Search

In the course of a study on the pathogenesis of neuronal necrosis in severe hypoglycemia, the morphological characteristics reflecting reversible and irreversible neuronal lesions were examined as a function of time following normalization of blood glucose. To that end, closely spaced time intervals were studied in the rat cerebral cortex before, during, and up to 1 year after standardized pure hypoglycemic insults of 30 and 60 min of cerebral isoelectricity. Both the superficial and deep layers of the cerebral cortex showed dark and light neurons during and several hours after the insult. By electron microscopy (EM) the dark neurons were characterized by marked condensation of both karyoplasm and cytoplasm, with discernible, tightly packed cytoplasmic organelles. The light neurons displayed clustering of normal organelles around the nucleus with clearing of the peripheral cytoplasm. Some cells, both dark neurons and neurons of normal electron density, contained swollen mitochondria with fractured cristae. Light neurons disappeared from the cerebral cortex by 4 h of recovery. Some dark neurons in the superficial cortex and almost all in the deep cortex evolved through transitional forms into normal neurons by 6 h recovery. Another portion of the dark neurons in the superficial cortex became acidophilic between 4 and 12 h, and by EM they demonstrated karyorrhexis with stippled electron-dense chromatin. The plasma membrane was disrupted, the cytoplasm was composed of amorphous granular debris, and the mitochondria contained flocculent densities. These definitive indices of irreversible neuronal damage were seen as early as 4-8 h recovery. Subsequently, the acidophilic neurons were removed from the tissue, and gliosis ensued. Thus, even markedly hyperchromatic "dark" neurons are compatible with survival of the cell, as are neurons with conspicuous mitochondrial swelling. Definite nerve cell death is verified as the appearance of acidophilic neurons at which stage extensive damage to mitochondria is already seen in the form of flocculent densities, and cell membranes are ruptured. Our previous results have shown that hypoglycemic neocortical damage affects the superficial laminae, chiefly layer 2. The present results demonstrate that, following the primary insult, this damage evolves relatively rapidly within the first 4-12 h. We have obtained no evidence that additional necrotic neurons are recruited after longer recovery periods.

Rats were exposed to insulin-induced hypoglycemia resulting in periods of cerebral isoelectricity ranging from 10 to 60 min. After recovery with glucose, they were allowed to wake up and survive for 1 week. Control rats were recovered at the stage of EEG slowing. After sub-serial sectioning, the number and distribution of dying neurons was assessed in each brain region. Acid fuchsin was found to stain moribund neurons a brilliant red. Brains from control rats showed no dying neurons. From 10 to 60 min of cerebral isoelectricity, the number of dying neurons per brain correlated positively with the number of minutes of cerebral isoelectricity up to the maximum examined period of 60 min. Neuronal necrosis was found in the major brain regions vulnerable to several different insults. However, within each region the damage was not distributed as observed in ischemia. A superficial to deep gradient in the density of neuronal necrosis was seen in the cerebral cortex. More severe damage revealed a gradient in relation to the subjacent white matter as well. The caudatoputamen was involved more heavily near the white matter, and in more severely affected animals near the angle of the lateral ventricle. The hippocampus showed dense neuronal necrosis at the crest of the dentate gyrus and a gradient of increasing selective neuronal necrosis medially in CA1. The CA3 zone, while relatively resistant, showed neuronal necrosis in relation to the lateral ventricle in animals with hydrocephalus. Sharp demarcations between normal and damaged neuropil were found in the hippocampus. The periventricular amygdaloid nuclei showed damage closest to the lateral ventricles. The cerebellum was affected first near the foramina of Luschka, with damage occurring over the hemispheres in more severely affected animals. Purkinje cells were affected first, but basket cells were damaged as well. Rare necrotic neurons were seen in brain stem nuclei. The spinal cord showed necrosis of neurons in all areas of the gray matter. Infarction was not seen in this study. The possibility is discussed that a neurotoxic substance borne in the tissue fluid and cerebrospinal fluid (CSF) contributes to the pathogenesis of neuronal necrosis in hypoglycemic brain damage.

The neuropathological findings in two cases of irreversible hypoglycemic brain injury are described. A 26-year-old diabetic man injected insulin without adequate food intake and died after 2 months in coma. An 84-year-old nondiabetic man accidentally received 10 mg of glibenclamide and died after 3 months in relatively superficial coma. In the first case, an extensive necrotizing injury with gliosis was present in the cerebral cortex with temporal preponderance, as well as in the amygdalae and hippocampus. Lesions were also present in the putamen and caudate nucleus whereas the globus pallidus and thalamus were less severely destroyed. The distribution of the lesions was therefore somewhat different from that commonly seen in hypoxic-ischemic brain injury, which, together with some previously published data, suggests some difference in the pathogenesis of hypoglycemic vs. hypoxic-ischemic brain injury. In the second case only a slight loss of cortical neurons with secondary gliosis could be attributed to the hypoglycemic insult. This case demonstrates the danger of accidental intake of sulfonylurea preparations, which can cause an irreversible brain injury due to their prolonged hypoglycemic effect.

Part I of this paper has documented the evolution of dark neurons into acidophilic neurons in the superficial laminae as well as the reversion of dark neurons to normal neurons in the deep laminae of the cerebral cortex in hypoglycemic brain damage. The present study describes the temporal evolution of hypoglycemic brain damage in the hippocampus. The evolution of dark neurons to acidophilic neurons was confirmed in this brain region. Four additional problems were addressed: Firstly, delayed neuronal death was looked for, and was found to occur in areas of CA1 undergoing mild damage. However, it was not preceded by a morphological free interval, had ultrastructural characteristics distinct from delayed neuronal death in ischemia, and hence should be considered a distinct phenomenon. Secondly, the gradient in the density of neuronal necrosis in the rat hippocampal pyramidal cell band was exploited to test the hypothesis that a more severe insult causes a more rapid evolution of neuronal changes. This was found to be the case, with a temporal spectrum in the timing of neuronal death: Necrosis occurred already after 2 h medially in the subiculum, and was delayed by up to several weeks laterally in CA1. Thirdly, the almost universal sparing of CA3 pyramidal neurons after 30 min hypoglycemic isoelectricity was exploited to address the question of whether reactive changes, which could with certainty be deemed reversible, occur in CA3. Mitochondrial injury was seen in these cells, and was found to be recoverable. No reactive changes of the type previously described following ischemic insults were observed. Fourthly, the astrocytic and vascular response of the tissue was studied. A sequence of astrocytic changes representing structural and probably metabolic activation of astrocytes was seen, consisting of morphological indices of increased turnover of cellular components. Capillaries demonstrated endothelial pits, vesicles, and prominent microvilli hours to days after recovery. The results demonstrate that, in the hippocampal gyrus as in other brain regions, hypoglycemic brain damage is distinct from ischemic brain damage and likely has a different pathogenesis.

Thirty-eight male Wistar rats were exposed to insulin-induced hypoglycemia resulting in periods of cerebral isoelectricity ranging from 10 to 60 min. Plasma glucose levels during cerebral isoelectricity ranged from 0.12 mM to 1.36 mM. Control rats were injected with insulin, but hypoglycemia was terminated with glucose at the stage of large delta-wave EEG slowing. After recovery, the rats were allowed to wake up and survive for 1 wk. The number of dying neurons was assessed with acid-fuchsin/cresyl-violet-stained, whole-brain, subserial sections using direct visual counting of acidophilic, cytoclastic neurons. Brains from control rats that were not allowed to become isoelectric showed no dying neurons. Ten minutes of cerebral isoelectricity produced very minimal brain damage. The density of neuronal necrosis was positively related to the number of minutes of cerebral isoelectricity up to the maximum examined period of 60 min, but showed no correlation with the blood sugar levels. The cerebral cortex, hippocampus, caudate nucleus, spinal cord, and, to a lesser extent, cerebellar Purkinje cells were affected. The distribution of neuronal necrosis was not identical with that seen in ischemia, but, rather, suggested a CSF-borne neurotoxin operant in contributing to the pathogenesis of neuronal necrosis in hypoglycemic brain damage. Neuronal death does not occur in hypoglycemia unless the EEG becomes isoelectric, whatever the blood sugar level. Serious brain damage does not occur until electrocerebral silence has been established for at least several minutes. Neuronal death accelerates after 30 min of EEG isoelectricity in the rat.(ABSTRACT TRUNCATED AT 250 WORDS)

A detailed light- and electron-microscopic study of the damage to the rat dentate gyrus in hypoglycemia was undertaken, in view of the previously advanced hypothesis that hypoglycemic nerve cell injury is mediated by a released neurotoxin. The distribution of neuronal necrosis showed a relationship to the subarachnoid cisterns. Electron microscopy of the dentate granule cells and their apical dendrites revealed dendrosomal, axon-sparing neuronal pathology. Dentate granule cells were affected first in the dendrites in the outer layer of the stratum moleculare, sparing axons of passage and terminal boutons. Subsequently, the neuronal perikarya were affected, and Wallerian degeneration of axons followed. Cell membrane abnormalities preceded the appearance of mitochondrial flocculent densities and degradation of the cytoskeleton, and are suggested to be early lethal changes. The observed early dendrotoxic changes, and the dendrosomal, axon-sparing nature of the lesion implicate an excitotoxin-mediated neuronal necrosis in hypoglycemia.

Previous results have shown that severe, prolonged hypoglycemia leads to neuronal cell damage in, among other structures, the cerebral cortex and the hippocampus but not the cerebellum. In order to study whether or not this sparing of cerebellar cells is due to preservation of cerebellar energy stores, hypoglycemia of sufficient severity to abolish spontaneous EEG activity was induced for 30 and 60 min. At the end of these periods of hypoglycemia, as well as after a 30 min recovery period, cerebellar tissue was sampled for biochemical analyses or for histopathological analyses or for histopathological analyses by means of light and electron microscopy. After 30 min of hypoglycemia. the cerebellar energy state, defined in terms of the phosphocreatine, ATP, ADP, and AMP concentrations, was better preserved than in the cerebral cortex. After 60 min, gross deterioration of cerebellar energy state was observed in the majority of animals, and analyses of carbohydrate metabolites and amino acids demonstrated extensive consumption of endogenous substrates. In spite of this metabolic disturbance, histopathologic alterations were surprisingly discrete. After 30 min, no clear structural changes were observed. After 60 min, only small neurons in the molecular layer (basket cells) were affected, while Purkinje cells and granule cells showed few signs of damage. The results support our previous conclusion that the pathogenesis of cell damage in hypoglycemia is different from that in hypoxia-ischemia and indicate that other mechanisms than energy failure must contribute to neuronal cell damage in the brain.