Most endocrine glands are governed by both neuro-endocrine and neural factors. As for the thyroid gland, it is well established that the trophic hormone TSH, secreted by the anterior pituitary, is the principal endocrine regulator of thyroid function. The serum concentration of TSH is regulated within the context of the hypothalamus-pituitary-thyroid (HPT) axis, which has received a great deal of attention in the literature over the past decades. In addition to this neuro-endocrine regulation, it has been known for many years that the thyroid gland is also richly innervated by both sympathetic and parasympathetic nerve fibers. In this short review, we will focus on neuro-anatomical aspects and functional implications of autonomic innervation of the thyroid gland highlighting recent advances in this research area.

Neuro-anatomy

As early as in the 1930s, nerve-endings surrounding blood vessels and follicles in the thyroid were reported, suggesting a role in thyroid function (1). More recently, it appeared that several ganglia contribute to the innervation of the thyroid gland. Tracing studies using True Blue as a retrograde tracer indicated that the thyroid ganglion and superior cervical ganglion (SCG) contribute most to the nerve supply of the thyroid, with more moderate contributions from the jugular nodose and cervical dorsal root ganglia (2). By combining surgical resection of the SCG and immunohistochemistry, the SCG was identified as the major sympathetic ganglion projecting to the thyroid (3).
As for parasympathetic innervation of the thyroid, the picture is less clear. Acetylcholinesterase (AChE)-positive fibers have been identified in the rat thyroid. However, the assumption that these represent postganglionic parasympathetic fibers is uncertain, as also adrenergic and sensory neurons may contain AChE. The thyroid is thought to receive its parasympathetic innervation via vagal branches originating in the brain stem, namely the superior laryngeal nerve (thyroid nerve), and -although to a much lesser extent- the recurrent nerve, which anastomoses with the thyroid nerve (2). The thyroid nerve projects to the thyroid ganglion which contains mostly vasoactive intestinal peptide (VIP) and neuropeptide Y (NPY) expressing neurons, innervating the thyroid gland.
Until recently, no data were present as to the central nuclei within the central nervous system (CNS) contributing to the autonomic innervation of the thyroid. In order to further investigate this, we performed viral tracing studies using pseudorabies virus (PRV) as a retrograde, transsynaptic tracer. PRV was injected into the thyroid gland of rats and by applying different times of survival, successive stages of infection were visualized using immunocytochemistry. Following the virus as it traveled towards the CNS via neural pathways enabled us to visualize polysynaptic autonomic pathways from the (pre-autonomic) hypothalamus and brain stem to the thyroid gland (4). After two days of survival, infection of sympathetic motorneurons occurred in the intermediolateral nucleus (IML), while parasympathetic motorneurons were labeled in the dorsal vagal complex (DMV). This confirmed the existence of both sympathetic and parasympathetic thyroid innervation. Furthermore, at three days of survival, infected neurons in the hypothalamic paraventricular nucleus (PVN) became apparent. These findings provided evidence for the existence of multisynaptic neuronal pathways between the hypothalamic PVN and the thyroid, probably via both branches of the autonomic nervous system. A schematic representation of these pathways is presented in Figure 1.

Fig 1. Schematic representation of endocrine and neural connections between the central nervous system and the thyroid gland. The hypothalamic suprachiasmatic nucleus (SCN) represents the biological clock, projecting to the paraventricular nucleus (PVN). The PVN harbours TRH expressing neurons that project to the median eminence, where TRH is released into the portal venous system. In the anterior pituitary, TRH stimulates the release of TSH, which acts on the thyroid gland to stimulate the synthesis and release of thyroid hormones. Negative feedback action by these hormones occurs both at the level of the anterior pituitary and the PVN. The hypothalamic arcuate nucleus is an additional site with abundant expression of thyroid hormone receptors. In addition to this well established neuro-endocrine axis, the PVN projects to the dorsal motor nucleus of the vagus nerve (DMV) in the brain stem, where parasympathetic motor neurons innervating the thyroid gland are located. An additional autonomic projection from the PVN targets the intermediolateral column (IML) in the spinal cord, which contains sympathetic motor neurons projecting to the thyroid gland.

Functional implications

The potential of sympathetic input to the thyroid as to altering thyroid function has been shown in older studies. For example, unilateral electrical stimulation of the postganglionic cervical sympathetic trunk induced ipsilateral colloid droplet formation and a marked increase in plasma radiolabeled iodine in mice pre-treated with thyroxine in order to suppress TSH (5). More recently, electrical stimulation of the cervical sympathetic trunk was shown to induce NE release into the thyroid vein as well as a sharp decrease in thyroid blood flow in rats, suggesting indirect neural control of thyroid function by modulating thyroid blood flow (6). A functional role for parasympathetic innervation of the thyroid in modulating thyroid blood flow was shown in experimental thyrotoxicosis. In this condition, electrical stimulation of the thyroid nerve resulted in an increase in thyroid blood flow. This effect was partly blocked by atropine pre-treatment (7).
In addition to classical autonomic neurotransmitters such as noradrenaline and acetylcholine, neurons innervating the thyroid contain a variety of neuropeptides including NPY and VIP. These substances often appear to coexist and may be co-released by the same neuron (3). For example, NPY is co-expressed by noradrenergic neurons projecting to the thyroid from the SCG, and presumably parasympathetic fibers from the thyroid ganglion containing VIP may express NPY as well. VIP was shown to induce accumulation of cAMP in human thyroid cells in vitro and this was unaffected both by anti-adrenergic and by anticholinergic substances. In addition, VIP stimulates release of thyroxine (T4) from human thyroid slices (8). When administered exogenously in rats, VIP increases thyroid iodine uptake (9), blood flow (10) and thyroid hormone secretion (11).
Both NPY and VIP are potent vaso-active substances. In isolated rabbit blood vessels, NPY induces vasoconstriction and enhances noradrenaline-induced vasoconstriction (12). Also when administered exogenously in vivo, NPY -like noradrenaline- decreases thyroid vascular conductance while potentiating the vasoconstricting effect of noradrenaline in rats (13), suggesting a possible indirect control of thyroid function by modulating thyroid blood flow.
It is questionable if NPY and VIP are important players in vivo with respect to nervous signaling to the thyroid gland. In physiological experiments in rats, NA (but not NPY) appeared to be the primary mediator of the acute thyroid blood flow response to sympathetic nerve stimulation (6,13). In addition, decreased concentrations of triiodothyronine (T3) and T4 in thyroid venous plasma induced by iodine deficiency were not paralleled by changes in thyroidal or neural (ganglial) VIP or NPY protein or mRNA content (6). Also blocking VIPergic signaling by VIP antibodies had no effect on thyroid function or thyroid blood flow in rats (14). Therefore, the importance of these neuropeptides for neural regulation of the thyroid gland is probably less than the classical neurotransmitters.
Altered neural regulation should perhaps be considered in situations where a discrepancy exists between serum thyroid hormone concentrations on the one hand and serum TSH on the other. For example, when studying the role of the suprachiasmatic nucleus (SCN), i.e., the endogenous biological clock, in daily rhythmicity of plasma TSH and thyroid hormones, it appeared that SCN ablation had only minor effects on daily plasma TSH levels. However, the daily fluctuations in plasma levels of thyroid hormones, in particular T4, were clearly abolished (4) (Figure 2).

Figure 2. 24-hour rhythms in plasma concentrations of T3, T4 and TSH in SCN-lesioned male rats (Δ). The shaded area indicates the mean ± SEM for control (SCN-intact) animals. Asterisks indicate time points that differ significantly between SCN-lesioned and SCN-intact animals. Copyright © 2000 by The Endocrine Society

This observation led us to speculate on alternative mechanisms by which the SCN could influence thyroid hormone secretion. Next to control of TSH release via its input to thyrotropin-releasing hormone (TRH) expressing neurons in the PVN (the neuroendocrine pathway), the SCN may influence thyroid function by its input to pre-autonomic neurons in the PVN that project to the thyroid via multisynaptic autonomic pathways, as revealed earlier by our tracing studies.

Interestingly, double-labelling experiments using confocal laser scanning microscopy showed TRH immunoreactivity in some PRV-infected pre-autonomic neurons in the PVN after inoculation of PRV into the thyroid gland. Moreover, numerous TRH-immunoreactive fibers were seen in close approximation to a large proportion of PRV-infected preganglionic neurons in the IML, suggesting an extensive TRH input to sympathetic motor neurons in the spinal cord. This also holds, although to a lesser extent, for the DMV area in the brain stem harbouring parasympathetic preganglionic neurons. (Figure 3)(4).

Figure 3 Composite two-color confocal laser scanning microscopical images of the same optical sections in the hypothalamus and spinal cord. Visualisation of TRH immunoreactive perikaryal profiles and axonal endings in red and PRV-FITC staining in green.
A, PRV (green) and TRH (red) staining in the paraventricular nucleus of the hypothalamus; B, a TRH-containing PVN neuron that is also infected by PRV (i.e. colocalization); (E) PRV and TRH (red) staining in the parasympathetic preganglionic nucleus of the solitary tract and DMV area; (C) Shows a PRV-infected DMV neuron receiving two TRH-immunoreactive contacts; (D,F) PRV and TRH (red) staining in the sympathetic preganglionic IML. Scale bar, 12 µm in F.

The origin of these TRH fibers is uncertain at present. The colocalization of PRV and TRH in the hypothalamic PVN indicates the possibility of a TRH containing projection to the IML and DMV. However, it seems more likely that the TRH containing IML/DMV projections originate from the raphe nucleus and ventral medulla, where a population of TRH containing neurons has been shown to project to these autonomic motornuclei (15,16). Nevertheless, the colocalization of TRH and PRV in PVN neurons after PRV inoculation in the thyroid gland gives rise to the notion of a second population of TRH containing neurons in the PVN projecting to the thyroid via multisynaptic autonomic pathways, in addition to the well-known TRH containing neurons projecting from the PVN to the median eminence.
Theoretically, there are several mechanisms by which the autonomic input to the thyroid may interfere with T3 and T4 plasma concentrations. It could influence thyroid hormone release directly, it could alter thyroid responsiveness to TSH and it could interfere with the deiodination of the pro-hormone T4 to the biologically active hormone T3. In our earlier SCN lesioning experiments, there was a clear lowering effect on mean plasma concentrations of T4, while plasma TSH and T3 were unaffected (4) (Figure 2), suggesting impaired thyroid sensitivity to TSH in SCN-lesioned animals in keeping with the second possibility. The idea of autonomic nervous system involvement in the process of deiodination is supported by our recent experiments on type 2 deiodinase activity in the rat pineal gland, which also receives sympathetic input via the SCG. We showed that the large diurnal variation of type 2 deiodinase activity was abolished by lesioning of the SCN (17).
In several endocrine organs, including the adrenal cortex and the testis, a role for autonomic innervation in modulating the responsiveness to the corresponding pituitary hormones has been demonstrated (18,19). The notion of CNS control of thyroid responsiveness to TSH is supported by studies in mice showing that exogenous noradrenaline inhibits the thyroid response to TSH (20). Increased sympathetic tone of thyroid innervation has also been proposed as a partial explanation of discordant serum T3 and TSH in the framework of non-thyroidal illness (NTI) (21). Finally, recent studies in rats confirmed a functional role for sympathetic thyroid innervation by showing that unilateral SCG lesions induced a reduction in thyroid weight, and in TSH-stimulated radioiodine uptake (22).

Conclusion

Pre-autonomic neurons in the hypothalamic PVN, partly expressing TRH, project to the thyroid via sympathetic as well as parasympathetic pathways. There is growing evidence suggesting a role for the autonomic nervous system in modulating thyroid function, and possibly thyroid size. Thus, in addition to the well-established neuro-endocrine HPT axis, neural control of the thyroid gland may prove to be an important modulator of thyroid function in health and disease.