Neuron/glia relationships observed over intervals of several months in living mice

Role of Estrogens in the Size of Neuronal Somata of Paravaginal Ganglia in Ovariectomized Rabbits

We aimed to determine the role of estrogens in modulating the size of neuronal somata of paravaginal ganglia. Rabbits were allocated into control (C), ovariectomized (OVX), and OVX treated with estradiol benzoate (OVX + EB) groups to evaluate the neuronal soma area; total serum estradiol (E2) and testosterone (T) levels; the percentage of immunoreactive (ir) neurons anti-aromatase, anti-estrogen receptor (ERα, ERβ) and anti-androgen receptor (AR); the intensity of the immunostaining anti-glial cell line-derived neurotrophic factor (GDNF) and the GDNF family receptor alpha type 1 (GFRα1); and the number of satellite glial cells (SGCs) per neuron. There was a decrease in the neuronal soma size for the OVX group, which was associated with low T, high percentages of aromatase-ir and neuritic AR-ir neurons, and a strong immunostaining anti-GDNF and anti-GFRα1. The decrease in the neuronal soma size was prevented by the EB treatment that increased the E2 without affecting the T levels. Moreover, there was a high percentage of neuritic AR-ir neurons, a strong GDNF immunostaining in the SGC, and an increase in the SGCs per neuron. Present findings show that estrogens modulate the soma size of neurons of the paravaginal ganglia, likely involving the participation of the SGC.

Role of neuron-glia interactions in developmental synapse elimination

During the embryonic development of the nervous system there is a massive formation of synapses. However, the exuberant connectivity present after birth must be pruned during postnatal growth to optimize the function of neuronal circuits. Whilst glial cells play a fundamental role in the formation of early synaptic contacts, their contribution to developmental modifications of established synapses is not well understood. The present review aims to highlight the various roles of glia in the developmental refinement of embryonic synaptic connectivity. We summarize recent evidences linking secretory abilities of glial cells to the disassembly of synaptic contacts that are complementary of a well-established phagocytic role. Considering a theoretical framework, it is discussed how release of glial molecules could be relevant to the developmental refinement of synaptic connectivity. Finally, we propose a three-stage model of synapse elimination in which neurons and glia are functionally associated to timely eliminate synapses.

Stereological and allometric studies on neurons and axo-dendritic synapses in superior cervical ganglia

The superior cervical ganglion (SCG) plays an important role in neuropathies including Horner's syndrome, stroke, and epilepsy. While mammalian SCGs seem to share certain organizational features, they display natural differences related to the animal size and side and the complexity and synaptic coverage of their dendritic arborizations. However, apart from the rat SCG, there is little information concerning the number of SCG neurons and synapses, and the nature of relationships between body weight and the numbers and sizes of neurons and synapses remain uncertain. In the recognition of this gap in the literature, in this chapter, we reviewed the current knowledge on the SCG structure and its remodeling during postnatal development across a plethora of large mammalian species, focusing on exotic rodents and domestic animals. Instrumentally, we present stereology as a state-of-the-art 3D technology to assess the SCG 3D structure unbiasedly and suggest future research directions on this topic.

Glia keep synapse distribution under wraps

The wiring of the nervous system requires that axons navigate to the correct targets and maintain their correct positions during developmental growth. In this issue, Shao et al. (2013) now reveal a crucial new role for glia in preserving correct synaptic connectivity during developmental growth.

Increased coupling and altered glutamate transport currents in astrocytes following kainic-acid-induced status epilepticus

Profound astrogliosis coincident with neuronal cell loss is universally described in human and animal models of temporal lobe epilepsy (TLE). In the kainic acid-induced status epilepticus (SE) model of TLE, astrocytes in the hippocampus become reactive soon after SE and before the onset of spontaneous seizures. To determine if astrocytes in the hippocampus exhibit changes in function soon after SE, we recorded from SR101-labeled astrocytes using the whole-cell patch technique in hippocampal brain slices prepared from control and kainic-acid-treated rats. Glutamate transporter-dependent currents were found to have significantly faster decay time kinetics and in addition, dye coupling between astrocytes was substantially increased. Consistent with an increase in dye coupling in reactive astrocytes, immunoblot experiments demonstrated a significant increase in both glial fibrillary acidic protein (GFAP) and connexin 43, a major gap junction protein expressed by astrocytes. In contrast to what has been observed in resected tissue from patients with refractory epilepsy, changes in potassium currents were not observed shortly after KA-induced SE. While many changes in neuronal function have been identified during the initial period of low seizure probability in this model of TLE, the present study contributes to the growing body of literature suggesting a role for astrocytes in the process of epileptogenesis.

Satellite glial cells in sympathetic and parasympathetic ganglia: in search of function

Glial cells are established as essential for many functions of the central nervous system, and this seems to hold also for glial cells in the peripheral nervous system. The main type of glial cells in most types of peripheral ganglia - sensory, sympathetic, and parasympathetic - is satellite glial cells (SGCs). These cells usually form envelopes around single neurons, which create a distinct functional unit consisting of a neuron and its attending SGCs. This review presents the knowledge on the morphology of SGCs in sympathetic and parasympathetic ganglia, and the (limited) available information on their physiology and pharmacology. It appears that SGCs carry receptors for ATP and can thus respond to the release of this neurotransmitter by the neurons. There is evidence that SGCs have an uptake mechanism for GABA, and possibly other neurotransmitters, which enables them to control the neuronal microenvironment. Damage to post- or preganglionic nerve fibers influences both the ganglionic neurons and the SGCs. One major consequence of postganglionic nerve section is the detachment of preganglionic nerve terminals, resulting in decline of synaptic transmission. It appears that, at least in sympathetic ganglia, SGCs participate in the detachment process, and possibly in the subsequent recovery of the synaptic connections. Unlike sensory neurons, neurons in autonomic ganglia receive synaptic inputs, and SGCs are in very close contact with synaptic boutons. This places the SGCs in a position to influence synaptic transmission and information processing in autonomic ganglia, but this topic requires much further work.

Fluorescence-based monitoring of in vivo neural activity using a circuit-tracing pseudorabies virus

The study of coordinated activity in neuronal circuits has been challenging without a method to simultaneously report activity and connectivity. Here we present the first use of pseudorabies virus (PRV), which spreads through synaptically connected neurons, to express a fluorescent calcium indicator protein and monitor neuronal activity in a living animal. Fluorescence signals were proportional to action potential number and could reliably detect single action potentials in vitro. With two-photon imaging in vivo, we observed both spontaneous and stimulated activity in neurons of infected murine peripheral autonomic submandibular ganglia (SMG). We optically recorded the SMG response in the salivary circuit to direct electrical stimulation of the presynaptic axons and to physiologically relevant sensory stimulation of the oral cavity. During a time window of 48 hours after inoculation, few spontaneous transients occurred. By 72 hours, we identified more frequent and prolonged spontaneous calcium transients, suggestive of neuronal or tissue responses to infection that influence calcium signaling. Our work establishes in vivo investigation of physiological neuronal circuit activity and subsequent effects of infection with single cell resolution.

Neuronal-glial and synaptic remodelling in the adult hypothalamus in response to physiological stimuli

Activation of certain neurosecretory systems of the mammalian hypothalamus induces remodelling of the conformation of their neurons and glial cells. During stimulation of the hypothalamo-neurohypophysial system, astrocytic coverage of oxytocinergic somata and dendrites diminishes and their surfaces become extensively juxtaposed. In the neurohypophysis and median eminence, stimulation evokes a retraction of glial processes and an increase in the contact area between neurosecretory terminals and the perivascular space. These changes are reversible and glial coverage returns to normal upon cessation of stimulation. Neuronal-astrocytic rearrangements also occur in the arcuate nucleus in response to changes in sex steroid levels. The significance of such modifications is a matter of speculation. In the hypothalamic nuclei they may permit synaptic remodelling that takes place concurrently; in the neurohaemal structures they may facilitate neuropeptide release. We know little about the cellular mechanisms involved but glia and neurons of these systems express certain molecules implicated in cell-cell interactions in the developing central nervous system, such as the polysialylated isoform of the neural cell adhesion molecule; this may allow them to manifest their capacity for morphological plasticity in adulthood. The factors inducing the changes vary in the different structures: while oxytocin, in synergy with steroids, appears essential to the induction of the changes in the oxytocinergic system, oestrogen alone is critical in the arcuate nucleus; in the neurohypophysis noradrenaline appears important.

Structure of peripheral synapses: autonomic ganglia

Final motor neurons in sympathetic and parasympathetic ganglia receive synaptic inputs from preganglionic neurons. Quantitative ultrastructural analyses have shown that the spatial distribution of these synapses is mostly sparse and random. Typically, only about 1%-2% of the neuronal surface is covered with synapses, with the rest of the neuronal surface being closely enclosed by Schwann cell processes. The number of synaptic inputs is correlated with the dendritic complexity of the target neuron, and the total number of synaptic contacts is related to the surface area of the post-synaptic neuron. Overall, most neurons receive fewer than 150 synaptic contacts, with individual preganglionic inputs providing between 10 and 50 synaptic contacts. This variation is probably one determinant of synaptic strength in autonomic ganglia. Many neurons in prevertebral sympathetic ganglia receive additional convergent synaptic inputs from intestinofugal neurons located in the enteric plexuses. The neurons support these additional inputs via larger dendritic arborisations together with a higher overall synaptic density. There is considerable neurochemical heterogeneity in presynaptic boutons. Some synapses apparently lack most of the proteins normally required for fast transmitter release and probably do not take part in conventional ganglionic transmission. Furthermore, most preganglionic boutons in the ganglionic neuropil do not form direct synaptic contacts with any neurons. Nevertheless, these boutons may well contribute to slow transmission processes that need not require conventional synaptic structures.

Fluorescent proteins expressed in mouse transgenic lines mark subsets of glia, neurons, macrophages, and dendritic cells for vital examination

To enable vital observation of glia at the neuromuscular junction, transgenic mice were generated that express proteins of the green fluorescent protein family under control of transcriptional regulatory sequences of the human S100B gene. Terminal Schwann cells were imaged repetitively in living animals of one of the transgenic lines to show that, except for extension and retraction of short processes, the glial coverings of the adult neuromuscular synapse are stable. In other lines, subsets of Schwann cells were labeled. The distribution of label suggests that Schwann cells at individual synapses are clonally related, a finding with implications for how these cells might be sorted during postnatal development. Other labeling patterns, some present in unique lines, included astrocytes, microglia, and subsets of cerebellar Bergmann glia, spinal motor neurons, macrophages, and dendritic cells. We show that lines with labeled macrophages can be used to follow the accumulation of these cells at sites of injury.

Astrocytic proliferation in the piriform cortex of amygdala-kindled subjects: a quantitative study in partial versus fully kindled brains

Complex partial epilepsy is a seizure disorder in which attacks frequently arise from foci located in the temporal lobes. The amygdala-kindling model is a widely used model of complex partial epilepsy with secondary generalization. The present study was designed to quantitatively assess astrocytic changes in the rat piriform cortex in the amygdala-kindling model of epilepsy. Bromodeoxyuridine-injected subjects were sacrificed 24 h after the first stage 1 or fifth stage 5 seizure. Brain sections were prepared and examined quantitatively. A significantly higher number of dividing astrocytes (identified by co-labeling with antibodies to bromodeoxyuridine and to one of the astrocytic intermediate filament proteins glial fibrillary acidic protein or vimentin) was found in both partially kindled (stage 1) and fully kindled (stage 5) brains. The partially kindled brains had a significantly higher number of double-labeled cells on the side ipsilateral to stimulation. The opposite trend was observed in the fully kindled brains. Differences between the ipsilateral and contralateral sides of the kindled brain may suggest different role(s) for astrocytes in the development and progression of the seizure-prone state.

In vivo mammalian brain imaging using one- and two-photon fluorescence microendoscopy

One of the major limitations in the current set of techniques available to neuroscientists is a dearth of methods for imaging individual cells deep within the brains of live animals. To overcome this limitation, we developed two forms of minimally invasive fluorescence microendoscopy and tested their abilities to image cells in vivo. Both one- and two-photon fluorescence microendoscopy are based on compound gradient refractive index (GRIN) lenses that are 350-1,000 microm in diameter and provide micron-scale resolution. One-photon microendoscopy allows full-frame images to be viewed by eye or with a camera, and is well suited to fast frame-rate imaging. Two-photon microendoscopy is a laser-scanning modality that provides optical sectioning deep within tissue. Using in vivo microendoscopy we acquired video-rate movies of thalamic and CA1 hippocampal red blood cell dynamics and still-frame images of CA1 neurons and dendrites in anesthetized rats and mice. Microendoscopy will help meet the growing demand for in vivo cellular imaging created by the rapid emergence of new synthetic and genetically encoded fluorophores that can be used to label specific brain areas or cell classes.

In vivo electrophysiological evidences for cortical neuron-glia interactions during slow (<1 Hz) and paroxysmal sleep oscillations

The cortical activity results from complex interactions within networks of neurons and glial cells. The dialogue signals consist of neurotransmitters and various ions, which cross through the extracellular space. Slow (<1 Hz) sleep oscillations were first disclosed and investigated at the neuronal level where they consist of an alternation of the membrane potential between a depolarized and a hyperpolarized state. However, neuronal properties alone could not account for the mechanisms underlying the oscillatory nature of the sleeping cortex. Here I will show the behavior of glial cells during the slow sleep oscillation and its relationship with the variation of the neuronal membrane potential (pairs of neurons and glia recorded simultaneously and intracellularly) suggesting that, in contrast with previous assumptions, glial cells are not idle followers of neuronal activity. I will equally present measurements of the extracellular concentration of K(+) and Ca(2+), ions known to modulate the neuronal excitability. They are also part of the ionic flux that is spatially buffered by glial cells. The timing of the spatial buffering during the slow oscillation suggests that, during normal oscillatory activity, K(+) ions are cleared from active spots and released in the near vicinity, where they modulate the excitability of the neuronal membrane and contribute to maintain the depolarizing phase of the oscillation. Ca(2+) ions undergo a periodic variation of their extracellular concentration, which modulates the synaptic efficacy. The depolarizing phase of the slow oscillation is associated with a gradual depletion of the extracellular Ca(2+) promoting a progressive disfacilitation in the network. This functional synaptic neuronal disconnection is responsible for the ending of the depolarizing phase of the slow oscillation and the onset of a phasic hyperpolarization during which the neuronal network is silent and the intra- and extracellular ionic concentrations return to normal values. Spike-wave seizures often develop during sleep from the slow oscillation. Here I will show how the increased gap junction communication substantiates the facility of the glial syncytium to spatially buffer K(+) ions that were uptaken during the spike-wave seizures, and therefore contributing to the long-range recruitment of cortical territories. Similar mechanisms as those described during the slow oscillation promote the periodic (2-3 Hz) recurrence of spike-wave complexes.

The neuronal naturalist: watching neurons in their native habitat

Dynamic processes of neural development, such as migrations of precursor cells, growth of axons and dendrites, and formation and modification of synapses, can be fully analyzed only with techniques that monitor changes over time. Although there has been long-standing motivation for following cellular and synaptic events in vivo (intravital microscopy), until recently few preparations have been studied, and then often only with great effort. Innovations in low-light and laser-scanning microscopies, coupled with developments of new dyes and of genetically encoded indicators, have increased both the breadth and depth of in situ imaging approaches. Here we present the motivations and challenges for dynamic imaging methods, offer some illustrative examples and point to future opportunities with emerging technologies.

Alzheimer's disease as a disorder of mechanisms underlying structural brain self-organization

Mental function has as its cerebral basis a specific dynamic structure. In particular, cortical and limbic areas involved in "higher brain functions" such as learning, memory, perception, self-awareness and consciousness continuously need to be self-adjusted even after development is completed. By this lifelong self-optimization process, the cognitive, behavioural and emotional reactivity of an individual is stepwise remodelled to meet the environmental demands. While the presence of rigid synaptic connections ensures the stability of the principal characteristics of function, the variable configuration of the flexible synaptic connections determines the unique, non-repeatable character of an experienced mental act. With the increasing need during evolution to organize brain structures of increasing complexity, this process of selective dynamic stabilization and destabilization of synaptic connections becomes more and more important. These mechanisms of structural stabilization and labilization underlying a lifelong synaptic remodelling according to experience, are accompanied, however, by increasing inherent possibilities of failure and may, thus, not only allow for the evolutionary acquisition of "higher brain function" but at the same time provide the basis for a variety of neuropsychiatric disorders. It is the objective of the present paper to outline the hypothesis that it might be the disturbance of structural brain self-organization which, based on both genetic and epigenetic information, constantly "creates" and "re-creates" the brain throughout life, that is the defect that underlies Alzheimer's disease (AD). This hypothesis is, in particular, based on the following lines of evidence. (1) AD is a synaptic disorder. (2) AD is associated with aberrant sprouting at both the presynaptic (axonal) and postsynaptic (dendritic) site. (3) The spatial and temporal distribution of AD pathology follows the pattern of structural neuroplasticity in adulthood, which is a developmental pattern. (4) AD pathology preferentially involves molecules critical for the regulation of modifications of synaptic connections, i.e. "morphoregulatory" molecules that are developmentally controlled, such as growth-inducing and growth-associated molecules, synaptic molecules, adhesion molecules, molecules involved in membrane turnover, cytoskeletal proteins, etc. (5) Life events that place an additional burden on the plastic capacity of the brain or that require a particularly high plastic capacity of the brain might trigger the onset of the disease or might stimulate a more rapid progression of the disease. In other words, they might increase the risk for AD in the sense that they determine when, not whether, one gets AD. (6) AD is associated with a reactivation of developmental programmes that are incompatible with a differentiated cellular background and, therefore, lead to neuronal death. From this hypothesis, it can be predicted that a therapeutic intervention into these pathogenetic mechanisms is a particular challenge as it potentially interferes with those mechanisms that at the same time provide the basis for "higher brain function".

Three-dimensional relationships between hippocampal synapses and astrocytes

Recent studies show that glutamate transporter-mediated currents occur in astrocytes when glutamate is released from hippocampal synapses. These transporters remove excess glutamate from the extracellular space, thereby facilitating synaptic input specificity and preventing neurotoxicity. Little is known about the position of astrocytic processes at hippocampal synapses. Serial electron microscopy and three-dimensional analyses were used to investigate structural relationships between astrocytes and synapses in stratum radiatum of hippocampal area CA1 in the mature rat in vivo and in slices. Only 57 +/- 11% of the synapses had astrocytic processes apposed to them. Of these, the astrocytic processes surrounded less than half (0.43 +/- 22) of the synaptic interface. Other studies suggest that astrocytes extend processes toward higher concentrations of glutamate; thus the presence of astrocytic processes at particular hippocampal synapses might signal which ones are releasing glutamate. The distance between nearest neighboring synapses was usually (approximately 95%) <1 microgram. Astrocytic processes occurred along the extracellular path between 33% of the neighboring synapses, neuronal processes occurred along the path between another 66% of the neighboring synapses, and only 1% of the synapses were close enough such that neither astrocytic nor neuronal processes occurred between them. These morphological arrangements suggest that the glutamate released at approximately two-thirds of hippocampal synapses might diffuse to other synapses, unless neuronal glutamate transporters are more effective than previously reported. The findings also suggest that physiological recordings made from hippocampal astrocytes do not uniformly sample the glutamate released from all hippocampal synapses.

Three-dimensional relationships between hippocampal synapses and astrocytes

Recent studies show that glutamate transporter-mediated currents occur in astrocytes when glutamate is released from hippocampal synapses. These transporters remove excess glutamate from the extracellular space, thereby facilitating synaptic input specificity and preventing neurotoxicity. Little is known about the position of astrocytic processes at hippocampal synapses. Serial electron microscopy and three-dimensional analyses were used to investigate structural relationships between astrocytes and synapses in stratum radiatum of hippocampal area CA1 in the mature rat in vivo and in slices. Only 57 +/- 11% of the synapses had astrocytic processes apposed to them. Of these, the astrocytic processes surrounded less than half (0.43 +/- 22) of the synaptic interface. Other studies suggest that astrocytes extend processes toward higher concentrations of glutamate; thus the presence of astrocytic processes at particular hippocampal synapses might signal which ones are releasing glutamate. The distance between nearest neighboring synapses was usually (approximately 95%) <1 microgram. Astrocytic processes occurred along the extracellular path between 33% of the neighboring synapses, neuronal processes occurred along the path between another 66% of the neighboring synapses, and only 1% of the synapses were close enough such that neither astrocytic nor neuronal processes occurred between them. These morphological arrangements suggest that the glutamate released at approximately two-thirds of hippocampal synapses might diffuse to other synapses, unless neuronal glutamate transporters are more effective than previously reported. The findings also suggest that physiological recordings made from hippocampal astrocytes do not uniformly sample the glutamate released from all hippocampal synapses.

Schwann cells proliferate at rat neuromuscular junctions during development and regeneration

Terminal Schwann cells (TSCs) cover neuromuscular junctions and are important in the repair and maintenance of these synapses. We have examined how these cells are generated at developing junctions and how their number is regulated during repair of nerve injury. At birth, approximately half of the junctions in rat soleus and extensor digitorum longus muscles have one TSC soma. Somata are absent from the remainder, although Schwann cell (SC) processes arising from somata along the preterminal axon cover almost all of these synapses. By 2 months of age, junctions have gained an additional two to three TSCs. Most of this gain occurs during the first 2 postnatal weeks and largely precedes the expansion of endplate size. Although the initial addition is caused by cell migration, mitotic labeling shows extensive division of TSCs at junctions. A slower addition of TSCs occurs in adult muscles, and TSC number in the adult is correlated with endplate size. During repair of nerve injury, TSC number is regulated by a combination of signals from motor neurons and denervated tissue. As shown previously (Connor et al., 1987), denervation of adult muscles did not, in itself, cause TSC mitosis. However, TSCs became mitotic during reinnervation. Partial denervation induced division of TSCs at innervated but not denervated endplates. A disproportionate number of these mitotic cells were found at endplates contacted by TSC processes extended from nearby denervated endplates, contacts known to promote nerve sprouting. These results show an association between TSC mitotic activity and alterations in synaptic structure during development, sprouting, and reinnervation.

Schwann cells proliferate at rat neuromuscular junctions during development and regeneration

Terminal Schwann cells (TSCs) cover neuromuscular junctions and are important in the repair and maintenance of these synapses. We have examined how these cells are generated at developing junctions and how their number is regulated during repair of nerve injury. At birth, approximately half of the junctions in rat soleus and extensor digitorum longus muscles have one TSC soma. Somata are absent from the remainder, although Schwann cell (SC) processes arising from somata along the preterminal axon cover almost all of these synapses. By 2 months of age, junctions have gained an additional two to three TSCs. Most of this gain occurs during the first 2 postnatal weeks and largely precedes the expansion of endplate size. Although the initial addition is caused by cell migration, mitotic labeling shows extensive division of TSCs at junctions. A slower addition of TSCs occurs in adult muscles, and TSC number in the adult is correlated with endplate size. During repair of nerve injury, TSC number is regulated by a combination of signals from motor neurons and denervated tissue. As shown previously (Connor et al., 1987), denervation of adult muscles did not, in itself, cause TSC mitosis. However, TSCs became mitotic during reinnervation. Partial denervation induced division of TSCs at innervated but not denervated endplates. A disproportionate number of these mitotic cells were found at endplates contacted by TSC processes extended from nearby denervated endplates, contacts known to promote nerve sprouting. These results show an association between TSC mitotic activity and alterations in synaptic structure during development, sprouting, and reinnervation.

Synaptic efficacy enhanced by glial cells in vitro

In the developing nervous system, glial cells guide axons to their target areas, but it is unknown whether they help neurons to establish functional synaptic connections. The role of glial cells in synapse formation and function was studied in cultures of purified neurons from the rat central nervous system. In glia-free cultures, retinal ganglion cells formed synapses with normal ultrastructure but displayed little spontaneous synaptic activity and high failure rates in evoked synaptic transmission. In cocultures with neuroglia, the frequency and amplitude of spontaneous postsynaptic currents were potentiated by 70-fold and 5-fold, respectively, and fewer transmission failures occurred. Glial cells increased the action potential-independent quantal release by 12-fold without affecting neuronal survival. Thus, developing neurons in culture form inefficient synapses that require glial signals to become fully functional.

Nerve terminal withdrawal from rat neuromuscular junctions induced by neuregulin and Schwann cells

Schwann cells (SCs) that cap neuromuscular junctions (nmjs) play roles in guiding nerve terminal growth in paralyzed and partially denervated muscles; however, the role of these cells in the day-to-day maintenance of this synapse is obscure. Neuregulins, alternatively spliced ligands for several erbB receptor tyrosine kinases, are thought to play important roles in cell-cell communication at the nmj, affecting synapse-specific gene expression in muscle fibers and the survival of terminal SCs during development. Here we show that application of a soluble neuregulin isoform, glial growth factor II (GGF2), to developing rat muscles alters terminal SCs, nerve terminals, and muscle fibers. SCs extend processes and migrate from the synapse. Nerve terminals retract from acetylcholine receptor-rich synaptic sites, and their axons grow, in association with SCs, to the ends of the muscle. These axons make effective synapses only after withdrawal of GGF2. These synaptic alterations appear to be induced by the actions of neuregulin on SCs, because SC transplants growing into contact with synaptic sites also caused withdrawal of nerve terminal branches. These results show that SCs can alter synaptic structure at the nmj and implicate these cells in the maintenance of this synapse.

Nerve terminal withdrawal from rat neuromuscular junctions induced by neuregulin and Schwann cells

Schwann cells (SCs) that cap neuromuscular junctions (nmjs) play roles in guiding nerve terminal growth in paralyzed and partially denervated muscles; however, the role of these cells in the day-to-day maintenance of this synapse is obscure. Neuregulins, alternatively spliced ligands for several erbB receptor tyrosine kinases, are thought to play important roles in cell-cell communication at the nmj, affecting synapse-specific gene expression in muscle fibers and the survival of terminal SCs during development. Here we show that application of a soluble neuregulin isoform, glial growth factor II (GGF2), to developing rat muscles alters terminal SCs, nerve terminals, and muscle fibers. SCs extend processes and migrate from the synapse. Nerve terminals retract from acetylcholine receptor-rich synaptic sites, and their axons grow, in association with SCs, to the ends of the muscle. These axons make effective synapses only after withdrawal of GGF2. These synaptic alterations appear to be induced by the actions of neuregulin on SCs, because SC transplants growing into contact with synaptic sites also caused withdrawal of nerve terminal branches. These results show that SCs can alter synaptic structure at the nmj and implicate these cells in the maintenance of this synapse.

A putative role for platelet-derived growth factor in angiogenesis and neuroprotection after ischemic stroke in humans

BACKGROUND AND PURPOSE

Growth factors control two important processes in infarcted tissue, ie, angiogenesis and gliosis. We recently reported that transforming growth factor-beta1 (TGF-beta1) might be involved in angiogenesis after ischemic stroke in humans; here we present data of an extensive study on platelet-derived growth factor (PDGF) and its receptors.

METHODS

We studied brain samples from patients who suffered from ischemic stroke for the expression of mRNA encoding PDGF-A, PDGF-B, and PDGF receptors (PDGF-R). Proteins were examined by Western blotting and immunohistochemistry using the antibodies to PDGF-AB, PDGF-BB, PDGF-R alpha, and PDGF-R beta.

RESULTS

At the mRNA level, PDGF-A and PDGF-B were expressed mainly in neurons in penumbra. PDGF-R mRNA was strongly expressed in some astrocytes but mainly in type III/IV neurons in infarct and penumbra. The least expression was seen in the contralateral hemisphere (P<.001). In contrast, both PDGF-AB and PDGF-BB immunoreactive products were present in most cell types: PDGF-R alpha and PDGF-R beta mainly on neurons, and PDGF-R beta on some endothelial cells, with less staining of all the isoforms in the contralateral hemisphere. On Western blots, PDGF-AB and -BB were expressed more within white matter than gray matter of infarct/penumbra, whereas both isoforms of receptor were expressed mainly in gray matter compared with contralateral hemisphere. There was no or very weak expression of the receptor in white matter.

CONCLUSIONS

PDGF proteins are highly expressed in white matter, suggesting that PDGF may exert its function in white matter participating either in regeneration of damaged axons or in glial scar formation. PDGF-BB and its receptor expressed on microvessel endothelial cells might be involved in angiogenesis after stroke. Thus, PDGF is likely to be angiogenic and neuroprotective in stroke.

Postnatal addition of satellite cells to parasympathetic neurons

We have examined the postnatal development of satellite cells associated with parasympathetic neurons of mouse salivary duct ganglia. The number of satellite cells associated with each neuron was found to increase during the first 8 weeks after birth but remained constant thereafter. This corresponds to the period of maximal growth of the salivary gland that serves as the target organ innervated by these neurons. At all ages examined, the number of satellite cells associated with each neuron was found to be highly correlated with neuronal volume. The development of satellite cells associated with individual identified neurons was followed directly by in vivo video microscopy over several months, and the number of satellite cell nuclei was found to increase in regions of the neuronal surface with increasing numbers of synaptic boutons. These results indicate that the postnatal addition of satellite cells to parasympathetic neurons is linked to neuronal enlargement and that synaptic remodeling occurs in concert with satellite cell development.

New views on synapse-glia interactions

Although glial cells ensheath synapses throughout the nervous system, the functional consequences of this relationship are uncertain. Recent studies suggest that glial cells may promote the formation of synapses and help to maintain their function by providing nerve terminals with energy substrates and glutamate precursors.

Axosomatic synapses of the rat ciliary ganglion

We have identified four different types of axosomatic synapses within the rat ciliary ganglion, and present the three-dimensional relationships of both pre- and postsynaptic elements. The majority of axosomatic synapses are situated on small postsynaptic spines that simply appose the axon (termed somatic spine), or are situated within an axonal invagination (termed invaginating somatic spine). The somatic spine synapse predominates, composing 70% of the population, which may be due to simplicity of construction as it usually forms only one active zone. In contrast, the invaginating somatic spine forms multiple active zones and accounts for only 22% of the population. Synapses involving a regular nonspinous portion of the cell membrane were rarely encountered (6%; termed somatic), as were those of axon branches situated within tubular invaginations of the cell body (2%; termed tunnelling). Synapses were differentially distributed, occurring four times more frequently on that portion of neuronal cell body membrane adjacent to the glial cell perinuclear area. However, there was no preferred location by synapse type, suggesting that this unequal distribution was the result of a general mechanism. The neuronal cells of the rat ciliary ganglion apparently constitute a single population, at least on the basis of cell size, shape, and organelle content.

Reduced retinal activity increases GFAP immunoreactivity in rat lateral geniculate nucleus

Dynamic regulation of astrocytic processes by the electrical activity of local neurons has been previously described in chick cochlear nucleus. The present study extends this observation by showing that astrocytes in the rat lateral geniculate nucleus (LGN) also increase their immunoreactivity for glial fibrillary acidic protein (GFAP) soon after deprivation of afferent visual neuronal activity. Within 6 h of enucleation, which eliminates a major source of afferent input to the contralateral LGN, GFAP immunoreactivity increases relative to the ipsilateral LGN. A similar increase in GFAP immunoreactivity can be induced by intraocular injections of tetrodotoxin, demonstrating that a reversible manipulation of optic nerve electrical activity is sufficient to regulate LGN astrocytes. This rapid response to activity deprivation is less dramatic than the gliotic reaction observed 3 weeks following deafferentation, by which time afferent terminals have degenerated. These results support the notion that regulation of astrocytic processes by neural activity may play an important role in activity-dependent synaptic regulations in the various sensory systems of vertebrates.

Synaptic regulation of glial protein expression in vivo

We investigated signaling between individual nerve terminals and perisynaptic Schwann cells, the teloglial cells that cover neuromuscular junctions. When deprived of neuronal activity in vivo, either by motor nerve transection or tetrodotoxin injection, perisynaptic Schwann cells rapidly up-regulated glial fibrillary acidic protein. Addition of transcription or translation inhibitors to excised muscles prevented this increase. Stimulation of cut nerves prevented glial fibrillary acidic protein increases even when postsynaptic nicotinic receptors were blocked, but not when neurotransmitter release was blocked with omega-conotoxin GVIA. We conclude that there is a nerve terminal to glial signal, requiring presynaptic neurotransmitter release, which regulates perisynaptic Schwann cell genes. This may be a general principle since many types of glial are sensitive to transmitters applied in vitro or released in situ.

Ultrastructure of the macaque ciliary ganglion

The primate ciliary ganglion is an obligatory relay in the pathways that control the lens and pupil for the near response and the light reflex, two functions which have been the target of increasing inquiry in behavioural physiology paradigms. This investigation provides a comprehensive description of the ultrastructure of the ciliary ganglion in the rhesus monkey (Macaca mulatta). The results indicate that the ciliary ganglion contains a heterogeneous population of neurons in terms of somatic size, cytoplasmic contents and somatodendritic distribution of terminals. Variations in the clear and dense-cored vesicle content of the synaptic profiles present in the ganglion suggest that the synaptic inputs are also heterogeneous and may mediate separate functions. Several characteristic ultrastructural features of the macaque ciliary ganglion are noteworthy. Despite the large size of the neuronal somata, most cells do not exhibit contacts directly onto the somatic membrane. However, the few somata that do receive direct input often display several axosomatic contacts. The vast majority of synaptic interactions occur in the perisomatic neuropil, where the postsynaptic elements consist of simple and complex somatic appendages, as well as dendrites with their appendages. There is little neuropil independent of these immediately perisomatic regions. In some cases, axonal terminals form the central element of complex glomeruli, in which they are presynaptic to numerous spine-like profiles. In other cases, axon terminals and their postsynaptic targets are found within shallow depressions in the somatic membrane or, occasionally, deeply embedded within the borders of the postganglionic neuron. The somata and all the non-myelinated neuronal elements are surrounded by interdigitating, electron-dense processes of satellite cells. These glial cells are sometimes found in shallow recesses, or deeply embedded within the borders of the neuronal somata. The complexity of the ultrastructure of the ciliary ganglion in the macaque suggests that this ganglion may not be a simple relay in the parasympathetic outflow to the eye, but may instead be the site of neuronal processing of the preganglionic input.

The percentage of nerve cell bodies arranged in clusters decreases with age in the spinal ganglia of adult rabbits

In the spinal ganglia of the rabbit the nerve cell bodies, which in early developmental stages are mutually in contact, come to be completely isolated from each other by a satellite cell sheath and by a connective envelope before birth. The present study demonstrates that in the early postnatal months some nerve cell bodies are still arranged in clusters, and that the percentage of these decreases progressively throughout adult life. This decrease probably arises because in some of the ganglion neurons the process of envelopment of the perikaryon by an individual sheath begins later, or takes place more slowly, than in the majority of cases. Therefore, the relationship between neurons and between neurons and satellite cells may change in certain clusters of nerve cell bodies under normal circumstances during adult life.

Localization of synapses in rat cortical cultures

Astrocyte-rich and astrocyte-poor cultures derived from embryonic rat cerebral cortex were compared to determine whether differences in the location of neuronal somas, dendrites, axons, synapses or astrocytes, relative to the bulk culture medium, could help to explain the large difference in neuronal susceptibility to glutamate toxicity between the two culture systems. The cultures were processed for electron microscopy, thin sectioned across their depths, and photomontaged. In astrocyte-rich cultures, most of the dendrites, axons and synapses were sequestered from the medium by a nearly continuous layer of astrocyte cell bodies and processes. In contrast, astrocytes did not cover the synapses or neuronal processes in astrocyte-poor cultures. In neither culture system were neuronal cell somas covered by glia. Since neuronal cell somas are freely exposed to the medium in both culture conditions, it seems unlikely that receptors on the somal membrane mediate the greater susceptibility of neurons in astrocyte-poor cultures to glutamate toxicity. The layer of astrocytes in the astrocyte-rich cultures may provide a physical buffer that could hinder diffusion of substances from the medium to the interstitium of the neuropil. This physical buffer combined with avid glutamate uptake mechanisms might allow astrocytes to maintain a sufficiently low concentration of glutamate in the local extracellular space to protect dendrites and synapses in the astrocyte-rich, but not in the astrocyte-poor cultures, from the excitotoxic effects of glutamate. The results of this study demonstrate that local sequestering of neurites and synapses by a physical buffer of astrocytes may help to explain the relative resistance of neurons cultured with astrocytes to glutamate toxicity. A similar physical sequestering by astrocytes, of sensitive regions of neurons in the brain, may help protect neurons from glutamate toxicity in vivo.

Astrocyte-regulated synaptogenesis: an in vitro ultrastructural study

Striatal neurons from E15 rat embryos were dissociated, plated at low cell density on polyornithine or on astrocyte monolayers derived from the striatum (homotopic) or mesencephalon (heterotopic), and cultured in a chemically defined medium. Dendrites developing in homotopic co-cultures could reach a state of maturation allowing the establishment of synapses with axons from mesencephalic explants. This culture system thus partially reproduces the in vivo conditions in which striatal neurons developing in an homotopic glial environment can serve as synaptic targets for afferent mesencephalic axons.

Ultrastructural development of the medial superior olive (MSO) in the ferret

When ferrets are born, four weeks before the onset of hearing, few synapses are evident in the medial superior olive (MSO). The synapses present are immature and almost exclusively found in the neuropil. The MSO somata are virtually devoid of synaptic contacts but are contacted by fine glial processes that increasingly ensheathe the somata during the first postnatal week. By P12, somatic synaptogenesis in the MSO is evident. Initially the terminals contain vesicles of irregular shape, size, and distribution. The glial lamellae appear to withdraw as the synaptic contacts form but continue to cover the asynaptic portions of the cell surface. The lamellae frequently extend from ensheathing the soma to encapsulate the immature terminals. During the next two weeks, synaptic density and terminal encapsulation proceed until the somata is surrounded by encapsulated synaptic terminals as in the adult ferret MSO. While most immature terminals contain round vesicles, during the first postnatal week some terminals with nonround vesicles can be distinguished. The first distinction between types of nonround vesicle-containing terminals, i.e., pleiomorphic and ovoid, is in the second postnatal week. This distinction becomes increasingly clear and by the end of the first postnatal month, terminal types can be reliably categorized. These observations indicate that: (1) synapses are present in the MSO neuropil one month prior to the onset of hearing, (2) the major period of synaptogenesis begins approximately two weeks prior to the onset of hearing, and (3) glial lamellae ensheathe MSO somata prior to the onset of somatic synaptogenesis, withdraw as synapses form, and subsequently re-extend to encapsulate newly formed synapses.

Division and migration of satellite glia in the embryonic rat superior cervical ganglion

While distinct precursors committed to a neuronal or glial cell fate are generated from neural crest cells early in peripheral gangliogenesis, little is known about the subsequent generation and maturation of young satellite glia from restricted glial precursor cells. To examine the division and migration of glial precursor cells and their satellite cell progeny, morphological, immunocytochemical and culture techniques were applied to the developing rat superior cervical ganglion. At embryonic day (E)18.5, numerous clusters of nonneuronal cells appeared transiently in the ganglion. Individual cells with a similar morphology were present in E16.5 ganglia, and are likely to represent the precursor cells which generate these clusters. The clustered cells were distinguishable from neighbouring neurons as well as from endothelial cells and fibroblasts. Morphologically similar cells were present in nerve bundles at E18.5 and surrounding principal neurons and nerve bundles in the adult ganglion. Double-label studies of the E18.5 ganglion with tyrosine hydroxylase to identify noradrenergic neurons and propidium iodide counterstaining to visualize all cell nuclei revealed that the cells in clusters stained with propidium iodide but lacked tyrosine hydroxylase immunoreactivity. To determine if cell clusters arose from division, bromodeoxy-uridine, a thymidine analogue, was administered to pregnant mothers between E16.5-E18.5, and ganglionic cells examined at E18.5 both in vivo and in vitro. Numerous non-neuronal cells divided during this period in situ and composed portions of clusters. When dissociated, superior cervical ganglion satellite glia reacted with an NGF-receptor antibody (MAb 217c) and possessed a flattened shape, in contrast to bipolar Schwann cells. Over half of the 217c-immunoreactive glia at E18.5 had incorporated bromodeoxyuridine during E16.5-18.5 in vivo. At birth, non-neuronal cells were no longer grouped in clusters, but were associated with neuronal cell bodies and processes. These findings suggest that, between E16.5-E18.5, glial precursors divide rapidly to form clusters, and that, after the peak of neurogenesis, daughter cells migrate within the ganglion to associate with nerve cell bodies and processes where proliferation continues at a slower rate. Distinct cellular and molecular interactions are likely to trigger the initial rapid division of glial precursors, initiate their migration and association with neuron cell bodies, and control their subsequent slower division.

A role for glial cells in activity-dependent central nervous plasticity? Review and hypothesis

Activity-dependent plasticity relies on changes in neuronal transmission that are controlled by coincidence or noncoincidence of presynaptic and postsynaptic activity. These changes may rely on modulation of neural transmission or on structural changes in neuronal circuitry. The present overview summarizes experimental data that support the involvement of glial cells in central nervous activity-dependent plasticity. A role for glial cells in plastic changes of synaptic transmission may be based on modulation of transmitter uptake or on regulation of the extracellular ion composition. Both mechanisms can be initiated via neuronal-glial information transfer by potassium ions, transmitters, or other diffusible factor originating from active neurons. In addition, the importance of changes in neuronal circuitry in many model systems of activity-dependent plasticity is summarized. Structural changes in neuronal connectivity can be influenced or mediated by glial cells via release of growth or growth permissive factors on neuronal activation, and by active displacement and subsequent elimination of axonal boutons. A unifying hypothesis that integrates these possibilities into a model of activity-dependent plasticity is proposed. In this model glial cells interact with neurons to establish plastic changes; while glial cells have a global effect on plasticity, neuronal mechanisms underlie the induction and local specificity of the plastic change. The proposed hypothesis not only explains conventional findings on activity-dependent plastic changes, but offers an intriguing possibility to explain several paradoxical findings from studies on CNS plasticity that are not yet fully understood. Although the accumulated data seem to support the proposed role for glial cells in plasticity, it has to be emphasized that several steps in the proposed cascades of events require further detailed investigation, and several "missing links" have to be addressed by experimental work. Because of the increasing evidence for glial heterogeneity (for review see Wilkin et al., 1990) it seems to be of great importance to relate findings on glial populations to the developmental stage and topographical origin of the studied cells. The present overview is intended to serve as a guideline for future studies and to expand the view of "neuro" physiologists interested in activity-dependent plasticity. Key questions that have to be addressed relate to the mechanisms of release of growth and growth-permissive factors from glial cells and neuronal-glial information transfer. It is said that every complex problem has a simple, logical, wrong solution. Future studies will reveal the contribution of the proposed simple and logical solution to the understanding of central nervous plasticity.

Synthesis and release of neuroactive substances by glial cells

Glia contain, synthesize, or release more than 20 neuroactive compounds including neuropeptides, amino acid transmitters, eicosanoids, steroids, and growth factors. The stimuli that elicit release differ among compounds but include neuropeptides, neurotransmitters, receptor agonists, and elevated external [K+]. The mechanisms of release are poorly understood in most cases. Many of the neuroactive compounds are localized in discrete subpopulations of glia. Thus, glia are equipped to send as well as receive chemical messages and appear to be present as classes of cells with differing abilities to communicate chemically. It is possible that glia are as diverse as neurons in their functional characteristics.

PDGF A-chain gene is expressed by mammalian neurons during development and in maturity

Platelet-derived growth factor (PDGF) may be a critical factor in the temporal differentiation of glial elements in the mammalian central nervous system. We have used in situ hybridization and immunoperoxidase staining to investigate the localization of PDGF A and have observed high levels of PDGF A-chain mRNA and immunoreactive PDGF A in neurons of embryonic and adult mice. PDGF A-chain expression was shown to be developmentally regulated and tissue specific. Every neuronal population examined in the central and peripheral nervous systems expresses PDGF A transcripts. Variable, significantly weaker signals are observed in glial cells. In contrast to known neurotrophic factors, the PDGF A transcripts are widely distributed among neurons. This generalized distribution of PDGF A transcripts, together with the known effects of PDGF on glial cells in vitro, suggests a unique role of neurons in regulating the proliferation and differentiation of glial cells in vivo.

Glial cell lineages in the neural crest

We have been studying how and when the different peripheral glial cell lineages individualize during avian embryonic development. Three different and complementary experimental approaches were used for this purpose: 1) the quail/chick chimera system allowed the tracing in vivo of the origin of the various types of peripheral glial cells (Schwann cells of nerves, satellite glial cells of sensory and autonomic ganglia, and enteric glial cells), and the analysis of the non-neuronal cell population of ganglia; 2) characterisation of early cell-type specific markers that discriminate between the different glial cell subpopulations; and 3) analysis of the progeny of neural crest cells in clonal cultures. As a result of these approaches, two novel glial-specific markers, expressed earlier than any previously described myelin components, have been identified and partly characterised. The divergence of glial and neuronal cell lineages is a process that is not completely terminated during the phase of neural crest migration. Whereas some cells are apparently already totally committed to a glial fate at this stage, others retain dual neuronal/glial potentialities.

Boundaries during normal and abnormal brain development: in vivo and in vitro studies of glia and glycoconjugates

This paper focuses on transient boundaries of glia and glycoconjugates during development of the mouse central nervous system (CNS). Lectin-bound glycoconjugates, glial fibrillary acidic protein, and the J1/tenascin glycoprotein are distributed coextensively within boundaries around developing substructural arrangements (e.g., developing nuclei, and at a finer level, somatosensory cortical "barrels" related to individual facial vibrissae) throughout the CNS during pattern formation events. Electron microscopy has shown that the J1/tenascin glycoprotein, for example, is present in immature astrocytes, on glial and neuronal plasma membranes, and within the pericellular space that could be extracellular matrix (ECM). The findings presented on the expression of this well-characterized ECM molecule suggest that previously described glial and glycoconjugate boundaries reported by our group are in part composed of specific recognition molecules. The J1/tenascin glycoprotein, a chondroitin sulfate-containing antigen termed the 473 proteoglycan, and the adhesion molecule on glia are expressed within discrete boundary regions and associated axonal pathways. There, they may sculpture fine aspects of functional cytoarchitectonic arrangements and help guide axons to specific targets. The expression and developmental regulation of glycoproteins such as J1/tenascin may thus be integral events during pattern formation and synaptogenesis in the CNS. The presence of abnormal glial arrangements and glycoconjugate boundaries in the cortices of the genetic mutant mouse reeler, and findings on plasticity of boundaries following various perturbations, suggest that boundary expression is controlled by both genetic and epigenetic factors. Some future directions for studying developmental boundaries, including use of cultured explants for in vitro "bioassays," are also discussed.

The use and effects of vital fluorescent dyes: observation of motor nerve terminals and satellite cells in living frog muscles

Several different fluorescent mitochondrial dyes were tested as vital stains for motor nerve terminals and other cells in frog skeletal muscles. It was found that 3,3' diethyloxadicarbocyanine iodide and 4-(4-diethylaminostyryl)-N-methylpyridinium iodide were most useful. Both dyes labelled motor nerve terminals with high reliability. Electrophysiological and morphological control experiments showed that these dyes could be used to repeatedly observe neuromuscular junctions in living animals without affecting synaptic growth or remodelling. The importance of appropriate controls was emphasized by the finding that illumination, if excessively intense or prolonged, can cause physiological and structural damage to nerve terminals. Additional observations indicated that these dyes may be useful for determining the mitochondrial content, and therefore oxidative capacity, of living muscle fibres. It was also found that the fluorescent dyes labelled cells identified as muscle satellite cells, and that these myoblast precursors could be visualized in fixed whole mounts with a nitroblue tetrazolium stain. Both methods were used to study reactive cells that were closely associated with muscle fibres in lesioned muscles. Mitochondrial dyes also labelled the microvasculature, associated axons and other cells.

Repeated, in vivo observation of frog neuromuscular junctions: remodelling involves concurrent growth and retraction

The fluorescent dye 4-(4-diethylaminostyryl)-N-methylpyridinium iodide was used as a vital stain to study remodelling of motor nerve terminals in sartorius muscles of living frogs (Rana pipiens). Identified terminals were observed twice in vivo at intervals of 87-192 days. After the second observation, muscles were fixed and stained with the nitroblue tetrazolium method for nerve terminals and with cholinesterase stain. Observations were made of 243 junctions in 26 frogs. Most nerve terminals grew during the observation interval, with an average increase in total terminal length of 29%. This growth involved substantial remodelling. Within single junctions, the change in size was the net result of differing degrees of growth or shrinkage in individual nerve terminal branches. At least one new terminal branch appeared in 25% of the junctions. Terminal retraction was also common, with branch shortening seen in 60% of junctions and the complete disappearance of a branch in 12%. In one case the original axonal input retracted completely and the junction was partially reinnervated by a terminal sprout from a junction on an adjacent fibre. Some discrepancies between histological and in vivo observations of remodelling were noted. These observations confirm that frog neuromuscular junctions are highly dynamic synapses, subject to profound structural remodelling throughout adult life.

In vitro regulation of neuronal morphogenesis and polarity by astrocyte-derived factors

Mesencephalic neurons were cultured from 2 to 5 days in mesencephalic (CM Gmes) or striatal (CM Gstr) astrocyte conditioned media or in the soluble (S100) and insoluble (P100) fractions prepared from these media by ultracentrifugation. CM Gmes as well as all soluble fractions induced dendritic and axonal elongation, whereas CM Gstr and the insoluble fractions promoted axonal growth only. The study of the shape of the neuronal cell bodies and the measurement of their adhesion to the substratum revealed that axons elongated under low adhesion conditions, but that dendrite growth was highly dependent upon adhesion and spreading of the neuronal soma. This different dependency of axonal and dendritic elongation upon spreading is explained by a model in which we consider the respective viscosities of axons and dendrites. From these observations and speculations we propose that axons and dendrites have different modes of elongation and that the primary effect of the astrocyte-derived factors capable of regulating neuronal polarity is to modify the adhesion of the neurons to their culture substratum.