This review presents the findings that led to the discovery of TRH and the understanding of the central mechanisms that control hypothalamus–pituitary–thyroid axis (HPT) activity. The earliest studies on thyroid physiology are now dated a century ago when basal metabolic rate was associated with thyroid status. It took over 50 years to identify the key elements involved in the HPT axis. Thyroid hormones (TH: T₄and T₃) were characterized first, followed by the semi-purification of TSH whose later characterization paralleled that of TRH. Studies on the effects of TH became possible with the availability of synthetic hormones. DNA recombinant techniques permitted the identification of all the elements involved in the HPT axis, including their mode of regulation. Hypophysiotropic TRH neurons, which control the pituitary–thyroid axis, were identified among other hypothalamic neurons which express TRH. Three different deiodinases were recognized in various tissues, as well as their involvement in cell-specific modulation of T₃concentration. The role of tanycytes in setting TRH levels due to the activity of deiodinase type 2 and the TRH-degrading ectoenzyme was unraveled. TH-feedback effects occur at different levels, including TRH and TSH synthesis and release, deiodinase activity, pituitary TRH-receptor and TRH degradation. The activity of TRH neurons is regulated by nutritional status through neurons of the arcuate nucleus, which sense metabolic signals such as circulating leptin levels.Trhexpression and the HPT axis are activated by energy demanding situations, such as cold and exercise, whereas it is inhibited by negative energy balance situations such as fasting, inflammation or chronic stress. New approaches are being used to understand the activity of TRHergic neurons within metabolic circuits.

Energy homeostasis relies on a concerted response of the nervous and endocrine systems to signals evoked by intake, storage, and expenditure of fuels. Glucocorticoids (GCs) and thyroid hormones are involved in meeting immediate energy demands, thus placing the hypothalamo–pituitary–thyroid (HPT) and hypothalamo–pituitary–adrenal axes at a central interface. This review describes the mode of regulation of hypophysiotropic TRHergic neurons and the evidence supporting the concept that they act as metabolic integrators. Emphasis has been be placed on i) the effects of GCs on the modulation of transcription ofTrhin vivoandin vitro, ii) the physiological and molecular mechanisms by which acute or chronic situations of stress and energy demands affect the activity of TRHergic neurons and the HPT axis, and iii) the less explored role of non-hypophysiotropic hypothalamic TRH neurons. The partial evidence gathered so far is indicative of a contrasting involvement of distinct TRH cell types, manifested through variability in cellular phenotype and physiology, including rapid responses to energy demands for thermogenesis or physical activity and nutritional status that may be modified according to stress history.

Central and peripheral mechanisms that modulate energy intake, partition and expenditure determine energy homeostasis. Thyroid hormones (TH) regulate energy expenditure through the control of basal metabolic rate and thermogenesis; they also modulate food intake. TH concentrations are regulated by the hypothalamus-pituitary-thyroid (HPT) axis, and by transport and metabolism in blood and target tissues. In mammals, hypophysiotropic thyrotropin-releasing hormone (TRH) neurons of the paraventricular nucleus of the hypothalamus integrate energy-related information. They project to the external zone of the median eminence (ME), a brain circumventricular organ rich in neuron terminal varicosities and buttons, tanycytes, other glial cells and capillaries. These capillary vessels form a portal system that links the base of the hypothalamus with the anterior pituitary. Tanycytes of the medio-basal hypothalamus express a repertoire of proteins involved in transport, sensing, and metabolism of TH; among them is type 2 deiodinase, a source of 3,3',5-triiodo-L-thyronine necessary for negative feedback on TRH neurons. Tanycytes subtypes are distinguished by position and phenotype. The end-feet of β2-tanycytes intermingle with TRH varicosities and terminals in the external layer of the ME and terminate close to the ME capillaries. Besides type 2 deiodinase, β2-tanycytes express the TRH-degrading ectoenzyme (TRH-DE); this enzyme likely controls the amount of TRH entering portal vessels. TRH-DE is rapidly upregulated by TH, contributing to TH negative feedback on HPT axis. Alterations in energy balance also regulate the expression and activity of TRH-DE in the ME, making β2-tanycytes a hub for energy-related regulation of HPT axis activity. β2-tanycytes also express TRH-R1, which mediates positive effects of TRH on TRH-DE activity and the size of β2-tanycyte end-feet contacts with the basal lamina adjacent to ME capillaries. These end-feet associations with ME capillaries, and TRH-DE activity, appear to coordinately control HPT axis activity. Thus, down-stream of neuronal control of TRH release by action potentials arrival in the external layer of the median eminence, imbricated intercellular processes may coordinate the flux of TRH into the portal capillaries. In conclusion, β2-tanycytes appear as a critical cellular element for the somatic and post-secretory control of TRH flux into portal vessels, and HPT axis regulation in mammals.

The hypothalamic-pituitary thyroid (HPT) axis modulates energy homeostasis. Its activity decreases in conditions of negative energy balance but the effects of chronic exercise on the axis are controversial and unknown at hypothalamic level. Wistar male rats were exposed for up to 14 days to voluntary wheel running (WR), or pair-feeding (PF; 18% food restriction), or to repeated restraint (RR), a mild stressor. WR and RR diminished food intake; body weight gain decreased in the 3 experimental groups, but WAT mass and serum leptin more intensely in the WR group. WR, but not RR, produced a delayed inhibition of central markers of HPT axis activity. At day 14, in WR rats paraventricular nucleus-pro-TRH mRNA and serum TSH levels decreased, anterior pituitary TRH-receptor 1 mRNA levels increased, but serum thyroid hormone levels were unaltered, which is consistent with decreased secretion of TRH and clearance of thyroid hormones. A similar pattern was observed if WR animals were euthanized during their activity phase. In contrast, in PF animals the profound drop of HPT axis activity included decreased serum T3 levels and hepatic deiodinase 1 activity; these changes were correlated with an intense increase in serum corticosterone levels. WR effects on HPT axis were not associated with changes in the activity of the hypothalamic-pituitary adrenal axis, but correlated positively with serum leptin levels. These data demonstrate that voluntary WR adapts the status of the HPT axis, through pathways that are distinct from those observed during food restriction or repeated stress.