Light and Electron Microscopic Evaluation of Hydrogen Ion-Induced Brain Necrosis

Journal of Cerebral Blood Flow and Metabolism 7:625-632 1987 Raven Press, Ltd" New York Light and Electron Microscopic Evaluation of Hydrogen Ion-Induced Brain Necrosis *c. K. Petito, tr. P. Kraig, and two A. Pulsinelli *Departments of Pathology-Neuropathology and tneurology. New York Hospital-Cornell University Medical College Summary: Excessive accumulation of hydrogen ions in the brain may play a pivotal role in initiating the necrosis seen in infarction and following hyperglycemic augmentation of ischemic brain damage. To examine possible mechanisms involved in hydrogen ion-induced necrosis, sequential structural changes in rat brain were examined following intracortical injection of sodium lactate solution (ph 4.5), as compared with injections at ph 7.3. Following ph 7.3 injection, neuronal swelling developed between 1 and 6 h, but only a needle track wound surrounded by a thin rim of necrotic neurons and vacuolated neuropil was present 24 h after injection. In contrast, ph 4.5 injection produced neuronal necrosis as soon as I h after injection, followed by necrosis of astrocytes and intravascular thrombi at 3 and 6 h. Alterations common to both groups included vascular permeability to horseradish peroxidase, dilation of extracellular spaces, astrocyte swelling, capillary compression, and vascular stasis. These data suggest that neurons, astrocytes, and endothelia can be directly damaged by increased acid in the interstitial space. Lethal injury initially appeared to affect neurons, while subsequent astrocyte necrosis and vascular occlusion may damage tissue by secondary ischemia. Key Words: Astrocytes-Intracortical injection Ischemia-Lactic acid- N eurons-ultrastructure. Excessive accumulation of hydrogen ions in brain may play a pivotal role in initiating the necrosis seen in infarction or following hyperglycemic augmentation of ischemic brain damage in experimental animals (Myers and Yamaguchi, 1977; Myers, 1979; Ginsberg et ai., 1980; Welsh et ai., 1980; Kalimo et ai., 1981; Rehncrona et ai., 1981; Pulsinelli et ai., 1982; Siesjo, 1985; Kraig et ai., 1986; Smith et ai., 1986). Similarly, elevated blood glucose may worsen stroke outcome in man (Pulsinelli et ai., 1983). Furthermore, hydrogen ions are directly neurotoxic in vivo and in vitro. Intracortical injections of hydrogen ions produced brain necrosis when a specific ph threshold had been reached (Pulsinelli and Petito, 1982; Kraig et ai., 1987). A ph-dependent threshold of necrosis also occurs in cultured astrocytes (Norenberg et ai., 1987) as well as dissociated brain cells and explants Received May 8, 1987; accepted May 19, 1987. Address correspondence and reprint requests to Dr. C. K. Petito at Department of Pathology, Cornell University Medical College, 1300 York Avenue, New York, NY 10021, U. S. A. Abbreviation used: BBB, blood-brain barrier. exposed to graded concentrations of lactic acid (Goldman et ai., 1987). The mechanisms underlying hydrogen ion-induced brain necrosis are unknown. Low tissue ph may denature proteins and nucleic acids (Kalimo et ai., 1981), although this may not occur at levels reached in ischemia (Kraig and Wagner, 1987). Astrocyte necrosis (Plum, 1983), caused by cellular acidosis and concomitant failure of plasma membrane transport systems (Kraig et ai., 1987) or irondependent, free radical formation (Pulsinelli et ai., 1985) may be alternative explanations. This present study investigated mechanisms involved in hydrogen-ion induced changes in rat brain by examining structural brain changes following intracortical injection of sodium lactate solution at ph 4. 5. Data showed the injection produced a maximal decrease in interstitial ph of 5.16 ± O. 14 (Kraig et ai., 1987), similar to that reached in hyperglycemic ischemia (Kraig et ai., 1985). These results were then compared with intracortical injections of sodium lactate at ph 7. 3. Prior studies have indicated that cortical necrosis occurs 24 h after a ph 4. 5 injection, whereas the ph 7. 3 625

626 C. K. P ETlTO ET AL. injection only produces damage indistinguishable from a needle tract wound alone (Pulsinelli and Petito, 1982; Kraig et ai., 1987). MATERIALS AND METHODS Adult male Wi star rats, 225-250 g, were anesthetized with halothane and placed into a stereotaxic head mount. A midline skin incision was made over the cranium, the soft tissues removed, and two small burr holes made through the parasagittal calvaria with a dental drill. A tail artery cannula was placed for continuous measurement of systemic blood pressure and intermittent sampling of arterial blood gases and ph levels. A rectal temperature probe was placed, and body temperature was maintained at 37 C during and after surgery as needed. Ten microliters of 150 mm sodium lactate solution, ph 4.5 or ph 7.3, was continuously injected at 0.5 /-LI/min for 20 min first into the left and then into the right parietal cortex with a fused silica injection needle (170 /-Lm diameter; Supelco, Inc., #2 1361) connected to a microinfusion pump (Raze I A-99, Stamford, CT, U.S.A.). Following the injection into the right parietal cortex, the needle was withdrawn, the skin defect closed with a metal clip, and the animal allowed to recover from anesthesia. Animals in both groups were allowed to survive I, 3, and 6 h following intracortical injections, and the animals injected with ph 7.3 solution were also examined at 24 h. Twenty minutes prior to death, the animals were injected intramuscularly with 0.5 mg/ioo g diphenhydramine hydrochloride (Cotran and Karnovsky, 1967; Deimann et al., 1976). Five minutes prior to death, animals received 100 mg horseradish peroxidase (Sigma Chemical Co., St. Louis, MO. Type II) dissolved in I ml of 2% Evans blue via a tail artery cannula. At the time of death, the animal was deeply anesthetized with 1.6 gm/kg intraperitoneal urethane. Following a thoracotomy, 0.1 ml of heparin (10,000 U/m!), was injected into the left ventricle of the heart, and the animal was perfusion-fixed with 0.1 M sodium cacodylate buffer, ph 7.4, for 30 s, followed by 2% paraformaldehyde-2.5% glutaraldehyde in 0.1 M cacodylate buffer, ph 7.4, for 15 min. The animals were decapitated, and the brains removed after 2 h or overnight and placed in cacodylate buffer. A l-cm coronal section surrounding the needle track was cut, serially sectioned at 40 /-L with a vibratome, and incubated with 3,3'diaminobenzidine for peroxidase activity (Graham and Karnovsky, 1966). A representative 40-/-L section was lightly stained with hematoxylin alone, or with eosin, and mounted on a glass slide for LM. The area surrounding the needle track was dissected from other 40-/-L sections, post-fixed in 1% osmium tetroxide, dehydrated through graded alcohols, and embedded in Epon. One-micron thick sections were cut and stained with Toluidine Blue. Appropriate areas were selected for ultrathin sections, stained with lead citrate and uranyl acetate, and examined with a JEOL 100S electron microscope (JEOL Co., Tokoyo, Japan). Although 4 hemispheres were examined per time period per ph, technical difficulties in horseradish peroxidase injection or tissue preparation resulted in smaller numbers for the evaluation of area of extravasated horseradish peroxidase on the 40-/-L sections. RESULTS Blood physiologic variables (means ± SEM) were similar for all animals and were also similar to other nonoperated, spontaneously-ventilating, halothane-anesthetized animals (Kraig et ai., 1987). Variables included arterial ph of 7.30 ± 0.01, Pao2 of 102 ± 3 mm Hg; Paco2 of 68 ± 1 (mm Hg), arterial glucose of 8.0 ± 0.3 mm, hematocrit of 42 ± 0%, systolic blood pressure of 102 ± 2 mm Hg, and body temperature of 37 ± O C. The brain changes were grouped according to their distance from the central needle track defect since previous studies (Kraig et ai., 1987) showed that injection at ph 4.5 produced a maximal decrease in interstitial ph to 5.16 ± 0.14. Following 24-h survival the frozen sections of brain showed an area of necrosis measuring from 0.3 to 0.4 mm in diameter. In contrast, injection at ph 7.3 resulted in no change in interstitial ph and only a thin rim of necrotic neurons around the needle track defect that together measured <0.1 mm in diameter. Consequently, three regions were examined: (a) needle track defect, (b) central zone surrounding the needle tract (approximately measuring 0.1 to 0.2 mm in diameter), and (c) a peripheral zone measuring from 0.2 to 0.4-0.5 mm around the needle tract. Regions 1 and 2 represented the area damaged by needle tract plus injected solution, whereas Region 3 (peripheral zone) represented the additional area destined to become necrotic following ph 4.5 injection. Vascular permeability following injection of ph 7.3 and ph 4.5 solutions Forty micron thick sections following both injections of ph 7.3 and ph 4.5 solutions showed similar horseradish peroxidase extravasation around the needle track. After injection of ph 7.3 solution, the area of horseradish peroxidase measured 0.5-0.6 mm (n = 2), 0.4-0.6 mm (n = 3), and 0.36-1.66 mm (n = 4) in diameter at 1, 3, and 6 h respectively (Fig. la). It was confined to the needle track defect at 24 h and measured 0.14-0.2 mm (n = 2) in diameter (Fig. IB). After injection of ph 4.5 solution, the area of extravasated horseradish peroxidase measured 0.2-0.66 mm, 0.44-0.7 mm (n = 3), and 0.54-0.7 mm (n = 3) at 1, 3, and 6 h respectively (Fig. IC). One of the 1 h animals was exposed to circulating horseradish peroxidase for 1 rather than 5 min prior to death, thereby making the diameter of horseradish peroxidase extravasation smaller than expected. Figure 10, obtained from paraffinembedded brain from a previous study (Kraig et ai., 1987), illustrates the well-circumscribed area of J Cereb Blood Flow Metab. Vol. 7. No.5, 1987

BRAIN NECROSIS 627 FIG. 1. A: Forty micron section, 6 h after ph 7.3 injection, shows area of extravasated horseradish peroxidase measuring 0.7 cm in diameter. Neurons are surrounded by but do not stain with horseradish peroxidase. B: Forty micron section, 24 h after ph 7.3 injection, shows needle track containing horseradish peroxidase. Adjacent parenchyma is slightly vacuolated. C: Forty micron section, 3 h after ph 4.5 injection, shows area of extravasated horseradish peroxidase that surrounds the central needle track and that measures 0.66 mm in diameter. 0: Paraffin section, 24 h after ph injection (see Kraig et ai., 1987), shows a well-circumscribed area of necrosis that surrounds central needle track (arrow). Hematoxylin-eosin; x 160. brain necrosis seen 24 h after injection of ph 4.5 solution. Horseradish peroxidase was present in the extracellular spaces, vascular basement membranes, and endothelial pinocytotic vesicles. Needle track and central zone after injection of ph 7.3 solution (Regions 1 & 2) Light and electron microscopic examination of the needle track (Region 1) showed marked vacuolation of the neuropil with dilated extracellular spaces and swollen, often fragmented cells and processes. Myelin sheaths were relatively intact. Neurons were shrunken and hyperchromatic, had pyknotic nuclei, and were flooded with horseradish peroxidase at 3 and 6 h. Disaggregation of polyribosomes and dilatation of Golgi vesicles preceded or accompanied progressive electron density of the cytoplasm, cell shrinkage, and nuclear pyknosis and karyorrhexis. Examination of Region 2 showed dilated extracellular spaces containing horseradish peroxidase. At 1, 3, and 6 hours after injection, neurons were pale and swollen and were encircled by but not stained with extracellular horseradish peroxidase (Fig. la). Residual organelles, including polyribosomes, rough endoplasmic reticulum, mitochondria, and Golgi apparatus clustered around the nucleus (Fig. 2A). Mitochondria were not noticeably J Cereb Blood Flow Me/ab, Vol. 7, No.5, 1987

628 C. K. PETITO ET AL. enlarged, although there was a mild decrease in matrix density and increase in intracristal spaces. Some neurons had disaggregated polyribosomes. Astrocytes showed progressive cytoplasmic swelling between I and 6 h. Nuclei were enlarged and contained finely dispersed chromatin with loss of the usual thin rim of condensed chromatin at the nuclear membrane. Necrotic glial cells were not seen. At 3 and 6 h, capillaries often were filled with red blood cells or were markedly compressed and surrounded by swollen perivascular astrocyte processes (Fig. 2B). Twenty-four hours after injection (Fig. IB), the needle track contained RBCs, monocytes, and horseradish peroxidase. Immediately adjacent brain was vacuolated and contained shrunken necrotic neurons, and fragmented cell processes. Astrocytes showed reactive changes with mildly enlarged nuclei and increased numbers of intermediate filaments. Remaining neurons appeared normal and no longer showed peripheral chromatolysis. Polymorphonuclear leukocytes and monocytes were present in the central necrotic zone and adjacent brain at 3 and 6 h, and numerous phagocytes containing myelin debris were present at 24 h. Needle track and central zone after ph 4.5 injection (Regions 1 & 2) Region 1 (needle tract and adjacent brain) showed changes similar to those occurring after injections of ph 7. 3 solution. The needle tract was surrounded by markedly vacuolated brain; necrotic neurons diffusely stained with horseradish peroxidase; markedly dilated extracellular spaces; and swollen, fragmented cell processes. However, Region 2, (the remaining central zone), showed changes unique to acid injection. Most neurons were shrunken, hyperchromatic, and contained pyknotic or fragmented nuclei. They were FIG. 2. A: A cortical neuron, 1 h after ph 7.3 injection, shows perinuclear clustering of organelles and clearing of peripheral cytoplasm. Adjacent extracellular space on the left is filled with horseradish peroxidase and cell processes in this region are swollen. x 4,800. B: Capillary lumen 3 h after ph 7.3 injection, is markedly narrowed to a slit-like lumen and is surrounded by markedly swollen astrocyte processes. x 7,000. J Cereb Blood Flow Metab, Vol. 7. No. 5, 1987

BRAIN NECROSIS 629 surrounded by swollen astrocyte processes and were frequently flooded with horseradish peroxidase (Fig. 1C). Number and degree of shrinkage increased with post-injection survival. The earliest ultrastructural changes in these neurons was a mild increase in electron density due to shrinkage of the cell body and nucleus, disaggregation of polyribosomes and contracted mitochondria with increased matrix density, and dilated intracristal spaces (Fig. 3A). This was followed by a loss of Golgi apparatus cisterns, progressive dilatation of Golgi apparatus vesicles and increased shrinkage, chromatin clumping, and nuclear fragmentation. A second type of neuronal necrosis was characterized by focal cytoplasmic lucency with loss of organelles and fragmentation of adjacent plasma membrane (Fig. 3B). These changes tended to be seen first at the dentritic hillock and to be associated with loss of dentritic neurotubules. Progressive involvement of the neuronal cell body led to neuronal necrosis characterized by fragmented cytoplasmic and nuclear membranes and clumped chromatin. This change was present in the central zone at 1 and 3 h, and in the more peripheral regions at 6 h after injection. Many astrocytes had enlarged nuclei and dilated cytoplasm containing contracted mitochondriachanges similar to those occurring following ph 7.3 injection. In addition, Region 2 contained necrotic astrocytes at 3 and 6 h that were identified by coarse chromatin clumps (Jenkins et ai., 1979; Kalimo et al., 1977) and focal cytoplasmic staining by horseradish peroxidase. These changes appeared to follow cytoplasmic swelling and nuclear enlargement. Capillaries were usually patent at 1 h after injection; but at 3 and 6 h, they were often filled with red blood cells or had markedly compressed lumens and were surrounded by massively dilated astrocyte processes (Fig. 4A). The endothelium contained variable numbers of horseradish peroxidasefilled vesicles. In contrast to the injection of ph 7.3 FIG. 3. A: Neuron from the central zone (Region 2), 1 h after ph 4.5 injection, has a relatively electron dense cytoplasm that contains ribosomes and dilated vesicles of the Golgi apparatus (G). Polyribosomes are absent. A peri neuronal astrocyte process (A) is swollen. x 30,000. B: Neuron from Region 2, 6 h after ph 4.5 injection, has focal loss of organelles and fragmented plasma membrane (arrow). Remaining cytoplasm contains usual complement of polyribosomes and endoplasmic reticulum. x 5,600. J Cereb Blood Flow Metab, Vol. 7, No.5, 1987

630 C. K. PETITO ET AL. FIG. 4. A: Lumens of two capillaries, 3 h after ph 4.5 injection, are markedly narrowed and contain circulating horseradish peroxidase (arrows). Perivascular astrocyte processes are markedly swollen. x 4,000. B: Peripheral region, 3 h after ph 4.5 injection, containing two astrocytes (upper right) that have enlarged nuclei containing small chromatin clumps and mildly dilated cytoplasm. Capillary is slightly compressed and is surrounded by markedly swollen astrocyte foot processes. Two neurons on the bottom are normal. x 3,000. solution, focal endothelial defects were seen at 1,3, and 6 h and included endothelial vacuoles and microvillus formation, focal areas of endothelium flooded with horseradish peroxidase, and endothelial defects with exposure of the underlying basement membrane. Occasionally, platelet thrombi were present in the peripheral zone at 6 h after injection. Peripheral zone after injections Astrocyte swelling and nuclear enlargement, red blood cell stasis, and capillary compression with indentation by swollen astrocyte processes were found after both injection with ph 4.5 and ph 7. 3 solutions. However, the severity of astrocyte swelling and capillary compression was greatest in the animals injected with ph 4.5 solution and, in this group, was more pronounced at 3 and 6 h than at 1 h. Swollen astrocyte nuclei containing coarse chromatin clumps indicative of cell necrosis, were seen at 3 and 6 h after ph 4.5 injection (Fig. 4B) but were rare following ph 7. 3 injection. White blood cells and/or platelets were adherent to or occluded occasional venules and capillaries at 3 and 6 h after ph 4.5 injection. However, white blood cells and/or platelets were not observed following ph 7. 3 injection. Neurons in both groups appeared normal on both light and electron microscopy at 1, 3, and 6 h. Occasionally, mildly enlarged proximal dendrites were observed. DISCUSSION This study shows that excessive acidosis produces early morphological changes in brain that can be distinguished from traumatic injury caused by an intracortical injection of fluid at physiological ph and molality. Specific alterations in neurons, astrocytes, and blood vessels precede the development of cortical necrosis from a lethal injection at ph 4.5. Fluid injection produces morphologic alterations that are especially prominent during the early post- J Cereb Blood Flow Metab, Vol. 7, No.5, 1987

BRAIN NECROSIS 631 injection period and must be distinguished from those due to acid injury itself. Needle injury alone produces a linear zone of cortical necrosis, <0.1 mm in diameter, focal hemorrhage within the needle track defect and overlying subarachnoid space, and a thin rim of necrotic neurons adjacent to the needle tract (Pulsinelli and Petito, 1982). Microvessels in the subarachnoid space may be constricted and contain platelet aggregates (Rosenblum and EI-Sabban, 1978). Although the additional insult of 10 I.Ll of fluid at physiological ph and molality is indistinguishable from a needle track defect at 24 h (Kraig et ai., 1987), this study demonstrates that neurons, glia, and blood vessels undergo severe alterations during the early post-injection period. Abnormal blood-brain barrier (BBB) permeability, capillary compression, marked dilation of the extracellular spaces, and swelling of both astrocytes and neurons occur during the initial 6 h. However, by 24 h, the BBB abnormalities were confined to the needle tract defect; cytoplasmic swelling of neurons and astrocytes was absent, and a narrow band of necrotic neuropil around the needle defect was undergoing organization by phagocytes. Neurons were the first cell to be irreversibly injured by the acid injection. The ultrastructural changes that preceded the neuronal necrosis seen by 1 h after injection were similar to changes that precede neuronal death following transient ischemia (Brown and Brierley, 1972; Petito and Pulsinelli, 1984a) and included disaggregation of polyribosomes, loss of Golgi apparatus cisterns, and dilatation of Golgi apparatus vesicles. Although the mitochondrial swelling typical of transient ischemia was not observed (Brown and Brierley, 1972), its occurrence may have been missed since the earliest post-injection interval was 1 h, whereas mitochondrial swelling following transient ischemia can be severe but reversible within 15 min (Petito and Pulsinelli, 1984a,b). A second type of neuronal necrosis was noted between 3 and 6 h after injection and resembled that seen with permanent ischemia (Kalimo et ai., 1977; Jenkins et ai., 1979). This neuronal necrosis was associated with mild cytoplasmic swelling rather than with progressive shrinkage. Focal and then diffuse cytoplasmic lucency, loss of organelles, membrane fragmentation, and chromatin clumping preceded neuronal cell death. The mechanisms involved in the production of neuronal necrosis in this model are unclear. Severe acidosis of the interstitial fluid may initiate or contribute to neuronal death during the first hour after injection. Vascular occlusion by extrinsic capillary compression or by intravascular thrombi during the late post-injection period (see below) may play an important role, and the ultrastructural evolution of cell death parallels that following transient or incomplete ischemia. Necrosis of astrocytes was common after the ph 4.5 injections although cytoplasmic swelling and nuclear enlargement also accompanied injection of ph 7.3 solution. Necrosis was first seen in the central zone at 1 and 3 h and was present in the more peripheral zone at 6 h. Astrocytic necrosis appeared to accompany rather than precede neuronal necrosis in the central zone. Although it was difficult to identify dead astrocytes in the peripheral zone, minimal nuclear swelling and minor chromatin clumping as shown in Figure 4B may have indicated early signs of cell death in regions where neurons and blood vessels appeared normal. Astrocytes play an important role in maintaining physiological ph of brain interstitial fluid (Kimmelberg et ai., 1982; Kraig et ai., 1986). Recent studies suggest that astrocytes selectively accumulate lactic acid following complete and hyperglycemic ischemia (Kraig et ai., 1986; Kraig and Chesler, 1987; Kraig and Nicholson, 1987). Similar astrocytic acidification may be related to an early and a lethal injury to these cells following acid injection. Furthermore, early irreversible injury to astrocytes may play an important role in the slow recovery of interstitial ph that correlates with production of tissue necrosis (Kraig et ai., 1987). Delayed circulatory impairment may be another important contributory mechanism whereby excessive acidosis produces brain necrosis in this model for several reasons. First, many of the morphological changes associated with acid injection resemble those accompanying irreversible neuronal injury following transient or permanent cerebral ischemia. Second, intravascular thrombi and endothelial cell defects were only seen in pre-necrotic brain regions following ph 4.5 injection, although red blood cell stasis was observed in animals injected with ph 4.5 and 7.3 solutions. A preliminary report of lactic acid-induced myelopathy describes vascular thrombi that may be similar to those described here (Balentine et ai., 1984). Although exposure to basement membrane is considered a prerequisite for platelet aggregation following endothelial injury, platelet aggregation has been demonstrated in blood vessels in which endothelia are damaged but basement membrane is not exposed (Hovig et ai., 1974; Povlishock et ai., 1983; Hermann and Voight, 1984; Rosenblum, 1986). Endothelial cells are sensitive to ph change and, in acidified cultures, develop focal defects between ph 7. 4 and ph 6 and J Cereb Blood Flo\\' Metab, Vol. 7, No.5, 1987

632 C. K. PETITO ET AL. severe necrosis at lower ph levels (Fox and DiCorleto, 1984). Last, preliminary studies show that injection of lactate, ph 4. 5 does not alter CBF at 15 min after injection but produces a fall in regional CBF at 3 h (personal observations). Acknowledgment: This study was supported by the National Institutes of Neurological and Communicative Disorders and Stroke Grant NS-03346. Dr. Kraig was supported by National Institutes of Neurological and Communicative Disorders Grant NS 19108 and a Teacher Investigator Development Award (NS-00767) as well as a Redel Foundation and DuPont Foundation Grant. The expert technical help of Mrs. Helen Hanes is gratefully acknowledged. REFERENCES Balentine JD, Green WB, Bonner MJ (1984) Lactic acid induced experimental myelopathy. J Neuropathol Exp Neuro143:324 Brown AW, Brierley JB (1972) Anoxic-ischemic cell changes in rat brains: Light microscopic and fine structural observations. J Neurol Sci 16:59-84 Cotran RS, Karnovsky MJ (1967) Vascular leakage induced by horseradish peroxidase in the rat. Proc Soc Exp Bioi Med 126:557-561 Deimann W, Taugner R, Fahini HD (1976) Arterial hypotension induced by horseradish peroxidase in various rat strains. J Histochem Cytochem 24:1213-1217 Fox PL, DiCorleto PE (1984) Regulation of production of a platelet-derived growth factor-like protein by cultured bovine aortic endothelial cells. J Cell Physiol 121 :298-308 Ginsberg MD, Welsh FA, Budd WW (1980) Deleterious effect of glucose pretreatment on recovery from diffuse cerebral ischemia in the cat. 1. Local cerebral blood flow and glucose utilization. Stroke 11:347-54 Goldman SA, Pulsinelli W, Plum F (1987) Differential survival of neurons and glia following acid exposure in primary culture. Neurology 37:250 Graham RC, Karnovsky MJ (1966) The early stages of absorption of injected horseradish peroxidase in the proximal tubules of mouse kidney: Ultrastructural cytochemistry by a new technique. J Histochem Cytochem 14:291-302 Hermann KS, Voight WH (1984) Ultrastructural observation of an electron dense amorphous layer on selectively damaged endothelial cells, a possible trigger of thrombogenesis in vivo and its inhibition by nafazatron. Thrombo Res 36:205-15 Hovig T, McKenzie FN, Arfors KE (1974) Measurement of the platelet response to laser-induced microvascular injury. Thromb Diath Haemorhh 32:695-703 Jenkins LW, Povlishock JT, Becker DP, Miller JD, Sullivan HG (1979) Complete cerebral ischemia: An ultrastructural study. Acta Neuropathol 48: 113-125 Kalimo H, Garcia JH, Kamijyo Y, Tanaka J, Trump BF (1977) The ultrastructure of "brain death" II. Electron microscopy of feline cortex after complete ischemia. Virchows Arch B Cell Path 25:207-220 Kalimo H, Rehncrona S, Soderfeldt B, Olsson Y, Siesj6 BK (1981) Brain lactic acidosis and ischemic cell damage: 2. Histopathology. J Cereb Blood Flow an d Metab 1:313-327 Kimelberg HK, Bourke RS, Steig PE, Barron KD, Hirata H, Pelton EW, Nelson LR (1982) Swelling of astroglia after injury to the central nervous system: mechanisms and consequences. In: Head Injury: Basic an d Clin ical Aspects (Grossman RG, Gildenberg PL, eds), New York, Raven Press, pp 31-44 Kraig RP, Pulsinelli WA, Plum F (1985) Heterogeneous distribution of hydrogen and bicarbonate ions during complete brain ischemia. In: Progress in Brain Research, Vol. 63, (Kogure K, Hossmann K-A, Siesj6 BK, Welsh FA, eds), Amsterdam, Elsevier, pp 155-166 Kraig RP, Pulsinelli WA, Plum F (1986) Carbonic acid buffer changes during complete brain ischemia. Am J Physiol 250:R 348-357 Kraig RP, Chesler M (1987) Glial acid-base behavior in brain ischemia. J Cereb Blood Flow an d Metab 7:S126 Kraig RP, Nicholson C (1987) Profound acidosis in presumed glia during ischemia. In: Cerebrovascular D iseas es, 15th Princeton-Williamsburg Conference (Raichle ME and Powers WJ, eds), New York, Raven Press, pp 97-102 Kraig RP, Petito CK, Plum F, Pulsinelli WA (1987) Hydrogen ions kill brain at concentrations reached in ischemia. J Cereb Blood Flow and Metab 7:379-386 Kraig RP and Wagner RJ (1987) Acid-induced changes of brain protein buffering. Brain Res 410:390-394 Myers RE (1979) A unitary theory of causation of anoxic and hypoxic brain pathology. In: Cerebral Hypoxia an d Its Con sequen ces. (Fahn S, Davis J, Rowland L, eds), Adv Neurol 26: 195-213 Myers RE, Yamaguchi M (1977) Nervous system effects of cardiac arrest in monkeys. Arch Neurol 34:65-74 Norenberg MD, Mozes LW, Gregorios JB, Norenberg LOB (1987) Effect of lactic acid on primary astrocyte cultures. J Neuropathol Exp NeuroI46:154-166 Petito CK, Pulsinelli WA (1984a) Sequential development of reversible and irreversible neuronal damage following cerebral ischemia. J Neuropathol Exp NeuroI43:141-53 Petito CK, Pulsinelli WA (1984b) Delayed neuronal recovery and neuronal death in rat hippocampus following severe cerebral ischemia: Possible relationships to abnormalities in neuronal processes. J Cereb Blood Flow and Metab 4:194-205 Plum F (1983) What causes infarction or ischemic brain? Neurology 33:222-233 Povlishock JT, Rosenblum WI, Shelley MM, Wei EP (1983) An ultrastructural analysis of endothelial change parelleling platelet aggregation in a light/dye model of microvascular insult. Am J Pathol 110:148-160 Pulsinelli WA, Waldman S, Rawlinson D, Plum F (1982) Moderate hyperglycemia augments ischemic brain damage: A neuropathologic study. Neurology 32: 1239-1246 Pulsinelli WA, Levy DE, Sigsbee B, Scherer P, Plum F (1983) Increased damage after ischemic stroke in patients with hyperglycemia with or without diabetes mellitus. Am J Med 74:540-544 Pulsinelli WA, Petito CK (1983) The neurotoxicity of hydrogen ions. Stroke 14: 130 Pulsinelli WA, Kraig RP, Plum F (1985) Hyperglycemia, cerebral acidosis, and ischemic brain damage. In: Cerebrovascular Diseases (Plum F and Pulsinelli WA, eds), New York, Raven Press, pp 201-210 Rehncrona S, Rosen I, Siesj6 BK (1981) Brain lactic acidosis and ischemic cell damage: 1. Biochemistry and neurophysiology. J Cereb Blood Flow an d Metab 1:297-311 Rosenblum WJ (1986) Aspects of endothelial malfunction and function on cerebral microvessels. Lab Invest 55:252-268 Rosenblum W J, El Sabban F (1978) Platelet aggregation and vasoconstriction in undamaged microvessels on cerebral surface adjacent to brain traumatized by a penetrating needle. Microvasc Res 15:299-307 Siesj6 BK (1985) Acid-base homeostasis in the brain: Physiology, chemistry, and neurochemical pathology. In: Progress in Brain Research Vol 63 (Kogure K, Hossmann K-A, Siesj6 BK, Welsh FA, eds), Amsterdam, Elsevier, pp 121-154 Smith ML, von Hanwehr R, Siesj6 BK (1986) Changes in extraand intracellular ph in the brain during and following ischemia in hyperglycemia and in moderately hypoglycemic rats. J Cereb Blood Flow an d Metab 6:574-583 Welsh FA, Ginsberg MD, Rieder W, Budd WW (1980) Deleterious effect of glucose pretreatment on recovery from diffuse cerebral ischemia in the cat. Stroke 11:355-363 J Cereb Blood Flow Metab, Vol. 7, No. 5, 1987