Cellular and genetic regulation of the development of the cerebellar system

Abstract

Recent advances in molecular biology have drastically changed our vision on the development of the nervous system, the cerebellum in particular. After a classical descriptive period, we are now in a modern mechanistic epoch as we begin to answer crucial questions in our quest to understand the mechanisms underlying the emergence of brain complexity. This review begins with an analysis of the role of the “isthmic organizer” in the induction and specification of the cerebellar territory and progresses through cerebellar development to the formation of cerebellar maps. It gathers information about the control of the proliferation of granule cell precursors by Purkinje cells and the role of Shh/Gli-patched signaling. The migratory routes for cerebellar and precerebellar neurons, together with the long-range and short-range cues guiding gliophilic and, particularly, neurophilic migrations, are also discussed. Because these cues are similar to those involved in axon guidance, both processes are under the same molecular constraints. Finally, using primarily the olivocerebellar projection as a model, the cellular and molecular mechanisms involved in the formation of cerebellar maps are discussed. During embryonic development, Purkinje cells in the cerebellum and neurons in the inferior olive follow a simultaneous, but independent, process of intrinsic parcellation, giving rise to subsets of biochemically different cortical compartments. The occurrence of positional information shared between olivary axons and their postsynaptic targets, the Purkinje cells, provides a molecular code for the formation of coarse-grained maps. Activity-dependent mechanisms are required for the transition from crude to fine-grained maps. This important refinement, which confers ultimate specificity to the maps, is under the regulation of parallel fiber–Purkinje cell synaptic activity.

Introduction

The cerebellum, especially the cerebellar cortex, is one of the central regions in which ordered organizational patterns are most obvious. Its apparent simplicity and geometrical disposition have attracted many investigators to its potential for providing an understanding of the mechanisms involved in the development of the nervous system. In the past, the study of cerebellar development allowed Cajal (1890) to obtain the first direct proof of the independence of nerve cells, leading to the formulation of the “Neuron Theory.” Cajal also accurately described for the first time the progression of neuronal migration and neuronal differentiation in the cerebellum, as well as the role of regressive processes during the formation of specific neuronal networks. From these and other studies, it has been concluded that the precise spatial positioning of neurons–perceived as a static arrangement in the adult animal–is the result of a strict temporal organization during development. Neuronal proliferation and migration, together with neuronal differentiation and oriented axon growth, proceed according to strictly defined kinetics, resulting in the establishment of complex neuronal networks with a high degree of specificity.

Most of the data discussed here are centered on the work done in my laboratory, with or without my personal contribution, during the last 15 years. Thus, after a short anatomical reminder of the organization of the adult cerebellum, its development will be presented following an ideal pace. I will start with the specification of the region of the neural tube at the origin of the cerebellum and end with the refinement of neural maps in the cerebellar cortex.

The cortex of the cerebellum is the region of the brain in which the most precise correlation between structure and function has been established. This broad zone, with its important role in the control of motor coordination, is composed of only eight different neuronal populations: six classical ones, already described by Cajal (1911), Purkinje cells, Golgi cells, Lugaro cells, granule cells, basket cells, stellate cells, and two more recently reported ones, unipolar brush cells and candelabrum cells (Laine and Axelrad, 1994, Mugnaini and Floris, 1994). With the exception of unipolar brush cells that are much more numerous in ventral than dorsal lobules (Hockfield, 1987, Mugnaini and Floris, 1994), the others repeat their specific pattern in an almost monotonous manner all along the numerous folia composing the cortex. This characteristic neuronal arrangement consists of a strict positioning of neurons and afferent fibers, conferring to the cortex a stereotyped three-dimensional geometry, which has been very helpful in determining the properties of most of these types of neurons and understanding its intrinsic connectivity (Ito, 1984).

The fundamental anatomical organization of the cerebellar circuit was established more than a century ago by Cajal, 1888a, Cajal, 1888b, whose detailed description has been only slightly modified thereafter. In adult mammals and birds, the cortex of the cerebellum contains three main layers. The most superficial one is the molecular layer, a region of low cellular density but of high synaptic incidence. The deepest layer exhibits, on the contrary, the highest cellular density of any central structure; it contains the granule cells and is called the granular layer. At the interface between these two layers, a thin middle zone in which the cell bodies of the Purkinje cells are aligned into a single row can be found; this zone is the Purkinje cell layer. From the eight classes of neurons encountered in the cerebellar cortex only one, the Purkinje cell, has an axon which projects outside the cortex. The other seven populations are local circuit neurons (granule and unipolar brush cells are glutamatergic, and the five others GABAergic), and their axons never reach extracortical structures (Cajal, 1911, Laine and Axelrad, 1994, Laine and Axelrad, 1996, Laine and Axelrad, 2002, Mugnaini and Floris, 1994, Nunzi et al., 2001). Purkinje cells are, therefore, the pivotal elements around which all the cerebellar circuit is organized, receiving extracortical synaptic information and, after processing it, channeling towards cortical efferent pathways, the deep cerebellar nuclei and the brain stem nuclei, mainly but not exclusively the vestibular nuclei (De Camilli et al., 1984).

The two main afferent systems conveying information to the cerebellar cortex are the climbing and mossy fiber systems, which direct their impulses to the single efferent system, the Purkinje cells. The way the two systems reach these neurons is very different. Only climbing fibers directly contact Purkinje cells. These fibers originate from the inferior olivary nucleus (Desclin, 1974, Sotelo et al., 1975) and establish extensive synaptic contacts with Purkinje cell dendrites (dendritic spines in the proximal dendritic compartment). The climbing fiber/Purkinje cell synapse is unique in the CNS because of the perfect numerical matching between both partners; there is only one climbing fiber per Purkinje cell (Eccles et al., 1966, Cajal, 1911). The mossy fibers constitute a composite population of afferent fibers arising from various nuclei in the spinal cord (Matsushita et al., 1979), in the brain stem (Gould, 1980), and even in the deep cerebellar nuclei (Gould, 1979, Tolbert et al., 1978). They reach the Purkinje cells indirectly through relay cells, the granule cells. The axons of these interneurons, called parallel fibers, extend in the molecular layer, perpendicularly oriented to the plane of the flattened Purkinje cell dendrites. At each intersection, the parallel fiber is provided with a varicosity, which represents a bouton “en passant” synapsing on the Purkinje cell dendrite (dendritic spines in the distal dendritic compartment). Thus, contrary to the exquisite convergence of inputs of the climbing fiber system, the mossy fiber—granule cell—Purkinje cell system shows a great degree of divergence. Through hundreds of parallel fibers, one mossy fiber can affect thousands of Purkinje cells. Superimposed on this excitatory network are the inhibitory interneurons, disposed either in the molecular (stellate, basket cells), in the Purkinje cell layer (candelabrum cells) or in the granular layer (Lugaro, and Golgi cells). All these inhibitory interneurons have all or most of their dendrites in the molecular layer, where they receive synapses from the parallel fibers.