JNK3 is abundant in insulin-secreting cells and protects against cytokine-induced apoptosis.

AIMS/HYPOTHESIS: In insulin-secreting cells, activation of the c-Jun NH(2)-terminal kinase (JNK) pathway triggers apoptosis. Whereas JNK1 and JNK2 are ubiquitously produced, JNK3 has been described exclusively in neurons. This report aims to characterise the expression and role in apoptosis of the three JNK isoforms in insulin-secreting cells exposed to cytokines. METHODS: Sections of human and mouse pancreases were used for immunohistochemistry studies with isoform-specific anti-JNK antibodies. Human, pig, mouse and rat pancreatic islets were isolated by enzymatic digestion and RNA or protein extracts were prepared. RNA and protein levels were determined by quantitative RT-PCR and western blotting respectively, using JNK-isoform-specific primers and isoform-specific antibodies; activities of the three JNK isoforms were determined by kinase assays following quantitative immunoprecipitation/depletion of JNK3. JNK silencing was performed with small interfering RNAs and apoptotic rates were determined in INS-1E cells by scoring cells displaying pycnotic nuclei. RESULTS: JNK3 and JNK2 mRNAs are the predominant isoforms expressed in human pancreatic islets. JNK3 is nuclear while JNK2 is also cytoplasmic. In INS-1E cells, JNK3 knockdown increases c-Jun levels and caspase-3 cleavage and sensitises cells to cytokine-induced apoptosis; in contrast, JNK1 or JNK2 knockdown is protective. CONCLUSIONS/INTERPRETATION: In insulin-secreting cells, JNK3 plays an active role in preserving pancreatic beta cell mass from cytokine attacks. The specific localisation of JNK3 in the nucleus, its recruitment by cytokines, and its effects on key transcription factors such as c-Jun, indicate that JNK3 is certainly an important player in the transcriptional control of genes expressed in insulin-secreting cells.

Role of the JNK pathway in NMDA-mediated excitotoxicity of cortical neurons.

Excitotoxic insults induce c-Jun N-terminal kinase (JNK) activation, which leads to neuronal death and contributes to many neurological conditions such as cerebral ischemia and neurodegenerative disorders. The action of JNK can be inhibited by the D-retro-inverso form of JNK inhibitor peptide (D-JNKI1), which totally prevents death induced by N-methyl-D-aspartate (NMDA) in vitro and strongly protects against different in vivo paradigms of excitotoxicity. To obtain optimal neuroprotection, it is imperative to elucidate the prosurvival action of D-JNKI1 and the death pathways that it inhibits. In cortical neuronal cultures, we first investigate the pathways by which NMDA induces JNK activation and show a rapid and selective phosphorylation of mitogen-activated protein kinase kinase 7 (MKK7), whereas the only other known JNK activator, mitogen-activated protein kinase kinase 4 (MKK4), was unaffected. We then analyze the action of D-JNKI1 on four JNK targets containing a JNK-binding domain: MAPK-activating death domain-containing protein/differentially expressed in normal and neoplastic cells (MADD/DENN), MKK7, MKK4 and JNK-interacting protein-1 (IB1/JIP-1).

Relationships between neuronal death and the cellular redox status. Focus on the developing nervous system.

During the development of the nervous system, a large number of neurons are eliminated through naturally occurring neuronal death. Many morphological and biochemical properties of such dying neurons are reminiscent of apoptosis, a type of death involving the action of genetically-programmed events but also epigenetic phenomena including oxidative stress. The following review contains three parts focusing respectively on basic knowledge of neuronal death and redox regulation, the mechanisms involved in neuronal death which are ordered in three sequential phases, and on the complex relations between neuronal fate and the redox status. Finally, we point out that oxidants are not always detrimental for neuronal survival. On the one hand, dying neurons often display signs of oxidative stress, including an elevation of their intracellular concentration of free radicals. Antioxidants may reduce the extent of neuronal death, suggesting a causal implication of free radicals in the death-process. On the other hand, at high concentrations antioxidants may lose their protective effects on developing neurons, and a non-lethal oxidative stress may potentiate the protective effects of other agents. These data suggest that free radicals, perhaps through their effects on cellular signalling pathways, may have positive effects on neuronal survival, provided that their intraneuronal concentrations are maintained at low levels. Much evidence suggests that the neuronal redox status must be maintained within a narrow range of values compatible with survival. Antioxidants may protect neurons subjected to an oxidative stress following axotomy or trophic factor-deprivation; but excessive reduction may become equally detrimental for neurons.

Neuronal death after perinatal cerebral hypoxia-ischemia: Focus on autophagy-mediated cell death.

Neonatal hypoxic-ischemic encephalopathy is a critical cerebral event occurring around birth with high mortality and neurological morbidity associated with long-term invalidating sequelae. In view of the great clinical importance of this condition and the lack of very efficacious neuroprotective strategies, it is urgent to better understand the different cell death mechanisms involved with the ultimate aim of developing new therapeutic approaches. The morphological features of three different cell death types can be observed in models of perinatal cerebral hypoxia-ischemia: necrotic, apoptotic and autophagic cell death. They may be combined in the same dying neuron. In the present review, we discuss the different cell death mechanisms involved in neonatal cerebral hypoxia-ischemia with a special focus on how autophagy may be involved in neuronal death, based: (1) on experimental models of perinatal hypoxia-ischemia and stroke, and (2) on the brains of human neonates who suffered from neonatal hypoxia-ischemia.

Neuronal death during development in the isthmo-optic nucleus of the chick: sustaining role of afferents from the tectum.

Neurons have been counted in the isthmo-optic nucleus following lesions of the optic tectum, its main source of afferents. Late lesions, made at 10.8-12.2 days of incubation, were employed as they cause the fewest non-specific side effects. The lesions spared the isthmo-optic tract, and although they caused many retinal ganglion cells to die, the degeneration did not spread to the inner nuclear layer, which contains the target cells of the isthmo-optic fibers. Hence the effects on the isthmo-optic nucleus were due to its being deprived of afferents. Even in unoperated embryos, 60% of the isthmo-optic neurons are known to die between embryonic days 12 and 17. The tectal lesions greatly increased the cell loss ipsilaterally; this was due to cell death, since other explanations such as migration away or differential cellular shrinkage have been ruled out. The fact that additional neuronal death occurred mainly during the latter half of the period of natural cell death implies that the tectal afferents are important for the survival of the isthmo-optic neurons during this latter half, but not before.

Some visual and other connections to the cerebellum of the pigeon.

By anatomical techniques it has been shown that folia VIc-IXc of the pigeon cerebellum receive inputs from the following groups of neurons: the medial and lateral pontine nuclei, the superficial synencephalic nucleus, the medial spiriform nucleus, the inferior olive, and the deep cerebellar nuclei. From all but the last of these, the projection is mainly crossed, though the uncrossed component from the lateral pontine nucleus is not insubstantial. The input from the superficial synencephalic nucleus provides a direct pathway from the retina to the cerebellum (folia VIc, VII, VIII and IXc). Less direct visual pathways reach the cerebellum via the following routes: (i) the superficial synencephalic nucleus projects ipsilaterally to the lateral pontine nucleus and sparsely to the inferior olive; (ii) the tectum projects ipsilaterally to the lateral and medial pontine nuclei, though the latter connection is sparse. In electrophysiological experiments, the importance of the tecto-pontine component of the projection has been demonstrated by cooling the tectum while recording visual responses from the cerebellum. The visual receptive fields of pontine cells have been analysed. They vary in extent from 10 degrees to the whole monocular field. They respond best to moving targets, preferring speeds of 20 to 60 degrees/second, and are usually direction-selective.

The generation and migration of the chick's isthmic complex.

By thymidine autoradiography it is shown that the entire isthmic complex is arranged according to a single and very finely grained gradient of proliferation that manifests itself in each of the component cell groups individually. Ventral neurons are generated first, dorsal last. The neurons in the various isthmic cell groups complete their final phase of DNA synthesis mainly between the following stages (of Hamburger and Hamilton, '51): N. semilunaris, HH-22 to HH-24; N. lemnisci lateralis pars ventralis, HH-23 to HH-25; N. lemnisci lateralis pars dorsalis, HH-24 to HH-27; N. isthmi partes principales magnocellularis and parvocellularis, HH-24 to HH-29; N. isthmo-opticus, HH-27 to HH-31. By thymidine autoradiography combined with the peroxidase retrograde transport technique, it is shown that the retinopetal neurons ventral to N. isthmo-optic neurons and substantially later than the nonretinopetal neurons amongst which they are scattered. Using similar autoradiographic methods but in embryos fixed at various ages, it is then shown that the isthmic alar plate gives rise to all the cell groups of the isthmic complex except possibly N. isthmi pars principalis magnocellularis. In confirmation of Vaage's ('73) account, it is shown that two main migrations, "dorsal" and dorsolateral", orignate from the dorsal and dorsolateral parts of the germinal epithelium, traveling first laterally and then ventrally, and that some cells from the former subsequently join the latter to create the mixed migration. The dorsal migration gives rise to N. semilunaris, N. isthmi pars principalis parvocellularis and N. isthmo-opticus, whereas the mixed gives rise to N. lemnisci lateralis partes dorsalis and ventralis and to much of the adjacent reticular formation. The migration is extremely rapid (15-30 micrometer per hour for the dorsal migration), but much of this speed may be due to passive bulk displacement, since there seems to be little relative movement between adjacent cells.

Abrupt loss of dependence of retinopetal neurons on their target cells, as shown by intraocular injections of kainate in chick embryos.

We have examined the capacity of neurons in the chick isthmo-optic nucleus (ION) to survive when their target neurons in the contralateral retinal are destroyed by intraocular injections of kainate (KA) at different stages in development. The retinal vulnerability to KA builds up progressively from embryonic day 10 (E10) until a plateau is reached at E15 (see accompanying paper); and the effects on the ION increase in parallel, almost all the ION neurons being rapidly lost after the E15 injections. KA injection before E15 lesioned only part of the retina and caused degeneration only in the topographically corresponding region of the ION. Near the end of the natural cell death period in the ION (E17), this initial dependence on the target cells is rapidly lost. Already at E16 the injections kill less ION neurons, and by E19 they kill none of them. The ION neurons have become completely insensitive to the KA injections and appear normal more than 4 months later, although axotomy (by eye removal) at a similar age would by then have killed them. The ectopic ION neurons, scattered outside the ION but projecting to the retina, are never affected by KA injections at any age.

Changes in the nuclei of dying neurons as studied with thymidine autoradiography.

Thymidine autoradiography has been used at light and electron microscopic levels to elucidate intracellular events during the death in chick embryos of isthmo-optic neurons deprived of trophic maintenance from their axonal target organ, the retina. When the intense cytoplasmic vacuolization described in the accompanying paper (Hornung, Koppel, and Clarke, J. Comp. Neurol. 283:425-437, '89) was beginning, the nuclei also underwent profound changes. They became more electron dense and shrank; their membranes became more sharply defined and convoluted; they sometimes contained pyknotic balls, but apparently only in the early stages of cell death; all lost more than half of their content of DNA, some of which was transferred to the largest kind of cytoplasmic vacuole. This transfer may have involved the budding off of nuclear regions containing pyknotic balls. The cells continued to survive for a day or 2 after these severe losses of nuclear DNA, sustaining intense endocytic activity. Pronounced unscheduled DNA synthesis occurred in the nuclei, but this was insufficient to replace the lost DNA.

Changes in lamination and neuronal survival in the isthmo-optic nucleus following the intraocular injection of tetrodotoxin in chick embryos.

We have studied how the development of the isthmo-optic nucleus (ION) is affected by electrical activity in the ION's axonal target territory, the contralateral retina. Electrical activity was blocked or reduced in the retina for various periods by tetrodotoxin injected intraocularly in different doses. The effects on the morphology of the retina appear to have been minor. During the ION's period of naturally occurring neuronal death (embryonic days 12 to 17), the injections substantially reduced this neuronal death and disrupted the development of lamination in the contralateral ION; there was also a lesser reduction in neuronal death in the ipsilateral ION. The dose of tetrodotoxin required to affect lamination was lower than that affecting neuronal death. Thus, the effects on neuronal death and on lamination were independent, since either could occur without the other. These effects were mediated by retrograde signals (probably two or more) from the eye; they occurred too early for the alternative anterograde route via the optic tectum (which projects to the ION) to be responsible. After embryonic day 17, the ION's response to intraocular tetrodotoxin changes abruptly from increased survival to total and rapid degeneration.

Maintenance of targeting errors by isthmo-optic axons following the intraocular injection of tetrodotoxin in chick embryos.

We have studied the role of electrical activity in the elimination of axonal targeting errors, which is a normal process in brain development. The experiments were focused on the isthmo-optic nucleus (ION), which, in adults, projects in topographical order on the contralateral retina. During embryogenesis, however, a few isthmo-optic neurons project to the ipsilateral retina, and many project to topographically inappropriate parts of the contralateral one; both kinds of targeting error are known to be eliminated by the deaths of the parent neurons. We injected tetrodotoxin (TTX) intraocularly at embryonic days 13 and 15 and, on the latter, applied a retrograde label to the retina of the same eye. Embryos were fixed at embryonic day 17. In some embryos, the label was a peripherally placed fleck of the carbocyanine dye "diI"; the resulting retrogradely labeled neurons in the contralateral ION were much more widely scattered in the TTX-injected embryos than in controls (errors in topography). In other embryos, the label was a solution of rhodamine-B-isothiocyanate (RITC) injected into the vitreous body; this yielded several ipsilaterally labeled isthmo-optic neurons in the TTX-injected embryos, but virtually none in the controls. The numbers of both kinds of aberrantly projecting neuron approached those previously reported near the beginning of the ION's period of neuronal death. We conclude that electrical activity plays an important role in the elimination of axonal targeting errors in the chick embryo's isthmo-optic system.

Spatiotemporal gradients of kainate-sensitivity in the developing chicken retina.

We have studied the age-dependence of the effects of kainate (KA) on the chick retina as a prelude to the accompanying paper on the effects of target-removal on the isthmo-optic nucleus. KA was injected into the eyes of chick embryos and chicks at different ages, and the retinas were fixed a few hours or several days later. The former group of retinas was scanned for pyknotic cells. The earliest age at which KA caused pyknosis was embryonic day 10 (E10), when pyknotic cells appeared in a ventrotemporal patch in the amacrine sublayer near the fundus. Over the next two days the sensitive region expanded tangentially, reaching the periphery first temporally, then nasally. Only after E12 did the KA cause pyknotic cells to occur also in the bipolar sublayer, where the sensitivity spread in the same spatiotemporal sequence as the initial wave, but two days later. Cell loss was examined in embryos that survived a week or more after the KA injection. Substantial cell depletion was found in both the inner nuclear and ganglion cell layers, but only when the injection had been made after E12. With progressively later injections, the depleted zone expanded in the same spatiotemporal sequence as described above, until at E15 the injections caused depletion throughout the entire extent of the retina. The reasons for the lack of cell depletion after KA injections made before E12 are discussed. Cell counts in the ganglion cell layer and studies of anterograde transport of intravitreally injected peroxidase along the retinofugal fibers showed that about half the ganglion cells (including the displaced ganglion cells) pass through a period of vulnerability to the KA injections, to which they subsequently become sensitive.

Dendritic reorientation and cytolamination during the development of the isthmo-optic nucleus in chick embryos.

In the mature isthmo-optic nucleus (ION, source of efferents to the contralateral retina), the neuronal perikarya are generally described as being arranged in a single convoluted lamina surrounding a U-shaped region of neuropil, into which their highly polarized (unidirectional) dendritic arbors project perpendicularly. We find, however, that the details are more complicated than this description suggests, and are variable, as might be expected if the ION is self-organized through neuron-to-neuron interactions in development. The laminated conformation of the ION first appears at embryonic day (E) 14. Our previous experiments indicate that this involves the displacement of perikarya and is not due to sculpting by neuronal death. We here present a quantitative demonstration that the dendritic arbors reorient during the period of lamination. At E11, they are already highly polarized, but their directions are different from those in the adult, being mostly medio-rostro-ventral. Then, between E11 and E13, the arbors in the border region of the ION undergo major changes in their direction of polarization, projecting towards the center of the ION. The arbors within the core of the ION make more subtle changes. The dendritic reorganization seems to be intrinsically linked to the process of cytolamination, since the two events occur synchronously and disruption of either affects the other. Mechanisms are discussed; interaction with afferents is not responsible for lamination.

The development of the isthmo-optic tract in the chick, with special reference to the occurrence and correction of developmental errors in the location and connections of isthmo-optic neurons.

The development of the centrifugal projection to the chick retina in the isthmo-optic tract (IOT) has been studied by the retrograde transport of the enzyme marker horseradish peroxidase (HRP) injected into the eye at various times during the incubation period. Some neurons in the isthmo-optic nucleus (ION--the cells of origin of the IOT) can first be labeled from the eye following injections on the tenth day of incubation; after injections on the twelfth day or later, about 95% of the neurons can be so labeled. It follows from this that the axons of virtually all the neurons in the ION (including the 60% which normally degenerate between the thirteenth and seventeenth days of incubation) reach the contralateral eye. Since in 12-day old embryos the IOT is between six and seven millimeters in length and HRP can be identified in the perikarya of ION neurons within three and one-half hours, the rate of retrograde transport in the system must be of the order of 48 mm/day. A similar time is required for HRP to appear in the perikarya of ION neurons in post-hatched chicks in which the length of the IOT is estimated to be about 14 mm. This suggests that at some time during the latter half of the incubation period there is a significant acceleration in the rate of retrograde transport, similar to that found for anterograde axonal transport in the chick and rabbit visual systems.

Endocytosis and autophagy in dying neurons: an ultrastructural study in chick embryos.

In an effort to understand naturally occurring neuronal death in the developing isthmo-optic nucleus, we have accentuated one of its most probably causes, failure to receive adequate trophic maintenance from the axonal terminal zone in the retina, and have studied the dying neurons ultrastructurally. Retrograde trophic maintenance was blocked by means of intraocularly injected colchicine, which caused all the isthmo-optic neurons to die by just one of the two or more kinds of cell death that they undergo during normal development. The present paper deals with the very prominent cytoplasmic aspects of this kind of cell death, notably the uptake of exogeneous horseradish peroxidase and autophagy. There were also nuclear changes, which are dealt with mainly in the accompanying paper (Clarke and Hornung, J. Comp. Neurol. 283:438-449,'89). Numerous cytoplasmic vacuoles occurred in both soma and dendrites, and they were of three main kinds, of which the smallest (less than 0.5 microns diameter) had unstructured contents, whereas the larger two (1-2 microns and 2-7 microns) were secondary lysosomes (mostly residual bodies). Intravascularly injected horseradish peroxidase labeled all three kinds of vacuole but not the free cytoplasm, indicating that the uptake was by endocytosis rather than by leakage through holes in the membrane, as is confirmed by our failure to detect any such holes. We suspect that the smallest vacuoles are the primary endosomes, that these subsequently fuse with vacuoles of the intermediate kind, and that the largest vacuoles are formed by the fusion of these latter. The purpose of the endocytosis may be to channel the plasma membrane piecemeal into the lysosomes for destruction.

Development of two morphological types of retinopetal fibers in chick embryos, as shown by the diffusion along axons of a carbocyanine dye in the fixed retina.

Centrifugal fibers to the retinas of chick embryos and hatched chicks have been examined and traced following staining by diffusion along their axonal membranes of the carbocyanine dye DiI in fixed tissue. In the older embryos and hatched chicks, the report of Dogiel (Arch. Mikrosk. Anat. 44:622-648, 1895) has been confirmed that there are two very different morphological types of centrifugal fiber. The restricted type ends as a relatively thick fiber, lacking varicosities, that runs for a short distance in the most sclerad level of the inner plexiform layer before terminating in a pericellular nest overlying the flask-shaped body of a single amacrine cell. Thin filaments occasionally leave the pericellular net, apparently to terminate on adjacent cells. The widespread type also runs in the most sclerad level of the inner plexiform layer, but it is thin, varicose, and highly branched, and its terminal arbor may span more than 1 mm, remaining at the same level. Both types of terminal arbor issue from parent axons in the optic fiber layer of the retina. A single parent axon gives either a single terminal fiber of the restricted type or several terminals of the widespread type, but never a mixture of the two. It is argued that the restricted and widespread types originate respectively from the neurons of the contralateral isthmo-optic nucleus and from the "ectopic" neurons scattered outside the isthmo-optic nucleus. In development, the centrifugal fibers reach the retina between E9 and E10 and initially run radially in the optic fiber layer, parallel to the retinofugal fibers but avoiding the dorsal retina. They dive into the inner plexiform layer at about E12. By E13, the terminal arbors are forming, and the widespread and restricted types can already be distinguished. The widespread type continues to increase its territory until about E18, and then appears to remain stable, whereas the restricted type attains its maximum ramification between E13 and E15 and then contracts. Prior to the retraction, the terminal territories of the restricted type fibers overlap, which may provide the anatomical basis for the interaxonal competition that apparently contributes to neuronal death in the isthmo-optic nucleus between E13 and E16. Axons of ganglion cells exhibit transient side branches between E11 and E13; these never reach as deep as the level where the centrifugal fibers run.

The time of origin and the pattern of survival of neurons in the isthmo-optic nucleus of the chick.

The time of origin of the cells in the isthmo-optic nucleus (ION--the nucleus of origin of centrifugal fibers to the avian retina) has been determined in the chick by a variant of the cumulative labeling method, using 3H-thymidine autoradiography. All the neurons of the ION are generated (i.e., pass through their last phase of DNA synthesis) over a 50-hour period between the latter part of the fifth and the seventh days of incubation (stages 28--31 of the Hamburger and Hamilton ['51] series) but the cells come to be assembled within the nucleus along a distinct temporo-spatial gradient. The earliest-formed cells occupy the ventrolateral part of the nucleus while the last neurons to be generated come to lie along its dorsomedial margin. When the nucleus is numerically complete, around the eleventh day of incubation, it contains about 22,000 neurons, but between the thirteenth and seventeenth days (stages 39 through 43) this number is produced by nearly 60% to about 9,500 cells. Following the radical extirpation of the one optic cup or the circumscribed removal of the neural retina early on the third day of incubation, the cell loss in the contralateral ION is greatly accentuated, so that by the eighteenth day of incubation no cells remain in the nucleus. Serial counts of the numbers of cells in the nucleus on the side of the eye (or retinal) removals show that it, too, undergoes additional cell degeneration, so that by the end of the phase of naturally occurring neuronal loss, only about half the normal number of cells persist in these experimental animals.

Response of macrophage/microglial cells to experimental neuronal degeneration in the avian isthmo-optic nucleus during development.

Blockade of the retrograde axonal transport of isthmo-optic nucleus (ION) neurons in the avian embryo results in their massive degeneration. We used this system to investigate the response of macrophage/microglial cells to neuronal degeneration in the embryonic brain. Colchicine was injected into the right eye of quail or chick embryos at a time when the survival of ION neurons depends on retrograde trophic support from the retina, and the chronology of the subsequent macrophage/microglial response in the ION was analyzed. This response was restricted to the ION contralateral to the injected eye; no modifications of the normal state were observed in the surrounding parenchyma or in the opposite ION, used as control. The response was first detected 18 hours after the colchicine injection (18 hours pi), when an increase of the macrophage/microglial cell number was evident. The number of these cells in the affected ION increased, peaking at 40-48 hours pi. At later survival times, macrophage/microglial cells were progressively less abundant in the affected ION, which gradually diminished in size. At 120 hours pi the only remnant of the ION was a small cluster of macrophage/microglial cells, surrounded by a clear area with scarce nonmicroglial cells, in the region formerly occupied by the ION. This study reveals that a strong macrophage/microglial response occurs in the embryonic brain in response to neuronal degeneration but that these cells do not trigger the neuronal death, as they only appear after pyknotic fragments are already observable.

A stain for ischaemic or excessively stimulated neurons.

Loyez' myelin stain sometimes impregnates the perikarya and processes of neurons. This is rare in normal adult mice, but can be provoked by various means including electrical stimulation and mild ischaemia shortly before fixation. Passing a direct current or a train of pulses across the dura of the cerebral cortex produces a patch at the site of stimulation in which the neurons, but not the glia, are Loyez-impregnated; pulse-trains evoke such impregnation even at remote sites, presumably through orthodromic and/or antidromic activation. The strength and duration of stimulation necessary to provoke impregnation are below the threshold for causing an overt lesion. Local ischaemia produced by irreversibly occluding a small subarachnoid artery for as little as 7 min evokes a small patch of impregnated neurons in the superficial layers adjacent to the occluded artery. After slightly longer-lasting (e.g. 17 min) ischaemia, the impregnated zone spans most of the cortical depth. Mild global ischaemia produced by irreversibly occluding one or both common carotid arteries 30 min before fixation causes large numbers of neurons to be impregnated in many parts of the brain. That interventions as different as electrical stimulation and arterial occlusion both lead to neuronal impregnation may most plausibly be explained by the fact that they both cause cells to be depolarized. The method will be useful for visualizing overstimulated or ischaemic neurons, and may be applicable to the tracing of neural pathways.

Cooperation between glutathione depletion and protein synthesis inhibition against naturally occurring neuronal death.

It is generally agreed that naturally-occurring neuronal death in developing animals is dependent on the synthesis of proteins. Oxidative stress, as when intracellular concentrations of free radicals are raised or when cell constituents such as membrane lipids or protein thiols are oxidized, is also involved in various types of neuronal death. In the present report, we show that the number of naturally dying retinal cells in the chick embryo can be reduced by intraocular injections of cycloheximide, an inhibitor of protein synthesis. L-buthionine-[S,R]-sulfoximine, an inhibitor of glutathione synthesis, can either enhance or diminish the cell death, depending on the conditions of treatment. Moreover, when the two inhibitors are combined, L-buthionine-[S,R]-sulfoximine potentiates the neuroprotective effects of cycloheximide. Measurements of retinal glutathione concentration and protein synthesis show the specificity of the treatments: buthionine-sulfoximine diminishes glutathione concentrations but not protein synthesis whereas cycloheximide inhibits protein synthesis without decreasing glutathione concentrations. Naturally-occurring neuronal death thus seems to involve the synthesis of proteins, and is also influenced by oxidative phenomena. Our results extend previous data in tectal-lesioned embryos, and suggest that a moderate, non-lethal oxidative stress can enhance the resistance of ganglion cells that might otherwise have died (spontaneously or following axotomy) owing to insufficient retrograde trophic support.