Cell Tissue Res (2002) 309:139–150 DOI 10.1007/s00441-002-0580-5

REVIEW

Morten Møller · Florian M. M. Baeres

The anatomy and innervation of the mammalian pineal gland

Received: 3 January 2002 / Accepted: 2 April 2002 / Published online: 18 May 2002 © Springer-Verlag 2002

Abstract The parenchymal cells of the mammalian pineal gland are the hormone-producing pinealocytes and the interstitial cells. In addition, perivascular phagocytes are present. The phagocytes share antigenic properties with microglial and antigen-presenting cells. In certain species, the pineal gland also contains neurons and/or neuron-like peptidergic cells. The peptidergic cells might influence the pinealocyte by a paracrine secretion of the peptide. Nerve fibers innervating the mammalian pineal gland originate from perikarya located in the sympathetic superior cervical ganglion and the parasympathetic sphenopalatine and otic ganglia. The sympathetic nerve fibers contain norepinephrine and neuropeptide Y as neurotransmitters. The parasympathetic nerve fibers contain vasoactive intestinal peptide and peptide histidine isoleucine. Recently, neurons in the trigeminal ganglion, containing substance P, calcitonin gene-related peptide, and pituitary adenylate cyclaseactivating peptide, have been shown to project to the mammalian pineal gland. Finally, nerve fibers originating from perikarya located in the brain containing, for example, GABA, orexin, serotonin, histamine, oxytocin, and vasopressin innervate the pineal gland directly via the pineal stalk. Biochemical studies have demonstrated numerous receptors on the pinealocyte cell membrane, which are able to bind the neurotransmitters located in the pinealopetal nerve fibers. These findings indicate that the mammalian pinealocyte can be influenced by a plethora of neurotransmitters. Keywords Superior cervical ganglion · Sphenopalatine ganglion · Otic ganglion · Trigeminal ganglion · Central innervation

M. Møller (✉) · F.M.M. Baeres Institute of Medical Anatomy, Panum Institute, University of Copenhagen, Blegdamsvej 3, 2200 Copenhagen N, Denmark e-mail: [email protected] Tel.: +45-35327258, Fax: +45-35369612

Introduction The mammalian pineal gland is a neuroendocrine gland secreting the hormone melatonin (Lerner et al. 1959; Ganguly et al. 2002; Stehle et al. 2002). The gland is derived from the neural tube and located at the border between the mesencephalon and the diencephalon of the brain. Neuroanatomically, the pineal is described as a part of the epithalamus and thereby as a part of the diencephalon. The structure of the gland has been extensively studied at the light- and electron-microscopic levels (see Bargmann 1943; Vollrath 1981, 1984). However, since the extensive review by Vollrath in 1981, anatomical studies have provided new information regarding the classification of cell types of the gland as well as the innervation of the gland.

Anatomy of the mammalian pineal gland The parenchyma of the mammalian pineal gland consists of cords of cells separated by capillaries with fairly wide perivascular spaces (Fig. 1). In certain larger species, for example the human, the parenchyma is divided into lobules separated by loose connective tissue septa derived from the pial tissue. A pial capsule surrounds the gland in all species. In most mammals, for example in the human, monkey, horse, cow, sheep, pig, mink, and hedgehog, the pineal gland exhibits a pyramidal shape and is located directly on the dorsal part of the brain stem, i.e., just rostral to the superior colliculi. However, in several other species, the pineal has an elongated structure, and in rodents the gland is subdivided into two parts: (1) a superficial pineal gland located at the dorsal surface of the brain and (2) a deep pineal gland located on the brain stem. The two parts are connected through the pineal stalk, which in most species contains pinealocytes (for an extensive description of the macroscopic appearance of the pineal gland of different mammalian species, see Vollrath 1981).

140

Fig. 1 Photomicrograph of a part of a survey section of the rat pineal gland. The ovoid pinealocytes are arranged in cords separated by the capillaries (c) and the perivascular spaces. The star-shaped interstitial cells (arrows) are characterized by their triangular dark nucleus and the dark cytoplasm. A phagocyte (phagocyte) is seen in the perivascular space. Two-µm-thick Epon-embedded section. Toluidine-blue staining. Scale bar 25 µm Fig. 2 Electron micrograph of the rat pineal gland showing an interstitial cell (Interst. cell) located between pinealocytes with more electron-lucent cytoplasm. A club-shaped terminal of a pinealocyte process is seen (arrow). Scale bar 5 µm

Fig. 3 Electron micrograph of the rat pineal gland showing a perivascularly located phagocyte (arrow). Note the high number of dense bodies in the cytoplasm. Scale bar 3.5 µm Fig. 4 Electron micrograph of a perivascular space of the rat pineal gland. Horseradish peroxidase was injected into the lateral ventricle of the brain 30 min before the rat was killed. The peroxidase was visualized by incubation in diaminobenzidine and hydrogen peroxide. Note the high uptake of the peroxidase into large (arrow) and small (arrowheads) vesicles of a phagocyte located in the perivascular space. Scale bar 10 µm

141

Fig. 5 Frontal section of the deep pineal gland of the mouse stained for S-antigen. The pinealocytes (arrowheads) of the deep pineal gland exhibit strong immunoreactivity. S-antigen-immunoreactive processes (arrows) project toward the habenula (PR pineal recess, SCO subcommissural organ, A cerebral aqueduct). Scale bar 50 µm. (From Korf et al. 1990, with permission) Fig. 6 Electron micrograph of an S-antigen immunoreactive pinealocyte (asterisk) in the mouse habenula. The black rectangle corresponds to the area shown in higher magnification in the inset. Scale bar 1.5 µm. Inset: A nerve terminal, without immunoreactivity, making synaptic contact (arrowheads) with an S-antigenimmunoreactive process. Scale bar 0.15 µm. (From Korf et al. 1990, with permission)

Pineal cell types In all mammals investigated five cell types have been described: (a) the hormone-producing pinealocyte, (b) the interstitial cell, (c) the perivascular phagocyte, (d) neurons, and (e) peptidergic neuron-like cells (Figs. 1, 2, 3, 4). The pinealocyte (Figs. 1, 2) consists of a cell body (7–12 µm in diameter) from which three to five processes emerge. These processes are rarely seen in the light microscope but can be seen after silver impregnation of the sections. In the electron microscope, dense-core granules are present with the highest density of granules in club-shaped terminals (Fig. 2) of the cellular processes. In the pineal gland, the so-called synaptic ribbon (Vollrath 1981) has been considered a specific marker for the pinealocyte. Although this organelle is common in the pinealocytes of several mammalian species, for example the rat and guinea pig, the number of synaptic ribbons is low or absent in many other species. The ultrastructure of the mammalian pinealocyte has also been reviewed in several extensive surveys (see, for example, Vollrath 1981, 1984) to which the reader is referred. Although specific antibodies against melatonin have been available for many years, immunocytochemistry has never succeeded to demonstrate melatonin in any specific organelle.

The pinealocytes express several retinal antigens (Schomerus et al. 1994), for example opsin (Korf et al. 1985a; Huang et al. 1992), recoverin (Korf et al. 1992), and the retinal S-antigen (Fig. 5; Korf et al. 1985b, 1990). The S-antigen has also been used for characterization of pineal parenchymal tumors (Korf et al. 1986; Perentes et al. 1986). By using the S-antigen as a marker, displaced pinealocytes were found in the habenular (Fig. 5) and pretectal areas (Korf et al. 1990), and some of these pinealocytes establish synapse-like contacts with neurons (Fig. 6) in the brain (Korf et al. 1990) suggesting a direct communication between pinealocytes and neurons in the brain. Two important enzymes of the melatonin synthesis, arylalkylamine-N-acetyltransferase (Coon et al. 1995) and hydroxyindole-O-methyltransferase (Ishida et al. 1987) have been cloned and pinealocytes can be identified by in situ hybridization using antisense probes recognizing nucleotide sequences in the mRNA encoding these enzymes (Pfeffer and Stehle 1998; Ribelayga et al. 1999). Detection of mRNA encoding hydroxyindole-Omethyltransferase has been used to characterize pineal tumors (Tsumanuma et al. 2000). The interstitial cell is smaller than the pinealocyte and exhibits, in the light and electron microscope, a darker triangular nucleus and also a darker cytoplasm (Figs. 1, 2) than the pinealocyte. The cell is star-shaped with several long and slender processes. In most species, the interstitial cell contains a high number of filaments. Several, but not all of the interstitial cells are immunoreactive for the glial fibrillary acid protein (Møller et al. 1978a), which is also found in the fibrillary astrocytes of the central nervous system. The third cell type, the phagocytic cell (Figs. 3, 4), mostly confined to the perivascular spaces, has been described during the last decade. These cells show a high uptake of exogenously applied tracer, for example horseradish peroxidase (Fig. 4) and might be a perivascular microglial-like cell like the ones present in other parts of

142

Fig. 7 Schematic representation of results from in vivo neuronal tracings and immunohistochemistry on the innervation of the rat pineal gland. Information from the suprachiasmatic nucleus (SCN) is transmitted to the paraventricular nucleus of the hypothalamus (PV). The paraventricular nucleus projects via a dorsal and ventral pathway to the intermediolateral nucleus (iml) of the thoracic spinal cord. The intermediolateral nucleus of the spinal cord projects to the superior cervical ganglion (SCG), from where sympathetic neurons project to the pineal gland via the conarian nerves. The parasympathetic innervation originates from the sphenopalatine (Sphenopal. gangl.) and otic ganglia (Otic gangl.). Nerve fibers reaching the pineal from these ganglia might enter the cranial cavity via the ethmoidal nerve through the ethmoidal foramen in the orbita of the rat. In the trigeminal ganglion (Trig. gangl.) small perikarya are present projecting to the pineal gland. The central innervation of the rat pineal is illustrated by a projection from neurons located in the medial part of the paraventricular nucleus. It enters the deep pineal gland and the pineal stalk (Centr. inn) and passes through the periventricular and epithalamic region. In this drawing of the median plane of the rat brain, other areas of the brain projecting to the deep pineal (for example the lateral geniculate nucleus, the lateral hypothalamus, habenula, and the dorsal raphe nucleus) are not illustrated

the central nervous system. Immunocytochemical studies have revealed that such perivascular cells often contain surface proteins, which are markers for macrophages and microglial cells. The phagocytes are also antigenpresenting cells because they are immunoreactive for the class II major histocompatibility system (Pedersen et al. 1993; Sato et al. 1996). In several mammalian species, for example human, rabbit, monkey, ferret, and cotton rat the pineal contains classic neurons characterized by Nissl substance, which receive an input via synapse-like contacts (Matsushima et al. 1994). In some species, these neurons form real ganglia within the pineal gland. The current concept is that most of the intrapineal neurons are parasympathetic neurons innervated by a peripheral ganglion. However, in the ferret the intrapineal ganglion receives a synaptic input from the neurons located in the habenular nucleus of the epithalamus (David and Herbert 1973). Finally, a peptidergic neuron-like cell has been shown in several rodent and non-rodent pineals. Thus, in the

European hamster (Coto-Montes et al. 1994) and human (own observations), neuron-like cells immunoreactive to enkephalin are present. Immunoelectron microscopy of the enkephalinergic cells in the hamster showed that these cells were pinealocytes without a synaptic input. In other species, neuron-like cells are found to be immunoreactive for vasopressin (Badiu et al. 1999) and oxytocin (Badiu et al. 2001). These cells, however, vary among the species, and more investigations are needed to elucidate the frequency, presence, and role of these cells. However, their presence in the pineal of some species suggests a paracrine regulatory function of the pinealocyte (Møller 1997). The mammalian pineal gland and the blood–brain barrier In most species, for example in the human fetus, Mongolian gerbil, Golden hamster, and sheep, the capillaries in the pineal gland are endowed with continuous endothelium. Contrarily, the endothelial cells of the capillaries in the rat pineal possess few fenestrations and in the mouse and Djungarian hamster many fenestrated capillaries are present (Vollrath 1981; Møller and van Veen 1981). After intravascular injection of hydrophilic tracers with low molecular weights, such as fluorescein or Lissamine green, the pineal glands displaying either continuous or fenestrated capillaries were stained by the tracers (Møller et al. 1978b; Møller and van Veen 1981). Also high molecular weight tracers, for example horseradish peroxidase (MW 40,000) permeate both the continuous capillaries of the gerbil pineal (Welsh and Beitz 1981) and the fenestrated capillaries of the mouse pineal (Møller et al. 1978b). Some studies have suggested that the capillaries of the deep pineal are less permeable to high molecular weight tracers (Chen et al. 1994). However, one must conclude that the majority of pinealocytes can be targeted by hydrophilic substances, for example peptides circulating in the vascular system (Reuss and Schröder 1987).

143

Fig. 8 Electron micrograph of neuropeptide Y (NPY)-immunoreactive nerve terminals (arrows) in the perivascular space of a rat pineal gland intermingled with non-reactive nerve terminals. Scale bar 3 µm Fig. 9 Electron micrograph of NPY-immunoreactive nerve terminals in the perivascular space of the rat. One immunoreactive terminal (arrow) is located intraparenchymally. Scale bar 3 µm

Innervation of the mammalian pineal gland The melatonin synthesis and secretion in the pineal gland has for more than four decades been known to be inhibited by light received by retinal photoreceptors and transmitted to the brain via the optic nerves. After the suprachiasmatic nucleus was identified as the circadian pacemaker for the melatonin rhythm in the pineal gland, many neuroanatomical tracing studies have tried to map the connections between the suprachiasmatic nucleus and the pineal gland. By use of both retrograde and anterograde tracers, a projection from the suprachiasmatic nucleus to the paraventricular nucleus of the hypothalamus was demonstrated (Vrang et al. 1995; Kalsbeek et al. 2000; Munch et al. 2002). Recently, by using the retrograde in vivo transsynaptic virus-tracing technique, the total pathway was retrogradely mapped (Fig. 7) from the pineal gland, via the superior cervical ganglion, to the intermediolateral nucleus of the thoracic spinal cord and further to the paraventricular nucleus and finally to the suprachiasmatic nucleus (Larsen et al. 1998; Buijs et al. 1999).

Sympathetic innervation of the pineal gland The final neuron of the above-described retrogradely mapped pathway to the pineal gland is a sympathetic neuron located in the superior cervical ganglion. This sympathetic innervation has been known for more than half a century (Bargmann 1943; Kappers 1960) and the functional importance of this innervation for pineal melatonin secretion (Yuwiler et al. 1977) has been known for decades. The sympathetic neurons of the superior cervical ganglion contain, in addition to norepinephrine, neuropeptide Y (NPY; Figs. 8, 9, 10; Zhang et al. 1991; Cozzi et al. 1992; Mikkelsen and Møller 1999). However, in several species, few NPY-immunoreactive nerve fibers persist in the gland after superior cervical ganglionectomy. In some species, for example the rat, these fibers are predominantly located in the deep pineal gland (Matsushima et al. 1999; Mikkelsen and Møller 1999), which suggests an origin of these fibers from perikarya in the brain. Non-sympathetic innervation of the pineal gland Numerous neuroanatomical tracing studies, often combined with immunohistochemistry for neurotransmitters, have consistently shown that pinealopetal nerve fibers also originate from parasympathetic ganglia and in the trigeminal ganglion (see Møller 1999). In addition, nerve fibers originating from the brain innervate the mammalian pineal directly via the pineal stalk. The direct innervation with nerve fibers originating from the brain is called the central innervation of the pineal gland.

144

145

The generation of specific antibodies with high affinity against different neurotransmitters, especially the neuropeptides, has been an important step for the elucidation of the non-sympathetic innervation of the mammalian pineal gland. Strong anatomical evidence for the presence of a non-sympathetic innervation of the mammalian pineal gland was provided by studies demonstrating that, after removal of the superior cervical ganglion, several of the peptidergic nerve fibers remained in the pineal gland (see Møller et al. 1996). Receptors were found on the pinealocyte membrane for several neuropeptides present in the pinealopetal nerve fibers and several neuropeptides were shown to influence pinealocyte biochemistry (see Simonneaux 1995). Parasympathetic innervation of the pineal gland

▲

Already 40 years ago, the presence of a parasympathetic innervation of the monkey pineal was strongly indicated by the study of Kenny (1961) who lesioned the greater petrosal nerve and observed degeneration of nerve fibers in the pineal gland. Corresponding results were later obtained in the rabbit (Romijn 1975). With regard to the parasympathetic pineal innervation, two neuropeptides appear to be important: (a) vasoactive intestinal peptide (VIP) and (b) peptide histidine isoleucine (PHI), both present in pinealopetal nerve fibers of mammals (Møller et al. 1985; Møller and Mikkelsen 1989). Vasoactive intestinal peptide and PHI are derived from the same pre-promolecule, transcribed from the same gene. The origin of the immunoreactive nerve fibers containing VIP was traced back to the sphenopalatine ganglion by a combination of in vivo tracing experiments and immunohistochemistry (Shiotani et al. 1986). The parasympathetic otic ganglion also contains many neurons immunoreactive for VIP and PHI. However, to date immunohistochemical demonstration of these peptides has not yet been combined with neuronal tracing techniques. Although VIP and PHI are considered classic “parasympathetic” neuropeptides and NPY appears as a “sympathetic” neuropeptide, few NPY-immunoreactive perikarya are also present in the parasympathetic sphenoFig. 10 Micrograph of a part of the rat pineal gland showing immunofluorescence for NPY. Note the dense network of immunoreactive peptidergic nerve fibers (arrows) in the gland. Scale bar 20 µm Fig. 11 Fluorescence micrograph of the trigeminal ganglion of the rat. Two green-fluorescent pituitary adenylate cyclase-activating peptide (PACAP)-immunoreactive perikarya (arrows) are seen. Neurons retrogradely labeled after injection of Fluoro-gold into the superficial pineal gland exhibit a red fluorescence (arrowheads). Not all neurons are double labeled. Scale bar 10 µm Fig. 12 Retrogradely Fluoro-gold labeled neuron in the trigeminal ganglion. In Fig. 13, the same neuron is seen to exhibit immunofluorescence for PACAP. Scale bar 5 µm Fig. 13 The same neuron as shown in Fig. 12 exhibiting immunofluorescence for PACAP. Scale bar 5 µm

palatine ganglion, and after axotomy some neurons in the superior cervical ganglion were found to express VIP (Shadiack et al. 2001). Innervation of the pineal gland from the trigeminal ganglion New anatomical studies have shown that also perikarya in the trigeminal ganglion project to the pineal gland of mammals (Reuss 1999). The trigeminal ganglion is the classic sensory ganglion of the fifth cranial nerve. The ganglion has for many years been known to contain peptidergic perikarya. Thus, substance P (SP), calcitonin gene-related peptide (CGRP), and pituitary adenylate cyclase-activating peptide (PACAP) (Figs. 11, 12, 13) are present in cell bodies of the trigeminal ganglion. Both SP (Reuss 1999) and CGRP (Møller et al. 1999) are also present in intrapineal nerve fibers. Antibodies against SP have been available for several decades and nerve fibers immunoreactive for this peptide have been shown in the pineal gland of several rodent (Rønnekleiv and Kelly 1984; Shiotani et al. 1986; Matsushima et al. 1994) and non-rodent species (Korf and Møller 1984; Rønnekleiv 1988; Møller et al. 1993; Kado et al. 1999). Since SP immunoreactivity is found in the medial habenular nucleus and in nerve fibers of the pineal stalk, we have previously suggested that these fibers belong to the central innervation (Møller et al. 1993). However, studies combining neuronal tracings and transmitter immunohistochemistry have clearly shown that SP-immunoreactive nerve fibers also originate from the trigeminal ganglion. The peptide PACAP, belonging to the secretin family, has recently been demonstrated in intrapineal nerve fibers (Møller et al. 1999), and it stimulates melatonin secretion via specific receptors located on the pinealocyte cell membrane (Schomerus et al. 1996). The PACAPimmunoreactive nerve fibers also contain CGRP (Møller et al. 1999) but not VIP. In our laboratory we have traced the PACAP fibers back to perikarya in the trigeminal ganglion (Figs. 12, 13). From a neurobiological point of view it is interesting that the sensory trigeminal ganglion contains neurons that stimulate the hormone secretion of a neuroendocrine cell. The central innervation of the mammalian pineal gland The mammalian pineal gland is attached to the brain via a stalk connected to the habenular and posterior commissures. The first neuroanatomical studies of the brain–pineal connections, based on silver impregnations of tissue sections of the epithalamus, described numerous nerve fibers penetrating into the pineal gland from the brain via both commissures. From a functional point of view, a central innervation is logical as the posterior commissure contains numerous nerve fibers belonging to the optic system. Because the pineal melatonin secretion

146

Fig. 14 Electron micrograph of the deep pineal gland of the rat. Nerve terminals (arrows) immunoreactive for orexin are seen among the pinealocytes. Scale bar 2 µm. Inset: Intraventricular orexin-immunoreactive nerve fiber (arrowhead) located in the pineal recess. Scale bar 0.75 µm Fig. 15 Electron micrograph of the border between the deep pineal gland and the pineal recess of the rat. Intraventricular orexin-immunoreactive nerve terminals, cut in the longitudinal (arrow) and in the coronal (arrowhead) plane, are located close to the pineal surface. Scale bar 1 µm

is influenced by the photoperiod, such a direct input from the optic system via the posterior commissure would be the shortest anatomical pathway. However, 40 years ago, the presence of these central nerve fibers was negated. The reason for this was the demonstration of the dense, functionally important sympathetic innervation of the gland, and also that some of the nerve fibers from the habenular and posterior commissures returned to the brain after looping in the gland (Kappers 1965). However, modern neuronal tracing methods and immunohistochemistry have verified the presence of such a central innervation (for surveys see Korf and Møller 1984, 1985; Møller 1999). By use of in vivo tracing methods, the origin of the central nerve fibers has been demonstrated in the paraventricular nucleus of the hypothalamus (Korf and Wagner 1980; Reuss and Møller 1986; Larsen et al. 1991), the lateral geniculate nucleus (Møller and Korf 1983b; Mikkelsen and Møller 1990), the lateral hypothalamus (Fink-Jensen and Møller 1990), and the dorsal raphe nucleus (Leander et al. 1998; Møller and Hay-Schmidt 1998). Classic transmission electron microscopy has strongly supported the presence of a central innervation of the

Fig. 16 Distribution of GABAergic nerve fibers in the pineal gland and habenular (HC) and posterior (PC) commissures of the tree shrew (Tupaia glis). Each immunoreactive profile seen in a single paraffin section was drawn. PR Pineal recess, SCO subcommissural organ. (From Sakai et al. 2001, reprinted with permission by Wiley and Sons)

147

Fig. 17 Electron micrograph of the mouse pineal gland showing GABA-immunoreactive nerve endings one of which (A) makes synapse-like contacts with a pinealocyte cell body (P). Two other nerve endings (B) establish synapse-like contacts with the process of a pinealocyte. Scale bar 0.4 µm. (From Sakai et al. 2001, reprinted with permission by Wiley and Sons) Fig. 18 Electron micrograph of the pineal gland of the tree shrew showing a synapse-like contact between a GABA-immunoreactive nerve ending (N) and the cell body of a pinealocyte (P). Scale bar 0.2 µm. (From Sakai et al. 2001, with permission by Wiley and Sons)

deep pineal gland and the pineal stalk in rodents. Thus, at the ultrastructural level many myelinated nerve fibers are observed in the deep pineal gland and these fibers continue into the rostral part of the pineal stalk. However, in the stalk the number of myelinated fibers decreases and only a few are present in its rostral part (Møller and Korf 1983a; Luo et al. 1984). Immunohistochemistry has indicated that central nerve fibers from the paraventricular nucleus contain vasopressin and oxytocin (Buijs and Pévet 1980; Nürnberger and Korf 1981). By use of immunohistochemistry, histaminergic nerve fibers, originating from the tuberomamillary magnocellular neurons of the posterior hypothalamus, have been observed to enter the deep pineal gland and the pineal stalk of the rat (Mikkelsen et al. 1992). Recently, a new peptide called orexin (hypocretin) that is involved in regulation of appetite and sleep, has been shown to be present in nerve fibers of the central innervation of the rat (Mikkelsen et al. 2001). In our laboratory we have also found such fibers in the pig (unpublished observations). In the rat, the orexin-immunoreac-

tive perikarya are located in the dorsolateral hypothalamic area and in the perifornical area. From these locations, the nerve fibers enter the deep pineal gland (Figs. 14, 15) via the epithalamic area. In the rat pineal gland, expression of the orexin receptor-2 has been demonstrated by RT-PCR, and in cultured pinealocytes orexin has been shown to inhibit the β-adrenergic stimulation of melatonin secretion (Mikkelsen et al. 2001). There are also immunohistochemical indications for the presence of cholinergic (Phansuwan-Pujito et al. 1991) and SP-immunoreactive central nerve fibers in the cow and the pig (Przybylska-Gornowicz et al. 2000). However, it is important to note that the central innervation in most rodents terminates in the deep pineal gland and the pineal stalk (Matsushima et al. 1999; Møller 1999). The paucity of nerve fibers of the central innervation in the superficial pineal gland of rodents has naturally created doubt about a functional influence of this innervation on the whole pineal. This might hold true for rodents, but in many primates, who only possess a deep pineal, nerve fibers of the central innervation are reaching all parts of the gland. This has also been shown in an extensive light- and electron-microscopic immunohistochemical study of the GABAergic innervation of the mammalian pineal in four rodent and one non-rodent species, the tree shrew (Sakai et al. 2001). In all species investigated, many GABAergic fibers could be followed from the habenular and the posterior commissures into the pineal gland (Sakai et al. 2001). In the rodents, the immunoreactive nerve fibers were located in the deep pineal and the pineal stalk. Contrarily, in the tree shrew,

148

gland with the posterior commissure (Fig. 19). This nerve, for which we have suggested the name nervus pinealis, disappears during fetal life. Concluding remarks The innervation of the pineal gland is very complex with inputs from many areas of the central nervous system and peripheral ganglia. Receptors for all the abovedescribed neurotransmitters are present on the pinealocyte cell membrane, and several of the receptors activate second messenger systems in the pinealocyte, thereby influencing the biochemistry of the pinealocyte. From a physiological point of view, the non-sympathetic innervation has not been investigated. This is due to the technical problems arising from the difficulties in a correct stimulation or lesion of the small parasympathetic ganglia and in performing correct stereotaxic lesions and microinjections into the central brain structures projecting to the pineal gland. However, with the strong anatomical documentation of the non-sympathetic nervous input, such studies will be performed in the future.

Fig. 19 Median section through the human fetal pineal gland (crown-rump length=156 mm). The fetal pineal nerve, nervus pinealis (arrows), is located caudal to the pineal gland (V third ventricle, Post. com. posterior commissure). Toluidine-blue staining. Scale bar 0.2 mm

the GABAergic nerve fibers were present in all parts of the gland (Fig. 16). By use of electron-microscopic immunocytochemistry, some GABAergic nerve fibers were found to make synapse-like contacts with the pinealocytes (Figs. 17, 18). The precise origin of the GABAergic nerve fibers has not been elucidated, but possible candidates might be GABAergic neurons located in the intergeniculate leaflet of the lateral geniculate nucleus. This nucleus projects to the pineal gland (see above) and contains GABAergic neurons. The fetal nervus pinealis of mammals In phylogenetic terms, the secretory pineal of mammals and the “third eye” of lower vertebrates develop from a common anlage. During ontogeny, embryonic pinealocytes possess morphological characteristics of photoreceptors (Vollrath 1981) and these cells still express photoreceptor-specific molecules in adult animals (Korf et al. 1985a, b, 1998). The photoreceptors of the “third eye”, for example the frontal organ of the frog, are connected to the brain via the pineal nerve and pineal tract (Oksche and Vaupel-von Harnack 1965). During ontogenetic development of the mammalian pineal, a fetal nerve, shown to be present in the human, rabbit, and sheep (Møller 1979) connects the caudal part of the pineal

References Badiu C, Coculescu M, Møller M (1999) Arginine vasotocin mRNA revealed by in situ hybridization in bovine pineal gland cells. Cell Tissue Res 295:225–229 Badiu C, Badiu L, Coculescu M, Vilhardt H, Møller M (2001) Presence of oxytocinergic neuronal-like cells in the bovine pineal gland: an immunocytochemical and in situ hybridization study. J Pineal Res 31:273–280 Bargmann W (1943) Die Epiphysis cerebri. In: Möllendorff W von (ed) Handbuch der mikroskopischen Anatomie des Menschen, vol VI/1. Springer, Berlin Heidelberg New York, pp 338–502 Buijs RM, Pévet P (1980) Vasopressin- and oxytocin-containing fibres in the pineal gland and subcommissural organ of the rat. Cell Tissue Res 205:11–17 Buijs RM, Wortel J, Van Heerikhuize JJ, Feenstra MG, Ter Horst GJ, Romijn HJ, Kalsbeek A (1999) Anatomical and functional demonstration of a multisynaptic suprachiasmatic nucleus adrenal (cortex) pathway. Eur J Neurosci 11:1535–1544 Chen YL, Chen WP, Huang BN, Lin HS (1994) Permeability of the pineal organ of the golden hamster (Mesocricetus auratus) to HRP with special reference to different types of blood capillaries. Arch Histol Cytol 57:175–186 Coon SL, Roseboom PH, Baler R, Weller JL, Namboodiri MA, Koonin EV, Klein DC (1995) Pineal serotonin N-acetyltransferase: expression cloning and molecular analysis. Science 270:1681–1683 Coto-Montes A, Masson-Pévet M, Pévet P, Møller M (1994) The presence of opioidergic pinealocytes in the pineal gland of the European hamster (Cricetus cricetus, L.). An immunocytochemical study. Cell Tissue Res 278:483–491 Cozzi B, Mikkelsen JD, Ravault J-P, Møller M (1992) Neuropeptide Y (NPY) and C-flanking peptide of NPY in the pineal gland of normal and ganglionectomized sheep. J Comp Neurol 316:238–250 David GFX, Herbert J (1973) Experimental evidence for a synaptic connection between habenula and pineal ganglion in the ferret. Brain Res 64:327–343 Fink-Jensen A, Møller M (1990) Direct projections from the anterior and tuberal regions of the lateral hypothalamus to the rostral

149 part of the pineal complex of the rat. An anterograde neurontracing study by using Phaseolus vulgaris leucoagglutinin. Brain Res 522:337–341 Ganguly S, Coon SL, Klein DC (2002) Regulation of melatonin biosynthesis in the mammalian pineal organ. Cell Tissue Res (this issue) Huang SK, Klein DC, Korf HW (1992) Immunocytochemical demonstration of rod-opsin, S-antigen, and neuron-specific proteins in the human pineal gland. Cell Tissue Res 267: 493–498 Ishida I, Obinata M, Deguchi T (1987) Molecular cloning and nucleotide sequence of cDNA encoding hydroxyindole O-methyltransferase of bovine pineal glands. J Biol Chem 262:2895–2899 Kado M, Yoshida A, Hira Y, Sakai Y, Matsushima S (1999) Light and electron microscopic immunocytochemical study on the innervation of the pineal gland of the tree shrew (Tupaia glis), with special reference to peptidergic synaptic junctions with pinealocytes. Brain Res 842:359–375 Kalsbeek A, Garidou ML, Palm IF, Van Der Vliet J, Simonneaux V, Pévet P, Buijs RM (2000) Melatonin sees the light: blocking GABA-ergic transmission in the paraventricular nucleus induces daytime secretion of melatonin. Eur J Neurosci 12: 3146–3254 Kappers JA (1960) The development, topographical relations and innervation of the epiphysis cerebi in the albino rat. Z Zellforsch Mikrosk Anat 52:163–215 Kappers JA (1965) Survey of the innervation of the epiphysis cerebri and the accessory pineal organs in vertebrates. Prog Brain Res 10:87–151 Kenny GCT (1961) The “nervus conarii” of the monkey (an experimental study). J Neuropathol Exp Neurol 20:563–570 Korf HW, Møller M (1984) The innervation of the mammalian pineal gland with special reference to central pinealopetal projections. Pineal Res Rev 2:41–86 Korf HW, Møller M (1985) The central innervation of the mammalian pineal organ. In: Mess B, Rúzsas C, Tima L, Pévet P (eds)The pineal gland. Akadémiai Kiadó, Budapest, pp 47–69 Korf HW, Wagner U (1980) Evidence for a nervous connection between the brain and the pineal organ in the guinea pig. Cell Tissue Res 209:505–510 Korf HW, Foster RG, Ekström P, Schalken JJ (1985a) Opsin-like immunoreaction in the retinae and pineal organs of four mammalian species. Cell Tissue Res 242:645–648 Korf HW, Møller M, Gery I, Zigler JS, Klein DC (1985b) Immunocytochemical demonstration of retinal S-antigen in the pineal organ of four mammalian species. Cell Tissue Res 239:81–85 Korf HW, Klein DC, Zigler JS, Gery I, Schachenmayr W (1986) S-antigen-like immunoreactivity in a human pineocytoma. Acta Neuropathol (Berl) 69:165–167 Korf HW, Sato T, Oksche A (1990) Complex relationships between the pineal organ and the medial habenular nucleuspretectal region of the mouse as revealed by S-antigen immunocytochemistry. Cell Tissue Res 261:493–500 Korf HW, White BH, Schaad NC, Klein DC (1992) Recoverin in pineal organs and retinae of various vertebrate species including man. Brain Res 595:57–66 Korf HW, Schomerus C, Stehle JH (1998) The pineal organ, its hormone melatonin, and the photoneuroendocrine system. Adv Anat Embryol Cell Biol 146:1–100 Larsen PJ, Møller M, Mikkelsen JD (1991) Efferent projections from the periventricular and medial parvocellular subnuclei of the hypothalamic paraventricular nucleus to circumventricular organs of the rat. A Phaseolus vulgaris leucoagglutinin (PHA-L) tracing study. J Comp Neurol 306:462–479 Larsen PJ, Enquist LW, Card JP (1998) Characterization of the multisynaptic neuronal control of the rat pineal gland using viral transneuronal tracing. Eur J Neurosci 10:128–145 Leander P, Vrang N, Møller M (1998) Neuronal projections from the median and dorsal raphe nucleus to the suprachiasmatic nucleus and the deep pineal gland of the golden hamster (Mesocricetus auratus). A combined in vivo tracing and

immunohistochemical neurotransmitter study. J Comp Neurol 399:73–93 Lerner AB, Case JD, Heinzelman RV (1959) Structure of melatonin. Chem Soc 81:6084–6085 Luo ZR, Schultz RL, Whitter EF, Vollrath L (1984) The ultrastructure of the nerve fibers and pinealocytes in the rat pineal stalk. J Pineal Res 1:323–337 Matsushima S, Sakai Y, Hira Y, Oomuri Y, Daikoku S (1994) Immunohistochemical studies on sympathetic and nonsympathetic nerve fibers and neuronal cell bodies in the pineal gland of cotton rats, Sigmodon hispidus. Arch Histol Cytol 57:47–58 Matsushima S, Sakai Y, Hira Y (1999) Peptidergic peripheral nervous systems in the mammalian pineal gland. Microsc Res Tech 46:265–280 Mikkelsen JD, Møller M (1990) A direct neuronal projection from the intergeniculate leaflet of the lateral geniculate nucleus to the deep pineal gland of the rat, demonstrated with Phaseolus vulgaris leucoagglutinin (PHA-L). Brain Res 520:342–347 Mikkelsen JD, Møller M (1999) Neuropeptide Y in the mammalian pineal gland. Microsc Res Tech 46:239–256 Mikkelsen JD, Panula P, Møller M (1992) Histamine-immunoreactive nerve fibers in the rat pineal gland: evidence for a histaminergic central innervation. Brain Res 597:200–208 Mikkelsen JD, Hauser F, deLecea L, Sutcliffe JG, Kilduff TS, Calgari C, Pévet P, Simonneaux V (2001) Hypocretin (orexin) in the rat pineal gland: a central transmitter with effects on noradrenaline-induced release of melatonin. Eur J Neurosci 14:419–425 Møller M (1979) Presence of a pineal nerve (nervus pinealis) in fetal mammals. Prog Brain Res 52:103–106 Møller M (1997) Peptidergic cells in the mammalian pineal gland. Morphological indications for a paracrine regulation of the pinealocyte. Biol Cell 89:561–567 Møller M (1999) Introduction to mammalian pineal innervation. Microsc Res Tech 46:235–238 Møller M, Hay-Schmidt A (1998) Direct neuronal projection from the dorsal raphe nucleus to the pineal complex of the rat: a Phaseolus vulgaris-leucoagglutinin in vivo neuronal tracing study. J Pineal Res 25:19–23 Møller M, Korf HW (1983a) Central innervation of the pineal organ of the Mongolian gerbil. A histochemical and lesion study. Cell Tissue Res 230:259–272 Møller M, Korf HW (1983b) The origin of central pinealopetal nerve fibers in the Mongolian gerbil as demonstrated by the retrograde transport of horseradish peroxidase. Cell Tissue Res 230:273–287 Møller M, Mikkelsen JD (1989) Vasoactive intestinal peptide (VIP) and peptide histidine isoleucine (PHI) in the mammalian pineal gland. Adv Pineal Res 3:l–l0 Møller M, Veen T van (1981) Vascular permeability in the pineal gland of the Mongolian gerbil and the Djungarian hamster. In: Semm K, Mettler L (eds) Human reproduction. Intern Congr Series 551. Elsevier/North-Holland, Amsterdam, pp 539–543 Møller M, Ingild A, Bock E (1978a) Immunohistochemical demonstration of S-100 protein and GFA protein in interstitial cells of rat pineal gland. Brain Res 140:1–13 Møller M, Deurs B van, Westergård E (1978b) Vascular permeability to proteins and peptides in the mouse pineal gland. Cell Tissue Res 195:1–15 Møller M, Mikkelsen JD, Fahrenkrug J, Korf H-W (1985) The presence of vasoactive intestinal polypeptide (VIP)-likeimmunoreactive nerve fibres and VIP-receptors in the pineal gland of the Mongolian gerbil (Meriones unguiculatus). An immunohistochemical and receptor-autoradiographic study. Cell Tissue Res 241:333–340 Møller M, Phansuwan-Pujito P, Govitrapong P, Schmidt P (1993) Indications for a central innervation of the bovine pineal gland with substance P-immunoreactive nerve fibers. Brain Res 611: 347–351 Møller M, Ravault J-P, Cozzi B (1996) The chemical neuroanatomy of the mammalian pineal gland: neuropeptides. Neurochem Int 28:23–33

150 Møller M, Fahrenkrug J, Hannibal J (1999) Innervation of the rat pineal gland by pituitary adenylate cyclase-activating polypeptide (PACAP)-immunoreactive nerve fibres. Cell Tissue Res 296:247–257 Munch IC, Møller M, Larsen PJ, Vrang N (2002) Light induced c-Fos expression in SCN neurons targeting the paraventricular nucleus of the hamster hypothalamus: phase dependency and immunochemical identification. J Comp Neurol 442:48–62 Nürnberger F, Korf HW (1981) Oxytocin- and vasopressinimmunoreactive nerve fibers in the pineal gland of the hedgehog, Erinaceus europaeus L. Cell Tissue Res 220:87–97 Oksche A, Vaupel-von Harnack M (1965) Electron microscopic studies on the nerve tracts of the pineal complex of Rana esculenta L. Z Zellforsch Mikrosk Anat 68:389–426 Pedersen EB, Fox LM, Castro AJ, McNulty JA (1993) Immunocytochemical and electron-microscopic characterization of macrophage/microglia cells and expression of class II major histocompatibility complex in the pineal gland of the rat. Cell Tissue Res 272:257–265 Perentes E, Rubinstein LJ, Herman MM, Donoso LA (1986) S-antigen immunoreactivity in human pineal glands and pineal parenchymal tumors. A monoclonal antibody study. Acta Neuropathol (Berl) 71:224–227 Pfeffer M, Stehle JH (1998) Ontogeny of a diurnal rhythm in arylalkylamine-N-acetyltransferase mRNA in rat pineal gland. Neurosci Lett 248:163–166 Phansuwan-Pujito P, Mikkelsen JD, Govitrapong P, Møller M (1991) Cholinergic innervation of the bovine pineal gland visualised by immunohistochemical detection of choline acetyltransferase (ChAT)-immunoreactive nerve fibers. Brain Res 545:49–58 Przybylska-Gornowicz B, Lewczuk B, Møller M (2000) Immunohistochemical localization of substance P in the pineal gland of the domestic pig. Folia Histochem Cytobiol 38:85–90 Reuss S (1999) Trigeminal innervation of the mammalian pineal gland. Microsc Res Tech 46:305–309 Reuss S, Møller M (1986) Direct projections to the rat pineal via the stria medullaris thalami: a horseradish peroxidase anterograde tracing study. Cell Tissue Res 244:691–694 Reuss S, Schröder H (1987) Neuropeptide Y effects on pineal melatonin synthesis in the rat. Neurosci Lett 74:158–162 Ribelayga C, Garidou ML, Malan A, Gauer F, Calgari C, Pévet P, Simonneaux V (1999) Photoperiodic control of the rat pineal arylalkylamine-N-acetyltransferase and hydroxyindole-Omethyltransferase gene expression and its effect on melatonin synthesis. J Biol Rhythms 14:105–115 Romijn HJ (1975) Structure and innervation of the pineal gland of the rabbit, Oryctolagus cuniculus (L.). III. An electron microscopic investigation of the innervation. Cell Tissue Res 157:25–51 Rønnekleiv OK (1988) Distribution in the macaque pineal of nerve fibers containing immunoreactive substance P, vasopressin, oxytocin, and neurophysins. J Pineal Res 5:259–271 Rønnekleiv OK, Kelly MJ (1984) Distribution of substance P neurons in the epithalamus of the rat: and immunohistochemical investigation. J Pineal Res 1:355–370

Sakai Y, Hira Y, Matsushima S (2001) Central GABAergic innervation of the mammalian pineal gland: a light and electron microscopic immunocytochemical investigation in rodent and nonrodent species. J Comp Neurol 430:72–84 Sato T, Kaneko M, Hama A, Kusakari T, Fujieda H (1996) Expression of class II MHC molecules in the rat pineal gland during development and effects of treatment with carbon tetrachloride. Cell Tissue Res 284:65–76 Schomerus C, Ruth P, Korf HW (1994) Photoreceptor-specific proteins in the mammalian pineal organ: immunocytochemical data and functional considerations. Acta Neurobiol Exp (Warsz) 54(suppl):9–17 Schomerus C, Maronde E, Laedtke E, Korf HW (1996) Vasoactive intestinal peptide (VIP) and pituitary adenylate cyclase-activating polypeptide (PACAP) induce phosphorylation of the transcription factor CREB in subpopulations of rat pinealocytes: immunocytochemical and immunochemical evidence. Cell Tissue Res 286:305–313 Shadiack AM, Sun Y, Zigmond RE (2001) Nerve growth factor antiserum induces axotomy-like changes in neuropeptide expression in intact sympathetic and sensory neurons. J Neurosci 21:363–371 Shiotani Y, Yamano M, Shiosaka S, Emson PC, Hillyard CJ, Girgis S, MacIntyre I (1986) Distribution and origins of substance P (SP)-, calcitonin gene-related peptide (CGRP)-, vasoactive intestinal polypeptide (VIP)- and neuropeptide Y (NPY)-containing nerve fibers in the pineal gland of gerbils. Neurosci Lett 70:187–192 Simonneaux V (1995) Neuropeptides of the mammalian pineal gland. Neuroendocrinol Lett 17:115–130 Stehle J, Gall C von, Korf HW (2002) Organization of the circadian system of melatonin-deficient C57Bl and melatonin-proficient C3H mice: a comparative investigation. Cell Tissue Res (this issue) Tsumanuma I, Tanaka R, Ichikawa T, Washiyama K, Kumanishi T (2000) Demonstration of hydroxyindole-O-methyltransferase (HIOMT) mRNA expression in pineal parenchymal tumors: histochemical in situ hybridization. J Pineal Res 28:203–209 Vollrath L (1981) The pineal organ. In: Oksche A, Vollrath L (eds) Handbuch der mikroskopischen Anatomie des Menschen, vol VI/7. Springer, Berlin Heidelberg New York, pp 1–665 Vollrath L (1984) Functional anatomy of the human pineal gland. In: Reiter RJ (ed) The pineal gland. Raven Press, New York, pp 285–322 Vrang N, Larsen PJ, Møller M, Mikkelsen JD (1995) Topographical organization of the rat suprachiasmatic-paraventricular projection. J Comp Neurol 353:585–603 Welsh MG, Beitz AJ (1981) Modes of protein and peptide uptake in the pineal gland of the Mongolian gerbil: an ultrastructural study. Am J Anat 162:343–355 Yuwiler A, Klein DC, Buda M, Weller JL (1977) Adrenergic control of pineal N-acetyltransferase activity: developmental aspects. Am J Physiol 233:E141–E146 Zhang E, Mikkelsen JD, Møller M (1991) Tyrosine hydroxylaseand neuropeptide Y-immunoreactive nerve fibers in the pineal complex of untreated rats and rats following removal of the superior cervical ganglia. Cell Tissue Res 265:63–71