ARHGAP4 is a novel RhoGAP that mediates inhibition of cell motility and axon outgrowth.

This report examines the structure and function of ARHGAP4, a novel RhoGAP whose structural features make it ideally suited to regulate the cytoskeletal dynamics that control cell motility and axon outgrowth. Our studies show that ARHGAP4 inhibits the migration of NIH/3T3 cells and the outgrowth of hippocampal axons. ARHGAP4 contains an N-terminal FCH domain, a central GTPase activating (GAP) domain and a C-terminal SH3 domain. Our structure/function analyses show that the FCH domain appears to be important for spatially localizing ARHGAP4 to the leading edges of migrating NIH/3T3 cells and to axon growth cones. Our analyses also show that the GAP domain and C-terminus are necessary for ARHGAP4-mediated inhibition of cell and axon motility. These observations suggest that ARHGAP4 can act as a potent inhibitor of cell and axon motility when it is localized to the leading edge of motile cells and axons.

Glial fibrillary acidic protein is necessary for mature astrocytes to react to beta-amyloid.

Upregulation of the glial fibrillary acidic protein (GFAP) in astrocytes is a hallmark of the phenomenon known as reactive gliosis and, yet, the function of GFAP in this process is largely unknown. Our previous studies have shown that mature astrocytes react vigorously to substrate bound beta-amyloid protein (BAP) in a variety of ways (i.e., increased GFAP, enhanced motility, unusual aggregation patterns, inhibitory ECM production). In order to uncover which, if any, of these phenomena are causally related to the function of GFAP, primary cortical astrocytes from transgenic mice lacking GFAP were cultured on BAP substrates at low or high density and at various lengths of time following in vitro maturation. Differences between mutant and control cells became progressively more obvious when cells were matured in vitro for two weeks or longer and especially in cultures that were at high density. Mature control astrocytes show a dramatic response to BAP by aggregating into a meshwork of rope-like structures that completely bridge over the peptide surface. In marked contrast, mature GFAP-null astrocytes initiate the response much more slowly and had a much reduced ability to aggregate tightly. Furthermore, we prepared hippocampal slice cultures from GFAP-/- and GFAP+/+ mice and compared their astrocytic responses to injected BAP. GFAP-/- astrocytes of hippocampal slice cultures failed to form a barrier-like structure around the edge of the BAP deposit as did GFAP+/+ astrocytes. Our data suggest that GFAP may be essential for mature astrocytes to constrain certain types of highly inflammatory lesions in the brain.

Glycine causes increased excitability and neurotoxicity by activation of NMDA receptors in the hippocampus.

Glycine is an inhibitory neurotransmitter in the spinal cord and also acts as a permissive cofactor required for activation of the N-methyl-D-aspartate (NMDA) receptor. We have found that high concentrations of glycine (10 mM) cause marked hyperexcitability and neurotoxicity in organotypic hippocampal slice cultures. The hyperexcitability, measured using intracellular recording in CA1 pyramidal neurons was completely blocked by the NMDA receptor antagonist MK-801 (10 microM), but not by the AMPA receptor antagonist DNQX (100 microM). The neurotoxicity caused by glycine occurred in all regions of hippocampal cultures but was most marked in area CA1. There was significant CA1 neuronal damage in cultures exposed to 10 mM glycine for 30 min or longer (P < 0.01) or those exposed to 4 mM glycine for 24 h compared to control cultures (P < 0.01). The NMDA antagonists MK-801 (10 microM) and APV (100 microM) significantly reduced glycine-induced neuronal damage in all hippocampal subfields (P < 0.01). The AMPA antagonists CNQX, DNQX, and NBQX (100 microM) had no effect on glycine-induced neuronal damage. High concentrations of glycine therefore appear to enhance the excitability of hippocampal slices in an NMDA receptor-dependent manner. The neurotoxic actions of glycine are also blocked by NMDA receptor antagonists.

Reinnervation of stratum lucidum by hippocampal mossy fibers is developmentally regulated.

Mossy fibers from dentate gyrus granule cells establish synapses on CA3 pyramidal neurons during the first 3 postnatal weeks in the rat. Mossy fiber synapses are primarily restricted to the stratum lucidum. When examined by Timm stain after 10-14 days in vitro, cultured hippocampal slices from postnatal day 4 rat pups show a similar mossy fiber termination pattern in stratum lucidum. Thus, axon guidance cues used by mossy fibers in vivo appear to be preserved in these cultured slices. Three experimental manipulations were performed on hippocampal slice cultures to examine whether the axon guidance cues used by mossy fibers are developmentally regulated. First, mossy fibers were transected on the day of culture or day 7 in vitro. Mossy fibers transected on either day were able to reestablish their synaptic pattern in stratum lucidum of CA3. Second, dentates and hippocampi of same age or different age were co-cultured. Same age co-cultures (P4 dentates to P4 hippocampi or P11 dentates to P11 hippocampi) showed good mossy fiber reinnervation of stratum lucidum, as did different age co-cultures from P4 dentates to P11 hippocampi. However, P11 dentates to P4 hippocampi co-cultures showed little mossy fiber reinnervation of stratum lucidum. Third, new P4 or P11 dentates were co-cultured onto hippocampal slices in which mossy fibers had been allowed to degenerate. New mossy fibers reinnervated these hippocampi, but did not reestablish their normal synaptic pattern in stratum lucidum. These three experimental manipulations suggest that mossy fiber axon guidance mechanisms are developmentally regulated, and that existing mossy fibers play a role in directing mossy fiber reinnervation of stratum lucidum.

Slice cultures of the imprinting-relevant forebrain area medio-rostral neostriatum/hyperstriatum ventrale of the domestic chick: immunocytochemical characterization of neurons containing Ca(2+)-binding proteins.

The forebrain area medio-rostral neostriatum/hyperstriatum ventrale, a presumed analogue to the mammalian prefrontal cortex, displays a variety of synaptic changes during auditory filial imprinting. In order to study the underlying basic mechanisms of this synaptic plasticity we developed slice cultures of the medio-rostral neostriatum/hyperstriatum ventrale from newly hatched chicks. As a prerequisite for these investigations and in order to test the suitability of this system for future studies, we performed a thorough characterization of the in vitro tissue, of its cellular components and some of their biochemical features in comparison with in situ tissue. Since in situ the medio-rostral neostriatum/hyperstriatum ventrale has been previously shown to contain three distinct neuron populations characterized by the activity-regulated Ca(2+)-binding proteins parvalbumin, calbindin D28K and calretinin, we used these proteins as neuronal markers to study the survival and preservation of the morphological features of medio-rostral neostriatum/hyperstriatum ventrale neurons in vitro. In agreement with in vivo studies the three Ca(2+)-binding proteins are confined to neuronal cells and they are not colocalized, i.e. they appear to characterize three different neuron populations. The immunoreactive neurons in medio-rostral neostriatum/hyperstriatum ventrale cultures to a certain extent appear to form synaptic contacts with each other, shown by the double immuncytochemical experiments. One difference between cells in vivo and in vitro is their soma size, which is much larger in vitro than in vivo. This and our previous study on neuronal morphology demonstrates that morphologically and biochemically intact neurons can be maintained in medio-rostral neostriatum/hyperstriatum ventrale slice cultures, which may thus provide a suitable in vitro system for further studies of neuronal and synaptic plasticity in vitro.

Glutamate and non-glutamate receptor mediated toxicity caused by oxygen and glucose deprivation in organotypic hippocampal cultures.

In vitro ischemia models have utilized oxygen, or oxygen and glucose deprivation to simulate ischemic neuronal injury. Combined oxygen and glucose deprivation can induce neuronal damage which is in part mediated through NMDA receptors. Severe oxygen deprivation alone however can cause neuronal injury which is not NMDA mediated. We tested the hypothesis that NMDA, or non-NMDA receptor mediated mechanisms may predominate, to induce neuronal injury following severe oxygen deprivation depending on the presence of glucose. We found that NMDA receptor blockade using dizocilpine (MK-801), DL-2-amino-5-phosphonovaleric acid (APV), or CGS 19755, was highly effective in reducing CA1 injury in organotypic hippocampal cultures, caused by complete oxygen and glucose deprivation. Complete oxygen deprivation alone however, caused CA1 neuronal injury which was not diminished using NMDA receptor blockade alone with MK-801 or APV, or in combination with AMPA/kainate receptor blockade using 6-cyano-7-dinitroquinoxalone-2,3-dione (CNQX). Neuronal protective strategies which act primarily through non-glutamate dependent mechanisms, including hypothermia, low chloride and calcium, and the free radical scavenger, alpha-phenyl-tert-butyl nitrone (PBN), provided neuronal protection against complete oxygen, as well as combined oxygen/glucose deprivation. Raising the pH using Hepes buffer during complete oxygen deprivation did not result in neuronal protection by NMDA receptor blockade. Partial oxygen deprivation alone, partial oxygen deprivation combined with glucose deprivation, glucose deprivation alone, and also glutamate exposure, all produced neuronal damage that was reduced by NMDA receptor blockade. The presence of glucose during complete oxygen deprivation appears to prevent glutamate receptor blockade from reducing neuronal injury in organotypic hippocampal cultures.

Role of calcium channel subtypes in calcium transients in hippocampal CA3 neurons.

Multiple subtypes of voltage-gated calcium channels are differentially localized in brain neurons suggesting that they serve distinct roles in neuronal excitation and signaling. In organotypic hippocampal slice cultures, class D (L-type) calcium channels are predominantly located in the cell bodies of CA3 neurons while class B (N-type) and class A (P or Q-type) are localized in dendrites and associated presynaptic terminals with relatively low somal expression. Using specific antagonists to inhibit calcium transients recorded in CA3 neuronal cell bodies, we found that L-type calcium channels have a predominant role in somal calcium transients elicited by trains of strong stimuli applied to either the soma or the distal apical dendrite while class A calcium channels make a smaller contribution. Presynaptic class B (N-type) and class A (P- and/or Q-type) calcium channels are critical for glutamate-mediated synaptic transmission onto the dendrites of CA3 neurons. Postsynaptic class A and B calcium channels detected on the dendritic shaft by immunocytochemistry were not found to contribute substantially to somal calcium transients during repetitive stimulation of distal dendrites, but sodium channels were required for calcium transients elicited by somatic or dendritic stimulation. Our results show that the different calcium channel subtypes serve distinct roles in cellular activation and transmission of signals in CA3 neurons, consistent with their differential subcellular localization.

Beta amyloid is neurotoxic in hippocampal slice cultures.

We examined the neurotoxicity of the 40 amino acid fragment of beta amyloid peptide (A beta 1-40) in cultured hippocampal slices. When injected into area CA3, A beta 1-40 produced widespread neuronal damage. Injection of the reverse sequence peptide, A beta 40-1, or vehicle alone produced little damage. The distribution A beta 1-40 was highly correlated with the area of neuronal damage. Thioflavine S and electron microscopic analysis confirmed that injected A beta 1-40 formed 7-9 nm AD type amyloid fibrils in the cultures. A beta 1-40 also altered the number of GFAP immunoreactive astrocytes and ED-1 immunoreactive microglia/macrophages within and around the A beta 1-40 deposit. The observed neurotoxicity of A beta 1-40 in hippocampal slice cultures provides evidence that this peptide may be responsible for the neurodegeneration observed in AD.

Differential effects of zinc on hyperpolarizing and depolarizing GABAA synaptic potentials in hippocampal slice cultures.

We have examined the changes in GABAA-mediated synaptic potentials recorded from CA3 pyramidal neurons in hippocampal slice cultures following application of zinc (Zn2+). Unlike 4-AP, Zn2+ did not enhance fast hyperpolarizing potentials but primarily enhanced depolarizing GABAA potentials. Zn2+ did not alter the postsynaptic response of pyramidal neurons to pressure applied GABA, consistent with previous reports that Zn2+ enhances the release of GABA from presynaptic terminals. To examine the role of local circuitry in the production of Zn2+ responses, we recorded from cultures maintained for 7-10 days following removal of the dentate and hilus to allow complete degeneration of the mossy fibers (DGX cultures). Zn2+ produced giant depolarizing potentials (GDPs) in DGX cultures that were identical to those in intact cultures. In contrast, the 4-AP response was dramatically altered in DGX cultures. In DGX cultures, Zn2+ co-applied with 4-AP appeared to inhibit the production of fast hyperpolarizing GABAA synaptic potentials produced by 4-AP alone. This inhibition of fast hyperpolarizing potentials suggests that Zn2+ may reduce the release of GABA onto pyramidal cell somata. These observations suggest that Zn2+ enhances GABA release from local circuit neurons that synapse onto pyramidal cell dendrites, and inhibits GABA release onto pyramidal cell somata.

Glycine site NMDA receptor antagonists provide protection against ischemia-induced neuronal damage in hippocampal slice cultures.

Ischemia-induced neuronal injury can be reduced by glutamate antagonists acting at the N-methyl-D-aspartate (NMDA) receptor. 7-Chlorokynurenic acid and the recently synthesized compound Acea 1021 block NMDA receptors by acting at the strychnine-insensitive glycine site. The anti-ischemic properties of these compounds were tested by evaluating their ability to reduce CA1 neuronal damage in hippocampal slice cultures deprived of oxygen and glucose. Acea 1021 and 7-chlorokynurenic acid significantly reduced CA1 injury produced by oxygen and glucose deprivation in a dose-dependent manner. The neuroprotective effect of these compounds was reversed by the addition of glycine. The phencyclidine site NMDA antagonist MK-801 also provided significant protection to CA1 neurons against the same insult, and this protection was not affected by the addition of glycine. These results indicate that Acea 1021 and 7-chlorokynurenic acid can provide protection to CA1 neurons against ischemia-induced injury by a glycine-sensitive mechanism.

Adenosinergic modulation of CA1 neuronal tolerance to glucose deprivation in organotypic hippocampal cultures.

Glucose deprivation produced neuronal degeneration of CA1 pyramidal neurons in hippocampal slice cultures. The effects of the adenosine agonist cyclohexyladenosine (CHA) and antagonist cyclopentylxanthine (CPX) on CA1 neuronal loss following hypoglycemia was examined using propidium iodide fluorescence as an indicator of cell death. The intensity of propidium iodide fluorescence in hippocampal area CA1 was quantified using Optimas image analysis software. Following 2 or 3 h of glucose deprivation, CPX significantly enhanced injury in the CA1 region while CHA provided significant protection. These results suggest that adenosine plays an important role in endogenous neuronal protection during hypoglycemic injury, and also supports a role for the use of adenosine agonists as neuroprotective agents.

Somatostatin-containing neurons in rat organotypic hippocampal slice cultures: light and electron microscopic immunocytochemistry.

Light and electron microscopic immunocytochemical techniques were used to study the interneuron population staining for somatostatin (SRIF) in cultured slices of rat hippocampus. The SRIF immunoreactive somata were most dense in stratum oriens of areas CA1 and CA3, and in the dentate hilus. Somatostatin immunoreactive cells in areas CA1 and CA3 were characteristically fusiform in shape, with dendrites that extended both parallel to and into the alveus. The axonal plexus in areas CA1 and CA3 was most dense in stratum lacunosum-moleculare and in stratum pyramidale. Electron microscopic analysis of this area revealed that the largest number of symmetric synaptic contacts from SRIF immunoreactive axons were onto pyramidal cell somata and onto dendrites in stratum lacunosum-molecular. In the dentate gyrus, SRIF somata and dendrites were localized in the hilus. Hilar SRIF immunoreactive neurons were fusiform in shape and similar in size to those seen in CA1 and CA3. Axon collaterals coursed throughout the hilus, projected between the granule cells and into the outer molecular layer. The highest number of SRIF synaptic contacts in the dentate gyrus were seen on granule cell dendrites in the outer molecular layer. Synaptic contacts were also observed on hilar neurons and granular cell somata. SRIF synaptic profiles were seen on somata and dendrites of interneurons in all regions. The morphology and synaptic connectivity of SRIF neurons in hippocampal slice cultures appeared generally similar to intact hippocampus.

Development of dopamine-beta-hydroxylase-positive fiber innervation in co-cultured hippocampus-locus coeruleus organotypic slices.

Development of the noradrenergic innervation of the rat hippocampus by the nucleus locus coeruleus was examined immunohistochemically in the roller tube organotypic cultured slice preparation. Slices of rat hippocampus and locus coeruleus were co-cultured on glass coverslips for 2-6 weeks and evaluated for the presence of dopamine-beta-hydroxylase (DBH) and tyrosine hydroxylase (TH) immunoreactive cells and fibers. Large, multipolar DBH- and TH-positive cells were visible within the locus coeruleus; an occasional cell appeared near or just within co-cultured hippocampal tissue and in connecting fiber tracts. DBH-positive cells tended to concentrate near the edges of locus coeruleus tissue. Locus coeruleus slices cultured alone showed little indication of fiber outgrowth in any direction. In co-cultures, however, beaded DBH- and TH-positive fibers were directed toward the hippocampus. The majority of these fibers entered the hippocampus in the hilar/CA3 region and formed extensive collateral branches. Light microscopy suggests that DBH-positive fiber growth was densest at or near the pyramidal cell layer in CA3b and CA3c and in the infragranular region of the dentate hilus. This pattern of noradrenergic innervation of hippocampus by co-cultured locus coeruleus in vitro appears very similar to the pattern established in vivo (see Moudy et al., companion article, this issue).

Colchicine is selectively neurotoxic to dentate granule cells in organotypic cultures of rat hippocampus.

Selective damage to dentate granule cells of the rat hippocampus was produced in organotypic hippocampal cultures by the addition of colchicine. Neuronal toxicity was assessed by quantifying the intensity of propidium iodide fluorescent staining of dying cells. Exposure to 1 microM and 10 microM doses of colchicine for 1 hour caused highly specific dentate granule cell toxicity after 24 hours whereas 100 microM was also slightly toxic to pyramidal cells. Forty eight hours following colchicine exposure additional damage to dentate granule cells and also slight damage to pyramidal cells was seen at all doses. This study demonstrates that dentate granule cells in organotypic hippocampal cultures are selectively vulnerable to colchicine and that colchicine can be used as a tool to selectively destroy the dentate in this preparation.

Effect of beta amyloid peptides on neurons in hippocampal slice cultures.

Several investigators have described the neurotrophic and neurotoxic effects of beta amyloid peptide fragments on dissociated hippocampal neurons in culture. In these prior studies, the peptides were added to dissociated cultures between day 0 and day 4 in vitro, before hippocampal neurons are fully mature. We have analyzed the neurotrophic and neurotoxic effects of beta amyloid fragments beta 1-28, beta 25-35 and beta 1-40 on hippocampal slice cultures, whose physiology and morphology resembles the intact hippocampus. Addition of beta 1-28 or beta 25-35 to the growth medium did not produce significant changes in dendritic length or number of branches. Nerve growth factor, previously reported to enhance the neurotoxic effects of beta 1-40 on dissociated hippocampal neurons in culture, did not significantly enhance the neurotrophic effects of beta 1-28. To achieve high local concentrations of peptides and to avoid potential access problems in the cultures, we injected beta 1-28, beta 25-35, and beta 1-40 directly into the cultures. Amyloid-mediated neurotoxicity was not observed for beta 1-28 or beta 25-35, but beta 1-40 appeared to produce neurodegeneration around the site of injection.

Glutamate-mediated selective vulnerability to ischemia is present in organotypic cultures of hippocampus.

Ischemic damage to the brain, whether induced experimentally or observed clinically, often produces a pattern of delayed selective cell death in subfield CA1 of hippocampus which has been associated with significant neurologic deficits. The present study demonstrates that this selective vulnerability of CA1 neurons to ischemia, with relative preservation of their neighbors, is expressed in organotypic tissue culture and is prevented by the N-methyl-D-aspartate (NMDA) receptor blocker, MK-801. These data provide conclusive evidence that this selective cell death does not have a vascular etiology but is mediated by factors intrinsic to the hippocampal neurons and/or local circuitry. This model system provides an opportunity both to examine mechanisms of ischemic cell death in an avascular environment and to study methods of prevention in the absence of systemic variables.

Phaclofen inhibition of the slow inhibitory postsynaptic potential in hippocampal slice cultures: a possible role for the GABAB-mediated inhibitory postsynaptic potential.

The GABAA antagonist bicuculline methiodide and the GABAB antagonist phaclofen were used to examine the function of the fast inhibitory postsynaptic potential and slow inhibitory postsynaptic potential, in hippocampal slice cultures in the rat. These cultures form easily-visualized monolayers of nerve cells which maintain the structure and synaptic organization of transverse hippocampal slices. The present study shows that the cellular and synaptic physiological properties of slice cultures are very similar, but not identical, to those observed in acutely-prepared hippocampal slices. The major difference is a higher incidence of fast excitatory postsynaptic potentials and inhibitory postsynaptic potentials compared to slices, and the appearance of spontaneous slow inhibitory postsynaptic potentials. This increase in synaptic drive has been useful for our investigation of the role of GABA-mediated inhibitory postsynaptic potentials. Bath application of 10 microM bicuculline blocked the fast inhibitory postsynaptic potentials and gave rise to bursts 1-11 s in duration. The presence of the slow inhibitory postsynaptic potentials did not prevent bicuculline-induced burst activity. Phaclofen (1 mM) perfused in the bath reversibly blocked the slow inhibitory postsynaptic potential, but did not result in the formation of large paroxysmal depolarizing shift-like bursts as seen with bicuculline. Rather, block of the slow inhibitory postsynaptic potential resulted in the formation of repetitive "afterdischarge bursts". These afterdischarge potentials typically appeared with a delay of 2-15 min following block of the slow inhibitory postsynaptic potential, during which time there was a gradual increase in non-synchronized excitatory activity. Once established, this cycle of increasing excitatory activity culminating in afterdischarge potentials recurred at 2-4 min intervals while phaclofen was present.(ABSTRACT TRUNCATED AT 250 WORDS)

Epileptiform activity in hippocampal slice cultures with normal inhibitory synaptic drive.

The synaptic events responsible for epileptiform burst discharge are often difficult to define. Blockade of inhibition has been used to produce epileptiform events, but it is unclear whether increased excitatory activity in the presence of normal inhibition can also result in burst discharge. In the hippocampal slice culture preparation, a small percentage of cultures exhibit spontaneous bursts. To determine whether the absence of inhibitory postsynaptic potentials (IPSPs) is responsible for these spontaneous bursts, we applied the glutamate antagonist, kynurenic acid (KYN) to block burst activity, and unmask any underlying IPSPs. KYN (10 mM) quickly reduced synaptic activity with concomitant loss of burst discharge. Washout of KYN resulted in a gradual return of synaptic activity, during which time both fast and slow IPSPs were clearly observed. As burst activity returned to control levels, excitatory postsynaptic potentials (EPSPs) were increasingly superimposed within the inhibitory events, obscuring (but not eliminating) the IPSPs. In these hippocampal slice cultures, therefore, epileptiform bursts appear to be the result of an abnormally high level of excitatory synaptic drive, not a reduction in inhibition.

Comparison of the actions of phencyclidine and sigma ligands on CA1 hippocampal pyramidal neurons in the rat.

To compare the actions of prototypic drugs which are selective for phencyclidine and sigma receptors, the electrophysiological effects of phencyclidine (PCP),3-[3-hydroxyphenyl]-N-(1-propyl)piperidine [+)3-PPP), and 1,3-di(2-tolyl)guanidine (DTG) on CA1 hippocampal pyramidal neurons were examined. A wide range of concentrations of drug was tested to differentiate specific, receptor-mediated effects from nonselective, anesthetic-like actions. At relatively large concentrations (0.1-1 mM), each compound reversibly increased the threshold of action potentials driven by Schaffer collaterals, the duration of action potentials and membrane resistance. The low potencies and rank order of potency suggested that phencyclidine, (+)3-PPP, and DTG were not acting through either high affinity sigma or phencyclidine receptors. These compounds did have receptor-mediated effects at smaller concentrations. Since none of the compounds affected evoked excitatory or inhibitory postsynaptic potentials (EPSP or IPSP) or driven action potentials at subanesthetic concentrations (less than 100 microM), no evidence was found to support the hypothesis that the actions of phencyclidine result from enhanced release of transmitter, caused by the inhibition of a presynaptic potassium conductance. As observed in other neurons, phencyclidine blocked excitations in CA1 pyramidal cells mediated by N-methyl-D-aspartic acid (NMDA) at behaviorally relevant concentrations (1-10 microM). However, (+)3-PPP (1 microM-1 mM) enhanced the pyramidal cell response to NMDA. Alone, DTG did not effect the NMDA-induced response but did inhibit the enhancement induced by (+)3-PPP. The agonist and antagonist actions of the sigma-selective ligands, (+)3-PPP and DTG, suggests that they modify NMDA-induced responses by acting at the sigma receptor.