BRIEF COMMUNICATIONS Progressive Loss of Cardiac Sympathetic Innervation in Parkinson’s Disease Sheng-Ting Li, MD, PhD, Raghuveer Dendi, MD, Courtney Holmes, CMT, and David S. Goldstein, MD, PhD This study addressed whether cardiac sympathetic denervation progresses over time in Parkinson’s disease. In 9 patients without orthostatic hypotension, 6-[18F]fluorodopamine positron emission tomography scanning was repeated after a mean of 2 years from the first scan. 6-[18F]fluorodopamine-derived radioactivity was less in the second scan than in the first scan, by 31% in the left ventricular free wall and 16% in the septum. In Parkinson’s disease, loss of cardiac sympathetic denervation progresses in a pattern of loss suggesting a dying-back mechanism. Ann Neurol 2002;52:220 –223 All of at least a dozen studies have agreed that patients with Parkinson’s disease have a high prevalence of neuroimaging evidence for decreased sympathetic innervation of the heart. Low myocardial concentrations of radioactivity have been noted after injection of the sympathoneural imaging agents 123I-metaiodobenzylguanidine1–12 and 6-[18F]fluorodopamine.13 Neurochemical assessments during right heart catheterization have confirmed that low concentrations of radioactivity result from loss of functional cardiac sympathetic nerve terminals.13 Although all patients with Parkinson’s disease and orthostatic hypotension have diffusely decreased sympathetic innervation throughout the left ventricular myocardium, among patients who do not have orthostatic hypotension, about half have diffusely decreased innervation and about half have normal or only locally From the Clinical Neurocardiology Section, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, MD. Received Nov 28, 2001, and in revised form Feb 28 and Mar 7, 2002. Accepted for publication Mar 7, 2002. Published online Jun 23, 2002, in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ana.10236 Address correspondence to Dr Li, Building 10, Room 6N252, National Institute of Neurological Disorders and Stroke, National Institutes of Health, 10 Center Drive, MSC-1620, Bethesda, MD 20892-1620. E-mail: email@example.com This article is a US Government work and, as such, is in the public domain in the United States of America. 220 Published 2002 by Wiley-Liss, Inc. decreased innervation.13 The latter finding afforded an opportunity to determine whether the loss of cardiac sympathetic innervation in Parkinson’s disease progresses over time and, if so, with what timing, pattern, and consistency across patients. This report describes the results of retesting such patients with 6-[18F]fluorodopamine positron emission tomography scanning after an average of 2 years. Patients and Methods The study protocol was approved by the Intramural Research Board of the National Institute of Neurological Disorders and Stroke. Each patient gave informed, written consent. Patients Thoracic positron emission tomography scanning was performed after intravenous injection of 6-[18F]fluorodopamine in 9 patients with Parkinson’s disease (age: mean, 60 years; SEM, 3 years). None of the patients had orthostatic hypotension, which was defined as a decrease in systolic blood pressure greater than 20mm Hg and decrease in diastolic pressure greater than 5mm Hg between the supine position and standing for 5 minutes. Caffeinecontaining beverages, cigarettes, and alcohol were not allowed for at least 24 hours before the scanning. Patients were allowed to take their usual medications, including L-dopa, except for medications known to inhibit neuronal uptake of catecholamines. Positron Emission Tomography Scanning 6-[18F]fluorodopamine, synthesized as described previously,14 was infused intravenously at a constant rate for 3 minutes. Tomography images (35 contiguous transaxial slices 4.25mm apart) were acquired for up to 30 minutes. Threedimensional positron emission tomography scans were obtained with an Advance whole-body scanner (General Electric, Milwaukee, WI). Transmission scans of 2 minutes and 8 minutes, with rotating 68Ge/68Ga pin sources, were obtained for attenuation correction and for confirming proper positioning in the scanner. Follow-up scanning was performed 1 to 4 years (mean, 2.0 years; SEM, 0.3 years) after the first scan, with the identical scanning procedure. None of the patients had orthostatic hypotension at the time of either test. Of the 9 patients, 2 had normal 6-[18F]fluorodopamine-derived radioactivity and 7 had locally decreased 6-[18F]fluorodopamine-derived radioactivity in the left ventricular myocardium at the time of the first test. Data Analysis and Statistics Tomography images were reconstructed after correction for attenuation and for physical decay of 18F. Cardiac images were analyzed as described previously.14 Briefly, circular regions of interest approximately half the ventricular wall thickness were placed on images of the septum, with time- averaged pictures of a single slice. Left ventricular septal radioactivity was averaged from two regions of interest for the 5-minute scanning interval with a midpoint about 8 minutes after initiation of the infusion. The same time interval was used for radioactivity in the liver and kidney. For radioactivity in structures of the head and neck, static three-dimensional data were obtained for 10 to 15 minutes. Images of noncardiac structures, including the liver, spleen, renal cortex, renal pelvis, salivary glands, nasopharyngeal mucosa, and thyroid, were reconstructed and analyzed by manual drawing of the regions of interest outlining the structures. Radioactivity concentrations were normalized by correction for the radioactivity concentration for the administered dose of radioactive drug per unit of body mass of the subject and were expressed as nCi-kg/ccmCi.14 Mean values for 6-[18F]fluorodopamine-derived radioactivity were compared with paired t tests. Differences between groups in trends over time of 6-[18F]fluorodopamine-derived radioactivity were assessed by analyses of variance for repeated measures. A p value of less than 0.05 defined statistical significance. Results At the time of the first scan, patients had had Parkinson’s disease for 5.7 ⫾ 1.2 years (range, 0.3–13 years). Disease severity averaged 2.2 ⫾ 0.2 (range, 1–3) of a maximum of 5. Heart rate and beat-to-beat systolic blood pressure changes during phase II of the Valsalva maneuver averaged 11 ⫾ 2bpm and ⫺43 ⫾ 5mm Hg for a mean arterial baroreflex-cardiovagal gain of 2.3 ⫾ 0.4ms/mm Hg (normal: mean, 8.5ms/mm Hg; SEM, 2.2ms/mm Hg). Plasma catecholamine levels (norepinephrine: mean, 2.34nmol/L; SEM, 0.37nmol/L; epinephrine: mean, 0.23nmol/L; SEM, 10nmol/L) were approximately normal. Most patients had tremor and urinary frequency. None of the patients developed orthostatic hypotension between the first and second scans. At the time of the first scan, 2 patients had normal myocardial 6-[18F]fluorodopamine-derived radioactivity, and 7 had decreased myocardial 6-[18F]fluorodopamine-derived radioactivity confined to the lateral wall or apex, so that no patient had 6-[18F]fluorodopamine-derived radioactivity more than two standard deviations below the normal mean in both the lateral wall and interventricular septum. All 9 patients had lower lateral wall concentrations of 6-[18F]fluorodopamine-derived radioactivity in the second scan than in the first scan (Fig 1). Left ventricular myocardial mean concentrations of 6-[18F]fluorodopamine-derived radioactivity decreased by 23% between the first scans (mean, 5,122nCi-kg/ cc-mCi; SEM, 564nCi-kg/cc-mCi) and second scans (mean, 6,634nCi-kg/cc-mCi; SEM, 447nCi-kg/cc-mCi; p ⫽ 0.003). Lateral wall radioactivity decreased by 31% between the first scans (mean, 4,107nCi-kg/ cc-mCi; SEM, 535nCi-kg/cc-mCi) and second scans (mean, 5,991nCi-kg/cc-mCi; SEM, 537nCi-kg/cc-mCi; p ⫽ 0.003), and septal radioactivity decreased by 16% between the first scans (mean, 6,137nCi-kg/cc-mCi; SEM, 716nCi-kg/cc-mCi) and second scans (mean, 7,278nCi-kg/cc-mCi; SEM, 385nCi-kg/cc-mCi; p ⫽ 0.05; Fig 2, Table). In 1 patient, the lateral ventricular wall was not visualized in the second scan, and the tissue concentration was assumed to be equal to the left ventricular chamber concentration. With the exclusion of data from the patient for whom the left ventricular wall was not visualized in the second scan, the percentage decrease in lateral wall radioactivity between the first and second scans (mean, Fig 1. Progressive loss of myocardial 6-[18F]fluorodopamine-derived radioactivity in a patient with Parkinson’s disease. Li et al: Sympathetic Innervation in Parkinson’s Disease 221 Fig 2. Concentrations (mean ⫾ standard error of the mean) of 6-[18F]fluorodopamine-derived radioactivity in the (top) lateral left ventricular wall and (bottom) interventricular septum in normal control subjects (open circles), patients with Parkinson’s disease at the time of the first scan (filled squares), and the same patients at the time of the second scan an average of 2 years later (open squares). 32%; SEM, 7%) exceeded the percentage decrease in septal wall radioactivity (mean, 13%; SEM, 7%; t ⫽ 2.5, p ⫽ 0.04). Tissue concentrations of 6-[18F]fluorodopaminederived radioactivity in the liver, renal cortex, renal pelvis, salivary glands, and thyroid did not change between the two scans (see Table). Radioactivity in the spleen, however, was lower in the second scan (mean, 5,020nCi-kg/cc-mCi; SEM, 137nCi-kg/cc-mCi) than in the first scan (mean, 6,495nCi-kg/cc-mCi; SEM, 364nCi-kg/cc-mCi; p ⫽ 0.001). Discussion These findings, based on sympathetic neuroimaging with 6-[18F]fluorodopamine positron emission tomography scanning, indicate that in patients with Parkinson’s disease who have normal or only locally decreased cardiac sympathetic innervation, the loss of innervation progresses over time, especially in the lateral ventricular wall, in which 6-[18F]fluorodopamine-derived radioactivity decreased by about 30% over an average of 2 222 Annals of Neurology Vol 52 No 2 August 2002 years. This rate of loss of sympathetic terminals appears to be at least as high as the rate of loss of nigrostriatal dopamine terminals.15 So far in our ongoing series, all patients with Parkinson’s disease and orthostatic hypotension have had evidence of diffuse loss of cardiac sympathetic innervation at the time of initial testing.13 About half of patients with Parkinson’s disease without orthostatic hypotension have also had evidence of diffuse cardiac sympathetic denervation, and because of the likelihood of a floor effect, data from these patients were not included in this study. Given these results, and the present findings, based on the remaining patients with Parkinson’s disease who did not have either orthostatic hypotension or diffuse cardiac sympathetic denervation at the time of initial testing, indicate that progressive loss of cardiac sympathetic innervation characterizes the disease. As of the time of the second scan, none of the patients developed orthostatic hypotension. The sympathetic innervation of the myocardium travels with the coronary arteries. The finding of more severely decreased 6-[18F]fluorodopamine-derived radioactivity in the lateral wall than in the interventricular septal wall leads to a suggestion of a dying-back mechanism for the loss of sympathetic terminals, as opposed to death of the cell bodies followed by loss of the terminals (as in Wallerian degeneration). In agreement with this notion, about one half of patients with Parkinson’s disease who do not have orthostatic hypotension already have decreased concentrations of 6-[18F]fluorodopamine-derived radioactivity throughout Table. Tissue Concentrations of 6-[18F]FluorodopamineDerived Radioactivity (nCi-kg/cc-mCi) in Patients with Parkinson’s Disease First Scan Organ Mean Left lateral wall Septum Right myocardium Right chamber Left chamber Liver Spleen Renal cortex Renal pelvis Nasophanrynx Parotid Submandibular gland Thyroid 5,991 ⫾ 537 7,278 ⫾ 385 5,889 ⫾ 564 4,799 ⫾ 411 4,971 ⫾ 517 6,897 ⫾ 724 6,495 ⫾ 364 23,564 ⫾ 2,412 27,667 ⫾ 3,874 1,431 ⫾ 162 1,832 ⫾ 170 1,870 ⫾ 239 1,933 ⫾ 187 a Significantly different from first scan. SEM ⫽ standard error of the mean SEM Second Scan Mean SEM 4,107 ⫾ 535a 6,137 ⫾ 716a 5,390 ⫾ 485 4,240 ⫾ 263 4,725 ⫾ 423 7,165 ⫾ 869 5,020 ⫾ 137a 20,716 ⫾ 2,070 23,278 ⫾ 5,203 1,501 ⫾ 163 1,698 ⫾ 221 1,782 ⫾ 267 1,624 ⫾ 250 the left ventricular myocardium, and of the remaining half, most have decreased 6-[18F]fluorodopaminederived radioactivity in the lateral ventricular wall or apex, with relative sparing of the interventricular septum.13 Because the patients did not have a progressive loss of 6-[18F]fluorodopamine-derived radioactivity in the liver or renal cortex but did in the heart and spleen, rates of loss of sympathetic innervation appear to vary across organs. This fits with the notion of cardioselective sympathetic denervation in Parkinson’s disease.11,12 The results lead to the general inference that Parkinson’s disease features progressive neurodegeneration not only in the nigrostriatal dopaminergic system but also in the sympathetic noradrenergic system. 10. Satoh A, Serita T, Seto M, et al. Loss of 123I-MIBG uptake by the heart in Parkinson’s disease: assessment of cardiac sympathetic denervation and diagnostic value. J Nucl Med 1999;40: 371–375. 11. Reinhardt MJ, Jungling FD, Krause TM, Braune S. Scintigraphic differentiation between two forms of primary dysautonomia early after onset of autonomic dysfunction: value of cardiac and pulmonary iodine-123 MIBG uptake. Eur J Nucl Med 2000;27:595– 600. 12. Taki J, Nakajima K, Hwang EH, et al. Peripheral sympathetic dysfunction in patients with Parkinson’s disease without autonomic failure is heart selective and disease specific. Eur J Nucl Med 2000;27:566 –573. 13. Goldstein DS, Holmes C, Li ST, et al. Cardiac sympathetic denervation in Parkinson disease. Ann Intern Med 2000;133: 338 –347. 14. Goldstein DS, Eisenhofer G, Dunn BB, et al. Positron emission tomographic imaging of cardiac sympathetic innervation using 6-[18F]fluorodopamine: initial findings in humans. J Am Coll Cardiol 1993;22:1961–1971. 15. Poewe WH. The natural history of Parkinson’s disease. Ann Neurol 1998;44(suppl 1):1–9. We gratefully acknowledge the assistance of Sandra Brentzel, RN, and the Positron Emission Tomography Department of the National Institutes of Health. References 1. Braune S, Reinhardt M, Bathmann J, et al. Impaired cardiac uptake of meta-[123I]iodobenzylguanidine in Parkinson’s disease with autonomic failure. Acta Neurol Scand 1998;97: 307–314. 2. Braune S, Reinhardt M, Schnitzer R, et al. Cardiac uptake of [123I]MIBG separates Parkinson’s disease from multiple system atrophy. Neurology 1999;53:1020 –1025. 3. Takatsu H, Nishida H, Matsuo H, et al. Cardiac sympathetic denervation from the early stage of Parkinson’s disease: clinical and experimental studies with radiolabeled MIBG. J Nucl Med 2000;41:71–77. 4. Takatsu H, Nagashima K, Murase M, et al. Differentiating Parkinson disease from multiple-system atrophy by measuring cardiac iodine-123 metaiodobenzylguanidine accumulation. JAMA 2000;284:44 – 45. 5. Orimo S, Ozawa E, Nakade S, et al. (123)I-metaiodobenzylguanidine myocardial scintigraphy in Parkinson’s disease. J Neurol Neurosurg Psychiatry 1999;67:189 –194. 6. Yoshita M, Hayashi M, Hirai S. Iodine 123-labeled metaiodobenzylguanidine myocardial scintigraphy in the cases of idiopathic Parkinson’s disease, multiple system atrophy, and progressive supranuclear palsy. Rinsho Shinkeigaku 1997;37: 476 – 482. 7. Yoshita M. Differentiation of idiopathic Parkinson’s disease from striatonigral degeneration and progressive supranuclear palsy using iodine-123 meta-iodobenzylguanidine myocardial scintigraphy. J Neurol Sci 1998;155:60 – 67. 8. Druschky A, Hilz MJ, Platsch G et al. Differentiation of Parkinson’s disease and multiple system atrophy in early disease stages by means of I-123-MIBG-SPECT. J Neurol Sci 2000; 175:3–12. 9. Satoh A, Serita T, Tsujihata M. Total defect of metaiodobenzylguanidine (MIBG) imaging on heart in Parkinson’s disease: assessment of cardiac sympathetic denervation. Nippon Rinsho 1997;55:202–206. Li et al: Sympathetic Innervation in Parkinson’s Disease 223 No Mutation in the TRKA (NTRK1) Gene Encoding a Receptor Tyrosine Kinase for Nerve Growth Factor in a Patient with Hereditary Sensory and Autonomic Neuropathy Type V Ennio Toscano, MD, PhD,1 Alessandro Simonati, MD,2 Yasuhiro Indo, MD, PhD,3 and Generoso Andria, MD1 Hereditary sensory and autonomic neuropathy type IV (HSAN-IV) and type V (HSAN-V) are autosomal recessive genetic disorders, both characterized by a lack of pain sensation. We report a girl with clinical and neurophysiological findings consistent with a diagnosis of HSAN-V. We sequenced her TRKA gene, encoding a receptor tyrosine kinase for nerve growth factor and responsible for HSAN-IV, but we could not detect any mutation. These data indicate that a gene (or genes) other than TRKA is probably responsible for HSAN-V in some patients. Ann Neurol 2002;52:224 –227 Hereditary sensory and autonomic neuropathy (HSAN) are classified into 5 different types according to Dyck.1 We have demonstrated that the TRKA (NTRK1) gene, encoding a receptor tyrosine kinase that is phosphorylated in response to nerve growth factor, is responsible for HSAN-IV.2 HSAN-IV is characterized by febrile episodes, anhidrosis, insensitivity to pain, self-mutilating behavior, and mental retardation. Pathological features of HSAN-IV are severe reduction of small-diameter afferent neurons, which are activated by tissue-damaging stimuli,3,4 and a loss of sympathetic neurons innervating eccrine sweat glands.5 HSAN-V also is characterized by absent reaction to From the 1Department of Pediatrics, “Federico II” University, Naples, Italy; 2Section of Clinical Neurology, Department of Neurological and Visual Sciences, University of Verona, Verona, Italy; and 3 Department of Pediatrics, Kumamoto University, Honjo, Kumamoto, Japan. Received Aug 21, 2001, and in revised form March 11, 2002. Accepted for publication March 11, 2002. Published online Jun 23, 2002, in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ana.10245 Address correspondence to Dr Andria, Department of Pediatrics, “Federico II” University, Via S. Pansini 5, I-80131 Naples, Italy. E-mail: firstname.lastname@example.org 224 © 2002 Wiley-Liss, Inc. noxious stimuli but usually lacks anhidrosis and mental retardation. Some patients with possible HSAN-V were described before 1960 (see Landrieu and colleagues6 for references), and 3 families with HSAN-V subsequently were described in which were demonstrated a selective severe decrease of the small myelinated fibers7,8 and a small reduction in unmyelinated fibers9 of the sural nerve. HSAN-V is likely an autosomal recessive disorder, but Landrieu and colleagues6 reported in a family 2 dominantly transmitted cases with normal nerve biopsy. Therefore, a clinical entity, also called congenital indifference to pain, is genetically heterogeneous and probably includes HSAN-IV or HSAN-V, as suggested by Dyck.1 Recently, Houlden and colleagues10 described a boy with febrile episodes and a lack of pain sensation that they diagnosed as HSAN-V, according to the typical findings in a nerve biopsy.9 DNA analysis showed that he had a homozygous mutation in the TRKA gene. They stated that HSAN-IV and HSAN-V are likely to be allelic. Here we present a girl with clinical features consistent with HSAN-V, namely, a lack of pain sensation, but no febrile episodes and normal sural nerve biopsy. We sequenced her TRKA gene and detected no mutation. Therefore, these data indicate that a gene (or genes) other than TRKA is probably responsible for HSAN-V in some patients, arguing against a previously cited report.10 We also discuss the differential diagnosis of HSAN-IV versus HSAN-V. Case Report I.F., a girl, was the second child of consanguineous parents (first cousins). She was born at term (birth weight, 3,400gm) with Apgar scores of 9 and 10. She was found to have insensitivity to painful stimuli, and multiple traumatic episodes were experienced from early life. She bit off the tip of her tongue, and she did not cry when she fell or during blood sampling. At the ages of 3, 4, and 5 years, she had episodes of left hip dislocation without pain perception. At the age of 6 years, she suffered from bilateral osteochondritis of the feet. Her psychomotor development was apparently normal, and she never exhibited behavioral problems or attention deficit problems. She never had thermoregulatory or feeding problems, vomiting, bowel dysfunction, or other signs or symptoms of gastrointestinal dysmotility. Results Clinical examination of the girl when she was 11 years old showed reduced response to painful stimuli and anosmia; thermic sensation appeared to be conserved. Fungiform papillae of the tongue, corneal reflexes, and overflow tears were present. Intelligence and deep tendon reflexes were normal. Blood pressure was normal in both supine and standing positions. An intradermal injection of histamine evoked axonal reflex as observed in a normal control. She could control her body tem- perature well under hot environmental conditions. Pilocarpine ionophoresis and sympathetic skin response were normal. A right sural nerve biopsy was performed and processed according to routine procedures for both morphological and morphometrical investigations.11 Features of the nerve fascicle were normal. At the ultrastructural level, unmyelinated fibers were regularly detected; neither denervated Schwann cells nor collagen pockets were seen (Fig 1). Myelinated fiber density was within normal ranges; the shape of the histogram was bimodal, showing the small-caliber fiber peak (Fig 2). Skin biopsy was examined on plastic sections only. Glands were present; both myelinated and unmyelinated fibers were present in the intradermal nerve fascicles. We sequenced all 17 exons of the TRKA gene of the patient, including their flanking intronic sequences,12 and detected no putative mutation. Furthermore, we analyzed 8 polymorphic sites in this gene and found that the patient is heterozygous at the intragenic polymorphic site (c. 1,953 cytosine/thymine).13 This furFig 1. Sural nerve biopsy. (A) Semithin section; toluidine blue stain. Normal features of the nerve fascicle showing myelinated fibers of both small and large calibers. Bar ⫽ 12.5m. (B) Thin section; uranyl acetate and lead citrate stain. Representative clusters of normal unmyelinated fibers. Bar ⫽ 3m. Fig 2. Size distribution (X-axis) of the myelinated fibers of the index case and an age-matched control. Note the normal, bimodal pattern of the histogram; small-caliber fibers are similarly represented in both the patient and the control. Numbers on the Y-axis refer to the absolute figures of the measured fibers. ther supports that the patient has no mutation in the TRKA locus because the parents are consanguineous. Discussion The patient reported here probably presents a hereditary autosomal recessive sensory neuropathy with selective loss of pain sensation. Autonomic abnormalities apparently were not observed. Reduced sweating was mentioned once, but she can maintain her body temperature under hot environmental conditions. This is compatible with the findings in her nerve and skin biopsies. These clinical and laboratory data suggest that the patient suffers from HSAN-V. Similar features, together with apparently intact peripheral nerves, were described in some classic reports of congenital indifference to pain.6 The patient reported by Low and colleagues,7 presenting indifference to pain and selective loss of small myelinated fibers without sweating abnormality, was classified as HSAN-V. Dyck and colleagues9 described a patient with indifference to pain, sweating abnormality, and a severe decrease in smalldiameter myelinated fibers with a mild reduction of unmyelinated fibers. It is their view that most earlier cases of indifference to pain may have had either HSAN-IV or HSAN-V.1,9 In contrast, HSAN-IV is a distinct clinical entity characterized by recurrent episodic fever, anhidrosis, insensitivity to pain, self-mutilating behavior, and mental retardation.3,14 Mental retardation is variable, Toscano et al: No Mutation in TRKA Gene in HSAN-V 225 from severe to mild, and some patients were apparently normal, but later mild retardation was showed by a formal assessment (patients KI-108 and KI-116 reported by Mardy and colleagues12 and Indo and colleagues,15 respectively). Sweating may be variable but should be evaluated cautiously, as we recently reported.15 We think that a fundamental phenotype of HSAN-IV consists of insensitivity to pain, anhidrosis, and mental retardation, each of variable degree. Recurrent hyperpyrexia, self-mutilating behaviors, traumas, and bone fractures can be devastating and often lead to crippling or fatal consequences. Nerve growth factor supports the survival of sympathetic ganglion neurons and nociceptive sensory neurons in dorsal root ganglia and ascending cholinergic neurons of the basal forebrain.16 Eccrine sweat glands, innervated by sympathetic cholinergic fibers, are well developed in humans. Therefore, the nerve growth factor-TRKA system has a crucial role in the development and function of the nociceptive reception and establishment of thermoregulation via sweating.2 A negative result of the intradermal histamine test, an important diagnostic criterion in HSAN-IV, probably also can be explained by defective peripheral sympathetic neurons. Therefore, anhidrosis and associated failure to maintain body temperature are characteristic features of HSAN-IV. Recently, Houlden and colleagues10 described a boy 9 years of age, born to healthy consanguineous parents, presenting anhidrosis and loss of pain and temperature sensation. Sural nerve biopsy demonstrated severe reduction in small-caliber myelinated fiber density but only modest reduction in unmyelinated axons. The authors did not mention either a skin biopsy or a histamine test, which provide important diagnostic criteria for the group of hereditary peripheral neuropathies.17 They detected a homozygous missense mutation in the TRKA gene, changing a tyrosine to a cysteine at codon 359. According to this case, they concluded that the 2 disorders, HSAN-IV and HSAN-V, are likely to be allelic. However, we propose an alternative interpretation of Houlden and colleagues’ report. Their patient may suffer from HSAN-IV, but not HSAN-V. They argued for the diagnosis of HSAN-V, mainly basing their argument on the finding of nerve biopsy with modest reduction of unmyelinated fibers. However, anhidrosis observed in the patient strongly indicates dysfunction of the most distal intradermal portions of the sympathetic axons. That also might account for the normal appearance of the more proximal unmyelinated fibers observed after sural nerve. Furthermore, it would be important to confirm the functional significance of the missense mutation by an expression study, such as that reported recently.18 We stress the importance of the molecular analysis of the TRKA gene in all cases with lack of pain sensa- 226 Annals of Neurology Vol 52 No 2 August 2002 tion to distinguish overlapping phenotypes of HSAN-IV and HSAN-V, both characterized by a lack of pain sensation. Most patients with HSAN-IV probably have mutations in the TRKA gene,2,12,13 although we cannot rule out the possibility that a mutation (or mutations) in another gene (or genes) is responsible for similar clinical phenotypes. The patient presented here suffers from HSAN-V, but not HSAN-IV, because she does not show anhidrosis or mental retardation. The presence of myelinated and unmyelinated fibers in nerve biopsy and the normal intradermal histamine test further support this diagnosis. We could not detect any putative mutation in the TRKA gene, whereas we found that the patient is heterozygous for this locus, despite the parental consanguinity. These findings strongly indicate that defects of a gene (or genes) other than TRKA is likely responsible for at least some patients with HSAN-V. It remains unknown whether peripheral nociceptive transmission or central processing might be involved in our case and in some cases with similar phenotypes reported as having HSAN-V. Anosmia observed in our case remains to be examined and may give us some clue for studying a mechanism underlying the defect in pain sensation. Some patients diagnosed as having HSAN-V do not show an apparent abnormality of peripheral nerve fibers. Therefore, their manifestations might be caused by defects in peripheral nociceptor or transduction and transmission of pain sensation or even central processing, such as a lack of concern for a painful stimulus well received by the peripheral nervous system, according to the classic definition of indifference to pain.19 References 1. Dyck PJ. Neuronal atrophy and degeneration predominantly affecting peripheral sensory and autonomic neurons. In: Dick PJ, Thomas PK, Griffin JW, Low PA, Podreslo JC, eds. Peripheral neuropathy. Philadelphia: Saunders, 1993:1065–1093. 2. Indo Y, Tsuruta M, Hayashida Y, et al. Mutations in the TRKA/NGF receptor gene in patients with congenital insensitivity to pain with anhidrosis. Nat Genet 1996;13:485– 488. 3. Swanson AG, Buchan GG, Alvord EC Jr, et al. Autonomic changes in congenital insensitivity to pain: absence of small primary sensory neurons in ganglia, roots and Lissauer’s tract. Arch Neurol 1965;12:12–18. 4. Rafel E, Alberca R, Bautista J, et al. Congenital insensitivity to pain with anhidrosis. Muscle Nerve 1980;3:216 –220. 5. Langer J, Goebel HH, Veit S. Eccrine sweat gland are not innervated in hereditary sensory neuropathy type IV: an electronmicroscopic study. Acta Neuropathol (Berl) 1981;54:199 –202. 6. Landrieu P, Said G, Allaire C. Dominantly transmitted congenital indifference to pain. Ann Neurol 1990;27:574 –578. 7. Low PA, Burke WJ, McLeod JG. Congenital sensory neuropathy with selective loss of small myelinated fibers. Ann Neurol 1978;3:179 –182. 8. Donaghy M, Hakin RN, Bamford JM, et al. Hereditary sensory and autonomic neuropathy with neurotrophic keratitis. Brain 1987;110:563–583. 9. Dyck PJ, Mellinger JF, Reagan TJ, et al. Not “indifference to pain” but varieties of hereditary sensory and autonomic neuropathy. Brain 1983;106:373–390. 10. Houlden H, King RHM, Hashemi-Nejad A, et al. A novel TRK A (NTRK1) mutation associated with hereditary sensory and autonomic neuropathy type V. Ann Neurol 2001;49: 521–525. 11. Simonati A, Fabrizi GM, Pasquinelli A, et al. Congenital hypomyelination neuropathy with Ser72Leu substitution in PMP22. Neuromuscul Disord 1999;9:257–261. 12. Mardy S, Miura Y, Endo F, et al. Congenital insensitivity to pain with anhidrosis: novel mutations in the TRKA (NTRK1) gene encoding a high-affinity receptor for nerve growth factor. Am J Hum Genet 1999;64:1570 –1579. 13. Miura Y, Mardy S, Awaya Y, et al. Mutation and polymorphism analysis of the TRKA (NTRK1) gene encoding a highaffinity receptor for nerve growth factor in congenital insensitivity to pain with anhidrosis (CIPA) families. Hum Genet 2000;106:116 –124. 14. Swanson AG. Congenital insensitivity to pain with anhidrosis. Arch Neurol 1963;8:299 –306. 15. Indo Y, Mardy S, Miura Y, et al. Congenital insensitivity to pain with anhidrosis (CIPA): novel mutations of TRKA (NTRK1) gene encoding the receptor tyrosine kinase for nerve growth factor, a putative uniparental disomy and a linkage of the mutant TRKA and PKLR genes in a family with CIPA and pyruvate kinase deficiency. Hum Mutat 2001;18:308 –318. 16. Levi Montalcini R. The nerve growth factor: thirty-five years later. EMBO J 1987;6:1145–1154. 17. Axelrod FB, Pearson J. Congenital sensory neuropathies. Am J Dis Child 1984;138:947–954. 18. Mardy S, Miura Y, Endo F, et al. Congenital insensitivity to pain with anhidrosis (CIPA): effect of TRKA (NTRK1) missense mutations on autophosphorylation of the receptor tyrosine kinase for nerve growth factor. Hum Mol Genet 2001; 10:179 –188. 19. Jewesburry ECO. Congenital indifference to pain. In: Vinken PJ, Bruyn GW, eds. Handbook of clinical neurology. Vol 8. Amsterdam: Elsevier, 1979:187–204. X-Linked Creatine Deficiency Syndrome: A Novel Mutation in Creatine Transporter Gene SLC6A8 Alberto Bizzi, MD,1 Marianna Bugiani, MD,2 Gajja S. Salomons, PhD,3 Donald H. Hunneman, PhD,4 Isabella Moroni, MD,2 Margherita Estienne, MD,2 Ugo Danesi, PhD,1 Cornelis Jakobs, PhD,3 and Graziella Uziel, MD2 Among creatine deficiency syndromes, an X-linked condition related to a defective creatine transport into the central nervous system has been described recently. Hallmarks of the disease are the absence of a creatine signal at brain spectroscopy, increased creatine levels in blood and urine, ineffectiveness of oral supplementation, and a mutation in the SLC6A8 (Online Mendelian Inheritance in Man [OMIM] 300036) creatine transporter gene. We report on a patient in whom a novel mutation (12211223delTTC) was identified. Ann Neurol 2002;52:227–231 Creatine deficiency syndromes are recently identified inborn errors of metabolism resulting in a progressive encephalopathy with early onset and mental retardation, extrapyramidal features, and drug-resistant epilepsy.1–3 Symptoms are related to a depletion of the creatine/phosphocreatine pool within the central nervous system, making this condition easily detectable by brain spectroscopy. Most of the cases reported so far were caused by a defect of the second enzyme involved in creatine biosynthesis: guanidino-acetate methyltransferase (GAMT; OMIM 601240). This defect results in increased guanidino-acetate (GAA) and reduced creatine levels in blood and urine.4,5 GAMT-deficient patients benefit from oral creatine monohydrate supplementation, which helps to control movement disorders and epilepsy, partially recovers mental impairment, and restores neurological development over time.6 The ab- From the 1Departments of Neuroradiology and 2Child Neurology, Istituto Nazionale Neurologico “C. Besta,” Milano, Italy; 3Metabolic Unit, VU Medical Center, Amsterdam, The Netherlands; and 4 University Kinderklinik, Gottingen, Germany. Received Nov 6, 2001, and in revised form Mar 11, 2002. Accepted for publication Mar 11, 2002. Published online Jun 21, 2002, in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ana.10246 Address correspondence to Dr Uziel, Department of Child Neurology, Istituto Nazionale Neurologico “C. Besta,” Via Celoria 11, 20133 Milano, Italy. E-mail: email@example.com © 2002 Wiley-Liss, Inc. 227 sence of complete recovery can be explained by neurotoxic GAA accumulation or by an imbalance of brain creatine and high-energy phosphates.7,8 Very recently, Item and colleagues9,10 described 2 sisters with mental impairment, low GAA levels in urine, and undetectable activity of arginine to glycine amidinotransferase (AGAT; OMIM 602360), the first enzyme involved in creatine biosynthesis. A homozygous nonsense mutation in AGAT gene was detected. Clinical symptoms and brain creatine deficiency were partially recovered by means of creatine supplementation. A different disease due to impairment of creatine transport into the brain was reported by Salomons and colleagues11,12 in a boy with language delay, short attention span, epilepsy, and increased creatine levels in blood and urine. A nonsense mutation in the X-linked creatine transporter gene SLC6A8 was demonstrated. Creatine supplementation in this boy was totally ineffective. This article reports on a second case caused by defective creatine transport into the central nervous system. Case Report The patient was the second-born son of healthy, nonconsanguineous parents. The child’s prenatal and perinatal history was unremarkable. Since the first months of life, he presented with motor delay, reduced interest in surroundings, and no language acquisition. At 8 months, he was admitted to the hospital after a febrile seizure. An electroencephalogram recording and a magnetic resonance imaging scan were normal. Routine blood and urine analysis and investigation for infectious (screening for rubella, cytomegalovirus, toxoplasma, herpes simplex, Coxsackie virus, and Mycoplasma pneumoniae) and neurometabolic disorders (plasma lactate and metabolic screening of a 24-hour urine sample with an assessment of amino acids, organic acids, and mucopolysaccharides) were normal. Since 16 months of age, he experienced complex partial seizures with secondary generalization, responding to sodium valproate. Serial electroencephalogram tracings showed a progressive instability of background activity with abundant fast activity and epileptic discharges from frontal and temporal leads during sleep. Physical examination at 3 years and 9 months showed a severe delay in speech and language functions with behavioral disturbances in agreement with an autistic disorder. He did not follow commands or speak, he presented with stereotypical motor behaviors, and he could not engage in any structured play. Gross and fine motor functions were normal. A second magnetic resonance imaging showed mild atrophy with signal abnormality in the right hippocampus, suggesting mesial temporal sclerosis. Proton magnetic resonance spectroscopic imaging (H-MRSI) was performed as part of the diagnostic workup for mesial temporal sclerosis conducted at our institution and showed a normal N-acetylaspartate/choline ratio in both hippocampi without abnormality in the asymmetry index. The surprising feature was the absence of the creatine peak in the whole brain (Fig), indi- Fig. Proton magnetic resonance spectroscopic imaging shows a complete absence of the creatine peak in both white and gray matter (1A, 2A, 3A, 4A). The position of selected voxels is indicated on the T2-weighted magnetic resonance image at the level of the centrum semiovale. There was no restoration of the creatine pool after 3 months of oral creatine monohydrate supplementation at 400mg/kg/day (1B) and after 8 months of supplementation at 700mg/kg/day (1C). There were no significant changes in choline and N-acetylaspartate levels. A normal spectrum from an age-matched control is given for comparison: the peaks of choline (Cho; 3.2ppm), creatine (Cr; 3.02ppm), and N-acetylaspartate (NAA; 2.02ppm) are indicated. 228 Annals of Neurology Vol 52 No 2 August 2002 cating a creatine deficiency syndrome, without any abnormal peak in the expected frequency of GAA (3.8ppm). Creatine and GAA, therefore, were measured in plasma and urine, and oral creatine monohydrate supplementation (400mg/kg/ day) was started. Because H-MRSI showed no creatine recovery after 3 months of therapy, supplementation was increased to 700mg/kg/day. Eight months later, a follow-up H-MRSI examination confirmed no appearance of brain creatine. By this time, clinical symptoms were not improved, and therapy was discontinued. Currently, the patient is 5 years old, and his clinical symptoms are grossly unchanged. The very recent identification of the first patient with a creatine transporter defect11,12 suggested that our patient suffered from a creatine uptake defect as well, which was in line with the ineffectiveness of creatine supplementation and with biochemical data. Mutational analysis of the SLC6A8 gene, therefore, was performed. Methods Before therapy, H-MRSI (two-dimensional phase encoding point resolved spectroscopy repetition (PRESS) technique: recovery time/echo time, 1,500/136msec; field of view, 160mm; matrix, 16 ⫻ 16; 20mm slice thickness) was performed from 3 separate sections at the level of the centrum semiovale, basal nuclei, and hippocampi. During therapy, H-MRSI was performed at the level of the centrum semiovale. Single-voxel spectra (PRESS: recovery time/echo delay time, 1,500/34msec) were acquired from a 30ml volume in the frontoparietal parasagittal cortical gray matter before and during therapy. Raw data were transferred to a SUN workstation and reconstructed with custom-made software.13 Mutational analysis complementary DNA–based sequence analysis was performed according to methods described elsewhere.12 DNA was isolated from blood cells with a QIAamp blood kit (Qiagen, Chatsworth, CA) for conformation of the mutation at the genomic level. Primers specific for exon 8 of the SLC6A8 gene were designed: forward 5⬘TCCCAGCCCCTGCCGCAC and reverse 5⬘TACAAACTGTGGCCAGGGC. Results Creatine and GAA levels before, during, and after the withdrawal of creatine supplementation are shown in Table 1. Sequence analysis of the creatine transporter gene SLC6A8 identified an hemizygous 3bp deletion in exon 8 involving nucleotides TTC in position 12211223 (1221-1223delTTC; GenBank accession number, NM 005629). This mutation resulted in the deletion of a single phenylalanine at residue 408 of the protein (delF408). The patient’s mother was heterozygous for the mutation. Discussion The failure of creatine supplementation to restore brain creatine by H-MRSI and to improve clinical symptoms and normal plasma and urine GAA levels ruled out a defect of creatine biosynthesis and prompted a search for molecular defects in the SLC6A8 gene.12,14 A novel hemizygous deletion located in a short repeat of 3 phenylalanines in exon 8 was detected. The repeat is part of transmembrane domain VIII, which is a very conserved region among the Na⫹- and Cl⫺-dependent neurotransmitter family.15 The mutation, resulting in a phenylalanine deletion at position 408, most likely causes a partial or even complete loss of creatine transport function. The activity of the transporter could not be tested in fibroblasts because the parents refused consent to a skin biopsy. To our knowledge, this is the first creatine deficiency case studied with the multivoxel H-MRSI technique, which has better spatial resolution and allows an evaluation of lesion heterogeneity in the shortest amount of time. We did not find any significant difference in creatine levels across the brain regions examined, suggesting that the transporter defect does not spare any brain area. Symptoms of creatine deficiency have been related previously to a defect in creatine transport by Salomons and colleagues,11,12 who described a male patient with mental retardation and severe delay in expressive speech and language function presenting a hemizygous nonsense mutation in the SLC6A8 gene. As in our case, a lack of creatine transport into the central nervous system resulted in the failure of creatine supplementation to reverse clinical symptoms and brain spectroscopy abnormality. Very recently, another unrelated family with a creatine transporter defect was recognized by the same authors.16 The biochemical profile of our patient showed a massive loss of creatine in urine, but in contrast with the first reported index case, creatine levels in plasma were normal even during creatine supplementation. Table 1. Laboratory Results before, during, and after Withdrawal of Oral Creatine Monohydrate Supplementation (700mg/kg/day) Blood Result Normal values Prior During After Urine GAA (mol/L) Creatine (mol/L) GAA (mmol/mol creatinine) Creatine (mmol/mol creatinine) 0.4–3 1.8 ND ND 10–200 83 174 66 10–125 79 28 46 40–360 4,519 21,315 3,102 GAA ⫽ guanidino-acetate; ND ⫽ not determined. Bizzi et al: Creatine Transporter Defect 229 Table 2. Creatine and GAA Levels in GAMT and AGAT Defects and in Creatine Transporter Gene (SLC6A8) Mutations Blood Defect or mutation GAMT AGAT SLC6A8 Urine Brain Creatine GAA Creatine GAA Creatine GAA Decrease Normal Normal/Increase Increase Normal Normal Decrease ND Increase Increase Decrease Normal Decrease Decrease Decrease Undetectable/Increase Undetectable Undetectable AGAT ⫽ arginine to glycine amidinotransferase; GAA ⫽ guanidino-acetate; GAMT ⫽ guanidino-acetate methyltransferase; ND ⫽ not determined. The mother underwent a brain H-MRSI that demonstrated a mildly reduced creatine signal compared with that of age-matched controls. Moreover, the results showed that she was heterozygous for the mutation identified in her son. However, no learning disabilities were reported for the patient’s female relatives, in contrast with what was encountered in 2 of 3 female carriers in the first family described.12 These data agree with skewed X inactivation (mosaic expression of mutant and wild-type alleles), resulting in a variably favorable mosaic expression of the wild-type allele. The 3 creatine deficiency syndromes described so far (ie, GAMT and AGAT defects and impairment of creatine transport) share overlapping symptoms of mental retardation with severe language impairment, autistic behavior, and epilepsy. Movement disorders have been reported only in patients with GAMT deficiency, suggesting that extrapyramidal features may result from neurotoxic GAA accumulation rather than from reduced creatine availability in brain. The major involvement of higher cortical functions and the frequent finding of epilepsy suggest that the cerebral cortex may be selectively vulnerable to creatine deficiency. How creatine deficiency adversely affects cortical functions is still to be established. It is conceivable that creatine plays a role in the latest stages of cortical organization, including synaptogenesis, a process that continues after birth. This could explain the homogeneity of clinical presentation and age at onset in the 3 creatine deficiency syndromes, even though in children with biosynthesis defects, creatine is supplied through the placenta during fetal life, whereas patients with a transporter defect suffer from creatine depletion already in utero. Spectroscopy alone cannot always distinguish between synthesis and transport defects. In GAMT deficiency, brain spectroscopy at a short echo time can identify an abnormal peak that is assigned to GAA (3.8ppm), but this elevation may be subtle, and it has been reported only in a few cases. Therefore, all patients in whom a diagnosis of creatine deficiency is reached should undergo a careful biochemical evaluation to assess creatine and GAA levels in blood and urine (Table 2). Brain spectroscopy is becoming more available; a quick automated spectrum acquisition can be performed at the 230 Annals of Neurology Vol 52 No 2 August 2002 time of conventional magnetic resonance imaging and may practically disclose creatine or other metabolites depletion: just recently, the first case of N-acetylaspartate brain deficiency was reported in a patient with mental retardation and severe language impairment.17 Clinical features of mental impairment, language delay, and autistic behavior with epilepsy frequently are encountered in infancy. Conceivably, creatine deficiency could be underdiagnosed, and if brain spectroscopy cannot be easily achieved, an assessment of blood and urinary creatine should always be performed. We thank Dr S. J. M. van Dooren and D. Brunea for excellent technical support to Dr G. S. Salomons. Electronic Database Information: OMIM. www.ncbi.nlm.nih.gov/ omim; GenBank. www.ncbi.nlm.nih.gov/genbank. References 1. Stockler S, Isbrandt D, Hanefeld F, et al. Guanidinoacetate methyltransferase deficiency: the first inborn error of creatine metabolism in man. Am J Hum Genet 1996;58:914 –922. 2. Stockler S, Marescau B, De Deyn PP, et al. Guanidinoacetate methyltransferase deficiency, a new inborn error of creatine synthesis. Metabolism 1997;46:1189 –1193. 3. Ganesan V, Johnson A, Connelly A, et al. Guanidinoacetate methyltransferase deficiency: new clinical features. Pediatr Neurol 1997;17:155–157. 4. Schulze A, Hess T, Wevers R, et al. Creatine deficiency syndrome caused by guanidinoacetate methyltransferase deficiency: diagnostic tools for a new inborn error of metabolism. J Pediatr 1997;131:626 – 631. 5. Ilas J, Muhl A, Stockler-Ipsiroglu S. Guanidinoacetate methyltransferase deficiency: non-invasive enzymatic diagnosis of a newly recognized inborn error of metabolism. Clin Chim Acta 2000;290:179 –188. 6. Stockler S, Hanefeld F, Frahm J. Creatine replacement therapy in guanidinoacetate methyltransferase deficiency, a novel inborn error of metabolism. Lancet 1996;348:789 –790. 7. van der Knaap MS, Verhoeven NM, Maaswinkel-Mooij P, et al. Mental retardation and behavioural problems as presenting signs of a creatine synthesis defect. Ann Neurol 2000;47:540 –543. 8. Leuzzi V, Bianchi MC, Tosetti M, et al. Brain creatine depletion: guanidinoacetate methyltransferase deficiency (improving with creatine supplementation). Neurology 2000;55: 1407–1409. 9. Bianchi MC, Tosetti M, Fornai F, et al. Reversible brain creatine deficiency in two sisters with normal blood creatine level. Ann Neurol 2000;47:511–513. 10. Item CB, Stockler-Ipsiroglu S, Stromberger C, et al. Arginine: glycine amidinotransferase deficiency: the third inborn error of creatine metabolism in humans. Am J Hum Genet 2001;69: 1127–1133. 11. Cecil KM, Salomons GS, Ball WS, et al. Irreversible brain creatine deficiency with elevated serum and urine creatine: a creatine transporter defect? Ann Neurol 2001;49:401– 404. 12. Salomons GS, van Dooren SJM, Verhoeven NM, et al. X-linked creatine-transporter gene (SLC6A8) defect: a new creatine-deficiency syndrome. Am J Hum Genet 2001;68: 1497–1500. 13. Soher BJ, van Zijl PCM, Duyn JH, et al. Quantitative proton spectroscopic imaging of the human brain. Magn Reson Med 1996;35:356 –363. 14. Gregor P, Nash SR, Caron MG, et al. Assignment of the creatine transporter gene (SLC6A8) to human chromosome Xq28 telomeric to G6PD. Genomics 1995;25:332–333. 15. Nash SR, Giros B, Kingsmore SF, et al. Cloning, pharmacological characterization and genomic localization of the human creatine transporter. Receptors Channels 1994;2:165–174. 16. Salomons GS, Dooren SJ, Verhoeven NM, et al. X-linked creatine transporter defect: the second family. J Inherit Metab Dis 2001;24(suppl 1):119. 17. Martin E, Capone A, Schneider J, et al. Absence of N-acetylaspartate in the human brain: impact on neurospectroscopy? Ann Neurol 2001;49:518 –521. Vertebrobasilar Stroke as a Late Complication of a Blalock-Taussig Shunt Philippe Gailloud, MD,1 Argye Hillis, MD,2 Bruce Perler, MD,3 and Kieran J. Murphy, MD1 We describe a 39-year-old patient with a cerebellar infarct and a history of a tetralogy of Fallot corrected during childhood. This is the first documented case of vertebrobasilar stroke occurring as a late complication of a Blalock-Taussig shunt followed by total cardiac repair. Ann Neurol 2002;52:231–234 From the Department of 1Radiology and Radiological Sciences, 2 Neurology, and 3Vascular Surgery, Johns Hopkins Hospital, Baltimore, MD. Received May 15, 2000, and in revised form Feb 28, 2002. Accepted for publication Mar 12, 2002. Published online Jun 21, 2002, in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ana.10249 Address correspondence to Dr Gailloud, Department of Radiology, Johns Hopkins University School of Medicine, 600 N. Wolfe Street, Radiology B-100, Baltimore, MD 21287. E-mail: firstname.lastname@example.org The Blalock-Taussig shunt, first performed in 1944 at the Johns Hopkins Hospital,1 is a surgical anastomosis established between a subclavian artery and the ipsilateral pulmonary artery. In patients with a tetralogy of Fallot, this procedure allows deferment of the definitive cardiac repair to a more robust stage of life by bypassing the pulmonary artery stenosis. The Blalock-Taussig shunt, however, creates the permanent anatomical condition of a subclavian steal phenomenon (Fig 1). The subclavian steal phenomenon, initially described by Reivich and colleagues in 1961,2 is related to a proximal subclavian artery stenosis or occlusion responsible for the occurrence of retrograde collateral flow in the ipsilateral vertebral artery. A few reports have suggested a possible link between a Blalock-Taussig shunt and the late occurrence of cerebrovascular disease. We report the first documented case of vertebrobasilar stroke occurring as a late complication of a Blalock-Taussig shunt followed by total cardiac repair during childhood. Case Report A 39-year-old man was admitted to our hospital because of an acute episode of nausea without vomiting, vertigo, and decreased hearing on the right side, rapidly followed by right facial numbness as well as weakness and decreased sensation in the left arm. The patient also described a visual disturbance consistent with nystagmus. These symptoms appeared without an identifiable precipitating factor and lasted for approximately 20 to 30 minutes. The patient was known to have tetralogy of Fallot treated by a Blalock-Taussig shunt at the age of 2 years followed by definitive surgical repair at the age of 9 years. He had been in good health since then. The review of risk factors for cerebrovascular disease, including obesity, tobacco use, hyperlipidemia, and hypertension, was negative. The social and familial histories showed only hypertension and hyperlipidemia in both parents. On admission, the general and neurological examinations were normal except for asymmetrical upper extremity blood pressure measurements (125/80mm Hg on the right, 146/ 95mm Hg on the left) and unpalpable distal arterial pulses in the right arm. According to the patient, this discrepancy had been known since his second cardiac surgery. The laboratory values were unremarkable; in particular, the lipid profile was normal, and there was no evidence of a hypercoagulable state (the protein C and S antigens, antithrombin III activity, dilute Russell’s viper venom time test, and homocysteine level were within normal ranges; the factor V Leiden mutation was negative). Computed tomography of the brain was normal. Magnetic resonance imaging showed signal anomalies in the inferior aspect of the right cerebellar hemisphere consistent with an acute ischemic lesion in the right posterior inferior cerebellar artery territory (Fig 2A). Magnetic resonance imaging incidentally showed inflammatory changes in the maxillary sinus and ethmoid cells bilaterally, whereas the sphenoid and frontal sinuses were unremarkable. The patient was not treated for this asymptomatic inflammatory sinus disease. A transesophageal echocardiogram, including a bubble study performed to © 2002 Wiley-Liss, Inc. 231 tery branches, including the right vertebral artery and the right ascending and deep cervical arteries (see Fig 2C). The cerebral vasculature was otherwise unremarkable. There was in particular no evidence of intracranial or extracranial atheromatous disease and no arterial or venous anomalies potentially associated with stroke. The patient was discharged without sequelae from his stroke. However, he continued to report frequent episodes of dizziness, associated once with facial numbness and diplopia, despite being anticoagulated with warfarin at a therapeutic level. At this stage, surgical correction of his vascular anomaly was advised to the patient, who wished to proceed with this option after the exposition of its potential risks and benefits. Interposition of a synthetic graft (Hemoshield Dacron Graft, Boston Scientific Corporation, Natick, MA) between the right common carotid and subclavian arteries was performed uneventfully, allowing antegrade flow to be reestablished in the right vertebral artery. After surgery, his episodic symptoms attributable to brainstem ischemia resolved. He has remained asymptomatic for more than 1 year. Fig 1. Schematic representation of the morphological changes associated with a Blalock-Taussig shunt. The pulmonary trunk and arteries are represented in light gray; the aortic arch (1) and supra-aortic trunks are in dark gray. The Blalock-Taussig shunt is established between the proximal right subclavian artery (6) and the right pulmonary artery (8) by interposition of a synthetic graft (7). The subclavian steal phenomenon, that is, antegrade flow in the left vertebral artery and retrograde flow in the right vertebral artery (4), provides blood supply to the right arm (direction of blood flow as indicated by the arrows). The figure also shows the right common carotid artery (2), the right carotid bifurcation (3), and the basilar artery (5). rule out potential cardioembolic sources, was unremarkable except for a mildly dilated right ventricular cavity. Reverse flow in the right vertebral artery was by documented transcranial Doppler. The transcranial Doppler bubble test was negative for embolic events at rest and during Valsalva maneuver and coughing. Cerebral digital subtraction angiography then was performed. The aortic arch study showed interruption of the right subclavian artery at its origin (see Fig 2B). Selective injections of the left vertebral artery confirmed the presence of a left-to-right subclavian steal phenomenon. Revascularization of the right upper extremity was provided by retrograde filling of several subclavian ar- 232 Annals of Neurology Vol 52 No 2 August 2002 Discussion We report the case of a 39-year-old man presenting with a vertebrobasilar stroke in the absence of identifiable cerebrovascular risk factors. The symptoms described by our patient pointed to a lesion in the territory of the right posterior inferior cerebellar artery, which was confirmed by neuroimaging. In our patient, the co-occurrence of a vertebrobasilar infarct and a subclavian steal phenomenon strongly suggest a causative relationship between the 2 events, that is, a subclavian steal syndrome secondary to the cardiac surgery undergone during childhood. Although an embolic mechanism cannot be completely excluded because of normal transesophageal echocardiogram and transcranial Doppler, both with negative bubble tests, that the involved branch (the right posterior inferior cerebellar artery) originates from the vertebral artery with reverse flow renders this explanation very unlikely. Conversely, this anatomical situation is consistent with an increased risk of preferential hypoperfusion in the right posterior inferior cerebellar artery territory. The possible association between a Blalock-Taussig shunt and a subclavian steal phenomenon initially was proposed by Folger and Shah in 1965.3 By reviewing 123 cardiac angiograms performed on patients with a Blalock-Taussig shunt, these authors could identify late opacification of a subclavian artery suggestive of a subclavian steal phenomenon in 12 instances. The steal phenomenon was associated with dizziness in 3 patients and with visual disturbances in 2 patients. However, in those patients and in other case reports describing early cerebrovascular events after a BlalockTaussig shunt,4,5 the vertebrobasilar manifestations occurred after the creation of the shunt but before definitive repair of the tetralogy of Fallot. In these pa- Fig 2. (A) Magnetic resonance imaging study of the brain. This axial T2-weighted image shows increased signal in the lower portion of the right cerebellar hemisphere associated with minimal mass effect on the right posterior-lateral aspect of the brainstem and subtle leftward midline shift. A small area of hypersignal also is seen in the right side of the myelencephalon (arrowhead). This pattern of abnormal signal is consistent with an ischemic lesion in the territory of the right posterior-inferior cerebellar artery. Note the incidental finding of chronic maxillary sinus inflammation. (B) Digital subtraction angiography of the aortic arch, left anterior oblique view. The aortic arch, the right common carotid artery (RCC), the left common carotid artery (LCC), and the left subclavian artery (LSC) are unremarkable. The right subclavian artery is interrupted close to its origin from the innominate artery (arrow). (C) Selective digital subtraction angiography of the left vertebral artery, anteroposterior view. The catheter tip (arrowhead) is placed at the origin at the left vertebral artery (LV). Retrograde flow in the right vertebral artery (RV) and in several cervical branches allows for late opacification of the right subclavian artery (arrow). tients, contributing factors related to cardiac dysfunction, such as hypoxia, polycythemia, bacterial endocarditis, and mural or valvular thrombosis, have to be considered.6 The association of a Blalock-Taussig shunt with delayed ischemic complications occurring after total cardiac repair of the tetralogy of Fallot was suggested in 1984 by Kurlan and colleagues.6 These authors described a 38-year-old patient developing transient vertebrobasilar ischemia 31 years after a Blalock-Taussig shunt and 4 years after total repair of Gailloud et al: Blalock-Taussig Shunt and Stroke 233 his tetralogy of Fallot. Our 39-year-old patient presented with a cerebellar infarct 37 years after the creation of the shunt and 30 years after total cardiac repair. This observation, the first report of a documented stroke after Blalock-Taussig shunt and total cardiac repair, appears to confirm that patients who underwent a Blalock-Taussig procedure in childhood have an increased risk of vertebrobasilar insufficiency later in life. The few cases reported so far that have attributed a late cerebrovascular event to a Blalock-Taussig shunt might reflect only the relatively young age of the concerned population. Blalock-Taussig shunts have been performed on infants and small children for approximately 50 years, so much of that population is now approaching middle age. It is conceivable that the morphological risk factor constituted by the surgically created subclavian steal phenomenon generally is not significant enough to become symptomatic in isolation, as appears to have been the case for our patient. However, the association of the Blalock-Taussig anatomical anomalies with age-related processes such as hypertension and atheromatous disease might predispose these patients to more precocious development of cerebrovascular diseases. If this assumption is correct, patients with a history of Blalock-Taussig shunt constitute an overlooked group at risk for cerebrovascular disease. Furthermore, these patients are now in good general health; most of them have lost contact with their cardiologist and are not followed up on a systematic basis. In view of the potentially devastating consequences of a vertebrobasilar stroke, patients who underwent a Blalock-Taussig shunt may benefit from prophylactic measures, such as the administration of platelet antiaggregant agents. More aggressive treatments may be necessary for patients who remain symptomatic under conservative management, such as the carotidsubclavian bypass performed in our patient. Follow-up studies of patients with a Blalock-Taussig shunt have been conducted mainly from a cardiological perspective.7 There is, as far as we know, no published long-term evaluation of this population from a specific neurological viewpoint. Such a study is required to better define the late cerebrovascular risk potentially associated with a history of Blalock-Taussig shunt. References 1. Blalock A, Taussig HB. The surgical treatment of the heart in which there is pulmonary stenosis or pulmonary atresia. JAMA 1945;128:189 –202. 2. Reivich M, Holling HE, Roberts B, Toole JF. Reversal of blood flow through the vertebral artery and its effect on cerebral circulation. N Engl J Med 1961;265:878 – 885. 3. Folger GM, Shah KD. Subclavian steal in patients with BlalockTaussig anastomosis. Circulation 1965;31:241–248. 4. Naito H, Kurokawa K, Kanno T, et al. Status epilepticus and cortical blindness due to subclavian steal syndrome in a girl with Blalock’s operation. Surg Neurol 1973;1:46 – 49. 234 © 2002 Wiley-Liss, Inc. 5. Sokol S, Narkiewicz M, Billewicz O. Subclavian steal syndrome after Blalock-Taussig anastomoses. J Cardiovasc Surg (Torino) 1969;10:350 –354. 6. Kurlan R, Krall RL, Deweese JA. Vertebrobasilar ischemia after total repair of tetralogy of Fallot: significance of subclavian steal created by Blalock-Taussig anastomosis. Stroke 1984;15: 359 –362. 7. Murphy JG, Gersh BJ, Mair DD, et al. Long-term outcome in patients undergoing surgical repair of tetralogy of Fallot. N Engl J Med 1993;329:593–599. Cerebral X-Linked Adrenoleukodystrophy in a Girl with Xq27-Ter Deletion Eli Hershkovitz, MD,1 Ginat Narkis, BA,2 Zamir Shorer, MD,1 Ann B. Moser, BA,3 Paul A. Watkins, MD, PhD,3 Hugo W. Moser, MD,3 and Esther Manor, PhD2 An 8.5-year-old girl with a pathogenic mutation (515insC) of the ATP-binding cassette, subfamily D, member 1 gene (ABCD1) on her maternally derived X chromosome showed clinical, biochemical, and magnetic resonance imaging abnormalities similar to those in affected males. Cytogenetic studies led to the surprise finding of a de novo deletion of Xq27 on the paternally derived X chromosome. A bone marrow transplant had an apparently favorable effect. Cytogenetic studies should be performed in all severely symptomatic X-linked adrenoleukodystrophy heterozygotes. Ann Neurol 2002;52:234 –237 X-linked adrenoleukodystrophy (X-ALD) is a progressive disorder that affects myelin because of a defect in the gene for the adenosine triphosphate (ATP)-binding cassette, subfamily D, member 1 (ABCD1).1 ABCD1 is located on Xq28. It codes for ALD protein (ALDP),2 a From the 1Pediatric Department and 2Genetic Institute, Soroka University Medical Centre, Faculty of Health Sciences, Ben-Gurion University of the Negev, Beer Sheva, Israel, and 3Kennedy Krieger Institute, Johns Hopkins University, Baltimore, MD. Received Dec 14, 2001, and in revised form Mar 12, 2002. Accepted for publication Mar 12, 2002. Published online Jun 23, 2002, in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ana.10248 Address correspondence to Dr Hershkovitz, Pediatric Department, Soroka University Medical Centre, POB 151, Beer Sheva 84101, Israel. E-mail: email@example.com peroxisomal membrane protein that belongs to the ATPbinding cassette superfamily of membrane transport proteins.3 X-ALD is associated with the accumulation of saturated, very long chain fatty acid in tissues, cultured skin fibroblasts,4 and plasma.5 In males, phenotypic expression ranges from the severe childhood or adolescent cerebral forms, which are associated with a white matter inflammatory reaction,6 to slowly progressive adrenomyeloneuropathy, which manifests most commonly in young adults and mainly affects the long tracts in the spinal cord, with the inflammatory reaction mild or absent.7 Approximately 70% of affected males have primary adrenocortical insufficiency. Adrenal steroid therapy is effective for the endocrine dysfunction but does not appear to alter neurological progression. Bone marrow transplantation (BMT) can be of long-term benefit to boys and adolescents with the cerebral forms of the disease when provided at a time when brain involvement is still relatively mild.8 There is no specific therapy for adrenomyeloneuropathy. Approximately 50% of women who are heterozygous for X-ALD develop a progressive syndrome that resembles adrenomyeloneuropathy but is milder and usually does not manifest until 35 years of age or later. Adrenocortical insufficiency is rare. Approximately 1% of heterozygous women have a more severe phenotype, in which there is progressive cerebral involvement, adrenocortical insufficiency, and earlier onset9 and which resembles what is seen in boys or adolescents with the cerebral phenotypes (H.W.M., unpublished observation). The cause of the severe manifestations in this small proportion of heterozygous women has not been determined definitively. Skewed inactivation of the normal X chromosome has been considered to be the most likely explanation.9,10 X-inactivation patterns in cultured skin fibroblasts or white blood cells can be assessed by an examination of the expression of ALDP by immunocytochemical techniques11 in women who are members of families in which affected males are known to lack immunoreactive material. In 15 heterozygous women, 23 to 86% of cells expressed ALDP.12 In contrast, ALDP was expressed in only 0 to 2% of cells from 2 women who had the adolescent cerebral phenotype,10 suggesting that in these patients the normal X chromosome failed to exert its protective effect. This conclusion must be considered tentative because findings in fibroblasts may not reflect those in the nervous system. We now report a patient with the severe phenotype who was known to be heterozygous for X-ALD because of a pathogenic mutation in the maternal cell line but who had in addition a deletion of Xq27-ter in the paternal chromosome, thereby rendering her totally deficient in ALDP. Case Report The patient is the first child of unrelated parents. Two maternal uncles have X-ALD, 1 with the adolescent cerebral phenotype and the other with the adrenomyeloneuropathy phenotype. Pregnancy and early development were normal. She started to walk at 14 months and to speak at 1 year. Several behavioral problems were noted in childhood. These included separation anxiety, with refusal to stay alone in the kindergarten and school; oversensitivity to noises; and isolation from peers and kindergarten staff. She continued to develop neurologically in agreement with her age, and she started to read at 5 to 6 years of age. At that time, she was referred to formal psychological assessment because of “school phobia,” but no diagnosis was made at that time. The deterioration of her academic performance and behavior was noted at 8.5 years. Physical and neurological examinations were normal. Formal psychological assessment (Wechsler intelligence scale for children-revised) showed that she functioned at the low normal range. The verbal intelligence quotient was 87, the performance intelligence quotient was 83, and the total intelligence quotient was 83. Her achievements were adversely affected by easy distractibility and visuospatial defects. Cortisol response to 1g of intravenous adrenocorticotropic hormone was normal at 30 minutes (26.2g/dl), ruling out clinically significant adrenal insufficiency. Brain magnetic resonance imaging demonstrated diffuse white matter involvement that was most prominent in the frontal regions. Results Biochemical and genetic studies were performed after informed consent had been obtained from the parents. Levels of very long chain fatty acid in plasma and cultured skin fibroblasts were measured by capillary gasliquid chromatography4,5 and found to be elevated (Table 1). DNA analysis of the patient’s white blood cells and her mother’s white blood cells showed a cytosine insertion in codon 515 (515insC) resulting in a frameshift after amino acid 171 (tyrosine). Immunocytochemical studies of the cultured skin fibroblasts11 showed that ALDP-reactive material was lacking in 99% of the patient’s cells (100 cells counted) and in all of the cells of the affected maternal uncles. Cytogenetic analyses showed a deletion at Xq27.23ter in the patient’s peripheral blood lymphocytes. Her karyotype was 46X;Xdel(q27-tel) (data not shown). Both parents’ karyotype analyses were normal. A deletion of the distal part of Xq, including the telomeric region in the patient’s fibroblasts, also was observed by fluorescence in situ hybridization with a specific Xq telomeric probe. This region was present in the X chromosomes of both parents (data not shown). DNA analysis with 23 polymorphic markers spanning the q26-28 region of the patient’s and her parents’ X chromosomes was performed to define the extent of the deletion, using methods that have been described previously.13 An analysis of 13 informative markers showed that a de novo deletion had occurred in the paternal chromo- Hershkovitz et al: Cerebral X-ALD 235 Table 1. Very Long Chain Fatty Acid Levels in Patient’s Plasma and Cultured Skin Fibroblasts Location Patient X-ALD Male X-ALD Heterozygote Control 1.81 0.91 0.043 ND ND ND ND ND ND ⬍1 0.68 ⫾ 0.15 0.018 ⫾ 0.009 0.762 0.434 0.457 0.64 ⫾ 0.32 0.41 ⫾ 0.15 0.69 ⫾ 0.19 0.72 ⫾ 0.26 0.27 ⫾ 0.17 0.40 ⫾ 0.23 0.90 ⫾ 0.50 0.09 ⫾ 0.07 0.08 ⫾ 0.03 Plasmaa C26:0 C24:0/C22:0 C26:0/C22:0 Cultured skin fibroblastsb C22:0 C26:0 C26:0/C22:0 The plasma assays were performed at the Sharee Zedek Hospital in Jerusalem (Dr Orly Elpeleg) and those in fibroblasts at the Kennedy Krieger Institute. The C26:0 levels in fibroblasts are similar to those of X-ALD male, and the C26:0/C22:0 ratios in fibroblasts are intermediate between those in male and heterozygous X-ALD patients and compatible with either genetic status. a Expressed as micrograms per milliliter of plasma. Expressed as micrograms per milligram of protein. b ND ⫽ not done; X-ALD ⫽ X-linked adrenoleukodystrophy. some. The deleted region is distal to marker DXS1227, which maps to Xq27.2 (Table 2). Therefore, the deleted segment spans part of Xq27.2 and all of Xq27.3 and Xq28. At 8.9 years of age, she received a bone marrow transplant. Her human leukocyte antigen–identical normal younger sister was the donor. The procedure was tolerated well. Full engraftment was documented. Although some deterioration in concentration and short-term memory were noticed during the early period after transplantation, her neurological and neuropsychological status were stable 18 months after BMT. Discussion The abnormal magnetic resonance imaging results and moderately abnormal neuropsychological dysfunction in this 8.5-year-old girl known to be heterozygous for X-ALD on the basis of family history, mutation analysis, and very long chain fatty acid studies indicate that she is among the few heterozygotes with a severe phenotype similar to that in boys with the cerebral forms of the disease. Mutation analysis indicates that the maternal side of her family has a mutation that abolishes the expression of ALDP. Studies in other heterozygotes have led to the hypothesis that severe disability, such as cerebral involvement in childhood, is caused by skewed X inactivation,9,10 and the absence of ALDP immunoreactive material would be compatible with this. However, cytogenetic studies led to a different conclusion, namely, the unexpected demonstration of a de novo deletion in her paternal X chromosome that involves all of Xq28 and part of Xq27. Combined with the abnormality on the maternal X chromosome, this leads to failure to express functional ALDP, similar to what occurs in affected males. Cytogenetic studies are performed only rarely in patients with X-ALD. To our knowledge, this type of abnormality has not been reported before in women heterozygous for X-ALD. The 236 Annals of Neurology Vol 52 No 2 August 2002 finding is of practical significance because under these circumstances therapeutic approaches are similar to those that would be used in affected males. Given this reasoning and because the patient met current criteria for BMT in affected males,8 we decided to perform the transplant. Her condition was stable 18 months later, and we hope that her long-term outcome will be favorable, as has been the case for male patients.8 We recommend that cytogenetic studies be performed in the approximately 1% of heterozygotes who show evidence of cerebral involvement by magnetic resonance imaging. A karyotype study should be the first step. A normal result should be followed by a search for microdeletions in chromosome X with specific DNA markers spanning the ALD gene region. This information is clinically significant because patients with cerebral involvement may be candidates for BMT. At this time, BMT is considered only for patients who have evidence of cerebral involvement.8 It is never recommended for males or females who have spinal cord Table 2. Haplotype DNA Analysis of Xq26-28 Using Informative DNA Markers Marker DXS8098 DXS1047 DXS994 DXS8050 DXS8072 DXS1062 DXS8094 DXS1192 DXS1227 DXS8084 DXS8043 DXS1200 DXS8091 Band Xq26.1 Xq26.3 Xq26.3 Xq26.3 Xq26.3 Xq27.1 Xq27.1 Xq27.2 Xq27.2 Xq27.3 Xq27.3 Xq27.3 Xq28 Maternal Allele Paternal Allele Present Present Present Present Present Present Present Present Present Present Present Present Present Present Present Present Present Present Present Present Present Absent Absent Absent Absent Absent involvement only. For female patients with cerebral involvement, BMT would be recommended only for those in whom both alleles are involved, thereby making them equivalent to hemizygotes. We do not believe that BMT would be indicated for women who do have 1 normal allele, where the pathogenesis of brain involvement is unknown (although skewed inactivation is the most likely cause), and we do not believe that in such patients the risk-benefit ratio would warrant BMT. Although it might be of some interest to do cytogenetic studies on all heterozygotes, the added expense is not warranted in women without cerebral involvement. References 1. Moser HW, Smith KD, Watkins PA, et al. X-linked adrenoleukodystrophy. In: Scriver CR, Beaudet AL, Sly WS, Valle D, eds. The metabolic and molecular bases of inherited disease. 8th ed. New York: McGraw-Hill, 2001:3257–3301. 2. Mosser J, Douar AM, Sarde CO, et al. Putative X-linked adrenoleukodystrophy gene shares unexpected homology with ABC transporters. Nature 1993;361:726 –730. 3. Higgins CF, Gallagher MP, Mimmack ML, Pearce SR. A family of closely related ATP-binding subunits from prokaryotic and eukaryotic cells. Bioessays 1988;8:111–116. fs 4. Moser HW, Moser AB, Kawamura N, et al. Adrenoleukodystrophy: elevated C26 fatty acid in cultured skin fibroblasts. Ann Neurol 1980;7:542–549. 5. Moser AB, Kreiter N, Bezman L, et al. Plasma very long chain fatty acids in 3,000 peroxisome disease patients and 29,000 controls. Ann Neurol 1999;45:100 –110. 6. Powers JM, Liu Y, Moser AB, Moser HW. The inflammatory myelinopathy of adreno-leukodystrophy: cells, effector molecules, and pathogenetic implications. J Neuropathol Exp Neurol 1992;51:630 – 643. 7. Powers JM, DeCiero DP, Ito M, et al. Adrenomyeloneuropathy: a neuropathologic review featuring its noninflammatory myelopathy. J Neuropathol Exp Neurol 2000;59:89 –102. 8. Shapiro E, Krivit W, Lockman L, et al. Long-term beneficial effect of bone marrow transplantation for childhood onset cerebral X-linked adrenoleukodystrophy. Lancet 2000;356: 713–718. 9. Heffungs W, Hameister H, Ropers HH. Addison disease and cerebral sclerosis in an apparently heterozygous girl: evidence for inactivation of the adrenoleukodystrophy locus. Clin Genet 1980;18:184 –188. 10. Naidu S, Washington C, Thirumalai S, et al. X-chromosome inactivation in symptomatic heterozygotes in X-linked adrenoleukodystrophy. Ann Neurol 1997;42:498 (Abstract). 11. Watkins PA, Gould SJ, Smith MA, et al. Altered expression of ALDP in X-linked adrenoleukodystrophy. Am J Hum Genet 1995;57:292–301. 12. Feigenbaum V, Lombard-Platet G, Guidoux S, et al. Mutational and protein analysis of patients and heterozygous women with X-linked adrenoleukodystrophy. Am J Hum Genet 1996; 58:1135–1144. 13. Parvari R, Mumm S, Galil A, et al. Deletion of 8.5 Mb, including the FMR1 gene, in a male with the fragile X syndrome phenotype and overgrowth. Am J Med Genet 1999;83: 302–307. A Novel Mutation in the Deoxyguanosine Kinase Gene Causing Depletion of Mitochondrial DNA Jan-Willem Taanman, PhD,1 Ihab Kateeb, MD,2 Ania C. Muntau, MD,3 Michaela Jaksch, MD,4 Nadine Cohen, MD,2 and Hanna Mandel, MD5 Recently, a homozygous single-nucleotide deletion in exon 2 of the deoxyguanosine kinase gene (DGUOK) was identified as the disease-causing mutation in 3 apparently unrelated Israeli-Druze families with depleted hepatocerebral mitochondrial DNA. We have discovered a novel homozygous nonsense mutation in exon 3 of DGUOK (313C3 T) from a patient born to nonconsanguineous German parents. This finding shows that mutations in DGUOK causing mitochondrial DNA depletion are not confined to a single ethnic group. Ann Neurol 2002;52:237–239 In 1991, Moraes and colleagues1 identified a group of infants with marked depletion of mitochondrial DNA (mtDNA) in association with defective mitochondrial respiratory chain function. This condition, often called mtDNA depletion syndrome (Online Mendelian Inheritance in Man [OMIM] 251880), now has been described for more than 50 patients;2–11 this suggests that it may be an important cause of mitochondrial dysfunction in neonates and infants. Most of the reported patients present in the neonatal period with muscle weakness, liver failure, and neurological abnormalities associated with lactic acidemia and die before 12 From the 1University Department of Clinical Neurosciences, Royal Free and University College Medical School, University College London, London, United Kingdom; 2Department of Genetics, Tamkin Human Molecular Genetics Research Facility, TechnionIsrael Institute of Technology, Bruce Rappaport Faculty of Medicine, Haifa, Israel; 3Dr von Hauner Children’s Hospital, LudwigMaximilians-Universität, Munich, Germany; 4Stoffwechselzentrum und Institut für Klinische-Chemie, Krankenhaus MünchenSchwabing, Munich, Germany; and 5Metabolic Disease Unit, Department of Pediatrics, Rambam Medical Center, Technion-Israel Institute of Technology, Bruce Rappaport Faculty of Medicine, Haifa, Israel. Received Dec 13, 2001, and in revised form Mar 12, 2002. Accepted for publication Mar 13, 2002. Published online Jun 23, 2002, in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ana.10247 Address correspondence to Dr Taanman, University Department of Clinical Neurosciences, Royal Free and University College Medical School, University College London, Rowland Hill Street, London NW3 2PF, United Kingdom. E-mail: firstname.lastname@example.org © 2002 Wiley-Liss, Inc. 237 months of age. Others present in infancy with isolated myopathy associated with motor regression or a slowly progressive encephalomyopathy. Inheritance appears to be autosomal recessive, suggesting that nuclear gene defects are responsible for mtDNA depletion. In agreement with this notion, cybrid studies have shown that the defect is not transmitted by patient mtDNA,3,12 and partial sequencing of patient mtDNA has failed to identify mutations.1,4 Recently, homozygosity mapping followed by candidate gene sequence analysis in 3 kindreds of IsraeliDruze origin with the early-onset hepatocerebral variant of the disease showed a homozygous mutation in the gene for deoxyguanosine kinase (DGUOK) on chromosome 2p13.13 In addition, mutations in the gene for mitochondrial thymidine kinase (TK2) on chromosome 16q22 were identified in patients with the late-onset muscle-specific variant of the disease.14 Deoxyguanosine kinase is responsible for phosphorylation of purine deoxyribonucleosides in the mitochondrial matrix compartment.15 The 3 apparently unrelated Israeli-Druze families carried the same singlenucleotide deletion in exon 2 of DGUOK.13 This mutation is expected to result in premature termination of translation. To investigate whether mutations in DGUOK are also present in patients from a different ethnic origin, we sequenced the exons and intron/exon boundaries of DGUOK from a German patient with the hepatocerebral form of the disease. The patient was found to harbor a novel homozygous nonsense mutation in DGUOK. This shows that mutations in DGUOK causing mtDNA depletion are not restricted to patients belonging to one particular ethnic group. Case Report The clinical details of the patient and histochemical findings are documented in detail elsewhere.11 In brief, the patient was the first child of healthy, nonconsanguineous German parents (birth weight, 2,570gm). The boy presented with lactic acidosis, hepatomegaly, hypoglycemia, and jaundice shortly after birth. He had a severe encephalopathy, characterized by marked muscle hypotonia, hyperreflexia, irritability, and pendular horizontal nystagmus. At the age of 2 months, neonatal giant cell hepatitis was diagnosed by light microscopy. Electron microscopy of the liver showed an accumulation of abnormal mitochondria and steatosis. Histochemistry and immunohistochemistry for cytochrome-c oxidase demonstrated a mosaic pattern of normal and deficient hepatocytes. Skeletal muscle was normal on both light and electron microscopy, but biochemical assays showed a minor cytochrome-c oxidase deficiency. Southern blot analysis of liver biopsies, taken at 2 and 3 months of age, showed that mtDNA was of normal size, but mtDNA levels were only 17 and 18%, respectively, of the mean of 6 age-matched control specimens (range controls, 66 –140%).11 The patient died of hepatic failure at the age of 5 months. A younger brother shows similar symptoms. 238 Annals of Neurology Vol 52 No 2 August 2002 After informed parental consent, in accordance with the guidelines of the local institution, DNA was extracted from liver of the proband, blood of the parents, and cultured fibroblasts of the second child.6 All exons of DGUOK were amplified by polymerase chain reaction, purified, and sequenced exactly as described earlier.13 Sequencing was performed in both directions. The human DGUOK messenger RNA sequence, with accession number U41668, and the human chromosome 2 working draft sequence segment, with accession number NT025651, were used to determine the intron/exon structure of DGUOK. Results Sequencing of the 7 exons and the intron/exon boundaries of the DGUOK gene from the proband showed a homozygous C3 T transition at nucleotide position 313 (numbering according to Johansson and Karlsson16) of exon 3. The same homozygous mutation was found in the clinically affected sibling. Both parents were heterozygous for the mutation (Fig). The mutation changes the arginine CGA codon 105 into a TGA stop codon (see Fig). This base change is predicted to result in a 173–amino acid residue truncation at the C terminus of the DGUOK protein product. The mutation was not found in 15 control subjects of European origin. Discussion In an earlier study, we identified an infant, born to healthy, nonconsanguineous German parents, with early-onset encephalopathy and rapidly progressing liver failure associated with severely depleted levels of mtDNA.11 The recent discovery of a single-nucleotide deletion in exon 2 of DGUOK in 3 unrelated IsraeliDruze families with depleted hepatocerebral mtDNA prompted us to screen the German patient for mutations in the gene. We found a homozygous nonsense mutation in exon 3 of the proband and his affected younger brother, whereas both parents were carriers. The mutation will lead to a shortening of the DGUOK protein product by more than half. The truncation includes 3 domains that are evolutionarily conserved between nucleoside kinases and are thought to be essential for catalysis.16 It is, therefore, highly unlikely that the mutated DGUOK gene is still functionally active in the patient. The absence of the mutation in control subjects of European origin further supports the pathogenicity of the mutation. The mutation in the DGUOK gene that we identified in a German family indicates that mutations in DGUOK causing mtDNA depletion are not limited to Israeli-Druze patients. However, mutation screening of DGUOK in 11 additional families with early-onset encephalopathy, liver failure, and mtDNA depletion of British (5 patients), Greek-Cypriot (3 patients), German (1 patient), French (1 patient), and Turkish (1 This study was supported by the Wellcome Trust (grant 048410, J.W.T.), the DFG (grant Ja 802/2-1, M.J.), and the Joseph Elias Fund/Technion VPR Fund (grant 181-421, H.M.). We thank Dr A. H. V. Schapira and Dr J. V. Leonard for helpful discussions. References Fig. Electropherograms showing part of the exon 3 sequences of the deoxyguanosine kinase gene (DGUOK) of a control, both parents, the proband (1st child), and his brother (2nd child). The position of the mutation is indicated with an arrow. Numbering is according to Johansson and Karlsson.16 patient) origin, including the 6 families described by us previously,3,6,8 did not show any mutations in the 7 exons and the intron/exon boundaries of DGUOK (not shown). The clinical, biochemical, and molecular findings of these patients were not markedly different from the case described here. These results suggest that mutations in the coding region of DGUOK are not common in patients with hepatocerebral mtDNA depletion. Although the disease may be genetically heterogeneous, our results do not rule out that these 11 families carry mutations elsewhere in DGUOK that affect gene expression. This possibility is likely in 2 additional families of Druze and Moroccan origin, respectively, in which the disease has been linked to DGUOK but in which mutations in the coding region of the gene could not be identified.13 More detailed mutation analysis and expression studies of DGUOK, therefore, are necessary to investigate the genetic heterogeneity of hepatocerebral mtDNA depletion. 1. Moraes CT, Shanske S, Trischler H-J, et al. mtDNA depletion with variable tissue expression: a novel genetic abnormality in mitochondrial diseases. Am J Hum Genet 1991;48:492–501. 2. Tritschler H-J, Andreetta F, Moraes CT, et al. Mitochondrial myopathy of childhood associated with depletion of mitochondrial DNA. Neurology 1992;42:209 –217. 3. Bodnar AG, Cooper JM, Holt IJ, et al. Nuclear complementation restores mtDNA levels in cultured cells from a patient with mtDNA depletion. Am J Hum Genet 1993;53:663– 669. 4. Mariotti C, Uziel G, Carrara F, et al. Early-onset encephalomyopathy associated with tissue-specific mitochondrial DNA depletion: a morphological, biochemical and molecular-genetic study. J Neurol 1995;242:547–556. 5. Macmillan CJ, Shoubridge EA. Mitochondrial DNA depletion: prevalence in a pediatric population referred for neurologic evaluation. Pediatr Neurol 1996;14:203–210. 6. Morris AAM, Taanman J-W, Blake J, et al. Liver failure associated with mitochondrial DNA depletion. J Hepatol 1998;28: 556 –563. 7. Vu TH, Sciacco M, Tanji K, et al. Clinical manifestations of mitochondrial DNA depletion. Neurology 1998;50:1783–1790. 8. Blake JC, Taanman J-W, Morris AMM, et al. Mitochondrial DNA depletion syndrome is expressed in amniotic fluid cell cultures. Am J Pathol 1999;155:67–70. 9. Barthélé my C, Ogier de Baulny H, Diaz J, et al. Late-onset mitochondrial DNA depletion: DNA copy number, multiple deletions, and compensation. Ann Neurol 2001;49:607– 617. 10. Mandel H, Hartman C, Berkowitz D, et al. The hepatic mitochondrial DNA depletion syndrome: ultrastructural changes in liver biopsies. Hepatology 2001;34:776 –784. 11. Müller-Höcker J, Muntau AC, Schäfer S, et al. Depletion of mitochondrial DNA in the liver of an infant with neonatal giant cell hepatitis. Hum Pathol 2002;33:247–253. 12. Taanman J-W, Bodnar AG, Cooper JM, et al. Molecular mechanisms in mitochondrial DNA depletion syndrome. Hum Mol Genet 1997;6:935–942. 13. Mandel H, Szargel R, Labay V, et al. The deoxyguanosine kinase gene is mutated in individuals with depleted hepatocerebral mitochondrial DNA. Nat Genet 2001;29:337–341. 14. Saada A, Shaag A, Mandel H, et al. Mutant mitochondrial thymidine kinase in mitochondrial DNA depletion myopathy. Nat Genet 2001;29:342–344. 15. Jüllig, M, Eriksson S. Mitochondrial and submitochondrial localization of human deoxyguanosine kinase. Eur J Biochem 2000;267:5466 –5472. 16. Johansson M, Karlsson A. Cloning and expression of human deoxyguanosine kinase cDNA. Proc Natl Acad Sci U S A 1996; 93:7258 –7262. Taanman et al: A Novel Mutation Depleting mtDNA 239 Dyskinesias and Grip Control in Parkinson’s Disease Are Normalized by Chronic Stimulation of the Subthalamic Nucleus Roland Wenzelburger, MD,1 Bao-Rong Zhang, MD,1 Meike Poepping, MD,1 Bettina Schrader, MD,2 Dieter Müller, PhD,3 Florian Kopper, MD,1 Urban Fietzek, MD,1 Hubertus M. Mehdorn, PhD,2 Günther Deuschl, PhD,1 and Paul Krack, PhD1 Deep-brain stimulation of the subthalamic nucleus appears to reduce levodopa-induced dyskinesias, but whether this effect is caused by the reduction of the total levodopa ingestion or represents a direct effect on the motor system is unknown. Precision grip force of grasping movements and levodopa-induced dyskinesias was analyzed in 10 parkinsonian patients before and after 3 months of deep-brain stimulation of the subthalamic nucleus. Peak grip force was abnormally increased before surgery in the off-drug state and, particularly, in the ondrug state (sensitization). This grip force upregulation normalized with chronic deep-brain stimulation in both conditions (desensitization). Peak-dose dyskinesias also improved, and off-dystonia was completely abolished. Mean dosage of dopaminergic drugs was reduced, but force overflow and dyskinesias were equally improved in 2 patients without a reduction. Despite the same single levodopa test dose, force excess and levodopa-induced dyskinesias were drastically reduced after 3 months of deep-brain stimulation of the subthalamic nucleus. This indicates that direct effects of deep-brain stimulation of the subthalamic nucleus on levodopa-induced dyskinesias are likely to occur. Grip force overflow is a promising parameter to study the desensitizing effect of chronic deep-brain stimulation on levodopa-induced dyskinesias. Ann Neurol 2002;52:240 –243 From the Departments of 1Neurology and 2Neurosurgery, Christian Albrechts University, Kiel; and the 3Department of Neurosurgery, University Hamburg, Hamburg, Germany. Received Dec 27, 2001, and in revised form Mar 14, 2002. Accepted for publication Mar 23, 2002. Published online Jun 23, 2002, in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ana.10254 Dr Zhang’s current address is the Department of Neuology, Second Affiliated Hospital, Zhejiang University, Hangzhou, P.R. China. Address correspondence to Prof Dr Deuschl, Neurologische Klinik of the University Hospital Kiel, Niemannsweg 147, 24105 Kiel, Germany. E-mail: email@example.com 240 © 2002 Wiley-Liss, Inc. L-Dopa–induced dyskinesias (LIDs) and motor fluctuations are major complications of therapy in longstanding Parkinson’s disease (PD). Once a patient develops LID, the sensitization to dopaminergic treatment cannot be reversed by drugs.1 Deep-brain stimulation of the subthalamic nucleus (STN DBS) is an effective treatment for both fluctuations and dyskinesias in L-dopa–responsive PD.2,3 Its beneficial effects on akinesia, rigidity, and tremor resemble that of 4 L-dopa, except for LIDs, which are promoted by acute STN stimulation, whereas they improve with chronic stimulation.5 This substantial improvement of LID by STN DBS even in the on-drug state (desensitization) is not fully understood.2,3,6 It has been ascribed to a reduction of dopaminergic dosage often allowed by the motor benefits of the procedure, but its pathophysiology has remained unclear. We followed the hypothesis that an inhibition of the excess of command sent to the muscle could underlie the antidyskinetic effects of chronic STN DBS and therefore measured the calibration of force in the precision grip, a function governed mainly by the motor cortex.7,8 An overshooting of grip force in late-stage PD has been described in this condition. Patients often apply excessive force in the ondrug state when grasping to lift a small object.9 –11 This force excess is closely related to the severity of LID in PD12 and therefore is suited to objectively study the beneficial effect of STN DBS on a surrogate parameter for LID. We also tried to elucidate if direct (related to stimulation) or indirect (related to reduction of dopaminergic drugs) effects account for therapeutic effects. Patients and Methods We studied a series of 10 advanced-stage patients with PD who underwent STN DBS. All suffered from severe motor fluctuations and LID (Table). Written informed consent was given by all patients. Ten age-matched healthy controls also were included in the study, which was approved by the local ethics committee. The assessment was conducted after a 12hour overnight withdrawal of dopaminergic drugs. Patients were assessed in 2 treatment conditions before (off-drug and on-drug) and in 4 conditions 3 months after surgery (offdrug/off-stimulation, on-drug/off-stimulation, off-drug/onstimulation, and on-drug/on-stimulation). A suprathreshold dose of L-dopa was applied.4 The motor score of the Unified Parkinson’s Disease Rating Scale (UPDRS) and a dyskinesia score of 7 body regions (score range 0 –28) were rated in all conditions.5 The amplitude of motor fluctuations was defined as the difference between UPDRS motor score off-drug and on-drug both before surgery and after surgery (onstimulation). The L-dopa equivalent daily dose (LEDD) was calculated as described previously.4 The subjects grasped an object (220gm, equipped with 3-dimensional force sensors) between the thumb and index finger and lifted it 15 times at a natural pace. We focused on the peak grip force (GFPEAK), a measure with a proven sensitivity for LID.12 Mean values Table. Mean Scores of the UPDRS Motor Score and the Dyskinesias Score Before and 3 Months after Surgery Before surgery After surgery Condition Motor Score ( ⫾ SD) Dyskinesia Score ( ⫾ SD) Off drug On drug ⌬ Off-on Off drug/off stim On drug/off stim Off drug/on stim On drug/on stim ⌬ Off-on drug 50.3 ⫾ 12.8 20.4 ⫾ 6.3 29.9 ⫾ 14.4 42 ⫾ 17.2 17.7 ⫾ 7.1 19.9 ⫾ 10.6b 13.8 ⫾ 5.7c 6.1 ⫾ 7.4d 3.4 ⫾ 5.4 9.1 ⫾ 5.6 0a 2.9 ⫾ 2.9c 2.9 ⫾ 1.9 4.9 ⫾ 4.2c p ⬍ 0.001, bp ⬍ 0.01, compared with presurgery off drug condition. p ⬍ 0.01 compared with presurgery drug condition. d p ⬍ 0.01 compared with amplitude of motor fluctuations before surgery (⌬ off-on drug, on stim). a c SD ⫽ standard deviation; stim ⫽ stimulation. of GFPEAK of the thumb were calculated from the lifting trials 6 to 15 of both sides (Fig 1). Grip force was compared between groups and treatment conditions by a general linear model (SPSS 10). Significance was assumed for p values less than 0.05 after Bonferroni correction. Clinical scores and LEDD were compared using the Wilcoxon test with a p level of less than 0.01. Results The LEDD was reduced in all but 2 patients, on average by 45% ( p ⬍ 0.01). The UPDRS motor score was improved by ⫺60% off-drug and by ⫺32% ondrug. The amplitude of motor fluctuations decreased by ⫺80% (see Table). Off-dystonia was abolished in all 9 patients involved. On-phase dyskinesias of the total body were completely abolished in 1 patient, and the score was reduced in all the other patients, on average by ⫺67%. Dyskinesias tended to increase on combined challenge with L-dopa and STN DBS, but the mean score was still ⫺44% lower than it was preoperatively. Peak grip force (GFPEAK) was significantly influ- enced by treatment. After 3 months of chronic stimulation, the force overflow vanished. No excess of GFPEAK was observed any longer regardless of the state of drug or stimulation, and the grip force was equivalent to controls (see Figs 1 and 2). The presurgical excess of GFPEAK in the off-drug condition was abolished. In the on-drug state GFPEAK decreased by ⫺110%. In the combined treatment state GFPEAK was still significantly lower than before surgery (see Fig 2). A reduction of initially exaggerated GFPEAK was observed not only in the 8 patients with a reduced LEDD after surgery, but also in the 2 patients on-stimulation whose LEDD was not reduced because of remaining minor fluctuations and positive effects on mood. In on-drug condition, the peak grip force decreased by ⫺51% and by ⫺48% in the first and second patient, respectively. The same was true in the off-drug state (GFPEAK, ⫺45% and ⫺25%). Thus, grip force excess resolved, although LEDD was not decreased in either of the patients. Furthermore, GFPEAK was not correlated significantly with LEDD, neither before nor after Fig 1. Peak grip force (GFPEAK in Newtons) in a healthy control subject and a representative patient who suffered from severe L-Dopa–induced dyskinesia. Force overshooting in on-state was abolished after chronic stimulation. (thick lines) Mean; (thin lines) standard error of mean from 10 trials. Scale is identical for all conditions. PD ⫽ Parkinson’s disease; N ⫽ Newton; S ⫽ seconds. Wenzelburger et al: Desensitization by DBS of the STN in PD 241 Fig 2. Mean peak grip force (GFPEAK ⫹ SD) before and 3 months after surgery. The overshooting of force was abolished in all conditions after surgery. (single dagger) p ⬍ 0.05, (double dagger) p ⬍ 0.01 compared with on-drug state before surgery. (asterisk) p ⬍ 0.05 compared with presurgical offdrug state. (single circle) p ⬍ 0.05, (double circle) p ⬍ 0.01 compared with controls. surgery (Spearman coefficient, ⬍ 0.4; p ⬎ 0.05). The reduction of GFPEAK closely matched the effect of STN DBS on peak-dose dyskinesias. Discussion We observed a remarkable normalization of force calibration in dexterous movements induced by chronic STN DBS which was found both during on-state and off-state. The preoperative L-dopa challenge caused an overshooting of grip force in line with previous findings,11,12 but this was no longer the case in the postoperative challenge with the same dosage. The initial grip force excess vanished almost completely under chronic STN DBS. In parallel with the reduction of grip force overshoot, the LID score during a suprathreshold L-dopa challenge was reduced substantially, which was within the range expected from recent reports.2,3,5,6 These findings extend our earlier observation that grip force excess in on-state is found in parkinsonian patients with motor fluctuations only, and overshooting of force correlates with the severity of LID.12 The close relationship between LID and force regulation is now supported by similar benefits of STN DBS on both. The reduction of LID in patients with STN DBS has been ascribed to a reduction of dopaminergic dosage,3,5,6 and a direct effect of DBS has been discussed mainly for off-dystonia.5 These observations make this view unlikely. We observed substantial benefits on LID in all patients, regardless whether the dopaminergic drugs were reduced. Furthermore, the LID-suppressing effect of STN is not simply part of a general effect on 242 Annals of Neurology Vol 52 No 2 August 2002 all the parkinsonian symptoms because switching off the stimulator led to a reoccurrence of severe akinesia and rigidity but left LID and force excess unchanged. Therefore, the LID-suppressing effect is more likely to reflect a desensitizing long-term effect of chronic STN DBS on LID and force regulation that outlasts even a temporary interruption of stimulation. The development of response fluctuations is believed to be caused by the discontinuous pharmacological stimulation of the dopamine-receptors by drug treatment. In contrast, STN DBS is continuously stimulating the motor system. We propose that such continuous stimulation may explain desensitization of both LID and grip force excess. This desensitization of involuntary motor activity might be explained by plastic changes in the motor system directly related to chronic STN DBS. It is not yet clear if this desensitization takes place within the basal ganglia or the cortex. The profound changes of cortical metabolism when comparing the on-state and the off-state with functional imaging might support the latter possibility but is far from proving this interpretation. This research was supported by the Deutsche Forschungsgemeinsschaft (01KO9811/7, R.W.) and the Kompetenznetzwerk Parkinson. B.-R.Z. was on sabbatical leave from the Department of Neurology, Second Affiliated Hospital, Zhejiang University, Hangzhou, P.R. China, and was supported by the Kiel University. We thank Mrs Witt for the excellent support of this investigation, Dr Johansson and Mr Bäckström for advice concerning the grip paradigm and the SC/ZOOM software, and Dr Pohl for valuable help with the statistics. References 1. Nutt JG. Clinical pharmacology of levodopa-induced dyskinesia. Ann Neurol 2000;47(suppl 1):S160 –S164; discussion, S164 –S166. 2. Limousin P, Krack P, Pollak P, et al. Electrical stimulation of the subthalamic nucleus in advanced Parkinson’s disease. N Engl J Med 1998;339:1105–1111. 3. Krack P, Limousin P, Benabid AL, Pollak P. Chronic stimulation of subthalamic nucleus improves levodopa-induced dyskinesias in Parkinson’s disease. Lancet 1997;350:1676. 4. Krack P, Pollak P, Limousin P, et al. Subthalamic nucleus or internal pallidal stimulation in young onset Parkinson’s disease. Brain 1998;121:451– 457. 5. Krack P, Pollak P, Limousin P, et al. From off-period dystonia to peak-dose chorea. The clinical spectrum of varying subthalamic nucleus activity. Brain 1999;122:1133–1146. 6. Bejjani BP, Arnulf I, Demeret S, et al. Levodopa-induced dyskinesias in Parkinson’s disease: is sensitization reversible? Ann Neurol 2000;47:655– 658. 7. Jeannerod M. The formation of finger grip during prehension. A cortically mediated visuomotor pattern. Behav Brain Res 1986;19:99 –116. 8. Lemon RN, Johansson RS, Westling G. Corticospinal control during reach, grasp, and precision lift in man. J Neurosci 1995; 15:6145– 6156. 9. Gordon AM, Ingvarsson PE, Forssberg H. Anticipatory control of manipulative forces in Parkinson’s disease. Exp Neurol 1997; 145:477– 488. 10. Fellows SJ, Noth J, Schwarz M. Precision grip and Parkinson’s disease. Brain 1998;12:1771–1784. 11. Gordon AM, Reilmann R. Getting a grasp on research: does treatment taint testing of parkinsonian patients? Brain 1999; 122:1597–1598. 12. Wenzelburger R, Zhang BR, Pohle S, et al. Force overflow and levodopa-induced dyskinesias in Parkinson’s disease. Brain 2002;125:871– 879. Demonstration of Acute Ischemic Lesions in the Fetal Brain by Diffusion Magnetic Resonance Imaging Cristina Baldoli, MD,1 Andrea Righini, MD,2 Cecilia Parazzini, MD,2 Giuseppe Scotti, MD,1 and Fabio Triulzi, MD2 The possibility of detecting acute hypoxic-ischemic brain lesions by prenatal magnetic resonance imaging or ultrasound is low. We present a case of a fetus with a vein of Galen arteriovenous malformation in whom prenatal diffusion-weighted magnetic resonance imaging at 33 weeks of gestation clearly detected cerebral acute ischemic lesions, associated with remarkable decrease of the average apparent diffusion coefficient, whereas T2weighted imaging was still not informative. Ann Neurol 2002;52:243–246 Prenatal magnetic resonance imaging (MRI) is widely used to confirm ultrasound findings in cases of complex fetal brain malformations.1 Prenatal MRI and ultrasound also can evaluate the end-stage sequelae of hypoxic-ischemic fetal brain damage2; however, the possibility of detecting these lesions by prenatal MRI or ultrasound in the acute phase is low.3 Diffusionweighted MRI (DWI) has been shown to be very sensitive in detecting hyperacute and acute hypoxicischemic brain damage in adults4 and in premature5 or term6 –10 newborns. We present the case of a fetus with a vein of Galen arteriovenous malformation (VGAM), in whom prenatal DWI at 33 weeks of gestation clearly detected acute ischemic lesions within the brain, whereas T2-weighted imaging was still not informative. Case Report A 27-year-old pregnant woman was referred to our institution at 33 weeks of gestation because of suspected fetal VGAM at Doppler ultrasonography. A prenatal MRI study at 1.5T (Eclipse; Marconi, Cleveland, OH) using the flex From the 1Neuroradiology Departments, Università Salute e Vita IRCCS–San Raffaele and 2Istituti Clinici di Perfezionamento, Milan, Italy. Received Jan 25, 2002, and in revised form Mar 20. Accepted for publication Mar 23, 2002. Published online Jun 23, 2002, in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ana.10255 Address correspondence to Dr Righini, Radiologia e Neuroradiologia, Ospedale dei Bambini V. Buzzi, Via castelvetro 32, 20154 Milan, Italy. E-mail: firstname.lastname@example.org © 2002 Wiley-Liss, Inc. 243 body surface coil and T2-weighted 5mm-thick single-shot fast spin-echo sequences was performed. Results T2-weighted images confirmed the presence of the VGAM, which appeared as an interhemispheric flow void mass in the region of the vein of Galen, connected to the torcular. Some irregular and prominent vessels were noted around the third ventricle; they were compatible with choroidal arteries feeding the malformation. The brain parenchyma did not exhibit clear signal alterations, and there was no ventricular enlargement (Fig 1). Because hypoxic-ischemic brain lesions can be associated with VGAM, a DWI acquisition was performed in the same occasion. The DWI study was based on an echo-planar 3-axis diffusion sensitized sequence, and 5mm-thick slices were acquired at fetal brain level (matrix, 128 ⫻ 128; field of view, 30cm; b-factor, 0 –1,000sec/mm2). DWI showed areas of hyperintense abnormal signal within the parietal and occipital lobe of the right hemisphere, and there were similar findings in the left temporal lobe; these foci of high signal encompassed not only the cortex, but also part of the adjacent subcortical and periventricular white matter (see Fig 1); in these areas, the rotationally averaged apparent diffusion coefficient (trace-ADC) was remarkably reduced (average value, 0.97; standard deviation [SD], 0.038m2/msec) both for apparently normal frontal lobe areas (average value, 1.83; SD, 0.18m2/msec) or for literature data on ADC in premature neonates.11,12 At 38 weeks of gestation, a cesarean section was performed, and a baby girl was delivered. She presented with severe muscular hypotonia and signs of moderate left ventricle failure at Doppler ultrasonography (23mm end diastolic diameter and 41% ejection fraction), whereas she did not show significant polypnea or pulmonary edema at chest x-rays. Four days after birth, an MRI study showed severe diffuse atrophy of both cerebral hemispheres with drastic enlargement of cortical liquoral spaces and ex vacuo dilatation of lateral ventricles; MRI showed also severe general cortical necrosis, most of the cortex being reduced to a thin rim, and diffuse loss of white matter, which presented several areas of high T1 signal. These were probably a sign of diffuse calcification, as they were too confluent to be hemorrhagic (Fig 2). No DWI or fluid-attenuated inversion recovery sequences were performed on this occasion. These findings were compatible with the endstage sequelae of severe ischemic insults. Discussion This report shows that DWI studies of fetal brain are feasible. This case highlights the potential of DWI in detecting acute fetal hypoxic-ischemic brain damage, Fig 1. (Top row) Single-shot fast spin-echo T2-weighted axial sections from the prenatal magnetic resonance imaging study at 33 weeks of gestation, showing no clear signal alterations within brain parenchima. The presence of the a vein of Galen arteriovenous malformation (arrowheads) and of possible abnormal choroidal arteries (arrow) is clearly noticeable. (Middle row) Diffusionweighted axial sections at similar levels depicting areas of markedly abnormal hyperintense signal in the occipital and parietal lobe of the right hemisphere (arrows). Some abnormal hyperintensity is visible also in the left temporal lobe (curved arrow). Both cortical and adjacent white matter areas are affected. (Bottom row) Corresponding trace–apparent diffusion coefficient (ADC) maps showing a clear ADC decrease in the same areas (arrows) of diffusion-weighted magnetic resonance imaging signal alterations. 244 Annals of Neurology Vol 52 No 2 August 2002 Fig 2. (Top row) Two T2-weighted axial sections (top left and top middle) and 1 axial T1-weighted section (top right) from the postnatal magnetic resonance imaging (MRI) study showing diffuse severe atrophy of both hemispheres, lateral ventricles ex vacuo dilatation, extensive white matter loss, and large areas of T1 hyperintensity within it, compatible with diffuse calcifications (arrows). (Bottom row) Two sagittal T1-weighted sections from the postnatal MRI study showing diffuse drastic enlargement of cortical liquoral spaces and thinning of corpus callosum. The presence of abnormal choroidal arteries feeding the VGAM is confirmed (arrows). confirming the well-known high accuracy of DWI in acute stroke diagnosis both in adults and in children. Although we cannot provide a direct demonstration of the ischemic nature of the lesions we detected, previous reports13–15 have extensively documented that ischemic and malacic lesions are the main complications that occur in subjects with VGAM, in both the prenatal and postnatal period. Moreover, the remarkable reduction in ADC that we noticed in the lesions has been reported only in a few other brain diseases, such as encephalitis, mature abscess, and status epilepticus16; in our case, the location of the lesions, their evolution, and the clinical course were poorly compatible with these conditions. Three mechanisms have been postulated for the pathogenesis of the hypoxic-ischemic brain injuries detectable in VGAM patients15: a steal phenomenon caused by blood shunting from the normal parenchymal arteries into those feeding the malformation; parenchymal hypoperfusion secondary to intracranial venous hypertension and congestion; and multiple organ hypoperfusion and hypoxia due to congestive heart failure. One or more of these mechanisms, acting together, might have produced the brain lesions that we observed in the fetus. We believe that the lesions were caused first by a steal phenomenon and second by a venous congestion mechanism, because the degree of cardiac failure was not so clinically important after birth. For these reasons, it is likely that the lesions of the cortex and white matter were mainly ischemic (because of reduce blood flow) rather than hypoxic (because of arrival of less oxygenated blood). The postnatal MRI showed diffuse cortical atrophy and leukomalacia, well beyond the areas of initial gray and white matter DWI alteration. A possible explanation for this discrepancy is that episodes of more extensive brain hypoperfusion repeatedly occurred after the prenatal MRI. The cerebral hemodynamics of our fetus was probably unstable, and the intracranial blood flow autoregulation mechanism was weak, so the prenatal MRI provided a snapshot of still evolving and spreading damage. The end stage of the damage included signs of calcific degeneration of the residual white matter at postnatal MRI; such calcifications are known to develop within the brain of newborns who have suffered hypoxic-ischemic insults.17 The explanation for the ADC decrease observed in our case resembles what is commonly reported for acute ischemia in adult humans or animal models18,19; it is based on energy failure of the cell membrane, with development of cytotoxic edema and consequent relative decrease of the extracellular water and increase of the intracellular one, which is more restricted and less diffusible. However, water diffusion in the brain differs according to maturity: the normal averaged ADC values in the white matter of premature neonates are much higher (range, approximately 1.8 –1.9m2/ msec)11,12 than those of adults (range, approximately 0.7– 0.9m2/msec)12,20; this finding could be related to the higher water content, to the lower macromolecules concentrations, or to the poorer neuronal and glial “packing” in developing and premyelinated brain tissue.12,21 This difference might account also for the abnormal ADC variations that were observed in our case. We noticed an ADC decrease to an average value of 0.97m2/msec (SD, 0.038m2/msec); similar ADC reductions have been reported recently in the acute Baldoli et al: Diffusion MRI of Fetal Brain Ischemia 245 stage of periventricular leukomalacia affecting premature babies.5 These values are almost double those commonly reached by mean diffusivity decrease in acute ischemic lesions of adults.20 However, in this case and in those of premature babies, the approximately 50% reduction in ADC for normal areas was similar to what has been observed with ischemia in adults. We hypothesize that the same tissular features (eg, water content, macromolecules concentrations, neuronal and glial “packing”), which produce the differences in ADC between the normal developing brain and adult brain, can play a role also in producing different ADC value decreases in pathological conditions, such as ischemia. To better understand these aspects, we need to study more cases of acute fetal or premature neonatal brain ischemia by using DWI, possibly correlating the results with neuropathological findings. Previous work8,9 on timing of the ADC modifications after acute perinatal hypoxic-ischemic damage suggests that a first ADC reduction can be noticed within 6 to 72 hours after the acute event, with a possible delayed and more widespread reduction in the following days, when conventional MRI also shows signal alterations due to developed vasogenic edema. Although in our case, we present only 1 time point of the ADC change time course, we hypothesize that we imaged mostly lesions associated with the first ADC decrease, because prenatal T2-weighted images did not show significant signal alterations. Although the prenatal detection of VGAM is rare, the early diagnosis by DWI of any associated destructive brain lesions can be important in therapeutic decision making, for example, in the decision to perform therapeutic intravascular procedures. Fetal brain hypoxic-ischemic damage also can occur in several other conditions, such as twin-to-twin transfusion syndrome, abruptio placentae, preeclampsia, severe maternal anemia, and severe hypovolemia. Under these circumstances, prenatal DWI is a promising technique that can detect acute hypoxic-ischemic complications, which could influence clinical and parental decisions. References 1. Girard N, Raybaud C, Gambarelli D, et al. Fetal brain MR imaging. Magn Reson Imaging Clin N Am 2001;9:19 –56. 2. De Laveaucoupet J, Audibert F, Guis F, et al. Fetal magnetic resonance imaging (MRI) of ischemic brain injury. Prenat Diagn 2001;21:729 –736. 3. Langer B, Boudier E, Gasser B, et al. Antenatal diagnosis of brain damage in the survivor after second trimester death of a monochorionic monoamniotic co-twin. Fetal Diagn Ther 1997; 12:286 –291. 4. Warach S, Chien D, Li W, et al. Fast magnetic resonance diffusion-weighted imaging of acute human stroke. Neurology 1992;42:1717–1723. 5. Inder T, Huppi PS, Zientara GP, et al. Early detection of periventricular leukomalacia by diffusion-weighted magnetic resonance imaging techniques. J Pediatr 1999;134:631– 634. 246 Annals of Neurology Vol 52 No 2 August 2002 6. Cowan FM, Pennock JM, Hanrahan JD, et al. Early detection of cerebral infarction and hypoxic-ischemic encephalopathy in neonates using diffusion-weighted magnetic resonance imaging. Neuropediatrics 1994;25:172–175. 7. Johnson AJ, Lee BC, Lin WL. Echoplanar diffusion-weighted imaging in neonates and infants with suspected hypoxicischemic injury: correlation with patient outcome. Am J Roentgenol 1999;172:219 –226. 8. Robertson RL, Ben-Sira L, Barnes PD, et al. MR line scan diffusion weighted imaging of term neonates with perinatal brain ischemia. AJNR Am J Neuroradiol 1999;20:1658 –1670. 9. Soul JS, Robertson RL, Tzika AA, et al. Time course of changes in diffusion-weighted magnetic resonance imaging in a case of neonatal encephalopathy with defined onset and duration of hypoxic-ischemic insult. Pediatrics 2001;108:1211–1214. 10. Barkovich AJ, Westmark KD, Bedi HS, et al. Proton spectroscopy and diffusion imaging of the first day of life after perinatal asphyxia: preliminary report. AJNR Am J Neuroradiol 2001;22: 1786 –1794. 11. Huppi PS, Maier SE, Peled S, et al. Microstructural development of human newborn cerebral white matter assessed in vivo by diffusion tensor magnetic resonance imaging. Pediatr Res 1998;44:584 –590. 12. Tanner SF, Ramenghi LA, Ridgway JP, et al. Quantitative comparison of intrabrain diffusion in adults and preterm and term neonates and infants. AJR Am J Roentgenol 2000;174: 1643–1649. 13. De Koning TJ, Gooskens R, Veenhoven R, et al. Arteriovenous malformation of vein of Galen in three neonates: emphasis on associated early ischaemic brain damage. Eur J Pediatr 1997; 156:228 –229. 14. Baeziger O, Martin E, Willi U, et al. Prenatal brain atrophy due to a giant vein of Galen malformation. Neuroradiology 1993;35:105–106. 15. Brunelle F. Arteriovenous malformation of the vein of Galen in children. Pediatr Radiol 1997;27:501–513. 16. Righini A, Pierpaoli C, Alger JR, Di Chiro G. Brain parenchyma apparent diffusion coefficient alterations associated with experimental complex partial staus epilepticus. Magn Reson Imaging 1994;12:865– 871. 17. Ansari MQ, Chincanchan CA, Armstrong DL. Brain calcification in hypoxic-ischemic lesions: an autopsy review. Pediatr Neurol 1990;6:94 –101. 18. Pierpaoli C, Righini A, Linfante I, et al. Histopathologic correlates of abnormal water diffusion in cerebral ischemia: diffusion-weighted MR imaging and light and electron microsopic study. Radiology 1993;189:439 – 448. 19. Lutsep HL, Albers GW, DeCrespigny A, et al. Clinical utility of diffusion weighted magnetic resonance imaging in the assessment of ischemic stroke. Ann Neurol 1997;41:574 –580. 20. Warach S, Gaa J, Siewert B, et al. Acute human stroke studied by whole brain echo planar diffusion-weighted magnetic resonance imaging. Ann Neurol 1995;37:231–241. 21. Baratti C, Barnett AS, Pierpaoli C. Comparative MR imaging study of brain maturation in kittens with T1, T2, and the trace of the diffusion tensor. Radiology 1999;210:133–142. Selective Loss of Cholinergic Sudomotor Fibers Causes Anhidrosis in Ross Syndrome Claudia Sommer, MD,1 Thies Lindenlaub, MD,1 Detlef Zillikens, MD,2 Klaus V. Toyka, MD,1 and Markus Naumann, MD1 Ross syndrome consists of segmental hyperhidrosis with widespread anhidrosis, Adie syndrome, and areflexia. The cause of this disorder is unknown. Selective degeneration of cholinergic fibers or of neural crest–derived structures has been suggested. We present clinical and skin biopsy data of 4 patients, providing evidence of reduced cholinergic sweat gland innervation in hypohidrotic skin by morphometric analysis. These findings indicate a selective degenerative process of the cholinergic sudomotor neurons. Ann Neurol 2002;52:247–250 The pathogenesis of Ross syndrome1 (tonic pupils, areflexia, and segmental hyperhidrosis) is as yet unknown. Wide overlap between Ross syndrome, Adie syndrome, Harlequin syndrome (isolated progressive segmental hypohidrosis), and a more widespread autonomic disease has been suggested.2 Here, we present findings from 4 patients with anhidrosis and segmental hyperhidrosis and variable expression of Adie’s pupils, hyporeflexia, and pathological neurophysiological findings. Importantly, the underlying pathology was the same in all patients, with reduction of cholinergic sweat gland innervation and normal epidermal and subepidermal nerve fibers. Patients and Methods Skin biopsies were obtained from a hyperhidrotic and an anhidrotic area of the patients’ back after informed consent. After fixation in 4% paraformaldehyde, 40m frozen sections were stained with antibodies to the panneuronal From the Departments of 1Neurology and 2Dermatology, University of Würzburg, Würzburg, Germany. Received Feb 21, 2002, and in revised form Mar 18. Accepted for publication Mar 23, 2002. Published online Jun 23, 2002, in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ana.10256 Current address for Dr Lindenlaub: Neurologische Klinik, Universitätskliniken des Saarlandes, 66424 Homburg, Germany. Address correspondence to Dr Sommer, Neurologische Universitätsklinik, Josef-Schneider-Strasse 11, D-97080 Würzburg, Germany. E-mail: email@example.com marker protein gene product 9.5 (PGP 9.5, 1:800; Ultraclone, Wellow, UK) and to vasoactive intestinal peptide (VIP, 1:500; Peninsula, San Carlos, CA). Immunofluorescence was performed using Cy3 or Cy2-labeled secondary antibodies (1:100; Amersham, Arlington Heights, IL). An ABC system (Vector, Burlingame, CA) was used with the same primary antibodies to verify the staining. For analysis of skin morphology (hematoxylin and eosin stain) and immune mediators, 6m-thick frozen sections were reacted with primary antibodies to T cells (CD3, CD4, CD8, CD45RO), B cells (CD20), and macrophages (CD68) or fluorescein isothiocyanate–labeled antibodies to human immunoglobulin (Ig) G, IgM, IgA, and complement component C3 (Dako, Hamburg, Germany). The sections were viewed with a Zeiss Axiophot 2 (Zeiss, Göttingen, Germany) microscope equipped with internal z-focus and a motorized scanning table by Märzhäuser, Wetzlar, Germany. Using Image Pro Plus 4.0 software, we generated 3-dimensional reconstructions of the epidermis and the sweat glands. Epidermal innervation was quantified by counting the nerve endings per millimeter of epidermal length and by determination of the epidermal nerve area with optical densitometry. Nerve endings were required to have a visible length greater than 20m to be considered. The subepidermal nerve plexus was measured in an area of 100m ⫻ 0.3mm along the subepidermal/epidermal border. Sweat gland innervation was quantified by measuring the complete nerve fiber area associated with sweat glands. Values from the ipsilateral and contralateral side were compared using Student t test; p values of less than 0.05 were considered significant. Neurophysiological Studies and Sudomotor Testing Nerve conduction studies and somatosensory and magneticevoked potentials were performed using our standard neurophysiological techniques.3 Hyperhidrotic and anhidrotic areas were visualized using Minor’s iodine starch test. Quantitative sensory testing (QST) was performed using a Peltier device (Medoc thermal analyzer) and the methods of limits.4 Heart rate variability was measured as described previously.5 Results Case 1 A 50-year-old man reported increased sweating on the right side of the trunk and on the left face for 18 years. Both hands and feet were dry. Physical examination showed hyperhidrosis in the right axilla, in the dermatomes T6 to T10 on the right and T11 to T12 on the left, and on the left face and neck. Ankle jerks were absent. Nerve conduction studies were normal. The H-reflex and the sympathetic skin response (SSR) on both feet were absent. QST showed normal thresholds in hyperhidrotic and anhidrotic areas of the back and on both hands, but increased warm thresholds on the dorsum of the right foot. The patient was treated with 200 units of botulinum toxin (Botox威) in the hyperhidrotic area in T6 to T10, with a 50% reduction in sweat secretion within 4 days. Treatment was repeated © 2002 Wiley-Liss, Inc. 247 3 months later, with a good effect lasting for 4 months (Table 1). Case 2 A 49-year-old man had hyperhidrosis in the right T5 to T7 dermatomes and in both knees for 7 years. Hands and feet did not sweat. The left pupil was wider, unreactive to light but had a tonic reaction to convergence, consistent with Adie’s pupil. Tendon jerks were elicitable only after reinforcement; right ankle jerk was absent. Latencies of somatosensory-evoked potentials of both tibial nerves were delayed, and the amplitude of the sural nerve action potential was slightly reduced. Further nerve conduction studies and SSRs (palms) were normal. The patient was injected with 1,000 units of botulinum toxin (Dysport) in the hyperhidrotic area of the trunk with marked reduction of sweating. Case 3 A 41-year-old woman had Adie’s pupils, areflexia, and hyperhidrosis in the left lower thoracic and gluteal area, and on the right ventral thigh and knee. The rest of the body was anhidrotic. Hypoesthesia was present on both lower legs; vibration thresholds were reduced at the ankles and knees. Heart rate variability was at the lower limit of normal. Nerve conduction studies showed delayed latencies of somatosensory-evoked potentials of the right tibial nerve. Sensory nerve potentials of the sural nerve could not be elicited. SSR of the right hand was absent. The patient opted for local treatment with aluminum chloride hexahydrate, which yielded satisfactory results. Case 4 A 30-year-old man reported anisocoria since age 17 years, hyperhidrosis on the right side of the trunk and the left temporal region since age 23 years, and heat intolerance. He had Adie’s pupils, normal tendon jerks, and hyperhidrosis of the dermatomes T4 to T6 on the right and in the inguinal area on the left. Palms and soles were dry. The SSR was absent. The patient was treated with 1,000 units of botulinum toxin (Dysport) in the hyperhidrotic area of the trunk. The area of excessive sweating decreased by approximately 60% for 6 months. Skin Biopsy Results No inflammatory infiltrates and no deposition of IgG or C3 were observed around sweat glands, which themselves appeared normal. Using PGP 9.5 immunohistochemistry, we found that epidermal innervation and the subepidermal plexus were normal, whereas sweat gland innervation was reduced in the anhidrotic areas (Fig, a–d). Immunoreactivity for VIP was present in the fibers innervating the sweat glands, but not in the epidermal fibers, and was markedly reduced around sweat glands of the anhidrotic side (Fig, e and f). Morphometry showed normal epidermal and subepidermal innervation density in hyperhidrotic and anhidrotic regions. Sweat gland innervation was reduced significantly in the anhidrotic compared with the hyperhidrotic regions ( p ⬍ 0.005; Table 2). Discussion Here, we show selective loss of cholinergic sweat gland innervation in 4 patients with Ross syndrome by using the panneuronal marker PGP 9.5 and a marker for cholinergic fibers, VIP.6 Because downregulation of neuropeptides such as VIP may occur in nerve lesions before degeneration of nerve fibers occurs, we used PGP 9.5 for quantification of nerve fibers. Interestingly, epidermal innervation remained intact, well in accordance with normal QST. Denervation as the cause of anhidrosis in patients with Ross syndrome has hitherto been suggested only on theoretical grounds.1 We previously presented qualitative data in another patient with Ross syndrome, also in whom a selective reduction of PGP 9.5 immunoreactive fibers around sweat glands could be found.7 Earlier skin biopsies Table 1. Clinical and Electrophysiological Data of 4 Patients with Ross Syndrome Case Age no. (yr) Gender Duration of Symptoms Segmental Adie’s (yr) Hyperhidrosis Pupils 1 50 M 18 2 49 M 7 3 41 F 6 4 30 M 7 T6–T10 R T11 L T5–T7 R T10–T12 Gluteal L Knee L T4–T6 No Yes Yes Yes Tendon Reflexes NCS/SEP SSR Ankle jerk Tibial nerve: H-reflex Absent missing missing Ankle jerk Sural nerve: A reduced Normal missing tSEP: A reduced Areflexia Sural nerve: no response Absent tSEP: Lat delayed Normal Normal Absent Heart Rate Variability QST Normal Normal Normal ND Normal ND Normal Normal A ⫽ amplitude; Lat ⫽ latency; L ⫽ left; NCS ⫽ nerve conduction studies; ND ⫽ not done; QST ⫽ quantitative sensory testing; R ⫽ right; SSR ⫽ sympathetic skin response; tSEP ⫽ somatosensory-evoked potentials of the tibial nerves. 248 Annals of Neurology Vol 52 No 2 August 2002 Fig. Forty-micrometer-thick frozen sections from skin biopsies of Patient 2, immunofluorescence, and secondary antibody Cy3. Sections from the hyperhidrotic side are on the left (a, c, e) and from the anhidrotic side on the right (b, d, f). (a, b) Immunoreaction with antibodies to the panneuronal marker protein gene 9.5 (PGP 9.5) shows normal epidermal innervation in the hyperhidrotic (a) and in the anhidrotic (b) skin. (c, d) PGP 9.5–immunoreactive fibers innervating sweat glands are abundant on the hyperhidrotic side (c), but depleted on the anhidrotic side (d). (e, f) Immunohistochemistry for vasoactive intestinal peptide shows normal cholinergic sweat gland innervation on the hyperhidrotic side (e) and loss of this innervation on the anhidrotic side (f). Bar ⫽ 50m. from Ross syndrome patients were reported as normal; however, immunohistochemical methods to visualize axons were not used.8,9 Segmental anhidrosis and hyperhidrosis are often the presenting symptoms in patients with Ross syndrome. Of our 4 patients with these symptoms, 3 had addi- tional Adie’s pupils and 3 also had areflexia or hyporeflexia. Nerve conduction studies were abnormal in 2 patients, and SSR was absent in 3 patients. Heart rate variability was normal in 3 and at the lower level of normal in 1 patient. QST was normal in the 2 patients in whom it was performed. Our observations are con- Table 2. Morphometric Analysis of Skin Biopsies Epidermal Nerve Fibers per mm Case no. 1 2 3 4 Mean ⫾ SD a Subepidemal Nerve Fiber Area per mm2 Area of Fibers Innervating Sweat Glands per 1,000m2 sweat gland area Hyperhidrosis Anhidrosis Hyperhidrosis Anhidrosis Hyperhidrosis Anhidrosis 16.3 18.4 16.2 17.1 17.0 ⫾ 1.0 18.4 16.3 18.5 19.0 18.1 ⫾ 1.2 1266 1165 1209 1356 1249 ⫾ 82.4 1194 1261 1231 1563 1312.3 ⫾ 169.4 109.9 70.0 126.7 90.1 99.2 ⫾ 24.5 39.9 40.6 39.9 51.2 42.9a⫾ 5.5 Significant difference from contralateral side, p ⬍ 0.005. SD ⫽ standard deviation. Sommer et al: Sudomotor Fiber Loss in Ross Syndrome 249 sistent with previous reports suggesting that segmental anhidrosis is part of the spectrum of distinct autonomic disorders due to a generalized injury to autonomic and dorsal root ganglia neurons.2,10 –12 It is unknown why these particular structures are involved. Both the ciliary nerves and the sympathetic innervation of sweat glands are cholinergic. However, a selective degeneration of cholinergic fibers does not explain areflexia, which has been suggested to be caused by loss of large diameter afferent fibers.13 Shin and colleagues suggested a combined lesion in these structures due to their derivation from the neural crest.2 Pigment cells and Schwann cells also are derived from the neural crest. No Schwann cell pathology has been described in Ross syndrome so far to our knowledge, but we previously described depigmentation in a patient with Ross syndrome.7 The sympathetic innervation of sweat glands is unusual, because it is initially catecholaminergic but becomes cholinergic after interactions with the target tissue. Catecholamines are necessary to induce secretory responsiveness of the sweat glands.14 The reason for the segmental hyperhidrosis in patients with Ross syndrome has not been discovered yet. The hyperhidrosis is not likely to be efficient as a thermoregulatory measure, and during treatment symptoms of heat intolerance did not occur to any greater degree than before. The area of hyperhidrosis was always localized in the lower thoracic to upper lumbar region. Why these ganglia/fibers are spared from the degenerative process is unclear. In sudomotor fibers of the rat, muscarinic M2 receptors are inhibitory presynaptic autoreceptors.15 The same function has been postulated for these receptors in human skin.16 We thus speculate that the cholinergic fibers innervating sweat glands first lose their M2 autoreceptors and that hyperhidrosis is caused by diminished presynaptic inhibition, before further degeneration leads to anhidrosis. This research was supported by research funds of the University of Würzburg. We thank Dr C. Rose for evaluating skin pathology and B. Dekant for excellent technical help. References 1. Ross AT. Progressive selective sudomotor denervation. Neurology 1958;8:809 – 817. 2. Shin RK, Galetta SL, Ting TY, et al. Ross syndrome plus: beyond horner, Holmes-Adie, and harlequin. Neurology 2000;55: 1841–1846. 3. Naumann M, Schalke B, Klopstock T, et al. Neurological multisystem manifestation in multiple symmetric lipomatosis: a clinical and electrophysiological study. Muscle Nerve 1995;18: 693– 698. 4. Claus D, Hilz MJ, Hummer I, Neundorfer B. Methods of measurement of thermal thresholds. Acta Neurol Scand 1987;76: 288 –296. 250 Annals of Neurology Vol 52 No 2 August 2002 5. Flachenecker P, Wermuth P, Hartung H-P, Reiners K. Quantitative assessment of cardiovascular autonomic function in Guillain-Barré syndrome. Ann Neurol 1997;42:171–179. 6. Schütz B, Schäfer MK, Eiden LE, Weihe E. VIP and NPY expression during differentiation of cholinergic and noradrenergic sympathetic neurons. Ann N Y Acad Sci 1998;865:537–541. 7. Bergmann I, Dauphin M, Naumann M, et al. Selective degeneration of sudomotor fibers in Ross syndrome and successful treatment of compensatory hyperhidrosis with botulinum toxin. Muscle Nerve 1998;21:1790 –1793. 8. Bartin RH, Schmutz JL, Cuny JF, et al. Le syndrome de Ross. A propos d’une observation. Ann Dermatol Venereol 1990;117: 113–114. 9. Caparros-Lefebvre D, Hache JC, Hurtevent JF, et al. Unilateral loss of facial flushing and sweating with contralateral anhidrosis: harlequin syndrome or Adie’s syndrome? Clin Auton Res 1993; 3:239 –241. 10. Jacobson DM, Hiner BC. Asymptomatic autonomic and sweat dysfunction in patients with Adie’s syndrome. J Neuroophthalmol 1998;18:143–147. 11. Bacon PJ, Smith SE. Cardiovascular and sweating dysfunction in patients with Holmes-Adie syndrome. J Neurol Neurosurg Psychiatry 1993;56:1096 –1102. 12. Drummond PD, Lance JW. Site of autonomic deficit in harlequin syndrome: local autonomic failure affecting the arm and the face. Ann Neurol 1993;34:814 – 819. 13. Pavesi G, Macaluso GM, Medici D, et al. On the cause of tendon areflexia in the Holmes-Adie syndrome. Electromyogr Clin Neurophysiol 1994;34:111–115. 14. Tian H, Habecker B, Guidry G, et al. Catecholamines are required for the acquisition of secretory responsiveness by sweat glands. J Neurosci 2000;20:7362–7369. 15. Haberberger RV, Bodenbenner M. Immunohistochemical localization of muscarinic receptors (M2) in the rat skin. Cell Tissue Res 2000;300:389 –396. 16. Cavanah DK, Casale TB. Cutaneous responses to anticholinergics: evidence for muscarinic receptor subtype participation. J Allergy Clin Immunol 1991;87:971–976. Normokalemic Periodic Paralysis Revisited: Does It Exist? Patrick F. Chinnery, PhD, MRCP,1 Timothy J. Walls, MD, FRCP,1 Michael G. Hanna, MD, MRCP,2 David Bates, MA, FRCP,1 and Peter R. W. Fawcett, BSc, FRCP3 of normoKPP confirmed earlier suspicions that these families actually had a variant of hyperKPP due to a SCN4A gene mutation.1,5 However, normoKPP remains enshrined in standard medical textbooks largely because the original detailed clinical description of the disorder by Poskanzer and Kerr2 in a family from the northeast of England remains unchallenged.6 This family recently became the subject of study again when a member of the next generation developed periodic paralysis, allowing us to revisit the original diagnosis. Case Report Normokalemic periodic paralysis (normoKPP) is well established in the literature, but there are doubts as to whether it exists as a discrete entity. Retrospective clinical and molecular analysis has confirmed suspicions that most normoKPP families actually have a variant of hyperkalemic periodic paralysis (hyperKPP) due to a mutation of the muscle-specific sodium channel gene (SCN4A). However, the original normoKPP family described by Poskanzer and Kerr (Poskanzer DC, Kerr DNS. A third type of periodic paralysis, with normokalemia and favourable response to sodium chloride. Am J Med 1961;31:328 –342) has remained unchallenged. We identified the Met1592Val mutation of SCN4A in an affected descendent of this original normoKPP family. This is the final piece in the puzzle: normoKPP is actually a variant of hyperKPP and is not a distinct disorder. Ann Neurol 2002;52:251–252 A 22-year-old man presented to the neurology department of Royal Victoria Infirmary, Newcastle Upon Tyne, because of an increase in the frequency and severity of his periodic muscle weakness. His symptoms began at 2 years of age. He experienced episodes of muscle weakness in all four limbs, sometimes associated with painful muscle stiffness. The episodes typically were precipitated by cold and damp weather and would begin during a period of rest after a period of strenuous exercise. There was no relationship with fasting or food intake. A severe attack typically would last for several days, followed by a prolonged recovery phase taking days to weeks. Two attacks had been so severe that he remained in bed and could only lift his head off the pillow. Although he occasionally experienced dysphagia during an attack, there had never been any respiratory symptoms, and he had not noticed any pigmenturia. His medical history included a camptodactyly correction to both hands but was otherwise unremarkable. There was no history to suggest cardiac dysrhythmias. Traditionally, three categories of periodic paralysis have been described: hypokalemic (hypoKPP), hyperkalemic (hyperKPP), and normokalemic (normoKPP).1 The last decade has seen major advances in our understanding of the molecular and electrophysiological basis of 2 of these groups. Three point mutations in the skeletal muscle dihydropyridine sensitive calcium channel (CACLN1A3) are responsible for most cases of hypoKPP, and mutations in the muscle-specific sodium channel gene (SCN4A) have been identified in a spectrum of disorders that includes hyperKPP, paramyotonia congenita, and potassium aggravated myotonia.1 Despite its prominence in standard texts, normoKPP has been reported only in a few families.2– 4 Clinical and molecular reanalysis of some of the original cases Results There were no abnormal findings on systemic or neurological examination. Routine hematological and biochemical tests were normal (including serum electrolytes and thyroid function tests), with the exception of an elevated alanine transaminase (53 U/L, normal ⬍45) and an elevated creatine kinase (807 U/L, normal ⬍190). A 12-lead electrocardiogram and chest x-ray were normal. A neurophysiological study showed normal peripheral nerve function, but several abnormalities on concentric needle electromyography. The most striking feature was the frequent myotonic discharges seen in every muscle sampled (Fig). There was also excess insertional activity and spontaneous activity in the form of fibrillation potentials and positive sharp waves recorded from biceps brachii, indicating minimal muscle fiber degeneration and early myopathic changes. Quantitative multi–motor unit potential analysis of the motor unit potentials in biceps showed normal motor unit potential durations, amplitudes, and number. From the 1Department of Neurology, University of Newcastle Upon Tyne, Newcastle Upon Tyne; 2Department of Neurology, Institute of Neurology, London; and 3Department of Clinical Neurophysiology, Royal Victoria Infirmary, Newcastle Upon Tyne, United Kingdom. Received Feb 15, 2002, and in revised form Mar 19, 2002. Accepted for publication Mar 25, 2002. Published online Jun 23, 2002, in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ana.10257 Address correspondence to Dr Chinnery, Department of Neurology, The Medical School, Framlington Place, Newcastle Upon Tyne, NE2 4HH, United Kingdom. E-mail: P.F.Chinnery@ncl.ac.uk. Family History The patient in this index case is the son of case V:10 who was studied in detail by Poskanzer and Kerr.2 Molecular Genetic Studies The typical pattern of the episodes of weakness, the muscle stiffness, and the prominent myotonic dis- © 2002 Wiley-Liss, Inc. 251 Fig. Continuous raster display of the concentric needle electromyography findings in the right biceps brachii of the index case. E ⫽ end of the myotonic discharge; I ⫽ insertion activity; O ⫽ onset of a myotonic discharge. Scale bar ⫽ 0.1mV per 100 milliseconds (ms). charges on electromyography pointed toward a muscle sodium channel disorder.1 Molecular genetic analysis by a standard polymerase chain reaction/restriction enzyme analysis identified the Met1592Val mutation of SCN4A in the case described above. Discussion We have shown that the original normoKPP family studied by Poskanzer and Kerr2 harbor a point mutation in SCN4A that is found in approximately 30% of families with hyperKPP.7 Molecular analysis of other reported normoKPP families also has shown mutations in SCN4A (Thr704Met).5 Based on these findings, our conclusion is that it is highly likely that normoKPP is actually a variant of hyperKPP, and it should not be considered as a distinct disease entity. In retrospect, the clinical and biochemical features of the original family of Poskanzer and Kerr are consistent with a diagnosis of hyperKPP.2 Twenty-one individuals were reported in that study, each developing symptoms in their first decade. The individuals experienced attacks of weakness every 1 to 3 months, each lasting from 2 days to 3 weeks. The attacks were provoked by rest after exertion, cold and damp conditions, and alcohol, particularly the local beer, which contained a high concentration of potassium. Oral potassium chloride exacerbated the attacks of muscle weakness in 4 members of the original family. Potassium sensitivity, the hallmark of muscle sodium channel disorders,1 therefore was clearly described in the original article.2 Although no significant change in serum potassium was ever documented during numerous spontaneous or precipitated attacks of weakness, it is now well recognized that this can occur in patients with hyperKPP.1,7 Perhaps the most interesting feature of this family is that before they received medical attention, they realized that a high intake of table salt reduced the frequency and severity of the attacks of muscle weakness. 252 Annals of Neurology Vol 52 No 2 August 2002 The muscle weakness in patients with hyperKPP is caused by muscle depolarization. In families with Met1592Val, this arises through impaired fast inactivation and slow inactivation of the sodium channel.8,9 Poskanzer and Kerr2 documented an objective improvement in muscle strength in 5 individuals who were given up to 1,750ml of triple normal saline over a 12-hour period. The mechanism responsible for this improvement is unclear, and neither the weakness nor the infusions were associated with a change in serum Na⫹ or K⫹ levels. Although no control experiments were conducted, this observation requires further investigation because of the possibility that an increased sodium load may also benefit other patients with hyperKPP. The issue of treatment is highly pertinent to our index case. Our patient’s father (V:10 in the original study2) now has generalized muscle wasting and a severe fixed vacuolar myopathy limiting normal daily activities. His son is currently not weak between attacks but has myopathic features on electromyography. Our objective is to prevent clinical progression in our new patient, and, although unsubstantiated at present, it is logical to take steps to reduce the frequency and severity of attacks. This could be achieved by alterations in his behavior and diet, or using a carbonic anhydrase inhibitor such as dichlorphenamide.10 We are very grateful to Dr D. N. S. Kerr for his comments on the original study and on this article. References 1. Ptacek LJ, Bendahhou S. Ion channel disorders of muscle. In: Karpati G, Hilton-Jones D, Griggs RC, eds. Disorders of voluntary muscle. 7th ed. Cambridge: Cambridge University Press, 2001:604 – 635. 2. Poskanzer DC, Kerr DNS. A third type of periodic paralysis, with normokalemia and favourable response to sodium chloride. Am J Med 1961;31:328 –342. 3. Mayers KR, Gilden DH, Rinaldi CF, Hansen JL. Periodic muscle weakness, normokalemia, and tubular aggregates. Neurology 1972;22:269 –279. 4. Danowski TS, Fisher ER, Vidalon C, et al. Clinical and ultrastructural observations in a kindred with normo-hyperkalemic periodic paralysis. J Med Genet 1975;12:20 –28. 5. Lehmann-Horn F, Rudel R, Ricker K. Workshop report. Nondystrophic myotonias and periodic paralyses. Neuromuscul Disord 1993;3:161–168. 6. Rudel R, Lehmann-Horn F. Muscle sodium channel and chloride channel diseases. In: Lane RJ, ed. Handbook of muscle disease. New York: Marcel Dekker, 1996:348. 7. Cannon SC. From mutation to myotonia in sodium channel disorders. Neuromuscul Disord 1997;7:241–249. 8. Cannon SC, Brown RH Jr, Corey DP. A sodium channel defect in hyperkalemic periodic paralysis: potassium-induced failure of inactivation. Neuron 1991;6:619 – 626. 9. Hayward LJ, Sandoval GM, Cannon SC. Defective slow inactivation of sodium channels contributes to familial periodic paralysis. Neurology 1999;52:1447–1453. 10. Tawil R, McDermott MP, Brown R Jr, et al. Randomized trails of dichlorphenamide in the periodic paralyses. Ann Neurol 2000;47:46 –53. Human Antibodies against Amyloid ␤ Peptide: A Potential Treatment for Alzheimer’s Disease Richard Dodel, MD,1 Harald Hampel, MD,2 Candan Depboylu, MD,1 Suizhen Lin, MD,3 Feng Gao, MD,3 Sabine Schock, MD,1 Steffi Jäckel, MD,1 Xing Wei, MD,3 Katharina Buerger, MD,2 Christine Höft, BSc, 1 Bernhard Hemmer, MD,1 Hans-Jürgen Möller, MD,2 Martin Farlow, MD,3 Wolfgang H. Oertel, MD,1 Norbert Sommer, MD,1 and Yansheng Du, PhD3 Naturally occurring antibodies directed against ␤-amyloid (A␤) were detected in intravenous immunoglobulin preparations. After intravenous immunoglobulin treatment in patients with different neurological diseases, total A␤ and A␤1-42 in the cerebrospinal fluid was reduced significantly compared with baseline values. In the serum, total A␤ levels increased after intravenous immunoglobulin treatment, whereas no significant change was observed in A␤1-42 levels. Antibodies against A␤ were found to be increased in the serum and cerebrospinal fluid after intravenous immunoglobulin treatment. This study provides evidence that intravenous immunoglobulin or purified A␤ antibodies may modify A␤ and A␤1-42 levels, suggesting potential utility as a therapy for Alzheimer disease. Ann Neurol 2002;52:253–256 The pathological hallmarks of Alzheimer’s disease (AD) are the occurrence of plaques in the neural parenchyma and the formation of neuronal tangles.1 ␤-Amyloid (A␤), a heterogenous 39 to 42–amino acid peptide, is the main constituent of senile plaques and cerebrovascular amyloid deposits. The origin of the A␤ deposited in vasculature and human brain is uncertain. According to the neuronal theory, A␤ is locally produced in the brain.2 In contrast, the vascular theory proposes that A␤ originates from the circulation, and that circulating soluble A␤ could contribute to neurotoxicity if it crosses the blood-brain barrier.3 Production of A␤ via amyloid precursor protein (APP) processing, however, is not the only factor that can influence the probability of A␤ brain deposition. Evidence has accumulated indicating that factors influencing A␤ catabolism, clearance4 and aggregation,5 are also critical in regulating A␤ metabolism.6 Recent data from transgenic mouse models of AD suggest that clearance via immune-mediated pathways may have a major impact on the development of plaques.7,8 Immunization against A␤ has prevented subsequent deposition of amyloid plaques. Furthermore, supportive data have shown that passive immunization with antibodies directed against A␤ also prevents amyloid deposition, ameliorates behavioral deterioration, and may even clear existing plaques.9,10 We hypothesized that if these results derived from animal experiments are transferable to humans, an immune-mediated A␤ degrading pathway may be physiologically present and its actions clinically significant in humans. Recently, we detected naturally occurring human antibodies against A␤ in both the cerebrospinal fluid (CSF) and serum of healthy subjects.11 These antibodies specifically recognize A␤. Furthermore, we detected significantly lower CSF titers of these anti–A␤ antibodies in AD patients compared with controls. Following these results, we hypothesized that a treatment using these naturally occurring antibodies might be beneficial as a therapeutic strategy for AD patients. Because these antibodies are predominantly present in the immunoglobulin G (IgG) fraction, we investigated whether they are detectable in commercially available IgG products. In addition, we investigated whether the administration of a selected IgG product (intravenous immunoglobulin [IVIG]) affects human CSF and serum A␤ levels. Patients and Methods From the 1Department of Neurology, Philipps University Marburg; 2 Dementia Research Section and Memory Clinic, Geriatric Psychiatry Brand and Alzheimer Memorial Center, Ludwig-Maximilian University, Munich, Germany; and 3Department of Neurology, Indiana University School of Medicine, Indianapolis, IN. Received Dec 11, 2001, and in revised form Mar 19, 2002. Accepted for publication Mar 23, 2002. Published online Jun 23, 2002, in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/ana.10253 Address correspondence to Dr Dodel, Department of Neurology, Philipps University, Rudolf-Bultmann Strasse 8, 35039 Marburg, Germany. E-mail: firstname.lastname@example.org or Dr Du, Department of Neurology, Indiana University School of Medicine, Indianapolis, IN 46202. E-mail: email@example.com Patients treated with IVIG were recruited at the Department of Neurology, Philipps University, Marburg. Each patient received IVIG (Octagam; Octapharma, Langenfeld, Germany) at a total dose of 0.4gm/kg body weight on 3 consecutive days. Seven patients (4 male, 3 female; mean age, 62.7 ⫾ 7.0 years) were included. During their regular evaluation and treatment, CSF and blood samples were withdrawn before infusion of IVIG and at the indicated times after the infusions. We took CSF samples in the morning at the L3/L4 or L4/L5 interspace by lumbar puncture after obtaining appropriate informed consent. We determined CSF leukocyte count and protein levels by standard methods. No contamination by erythrocytes was seen in any of the samples. Aliquots then were stored at ⫺80°C until biochemical analysis. Albumin and IgG concentrations were determined in the se- © 2002 Wiley-Liss, Inc. 253 rum, and the CSF by immunoprecipitation nephelometry. Only patients with an intact blood-brain barrier and regular CSF protein concentrations according to our laboratory reference were included (Table 1). The A␤ antibody enzyme-linked immunoadsorbent assay (ELISA), the purification of A␤ antibodies, and immunoprecipitation of A␤ were performed as described previously.11 The concentration of A␤ antibodies in commercially available IVIG Flebogamma, (Grifols, Langen), Germany; and Octagam; Octapharma) was determined using the A␤ antibody ELISA.11 Mean levels of anti-A␤ titers were compared by 2-way analysis of variance. A log transformation was conducted before statistical analysis to reduce skewness within each patient group.12 The mean log antibody titers for each patient group were back-transformed to give geometric mean and standard errors. Results We purified the anti–A␤ antibodies from IVIG preparations using affinity chromatography as previously described.11 A marked decrease in anti–A␤ antibody titer from the IgG fraction was observed when the flowthrough was tested in the ELISA assay (data not shown). We confirmed the ability of the affinitypurified anti–A␤ antibody from plasma to bind A␤ by immunoprecipitation of synthetic A␤ peptide (Fig 1A). Moreover, the affinity-purified anti–A␤ antibody from human IVIG readily detected a 4kDa A␤ peptide on Western blots prepared from PDAPP mouse13 hippocampal homogenates but not from cerebellar homogenates (see Fig 1B). In the next set of experiments, we incubated overnight different commercially available IVIG preparations (Flebogamma, [Grifols] and Octagam; Octapharma) with A␤1-40 (0.4gm/ml; Fig 2). Thereafter, IVIG was analyzed for anti–A␤ antibodies using ELISA. In both preparations, we could detect antibodies against A␤. In both IVIG preparations, we could detect a considerable decrease of the ELISA signal after incubation with A␤1-40 and agarose A. To evaluate our hypothesis that these antibodies may have an effect on A␤- peptide levels (total A␤ and A␤1Table 1. Clinical Data of the Patients Included in this Study Patient no. Age (yr) Gender Diagnosis QA1b 1 2 3 4 5 6 7 67 47 61 67 70 34 78 F F F M M M M MS MS PNP unknown origin PNP unknown origin PNP unknown origin Lambert-Eaton syndrome Dermatomyositis 4.5 5.9 3.5 4.3 6.3 7.9 7.8 None of the patients showed disturbance of the blood-brain barrier (QALb ⬍ 8; QALb ⫽ CSFALb/SerumALb).15 MS ⫽ Multiple Sclerosis; PNP ⫽ Polyneuropathy. 254 Annals of Neurology Vol 52 No 2 August 2002 Fig 1. Purification and characterization of human anti–A␤ antibody from human intravenous immunoglobulin (IVIG) samples (Octagam; Octapharma). (A)Anti–A␤ antibody after elution from the A␤ affinity column. We passed IVIG through an affinity sepharose column conjugated with the A␤1-40 peptide. Anti–A␤ antibody from the purified human plasma IgG was recovered after elution with buffer 1 (pH 2.5) followed by buffer 2 (pH 1.5). A␤1-40 was immunoprecipitated by affinity-purified human anti–A␤ antibody. (B) A␤ from PDAPP mouse hippocampal homogenates was immunoprecipitated by purified human anti–A␤ antibody and detected by monoclonal antibodies to A␤.13 A␤ ⫽ amyloid ␤; crb ⫽ cerebellum; hanti-A␤ ⫽ human purified A␤ antibody; hippo ⫽ hippocampus; IgG ⫽ immunoglobulin G (IVIG flowthrough). 42), we investigated the effects of IVIG administration on CSF and serum A␤ levels in patients with different neurological diseases. We detected a significant decrease in CSF levels of total A␤ and A␤1-42 after IVIG (Table 2). Similar to the recent results in PDAPP mice,14 a significant increase in total A␤ concentration was observed in the serum after IVIG treatment (see Table 2). Although there was a trend toward an increase of A␤1-42 in the serum, values did not reach statistical difference ( p ⫽ 0.06). The concentration of Fig 2. Human anti–A␤ antibody was preabsorbed with A␤ peptide. Experiments were performed using intravenous immunoglobulin (IVIG) before and after incubation with A␤1-40. Briefly, 100l of IVIG was incubated (overnight, 4°C) with protein A–agarose or A␤1-40 at the designated concentration. Protein A–agarose (Sigma, St. Louis, MO) was added to IVIG (1l, overnight, 4°C) and removed by low-speed centrifugation. The IVIG then was analyzed for anti–A␤ antibody using enzyme-linked immunoadsorbent assay. Antibody was recovered by incubation with eluting buffer (pH 1.5). This experiment was repeated 3 times with similar results. A␤ ⫽ amyloid ␤; Alp ⫽ Flebogamma, (Grifols); Oct ⫽ Octagam (Octapharma). A␤-antibodies in the serum increased after treatment with IVIG. Discussion On the basis of the results by Schenk and colleagues,7 we identified specific anti–A␤ antibodies (IgG) in both the serum and the CSF from nonimmunized humans, which may act in an immune-mediated A␤ clearance pathway.11 In an earlier study, human antibodies reactive with A␤ were isolated and cloned from human B-cell lines from AD patients; however, the role of these antibodies in AD pathogenesis remains unclear.6 In our study, we detected a significant difference in the amount of A␤ antibodies in AD patients compared with controls.11 These observations led us to ask whether these anti–A␤ antibodies are detectable in commercially available human IVIG preparations. After purification of these antibodies from IVIG, we found that they specifically recognize A␤. Furthermore, we investigated the effect of IVIG treatment on A␤, A␤1-42 levels in serum and CSF. Intravenous immunoglobulin is an accepted and routinely used treatment for several neurological and nonneurological immunemediated disorders.17 These patients are treated with IVIG on a routine basis in our department. Although we are aware that the patient selection for this study has several limitations (eg, age, immune status, different diagnoses), this was the most straightforward and feasible approach to evaluate our hypothesis. Treatment of IVIG resulted in a significant decrease of total A␤ and A␤1-42 in the CSF compared with baseline. Mean A␤ antibody concentration increased in the CSF. In contrast, the serum total A␤ and A␤ antibody concentration increased, but A␤1-42 remained unchanged. A␤ antibody concentration increased in the serum. These findings suggest that A␤ peptides may pass from the CSF to the blood and probably are metabolized locally. Recently, transporters at the blood-brain barrier have been reported, which control the central and peripheral exchange of smaller peptides including Table 2. Total A␤, A␤1– 42, and A␤. Antibody Concentrations in the CSF and Serum at Baseline and after IVIG CSF Total A␤ (pg/ml) Mean SEM A␤1–42(pg/ml) Mean SEM A␤ antibody (ng/ml) Mean SEM Serum Baseline After IVIG 1900.3 125.4 1545.7 290.1 a 209.5 37.4 159.1 45.0 a 2.5 3.3 10.7 5.1 a p Baseline 1–4 Days After IVIG 358.1 87.2 433.5 135.5 23.7 8.25 189.3 81.5 28.3 11.1 341.7 201.7 p a nsb a 7–14 Days After IVIG 426.4 123.5 p a 30.7 12.7 nsb 276.6 195.7 nsb Total A␤, A␤1– 42, and A␤-antibody concentrations in the CSF and serum were assessed at baseline and 1 to 4 days (serum), 7 to 14 days (serum), and 7 to 20 days (CSF) after IVIG (Octagam [Octapharma], 0.4 gm/kg body weight for 3 consecutive days). Values represent the mean of triplicate determinations from a single assay. a p ⬍ 0.05. Levels of significance ⬎0.05. b A␤ ⫽ ␤-amyloid; CSF ⫽ cerebrospinal fluid; IVIG ⫽ intravenous immunoglobulin; ns ⫽ not significant; SEM ⫽ standard error of the mean. Dodel et al: Human Antibodies against A␤ Peptide 255 A␤.2,3 A similar A␤ efflux from the central to the peripheral compartment has been found in PDAPP mice passively immunized with antibodies against A␤.14 Whether there is a specificity for shorter A␤ peptides to cross the blood-brain barrier, as seen in the aforementioned experiments, cannot be stated at this point. No data are available on the quantitative decrease of A␤ concentration necessary to reduce A␤ deposition. Therefore, one can only speculate whether the observed reduction may have an impact on plaque formation. Further studies, including careful dose studies in animals, are necessary. From earlier studies, however, some information can be deduced. First, a relatively modest A␤ clearance already reduced memory impairment in Tg2576 APP⫹PS1 mice.10 Second, an increase in A␤ concentration of approximately 1.5 times in familial AD patients because of mutations in the APP gene shifts the disease onset earlier by several decades. It can be assumed that already small changes in A␤ concentrations in the CSF may have an impact on A␤ deposition and plaque development.18 Our findings might have important implications for the understanding of immune-mediated clearance pathways of A␤ in humans. Moreover, our data support earlier in vitro and animal experiments that antibodies applied outside of the blood-brain barrier may serve to facilitate clearance of soluble peptides such as A␤ out of the central nervous system.2,19,20 Given the relatively large volume of distribution of the peripheral as compared with the central compartment, the addition of anti–A␤ antibodies or other antipeptide antibodies may considerably alter the clearance of biologically active peptides in the brain. Finally, further studies are warranted to investigate the role of A␤ antibodies in A␤ clearance; however, our findings suggest that IVIG-purified A␤ antibodies may be a potential therapeutic approach to AD. References 1. Selkoe DJ. Translating cell biology into therapeutic advances in Alzheimer’s disease. Nature 1999;399:A23–A31. 2. Shibata M, Yamada S, Kumar SR, et al. Clearance of Alzheimer’s amyloid-ss(1-40) peptide from brain by LDL receptorrelated protein-1 at the blood-brain barrier. J Clin Invest 2000; 106:1489 –1499. 3. Mackic JB, Stins M, McComb JG, et al. Human blood-brain barrier receptors for Alzheimer’s amyloid-beta 1- 40. Asymmetrical binding, endocytosis, and transcytosis at the apical side of brain microvascular endothelial cell monolayer. J Clin Invest 1998;102:734 –743. 256 Annals of Neurology Vol 52 No 2 August 2002 4. Ghersi-Egea JF, Strazielle N. Brain drug delivery, drug metabolism, and multidrug resistance at the choroid plexus. Microsc Res Tech 2001;52:83– 88. 5. Wisniewski T, Ghiso J, Frangione B. Biology of A beta amyloid in Alzheimer’s disease. Neurobiol Dis 1997;4:313–328. 6. Iwata N, Tsubuki S, Takaki Y, et al. Identification of the major Abeta1– 42-degrading catabolic pathway in brain parenchyma: suppression leads to biochemical and pathological deposition. Nat Med 2000;6:143–150. 7. Schenk D, Barbour R, Dunn W, et al. Immunization with amyloid-beta attenuates Alzheimer-disease-like pathology in the PDAPP mouse. Nature 1999;400:173–177. 8. Bard F, Cannon C, Burke R, et al. Peripherally administered antibodies against amyloid beta-peptide enter the central nervous system and reduce pathology in a mouse model of Alzheimer disease. Nat Med 2000;6:916 –919. 9. Janus C, Pearson J, McLaurin J, et al. A beta peptide immunization reduces behavioural impairment and plaques in a model of Alzheimer’s disease. Nature 2000;408:979 –982. 10. Morgan D, Diamond DM, Gottschall PE, et al. A beta peptide vaccination prevents memory loss in an animal model of Alzheimer’s disease. Nature 2000;408:982–985. 11. Du Y, Dodel RC, Hampel H, et al. Reduced CSF levels of amyloid beta-peptide antibody in Alzheimer’s disease. Neurology 2001;57:801– 805. 12. Armitage P, Berry G. Statistical methods in medical research. Oxford: Blackwell Scientific Publications, 1994:389 –390. 13. Games D, Adams D, Alessandrini R, et al. Alzheimer-type neuropathology in transgenic mice overexpressing V717F betaamyloid precursor protein. Nature 1995;373:523–527. 14. DeMattos RB, Bales KR, Cummins DJ, et al. Peripheral anti-A beta antibody alters central nervous system and plasma A beta clearance and decreases brain A beta burden in a mouse model of Alzheimer’s disease. Proc Natl Acad Sci USA 2001;98: 8850 – 8855. 15. Reiber H, Peter JB. Cerebrospinal fluid analysis: disease-related data patterns and evaluation programs. J Neurol Sci 2001;184: 101–122. 16. Gaskin F, Finley J, Fang Q, et al. Human antibodies reactive with beta-amyloid protein in Alzheimer’s disease. J Exp Med 1993;177:1181–1186. 17. Bril V, Allenby K, Midroni G, et al. IVIG in neurology—evidence and recommendations. Can J Neurol Sci 1999;26: 139 –152. 18. Hardy J, Allsop D. Amyloid deposition as the central event in the aetiology of Alzheimer’s disease. Trends Pharmacol Sci 1991;12:383–388. 19. Zlokovic BV, Yamada S, Holtzman D, et al. Clearance of amyloid beta-peptide from brain: transport or metabolism? Nat Med 2000;6:718 –719. 20. Strazielle N, Ghersi-Egea JF, Ghiso J, et al. In vitro evidence that beta-amyloid peptide 1– 40 diffuses across the blood brain barrier and affects its permeability. J Neuropathol Exp Neurol 2000;59:29 –38.