Neuron-specific mitochondrial degeneration induced by hyperammonemia and octanoic acidemia

Brain Research, 340 (1985) 2 l 1- 218

211

Elsevier BRE 10891

Neuron-Specific Mitochondrial Degeneration Induced by Hyperammonemia and Octanoic Acidemia LESTER R. DREWES I and RICHARD L. LEINO -~ Departments o f JBiochemistry and 2Biomedical Anatomy, School of Medicine, UniversiO, ~1' Minnesota - - Duluth, Duluth. M N 55812 (U. S. A. )

(Accepted October 30th, 1984) Key words: hyperammonemia - - mitochondria - - brain - - neuropathology - - Reye's syndrome - -

hepatic encephalopathy - - octanoic acidemia

The neuropathological consequences of acute exposure to the neurotoxicants ammonia and octanoic acid were investigated with the isolated, perfused canine brain preparation. After 1 h of combined hyperammonemia and octanoic acidemia, ultrastructural changcs were apparent in all brain regions examined. The cell bodies of neurons were the primary sites of these alterations. Neuronal mitochondria were distended, and the lamellae of the mitochondrial cristae were separated. In some cases the lamellae had completely dispersed, leaving only matrix remnants. Mitochondria of adjacent astrocytes appeared normal. Thus, a characteristic population of brain mitochondria is selectively vulnerable to a combination of hyperamrnonemia and octanoic acidemia and may he related to the biochemical mechanisms underlying encephalopathies of hepatic origin. INTRODUCTION

astrocytes. At the ultrastructural level, astrocytes

The brain is capable of maintaining chemical homeostasis and normal function despite being exposed to blood of relatively variable composition 4. The central nervous system is, however, sensitive to

initially appear to become enlarged. After 4 - 8 weeks, astrocytic hyperplasia is evident, and these cells exhibit increased n u m b e r s of cytoplasmic organelles. Most classical light microscopic Alzheimer type II changes, as reported earlier, are probably artifacts of immersion fixation 5,s. Few changes report-

conditions arising from i m p a i r m e n t of liver function. Neurologic disorders linked to hepatic dysfunction may range from confusion to coma and cerebral death. The causes of the cerebral abnormalities are not known, but a m m o n i a 17,26, short-chain fatty acids 21,33 and other potential toxicants 34 have been implicated because they become elevated when liver ureogenesis and fatty acid metabolism are impaired. Acute exposure to a m m o n i a or short-chain fatty acids is also known to alter cerebral metabolism and to produce coma in experimental animals 27,31. H y p e r a m m o n e m i a in humans occurs primarily as a result of liver dysfunction and can be produced in animals by portocaval anastomosis or with urease treatment. The neuropathological consequences of these diseases or conditions are detected primarily in the

edly occur in neurons, although their numbers may decrease in severe and protracted cases I. Endothelial cells appear unaffected. In Reye's syndrome, another liver disorder with acute encephalopathy, there is astrocyte swelling and partial loss of glycogen, but, in addition, characteristic injury to neuronal mitochondria. These organelles lose their matrix integrity and expand 24. Dilation and loss of rough endoplasmic reticulum are also reported. These morphological changes are substantially reversed in recovered patients 25. A m m o n i a , shortchain fatty acids and other blood-borne factors have been implicated as neurotoxicants in this disorder (see ref. 7 for review). Although ammonia, short-chain fatty acids and

Correspondence: L. R. Drewes, Department of Biochemistry, School of Medicine, University of Minnesota - - Duluth, Duluth, MN 55812, U.S.A.

212

Fig. 1. Light micrograph of motor cortex perfused under control conditions for 12(! min showing normal-appearing morphology Epon-Araldite section, methylene blue-azure II stain, x630.

other blood-borne toxic substances have been suggested to be responsible for the encephalopathies associated with liver disorders, most studies have examined the biochemical and pathological alterations which develop when the brain is exposed to only one of the putative neurotoxicants, In this study, the neuropathoiogic response to a combination of elevated blood ammonia and octanoic acid was investigated using the isolated canine brain preparation in which the composition of the perfusate was controlled and the influence of other organs was eliminated. After 1 h of hyperammonemia and octanoic acidemia, there were clear, substantial changes in brain cell ultrastructure. The principal alterations occurred in neuronal organelles and cytoplasm and were similar to those associated with Reye's syndrome. These results illustrate that a characteristic population of brain mitochondria is selectively vulnerable to a combination of hyperammonemia and octanoic acidemia and may be relevant to the underlying biochemical mechanisms involved in Reye's syn-

drome or similar encephalopathies. MATERIALS AND METHODS

Brain perfusion, The canine brain was isolated by the method of Gilboe et al. 13 and perfused as described previously9. The perfusate consisted of heparinized blood from a donor animal and was reduced in hematocrit to 33% with a high-molecular-weight Dextran solution in normal saline (Rheomacrodex). The pH was maintained at 7.4. the pCO, at 40 mm Hg and the pO2 at greater than 110 mm Hg. Blood glucose was maintained at 5-6 mM with continuous infusion of a concentrated glucose solution into the reservoir. It has been shown previously9,13 that under these conditions substrates of brain metabolism, particularly oxygen and glucose, are delivered in excess of the demand needed for normal brain activity. Blood temperature was held constant at 38 °C Cerebral blood flow rate was determined by measuring the volume of venous blood collected during a t0-min interval, and the flow was then adjusted to approxl-

213

Fig. 2. Electron micrograph of neuron from cerebellum perfused under control conditions showing relatively normal ultrastructure. × 17,400.

mately 65 ml/100 g'min ~ for the duration of the perfusion. Following isolation, the brain was perfused for a 60-min control period to establish steady-state conditions. Perfusion with blood containing 1.5 mM NH4CI and 2.0 mM sodium octanoate was conducted

by mixing the appropriate solution with the volume of blood (1-1.6 liter) in the reservoir. Sodium octanoate (0.5 M) was prepared by mixing equimolar amounts of octanoic acid and 0.5 M N a O H and adjusting the pH to 7.4. Perfusion for an additional 60-min experimental period was then continued. Per-

214

Fig. 3. Portion of a neuron from cerebellum perfused with ammonia and octanoate for 60 min showing damaged mitochondria (M) and Golgi bodies (G) with expanded cisternae. × 15,66(1. Fig. 4. Astrocyte processes (a) in cerebellum perfused with ammonia and octanoate showing normal mitochondria. Adjacent small neurons (n) have swollen mitochondria and expanded Golgi cisternae. × 17~775r

215

Fig. 5. Light micrograph of hippocampal white matter after ammonia and octanoate perfusion showing edematous changes resulting in increased spacing between axons, x567. Fig. 6. Light micrograph of hippocampus perfused under control conditions. White matter appears normal, x 567. Fig. 7. Electron micrograph of hippocampal white matter after ammonia and octanoate perfusion. Astrocytes (a) appear to be damaged or swollen, x 8820.

216 fusion during each of 3 control experiments lasted for a total of 120 min. The concentrations of neurotoxicants used in this study were chosen because they are in the high range of those concentrations observed in some disease states 20 and because they are similar to other experimental models27, 34. Fixation and microscopy. At the end of the 60-min experimental period, the brain was fixed without interruption of flow by perfusion for 10 min at 25 °C with 4% paraformaldehyde-2.5% glutaraldehyde in 0.1 M phosphate buffer, pH 7.2-7.4. Administration of fixative was preceded by a 10-s perfusion with phosphate-buffered saline to clear the brain vasculature of most blood components. Upon removal from the skull, the brain tissue was examined for uniform color and firmness, placed in fresh perfusion fixative and stored overnight at 4 °C. Blocks 1 mm × l mm x 0.5 mm were taken from areas of motor cortex, sensory cortex, hippocampus, thalamus, hypothalamus and cerebellum. The blocks were rinsed in two changes of 0.15 M phosphate buffer and post-fixed for 1.5 h in 1% osmium tetroxide in 0.1 M phosphate buffer containing 1.5% potassium ferricyanide 19. After dehydration in graded acetones, the tissues were embedded in Epon-Araldite and sectioned using an ultramicrotome (LKB-Huxley). Thick sections (1.5 /~m) for light microscopy were stained with toluidine blue or methylene blue-azure 1116. Silver or grey sections for electron microscopy were stained with uranyl acetate and lead citrate. RESULTS Three isolated brain preparations were perfused for 120 min under control conditions. All specimens obtained from these brains exhibited no observable abnormalities when examined by light and electron microscopy (Figs. 1, 2 and 6). Cerebral oxygen and glucose consumption and brain electrical activity (electroencephalographic ( E E G ) activity) in these preparations were constant and in the normal range for the duration of the perfusion period t0. In contrast, the cerebral cortex, thalamus, hippocampus and cerebellum of a brain perfused for 60 rain with a combination of elevated ammonia and octanoic acid exhibited marked ultrastructural abnormalities (Figs. 3, 4 and 7). The experimental group consisted of 6 isolated brain preparations, two of which were

prepared for microscopic studies. All preparations exhibited similar brain electrical (EEG) and metabolic abnormalities in response to the neurotoxicants 10. An isolated brain that was fixed after an experimental period of 80 min exhibited similar, but more extensive, ultrastructural abnormalities to those described here. The physical appearances of neuronal mitochondria were classified into 3 categories. The first group, present in a few neurons, consisted of substantially normal-looking mitochondria. The second group, comprising the majority of mitochondria, had normal cristae but rarefied matrices. The third group, also substantial, consisted of mitochondria with rarefied matrices and few or no cristae. Mitochondria of this group were often enlarged or distended into irregular shapes. Mitochondrial volume and density were normal in astrocytes (Figs. 4 and 7). The mitochondria of axons also remained normal in appearance, although some dendritic mitochondria were affected. Cisternae of the Golgi apparatus were frequently dilated (Figs. 3 and 4), and some dilation of the rough endoplasmic reticulum and disaggregation of polyribosomes were observed in neurons of the experimental group when compared to the controls. Although no chromatin clumping was evident, the profile of the neuronal nuclear membrane was often irregular and uneven. Oligodendrocytes appeared normal. Examination of astrocytes indicated a decrease in cytoplasmic density of some cells, suggesting initial stages of edema formation (Fig. 7). in the hippocampus astrocytic swelling was evident (Fig. 7), and enlarged perivascular spaces were common (Fig. 5L Myelinated axons in the experimental group showed evidence of separated laminae, although these alterations were not profound. Endothelial cells appeared normal in both the control and experimental groups DISCUSSION Studies of the biochemical and pathological events associated with encephalopathy resulting from liver dysfunction have focused primarily on ammonia because of its elevated plasma concentration when ureagenesis is inadequate and because some of its direct effects on cerebral metabolism are known. Several theories have been proposed to account for the pathophysiological consequences of hyperammone-

217 mia 2,3. These theories are based largely on altered brain ammonia concentrations, brain energy metabolism and neurotransmitters related to glutamate/ glutamine and tyrosine/tryptophan. However, shortchain fatty acids 21,33 and other blood-borne substances35 also are elevated in hepatic encephalopathy and in Reye's syndrome20, 28, and the former rapidly produce reversible coma in experimental animals27,31. Perfusion of the isolated canine brain with blood containing a combination of elevated ammonia and octanoic acid resulted in pathological alterations that were readily apparent at 60 min. In all of the brain regions examined the cell bodies of the neurons were primary sites for these alterations. Mitochondria of these cells were distended, and the lamellae of the mitochondrial cristae were separated. In some cases the lamellae had completely dispersed, leaving only matrix remnants. Mitochondria of adjacent astrocytes contained normally dense matrices and normal volumes. However, some signs of edema formation and disturbances in osmotic gradients were observed. Despite the apparent edema formation, astrocytic mitochondria remained normal in appearance (Fig. 7). In experimental encephalopathy by portocaval anastamosis, vacuolated mitochondria were evident only in astrocytes that had expanded cytoplasmic volumes 22,32. These observations, however, were recorded after experimental periods of several days instead of the 1-h period described in this report. Neuronal mitochondrial alterations similar to those observed here reportedly occur in ischemia 11. However, blood flow was held constant during our experiments, while venous oxygen tension increased and cerebrovascular resistance decreased 10. These observations preclude the possibility that insufficient blood flow and 0 2 delivery might explain the observed morphological disturbances. Selective injury to neuronal mitochondria similar to that reported here occurs in Reye's syndrome patients 24,25. The dilations of Golgi apparatus and rough endoplasmic reticulum and the nuclear changes may also indicate a similarity between Reye's syndrome and the experimental conditions in the present study. Whether other ultrastructural features of Reye's syndrome, such as marked swelling and deglycogenation of astrocytes and extensive

myelin bleb formation, would occur in longer term experiments can only be speculated. No disturbance in the ultrastructure of the vascular endothelium was detected in the current experiments. This concurs with the observation that functional alterations in the blood-brain barrier, assessed by blood-brain transport of D-glucose and amino acids, were also absent 10. A detailed discussion of the biochemical mechanism of the observed alterations is beyond the scope of this paper. However, any hypothesis must take into account: (1) the pathways of ammonia detoxification in brain including its compartmentation, and (2) the pathways of octanoate metabolism. The biochemical reactions by which ammonia is incorporated into brain glutamine are well known ts. The mechanisms by which short-chain fatty acids act or are integrated into brain metabolism are less clear. It is known that short-chain fatty acids such as octanoate rapidly penetrate the blood-brain barrier 23 and cause alterations in brain electrical acitivity 2v,3x and, reportedly, elevations in intracranial pressure 30. These responses may be due to the direct action of octanoate on an enzyme such as Na +,K +ATPase as previously reported~. 29. However, in view of the selective vulnerability of neuronal mitochondria observed in the present study, the roles of shortchain fatty acids in reactions of intermediary metabolism that may be localized to specific cell types or compartments must also be considered. Such selective metabolic interactions may also underlie the mechanisms of anticonvulsant activities of valproate ~2 and medium-chain triglyceride diets 15, as well as the neurotoxicity of 4-pentenoic acid in patients with Jamaican vomiting disease 14. Further investigation to characterize these biochemical interactions with toxicants is needed. ACKNOWLEDGEMENTS The authors wish to thank Audrey Comstock for preparation of the typescript and Carolyn Clark for assistance in editing and revising the manuscript. Marjorie Lindemann provided expert technical assistance. The Duluth Clinic Foundation is acknowledged for financial assistance.

218 REFERENCES 1 Adams, R. D. and Foley, J. M., The neurological disorder associated with liver disease. In H. H. Merritt and C. C. Hare (Eds.), Metabolic and Toxic Diseases of the Nervous System, Vol. 32, Proc. Assoc. Res. Nerv. and Ment. Dis., Williams and Wilkins, Baltimore, 1953, pp. t98-237. 2 Benjamin, A. M., Ammonia. In A. Lajtha (Ed.), Handbook c~fNeurochemistrv, Vol. 1. Plenum Press, New York, 1982, pp. 117-137. 3 Bernardini, P. and Fischer, J. E., Amino acid imbalance and hepatic encephalopathy, Ann. Rev. Nutr., 2 (1982) 419-454. 4 Bradbury, M., The Concept of a Blood Brain Barrier, Wiley, New York, 1979, p. 383. 5 Cavanagh, J. B. and Kyu, M. H., Type II Alzheimer change experimentally produced in astrocytes in the rat, J. Neurol. Sci., 12 (1971) 63-75. 6 Dahl, D. R., Short chain fatty acid inhibition of rat brain Na-K adenosine triphosphatase, J. Neurochem.. 15 (1968) 815-820. 7 DeLong, G. R. and Glick, T. H., Encephalopathy of Reye's syndrome: a review of pathogenetic hypotheses, Pediatrics, 69 (1982) 53-63. 8 Diemer, N. H., Glial and neuronal changes in experimental hepatic encephalopathy. A quantitative morphological investigation, Acta neurol, scand., 58, Suppl. 71 (1978) 7-144. 9 Drewes, L. R., An improved apparatus for blood perfusion of the canine cerebral vasculature, Neurochem. Res., 5 (1980) 553-562. 10 Drewes, L. R., Cerebral metabolism and blood-brain transport during elevated blood ammonia and octanoate, manuscript submitted. 11 Garcia, J. H., Kalimo, H., Kamijyo, Y. and Trump, B. F., Cellular events during partial cerebral ischemia. I. Electron microscopy of feline cerebral cortex after middle-cerebralartery occlusion, Virchows Arch. B, 25 (1977) 191-206. 12 Gerber, N., Dickinson, R., Harland, R., Lynn, R., Houghton, D., Antonias, J. and Schimschock, J., Reye-like syndrome associated with valproic acid therapy, J. Pedlar., 95 (1979) 142-144. 13 Gilboe, D. D., Betz, A. L. and Drewes, L. R., Use of the isolated canine brain in studies of cerebral metabolism, metabolite transport and cerebrovascular physiology. In N. Marks and A. Rodnight (Eds.), Research Methods" in Neurochemistrv, Vol. 3, Plenum Press, New York, 1975, pp. 3-42. 14 Glasgow, A. M. and Chase, H. P., Production of the features of Reye's syndrome in rats with 4-pentenoic acid, Pediat. Res., 9 (1975) 133-138. 15 Haidukewych, D., Forsythe, W. I. and Sills, M,, Monitoring octanoic and decanoic acids in plasma from children with intractable epilepsy treated with medium-chain trigtyceride diet, Clin. Chem., 28 (1982) 642-645. 16 Hollander, H., The section embedding (SE) technique. A new method for the combined light microscopic and electron microscopic examination of central nervous tissue, Brain Research, 20 (1970) 39-47. 17 Hoyumpa, A. M., Desmond, P. V., Avant, G. R., Roberts, R. K. and Schenker, S., Hepatic encephalopathy (Clinical

Conference), Gastroenterology, 76 (1979) 84- 195~ 18 Kvamme, E., Ammonia metabolism in 1he CNS, Prog. Neurobiol., 20 (1983) 109-132. 19 Langford, L.. A. and Coggeshall, R. E., The use of potassium ferricyanide in neural fixation, Anat. Rec.. 197 (19811) 297-303. 20 Mamunes, P., DeVries, G. H., Miller, C. D. and David, R. B., Fatty acid quantitation in Reye's syndromc. In J. D. Pollack (Ed.), Reye's Syndrome, Grune and Stratton, New York, 1975, pp. 245-254. 2l Muto, Y. and Takahashi, Y. J., Gas chromatographic determination of plasma short chain fatty acids in diseases of the liver, J. Jap. Soc. intern. Med., 53 (1964)828-839. 22 Norenberg, M. D. and Lapham, L. W., The astrocyte response in experimental portal-systematic encephalopathy: an electron microscope study, J. Neuropath. exp~ Neurol., 33 (1974) 422-435. 23 Oldendorf, W., Carrier-mediated blood-brain barrier transport of short-chain monocarboxylic organic acids, Amer. J. Physiol., 224 (1973) 1450-1453. 24 Partin, J. C., Partin, .1. S., Schubert, W. K. and McLaurin, R. J., Brain ultrastructure in Reye's syndrome. J. Neuropath. exp. Neurol., 34 (1975) 425-444. 25 Partin, J. S., McAdams. A. J., Partin. J., Schubert, W. K. and McLaurin, R. J.. Brain ultrastructure in Reye's disease. II. Acute injury and recovery processes in three children, J. Neuropath. exp. Neurol.. 37 C1978) 796-819. 26 Plum, F. and Hindfelt. B. In P. J. Vinken and G. W. Bruyn (Eds.), Handbook o f Clinical Neurology, Vol 27. NorthHolland, Amsterdam. 1976, pp. 349-377. 27 Rabinowitz, J. L.. Staeffen. J.. Aumonier. P.. Blanquet. P., Vincent, J. D.. Daviaud. R., Ballan. P.. Ferrer, J.. Terme, R., Series, C. and Myerson. R. M.. The effects of intravenous sodium octanoate on the rhesus monkey, Amer. J. Gastroenterol.. 69 11978) 187-190. 28 Trauner, D. A.. Nyhan. W L. and Sweetman. L.. Shortchain organic acidemia and Reye's syndrome. Neurol. (Minneap.), 25 (1975) 296-298 29 Trauner, D. A., Regional cerebral Na ~ .K ~-ATPase activity following octanoate administration. Pediatr. Res.. 14 (1980) 844-845. 30 Trauner, D. A. and Adams. H.. lntracranial pressure elevations during octanoate infusion in rabbits: an experimental model of Reye's syndrome. Pediatr. Res.. 15 ft981) 1097-1099. 31 White, R. P. and Samson. F. E.. Jr.. Effects of fatty acid anions on the electroencephalogram of unanesthetized rabbits, Amer. J. Phvsiol.. 186 (1956) 271-274. 32 Zamora, A. J., Cavanagh, J. and Kyu, M. H.. Ultrastructural responses of the astrocvtes to portocaval anastomosis in the rat, J. Neurol. Sci., 18 (1973) 25-45. 33 Zieve, L., Pathogenesis of hepatic coma, Arch. Intern. Med., 118 (1966) 211-223. 34 Zieve, L., Doizaki. W. M. and Zieve, F. J., Synergism between mercaptans and ammonia or fatty acids in the production of coma: a possible role for mercaptans in the pathogenesis of hepatic coma, J. Lab. clin. Med., 83 (1974) 16-28. 35 Zieve, L. and Nicoloff, D. M., Pathogenesis of hepatic coma, Ann. Rev. Meal., 26 (1975) 143-157.