MICROSCOPY RESEARCH AND TECHNIQUE 52:520–533 (2001) Neural Input and Neural Control of the Subcommissural Organ 1 2 1 ANTONIO J. JIME´ NEZ, * PEDRO FERNA´ NDEZ-LLEBREZ, AND JOSE MANUEL PE´ REZ-FI´GARES 1Departamento de Biologı´a Celular y Gene´tica, Facultad de Ciencias, Universidad de Ma´laga, Ma´laga, Spain 2Departamento de Biologı´a Animal, Facultad de Ciencias, Universidad de Ma´laga, Ma´laga, Spain KEY WORDS serotonin; GABA; monoamines; pineal organ ABSTRACT The neural control of the subcommissural organ (SCO) has been partially character- ized. The best known input is an important serotonergic innervation in the SCO of several mammals. In the rat, this innervation comes from raphe nuclei and appears to exert an inhibitory effect on the SCO activity. A GABAergic innervation has also been shown in the SCO of the rat and frog Rana perezi. In the rat, GABA and the enzyme glutamate decarboxylase are involved in the SCO innervation. GABA is taken up by some secretory ependymocytes and nerve terminals, coexisting with serotonin in a population of synaptic terminals. Dopamine, noradrenaline, and different neuropeptides such as LH- RH, vasopressin, vasotocin, oxytocin, mesotocin, substance P, ␣-neoendorphin, and galanin are also involved in SCO innervation. In the bovine SCO, an important number of fibers containing tyrosine hydroxylase are present, indicating that in this species dopamine and/or noradrenaline-containing fibers are an important neural input. In Rana perezi, a GABAergic innervation of pineal origin could explain the influence of light on the SCO secretory activity in frogs. A general conclusion is that the SCO cells receive neural inputs from different neurotransmitter systems. In addition, the possibility that neurotransmitters and neuropeptides present in the cerebrospinal fluid may also affect the SCO activity, is discussed. Microsc. Res. Tech. 52:520–533, 2001. © 2001 Wiley-Liss, Inc. INTRODUCTION 5,6-dyhydroxytriptamine or 5,7-dyhydroxytriptamine The subcommissural organ (SCO) is an ependymal (5,6-DHT; 5,7-DHT; Møllgard et al., 1978; Møllgard structure that covers and penetrates the posterior com- and Wiklund, 1979) have convincingly demonstrated missure (Rodrı´guez et al., 1987, 1992, 1998). The SCO the innervation of the SCO by serotonergic fibers. cells are arranged into two layers, ependyma and hy- In the rat SCO, serotonin is located in well-differen- pendyma. tiated axoglandular synapses with both ependymocytes The secretory activity of this brain gland, as other and hypendymocytes (Bouchaud and Arluison, 1977; body glands, is expected to be controlled by means of Bouchaud, 1979, 1993; Møllgard et al., 1978; Møllgard molecular signals. Two general ways of control for a and Wiklund, 1979). Synaptic contacts are on the basal gland have been classically considered: the humoral processes and laterobasal portions of the ependymal and the neural inputs. A confirmed way of control of the secretory cells. Bouchaud (1979), using high-resolution SCO is by synaptic contacts using neurotransmitters or radioautography in the rat, concluded that the seroto- neuropeptides. Although the SCO has a rich vascular- nergic terminals represent approximately 75% of the ization, it presents a blood-brain barrier that restricts synaptic contacts. Therefore, the remaining 25% of the access of blood factors (Bouchaud, 1975; Krisch, non-labeled synaptic contacts contain another neuro- 1993; Porier et al., 1983; Rodrı´guez et al., 1998). This transmitter. Mo¨llgard and Wiklund (1979), Wiklund gland presents another humoral domain with the in- and Møllgard (1979), Bouchaud (1979; 1993) and ternal (ventricular) and external (subarachnoidal) ce- Bouchaud and Bosler (1986) agree with respect to the rebrospinal fluid (CSF) compartments. Thus, the pos- ultrastructural characteristics of the serotonergic syn- sibility that the SCO receives signals from the CSF has aptic contacts in the rat SCO. Synaptic contacts dis- to be considered. played clear vesicles, 40–55 nm in diameter, frequently pleomorphic and less than 1% of large granular vesi- SEROTONERGIC INNERVATION cles. On the other hand, the non-serotonergic terminals OF THE SUBCOMMISSURAL ORGAN contained large granular vesicles and clear vesicles The first evidence of a serotonergic innervation of the with an average diameter of 50 nm, occasionally dis- SCO comes from histochemical studies carried out in playing a paracrystaline arrangement (Bouchaud, rats using formaldehyde-induced fluorescence that 1979, 1993; Bouchaud and Bosler, 1986). The postsyn- showed a dense serotonergic plexus (Bjo¨rklund et al., aptic ependymal cytoplasm showed electron-dense ma- terial attached to the postsynaptic membrane, confer- 1972; Bouchaud and Arluison, 1977; Fuxe, 1965; Møll- Ј gard and Wiklund, 1979; Wiklund, 1974). Such a mas- ring to it the aspect of an asymmetrical Gray s Type-I sive innervation was later confirmed by means of im- munocytochemical studies (Matsuura et al., 1989; Ueda et al., 1988). Ultrastructural radioautography Contract grant sponsor: DGICYT; Contract grant number: PB96-0696; Con- using 3H-serotonin (Bouchaud and Arluison, 1977; tract grant sponsor: FIS; Contract grant number: 98/1508. Bouchaud, 1979) and the ultrastructural demonstra- *Correspondence to: Antonio J. Jime´nez. Departamento de Biologı´a Celular y Gene´tica, Facultad de Ciencias, Universidad de Ma´laga, Ma´laga, Spain, tion of degeneration of nerve terminals after intraven- E-29071. E-mail: [email protected] tricular administration of neurotoxic drugs such as Received 3 March 2000; Accepted 26 May 2000 © 2001 WILEY-LISS, INC. INNERVATION OF THE SCO 521 synapse (Bouchaud and Bosler, 1986; Bouchaud, 1993; comes from the nuclei raphe centralis superior and Wiklund and Møllgard, 1979). raphe dorsalis, each nucleus contributing about one- third of the input (Le´ger et al., 1983). These experi- Comparative Studies of the Serotonergic ments also describe that the nucleus raphe centralis Innervation of the Subcommissural Organ superior contributes with thin fibers that innervate the The studies of the serotonergic innervation in differ- nuclear level of the ependymal layer of the SCO, ent mammalian species have resultes in contradictory whereas the nucleus raphe dorsalis contributes with results. Matsuura et al. (1989) distinguished three thicker fibers innervating the hypendymal cells. Later, morphological types of serotonergic innervation. Ac- a study using the neuronal tracer Phaseolus vulgaris- cording to these authors, type I is observed in the rat leucoagglutinin (PHA-L) applied in different raphe nu- and the Japanese squirrel, it is characterized by a clei confirmed the origin from the nucleus raphe dor- dense serotonergic plexus located at the basal portion salis (Mikkelsen et al., 1997). The innervation of the of the SCO ependymal cells (Fig. 1). In type II, the rat SCO by serotonergic axon terminals has also been innervation is distributed among the cells of the SCO partially prevented by X ray-irradiation on the brain- ependyma and can be found in the dog and cat. In the stem during the early postnatal period (Delhaye- dog, this innervation is more pronounced in the rostral Bouchaud and Bouchaud, 1993). portion of the SCO, and terminals resembling those According to Marcinkiewicz and Bouchaud (1986) described in the rat can be ultrastructurally observed and Bouchaud and Bosler (1986), the rat SCO is (Matsuura and Sano, 1987). Type III corresponds to the reached by the serotonergic innervation in a rostrocau- absence of such an innervation, and it has been de- dal direction from the lamina intercalaris. Mikkelsen scribed in the monkey, guinea pig, and adult mouse. et al. (1997) have reported, using PHA-L in the rat, Wiklund et al. (1977) have also observed that the SCO that the ascending projections from nucleus raphe dor- of the Mongolian gerbil and rabbit lacks of a serotoner- salis extended mainly within or along the ependyma of gic innervation. In 20-day-old mice of the C57BL/10J the dorsal part of the aqueduct directly toward the SCO strain we have found, using immunocytochemistry region. Finally, labeled fibers reach the SCO from the against serotonin at the light-microscopic level, the ependyma in its lateral portions and from the posterior existence of a dense plexus of labeled fibers in close commissure. contact with the SCO, displaying a similar pattern to the serotonergic innervation of the rat (Figs. 2–4). We Deafferentation and Reinnervation of the have also detected an important number of immunore- Subcommissural Organ active fibers contacting the basal portion of the bovine The effect of the serotonergic innervation has been secretory SCO ependymocytes (Fig. 5). Some of the studied in the rat by means of selective destruction of labeled fibers penetrated into the ependymal layer serotonergic fibers using neurotoxic indolamines such near the apical border. as 5,6-DHT (5,6-dihydroxytryptamine) or 5,7-DHT In non-mammal species such as the frog Rana perezi (5,7-dihydroxytryptamine). Mo¨llgard and Wiklund (Jime´nez et al., 2000) or the goldfish Carassius auratus (1979) carried out such a destruction and found that (Pe´rez-Fı´gares et al., 1993, Jime´nez et al., 1993) a after the first 3 days, ultrastructural changes were serotonergic innervation has not been detected. present in synaptic profiles innervating the SCO. Møll- gard and collaborators (Møllgard et al., 1978; Møllgard Ontogenetic Development of the and Wiklund, 1979) described a sequence of changes in Serotonergic Innervation the rat SCO ultrastructure