Y. Smith et al., SYNAPTIC INNERVATION OF NEURONS IN THE INTERNAL PALLIDAL SEGMENT BY THE SUBTHALAMIC NUCLEUS AND THE EXTERNAL PALLIDUM IN MONKEYS, Journal of comparative neurology, 343(2), 1994, pp. 297-318
In order to better understand the way by which the subthalamic nucleus
interacts with the globus pallidus to control the output of the basal
ganglia, we carried out a series of experiments to investigate the pa
ttern of synaptic innervation of the pallidal neurones by the subthala
mic terminals in the squirrel monkey. To address this problem we used
the anterograde transport of biocytin. Following injections of biocyti
n in the subthalamic nucleus, rich plexuses of labelled fibres and var
icosities formed bands that lay along the medullary lamina in both seg
ments of the ipsilateral pallidum. At the electron microscopic level,
two populations of biocytin-containing terminals were identified in th
e internal pallidum (GPi). A first group of small to medium-sized term
inals (type 1; mean cross-sectional area +/- S.D. = 0.41 +/- 0.04 mu m
(2)) contained round vesicles and formed asymmetric synapses with dend
ritic shafts (95%) of mixed sizes (maximum diameter ranging from 0.3 t
o 4.0 mu m) and spine-like structures (5%). The second group of termin
als (type 2) contained pleiomorphic vesicles, had a larger cross-secti
onal area (mean +/- S.D. = 0.9 +/- 0.4 mu m(2)) and formed symmetric s
ynapses predominantly with perikarya (41%) and large dendrites (57%).
In some cases, the two types of terminals converged at the level of si
ngle GPi neurones. Postembedding immunogold method revealed that the t
ype 2 terminals displayed gamma-aminobutyric acid (GABA) immunoreactiv
ity, whereas the type 1 terminals did not. In the external pallidum (G
Pe), injections in the subthalamic nucleus labelled both type 1 or typ
e 2 terminals. However, the labelled type 2 boutons were much less abu
ndant in GPe than in GPi. The presence of biocytin-labelled perikarya
in GPe and the fact that the type 2 terminals displayed GABA immunorea
ctivity led us to suspect that these terminals were derived from axons
of GPe neurones. In agreement with this hypothesis, injections of Pha
seolus vulgaris-leucoagglutinin (PHA-L) in GPe labelled terminals in G
Pi that displayed the morphological features and a pattern of synaptic
organization similar to the type 2 terminals. In conclusion, the resu
lts of our study demonstrate that the subthalamopallidal terminals for
m asymmetric synapses that are distributed along the dendritic tree of
GPe and GPi neurones. In contrast, the GPe projection to GPi gives ri
se to large GABA-containing terminals that form symmetric synapses pre
dominantly with the proximal region of pallidal neurones. Because the
GABAergic axon terminals from GPe form synapses onto the perikarya and
proximal dendrites of GPi neurones, the GPe input is in a strategic p
osition to reduce the excitatory influence generated more distally on
the dendritic tree by the subthalamic nucleus. (C) 1994 Wiley-Liss, In
c.