Cerebral lactic acidosis correlates with neurological impairment in MELAS

Transcription

1 Cerebral lactic acidosis correlates with neurological impairment in MELAS P. Kaufmann, MD, MSc; D.C. Shungu, PhD; M.C. Sano, PhD; S. Jhung, BS; K. Engelstad, BS; E. Mitsis, PhD; X. Mao, MS; S. Shanske, PhD; M. Hirano, MD; S. DiMauro, MD; and D.C. De Vivo, MD Abstract Objective: To evaluate the role of chronic cerebral lactic acidosis in mitochondrial encephalomyopathy, lactic acidosis, and stroke-like episodes (MELAS). Methods: The authors studied 91 individuals from 34 families with MELAS and the A3243G point mutation and 15 individuals from two families with myoclonus epilepsy and ragged red fibers (MERRF) and the A8344G mutation. Subjects were divided into four groups. Paternal relatives were studied as controls (Group 1). The maternally related subjects were divided clinically into three groups: asymptomatic (no clinical evidence of neurologic disease) (Group 2), oligosymptomatic (neurologic symptoms but without the full clinical picture of MELAS or MERRF) (Group 3), and symptomatic (fulfilling MELAS or MERRF criteria) (Group 4). The authors performed a standardized neurologic examination, neuropsychological testing, MRS, and leukocyte DNA analysis in all subjects. Results: The symptomatic and oligosymptomatic MELAS subjects had significantly higher ventricular lactate than the other groups. There was a significant correlation between degree of neuropsychological and neurologic impairment and cerebral lactic acidosis as estimated by ventricular MRS lactate levels. Conclusions: High levels of ventricular lactate, the brain spectroscopic signature of MELAS, are associated with more severe neurologic impairment. NEUROLOGY 2004;62: Mitochondrial myopathy, encephalopathy, lactic acidosis, and stroke-like episodes (MELAS) includes in its definition 1) encephalopathy, frequently with seizures and progressive dementia; 2) stroke-like episodes at a young age; and 3) biochemical or morphologic evidence for mitochondrial defects, such as lactic acidosis or ragged-red fibers (RRF) in the muscle biopsy. 1 Just 6 years after its first clinical description, an A-to-G transition at nucleotide (nt) 3243 in the mitochondrial DNA (mtdna) trna Leu(UUR) gene was associated with the MELAS syndrome. 2 Another mitochondrial encephalopathy, myoclonus epilepsy and RRF (MERRF), was associated with an A83444G transition within a transfer RNA (trna Lys ), yet its phenotype is quite different and its prognosis is more favorable. 3,4 Despite the progress in elucidating the genetic etiology of MELAS, its pathophysiology remains poorly understood. It has been suggested that the respiratory chain defect in MELAS disturbs the cerebral oxidationreduction potential. The elevated NADH H would in turn influence the reduction of pyruvate in glial cells and the oxidation of lactate by neurons. Chronic lactic acidosis is expected to increase reducing equivalents. This would facilitate the anaerobic reaction from pyruvate to lactate in the glial cell, but would oppose the re-oxidation of lactate to pyruvate in the neuron. 5 The neuron would thus be deprived of its principal fuel source. 5,6 Several studies have suggested a more direct harmful effect of lactate on neurons In animal models of ischemia, lactic acidosis increases the neuronal injury in the infarcted area. 13,14 Impaired glucose homeostasis together with chronic cerebral lactic acidosis may also contribute to brain injury in MELAS given that diabetes mellitus is a frequent complication of the A3243G mutation. We hypothesized that 1) lactic acidosis can cause neuronal injury, 2) lactic acidosis may exacerbate ischemic brain damage, and 3) hyperglycemia can worsen lactic acidosis and ischemic brain damage. We sought to evaluate the role of chronic cerebral lactic acidosis in the pathogenesis of MELAS. Methods. Subjects. We studied 91 individuals from 34 families. In each family, the proband had been diagnosed with MELAS. Fifteen individuals from two families with MERRF and the A8344G mutation were studied as disease controls. In addition to the probands, we studied matrilineal and paternal relatives. Subjects were divided into four groups according to their relationship to the proband and their clinical symptomatology: 11 paternal relatives from A3243G and A8344G families were studied as controls (Group 1). The matrilineal relatives were divided into three clinical groups: asymptomatic (no clinical evidence of neurologic disease) (Group 2, n 33 for MELAS and n 3 for MERRF); oligosymptomatic (neurologic symptoms but without the full clinical picture of MELAS or MERRF) (Group 3, n 20 for MELAS and n 4 for MERRF); and symptomatic (fulfilling MELAS or MERRF diagnostic criteria) (Group 4, n 30 for MELAS and n 5 for MERRF). The A3243G carriers were younger than the A8344G carriers studied (table). All subjects underwent a comprehensive evaluation, including standardized medical and neurologic history and examination, a From the Departments of Neurology (Drs. Kaufmann, Sano, Mitsis, Shanske, Hirano, DiMauro, and De Vivo, and S. Jhung and K. Engelstad), Pediatrics (S. Jhung and K. Engelstad, and Dr. De Vivo), and Radiology (Dr. Shungu and X. Mao), The Gertrude H. Sergievsky Center (Dr. Sano), Columbia University, New York, NY. Supported by grant # 5-PO1-HD32062 (S.D.M.), an Irving Scholar Award and K12 Award (P.K.), and the Colleen Giblin Foundation (D.C.D.V.). Received May 6, Accepted in final form December 15, Address correspondence and reprint requests to Dr. Darryl C. De Vivo, The Neurological Institute, Columbia University, 710 W 168th Street, New York, NY 10032; dcd1@columbia.edu Copyright 2004 by AAN Enterprises, Inc. 1297

2 Table Columbia Neurological Score (CNS), global neuropsychological score (GNP), MRS ventricular lactate estimates (MRS Lac, in institutional units), venous lactate (V Lac, in mm/l), age (in years), and fasting glucose (Gluc, in mg/dl) in families harboring the A3243G mutation and the A8344G mutation Clinical group CNS GNP MRS Lac V Lac Age Gluc Control group, n Mutation Asymptomatic Oligosymptomatic Symptomatic Values are mean SD. The paternal, noncarrier controls from all families were analyzed together. neuropsychological test battery, and estimation of brain and ventricular lactate by MRS. The MELAS probands were evaluated no earlier than 3 months after a stroke-like episode and during a period of neurologic stability and in the absence of concurrent, active medical disease. MRS. Following T1-weighted localizer images, multislice proton MRS was performed on each subject using a previously described method. 15 The data analyzed in this study were obtained from two 15-mm axial-oblique brain sections. The lower of the two slices traversed and encompassed the bodies of the lateral ventricles. The following acquisition sequences were used: echo time/ repetition time 280/2300 msec, field of view 240 mm, 32 x 32 phase-encoding steps with circular k-space, and 256 sample points. The strong pericranial lipid resonances were suppressed using octagonally tailored outer volume presaturation pulses. Because there are no neuronal structures in the lateral ventricles, the levels of N-acetylaspartate, creatine, and choline are expected to be undetectable. The lactate peaks are therefore quantitated as ratio of the lactate signal area over the mean square root of the background noise. 16 Figure 1 demonstrates exemplary MRS studies of three subjects from A3243G families. Neurologic examination. A new, quantitative tool was developed to summarize the clinical evaluations of patients with mitochondrial disease. The tool focuses on physical findings to complement the extensive neuropsychological testing and to capture the following domains: 1) height, weight, and head circumference; 2) general medical examination; 3) general neurologic examination; 4) cranial nerves; 5) stance and gait; 6) involuntary movements; 7) sensation; 8) cerebellar function; 9) muscle bulk, tone, and strength; 10) myotatic reflexes; 11) Babinski signs; and 12) other findings. Results of these domains were scored as normal or abnormal and summarized in the Clinical Neurologic Score (CNS), ranging from 0 to76, with 76 being normal. The CNS is clear, easily administered, and has good inter-rater reliability in a subset of 10 subjects who were evaluated by the two principal neurologists, D.C.D. and P.K., on the same day (kappa 0.95). Neuropsychological testing. Areas of cognitive function were assessed with a brief global mental status examination and specific tests to assess the following domains: abstract reasoning, verbal memory, visual memory, language (consisting of naming and fluency), executive function, attention, and visual-spatial abil- Figure 1. T1-weighted MR brain images at the level of the lateral ventricles for three subjects: on the left, a fully symptomatic mitochondrial encephalomyopathy, lactic acidosis, and stroke-like episodes (MELAS) proband; in the middle, an oligosymptomatic mutation carrier; and on the right, an asymptomatic control subject. Above each image are on the left 1 H MR spectra of a voxel in the left gray matter and on the right MR spectra from a voxel in the left lateral ventricle. The spectral resonances have been assigned to CHO (total choline), CR (total creatine), NAA (N-acetylaspartate), and LAC (lactate). All the spectra have been plotted on the same vertical scale NEUROLOGY 62 April (2 of 2) 2004

3 ities. Each domain was evaluated using standard tests appropriate to the age of the patient. Each test of the neuropsychological battery was scored according to standard guidelines. Individual domain scores were expressed as percentile of normative data and a global mean score was derived by dividing the individual domain results by the number of domains. Laboratory studies. The mitochondrial DNA point mutation was confirmed by leukocyte DNA studies and we analyzed the relative amount of mitochondrial genomes containing the A3243G or A8344G mutation. Total DNA was extracted according to published protocols and PCR amplification and restriction fragment length polymorphism (RFLP) analysis of total DNA extracted from leukocytes was performed. Appropriate primers and restriction enzymes were chosen to detect the A3243G 17 or the A8344G 18 mutations. Venous lactate and fasting glucose were measured according to standard methods in the Clinical Chemistry Laboratory of Columbia Presbyterian Medical Center. The normal ranges established in this laboratory are 0.5 to 2.2 mm/l for venous lactate and 70 to 105 mg/dl for fasting blood glucose. Figure 2. MRS ventricular lactate (expressed in institutional units [i.u.], mean SEM in institutional units) by clinical group in MELAS/3243 families. Results. Mutation analysis. Thirty-four families had the A3243G mutation. Two families harbored the A8344G mutation. Mutations were found in leukocyte DNA of all probands (Group 4). Mutations were also detected in 16 of 20 oligosymptomatic MELAS relatives, 21 of 33 asymptomatic MELAS relatives, 4 of 5 oligosymptomatic MERRF relatives, and 2 of 3 asymptomatic MERRF relatives. Eleven paternal relatives from MELAS and MERRF families were studied as controls and the mutations were not found (Group 1). In mutation carriers, there was no significant correlation between the relative amount of mutated mitochondrial genomes and MRS ventricular lactate estimates (r with p for A32343G carriers, and r with p 0.86 for A8344G carriers). Comparing MELAS and MERRF probands. The MELAS probands had higher ventricular lactate estimates on MRS than the MERRF probands (t 2.368, p 0.05) (see the table). The MELAS probands also showed lower scores on formal neuropsychological testing than the MERRF probands (t 4.307, p 0.001). The mean score for the degree of impairment on neurologic examination (as measured using the CNS instrument) was lower for the MERRF probands, but there was no significant difference between the MERRF and MELAS groups. Clinical analysis of A3243G mutation carriers. Within the MELAS/A3243G families, the CNS score was progressively lower in asymptomatic, oligosymptomatic, and fully symptomatic mutation carriers. There was an overall difference between groups by analysis of variance (ANOVA) (F , p 0.001) and a difference for all pairwise group comparisons except between the asymptomatic and control groups by Bonferroni-t-test (p 0.05). There was a similar progressive worsening of neuropsychological impairment in asymptomatic, oligosymptomatic, and symptomatic A3243G mutation carriers. Comparing the groups by ANOVA, there was an overall difference between groups (F , p 0.001), and a difference of means comparing symptomatic individuals with all other groups (p 0.05, Bonferroni-t-test). Venous lactate in A3243 mutation carriers. Venous lactate was elevated in the symptomatic and oligosymptomatic A3243G carriers when compared to controls (ANOVA, F 9.98, p 0.01). Venous and ventricular lactate were associated in A3243G mutation carriers (r 0.41, p 0.01). However, in 30 of 81 (37%) mutation carriers, the results were not concordant (i.e., 12/30 symptomatic subjects, 10/19 oligosymptomatic subjects, and 9/32 asymptomatic subjects had abnormal results on one test, but not on the other). The mean fasting blood glucose was elevated in symptomatic and oligosymptomatic A3243G mutation carriers when compared to the asymptomatic and control groups, but the difference was not significant. There was a weak correlation between fasting glucose (mg/dl) and ventricular lactate (MRS, r 0.32, p 0.01), and between fasting blood sugar (mg/dl) and venous lactate (mm/l, r 0.51, p 0.01). There was no correlation between fasting glucose (mg/ dl) and 1) daily living functional abilities (Karnofsky score, r 0.15, p 0.2), 2) neurologic examination (CNS, r 0.13, p 0.3), or 3) neuropsychological performance score (r 0.11, p 0.4). MRS ventricular lactate within MELAS/A3243G and MERRF/A8344G families. The MRS ventricular lactate differed significantly between groups, with lactate estimates rising from asymptomatic to oligosymptomatic and to symptomatic A3243G mutation carriers. Figure 1 demonstrates an abnormal MRI of the brain in a fully symptomatic A3243G mutation carrier and the associated MR spectra from a voxel over the cerebral gray matter and over the lateral ventricular fluid. The oligosymptomatic mutation carrier from the same family has a normal MR imaging study, but abnormal MR spectra. The overall difference of ventricular lactate between clinical groups was significant (F , p 0.001) and so were the pairwise comparisons between the following groups: control and oligosymptomatic, asymptomatic and oligosymptomatic, control and symptomatic, asymptomatic and symptomatic (Groups 1 and 2 vs 3, Groups 1 and 2 vs 4) (p 0.05) (figure 2). We noted a linear correlation between CNS and MRS ventricular lactate with r in matrilinear A3243G subjects (p 0.001) (figure 3A). We also identified a correlation between neuropsychological mean score and MRS April (2 of 2) 2004 NEUROLOGY

4 Figure 3. (A) MRS ventricular lactate (iu) vs Columbia Neurologic Scale (CNS) score (ranging from 0 to 76 normal) for all subjects from MELAS/3243 families by clinical group: asymptomatic, oligosymptomatic, symptomatic. (B) MRS ventricular lactate [iu] vs global neuropsychological score (0 normal to 2) for all subjects from MELAS/3243 families by clinical group: asymptomatic, oligosymptomatic, symptomatic. ventricular lactate values with r (p 0.001) in matrilinear A3243G subjects (figure 3B). Within MERRF families, the ventricular lactate ratings of clinical groups did not show a significant difference, but there was a trend toward higher lactate values in more symptomatic individuals (see the table). Discussion. Our data indicate that neuronal dysfunction correlates with cerebral lactic acidosis in families harboring the MELAS/A3243G mutation. Heteroplasmy and mitotic segregation can lead to mutated DNA levels below the detection threshold in leukocyte DNA. However, even in cases with undetectable mutation in lymphocytes, our pedigree analysis suggests that these maternal relatives are 1300 NEUROLOGY 62 April (2 of 2) 2004 obligate carriers. Furthermore, we have confirmed the presence of mutated mtdna in most maternal relatives studying additional tissues, such as urine sediment, fibroblasts, buccal mucosa cells, and hair follicles. 19 A previous report had suggested that MRS brain lactate was linearly related to the relative amount of mutated mitochondrial genomes (calculated as average of several tissues). 20 In our sample, we did not find a significant relationship between MRS ventricular lactate and the relative amount of mutated mitochondrial genomes in leukocyte DNA. This may be due in part to the fact that the amount of mutated mitochondrial genomes in leukocytes differs from that in brain tissue because mtdna are randomly distributed at cell division. We are collecting additional tissues from mutation carriers to further study the relationship between the amount and distribution of mutated mtdna and phenotypic expression. 19 To assess the functional impact of neuronal injury, we used quantitative measures of neurologic and cognitive impairment. We believe that the neuropsychological and neurologic deficits measured by our instruments are not merely the consequences of stroke-like episodes because differences were found not only between patients with or without clinical CNS involvement (i.e., history of seizures or strokes), but also between individuals and groups without clinical CNS involvement. It has been suggested that the A3243G mutation causes a progressive encephalopathy with dementia. 21 The course of this progressive encephalopathy is punctuated by mostly posterior stroke-like events causing episodic deficits with partial recovery and increasing disease burden. Our data support this notion and show that the progressive encephalopathy precedes the final stage of MELAS when mutation carriers become fully symptomatic and develop focal CNS involvement. Furthermore, the cerebral lactic acidosis correlates not only with the global neuropsychological score, but also with subscores measuring domains (e.g., frontal domains) that are not typically affected by the stroke-like insults of MELAS (which occur predominantly in posterior brain regions involved in visual processing). Cerebral lactic acidosis as estimated by MRS and venous lactate measurements is correlated in A3243G mutation carriers. However, since several subjects had elevated MRS brain lactate but normal venous lactate, we focused our analyses on MRS estimates, which would be expected to more closely reflect cerebral metabolism. Our hypothesis is further strengthened by the comparison with disease controls. Even though the number of individuals harboring the MERRF/ A8344G mutation in our sample is relatively small, there is a significant difference in cerebral lactic acidosis and neuropsychological performance between MELAS and MERRF probands. The lack of a major difference in neurologic examination may be because our instrument (designed to supplement the neuropsychological testing) is largely directed toward the

5 neuromuscular system, which is often more severely affected in MERRF patients. Both MELAS and MERRF are diseases caused by point mutations within mitochondrial transfer RNA, yet their phenotype and progression are very different. 22 MELAS is a more severe syndrome with younger onset and truncated course (mean survival from disease onset 6.5 years), often leading to severe disability and premature death within a few years of onset. MERRF, on the other hand, typically has a prolonged course with progressive disability, but a more normal life expectancy (no deaths in our cohort and a mean observation period from disease onset to assessment of 24.3 years). 4 It is difficult to explain such different clinical courses on the basis of the underlying molecular changes, since both trna mutations would be expected to interfere with mitochondrial transcription and protein synthesis. We believe that the observed clinical and radiologic differences between the two syndromes are due to a distinctive metabolic consequence; namely, that MELAS typically causes significant cerebral lactic acidosis. As our data show, the lactic acidosis is present not only in probands during episodes of worsening, but also in oligosymptomatic mutation carriers. MERRF, on the other hand, is associated with minimal lactic acidosis. Our hypothesis that cerebral lactic acidosis plays a key role in neuronal injury in MELAS is supported by an extensive body of in vitro and animal experiments in the literature. First, lactic acidosis caused cell swelling and death in neuronal cultures in vitro. 7 Second, in neuronal and glial cultures cell death occurred at higher ph when the cells were subjected to lactic acidosis than when they were exposed to HCl, suggesting that the relatively higher cell membrane permeability of lactic acid may contribute to its toxicity. 8,9 Third, cultured hippocampal rat neurons exposed to moderate lactic acidosis for 10 minutes showed neurodegeneration. 10 Fourth, dripping lactic acid onto the surgically exposed spinal cord of adult male rats caused severe necrotizing myelopathy, with an early effect on small blood vessel walls followed by nerve fiber alterations. 11 Fifth, in rats, lactate injections into the parietal cortex produced brain necrosis when the ph was less than or equal to Studies of experimental incomplete cerebral ischemia models suggest that lower intracellular ph and impaired mitochondrial function are related to greater infarct volume, 13 and that lactic acidosis prevents normalization of cortical energy metabolism as estimated from concentrations of phosphocreatine, ATP, ADP, and AMP. 14 This apparent exacerbation of ischemic injury may be particularly deleterious in MELAS. Even though most studies underscore the pathogenic importance of primary metabolic failure over vascular insufficiency, the controversy between vascular and metabolic hypotheses in MELAS remains open. This is mainly due to the evidence of histologic and histochemical mitochondrial changes in small blood vessel walls of MELAS Impaired glucose homeostasis may also contribute, together with lactic acidosis, to brain injury in MELAS. This is of great clinical significance and may in part explain the severity of phenotypes associated with A3243G, given that diabetes occurred in 15% of A3243G carriers, but only in 3% of A8344G carriers in a meta-analysis of 245 individuals. 22 Neuronal injury was aggravated in a rat model of severe incomplete ischemia in the presence of excessive tissue lactic acidosis due to glucose pretreatment. 26 Similarly, hyperglycemia exacerbated the neuronal injury in a rat cerebral ischemia model, possibly because the accumulation of glycolytically derived lactate caused acidosis. 27 However, the role of acidosis in neuronal injury is controversial and there are in vitro observations suggesting that acidosis may ameliorate neuronal injury caused by glutamate and anoxia. 28 Also, lactic acidosis at ph 7.3 did not produce neuronal damage and the deleterious effects of lactate injected into the cortices of animals were conditional on coexisting low ph. 29 Our study shows a strong association between clinical phenotype and cerebral lactic acidosis in MELAS and may provide some insight into the differences in clinical courses between MELAS and MERRF. In MELAS, neuronal dysfunction and lactic acidosis are likely due to the underlying disease process. The critical question to be answered is whether there is a causal relationship between elevated brain lactate and neuronal injury. In an attempt to address this question, we are currently investigating the efficacy of dichloroacetate, an experimental lactate-lowering agent, in mitigating the clinical progression of MELAS. References 1. Pavlakis SG, Phillips PC, DiMauro S, et al. Mitochondrial myopathy, encephalopathy, lactic acidosis, and stroke-like episodes: a distinctive clinical syndrome. Ann Neurol 1984;16: Goto Y-I, Nonaka I, Horai S. A mutation in the trna Leu(UUR) gene is associated with the MELAS subgroup of mitochondrial encephalomyopathies. Nature 1990;348: Shoffner JM, Lott MT, Lezza AMS, Seibel P, Ballinger SW, Wallace DC. Myoclonic epilepsy and ragged red fiber disease (MERRF) is associated with a mitochondrial DNA trna Lys mutation. Cell 1990;43: Sano M, Polanco Y, De Vivo DC. Comparative analysis of mitochondrial encephalopathy, lactic acidosis and stroke-like episodes and myoclonus epilepsy and ragged red fibers. Ann Neurol 1998;44: De Vivo DC. Cerebral energy failure. Curr Neurol Neurosci Rep 2001;1: Magistretti PJ. Cellular bases of functional brain imaging: insights from neuron-glia metabolic coupling. Brain Res 2000;886: Staub F, Mackert B, Kempski O, Peters J, Baethmann A. Swelling and death of neuronal cells by lactic acid. J Neurol Sci 1993;119: Goldman SA, Pulsinelli WA, Clarke WY, Kraig RP, Plum F. The effects of extracellular acidosis on neurons and glia in vitro. J Cereb Blood Flow Metab 1989;9: Nedergaard M, Goldman SA, Desai S, Pulsinelli WA. Acid-induced death in neurons and glia. J Neurosci 1991;8: Himmelseher S, Pfenninger E, Georgieff M. Basic fibroblast growth factor reduces lactic acid-induced neuronal injury in rat hippocampal neurons. Crit Care Med 1998;26: Balentine JD, Greene WB. Myelopathy induced by lactic acid. Acta Neuropathol 1987;73: Kraig RP, Petito CK, Plum F, Pulsinelli WA. Hydrogen ions kill brain at concentrations reached in ischemia. J Cereb Blood Flow Metab 1987; 7: Anderson RE, Tan WK, Martin HS, Meyer FB. Effects of glucose and PaO2 modulation on cortical intracellular acidosis, NADH redox state, and infarction in the ischemic penumbra. Stroke 1999;30: April (2 of 2) 2004 NEUROLOGY

6 14. Rehncrona S, Rosen I, Siesjo BK. Brain lactic acidosis and ischemic cell damage: 1. Biochemistry and neurophysiology. J Cereb Blood Flow Metab 1981;1: Duyn JH, Gillen J, Sobering G, van Zijl PC, Moonen CT. Multisection proton MR spectroscopic imaging of the brain. Radiology 1993;188: Shungu DC, Mao X, Kaufmann P, et al. Comparison of in vitro and in vivo CSF lactate in A3243G MELAS patients: a viable method for absolute quantitation of CSF lactate by 1 H MRSI. Proc Int Soc Magn Reson Med 2002;10: Kaufmann P, Koga Y, Shanske S, et al. Mitochondrial DNA and RNA processing in MELAS. Ann Neurol 1996;40: Masucci JP, Davidson M, Koga Y, Schon EA, King MP. In vitro analysis of mutations causing myoclonus epilepsy with ragged-red fibers in the mitochondrial trna Lys gene: two genotypes produce similar phenotypes. Mol Cell Biol 1995;15: Shanske S, Pancrudo J, Engelstad K, et al. Study of mitochondrial DNA heteroplasmy in multiple tissues. Neurology 2003;60:A Dubeau F, De Stefano N, Zifkin BG, Arnold DL, Shoubridge EA. Oxidative phosphorylation defect in the brains of carriers of the trna leu(uur) A3243G mutation in a MELAS pedigree. Ann Neurol 2000;47: Hirano M, Pavlakis SG. Mitochondrial myopathy, encephalopathy, lactic acidosis, and stroke like episodes (MELAS): current concepts. J Child Neurol 1994;9: Chinnery PF, Howell N, Lightowlers RN, Turnbull DM. Molecular pathology of MELAS and MERRF. The relationship between mutation load and clinical phenotypes. Brain 1997;120: Sakuta R, Nonaka I. Vascular involvement in mitochondrial myopathy. Ann Neurol 1989;25: Ohama E, Ohara S, Ikuta F, Tanaka K, Nishizawa M, Miyatake T. Mitochondrial angiopathy in cerebral blood vessels of mitochondrial encephalomyopathy. Acta Neuropathol 1987;74: Tanji K, Kunimatsu T, Vu TH, Bonilla E. Neuropathological features of mitochondrial disorders. Cell Dev Biol 2001;12: Paljarvi L. Brain lactic acidosis and ischemic cell damage: a topographic study with high-resolution light microscopy of early recovery in a rat model of severe incomplete ischemia. Acta Neuropathol 1984;64: Phillis JW, Song D, Guyot LL, O Regan MH. Lactate reduces amino acid release and fuels recovery of function in the ischemic brain. Neurosci Lett 1999;272: Tombaugh GC, Sapolsky RM. Mild acidosis protects hippocampal neurons from injury induced by oxygen and glucose deprivation. Brain Res 1990;506: Petito CK, Kraig RP, Pulsinelli WA. Light and electron microscopic evaluation of hydrogen ion-induced brain necrosis. J Cereb Blood Flow Metab 1987;7: Note that this issue of Neurology has a NeuroImage that does not appear in the print journal: Severe restless legs syndrome presenting as intractable insomnia I. Arnulf, E. Konofal, C. Gauthier, and F. Chedru 1302 NEUROLOGY 62 April (2 of 2) 2004