Hyperammonemic Encephalopathy
2002; Wolters Kluwer; Volume: 81; Issue: 3 Linguagem: Inglês
10.1097/00005792-200205000-00007
ISSN1536-5964
Autores Tópico(s)Alcoholism and Thiamine Deficiency
ResumoIntroduction Hyperammonemia is observed most often in patients with severe liver disease, the ninth leading cause of death in the United States (15). The incidence of hyperammonemia in hepatic encephalopathy (55,64) or fulminant hepatic failure (9) is estimated to be as high as 90% (19,63). However, assigning a role to ammonia in the pathophysiology of encephalopathy associated with liver disease is complicated by the many other central nervous system mechanisms implicated in liver disease or portal systemic shunting, for example, impairment of the blood-brain barrier, gut-derived neurotoxins, neurotransmitter abnormalities, disturbed energy supply, and others (9). The clinical picture of hyperammonemia is best examined under circumstances in which mechanisms attributable to severe liver impairment are absent. Urea cycle disorders may be the most appropriate model in that liver function is normal (apart from mild increases in transaminase levels) and clinical signs and symptoms are directly related to hyperammonemia. As will be demonstrated, a comparison of the encephalopathy of hyperammonemia as it occurs in urea cycle disorders with the encephalopathy of fulminant hepatic failure and hepatic encephalopathy reveals that the 3 conditions have similar biochemical and pathologic findings with abnormalities in glutamine metabolism and astrocyte morphology predominating. Also reviewed are studies of experimental hyperammonemia in animals that describe the pathophysiologic mechanisms by which hyperammonemia produces encephalopathy and suggest a pharmacologic intervention that may prevent or ameliorate hyperammonemic encephalopathy. Clinical Hyperammonemic Encephalopathy The urea cycle disorders represent a clinical setting in which hyperammonemia, in the absence of impairment of liver function or portal systemic shunting, regularly causes encephalopathy associated with cerebral edema (4,5). To illustrate the clinical, laboratory, and pathologic picture of hyperammonemia in the absence of severe liver impairment, the natural history of a composite case of fatal acute hyperammonemic encephalopathy in a patient with ornithine transcarbamylase deficiency has been constructed. The features included are based on literature case reports and the author’s experience with 545 cases of urea cycle disorders (5). Composite case This 13-year-old girl presented at a community hospital with a chief complaint of progressive irritability, lethargy, and vomiting. Past history revealed dietary protein avoidance and the neonatal death of her brother. Despite treatment with antibiotics, antiemetics and intravenous fluids, she developed ataxia and slurred speech and became progressively more obtunded. She had hyperactive deep-tendon reflexes and a positive Babinski response; a computer axial tomography (CT) examination of her head was within normal limits (Figure 1A). Laboratory studies were unremarkable except for venous blood gases: pH, 7.54; pCO2, 25 tor (3.33 kPa); HCO3−, 21 mmol/L (5,41). Routine analysis of cerebrospinal fluid (CSF) was within normal limits.Fig. 1: Computer axial tomographic scans of the head of hyperammonemic encephalopathy in the composite case of ornithine transcarbamylase deficiency. A. Image done upon admission to the community hospital. B. Image done 24 hours later demonstrates bilateral hemispheric edema with effacement of cerebrospinal fluid spaces.Twenty-four hours later, following a tonic clonic seizure that was treated with lorazepam, she was transferred to a tertiary care medical center where she was found to be hypotensive and unresponsive except to painful stimuli. She required mechanical ventilation. Her pupils were dilated to 6 mm and reacted sluggishly to light. She exhibited episodes of decorticate and decerebrate activity. The oculocephalic and oculovestibular responses were positive; papilledema was not detected. Admission laboratory findings were as follows: venous pH, 7.55; pCO2, 22 tor (2.93 kPa); HCO3−, 17 mmol/L; serum alanine aspartate, 54 U/L; serum alanine aminotransferases, 58 U/L; and serum bilirubin and prothrombin time were within normal limits. Plasma ammonia was 849 μmol/L (normal <50). Other routine analyses were normal. Plasma amino acids were (μmol/L): glutamine, 1,440 (normal, 591 ± 66); citrulline, 5 (normal, 33 ± 7); alanine, 508 (normal, 369 ± 88); arginine, 63 (normal, 84 ± 20). Cerebrospinal fluid obtained from the earlier spinal tap had a glutamine level of 4,700 μmol/L (normal, 509 ± 44), and proton magnetic resonance spectroscopy of the head revealed increased glutamine levels (8,10) and reduced myo-inositol levels (8). A repeat CT scan of the head revealed severe cerebral edema (Figure 1B). Electroencephalography revealed slow waves and the absence of triphasic activity. Over the ensuing 24 hours she developed a severe metabolic acidosis; her pupils became dilated and fixed, an intracranial pressure monitor recorded pressures as high as 60 mmHg (8 kPa). After loss of brain stem function became evident, life support systems were discontinued. Postmortem gross examination of the brain revealed flattening of the gyri, effacement of the sulci, and herniation of the cerebellar tonsils. Microscopic examination revealed swollen astrocytes (Type II Alzheimer cells, swollen astrocytes whose cytoplasmic water has been incorporated into the nucleus as a consequence of an immersion fixation artifact [7,51]) and preserved neurons (75). A diagnosis of ornithine transcarbamylase deficiency was confirmed after it was discovered she had oroticaciduria and a mutation at the ornithine transcarbamylase locus (delL81 ex3). Comment: Although most patients with urea cycle deficiencies recover from multiple episodes of hyperammonemic encephalopathy, patients may suffer brain damage and cognitive loss (20,40,49). The composite case demonstrates the typical clinical, laboratory, and pathologic features of acute hyperammonemic encephalopathy: a respiratory alkalosis (perhaps the earliest sign of encephalopathy) evolving to a metabolic acidosis as hemodynamic instability ensues; a typical pattern of cerebral edema (often not evident by imaging early in the course) progressing to intracranial pressure; increased levels of glutamine in plasma, CSF, and brain; decreased brain levels of myo-inositol; astrocyte swelling and few neuronal changes; and little evidence of impaired liver function. That these clinical and laboratory findings are a function of hyperammonemia and independent of liver disease is supported by similar clinical and laboratory findings in other hyperammonemic settings where liver function is normal or nearly so, for example, ureolysis in stagnant urine (18,60), cytoreductive and immunosuppressive therapy (14,39,46,62), and valproate therapy (56). Fulminant Hepatic Failure The course of the encephalopathy of fulminant hepatic failure is characterized by hyperammonemia (9), respiratory alkalosis (58), accumulation of brain glutamine (as measured on postmortem tissue [57] and in vivo by proton magnetic resonance spectroscopy [2,23]), decreased brain levels of myoinositol (measured by proton magnetic resonance spectroscopy) (23), astrocyte swelling (31), cerebral edema (9), increased intracranial pressure, and death from brain stem compression. A contribution of liver factors other than hyperammonemia to explain the encephalopathy would appear unlikely or coincidental because the characteristics of fulminant hepatic failure are identical to those in hyperammonemic encephalopathy, as described in the composite case of ornithine transcarbamylase deficiency and in experimental hyperammonemia, where liver disease is absent. If fulminant hepatic failure is a result of hyperammonemia, it may be successfully treated by hemodialysis before brain stem involvement, as is recommended for urea cycle disorders when medical therapy is unsuccessful (50). Hepatic Encephalopathy Hepatic encephalopathy as it occurs in chronic liver disease shares all but 1 (cerebral edema) of the characteristics of hyperammonemic encephalop-athy and fulminant hepatic failure: hyperammonemia (55,64), respiratory alkalosis (69), increased brain glutamine levels as measured on postmortem tissue (37) and in vivo by proton magnetic resonance spectroscopy (24), and astrocyte swelling (43). Also found by proton magnetic resonance spectroscopy in hepatic encephalopathy is a decrease in brain myo-inositol levels (24). Although typical cerebral edema is uncommon in hepatic encephalopathy (12,16), the finding of abundant Type II Alzheimer cells indicates an increase in astrocyte volume. This enlarged astrocyte compartment may be compensated by a combination of factors; intracranial compliance may play a role as might mild to moderate degrees of brain atrophy (48,74). A third compensatory mechanism may be other brain volume regulatory factors (45,65), for example, down-regulation of organic osmolytes such as myo-inositol. A compensatory decrease in myoinositol levels will reduce intracellular osmolarity and thereby ameliorate cerebral edema and its consequences. These compensatory mechanisms (brain compliance, brain atrophy, osmolar adjustments) in a chronically ill mildly hyperammonemic patient may account for the differences between the encephalopathic picture of hepatic encephalopathy and that of the acutely ill, severely hyperammonemic, fulminant hepatic failure patients. Table 1 summarizes the similarities of the clinical, biochemical, and pathologic features of the encephalopathy of acute hyperammonemia, hepatic encephalopathy, and fulminant hepatic failure.TABLE 1: The main features of hyperammonemia (in the absence of liver disease), fulminant hepatic failure (FHF), and hepatic encephalopathyExperimental Hyperammonemia (with Normal Liver Function) Clinicopathologic correlation A particularly convincing study (70) of the role of hyperammonemia was done in awake primates with normal liver function. These animals, when given ammonium acetate intravenously over a 24-hour period, exhibited all the signs and symptoms described in the composite case: vomiting, lethargy progressing to somnolence, seizures, absence of spontaneous movement, and apnea. Laboratory and pathologic studies were also similar to the composite case: peak mean plasma ammonia concentration of 900 (mol/L, respiratory alkalosis, slow wave electroencephalogram, gross edema of the brain with herniation of the cerebellar tonsils, swollen astrocytes, but unremarkable neurons, axons, dendrites, and synapses. The Astrocyte Astrocytes, once thought to be structural support elements, constitute an estimated one-third of brain volume and have been discovered to have important biochemical, neurochemical, and regulatory roles (42,47,52,68). They express these roles by interactions with all brain elements: neurons, oligodendroglia, synapses, endothelia, pia mater, and ependyma among themselves (33) (Figure 2). Because of the astrocytes’ ubiquitous relationship to other brain structures, it is proposed that changes in astrocyte size and function are responsible for many of the neurologic manifestations of hyperammonemia.Fig. 2: The astrocyte, demonstrating its relationship with other structures of the brain: the capillaries, neurons, synapses, synaptic vesicles, node of Ranvier, ependyma, and pia mater. Reprinted with permission from reference 46, Scientific American, vol. 260, Kimelberg HK, Norenberg MD. Astrocytes, pp 66–72, 1989.Glutamine: An astrocyte organic osmolyte Hyperammonemia, by itself, has been shown repeatedly to cause an increase in cerebral cortical glutamine content and astrocyte swelling (17,36). Glutamine is a reaction product of glutamine synthetase activity (equation 1), a cytosolic enzyme confined in the brain to the astrocyte (44,54). L-Methionine S-sulfoximine (MSO) (59), the only 1 of 4 stereoisomers of MSO that is an irreversible inhibitor (35) of glutamine synthetase, lowers glutamine content in normal brain (21) and prevents the increase in cortical glutamine in hyperammonemic animals (11). MSO has been shown to protect mice against ammonia toxicity (26,71); MSO-treated hyperammonemic animals had a significantly higher ammonia LD50 than untreated hyperammonemic animals, and none of the MSO-treated mice died at ammonia doses that were lethal for 50% of the non-MSO-treated animals (71). The Pathophysiology of Hyperammonemia The interrelationships among the clinical picture of hyperammonemia, cerebral edema, brain glutamine and glutamine synthetase, swollen astrocytes, and the protective effects of MSO remained obscure until it was suggested (3,6,72) that the increase in intraastrocyte glutamine concentration, the synthesis of which is activated by high ammonia levels, functions as an organic osmolyte which increases intracellular osmolarity. In response, cell volume increases as water enters the cell, creating an intracranial mass effect and astrocyte dysfunction. Quantitative consideration of these events is compelling: glutamine concentration, normally 5 mmol/kg of cerebrum, increases to 18 mmol/kg under hyperammonemic conditions (11). This increase, when confined to the smaller astrocyte compartment, may represent as much as 30 mOsm per kg. If astrocyte swelling is a function of glutamine accumulation and is the cause of cerebral edema in hyperammonemic states, inhibition of glutamine synthetase by MSO, long known to prevent increases in cerebral glutamine concentration, should prevent astrocyte swelling and brain water accumulation. Two studies, 1 morphologic (73) and the other physiologic (67), were performed to evaluate these possibilities. MSO prevents astrocyte swelling In the morphologic study, there were 2 groups of hyperammonemic rats, 1 group of which was pretreated with MSO (73). There were also 2 nonhyperammonemic control groups, 1 of which received vehicle and the other only MSO. To show that the astrocyte is the brain cell responsible for cerebral edema, the fine morphologic features of the astrocyte were studied. In the group of rats that received ammonia alone, the astrocyte cytoplasm and perivascular endfeet were moderately to markedly expanded, whereas pretreatment with MSO ameliorated these changes. Figure 3 demonstrates this effect of MSO on astrocyte foot process cell volume; its watery expansion in a hyperammonemic animal is mitigated by pretreatment with MSO.Fig. 3: The protective effective of methionine sulfoximine (MSO) on astrocyte swelling as determined by electron microscopy after perfusion fixation. A. Swollen astrocyte perivascular endfeet (arrowheads) in a hyperammonemic rat. B. Astrocyte perivascular endfeet (arrowheads) in a hyperammonemic rat pretreated with MSO; swelling is not apparent. Reprinted from reference 73, Neuroscience, vol. 71, Willard-Mack CL, Koehler RC, Hirata T, Cork LC, Takahashi H, Traystman RJ, Brusilow SW. Inhibition of glutamine synthetase reduces ammonia-induced astrocyte swelling in rat, pp 589–99, ©1996, with permission from Elsevier Science.MSO prevents cerebral edema Using the same protocol described in the morphologic study, the effect of MSO on plasma ammonia, brain glutamine synthetase, brain glutamine, and brain water was studied (67). Figure 4 demonstrates the effects of MSO; it inhibited glutamine synthetase (Figure 4B) and prevented the increase in cortical glutamine (Figure 4C) and cortical water (Figure 4D) in the hyperammonemic group. MSO also prevented the decrease in cortical specific gravity (67) (not shown). There are other conditions in which astrocyte swelling may occur (32), but the prevention of swelling by MSO by virtue of its glutamine synthetase inhibitory effect identifies glutamine accumulation as the mechanism for the swelling. To eliminate a possible role of MSO’s ability to also inhibit γ- glutamylcysteinegamma;-glutamylcysteine synthase (GGS), the specific GGS inhibitor, buthionine sulfoximine, was studied and found not to protect hyperammonemic animals from an increase in brain water (73), suggesting that the GGS inhibitory effect of MSO played little part in the results of these experiments. The 2 groups of animals that received MSO had higher plasma ammonia concentrations (see Figure 4A) than their respective controls, presumably as a consequence of glutamine synthetase inhibition in muscle and lung, both of which have glutamine synthetase activity and high rates of glutamine synthesis (13). These data suggest that glutamine accumulation in the astrocyte is necessary for the increase in brain water in hyperammonemic states and that, although ammonia is necessary, it alone is insufficient to cause hyperammonemic encephalopathy.Fig. 4: The protective effect of methionine sulfoximine (MSO) on hyperammonemia-induced cerebral edema. A. Plasma ammonia levels in 4 groups of animals; a group that received vehicle (Control), a group that received vehicle plus MSO (MSO), a group that received ammonia infusions (NH4), and a group that received ammonia and MSO (NH4 + MSO). B. The activity of brain glutamine (Gln) synthetase in the same 4 groups. C. Brain glutamine (Gln) levels in the same 4 groups. D. Brain water in the same 4 groups. The p values refer to the data immediately above the p value. ns = not significant. Reprinted with permission from reference 67, American Journal of Physiology, vol. 261, Takahashi H, Koehler RC, Brusilow SW, Traystman RJ. Inhibition of brain glutamine accumulation prevents cerebral edema in hyperammonemic rats, pp H825–9, 1991.In vitro studies In vitro studies of astrocytes in culture implicate the osmotic effect of glutamine on the increase in cell volume. In these experiments, an increase in medium ammonia concentrations caused astrocyte swelling (53) and increase of cell glutamine levels (53,76), both of which were prevented if MSO was included in the medium. Additional evidence implicating osmotic factors in hyperammonemia is the down-regulatory response of myo-inositol (76) when astrocytes are exposed to high concentrations of ammonia (29). These in vitro data support the concept that glutamine may function as an increased organic osmolyte and myo-inositol functions as a compensatory decreased organic osmolyte in hyperammonemic states, both exerting their effect in the absence of liver “toxins.” The clinical correlate of these in vitro changes is persuasive; in hepatic encephalopathy, an inverse relationship exists between the brain glutamine and myoinositol levels (as measured in vivo by proton mass spectroscopy [24]). The growing recognition of the importance of organic osmolytes (for example, myo-inositol, taurine, sorbitol, glycerophosphorylcholine, betaine) in cell volume regulation has been reviewed (48,65). Hyperammonia-Induced Changes in Brain Function That hyperammonemia results in functional changes in addition to a mass effect is demonstrated by its interference in the astrocytes’ role in regulating extracellular potassium activity (25). Using the protocol described earlier in evaluating cerebral edema in rats, plasma ammonia concentrations of approximately 700 μmol/L result in a progressive increase in brain extracellular potassium activity, and inhibition of glutamine synthetase by MSO attenuates this effect (Figure 5).Fig. 5: The protective effect of methionine sulfoximine (MSO) on hyperammonemia-induced changes in cortical extracellular potassium activity. A. A control period, during which a 5-hour intravenous infusion of sodium acetate was maintained, showing normal stability of potassium activity. B. During an ammonium acetate infusion, showing an increase in cortical potassium activity. C. During an ammonium acetate infusion after pretreatment with MSO, showing attenuation of the increase in ammonia-induced increase in cortical extracellular potassium activity. Reprinted with permission from reference 66, Journal of Cerebral Blood Flow and Metabolism, vol. 17, Sugimoto H, Koehler RC, Wilson DA, Brusilow SW, Traystman RJ. Methionine sulfoximine, a glutamine synthetase inhibitor, attenuates increased extracellular potassium activity during acute hyperammonemia, pp 44–9, ©1997 Lippincott Williams & Wilkins.MSO also prevents the accumulation of neutral amino acids in brain during hyperammonemia (30). Figure 6 demonstrates this effect for glutamine and 3 amino acids that are substrates for the synthesis of neurotransmitters: phenylalanine and tyrosine (precursors of dopamine and epinephrine) and tryptophan (precursor of serotonin and quinolinate).Fig. 6: The protective effect of methionine sulfoximine (MSO) in prevention of hyperammonemia-induced increases in brain amino levels in rats. The open bars represent a group that received vehicle only, the crosshatched bars represent rats that received an ammonia infusion only, and the horizontally hatched bars represent rats that were pretreated with MSO followed by an ammonia infusion. Differences between the control group and the ammonia-only infusion group and between the 2 ammonia-infusion groups (with and without pretreatment with MSO) were significant at p values of <0.05. Abbreviations: Gln, glutamine; Phe, phenylalanine; Trp, tryptophan; Tyr, tyrosine. Reprinted with permission from reference 30, Journal of Surgical Research, vol. 36, Jonung T, Rigotti P, Jeppsson B, James JH, Peters JC, Fischer JE. Methionine sulfoximine prevents the accumulation of large neutral amino acids in brain of hyperammonemic rats, pp 349–53, ©1984 Academic Press.That more than a mass effect may be a consequence of glutamine accumulation on astrocyte dysfunction is also suggested by a study (27) in which MSO prevented the loss of hyperammonia-induced hypercapnic reactivity of pial arteries, vessels not invested with astrocyte endfeet. Figure 7 summarizes the pathophysiology of hyperammonemic encephalopathy.Fig. 7: The pathophysiology of hyperammonemic encephalopathy. Hyperammonemia activates astrocyte glutamine synthesis and accumulation causing astrocytes to swell as a consequence of the osmotic effect of glutamine. Astrocyte swelling has 2 effects; it may have a mass effect causing cerebral edema and it may lead to astrocyte dysfunction. Regulatory organic osmolytes (for example, myo-inositol) are down-regulated to compensate for the increase in glutamine accumulation. Reprinted with permission from reference 4, Scriver CR, Beaudet AL, Sly WS, Valle D, eds. The Metabolic and Molecular Bases of Inherited Disease, pp 1909–63. ©2001, New York: McGraw-Hill. Reproduced with permission of the McGraw-Hill Companies.L-methionine S-sulfoximine: A Pharmacologic Agent? Although it has been thought that the convulsant effect of MSO was linked to its glutamine synthetase inhibitory effect, there is persuasive evidence to suggest that the 2 are unrelated and that MSO might be a useful drug. For example, the mean rat brain glutamine level (an indirect measure of glutamine synthetase activity) decreased to the same degree at MSO doses (mg/kg) of 200, 100, and 50; however, the rats receiving the lower doses “did not differ in any way from the control animals” whereas the animals receiving the highest dose had seizures (21). (The precise dosages of MSO in these and many other studies are uncertain because the stereoisomers administered were frequently not fully described.) Another study (61) demonstrated that L-methionine protected rats from the convulsant effect of MSO, but did not affect MSO’s inhibition of glutamine synthetase. The clinical effect of mixed stereoisomers of MSO has been evaluated in 2 primate studies. Monkeys were found to tolerate MSO doses of 50 and 100 mg/kg “without observable effect” (22); glutamine synthetase activity was not measured. When mixed stereoisomers of MSO were given in daily doses to terminally ill cancer patients, significant side effects were noted after 2 days: agitation, disorientation, and hallucinations (34). Interpretation of this study is hindered by the dire status of the patients and the potential cumulative effect of daily doses, a regimen probably not necessary for an irreversible inhibitor like MSO, which has an inhibition duration of at least 7 days (35). Because there are species-specific responses to MSO (22), a dose-response evaluation of the active stereoisomer, L-methionine S-sulfoximine, on glutamine synthetase activity and clinical symptoms in primates may reveal a dose that inhibits glutamine synthetase activity in the absence of a convulsant effect and thereby may be useful in treating hyperammonemic encephalopathy. Summary Hyperammonemia is most often observed in the clinical setting of liver failure. However, understanding the consequences of hyperammonemia under such conditions is hindered by the many other metabolic abnormalities associated with liver failure. To clarify the clinical picture of hyperammonemic encephalopathy, a patient with a urea cycle disorder is described, a setting in which mechanisms attributable to severe liver impairment are absent. Hyperammonemic encephalopathy in urea cycle disorders is compared with the encephalopathy of fulminant hepatic failure and hepatic encephalopathy. Such a comparison reveals that the following features are shared by all 3 conditions: hyperammonemia; respiratory alkalosis; increased levels of glutamine in plasma, cerebrospinal fluid, and brain; decreased brain levels of myo-inositol; and astrocyte swelling with few neuronal changes. These findings in patients, when added to data obtained from experimental animals made hyperammonemic but with normal liver function, support the proposal that hyperammonemic encephalopathy is a consequence of astrocyte swelling and dysfunction resulting from the osmotic effects of astrocyte glutamine synthesis (activated by ammonia) and accumulation. If this proposal is correct the physiologic changes induced by hyperammonemia should be prevented or ameliorated by inhibiting brain glutamine synthetase activity. Animal studies demonstrate that the astrocyte is the site of brain glutamine synthetase and that abnormalities induced by hyperammonemia, for example, cerebral edema, cerebral glutamine accumulation, astrocyte swelling, and brain functional changes, are prevented or ameliorated by treatment with the glutamine synthetase inhibitor, L-methionine S-sulfoximine (MSO). Therapeutic implications of MSO are discussed.
Referência(s)