The possibility that tanycytes are involved in hypothalamic functions was the basis for investigations of these cells by a number of investigators in the 1960s and 70s; these cells exhibited alterations of morphology or chemical staining as a result of the photoperiod, stress and sexual activity . Early work on tanycytes described them as a group of cells lining the lower third of the dorsal wall of the third ventricle and the floor of the infundibular recess [36, 37]. Cytological analysis showed morphology of tanycytes as nonciliated cells with long basal process extending to the basal lamina, terminating as endfeet that were in contact with fenestrated vessels of the portal plexus in the median eminence [38, 39] and, more recently, with fenestrated blood vessels that infiltrate the ventromedial arcuate nucleus (ARC) [40, 41].
In the seasonal context, plasticity has been demonstrated in two LD-breeding animals, the Siberian hamster and the Japanese quail. In the quail, tanycytic endfeet contact a larger surface area of the basal lamina in the external zone of the median eminence during SD exposure compared with LD exposure . This conforms to a role for tanycytes in controlling the seasonal release of GnRH from axon terminals. However, in Siberian hamsters held in constant darkness for 1 month or exposed to SD for a period of 2 months, the reverse situation occurs with substantially reduced innervation and fewer contacts between tanycytes and axon terminals . As a result, this does not support the notion for a function of tanycyte endfeet to solely impede release of GnRH. Thus, the response of tanycytic endfeet in the Siberian hamster would be expected to enhance GnRH release and promote reproductive physiology and behaviour. Therefore it is likely that the relationship between endfeet and release of hypophysiotrophic hormones is not a simple one and interactions between tanycyte endfeet and axon terminals containing other hypophysiotrophic hormones may be involved and have a bearing on the seasonal hypothalamic-pituitary axis.
Recently, Bolborea et al.  confirmed the observation of Kameda et al.  of SD-induced downregulation of the intermediate filament, vimentin, which may be involved in the alteration of tanycyte morphology. In addition, gene expression for neural cell adhesion molecule was also found to be downregulated in SD adding further support to structural plasticity of tanycytes with altered photoperiod. Importantly, both of these gene expression changes were shown to be effected by melatonin and independent of seasonal changes in sex steroids .
In addition to alterations in tanycyte morphology, many gene expression alterations have now been found to occur which are likely to contribute to the role tanycytes play in seasonal adaptations to physiology. Changes in expression of components of the thyroid hormone system are described later, but other changes include increase in expression of genes for glycogen and glucose metabolism, glutamine synthesis, lactate and glutamate transport . These changes are proposed to have an involvement in modulating glutamate and neurotransmitter availability in the hypothalamic environment .
Tanycytes – a source of hypothalamic thyroid hormone
Studies in birds established an important role for thyroid hormones (T3 and T4) in seasonal reproductive rhythms [47–50], but only recently has the molecular basis for thyroid hormone action and a tangible connection between the PT and central mechanisms regulating seasonal physiology been established. Yoshimura et al.  using a subtractive hybridization approach discovered higher levels of type II deiodinase (Dio2) responsible for converting inactive T4 to active T3 (Fig. 2) in the hypothalamus of the Japanese quail in LD-exposed birds compared with those in SD . In situ hybridization studies revealed the Dio2 expression occurred in tanycytes, suggesting tanycytes were the source of thyroid hormone. Subsequent studies found type III (Dio3) responsible for degrading T4 and T3 to inactive metabolites was expressed and reciprocally regulated by photoperiod in tanycytes of the quail , providing the basis for the measured lower T3 values in the hypothalamus of the quail in SD. Measurements of T3 and T4 in circulation showed no change between LD and SD photoperiod, providing evidence that local synthesis by tanycytes is the principal contributor to photoperiod-dependent hypothalamic T3 availability .
Figure 2. A schematic representation of the relationship between the pars tuberalis (PT), the site of melatonin action, tanycytes, the intermediary relay station and neurons of the hypothalamus which govern centrally mediated mechanisms underlying body weight and reproduction. TSH produced and secreted by melatonin responsive cells of the PT gains access to tanycytes of the third ventricle. This may be via cisterns, bathed in cerebrospinal fluid which connect directly with the third ventricle; direct communication via tanycyte endfeet abutting PT cells or retrograde transport via the capillary network system that infiltrates the medial basal hypothalamus. The boxed diagram shows a schematic relationship between TSH action on tanycytes and deiodinase expression. TSH stimulates expression of type 2 deiodinase (Dio2) for T3 production and availability to the hypothalamus where this has consequences for neuronal interactions and expression of key components leading to the establishment of long-day physiology. A decline in short days of TSH originating from the PT leads to a reduction in Dio2 expression. This may account for some or all of the reduction in T3 availability to the hypothalamus (e.g. in the Syrian hamster). However, in some species (e.g. the Siberian hamster), a concurrent expression of type III (Dio3) may occur for the metabolism of T3 and T4 which will limit the availability of bioactive T3 to the hypothalamus in short days. It is not known whether the mechanism of induction of Dio3 expression involves a stimulatory receptor-mediated input (?) or whether it occurs by another mechanism as a result of a reduction in TSH input.
Download figure to PowerPoint
Thyroidectomy and T4 replacement studies in sheep [53, 54] had previously demonstrated a central action of this thyroid hormone precursor for the spring termination of the breeding season in sheep, providing additional evidence for the involvement of the thyroid hormone system in the regulation of the seasonal reproductive physiology [55, 56]. Subsequently, led by the discoveries in the quail, regulation of Dio2 and Dio3 was found to occur in tanycytes of seasonal mammalian species, albeit with some variation in expression patterns. In the Syrian hamster (Mesocricetus auratus), for example, Dio2 is downregulated in SD and Dio3 does not appear to be expressed [57, 58]. In the Siberian hamster (Phodopus sungorus), Dio2 shows a partial downregulation by 8 wk of SD exposure, but then returns to full expression by 14 wk. Dio3, on the other hand, demonstrates interesting temporal kinetics, increasing in expression during the course of first 8 wk in SD, but declining thereafter . In the European hamster (Cricetus cricetus) and Fischer F344 strain rats, Dio2 and Dio3 are reciprocally regulated [59–61]. Dio2 is also elevated in the tanycytes of sheep in LD , but at present, there is no information available as to the involvement of Dio3. Together, these data point to a local depletion of T3 as the trigger to induce seasonal changes in physiology. This view is supported by T4 microimplant studies in sheep  and T3 microimplant studies in Siberian hamsters . In the latter, T3 microimplants into the hypothalamic region prevented the SD-induced reduction in body weight and testicular regression, but did not prevent the SD-mediated pelage moult, a response governed by a local PT–pituitary interaction .
In addition to the modulation of hypothalamic T3 availability via regulation of Dio2 and Dio3 expression, thyroid hormone transport may also be involved. MCT8, a specific thyroid hormone transporter , has been shown to be regulated by photoperiod in tanycytes of the Siberian hamster and F344 rat. But paradoxically, Mct8 shows increased expression in SD, a counter-intuitive response given that a reduction in hypothalamic T3 is required to drive the transition to SD physiology [60, 64]. The significance of this has yet to be determined, but given that Dio2 and Dio3 mRNAs exist within the same cells during the development of SD physiology, there is a possibility that a metabolite of T3 may be involved in the seasonal response. A more parsimonious explanation is that rT3, a product of Dio3 enzyme activity on T4, is required to inhibit Dio2 activity  as expression of Dio2 only partially decreases in the Siberian hamster.
Tanycytes of the F344 rat also show regulation of an alternative thyroid hormone transporter Oatp1c1, which is elevated in LD consistent with a role in increasing thyroid hormone supply . Taken together, these data support the view that regulation of T3 availability to the hypothalamus is a key factor in driving seasonal physiological responses.
The link between melatonin action in the PT and regulation of deiodinase activity in tanycytes was discovered independently using two different approaches [20, 66]. Nakao et al.  used the advantage of the rapid response in LH secretion in the Japanese quail when switched from SD to LD with concomitant differential expression of Dio2 and Dio3. Application of a genome wide expression analysis to identify temporally regulated waves of gene expression in the mediobasal hypothalamus underpinning the photoresponsiveness of the quail revealed a first wave of gene expression following the switch from SD to LD which included an increase in expression of TSHβ. This was followed by a second wave of gene expression 4–5 h later, which included upregulation of Dio2 expression. Regulated expression of TSHβ was localized to the PT leading to a hypothesis that PT-derived TSH may be the signal for photoperiod-mediated changes in physiology. Subsequently, in situ hybridization analysis demonstrated TSH receptor gene expression in tanycytes and intracerebral ventricular injection of TSH led to an increase in genes known to be regulated by cAMP including Creb, Icer and Dio2 .
A similar conclusion was derived from studies in Soay sheep, based on previous observations of regulation of TSHβ and αGSU by photoperiod in the PT of the Siberian hamster . Thus, in the Soay sheep, TSHβ is elevated in LD, TSH receptors are found both on cells of the ependymal layer (tanycytes) and on PT cells and TSH infused into the lateral ventricles induces Dio2 expression . Together, these data imply tanycytes are a portal between melatonin action in the PT and a central response leading to physiological adaptations in body weight and the reproductive axis.
Recently, a detailed ultrastructural analysis of the rat PT has provided a new insight as to how the PT may communicate a TSH signal with tanycytes of the third ventricle. This study reveals a complex architecture of the PT consisting of secretory and nonsecretory cells surrounding a cavity (cistern) that is in open communication with the CSF allowing direct access of PT-secreted products access to tanycytes (Fig. 2) . In addition, tanycytic process has been found to terminate on PT cells and thereby potentially facilitating direct communication between PT cells and tanycytes. Another potential route of access of TSH to tanycytes is release into the portal blood vessels with a retrograde blood flow to the hypothalamus. This concept is in opposition to the classical view of neuroendocrine communication between the hypothalamus and pituitary gland via anterograde flow of blood in the portal blood vessels leading from axon terminal of the median eminence to the pituitary gland. However, recent evidence supports the concept of retrograde blood flow to the basal ARC region of the hypothalamus , leaving open the potential for this route of communication between the PT and the brain.
Thyroid hormone replacement experiments performed in birds, sheep and Siberian hamster clearly show that thyroid hormone is an important agent for the translation of photoperiod. Furthermore, tanycytes are key cells in relaying the thyroid hormone to initiate the seasonal change in physiology. However, it should not be overlooked that tanycytes may have an important role in transport or responses to other biologically active or prohormone molecules. One strong candidate for which tanycytes may act as a gatekeeper is retinol. Studies on the Siberian hamster and rat have shown components of the retinoic acid signalling system to be present in tanycytes, including STRA6, CRBP1 (transport proteins for retinol), RADLH2 for the conversion of retinol to retinoic acid, and the receptor for retinoic acid, RAR. The expression of some of one or more of these components (CRBP1, RADLH2 and RAR) is downregulated in SD in tanycytes or hypothalamic structures in the Siberian hamster and the F344 rat, suggesting this signalling pathway is also important to hypothalamic responses underlying seasonal physiology [69–72].
The orphan G-protein-coupled receptor (GPR50) that shares significant homology with the melatonin receptor family, but does not bind melatonin [73–75], is highly expressed in tanycytes  and is downregulated in the SD-exposed Siberian hamster . This would suggest that the ligand for GPR50 is important in mediating an endocrine, neuropeptide or neurotransmitter signalling to tanycytes of LD-exposed hamsters, but the precise role of this receptor and whether downregulation may be involved in alteration in tanycyte functions in SD will require further studies once the ligand has been identified or technological approaches used to carry out tanycyte knock-down or over-expression studies become available.
The mouse is generally not considered to be a seasonal mammal. This could be attributed to the use of mouse strains that do not synthesize melatonin; however, even in melatonin proficient mice, changes in seasonal physiology have not been noted. Nevertheless, as mentioned earlier, mice are responsive to photoperiod at the level of the PT and demonstrate photoperiod regulation of Dio2 expression at the level of the tanycyte . This being the case, it suggests that as the molecular mechanisms exist within the PT and tanycytes are largely intact, the inability to respond to seasonal photoperiod lies downstream of the first two components required for the interpretation of photoperiod.