A new in vitro model for agonist-induced communication between microvascular endothelial cells.

Microvascular endothelial cells (MECs) grown in Matrigel form capillary-like structures. We hypothesized that these "capillaries" better mimic communication properties of microvessels than conventional cell monolayers. MECs were isolated from the rat hindlimb skeletal muscle. Functional communication was tested by visualizing the spread of microinjected 6-carboxyfluorescein (CF) dye and by measuring a conducted change of membrane potential after micropipette application of 500 mM KCl or 10 mM adenosine triphosphate (ATP) on the capillary and monolayer. MECs grown under both conditions were dye-coupled, as demonstrated by the spread of CF injected into a single cell. The membrane potential of cells grown in capillaries (-59 +/- 5 mV) was significantly greater than that of cells grown in monolayers (-24 +/- 2 mV). KCl and ATP caused local depolarization (18 +/- 3 mV) and hyperpolarization (21 +/- 3 mV) in capillaries that yielded conducted 13 +/- 3 mV depolarization and 15 +/- 5 mV hyperpolarization at a 300-microm distal site, respectively. In monolayers, local and distal responses to agonists were 3- to 6-fold and 9- to 10-fold less, respectively, than the corresponding responses in capillaries. Cells grown under both conditions expressed connexin 43, as demonstrated by immunohistochemistry and Western blotting. We conclude that cells grown in capillaries yield substantially larger local and communicated responses than cells in monolayers and thus offer a more sensitive model for mechanistic studies of MEC communication.

Subtractive cloning of complementary DNAs and analysis of messenger RNAs with regional heterogeneous distributions in primate cortex.

Morphological heterogeneity of individual neurons in the mammalian brain must ultimately result from differences between cells in their profiles of gene expression. However, the degree to which neurons located in different regions of the brain express different sets of genes is not known. Using differential cDNA cloning procedures, including subtractive hybridization and differential colony screening, a quantitative analysis of RNAs with heterogeneous distributions in the telencephalon of the rat and Old World monkey has been performed. The results suggest that no species of RNA exist with a distribution specific to the hippocampus or neocortex of the rat with an abundance greater than 0.05%. Additionally, the results with an abundance of 0.05% or greater which is present in neocortex of the monkey but absent from the cerebellum suggest that only one species of mRNA exists. A Northern blot analysis of cDNA clones representing mRNAs present in the neocortex but absent from the cerebellum of the monkey is presented, establishing the effectiveness of subtractive hybridization between different regions of the brain for cDNA cloning. Also presented is in situ hybridization analysis of the cellular distributions in primate neocortex of two mRNAs. One of these RNAs, detected by clone 1B4 is present in cortical laminae V and VI, and exhibits a high degree of heterogeneity in the overall density of labelled cells.

Ultrastructural study of remodeled rubral afferents following neonatal lesions in the rat.

Following neonatal hemicerebellectomy, an aberrant ipsilateral cerebellorubral projection develops that maintains the topographic specificity of the normal contralateral projection. Similarly, neonatal lesions of the sensorimotor cortex lead to the appearance of an aberrant contralateral corticorubral projection that mirrors the topographic specificity of the normal ipsilateral input. The specificity of synaptic localization in these aberrant projections was studied by use of ultrastructural visualization of anterogradely transported HRP-WGA. Following neonatal ablations, adults received HRP-WGA injections in the unablated deep cerebellar nuclei or sensorimotor cortex. After 48 hours, animals were sacrificed and processed for ultrastructural localization of anterogradely transported HRP-WGA. In hemicerebellectomized animals, both the contralateral and ipsilateral interpositorubral projections terminated on the somatic and proximal dendritic membrane of magnocellular neurons. Some of these labeled synaptic terminals were located on somatic and dendritic spines. Following HRP-WGA injection in the unablated nucleus lateralis, anterogradely labeled synaptic terminals were located bilaterally on small- to medium-sized dendrites of parvicellular neurons. Injection of HRP-WGA in the remaining sensorimotor cortex of animals that had undergone neonatal unilateral ablation of the sensorimotor cortex resulted in labeled corticorubral synaptic terminals that contacted distal dendrites of ipsilateral and contralateral parvicellular neurons. These results demonstrate that, following neonatal deafferentation of the rat red nucleus, the topographic specificity of the aberrant rubral afferents is accompanied by a specificity of synaptic localization on discrete membrane areas of rubral neurons.

Quantitative analysis of rubral degeneration following neonatal deafferentation.

The effect of neonatal hemicerebellectomy on the cytoarchitecture of the red nucleus was investigated in the rat. Quantitative analysis revealed a 31% loss of neurons in the magnocellular red nucleus contralateral to the lesion. This cell loss was accompanied by a similar decrease in the cross-sectional area of the red nucleus. To provide insight into this degenerative process, the neonatal status of the normal cerebellorubral projection was determined using anterograde transport of HRP-WGA which was injected into the deep cerebellar nuclei at various times during the first 10 postnatal days. A definitive cerebellorubral projection was not detectable before postnatal day 8. The data indicate that the rubral afferent fibers from the cerebellum were removed before they reached their target. These results are significant in light of the aberrant projections that develop after neonatal lesions.

An HRP-TMB ultrastructural study of rubral afferents in the rat.

The projections from the deep cerebellar nuclei and the sensorimotor cortex to the red nucleus were studied in the rat using anterograde transport of horseradish peroxidase conjugated with wheat germ agglutinin (HRP-WGA). The anterogradely transported HRP-WGA was visualized ultrastructurally by using a modification of the tetramethylbenzidine (TMB) histochemical technique of Carson and Mesulam ('82). Following injection of HRP-WGA into the sensorimotor cortex, ultrastructural examination of anterograde labeling in the ipsilateral red nucleus revealed labeled synaptic terminals located on small-diameter dendrites of the parvocellular region. These terminals made asymmetrical contacts and contained round vesicles. HRP-WGA placement in the nucleus lateralis resulted in anterograde labeling of synaptic terminals which made asymmetrical contacts with small- to medium-sized dendrites of the parvocellular red nucleus. Similar placements in the nucleus interpositus gave rise to anterograde labeling of synaptic terminals which made asymmetrical contacts with somata and proximal dendrites of magnocellular neurons. In addition, retrograde labeling of magnocellular neurons was also observed following HRP-WGA placements in the nucleus interpositus. Anterogradely labeled interpositorubral synaptic terminals were located on retrogradely labeled rubrocerebellar neurons. The rat red nucleus thus receives topographically organized afferents which are characterized by their specificity in location at the cellular level.

Topographic specificity of aberrant cerebellorubral projections following neonatal hemicerebellectomy in the rat.

Anterograde transport of horseradish peroxidase-wheat germ agglutinin (HRP-WGA) was used to examine the topographic specificity of ascending cerebellar efferent projections in adult rats which were hemicerebellectomized at birth. The results were compared to similar cerebellar projections in unlesioned adults. HRP-WGA placement in the nucleus interpositus of control rats resulted in a dense projection of labeled fibers which decussated in the midbrain, caudal to the red nucleus. In the red nucleus, dense terminal labeling was confined to the magnocellular region, while retrogradely labeled rubrocerebellar neurons were present throughout both parvo- and magnocellular areas. Similar HRP-WGA placements in the nucleus lateralis gave rise to fewer labeled fibers which terminated in the parvocellular red nucleus. In addition to the cerebellorubral projection, other areas of terminal labeling included the mid-brain reticular formation, nucleus parafascicularis prerubralis, zona incerta, fields of Forel and ventral thalamus. In neonatally lesioned adults, aberrant cerebellorubral and cerebellothalamic projections were observed deflecting ipsilaterally at the decussation of the normal contralateral projection. Topographic specificity of the aberrant ipsilateral cerebellorubral projection mirrored that of the normal contralateral fibers. In addition, an ipsilateral projection from the cerebellum could be followed rostral to the red nucleus, to terminate in the ipsilateral ventral thalamus. Lesioned animals also demonstrated marked cell loss in the red nucleus contralateral to the hemicerebellectomy.