79 In many parts of the vertebrate central nervous system, inputs of distinct types confine their synapses to individual laminae. Such laminar specificity is a major determinant of synaptic specificity. Recent studies of several laminated structures have begun to identify some of the cells (such as guidepost neurons in hippocampus), molecules (such as N-cadherin in optic tectum, semaphorin/collapsin in spinal cord, and ephrins in cerebral cortex), and mechanisms (such as activity-dependent refinement in lateral geniculate) that combine to generate laminar specificity. Addresses Department of Anatomy and Neurobiology, Washington University Medical School, 660 South Euclid Avenue, Campus Box 8108, St Louis, Missouri 63110, USA *e-mail: sanesj@thalamus.wustl.edu Current Opinion in Neurobiology 1999, 9:79–87 http://biomednet.com/elecref/0959438800900079 © Elsevier Science Ltd ISSN 0959-4388 Abbreviations GABA γ-aminobutyric acid NGF nerve growth factor NT neurotrophin VVA Vicia villosa agglutinin-B 4 Introduction Many parts of the vertebrate central nervous system are divided into histologically discrete parallel laminae. Each lamina bears a distinct complement of neuronal subtypes, distinguishable by morphology, molecular composition, projections and, in some cases, developmental histories. In addition, and of particular importance here, distinct popu- lations of afferent axons confine their terminal arbors and synapses to different subsets of laminae. This laminar specificity of synaptic connections is so striking and wide- spread that it appears to be a major determinant of specific connectivity in the central nervous system. Nonetheless, in contrast to a long-standing interest in the related issue of how laminae form (see [1]), determinants of lamina-spe- cific connectivity have been relatively little studied. This situation is now changing, however, and recent analyses of several systems have begun to provide insights into the cellular and molecular bases of laminar specificity. Here, we review studies of five laminated structures — hip- pocampus, optic tectum, lateral geniculate nucleus, spinal cord and cerebral cortex — and describe general insights that can be derived from them. General principles of lamina-specific synaptic connections Lamina-specific connectivity may arise in any of several ways (Figure 1). We thought it would be helpful to begin with a series of questions applicable to all systems. These provide a framework for considering the experimental findings summarized in subsequent sections. First, what are the synaptic targets of the axons that form lamina-specific connections? In some cases, axons synapse onto cells that are themselves confined to particular lami- nae. In other cases, axons synapse on lamina-restricted segments of dendrites that extend through many laminae. We refer to these two phenomena as cellular and subcellu- lar specificity, respectively. Second, where are the cues that instruct afferents to arborize or form synapses in specific laminae? In cases of cellular specificity, recognition molecules might be pre- sent on the surface of lamina-restricted neurons, and arise as part of the program that allocates cells to lami- nae. Subcellular specificity might reflect the presence of molecules that are transported to or secreted from dis- tinct domains within dendritic arbors. Alternatively, cues might be immobilized within the extracellular matrix or be present on lamina-restricted cells that are not defini- tive synaptic targets, such as glia [2] or guidepost cells (see below). The localization of the cues may give hints as to their functional significance; for example, cues on synaptic targets could promote lamina-specific synapse formation directly. Alternatively, if cues are present in non-neuronal cells or extracellular matrix, the instruction an axon receives might be something like ‘arborize with- in this lamina, then synapse on whatever receptive surfaces are present where you arborize.’ Third, are connections specified entirely by invariant mol- ecular cues or are they shaped by patterned electrical activity? This issue is a fundamental one for neural con- nections of all sorts, and clearly applicable to laminar specificity. A seemingly related issue is whether connec- tions are lamina-specific from the outset or refined from a diffuse pattern as development proceeds. Often, precise connections reflect ‘hard-wiring’, whereas progressive refinement results from an influence of activity. These cor- relations, however, are not absolute; gradual refinement could reflect a progressive restriction of molecular cues, and initial precision could mean that the system is already appropriately active as synapses form. Finally, what molecules determine lamina-specific connec- tions? Over the past few years, numerous proteins have been identified that regulate axonal behavior; they include membrane, matrix and soluble molecules that act at short and long ranges to promote or inhibit neurite outgrowth. Some molecules already well studied in other contexts may also guide lamina-specific axonal behaviors, perhaps in collaboration with other molecules yet to be discovered. Lamina-specific connections in the hippocampus Pyramidal neurons of the hippocampus have somata in a basal layer, the stratum pyramidale, and extend dendrites Formation of lamina-specific synaptic connections Joshua R Sanes* and Masahito Yamagata