Evidence for glutamatergic tectotectal neurons in the cat superior colliculus: a comparison with GABAergic tectotectal neurons E. Olivier, J. Corvisier, 1 Q. Pauluis and O. Hardy 1 Laboratory of Neurophysiology, School of Medicine, Catholic University of Louvain (UCL), Brussels, Belgium 1 Laboratoire de Neurobiologie des Re Âseaux Sensorimoteurs, CNRS, Universite Âs Paris 5 & 7, Paris, France Keywords: commissure, GABA, glutamate, orienting movements, saccade Abstract The tectotectal commissural pathway is commonly regarded as responsible for the reciprocal inhibition that takes place between the two superior colliculi (SC). Although this hypothesis has received strong support from electrophysiological studies, more recent investigations have suggested that some collicular cells, e.g. ®xation neurons, may establish excitatory connections with cells in the contralateral SC through the collicular commissure. The goal of the present study was to seek immunohistochemical evidence for glutamatergic tectotectal cells in the cat SC by using a double-labelling technique. Tectotectal cells were retrogradely labelled with wheat germ agglutinin (WGA) ±horseradish peroxidase (HRP) coupled to colloidal gold injected in the contralateral SC, and neurons containing glutamate or g-aminobutyric acid (GABA) were then identi®ed with immunohistochemical techniques. The present study providesevidencethat,inthecatSC,equalnumbersoftectotectalcellsareimmunopositivetoglutamateandGABA,suggestingthat the tectotectal pathway may consist of two distinct functional components. The ®nding that an equal number of tectotectal cells are GABAergic and glutamatergic is somewhat surprising as electrophysiological studies have invariantly indicated that the inhibitory component of the tectotectal projection predominates. Another striking feature of the GABAergic and glutamatergic tectotectal cell populations is their identical topographic distribution in the SC. These results suggest that not only cells in the rostral ®xation zone establish excitatory connections with the contralateral SC. Tectotectal projections could be potentially important to shape the spatial pattern of saccade-related activity that may occur simultaneously in the two SC during vertical and oblique orienting movements. Introduction Since the description of the Sprague effect (Sprague, 1966) in which a hemianopsia resulting from an ablation of one occipital lobe can be abolished by sectioning the collicular commissure, tectotectal projections are ordinarily regarded as responsible for mediating reciprocal inhibition between the two superior colliculi (SC). This hypothesis has received strong support from electrophysiological studies that con®rmed the inhibitory nature of the commissural tectotectal projection. Recordings of either local ®eld potentials or single cell activity in the collicular super®cial layers showed that visual responses are inhibited by stimulation of the contralateral SC in pigeons (Robert & Cue Ânod, 1969), rats (Goodale, 1973) and cats (Hoffmann & Straschill, 1971; Mascetti & Arriagada, 1981). The same pattern of mutual inhibition also applies for output collicular cells discharging for eye saccades because they were found inhibited during ipsiversive orienting movements (Infante & Leiva, 1986; Peck, 1990). More recently, Munoz & Istvan (1998) reported that, in the monkey SC, the activity of saccade-related cells can be inhibited by electrically stimulating the `saccadic zone' in the contralateral SC. The existence of direct inhibitory connections between the two SC was con®rmed by intracellular recordings of inhibitory postsynaptic potentials (IPSPs) from collicular cells in response to electrical stimulation of the contralateral SC; these IPSPs had a latency consistent with a monosynaptic inhibitory commissural connection (Maeda etal., 1981; Moschovakis & Karabelas, 1982). In agreement with these observations, Appell & Behan (1990) found that, in the cat, commissural cells constitute a major source of GABAergic (g- aminobutyric acid) projection to the contralateral SC. However, Behan (1985) showed that tectotectal synaptic terminals are not homogeneous, suggesting that all of them may not be inhibitory, and more recent electrophysiological studies have con®rmed this view. In monkeys, Moschovakis etal. (1988) found that a large proportion of saccade-related burst neurons ± presumably excitatory±participate in both the tectotectal pathway and the predorsal bundle; a similar pattern of connections was reported for efferent collicular cells in cats (Moschovakis & Karabelas, 1985; Olivier & May, 1996). Further evidence supporting the existence of excitatory tectotectal projections comes from the ®nding that cells in the ®xation zone of the two SC are likely to be coupled together via mutual excitatory connections (Munoz & Guitton, 1991; Munoz & Wurtz, 1993). This hypothesis was recently con®rmed by stimulating electrically one ®xation zone while recording simultaneously the activity of contralateral ®xation cells; the relatively short-latency facilitation (1±2 ms) observed in the discharge of these cells suggests that the two ®xation zones are coupled via commissural mono- synaptic excitatory projections (Munoz & Istvan, 1998). In light of evidence for heterogeneous functions of the tectotectal pathway, it appears important to determine whether some tectotectal Correspondence: Professor E. Olivier, as above. E-mail: olivier@nefy.ucl.ac.be Received 23 December 1999, revised 23 March 2000, accepted 3 April 2000 European Journal of Neuroscience, Vol. 12, pp. 2354±2366, 2000 Ó Federation of European Neuroscience Societies