Pathological implications of iNOS expression in central white matter: an ex vivo study of optic nerves from rats with experimental allergic encephalomyelitis

Excessive nitric oxide (NO) production from the inducible isoform of nitric oxide synthase (iNOS) has been invoked as a causative factor in many neurodegenerative disorders, including multiple sclerosis. This hypothesis has been supported by in vitro studies showing that glial iNOS expression results in toxic NO concentrations (near 1 microm). To investigate the relevance of such findings, experiments were carried out ex vivo on optic nerves from rats with exacerbated experimental allergic encephalomyelitis, a model of multiple sclerosis. The nerves displayed characteristic immunopathology and expression of iNOS in macrophages and/or microglia and there was overt axonal damage in localized regions of the optic chiasm. The resulting NO levels in the optic nerve were sufficient to cause activation of guanylyl cyclase-coupled NO receptors, resulting in marked cGMP accumulation in axons throughout the nerve. Nevertheless, calibration of cGMP levels against those evoked by exogenous NO indicated that the nerves were not compromised metabolically and that their ambient NO concentration was only approximately 1 nm. Consistent with this observation, electrophysiological tests indicated that there was no ongoing malfunctioning of the type that can be elicited by high exogenous NO concentrations. It is concluded that, with iNOS expressed in physiological locations and levels, the tissue levels of NO remain at concentrations far lower than those shown to have toxic effects, despite continuous NO synthesis. The fact that NO can rise to much higher levels in dispersed cultures in vitro may be attributable to a deficiency in NO inactivation in such preparations.

Nitric oxide toxicity in CNS white matter: an in vitro study using rat optic nerve

Excessive nitric oxide formation may contribute to the pathology occurring in diseases affecting central white matter, such as multiple sclerosis. The rat isolated optic nerve preparation was used to investigate the potential toxicity of the molecule towards such tissue. The nerves were exposed to a range of concentrations of different classes of nitric oxide donor for up to 23 h, with or without a subsequent period of recovery, and the damage assessed by quantitative histological methods. Degeneration of axons and macroglia occurred in a time- and concentration-dependent manner, the order of susceptibility being: axons>oligodendrocytes>astrocytes. Use of NONOate donors differing in half-life indicated that nitric oxide delivered in an enduring manner at relatively low concentration was more toxic than the same amount supplied rapidly at high concentration. The mechanism by which nitric oxide affects axons was studied using a donor [3-(n-propylamino)propylamine/NO adduct, PAPA/NO] with an intermediate half-life that produced selective axonopathy after a 2-h exposure (plus 2 h recovery). Axon damage was abolished if, during the exposure, Na(+) or Ca(2+) was removed from the bathing medium or the sodium channel inhibitors tetrodotoxin or BW619C89 (sipatrigine) were added. In electrophysiological experiments, the donor elicited a biphasic depolarisation. The second, larger component (occurring after 7-10 min) was associated with a block of nerve conduction and could be inhibited by tetrodotoxin. Coincident with the secondary depolarisation was a reduction in ATP levels by about 50%, an effect that was also inhibited by tetrodotoxin. It is concluded that nitric oxide, in submicromolar concentrations, can kill axons and macroglia in white matter. The findings lend support to the hypothesis that nitric oxide may be of importance to white matter pathologies, particularly those in which inducible nitric oxide synthase is expressed. The axonopathy, at least when elicited over relatively short time intervals, is likely to be caused by metabolic inhibition. As in anoxia and anoxia/aglycaemia, nitric oxide-induced destruction of axons is likely to be caused by the Ca(2+) overload that follows a reduction in ATP levels in the face of continued influx of Na(+) through voltage-dependent channels.

Soluble guanylyl cyclase activator YC-1 protects white matter axons from nitric oxide toxicity and metabolic stress, probably through Na(+) channel inhibition

In the rat isolated optic nerve, nitric oxide (NO) activates soluble guanylyl cyclase (sGC), resulting in a selective accumulation of cGMP in the axons. The axons are also selectively vulnerable to NO toxicity. The experiments initially aimed to determine any causative link between these two effects. It was shown, using a NONOate donor, that NO-induced axonal damage occurred independently of cGMP. Unexpectedly, however, the compound YC-1, which is an allosteric activator of sGC, potently inhibited NO-induced axonopathy (IC(50) = 3 microM). This effect was not attributable to increased cGMP accumulation. YC-1 (30 microM) also protected the axons against damage by simulated ischemia, which (like NO toxicity) is sensitive to Na(+) channel inhibition. Although chemically unrelated to any known Na(+) channel inhibitor, YC-1 was effective in two biochemical assays for activity on Na(+) channels in synaptosomes. Electrophysiological recording from hippocampal neurons showed that YC-1 inhibited Na(+) currents in a voltage-dependent manner. At a concentration giving maximal protection of optic nerve axons from NO toxicity (30 microM), YC-1 did not affect normal axon conduction. It is concluded that the powerful axonoprotective action of YC-1 is unrelated to its activity on sGC but is explained by a novel action on voltage-dependent Na(+) channels. The unusual ability of YC-1 to protect axons so effectively without interfering with their normal function suggests that the molecule could serve as a prototype for the development of more selective Na(+) channel inhibitors with potential utility in neurological and neurodegenerative disorders.

Mechanisms of ischaemic damage to central white matter axons: a quantitative histological analysis using rat optic nerve

The mechanism of ischaemic injury to white matter axons was studied by transiently depriving rat optic nerves in vitro of oxygen and glucose. Light and electron microscopic analysis showed that increasing periods of oxygen/glucose deprivation (up to 1 h) caused, after a 90-min recovery period, the appearance of increasing numbers of swollen axons whose ultrastructure indicated that they were irreversibly damaged. This conclusion was supported by experiments showing that the damage persisted after a longer recovery period (3 h). To quantify the axonal pathology, an automated morphometric method, based on measurement of the density of swollen axons, was developed. Omission of Ca2+ from the incubation solution during 1 h of oxygen/glucose deprivation (and for 15 min either side) completely prevented the axonopathy (assessed following 90 min recovery). Omission of Na+ was also effective, though less so (70% protection). The classical Na+ channel blocker, tetrodotoxin (1 microM), provided 92% protection. In view of this evidence implicating Na+ channels in the pathogenesis of the axonal damage, the effects of three different Na+ channel inhibitors, with known neuroprotective properties towards gray matter in in vivo models of cerebral ischaemia, were tested. The compounds used were lamotrigine and the structurally-related molecules, BW619C89 and BW1003C87. All three compounds protected the axons to varying degrees, the maximal efficacies (observed at 30 to 100 microM) being in the order: BW619C89 (>95% protection) > BW1003C87 (70%) > lamotrigine (50%). At a concentration affording near complete protection (100 microM), BW619C89 had no significant effect on the optic nerve compound action potential. Experiments in which BW619C89 was added at different times indicated that its effects were exerted during two distinct phases, one (accounting for about 50% protection) was during the early stage of oxygen/glucose deprivation itself and the other (also about 50%) during the first 15 min of recovery in normal incubation solution. The results are consistent with a pathophysiological mechanism in which Na+ entry through tetrodotoxin-sensitive Na+ channels contributes to Na+ loading of the axoplasm which then results in a lethal Ca2+ overload through reversed Na(+)-Ca2+ exchange. The identification of BW619C89 as a compound able to prevent oxygen/glucose deprivation-induced injury to white matter axons without affecting normal nerve function opens the way to testing the importance of this pathway in white matter injury in vivo.

Nitric oxide stimulates cGMP formation in rat optic nerve axons, providing a specific marker of axon viability

A major transduction pathway for nitric oxide (NO) is stimulation of soluble guanylyl cyclase and the generation of cyclic GMP (cGMP). In the central nervous system, the NO-cGMP pathway has previously been associated primarily with synapses, particularly glutamatergic synapses. We report here that NO caused a large increase in the levels of cGMP in a central white matter tract devoid of synapses, namely in the rat isolated optic nerve. Cyclic GMP immunohistochemistry indicated that this response was confined to the axons. Accordingly, nerves previously subjected to 1 h of oxygen/glucose deprivation, which leads to irreversible axonal damage, displayed an 80% reduction in their subsequent capacity to generate cGMP in response to NO and a corresponding reduction in the numbers of cGMP-immunostained axons. Protection of the axon cGMP response against this insult was achieved by omission of Ca2 + or Na + from the incubation medium, and by the pharmacological agents tetrodotoxin, lamotrigine, BW619C89 and BW1003C87, all of which protect axonal structure from oxygen/glucose deprivation-induced damage. The results suggest that the NO-cGMP pathway has a hitherto unsuspected function in the optic nerve. Additionally, the expression of NO-stimulated guanylyl cyclase in optic nerve axons provides a simple, sensitive and specific marker of their functional integrity that is likely to be valuable in investigating the mechanisms responsible for axon degeneration in ischaemia and other conditions.

Nitric oxide does not mediate acute glutamate neurotoxicity, nor is it neuroprotective, in rat brain slices

Nitric oxide (NO), generated upon glutamate receptor activation, elicits cyclic GMP accumulation through stimulation of guanylyl cyclase. NO is also a potential cytotoxin that has been suggested, on the basis of tissue culture experiments, to mediate neuronal damage associated with excessive activity of the N-methyl-D-aspartate (NMDA) subtype of glutamate receptor. We have investigated the involvement of NO in the toxicity of glutamate receptor agonists in brain slice preparations. Slices of cerebellum and hippocampus from the developing rat exhibited neuronal necrosis following exposure (5-30 min) to NMDA (100 microM or 1 mM). When the exposures were carried out in the presence of NO synthase inhibitors, at concentrations suppressing NMDA-induced NO formation (as judged by measurements of cyclic GMP accumulation), the extent of injury was unaffected. To determine if exogenous NO is able to replicate NMDA toxicity, the slices were exposed to high concentrations of NO donating compounds for up to 2 hr. No damage was detectable. NO donors, moreover, neither reduced NMDA toxicity, nor potentiated the degeneration caused by just suprathreshold NMDA concentrations. The toxicities of non-NMDA agonists, or of glutamate itself, were also unaltered by NO synthase inhibitors or NO donors. Similar results were obtained using hippocampal slices from more mature animals. We conclude that the acute neurodegeneration mediated by NMDA or non-NMDA receptors in the slice preparations is not mediated by NO, nor is NO neuroprotective under these conditions.

Morphological response of endoplasmic reticulum in cerebellar Purkinje cells to calcium deprivation

Mobilisable intracellular Ca2+ stores are highly enriched in the cerebellum, particularly in Purkinje cells. We have detected, by light and electron microscopy, striking morphological changes in the presumed Ca2+ stores of Purkinje cells when slices of eight-day-old rat cerebellum were incubated in Ca(2+)-deficient media. After 30 min under these conditions, the endoplasmic reticulum became thinned and elongated. By 2 h, it was transformed into multilamellar, whorl-like inclusions with electron-dense cores. These changes were reversed on reintroduction of Ca2+. Analogous changes in other neurons were not observed. The results suggest that Ca2+ storage sites within Purkinje cells are capable of dramatic morphological change depending on the availability of Ca2+. The transformations may reflect, initially, depletion of Ca2+ from the stores and then homeostatic alterations in their capacity.

NMDA receptor activation induces nitric oxide synthesis from arginine in rat brain slices

Activation of N-methyl-D-aspartate (NMDA) receptors in rat cerebellum leads to the release of endothelium-derived relaxing factor, now identified as nitric oxide (NO), a stimulator of soluble guanylate cyclase. L-NG-monomethylarginine (L-NMMA), which blocks NO synthesis from L-arginine in several tissues, including a crude synaptosomal preparation from brain, inhibited the elevation of cyclic GMP induced by NMDA in rat cerebellar slices. D-NMMA was ineffective. L-Arginine, but not its D enantiomer, augmented the response to NMDA and reversed the inhibition by L-NMMA. The results indicate that stimulation of NMDA receptors results in the activation of the enzyme which catalyzes the formation of NO from L-arginine.

Neurotoxicity of excitatory amino acid receptor agonists in young rat hippocampal slices

Hippocampal slices from young (8-day-old) rats were evaluated as a model for investigating the mechanisms underlying the neurotoxic action of excitatory amino acid receptor agonists. The slices were exposed to the agonists for up to 30 min and were then postincubated for 90 min in order to allow irreversibly damaged cells to become visibly necrotic. Under control conditions (greater than or equal to 3 h incubation) all regions of the hippocampus and dentate gyrus displayed good preservation. Exposure of the slices to N-methyl-D-aspartate (NMDA) resulted in widespread, oedematous necrosis of all neuronal types (except undifferentiated granule cells) which was maximal after 20 min exposure to a concentration of 100 microM. With 30 min exposure, the EC50 for NMDA was 30 microM; 10 min exposure to NMDA at a concentration of 100 microM was sufficient to destroy 50% of the neurones. Quisqualate produced a degeneration of most (98%) of the CA3 neurones, a proportion (65%) of CA1 neurons and some (25%) of the dentate granule cells. The occurrence of "dark cell degeneration" was prevalent. Half maximal effects on CA3 neurones were estimated to be produced by a concentration of 15 microM (with 30 min exposure) or by 8 min exposure (at 100 microM concentration). Incubation of the slices with kainate (100 microM for 30 min) did not cause widespread damage but led to the necrosis of a small population of cells scattered in all regions of the hippocampus and dentate gyrus. The patterns of toxicity of the different agonists resemble closely those found after their administration in vivo. It is suggested that the hippocampal slices provide a valuable new model system for studying excitatory amino acid toxicity.

Quisqualate neurotoxicity: a delayed, CNQX-sensitive process triggered by a CNQX-insensitive mechanism in young rat hippocampal slices

Exposure of slices of young rat hippocampus for 30 min to the glutamate receptor agonist, quisqualate (QA, 30 microM), led, after a 90 min recovery period, to severe 'dark cell degeneration' of pyramidal neurones, most extensively those in CA3. When present during the exposure, 6-cyano-2,3-dihydroxy-7-nitroquinoxaline (CNQX, 10 microM), an antagonist with preferential action on non-N-methyl-D-aspartate receptors, did not prevent this toxic effect of QA. However, it was effective when included either during the recovery period as well or, indeed, only during recovery. Comparable results were obtained with kynurenate (3 mM), but not with D,L-2-amino-5-phosphonovalerate (100 microM) or with tetrodotoxin (0.5 microM). Grease-gap recordings showed that CNQX markedly inhibited QA-induced depolarization. It is concluded that QA toxicity is not triggered by QA-induced depolarization but instead involves CNQX-resistant QA receptors, possibly those linked to phospholipid metabolism. The induction mechanism does not itself cause irreversible injury but subsequently, a delayed form of damage takes place which is mediated by activation of CNQX/kynurenate-sensitive receptors.

Differential dependence on Ca2+ of N-methyl-D-aspartate and quisqualate neurotoxicity in young rat hippocampal slices

Exposure of slices of young (8 days old) rat hippocampus to 100 microM N-methyl-D-aspartate (NMDA) for 20 min followed by 90 min recovery, resulted in widespread, oedematous necrosis of all classes of neurones. The NMDA antagonist, D,L-2-amino-5-phosphonovalerate (APV) or omission of Ca2+ from the exposing solution prevented this cell death, but a large reduction in Cl- was ineffective. Quisqualate (100 microM, 20 min) led to a different pathological pattern characterised most strikingly by large numbers of cells undergoing 'dark cell degeneration'. Numerically, the neurones were affected in the order CA3 greater than CA1 greater than dentate granule cells. Quisqualate toxicity was not prevented by APV nor by reducing Ca2+ or Cl-. It is concluded that, as in cerebellar slices (but unlike in cultures of hippocampal neurones) NMDA toxicity in hippocampal slices is Ca2+-dependent and Cl- -independent. However, quisqualate exerts its pathological effects through a different mechanism. This mechanism may be primarily metabolic rather than ionic.

Cyclic GMP and cell death in rat cerebellar slices

Incubated slices of young rat cerebellum were used to examine the possible relationship between the neurotoxic effects of excitatory amino acids and their ability to elicit large increases in the levels of cyclic GMP in this tissue. No cell death was detectable following exposure of the slices to the guanylate cyclase activator, nitroprusside (up to 0.3 mM), the phosphodiesterase inhibitor, isobutylmethylxanthine (0.5 mM), or to cyclic GMP (10 mM) and its dibutyryl and 8-bromo derivatives (0.5 mM). However, incubation of the slices with tbe guanylate cyclase inhibitors, N-methylhydroxylamine and hydroxylamine (0.1-1 mM), methylene blue (10-100 microM), ethacrynic acid (300 microM) and retinol (1 mM) caused a progressive destruction of the differentiating cells. The damage induced by N-methylhydroxylamine and hydroxylamine was inhibited by nitroprusside, cyclic GMP and isobutylmethylxanthine. It could also be reduced by lowering the partial pressure of oxygen, by oxygen radical scavenging enzymes and by omitting Ca2+ from the medium. Oxygen radical generating enzyme systems mimicked the pattern of toxicity of the guanylate cyclase inhibitors but their effects were not reduced by nitroprusside or omission of Ca2+. The results indicate that guanylate cyclase/cyclic GMP does not mediate amino acid neurotoxicity but, instead, may be part of a protective mechanism against oxygen free radicals.

Electron microscopic autoradiography of D-[3H]aspartate uptake sites in mouse cerebellar slices shows poor labelling of mossy fibre terminals

Recent evidence indicates that mossy fibres in the cerebellum may use glutamate as their transmitter. To determine if mossy fibre terminals accordingly express a high-affinity uptake mechanism for excitatory amino acids, mouse cerebellar slices were incubated for 3 h with 1 microM D-[3H]aspartate and examined by electron microscopic autoradiography. In the granule cell layer, granule cell bodies were clearly labelled but the majority of mossy fibre terminals were without silver grains.

Receptor-linked ionic channels mediate N-methyl-D-aspartate neurotoxicity in rat cerebellar slices

In young rat cerebellar slices, histological methods showed that the neurotoxic potency of N-methyl-D-aspartate (NMDA) towards granule cells and intracerebellar nucleus neurons was increased 2- to 3-fold on removal of Mg ions, which have a blocking effect on NMDA-activated ion channels. The depolarizing potency of NMDA on granule cells, recorded using a gap method, was similarly enhanced whereas that of kainate, a non-NMDA receptor agonist, was unchanged. The neurotoxic potency of kainate (towards Golgi cells) was also unaltered by removal of Mg2+. In Mg2+-containing medium, neuronal depolarization induced either by kainate or by high K+ potentiated NMDA toxicity, apparently by reducing the channel block by Mg2+. The results strongly support the hypothesis that excessive Ca2+ influx through NMDA/Mg2+-gated ion channels mediates NMDA toxicity. They also have clear implications regarding the likely mechanism of toxicity of agonists, such as glutamate, able to activate both NMDA and non-NMDA receptors.

Selective loss of Purkinje and granule cell responsiveness to N-methyl-D-aspartate in rat cerebellum during development

Depolarizing responses of Purkinje and granule cells to excitatory amino acid receptor agonists were recorded from rat cerebellar slices at various stages of postnatal maturation using a gap technique. No major developmental changes in relative potency or efficacy of kainate and quisqualate were observed. However, Purkinje and granule neurones both became less responsive to N-methyl-D-aspartate (NMDA) with age, most dramatically so between 14 and 21 days. This transient chemosensitivity to NMDA may reflect a special role of the NMDA receptor system in cerebellar development.

Quinolinate mimics neurotoxic actions of N-methyl-D-aspartate in rat cerebellar slices

Incubation of slices of immature rat cerebellum for 30 min with quinolinate (QUIN), an endogenous neurotoxin, resulted in the selective necrosis of granule cells and intracerebellar nucleus neurones. Concentration of QUIN in the millimolar range were needed for these effects. The same neuronal populations were also selectively killed by N-methyl-D-aspartate (NMDA) but the toxic potency of NMDA was 40-fold higher than that of QUIN. Depolarizing responses of granule cells to brief applications of QUIN and NMDA were recorded using a gap method. Dose-response curves to the two compounds appeared parallel but NMDA was 30-fold more potent than QUIN. The depolarizing and toxic actions of QUIN and NMDA were inhibited by the NMDA antagonist, 2-amino-5-phosphonopentanoate. We conclude that the selective toxicity of QUIN in this tissue arises from its activity on NMDA receptors.

Cellular origins of cyclic GMP responses to excitatory amino acid receptor agonists in rat cerebellum in vitro

Incubated slices and freshly dissociated cells from 8-day-old rat cerebellum were used to try to identify the cells that participate in the large increases in cyclic GMP levels that follow activation of excitatory amino acid receptors in this tissue. In the slices, cyclic GMP responses to L-glutamate and related excitants were unaffected by tetrodotoxin and could be replicated by the guanylate cyclase activator nitroprusside. Nitroprusside and the receptor agonists appeared to activate the same pool of the enzyme. Prior destruction of neuroblasts, deep nuclei, or Golgi neurones did not cause loss of responses to L-glutamate. If granule cells were rendered necrotic, however, the cyclic GMP responses to all excitants tested were reduced by greater than or equal to 90%. Substantial losses of responses to veratridine and high K+ levels also occurred, but the nitroprusside-induced elevations were unaffected. In dissociated cell suspensions, the magnitude of responses to receptor agonists, but not those to nitroprusside, was markedly dependent on cell concentration. Responses to L-glutamate were the same in cell suspensions that were Purkinje cell depleted and Purkinje cell enriched. It is concluded that granule cells are primarily involved in the cyclic GMP responses to excitatory amino acids but that the cyclic GMP accumulations occur elsewhere, probably in glial cells.

Amino acid neurotoxicity: intracellular sites of calcium accumulation associated with the onset of irreversible damage to rat cerebellar neurones in vitro

Electron microscopy and the combined oxalate-pyroantimonate technique were used to locate calcium in intracerebellar nucleus neurones of rat cerebellar slices subjected to a neurotoxic concentration of N-methyl-D-aspartate. After a sub-lethal exposure period (5 min) calcium pyroantimonate deposits were found in swollen cisterns of the Golgi apparatus and, in lesser amounts, in the nuclei. Deposits were more prominent in the nuclei after a just-lethal exposure (10 min) when they were additionally observed within a population of swollen mitochondria and also apparently free in the dendritic and somatic cytoplasm. The results support the proposal that amino acid neurotoxicity is a consequence of an intracellular Ca2+ overload brought about by excessive Ca2+ influx.

Reversible and irreversible neuronal damage caused by excitatory amino acid analogues in rat cerebellar slices

Slice preparations of the developing rat cerebellum were used to investigate the light and electron microscopic correlates of reversible and irreversible neuronal injury caused by the neurotoxic excitatory amino acid receptor agonists, kainate and N-methyl-D-aspartate. The slices were examined after various periods of exposure to the agonists (up to 30 min) with or without a 90 min recovery period in agonist-free medium. N-Methyl-D-aspartate (100 microM) caused necrosis of deep nuclear neurons and differentiating granule cells, the exposure times necessary to induce non-recoverable damage (leading to necrosis), being, respectively, 10 min and 20-30 min. Exposure periods of only 2-4 min with kainate (100 microM) were needed for Golgi cells to subsequently undergo necrosis. Other cell types (Purkinje, granule and deep nuclear neurons) were altered histologically by kainate but most recovered fully from 30 min exposures. Before the recovery period, the worst affected of these cells (deep nuclear neurons) displayed increased cytoplasmic and nuclear electron density and microvacuolation due to swelling of Golgi cisterns but little or no chromatin clumping or mitochondrial expansion. The neurons which were injured irreversibly by the agonists within 30 min displayed, near the time of lethal injury, increased cytoplasmic and nuclear electron lucency, marked focal aggregation of chromatin and swelling of Golgi apparatus. Mitochondrial swelling did not appear to precede lethal injury and even after exposure times sufficient, or more than sufficient, to lead to necrosis, large numbers of mitochondria remained in a condensed configuration. The significance of the histological changes is discussed and they are compared with those occurring in other pathological conditions. The time scales required for the receptor agonists to induce irreversible cellular lesions would be consistent with this being a process which is responsible for acute neuronal necrosis in the brain.

Ionic requirements for neurotoxic effects of excitatory amino acid analogues in rat cerebellar slices

The ionic requirements for the neurotoxic effects of N-methyl-D-aspartate and kainate in incubated slices of developing rat cerebellum were studied using light and electron microscopy. Under normal conditions, 30 min exposure to 100 microM N-methyl-D-aspartate followed by a 90 min recovery period in agonist-free medium resulted in the necrosis of differentiating granule cells and deep nuclear neurons, while the corresponding effect of 100 microM kainate was the death of Golgi cells. Substitution of 96% of the Cl- in the medium with isethionate did not prevent the toxicity of either agonist. However, all the ordinarily vulnerable cells survived and exhibited normal ultrastructure if the slices were exposed to the excitants in a Ca2+-free medium and were subsequently allowed to recover in a Ca2+-containing solution. Prior to this recovery period, granule, Golgi and deep nuclear neurons exposed to N-methyl-D-aspartate were markedly swollen but their mitochondria were hypercontracted and there was no clumping of chromatin or obvious swelling of the rough endoplasmic reticulum or Golgi apparatus, in contrast to observations made on slices exposed to this agonist in normal medium. Substitution of all the Na+ in the medium with a mixture of choline (118 mM) and Tris (25 mM) itself caused necrosis of granule cells and deep nuclear neurons and an intense microvacuolation of Purkinje cells, due, in large part, to high amplitude mitochondrial swelling. A low (25 mM) Na+ medium was well tolerated under control conditions. This medium protected granule cells but not deep nuclear neurons from the toxicity of N-methyl-D-aspartate and failed to prevent kainate-induced death of Golgi cells. It is concluded that the acute neurotoxic effects of the two excitatory amino acid receptor agonists in the slices are dependent on extracellular Ca2+ and are independent of extracellular Cl-. Where apparent, the protective effect of reducing extracellular Na+ on the toxicity of N-methyl-D-aspartate is likely to reflect the involvement of this ion in the primary depolarizing mechanism.

Amino acid neurotoxicity: relationship to neuronal depolarization in rat cerebellar slices

It has long been proposed that the excitatory and toxic properties of acidic amino acid receptor agonists are linked. To test this hypothesis, the depolarizing effects of quisqualate, kainate and N-methyl-D-aspartate in adult and immature rat cerebellar slices have been studied in relation to their neurotoxic effects in the same tissues (reported separately). A "grease-gap" method was used to measure the depolarizing responses of Purkinje cells and granule cells in lobule VI to the agonists. The depolarizing potencies of kainate and quisqualate were apparently similar on both cell types and at both ages studied although maximal responses to kainate were always larger. N-Methyl-D-aspartate was a very weak agonist in the adult slices but was much more effective in the immature tissues, apparently on both Purkinje cells and granule cells. Comparison of the depolarizing effects of the agonists with their neurotoxic effects on Purkinje cells and granule cells suggested that: (a) the ability to depolarize is a required condition for an agonist to be neurotoxic, (b) the magnitude of depolarization, rather than depolarizing potency, is the more pertinent determinant of neurotoxic potency and (c) resistance to the neurotoxicity of an agonist is not necessarily associated with resistance to its depolarizing actions. Histological studies indicated that the neurotoxicity of N-methyl-D-aspartate and kainate in immature cerebellar slices could largely not be replicated by veratridine (50 microM) or high extracellular K+ (124 mM) indicating that receptor-mediated ionic fluxes may be needed in addition to those caused by depolarization. Exposure of the slices to anoxia in the absence of glucose partially reproduced the toxicity of the receptor agonists. Application of ouabain for 30 min caused necrosis of all the cells which are vulnerable to the agonists but spared the cells which are not vulnerable. Profound ionic imbalance thus appears to be a sufficient explanation for amino acid neurotoxicity.

Neurotoxicity of excitatory amino acid receptor agonists in rat cerebellar slices: dependence on calcium concentration

In slices of developing rat cerebellum, a 30-min application of the excitatory amino acid receptor agonist, N-methyl-D-aspartate (NMDA), led to the necrosis of differentiating granule cells and deep nuclear neurones. The corresponding effect of another agonist, kainate, was the death of Golgi cells. The toxic effects of both agonists were prevented if the concentration of calcium in the exposing solution was reduced to 0.3 mM from the control level of 2.5 mM. A lesser reduction (to 1 mM) was enough to prevent 90% of the NMDA-induced necrosis of granule cells. The results indicate that an important component of the acute neurotoxic effects of excitatory amino acids is calcium-dependent and suggest reasons why this may not have been revealed in some previous studies.

In vitro neurotoxicity of excitatory acid analogues during cerebellar development

The neurotoxic effects of the selective excitatory amino acid receptor agonists, quisqualate, kainate and N-methyl-D-aspartate, were studied in slice preparations of cerebellum from rats at different stages of postnatal development. With increasing age, (i) Purkinje cells became more vulnerable to kainate and quisqualate but remained insensitive to N-methyl-D-aspartate (up to 300 microM); (ii) Golgi cells became more sensitive to kainate, quisqualate and N-methyl-D-aspartate; (iii) granule cells became more sensitive to kainate, less sensitive to N-methyl-D-aspartate and remained unaffected by quisqualate (up to 100 microM), and (iv) basket and stellate cells and, up to 14 days of age, neurones of the deep cerebellar nuclei, became more vulnerable to kainate and quisqualate, but their sensitivity to N-methyl-D-aspartate stayed the same. The neurotoxicity of N-methyl-D-aspartate, but not that of kainate in 8-day-old cerebellar slices was prevented by 2-amino-5-phosphonovaleric acid; tetrodotoxin did not affect the toxicity of the agonists in 8-day-old or adult slices. The results with kainate are consistent with other studies indicating an insensitivity of the immature brain to its neurotoxic effects, but suggest that this property is not a peculiarity of kainate. Alterations in excitatory potency can explain some of the observed developmental changes. However, other observations cannot readily be accounted for on the basis of either changes in excitatory potency, the functional maturation of cerebellar circuits, changes in synaptic density, or the developmental appearance of Ca2+ channels in susceptible cells, suggesting that additional factors play an important role in the neurotoxic effects of the excitants.

Sites of D-[3H]aspartate accumulation in mouse cerebellar slices

Slices of mouse cerebellar vermis, cut in the parasagittal plane, were incubated for various times (up to 3 h) in the presence of 1 microM D-[3H]aspartate, a non-metabolized substrate for the glutamate/aspartate carrier in brain tissue. Light microscopic autoradiography indicated that in regions away from the cut edges of the slices the amino acid accumulated in glia and granule cells. Relatively few grains were seen over Purkinje, Golgi, stellate and basket cells or over white matter. Grain counts over the granule cell layers in the middle parts of the slices indicated that after short (15 min) exposures to the labelled substrate, non-granule cell areas (which included glia) contained, on average, slightly more grains than granule cells but with longer exposures (1.5 and 3 h) the relative grain density over granule cells became much higher, possibly because glial uptake prevents D-[3H]aspartate gaining access to neuronal sites in adequate amounts during short incubations and/or because the longer incubations allow time for retrograde migration of the label from parallel fibre terminals to occur. The demonstration of selective uptake of D-[3H]aspartate into granule cells contrasts with previous autoradiographic results (possible reasons for which are discussed) and supports the notion that L-glutamate is the transmitter of granule cells. The results also have a bearing on the importance of the metabolic compartmentation of glutamate in relation to its proposed transmitter role.

Relative deficiency of serum IgA in the german shepherd dog: a breed abnormality

Serum immunoglobulin concentrations in clinically healthy German shepherd dogs were compared with those in mongrel dogs and Irish setters. Serum IgA concentrations were significantly lower in the German shepherd dogs than in the other two groups, while there were no differences in the concentrations of IgM or IgG. These findings suggest that production of IgA by gut associated lymphoid tissue may be abnormally low in the German shepherd dog.

Differential sensitivity of rat cerebellar cells in vitro to the neurotoxic effects of excitatory amino acid analogues

The neurotoxic effects of N-methyl-D-aspartate, quisqualate and kainate were studied in slice preparations of adult rat cerebellum. Only the inhibitory interneurones (basket, stellate and Golgi cells) were affected by N-methyl-D-aspartate in concentrations up to 300 microM. Quisqualate (10-100 microM) affected both Purkinje cells and inhibitory interneurones but spared granule cells. Kainate affected Purkinje cells and inhibitory interneurones at a concentration of 10 microM but at 30 microM also damaged granule cells. The relative potencies of these compounds and the vulnerability of the different cell types to their neurotoxic effects are in accordance with the predictions of the excitotoxic hypothesis and therefore do not support a special mechanism for kainate neurotoxicity in the cerebellum.

The mechanism of kainic acid neurotoxicity

The putative excitatory transmitters glutamate and aspartate, as well as their excitatory analogues, can kill neurones in the central nervous system and may thus be involved in the pathogenesis of various neurodegenerative disorders. Several studies have suggested that postsynaptic receptors are important in the mechanism of toxicity. However, presynaptic factors might also be involved because, in some brain areas, the neurotoxicity of kainate (a potent structural analogue of glutamate) is greatly reduced following elimination of afferent excitatory innervation, even though the postsynaptic excitatory potency of kainate may be unaltered in these conditions. The supply of glutamate from the afferent nerve endings has been suggested to be a necessary factor. Recently, Ferkany, Zaczec and Coyle concluded from studies on slices of mouse cerebellum that kainate activates presynaptic kainate receptors on parallel fibre terminals to release glutamate and that it is the postsynaptic interaction between kainate and the released amino acid that is instrumental in causing neuronal necrosis. The more direct evidence we report here does not support these conclusions.

Location of immunoglobulins and complement (C3) at the surface and within the skin of dogs

The serum proteins present at the skin surface in dogs were found to include the immunoglobulins (Ig) G, M and A, complement (C3) and albumin. Within the skin IgG and IgM were consistently found in the interstitial tissue throughout the dermis and were commonly present in the dermal blood vessels and hair papillae. IgA was undetectable or present in small amounts in the dermis in most of the 16 samples examined. It was found in the sweat glands, both in the lumen and the fundic epithelium, in some skin samples. C3 was demonstrated in the dermis and in the inter-cellular spaces of the stratum corneum. Elution of fresh skin specimens with phosphate buffered saline removed the majority of interstitial immunoglobulin and C3 from the dermis and some of the IgA from the sweat glands. IgM could still be demonstrated in the region of the epidermal basement membrane and C3 was not eluted from the stratum corneum. Removal of interstitial IgG and IgM facilitated the identification of immunoglobulin-bearing cells in the dermis. The distribution of IgG and IgM in dog skin is similar to that found in ruminants. Demonstration of IgA in the canine sweat gland fundus lends further support to the concept of IgA as a secretory immunoglobulin in the skin.

Epidermal structure and surface topography of canine skin

The stratum corneum of canine skin when measured in cryostat sections was found to have a mean thickness of 47.5 cell layers and measured 13.3 micrometers. The living epidermis was composed of three to six cell layers and measured 10.1 micrometers. Stratum corneum thickness was similar on the back and abdomen but was greater in the inguinal fold. Lipid was found in the distal intercellular spaces of the corneum to a mean depth of 34 cell layers. The surface topography of freeze dried canine skin was examined by scanning electron microscopy. The interfollicular stratum corneum was generally covered by a thin homogeneous film but in some areas hexagonal squames surrounded by amorphous globular material, which appeared to be oozing onto the surface, could be found. This material, which was also found sealing the bases of hair follicles and matting together the emerging hairs, appeared to be sebaceous lipid.