated patients, TRH levels decrease [65] and even older [66]. In contrast, Frye et al. [67] find no correlation between CSF TRH levels and depression but show gender differences since TRH levels are lower in females than males. Importantly, the difference is greater when patients are depressed [67].

Intra-thecal TRH injection in patients with major refractory depression ameliorates their symptoms, reduces their scores in the Hamilton scale in 50% of the sample as well as their suicidal attempts [4]; i.v. administration of TRH during the night is also successful in decreasing depressive symptoms in patients [68]. When the administration of TRH is oral or i.v. and intermittent for 14 days to depressed patients, their encephaloelectrographic pattern is altered along with enhanced interest, desire, and drive for work, food, and sex, which are symptoms of a mitigated depression [69].

We found five papers analyzing the role of TRH in models of fasting, food restriction and anorexia, three papers delve into the role of TRH in feeding during negative energy balance, and four studies focus on the role of TRH in different approaches that modify feeding such as, genetic or disease models, and by the modulation of the neuronal activity of specific brain regions.

The initial studies of TRH involvement in feeding include research done in the 1980s reporting an increase of a hypothalamic TRH metabolite (cyclo-His-Pro) after fasting, while no changes in TRH levels from this region are observed [70]; starvation causes a decrease in TRH release from the Hyp into the hypophyseal-portal blood [71]. A more recent finding shows that the reduction in PVN TRH mRNA levels after fasting remains low 4 and 12 h after refeeding [72]. A report where both food-restriction and a model of anorexia were analyzed shows the expected decrease in TRH mRNA levels in the PVN due to negative energy balance in food-restricted animals; nevertheless, anorectic rats do not display such adaptive response, and in turn, show an increase in TRH mRNA levels [34]. An additional report reveals that TRH expressed specifically in the anterior subdivision of the PVN possesses the specific role of inhibiting appetite, even in starved rodents in the model of dehydration-induced anorexia [73].

One report [74] shows hibernating ground squirrels have alterations in their hypothalamic transcriptome that prepare them for the winter. One of these alterations is a 55% decrease in their food intake before hibernation, associated with TRH mRNA upregulation [74].

During fasting, the highly abundant and proximity of NPYergic tonic inhibition to TRHergic neurons in the PVN decreases energy expenditure during negative energy balance [75]. Unexpectedly, an excitatory input to orexigenic AgRP neurons arising from a subset of TRHergic cells that coexpress the pituitary adenylate cyclase-activating polypeptide (PACAP) is necessary and sufficient to drive hunger [76].

On the other hand, the spontaneous hypophagia present in Lou/C rats, a strain resistant to diet-induced obesity, is due to high leptin sensitivity, which induces the overexpression of hypothalamic brain-derived neurotrophic factor (BDNF) and TRH [77].

Excitotoxic lesions of the dorsomedial hypothalamus (DMH) uncover another possible role of TRH in feeding, specifically in the circadian aspect of this behavior. Such lesions disrupt the animals’ circadian food intake distribution, possibly mediated by TRH-expressing neurons that project to the lateral hypothalamus, impinging on orexinergic neurons [78]. Recently, with new technologies of neuronal labeling, a report shows that activation of DMH afferents to PVN TRH receptor-expressing neurons, coexpressing other feeding-regulatory peptides such as cholecystokinin (CCK), prodynorphin and gastrin-releasing peptide, inhibits food intake [79].

In a model of diabetic rats, TRH mRNA expression is downregulated in the PVN correlating with hyperglycemia. This change, along with increases in ARC orexigenic neuropeptides that directly decrease TRH expression levels, is responsible for the hyperphagia observed in animals subjected to this model [80].

Mice with a mutation resulting in the loss of function of the glucose transporter 2 (GLUT2) as a receptor for extracellular glucose have alterations in their feeding patterns, characterized by an increase in their meal size. This is related to an altered neuropeptidergic expression in the Hyp, including downregulation of TRH mRNA, whose synthesis is dependent on the GLUT2 detection function, favoring actions of orexigenic peptides and contributing to defective control of food intake in this mouse strain [81].

Finally, we found one in vitro report that provides further support of the role of TRH in feeding: administration of TRH and one of its agonists hinders MCH neuronal activation through an increase in inhibitory GABAergic neurotransmission in the brain slices [82].

Eleven studies investigate the feeding effects of the peripheral or central administration of TRH or its metabolites, supporting its anorectic role even during starvation; moreover, TRH injection directly into the Hyp of rats causes anorexia and sometimes adipsia [7, 83, 84, 85, 86, 87, 88, 89, 90]. Chronic peripheral injection of TRH in obese rats causes an increase in their resting metabolic rate and a decrease in their food intake. Thus weight gain is reduced [91]. TRH administered i.c.v in rodents reverses the feeding-promoting effects of drugs such as diazepam [92].

The anorectic effect of an i.c.v. Administration of TRH is partially reversed in histamine depleted rats, and central injection of TRH increases the turnover of the NT. Histamine-synthetizing neurons express TRH-R2, and thus authors propose a hypothalamic histaminergic mediation of the anorectic actions of TRH [93]. By using a histamine H1 receptor KO mice followed by anti-TRH antibody administration, the anorectic effects of the peptide nesfatin 1 seem to be mediated by histamine and TRH [94].

Recent work has expanded the knowledge of the TRH system beyond the hypothalamic role of energy homeostasis regulation. Through intracisternal or lateral ventricle administration of a TRH analog, the role of brainstem TRH has been described in more detail. Intracisternal TRH administration increases food intake two to three-fold in ad libitum fed rats, while the changes after lateral ventricle administration are insignificant. In contrast to what is reported in Hyp, after fasting, rats display an enhanced TRH and TRH-R1 mRNA expression in the brainstem. Brainstem TRH can stimulate feeding through a vagal-ghrelin pathway and interact with NPY-Y1 receptors [95].

Moreover, the orexigenic response after an intracisternal administration of a TRH analog is reduced in Goto-Kakizaki rats, a model of type 2 diabetes and with an impaired central vagal function. Natural food intake in these rats is also diminished, supporting the role of altered brainstem TRH in preventing further weight gain in obesity and type 2 diabetes [96].

Finally, only two studies related to accumbal TRH participation in feeding. The direct injection of the peptide into the NAc shell decreases feeding motivation by augmenting DA release and turnover in food-restricted rats [8]. In contrast, accumbal injections of TRH can reduce palatable food intake in stressed animals while increasing serum glucocorticoid (GC) levels when GC signaling is blocked in the NAc of stressed animals, palatable food intake decreases and TRH expression increases. Overall, these results suggest that accumbal TRH is downstream of GCs activity and its effects on palatable food intake [97].

We found 11 studies related to the administration of hormones that regulate energy homeostasis, signaling mediators or receptor antagonists that modulate TRH levels of expression in different brain regions with consequent changes in feeding.

Altering TRHergic signaling through an intracisternal administration of an antisense ODN directed against TRH-R1, the cephalic phase of acid secretion decreases during sham-feeding. Authors propose that cholinergic activation is mediated by medullary TRH [98].

As mentioned in TRH and anxiety section, the i.p. injection of PDE7B inhibitor affects feeding and anxiety depending on the administered doses. While a lower concentration (0.2 mg) of PDE7B inhibitor increases PVN TRH expression and decreases feeding (without affecting anxiety-like behavior), a higher dose enhances accumbal TRH mRNA. It induces anxiolysis, with no changes in food intake [42].

Feeding related hormones such as ghrelin or leptin, when i.p. administered, have opposing effects on TRH levels in several brain regions, provoking an increase and decrease, respectively [57, 99]. Conversely, intra-VMH leptin administration increased TRH mRNA expression and other anorectic peptides in the arcuate-median eminence complex; this correlates with decreased food intake and is supposed to be mediated through a different pathway than when leptin is i.c.v. Administered [100]. I.p. co-administration of CCK and leptin can reduce food intake in rats, which is reversed by the i.c.v. Injection of antibodies against TRH and cocaine- and amphetamine-regulated transcript (CART) [101].

Secretin administration to mice inhibits food intake while increasing the expression of TRH and the melanocortin four receptor (MC4R) in the PVN, supporting the interaction between TRHergic and melanocortinergic systems [102]. Moreover, suppose TRH expression is blocked using an antisense ODN. In that case, the anorectic effect of $\alpha$-melanocyte-stimulating hormone ($\alpha$-MSH) in the NAc is blocked in fasted rats [103].

Two papers have analyzed early-life exposure to some chemicals in TRH and feeding in rats: fetal exposure to GCs causes a decrease in TRH immunoreactivity in the yuxtaparaventricular region of the lateral hypothalamus, which is associated with increased food intake in adult offspring from both sexes [104]. Also, exposure to tobacco smoke during lactation provokes a decrease in TRH content in the PVN, related to hyperphagia during adulthood [105].

The few studies evaluating TRH involvement in eating disorders only report an impaired response of TSH in the TRH-stimulation test and thus were omitted from the present review. Analysis of genetic variants in extremely obese patients revealed six novel and rare missense variants, including one in the hypothalamic TRH gene, which may probably damage the protein function [106].