Search

Glial cells associated with elements of central neuropils in several insect species were studied with conventional light and electron microscopical techniques, the Golgi procedure, and a combination of the latter with electron microscopy. Different types of cells located in the layer of cells covering the neuropil were found to send complex arborizations into synaptic neuropils. These arborizations grow in clusters that seem to represent discrete compartments circumscribing groups of synaptic terminals. The thinnest glial processes are found deep in the neuropil and consist of compact membrane leaflets lacking cell organelles and with reduced amounts of cytoplasmic matrix. Some of these glial processes also invest neuropil tracheoles in a manner reminiscent of the association between astrocyte end-feet and blood capillaries in the central nervous system of mammals. Other glial cells reside completely in the neuropil, where they enwrap fiber bundles in a similar manner to oligodendrocytes in the central nervous system of mammals.

Odor information is processed in spherical structures called glomeruli, which in all animals with differentiated olfactory systems are sites of densely spaced synaptic contacts between olfactory sensory axons and target central nervous system (CNS) neurons. Glomerulus development in the antennal (olfactory) lobe of the moth brain, which is initiated by the arrival of antennal receptor axons, requires interaction among three elements: glial cells, receptor axons, and their targets, the antennal-lobe neurons. Receptor axons form an array of protoglomeruli that become surrounded by glia and serve as a template for mature glomeruli. Previous experiments showed that when the number of glial cells is sharply reduced during development either by irradiation or by an anti-mitotic agent, receptor axons form protoglomeruli, but in the mature lobes, glomeruli are absent and central neurons lack the characteristic glomerular tufting of their arbors. The current investigation was conducted to determine which cellular events in the process of glomerulus formation are disrupted by severe reduction in glial-cell number. The branching patterns of receptor axons and antennal-lobe neurons were examined in animals that had been irradiated to produce glia-deficient antennal lobes at stages during which glomeruli normally would develop. We found that the receptor axons did form protoglomeruli, but that the protoglomeruli quickly disintegrated in glia-deficient antennal lobes; the receptor axons branched diffusely, except where several neighboring glia survived irradiation and together formed a wall of processes that appeared to block the passage of neuronal processes. Multi-glomerular antennal-lobe neurons never developed tufted arbors even at early stages. These results suggest that maintenance of protoglomeruli depends on the border of glia that forms around each protoglomerulus and that the subsequent tufting of antennal-lobe neurons depends on maintenance of the protoglomerular template during the period of dendritic growth.

The antennal system of the moth Manduca sexta is a useful model for studies of the development of olfactory glomeruli, the complex synaptic structures that typically underlie the initial processing of olfactory input in vertebrates and invertebrates. In this review, we summarize cellular events in the construction of glomeruli in Manduca and highlight experiments that reveal factors that influence glomerulus development. By methodically manipulating each of various cell types, both neuronal and glial, that contribute to glomerular architecture, we have found that: olfactory receptor axons lay a template for developing glomeruli, stabilization of the template by glial cells is necessary to permit subsequent steps in development of the glomeruli, and the hormone that regulates adult development causes production of adequate numbers of glial cells. Neither electrical activity nor the presence of a serotonin-containing neuron that persists throughout development is required for a glomerular pattern to develop; these factors might, however, influence the synaptic organization of individual glomeruli.

In Manduca sexta, the accessory planta retractor muscle (APRM), which retracts the larval proleg, is innervated by two excitatory motoneurons, the accessory planta retractor motoneurons (APRs). These muscles and motoneurons have been the focus of a number of developmental and behavioral studies. The present study investigated properties of the pair of APRs that innervate each APRM and determined their pattern of innervation of APRM fibers. Members of APR pairs could not be distinguished by their anatomical or electrical properties (resting membrane potential, input resistance and spike threshold). Spontaneous synaptic inputs to members of APR pairs were highly correlated, whereas spontaneous synaptic inputs to APRs and functionally dissimilar motoneurons were not well correlated. Synaptic inputs from identified mechanosensory neurons and interneurons to the two APRs were qualitatively similar, but the magnitude of the response to sensory stimulation sometimes differed within a pair. Both APRs produced large, rapidly rising excitatory junction potentials in APRM fibers. Within the APRM, some fibers were singly innervated by one or the other APR while the remaining fibers were dually innervated by both APRs. In dually innervated fibers, the motor terminals of the two APRs were spatially segregated. This innervation pattern appears to be unique among insects and shares some properties with the innervation of vertebrate muscle.

Octopod (Octo) is a mutation of the moth Manduca sexta, which transforms the first abdominal segment (A1) in the anterior direction. Mutant animals are characterized by the appearance of homeotic thoracic-like legs on A1. We exploited this mutation to determine what rules might be used in specifying the fates of sensory neurons located on the body surface of larval Manduca. Mechanical stimulation of homeotic leg sensilla did not cause reflexive movements of the homeotic legs, but elicited responses similar to those observed following stimulation of ventral A1 body wall hairs. Intracellular recordings demonstrated that several of the motoneurons in the A1 ganglion received inputs from the homeotic sensory hairs. The responses of these motoneurons to stimulation of homeotic sensilla resembled their responses to stimulation of ventral body wall sensilla. Cobalt fills revealed that the mutation transformed the segmental projection pattern of only the sensory neurons located on the ventral surface of A1, resulting in a greater number with intersegmental projection patterns typical of sensory neurons found on the thoracic body wall. Many of the sensory neurons on the homeotic legs had intersegmental projection patterns typical of abdominal sensory neurons: an anteriorly directed projection terminating in the third thoracic ganglion (T3). Once this projection reached T3, however, it mimicked the projections of the thoracic leg sensory neurons. These results demonstrate that the same rules are not used in the establishment of the intersegmental and leg-specific projection patterns. Segmental identity influences the intersegmental projection pattern of the sensory neurons of Manduca, whereas the leg-specific projections are consistent with a role for positional information in determining their pattern.

The central projections of olfactory receptor cells associated with two distinct types of antennal sensilla in the sphinx moth Manduca sexta were revealed by anterograde staining. In both sexes, receptor axons that arise from sexually isomorphic, type-II trichoid sensilla (and possibly some basiconic sensilla) project to the spheroidal glomeruli in the ipsilateral antennal lobe. Each axon terminates in one glomerulus. Axons from a limited region of the antenna project to glomeruli throughout the lobe, arguing against strict topographic mapping of antennal receptor cells onto the array of glomeruli. Axons of sex-pheromone-selective receptor cells in the male-specific type-I trichoid sensilla project exclusively to the sexually dimorphic macroglomerular complex (MGC). Axons from sensilla on the dorsal surface of the antenna are biased toward the medial MGC and those from ventral sensilla, toward the lateral MGC. Some receptor-cell axons branch before reaching the MGC, but their terminals are always confined to one of the two main glomerular divisions of the MGC, the cumulus and toroid. These findings confirm that primary-afferent information about pheromonal and non-pheromonal odors is segregated in the antennal lobe and suggest that there is a functional correspondence between particular olfactory receptor cells and specific glomeruli.