Sporadic fatal insomnia: A case study

BRIEF COMMUNICATIONS

Cortical Areas Differentially Involved in Multiplication and Subtraction: A Functional Magnetic Resonance Imaging Study and Correlation with a Case of Selective Acalculia Kyoung-Min Lee, MD, PhD*

A patient with an intracranial hemorrhage showed differential impairment among arithmetic types (impaired in multiplication but not in subtraction). A functional magnetic resonance imaging (fMRI) experiment using normal volunteers also revealed a differential activation between the two arithmetic types. The fMRI result could account for the selective acalculia of the patient in that the lesion shown by structural MRI included the region with multiplication-higher activation and spared subtractionhigher regions. These findings suggest that the cognitive mechanism and neural substrates differ among the simple arithmetic operations. Lee K-M. Cortical areas differentially involved in multiplication and subtraction: a functional magnetic resonance imaging study and correlation with a case of selective acalculia. Ann Neurol 2000;48:657– 661

Neuropsychological research has provided evidence that the arithmetic system can be broken down into separate components that may be selectively impaired or preserved.1–10 Recently, we encountered a patient who showed differential performance among arithmetic types; that is, impairment of multiplication with relative preservation of subtraction. Cognitive mechanisms underlying multiplication and subtraction may differ (ie, number-fact retrieval and quantity-code manipulation, respectively) and so may the neural substrates underlying such mechanisms.8 –10 Inspired by the findings of the case patient, we designed functional magnetic resonance imaging (fMRI) experiments to compare

From the *Department of Neurology, BK21 HLS Division, Neuroscience Research Institute, and Program of Cognitive Sciences, Seoul National University, Seoul, Korea. Received May 26, 1999, and in revised form Aug 2, 1999. Accepted for publication Sep 24, 1999. Address correspondence to Dr Lee, Department of Neurology, Seoul National University Hospital, 28 Yongon-Dong, ChongnoGu, Seoul, 110-744, Korea.

brain activation associated with the two arithmetic operations in an attempt to make correlation between functional neuroimaging and neuropsychology of arithmetic system in the brain. Patient and Methods Case Description The patient was a 56-year-old woman who suffered from an intracranial hemorrhage at the left parieto-temporal junction. Initially, she experienced mild difficulty in finding words, which cleared in a day or two. Neuropsychological evaluation on the 5th and 42nd days after onset found her to be within normal limits in attention, language, memory, and visuospatial functions. (The evaluation was done using the SNUH-Neuropsychology Battery that was developed from the Korean version of tests described by Mesulam.11 The test set included the modified Mini-Mental Status Examination, counting backward from 20, serial 3’s up to 40, color-word interference test, written alternating sequence test, semantic and phonemic word fluency, auditory/reading comprehension, auditory repetition and dictation, picture naming, three-word/three-shape memory test, drilled word-span test, line-bisection test, random letter cancellation test, gesture generation, body part naming, and clock drawing.) Her arithmetic performance, however, was abnormal, depending on the operation type. Multiplication was erratic (error rate, 37/81 and 22/81 on the first and second evaluations, respectively) even with pairs of small numbers, such as 3 ⫻ 2 and 2 ⫻ 5, and worse with larger numbers (Table 1). This was in contrast with subtraction, which was much better (error rate, 6/81 and 4/81 on the first and second evaluations) except for a peculiar error of consistently omitting the minus sign when the result was less than zero. Addition was normal and division was impaired to an extent comparable with multiplication (error rate, 31/81 and 25/81 on the first and second evaluations). She made no errors in choosing the larger or smaller number from a pair. She could also recall semantic knowledge associated with certain number combinations, for example, 6-2-5 for the Korean War (from the date of war breaking out, June 25, 1950) and 3-1 movement for the nationwide struggle against Japanese imperial occupation (again from the date, March 1, 1919).

Functional MRI Methods Eleven right-handed volunteers (6 men and 5 women between 25 and 35 years old) with college-level education participated with informed consent. Two single-digit numbers were presented with one operation symbol, ⫻ or ⫺, between them. The same set of number pairs was used twice, once for multiplication and once for subtraction, with presentation sequence randomized. Tests of an arithmetic type were presented at 1-per-second speed in a 30-second block. A run was composed of 10 blocks of alternating arithmetic types. Subjects were instructed to silently say the results to minimize head-motion frequently induced by speaking out loud. Although the patient showed abnormal performance in division as well, we decided to compare multiplication and subtraction only, for the following reasons: (1) division seems to be performed using similar mechanism as multiplication10; and (2) addition might be carried out using number-fact re-

Copyright © 2000 by the American Neurological Association

657

Table 1. Arithmetic Performance of the Patient 2nd 1st Subtraction 1 2 3 4 5 6 7 8 9 Multiplication 1 2 3 4 5 6 7 8 9

1

2

3

4

5

6

7

8

9

X F F F F F F F F

F F F F F F F F X

ΠΠF F F F F F F

ΠΠΠX F F X F X

ΠΠΠΠF F F F F

ΠΠΠΠΠF F F F

ΠΠΠΠΠΠF F F

ΠΠΠΠΠΠΠX F

ΠΠΠΠΠΠΠΠF

F X F X F F F F F

F F X F F F F F F

F X F F F X F X X

F X F F F X X X X

F X X F F F F X X

F F F F F X X X X

F F F X X X F X X

F X X X X X X X F

F F X X X X X F F

1st, 2nd ⫽ the first and second operands of arithmetic. F ⫽ correct; X ⫽ incorrect; Œ ⫽ correct magnitude without a minus sign.

trieval as well as quantity-code manipulation, whereas subtraction seems to use only the latter.10 A separate group of 7 subjects (4 men and 3 women, 25–33 years old) participated in two control experiments using the same setting as the main experiment. First, comparison was made between multiplication pairs whose results were relatively larger (⬎20) and those whose results were smaller (ⱕ20). Second, subtraction pairs yielding positive numbers were compared with those yielding negative ones. These controls were aimed at addressing issues related to differences between conditions in the main experiment other than the difference of arithmetic types (see later). The parameters for MR image acquisition were EPI gradient echo sequence in a GE 1.5T scanner, TR 3,000 msec, TE 60 msec, FA 90 degrees, matrix size 64⫻64, FOV 24⫻24 cm, and 20 slices 5 mm thick without separation parallel to the intercommissural plane. The images were aligned using AIR (Automated Image Registration, version 3.0) package12 and were smoothed and normalized according to Talairach and Tournoux coordinates13 using SPM96 (Statistical Parametric Mapping, 1996 version). The statistical analysis was then performed with SPM96 on pooled data by setting up a contrast between the two conditions compared.14 The resulting z-maps were then thresholded by the criteria of z-score height ( p ⬍ 0.001) and cluster size ( p ⬍ 0.05).14

Results Blood oxygen–level dependent MR signal was significantly higher during multiplication task than subtraction at the border between angular gyrus (BA39) and supramarginal gyrus (BA40) of the left hemisphere (AG/SMG) and the superior frontal gyrus of the same

658

Annals of Neurology

Vol 48

No 4

October 2000

hemisphere. Smaller areas were also found in the same comparison at the lingual and precentral gyri in the right hemisphere and anterior cingulate and precuneus in the midline. In contrast, the signal was significantly higher during subtraction than multiplication at areas in the intraparietal sulcus, superior and inferior frontal gyri, and posterior inferior temporal gyri of bilateral hemispheres (Table 2). Examination of the location of activation foci in comparison with that of the brain lesion found in the patient revealed a close match between the two (Fig). The multiplication-higher activation at left AG/SMG was included in the region with abnormally increased signal in T2-weighted MRI (see panels with z of 30.0 and 37.5 mm), whereas neighboring subtraction-higher activation fell just outside of the abnormality (see panel with z of 45.0 mm). In the two control experiments (ie, comparisons between larger and smaller multiplication pairs and between positive and negative subtraction results), no brain areas were found that showed significant difference between the conditions compared. Discussion Selective impairment of multiplication with relative preservation of subtraction has previously been described.3 The lesion of the case (Figs 1 and 2 of Lampl and colleagues3) appears to include the AG/SMG activation found in our fMRI experiment. Another case with parietal damage (Patient MAR in the report of

Table 2. Areas with Significant Difference between Subtraction and Multiplication Talairach Coordinatesa (mm) Cortical Area Multiplication ⬎ subtraction Left AG/SMG Superior frontal gyrus Right Lingual gyrus Precentral gyrus Midline Anterior cingulate gyrus Precuneus Subtraction ⬎ multiplication Left Intraparietal sulcus Superior frontal gyrus Inferior frontal gyrus Posterior ITG Right Intraparietal sulcus Superior frontal gyrus Inferior frontal gyrus Posterior ITG

z Score

Cluster Sizea

31.0 44.2

7.22 6.61

147 142

⫺74.9 ⫺9.1

5.3 23.6

7.38 4.74

75 24

49.5 ⫺44.5

15.7 49.1

6.95 6.79

82 59

BA

x

y

39/40 9/46

⫺49.7 ⫺23.3

⫺54.5 33.2

18 4

11.1 51.3

30 7

2.15 ⫺3.3

z

7/40 6 44 37

⫺31.5 ⫺24.4 ⫺47.0 ⫺55.8

⫺52.1 1.4 8.7 ⫺58.0

49.9 56.2 29.3 ⫺8.7

8.44 7.97 7.43 7.08

330 59 86 27

7/40 6 44 37

28.7 24.2 42.5 46.6

⫺54.6 1.8 10.8 ⫺57.3

52.4 54.8 30.8 ⫺1.8

8.29 7.43 7.82 5.07

368 68 130 42

Only those clusters with maximum z score higher than 3.09 ( p ⬍ 0.001, corrected) and cluster size greater than 15 voxels are listed. a

Talairach and Tournoux coordinates12 of the z score–weighted centroid of each cluster.

BA ⫽ Brodmann area; AG/SMG ⫽ the border area between angular gyrus and supramarginal gyrus; ITG ⫽ inferior temporal gyrus.

Cohen and Dehaene9) showed an opposite pattern (ie, more impaired with subtraction than multiplication). The lesion appears to extend more superiorly (Fig 2 of Cohen and Dehaene9), and the multiplication was also abnormal (error rate 22/81). The lesion in that case, we suspect, might include the intraparietal sulcus as well as supramarginal and angular gyri. A few questions were raised in interpreting the main result of the current fMRI data. First, answers are bigger for multiplication, mostly two-digit numbers, requiring more verbal load in articulating the answer, which might have caused the higher signal in AG/ SMG. However, no difference was observed between multiplication yielding larger results and that yielding smaller results in one of the control experiments. Although this negative observation cannot establish that the brain activity at the region is not related to the size effect of arithmetic results, it makes it unlikely that the AG/SMG difference in MR signal is due solely to the difference in answer size. Second, subtraction entails verbal generation of “minus” in addition to the answer in half of the tested pairs. This does not seem to account for the fMRI result either, since no difference was observed in another control comparison of subtraction yielding positive versus negative results. The current results provide support for the notion that different cognitive processes underlie multiplica-

tion and subtraction; that is, number-fact retrieval and quantity-code manipulation, respectively.8 –10 Numberfact retrieval is usually triggered by verbal recitation of overlearned multiplication tables and can be considered a form of rote verbal memory. In this regard, the finding that the left AG/SMG showed higher signal during multiplication than subtraction is in keeping with previous functional neuroimaging data that implicated a nearby region as a short-term phonological store.15,16 It is also plausible that retrieval of multiplication table uses a mechanism similar to retrieval of object names, as a damage to the angular gyrus frequently results in anomic aphasia. In the patient, however, initial anomia had already been resolved when she showed impaired retrieval of multiplication table, suggesting dissociation between the two processes. Quantity representation in the brain, on the other hand, is thought to be an analogue magnitude code on an oriented number-line subserving operations such as judgments of proximity and larger-smaller relations between numbers as well as subtraction.10 Previous neuropsychological research suggested that the magnitude code is represented by bilateral parietal regions,8 –10 which is consistent with the bilateral activation of the intraparietal sulcus in association with subtraction, as shown in the current study. A few areas also showed differential activity between

Brief Communication: Lee: fMRI Difference between Multiplication and Subtraction

659

Fig. (a) T2-weighted structural MRI of the patient is shown along with tracing of the high-signal lesion (hatched area). The brain volume was resliced parallel to the intercommissural plane, and the distance of each slice from the plane is indicated. (b) Areas with difference between multiplication and subtraction in the functional MRI experiment are shown on slices closest to those of the structural MRI. Green areas are where multiplication showed higher signal than subtraction, and red areas vice versa.

multiplication and subtraction, including those in the frontal and posterior temporal convexity and on the medial surface of the hemispheres. These may form functionally interconnected networks that are differentially involved in the two types of arithmetic. How each area of the networks contributes to the operation calls for further investigation. This work was supported by grant from the Korean Science and Engineering Foundation (no. 97-0403-04-01-3). I am grateful to Drs Kee-Hyun Chang, Jae-Kyu Roh, and Sang-Bok Lee for their generous support. I also thank Dr Kwang-Ki Kim, Soo-Hwa Lee, Chai-Youn Kim, and Na-Young Kim for their assistance in examination of the patient and fMRI experiment.

660

Annals of Neurology

Vol 48

No 4

October 2000

References 1. Sokol SM, McCloskey M, Cohen NJ, Aliminosa D. Cognitive representations and processes in arithmetic interference from the performance of brain damaged subjects. J Exp Psychol 1991;17:355–376 2. Dagenbach D, McCloskey M. The organization of arithmetic facts in memory: evidence from a brain-damaged patient. Brain Cogn 1992;20:345–366 3. Lampl Y, Eshel Y, Gilad R, Sarova-Pinhas I. Selective acalculia with sparing of the subtraction process in a patient with left parietotemporal hemorrhage. Neurology 1994;44:1759 –1761 4. Pesenti M, Seron X, Van der Linden M. Selective impairment as evidence for mental organization of arithmetical facts: BB, a case of preserved subtraction? Cortex 1994;30:661– 671 5. Delazer M, Benke T. Arithmetic facts without meaning. Cortex 1997;33:697–710 6. McNeil J, Warrington EK. A dissociation between addition and

7. 8. 9.

10.

11. 12.

13. 14.

15.

16.

subtraction with written calculation. Neuropsychologia 1994; 32:717–728 Hittmair-Delazer M, Semenza C, Denes G. Concepts and facts in calculation. Brain 1994;117:715–728 Cohen L, Dehaene S. Amnesia for arithmetic facts: a single case study. Brain Lang 1994;47:214 –232 Cohen L, Dehaene S. Cerebral pathways for calculation: double dissociation between rote verbal and quantitative knowledge of arithmetic. Cortex 1997;33:219 –250 Dehaene S, Cohen L. Levels of representation in number processing. In: Stemmer B, Whitaker HA, eds. Handbook of neurolinguistics. San Diego: Academic Press, 1998 Mesulam M-M. Principles of behavioral neurology. Philadelphia: FA Davis, 1985 Woods RP, Mazziotta JC, Cherry SR. MRI-PET registration with automated algorithm. J Comput Assist Tomogr 1993;17: 536 –546 Talairach J, Tournoux P. Co-planar stereotaxic atlas of the human brain. New York: Thieme, 1988 Friston KJ, Holmes AP, Worsley KJ, et al. Statistical parametric maps in functional imaging: a general linear approach. Hum Brain Map 1995;2:189 –210 Smith EE, Jonides J, Marshuetz C, Koeppe RA. Components of verbal working memory: evidence from neuroimaging. Proc Natl Acad Sci USA 1998;95:876 – 882 Paulesu E, Frith CD, Frackowiak RSJ. The neural correlates of the verbal component of working memory. Nature 1993;362: 342–344

Neurolisteriosis Presenting as Recurrent Transient Ischemic Attacks R. Staudinger, MD,* D. Levine, MD,* B. Swaminathan, PhD,† and D. Zagzag, MD, PhD‡

An elderly man experienced recurrent transient episodes of right arm weakness and expressive aphasia. He was initially treated with aspirin and then with coumadin. Thirteen days after initial presentation, he became febrile and had signs of meningitis. The illness progressed relentlessly to death 9 weeks after admission to the hospital. Necropsy showed prominent meningitis with vasculitis extending into the left frontal lobe. Polymerase chain reaction identified the organism as Listeria monocytogenes. Staudinger R, Levine D, Swaminathan B, Zagzag D. Neurolisteriosis presenting as recurrent transient ischemic attacks. Ann Neurol 2000;48:661– 665

Listeria monocytogenes (LM) has a predilection for the central nervous system (CNS), where involvement takes two forms (Table). Acute purulent meningitis is most common and is clinically indistinguishable from other pyogenic meningitis.1 The onset is abrupt and dramatic. Ataxia, tremor, myoclonic jerks, seizures, and fluctuating mental status may be prominent.2,3 Gram stain is often negative. Positive cultures may be confused with harmless “diphtheroids,” making diagnosis difficult. Recently, a polymerase chain reaction (PCR) assay for detection of LM in CSF samples has been developed and may greatly improve diagnosis.4 Very rarely, LM can also lead to chronic meningitis.5,6 The second, rare but well recognized form of neurolisteriosis is a parenchymal involvement of the CNS. LM encephalitis can affect the hemispheres (cerebritis) or focus on the brainstem (rhombencephalitis). Typically, the illness is biphasic with a prodrome of 5 to 15 days of malaise, headache, nausea, and fever (sepsis syndrome), followed by focal neurological deficit, suggesting either cerebral or pontomedullary involvement.7,8 Even more rarely, a brain abscess can develop, presumably having evolved from a focal encephalitis.9

From the *Departments of Neurology, and ‡Pathology and Neurosurgery, New York University, New York, NY; and †National Center for Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, GA. Received Feb 14, 2000, and in revised form May 15. Accepted for publication May 16, 2000. Address correspondence to Dr. Zagzag, Department of Pathology, New York University, 550 First Avenue, New York, NY 10016.

Copyright © 2000 by the American Neurological Association

661

Table. Central Nervous System Manifestations of Listeria monocytogenes Infection Meningitis Acute (indistinguishable from other bacteria) Chronic (exceedingly rare) Encephalitis Cerebritis Rhombencephalitis Abscess Cerebral Brainstem Spinal cord

CSF pleocytosis and elevated protein may be present, particularly in the brainstem form. CSF cultures are often negative, whereas blood cultures are positive. We report the case of a patient who presented with recurrent, persistent transient ischemic attacks (TIAs) of aphasia and right-sided weakness with no fever or meningeal signs. Two weeks later, fever developed. Meningitis with arteritis involving the leptomeningeal arteries of the left frontal convexity was suggested by gadolinium-enhanced magnetic resonance imaging (MRI) and by cerebrospinal fluid (CSF) examination and was confirmed on brain biopsy and autopsy. To our knowledge, such a presentation of listeria meningitis has not hitherto been reported. Case Report This 78-year-old ambidextrous man presented with episodes of right arm weakness and hesitant speech. The patient was well until 1 week before admission, when he experienced transient difficulty finding words, from which he recovered completely within 2 hours. Enteric-coated aspirin 325 mg daily was prescribed. He remained well for 1 week. On the day of admission, he experienced another episode of wordfinding difficulty and clumsiness of his right hand. These lasted for about 2 hours with complete recovery. The patient’s medical history was significant for hypertension and prostate carcinoma. He was afebrile, heart rate was 62 and regular, and blood pressure was 140/90 mm Hg. He appeared well and was not distressed. The neck was supple. He was alert; memory and attention were normal. Spontaneous speech, repetition, and comprehension were normal. The cranial nerves were intact. He had normal muscle tone and bulk, and there was no weakness. The deep tendon reflexes were 2⫹ and symmetrical, and the plantar responses were flexor. Sensation, coordination, and gait were normal. The hematocrit was 34.8, the white blood cell (WBC) count was 3,200 and, platelet count was 269,000. Prothrombin and partial thromboplastin time were normal, as were all of the blood chemistries. In the hospital, the patient continued to have one or two transient episodes of right hand weakness, right central facial nerve palsy, and hesitant speech per day, each lasting 2 to 4 hours. Despite satisfactory anticoagulation, the patient continued to have attacks, and residual hesitancy of speech eventually developed between attacks. On the sixth day of hos-

662

Annals of Neurology

Vol 48

No 4

October 2000

pital admission, he became febrile (38.9°C). A chest radiograph was normal, and blood cultures later turned out sterile. The attacks of speech hesitancy and right-sided weakness became more frequent. It was thought that the duration of symptoms was atypical for vascular disease, and gadolinium-enhanced brain MRI was ordered, which showed focal leptomeningeal enhancement over the left frontal convexity (Fig 1A). CSF analysis revealed 330 WBCs/␮l, with 75% polymorphonuclear leukocytes, and 240 RBCs/␮l. The total protein content was 185 mg/dl and glucose was 14 mg/ dl. Routine cultures showed no growth, and special cultures for acid-fast bacilli and fungi were later negative. The patient was treated with multiple antibiotics, including ceftriaxone, metronidazole, penicillin G, doxycycline, rifampicin, Isoniazid (INH), ethambutol, pyrazinamide, and amphotericin B without clinical benefit. Neither ampicillin nor trimethoprimsulfamethoxazole was used. Multiple lumbar punctures did not reveal an etiological event. A left frontal craniotomy with biopsy of the meninges and of the left frontal cortex showed inflammation consisting of polymorphonuclear neutrophils, lymphocytes, and plasma cells, affecting the leptomeninges and also extending along Virchow-Robin spaces (see Fig 1B). No etiological agent could be identified. The patient’s condition gradually deteriorated, and he died 9 weeks after admission.

Neuropathological Examination The inner surface of the dura over the left frontal area was focally thickened and had a yellowish discoloration (see Fig 1C). The leptomeninges were filled with grayish exudate, more prominent on the left side than on the right. On coronal sections, the leptomeninges were thickened and exudate filled the sulci (see Fig 1D). Microscopic examination of the left frontal cortex showed preserved laminar architecture. In the white matter, there were perivascular aggregates of macrophages and mononuclear cells, including plasma cells (see Fig 1E). The leptomeninges were filled with lymphocytes, plasma cells, and foamy histiocytes. A Gram stain identified gram-positive bacilli (see Fig 1F). Samples of fresh brain tissue were cultured, but cultures remained sterile.

PCR Amplification of L. monocytogenes DNA from Paraffinized Tissue DNA was recovered from paraffinized tissue as previously described10 and LM DNA was amplified by a nested two-step PCR protocol.11 The PCR products were separated by standard electrophoretic procedures and visualized by staining with ethidium bromide (Fig 2; the sample from our patient is shown in lane 10). Positive (1 ng/ml of LM DNA) and negative (Escherichia coli DNA) controls were routinely included in the PCR assays (lanes 8 and 9, respectively).

Discussion The patient’s illness was a meningitis caused by LM, and in that sense was not unusual. Meningitis is the most frequent CNS presentation of LM. The meningeal tropism is in part determined by surface characteristics of the bacteria. Compatible receptors on the

Fig 1. (A) T1-weighted gadoliniumenhanced MRI demonstrates focal left frontal meningeal enhancement (arrowheads). (B) Biopsy of the meninges and the left frontal lobe reveals a meningovasculitis extending into sulci (s). (C) The fresh postmortem specimen shows yellowish discoloration of the leptomeninges. The biopsy site is seen (B). (D) Coronal section (formalin fixed) reveals thickened leptomeninges. The exudate extends into the sulci (s). A necrotic focus can be seen (arrowhead). (E) Perivascular (V) aggregates of HAM-56 immunoreactive macrophages are seen in the white matter. (F) Gram stain shows numerous gram-positive bacilli, often in clusters.

membranes of meningeal cells and the capsular surface of LM have been postulated.8,12 Our case is most unusual because of the clinical presentation. There was no fever, and there were no signs of meningeal inflammation. Instead, the patient presented with recurrent symptoms of transient ischemia of the left frontal lobe, consisting of nonfluent aphasia and right hemiparesis. These proved to be caused by vasculitis of leptomeningeal arteries. Arteritis with transient ischemia or infarction is a common complication of acute meningitis. In such patients, angiographically documented cerebrovascular involvement is frequent.13,14 The patients are generally very severely ill. In one study,13 11 of 13 patients with bacterial meningitis and cerebrovascular complications were comatose on admission and presumably all had fever, headache, and a stiff neck. At the opposite extreme, meningitis that is exceedingly chronic may present clinically with transient cerebral ischemia or infarction in the absence of fever

and signs of meningeal irritation. Meningovascular syphilis is the most notable example. The meningitis is often completely asymptomatic, and stroke may occur as the initial complaint 6 months to 12 years after onset of the meningitis. The meninges are infiltrated by chronic inflammatory cells, which tend to aggregate around blood vessels. The vessels show chronic inflammation infiltrating the adventitia that may reach the media. There is fibrosis and thickening of the intima as well as thinning of the media with preservation of the elastica (Heubner arteritis). There is no indication that LM had caused a chronic meningitis in our patient. He had fever within 2 weeks of his presentation with TIAs, and an acute and eventually fulminant meningitis was evident on brain biopsy and later at autopsy. Neuropathological examination of the biopsy showed numerous polymorphonuclear neutrophils and mononuclear cells. Cerebral infarction is common in tuberculous meningitis and in one series was present in 41% of autop-

Brief Communication: Staudinger et al: CNS Listeria

663

Fig 2. Two-step nested PCR amplification of Listeria monocytogenes DNA from paraffinized tissue specimens. Lanes 1 and 14: Molecular size marker (123-bp ladder); lanes 3, 5, and 7 show positive signals for other patients; lane 8: positive control; lane 9: negative control; lane 10: tissue from our patient, and lanes 11 to 13 are blank lanes.

sied cases.15 Unlike the typical pyogenic meningitis, the inflammatory process in tuberculous meningitis frequently extends along pial vessels into the brain, constituting a meningoencephalitis. That was true in our case as well. However, tuberculous meningitis usually is a subacute illness marked by fever, headache, drowsiness confusion, and nuchal rigidity. Nevertheless, the initial symptoms may occasionally be those of cerebral ischemia. Rohr-LeFloch and colleagues16 reported on 2 patients presenting with transient cerebral ischemia. Their second case resembled ours. The patient was a 47-year-old man who presented with recurrent attacks of aphasia and right arm weakness or numbness, associated with headache. Three weeks later, he became slower in his thinking and spoke less. Arteriography was negative, and erythrocyte sedimentation rate was normal. CSF showed 610 leukocytes, of which 95% were lymphocytes. There was no hypoglycorrhachia. PCR was positive for Mycobacterium tuberculosis. He improved with antituberculous therapy, and the recurrent attacks of aphasia and right-sided weakness disappeared. We conclude that, on rare occasions, subacute meningoencephalitis can present with TIAs in the absence of fever or of signs of meningeal irritation. We report on the first such case with LM meningoencephalitis. In

664

Annals of Neurology

Vol 48

No 4

October 2000

a patient with unexplained persistent attacks of cerebral ischemia, lumbar puncture and gadolinium-enhanced MRI may be of use in pointing to a diagnosis of meningitis. We thank Daniella Asaro for her assistance in the preparation of this paper.

References 1. Gray ML, Killinger AH. Listeria monocytogenes and listeric infections. Bacteriol Rev 1966;30:309 –382 2. Ford PM, Herzberg L, Ford SE. Listeria monocytogenes: six cases affecting the central nervous system. Q J Med 1968;37:281– 290 3. Simpson JF. Listeria monocytogenes meningitis: an opportunistic infection. J Neurol Neurosurg Psychiatry 1971;34:657– 663 4. Jaton K, Sahli R, Bille J. Development of polymerase chain reaction assay for detection of Listeria monocytogenes in clinical cerebrospinal fluid samples. J Clin Microbiol 1992;30:1931– 1936 5. Heck AF, Hameroff SB, Hornick RB. Chronic Listeria monocytogenes meningitis and normotensive hydrocephalus. Neurol 1971;21:263–270 6. Whitty CM, Macaulay JD. Listeria monocytogenes meningoencephalitis in an adult. Br J Med 1965;1:634 – 638 7. Johnson ML, Colley E. Listeria monocytogenes encephalitis associated with corticosteroid therapy. J Clin Pathol 1969;22:465– 469

8. Uldry PA, Kuntzer T, Bogousslavsky, et al. Early symptoms and outcome of Listeria monocytogenes rhombencephalitis: 14 adult cases. J Neurol 1993;240:235–242 9. Dee RR, Lober B. Brain abscess due to Listeria monocytogenes: case report and literature review. Rev Inf Dis 1986;8:968 –977 10. Dyall-Smith M, Dyall-Smith D. Recovering DNA from pathology specimens: a new life for old tissues. Mol Biol Rep 1988; 6:1–2 11. Schuchat A, Lizano C, Broome CV, et al. Outbreak of neonatal listeriosis associated with mineral oil. Pediatr Infect Dis J 1991; 10:183–189 12. Garvey G. Current concepts of bacterial infections of the central nervous system: bacterial meningitis and bacterial brain abscess. J Neurosurg 1983;59:735–744 13. Pfister HW, Borasio GD, Dirnagl U, et al. Cerebrovascular complications of bacterial meningitis in adults. Neurology 1992;42:1497–1504 14. Pfister HW, Feiden W, Einhaupl KM. Spectrum of complications during bacterial meningitis in adults: result of a prospective clinical study. Arch Neurol 1993;50:575–581 15. Dastur DK, Lalitha VS, Udani PM, Parekh U. The brain and meninges in tuberculous meningitis: gross pathology in 100 cases and pathogenesis. Neurol 1970;18:86 –100 16. Rohr-Le Floch J, Myers P, Gauthier G. Accidents ischemique cerebraux et meningite tuberculeuse. Rev Neurol 1992;148: 779 –782

Sporadic Fatal Insomnia: A Case Study Francesco Scaravilli, MD,* Rebecca J. Cordery, MRCP,† Hans Kretzschmar, MD,‡ Pierluigi Gambetti, MD,§ Brian Brink, MBBCh,储 Vivian Fritz, FRCP,储 James Temlett, FRCP,储 Charles Kaplan, FCP(SA),储 David Fish, MD,† Shu F. An, MD,* Walter J. Schulz-Schaeffer, MD,‡ and Martin N. Rossor, MD†

A 58-year-old man died after a 27-month illness characterized by insomnia, confirmed by polysomnography. He was homozygous for methionine at codon 129 of the prion gene but had no mutation in the prion gene. Neuropathology showed thalamic and olivary atrophy and no spongiform changes. Paraffin-embedded tissue blotting demonstrated abnormal prion protein in the brain. This is the first case of the sporadic form of fatal familial insomnia with demonstration of the disorder by polysomnography. Scaravilli F, Cordery RJ, Kretzschmar H, Gambetti P, Brink B, Fritz V, Temlett J, Kaplan C, Fish D, An SF, Schulz-Schaeffer WJ, Rossor MN. Sporadic fatal insomnia: a case study. Ann Neurol 2000;48:665– 668

In 1992, the group of human prion diseases expanded to include fatal familial insomnia (FFI), an autosomal dominant disorder showing loss of ability to sleep, dysautonomia, and thalamic and olivary atrophy.1– 4 FFI is linked to a point mutation at codon 178 of the prion protein gene (PRNP) and to presence of the methionine codon at position 129 of the mutant allele.3,4 Recently, a sporadic form of fatal insomnia (sFI) has also been described.5–7 The diagnosis of FFI requires demonstration of the sleep disorder by polysomnography.8 We report the first proven case of sFI in which insomnia was proven and characterized by polysomnography. Patient and Methods A 58-year-old South African caucasian man presented with an 11-month history of discomfort in the legs and jerky

Departments of *Neuropathology and †Neurology, Institute of Neurology, University College London, London, UK; ‡Institut fu¨r Neuropathologie, Universita¨t Gu¨ttingrn, Go¨ttingen, Germany; §Division of Neuropathology, Institute of Pathology, Case Western Reserve University, Cleveland, OH; 储Neurology Unit, Department of Medicine, University of the Witwatersrand Medical School, Johannesburg, Republic of South Africa. Received Jan 28, 2000, and in revised form May 18. Accepted for publication May 19. Address correspondence to Dr Scaravilli, Department of Neuropathology, Institute of Neurology, Queen Square, London, WC1N 3BG, UK.

Copyright © 2000 by the American Neurological Association

665

movement of the legs in bed. His workload was taking longer to complete, and his sleep was increasingly impaired (up to 2 hours per night). He had undergone coronary angioplasty in 1992 with bypass surgery in 1993. There was no significant family history. On examination there was mild dysarthria, reduced blink frequency, and stooped gait with reduced arm swing. Magnetic resonance imaging (MRI) of the brain and spine, cerebrospinal fluid (CSF), elecroencephalogram (EEG), and electromyelogram (EMG) were normal. Four months later, he showed variable tremor at rest, hypomimia and dysarthria, brisk jaw jerk, brisk reflexes with equivocal plantars, and spastic gait, without fasciculation or myoclonus. Blood and CSF examination were normal, including negative oligoclonal bands. Concentration of S-100 and neuron specific enolase (NSE) was normal. Immunoassay for 14-3-3 protein9 was negative. Brain MRI showed white matter lesions, suggesting small vessel disease. ECG showed an old inferior infarct. Echocardiogram revealed good left ventricular function with mild mitral regurgitation. EEG showed alpha rhythm (9 Hz, responsive, posterior, symmetrical) alternating with episodes of moderate amplitude theta/slow activity at 4 to 6 Hz, bilaterally, maximal anteriorly. These became more prolonged with drowsiness and showed a predominantly central distribution. There were no periodic complexes. Formal polysomnography was performed (total of 16 hours, 16:46 to 08:00 AM). Concurrent finger pulse oximetry was normal. The patient settled for sleep at 22:59 and stopped trying at 6:50 AM, being asleep for 1 hour and 30 minutes (23% of the time). He tried reading and walking. On some occasions, when walking, he seemed confused; when asleep, he had almost continuous restless movements of either or both legs and feet, varying from small jerks to larger restless movements and not exhibiting the expected features of the periodic movement of sleep. There were also small jerks of the upper limbs. Sleep efficiency measures were as follows: latency to stage 2, 1 hour 9 minutes; latency to REM, 31 minutes and 20 seconds; time in light sleep, 86 minutes; time in REM sleep, 4 minutes; awakenings, more than 20 (lasting 2–162 minutes). There were three periods of REM (mean time, 1 minute; range, 1–2 minutes). The record showed severe disruption of the sleep–wake cycle, with little time spent asleep (efficiency 25%) and lack of sleep spindles and slowwave sleep. There were no clear periods of REM with atonia (possibly some brief REM-like periods with lack of atonia). Neuropsychology, 14 months after the onset, showed verbal (113) and performance (109) IQ reflecting intellectual deterioration (National Adult Reading Test [NART]: optimal level of functioning in the superior range). There was global impairment of memory function, naming, and frontal lobe tasks. The PRNP revealed no mutations in the open reading frame. The patient was homozygous for methionine at codon 129.10 A possible sporadic Creutzfeldt-Jakob disease was suspected, and a cerebral biopsy was considered but was declined. The patient became severely wasted, with generalized

666

Annals of Neurology

Vol 48

No 4

October 2000

fasciculation and further intellectual deterioration. He died 27 months after the onset of the disease. Histopathology of the CNS was carried out according to standard techniques, including tissue decontamination with formic acid. Immunohistochemistry was applied using antibodies anti-GFAP, 1/400; anti-tau, 1/150; anti-ubiquitin, 1/150; anti-A4, 1/60; and CD68, 1/150 (all Dako, UK); and anti-PrP (3F4, 1/2000 and KG9, 1/150). PrPSc detection was also performed with the paraffin-embedded tissue blot method using the monoclonal antibody Go¨ 138 to residues 138 to 152 of the human prion protein (PrP).11 The brain showed atrophy of both thalami (Fig 1). On histological examination, the cortex and deep gray nuclei showed diffuse increase of astrocytes and microglia. The thalamus, especially the anterior and dorsomedial regions (Fig 2a) and the pulvinar, as well as the olives (see Fig 2b), revealed severe nerve cell loss, diffuse gliosis (see Fig 2c), microgliosis, and presence of macrophages (see Fig 2d). In the thalamus, glial hyperplasia did not appear to correlate with the severity of neuronal loss. The cerebellum showed moderate Purkinje cell loss. No other lesions, including spongiform degeneration, were detected. Paraffin-embedded tissue blotting of PrP scrapie (PrPSc) on three sets of blocks stained with different procedures revealed mild to moderate diffuse staining in all layers of the entorhinal cortex (see Fig 2e). Granular staining was seen in the upper temporal, insular, and frontal layers, and in the cyngulate gyrus, with small foci present in the occipital cortex and the centrolateral thalamic nucleus. No PrPSc deposition was seen in the white matter or in the deep gray nuclei.

Discussion This patient presented clinical, polysomnographic, and histopathological features indistinguishable from those of FFI. Furthermore, in the absence of unfixed brain tissue for Western blotting, evidence of PrPSc in the brain was obtained with paraffin-embedded tissue blotting.11 Therefore, the present case belongs to the group of prion diseases. Up to 1998, FFI had been reported in 24 kindreds, all linked to a mutation at PRNP codon 178, resulting Fig 1. The thalamus of the patient (a) shows considerable reduction in size compared with one of a normal individual of same sex and comparable age (b) (Luxol fast blue–cresyl violet).

2. Photomicrographs of the anteromedial region of the Š Fig thalamus (a), showing reduction in number of neurons. (b) The neuronal loss in the inferior olive is virtually complete (hematoxylin and eosin, ⫻150). (c) GFAP antibody emphasises the degree of glial hyperplasia present in the thalamus. Similar appearances are seen in the olive (⫻150). (d) The inferior olive, showing increased number of microglia and presence of macrophages (CD68, ⫻240). (e) Immunoblotting shows PrPSc immunoreaction in the entorhinal cortex.

in a substitution of aspartic acid with asparagine (D178N).12 The D178N mutation is coupled to the methionine codon at position 129, the site of a common methionine/valine polymorphism. Both sexes are affected between at 20 and 72 years of age.13 Cardinal signs of the disease include loss of the ability to sleep, dysautonomia, and motor signs,8 and its duration is related to the genotype at the PRNP codon 129 of the prion gene3: methionine homozygous and methionine/ valine heterozygous subjects have mean disease duration of 12 and 21 months, respectively.12 On 24-hour video, polysomnography in homozygotes initially shows substantial absence of non-REM sleep episodes, whereas heterozygotes are less severely affected. Eventually, both groups show marked decrease of slow-wave and REM sleep, lack of spindle activity, and disruption of the sleep–wake cycle.8 All FFI subjects, regardless of the RPNP genotype at codon 129, have neuronal loss and gliosis of the thalamus and olives. Moreover, 129 homozygotes may have a normal cerebral cortex, whereas heterozygotes often show cortical spongiosis and gliosis.12 Western blotting demonstrates relatively low quantity of PrPSc, invariably of type 2, and underrepresentation of the unglycosylated form versus diglycosylated and monoglycosylated.12 A sporadic form of FFI was reported by Parchi and colleagues5,6 in 5 patients and in one by Mastrianni and others.7 All had similar clinicopathological phenotypes, and 5 had insomnia, although it was not the presenting sign in any. All were methionine homozygous at codon 129 and lacked the mutation at codon 178. Relatively low quantities of PrPSc type 2, as in FFI, were detected. However, at variance with FFI, the unglycosylated form of PrPSc was comparable in relative amounts to that seen in other forms of sporadic prion disease,6 a difference that was expected, since the underrepresentation of the unglycosylated form of PrP is related to the presence of the D178N mutation in FFI.14 Inoculating mice with extracts from the thalamus and cortex of the patient produced neurological signs and pathological lesions after 6 and 7 months, respectively.7 Polysomnography was not carried out in any of these patients.5–7 Our patient presented a phenotype similar to that of both FFI8 and sFI5–7 (Table), had no D178N muta-

Brief Communication: Scaravilli et al: Sporadic Fatal Insomnia

667

Table. Summary of the Salient Data regarding the Previous Patients with Sporadic Fatal Insomnia and the Present Sex/Age Duration (yr) (mo) Insomnia Spongiosis

Study Parchi et al6

7

Mastrianni et al Present

M/53 F/46 M/70 F/40 M/36 M/44 M/58

18 15 24 15 17 16 27

⫹ ⫹ ⫹ ⫺ ⫹ ⫹ ⫹

⫹ ⫹⫹ ⫹⫹ ⫹ ⫹ ⫹ ⫺

tion, and was similarly methionine homozygous at PRNP codon 129. However, mild neocortical spongiosis, observed in all 5 cases reported on by Parchi and associates,6 was lacking in our case and was only mild and circumscribed in the case by Mastrianni and colleagues.7 The lack of cortical spongiosis in our case indicates that the variability of severity of this lesion found in FFI is present also in sFI. Conventional immunostaining for PrPSc was positive in 3 of 5 cases described by Parchi and co-workers6 but negative in our case. However, PrPSc was demonstrated with paraffin-embedded tissue blotting. Our patient is the first with sFI in whom polysomnography has been performed, demonstrating findings indistinguishable from those observed in methionine homozygous affected by FFI8. Prion diseases have the unique characteristic of including forms with sporadic, familial, and an infectious pattern of transmission. These forms may share the same phenotype. Parchi and colleagues5,6 showed that a familial prion disease phenotype associated with malignant insomnia, dysautonomia, and thalamic atrophy may have a sporadic counterpart, as has the classical form of CJD. Mastrianni and associates7 showed that this new entity is transmissible to receptive animals. We now further confirm its similarities with the FFI by showing, for the first time, that in both conditions the sleep disorder is the same. Moreover, the detection of the present case in South Africa indicates that sporadic familial insomnia may have a worldwide distribution. We are grateful to Prof James Ironside for the donation of the antibody KG9.

References 1. Lugaresi E, Medori R, Montagna P, et al. Fatal familial insomnia and dysautonomia with selective degeneration of the thalamic nuclei. N Engl J Med 1986;315:997–1003 2. Manetto V, Medori R, Cortelli P, et al. Fatal familial insomnia. Clinical and pathologic study of five new cases. Neurology 1992;42:312–319 3. Goldfarb LG, Petersen RB, Tabaton M, et al. Fatal familial

668

Annals of Neurology

Vol 48

No 4

October 2000

4.

5.

6.

7.

8.

9.

10.

11.

12.

13. 14.

insomnia and familial Creutzfeldt-Jakob disease: disease phenotype determined by a DNA polymorphism. Science 1992;258: 806 – 808 Medori R, Tritschler HJ, LeBlanc A, et al. Fatal familial insomnia, a prion disease with a mutation at codon 178 of the prion protein gene. New Engl J Med 1992;326:444 – 449 Parchi P, Capellari S, Chin S, et al. Fatal sporadic insomnia (thalamic form of sporadic Creutzfeldt-Jakob disease). J Neuropathol Exp Neurol 1998;57:518 (Abstract) Parchi P, Capellari S, Chin S, et al. A subtype of sporadic prion disease mimicking fatal familial insomnia. Neurology 1999;52: 1757–1763 Mastrianni JA, Nixon R, Layzer R, et al. Prion protein conformation in a patient with sporadic fatal insomnia. N Engl J Med 1999;340:1630 –1638 Montagna P, Cortelli P, Avoni P, et al. Clinical features of fatal familial insomnia: phenotypic variability in relation to a polymorphism at codon 129 of the prion protein gene. Brain Pathol 1998;8:515–520 Hsich G, Kenney K, Gibbs CJ, et al. The 14-3-3 brain protein in cerebrospinal fluid as a marker for transmissible spongiform encephalopathies. N Engl J Med 1996;335:924 –930 Collinge J, Sidle KCL, Meads J, et al. Molecular analysis of prion strain variation and the aetiology of “new variant” CJD. Nature 1996;383:685– 690 Schulz-Schaeffer W, Tscho¨ke S, Kranefuss N, et al. The paraffinembedded tissue blot (PET blot) detects PrPSc early in the incubation time in prion disease. Am J Pathol 2000;156:51–56 Gambetti P, Goldfarb L, Gabizon R, et al. Inherited prion diseases. In: Prusiner S, ed. Prion Biology and Diseases, Cold Spring Harbor Laboratory Press, 1999:509 –583 Padovani A, D’Alessandro M, Parchi P, et al. Fatal familial insomnia in a new Italian family. Neurology 1998;51:1491–1494 Petersen RB, Parchi P, Richardson SL, et al. Effect of the D178N mutation and the codon 129 polymorphism on the metabolism of the prion protein. J Biol Chem 1996;271: 12661–12668

Aneurysm of a Dural Sigmoid Sinus: A Novel Vascular Cause of Pulsatile Tinnitus Emmanuel Houdart, MD,* Rene´ Chapot, MD,* and Jean-Jacques Merland, MD, PhD*

We report a newly evidenced cause of venous pulsatile tinnitus—the aneurysm of a dural sigmoid sinus. A 33year-old patient presented with an incapacitating pulsatile tinnitus of 6 months’ duration in the left ear. The radiological workup evidenced an aneurysm of the left sigmoid sinus. Selective endovascular coil occlusion of the aneurysm was followed by complete resolution of the tinnitus. Houdart E, Chapot R, Merland J-J. Aneurysm of a dural sigmoid sinus: a novel vascular cause of pulsatile tinnitus. Ann Neurol 2000;48:669 – 671

Pulsatile tinnitus is a rare but disabling symptom, which has several causes, both vascular and nonvascular.1,2 We describe here a novel venous cause, namely, the aneurysm of a dura mater sinus. Selective embolization of the aneurysm was followed by resolution of the tinnitus.

Fig 1. CT scan showing the cavity in the temporal bone (arrow). tom and the ease of selective occlusion, endovascular treatment of the aneurysm was proposed. Aspirin therapy, 160 mg daily, was initiated 5 days before the procedure. Embolization was performed under general anaesthesia. A 5F catheter was placed into the left sigmoid sinus after jugular punction, and the aneurysm was catheterized coaxially with a double-marker microcatheter MicroFerret (Cook, Inc, Bloomington, IN). A selective opacification of the aneurysm was performed to confirm the absence of draining channel from the sac. Three detachable platinum coils Detach 18 (Cook, Inc) were placed into the aneurysm,

Patient and Methods A 33-year-old woman presented with a complaint of left pulsatile tinnitus, which had started 6 months earlier. The murmur increased on physical activity and rotation of the head to the right and decreased on rotation of the head to the left. Otoscopy was normal. Auscultation evidenced a lowintensity murmur in the left retroauricular region, when the head was rotated to the right. The rest of the examination was normal. Medical history was unremarkable. The CT scan showed a smooth cavity in the left temporal bone, measuring 6 mm in its greatest diameter. This cavity communicated with the sigmoid sinus on its inner side, making a venous origin a possibility (Fig 1). The cerebral angiogram confirmed the presence of a left sigmoid sinus ectasia, which we termed “aneurysm” (Fig 2), and did not evidence any known vascular cause of pulsatile tinnitus. Because of the incapacitating nature of the symp-

Fig 2. Venous phase of the initial angiogram showing the left sigmoid sinus aneurysm (arrow).

From the *Department of Neuroradiology and Therapeutic Angiography, Hoˆpital Lariboisie`re, Paris, France. Received Mar 22, 2000. Accepted for publication May 26, 2000. Address correspondence to Dr Houdart, Service de Neuroradiologie et d’Angiographie The´rapeutique, Hoˆpital Lariboisie`re, 2 rue Ambroise Pare´, 75010 Paris, France.

Copyright © 2000 by the American Neurological Association

669

leading to its angiographic exclusion (Fig 3). The tinnitus disappeared postprocedure and had not reappeared at 8 months of follow-up. Aspirin therapy was continued for a month.

Results and Discussion Pulsatile tinnitus manifests as an intracranial murmur, synchronous with the heart beat. Depending on cranial auscultation results, the tinnitus is classified as objective (when the murmur is heard by the examiner) or subjective (when it is not). Manual compression of the vascular structures of the neck can help the diagnosis. In case of intracranial arteriovenous fistulas (AVF) or intracranial arterial stenosis, the tinnitus decreases with ipsilateral carotid artery compression. A more specific diagnostic maneuver for dural AVF of the lateral sinus consists in compressing the ipsilateral occipital artery in the retroaural region. This diminishes the occipital artery flow, which usually supplies the AVF and results in a decrease of the murmur’s intensity. Pulsatile tinnitus may be related to turbulences in a sigmoid sinus without associated intracranial AVF.3– 6 In these cases, the murmur’s intensity varies with maneuvers that modify the internal jugular vein (IJV) output. Compression of the ipsilateral IJV or rotation of the head in the murmur’s direction diminishes the bruit. Conversely, compression over the contralateral IJV or turning the head in the opposite direction increases the bruit by increasing the flow in the affected IJV. In our patient, the variations in the murmur’s in-

Fig 3. Venous phase of the postembolization angiogram showing the occlusion of the aneurysm.

670

Annals of Neurology

Vol 48

No 4

October 2000

tensity with these maneuvers oriented toward the existence of a left-sided venous lesion. Radiological investigations of a pulsatile tinnitus include at our institution high-resolution computerized tomography of the temporal bone, brain magnetic resonance imaging (MRI), and carotid ultrasonography. Cerebral angiography is done when these explorations raise the possibility of an intracranial AVF or if the clinical characteristics of the tinnitus are evocative of such a lesion. Some authors have proposed MR angiography (MRA) as a substitute for angiography,7 but the value of a normal MRA is unknown. Lumbar puncture is done when benign intracranial hypertension (BIH) is clinically suspected and if the radiological explorations have ruled out a tumor or an AVF. It is admitted that a vascular tinnitus is due to turbulences of the blood flow in vascular structures located close to the inner ear. In our patient, the only radiological anomaly found was the sigmoid sinus aneurysm. Because of its anatomical situation, we hypothesized that blood turbulences were causing the tinnitus. This hypothesis was also supported by the clinical characteristics of the tinnitus and the normality of the clinical examination. Known causes of venous tinnitus include BIH,6 compression of the IJV by the atlas lateral process,3 “high jugular bulb,”5 and presence of an abnormal mastoid emissary vein.8,9 In BIH syndrome, pulsatile tinnitus is the consequence of the raised intracranial pressure and resolves after weight reduction and administration of acetazolamide and furosemide.6 In the other causes, there is little evidence that a venous abnormality may be directly responsible for the tinnitus. Of the two published cases of abnormal mastoid emissary vein, only one reported regression of the tinnitus following ligature of the mastoid vein.9 The second patient was treated by IJV ligation.8 Sismanis6 and Meador10 suspect that most reported cases of venous tinnitus are actually due to intracranial hypertension that creates turbulences in the jugular bulb. According to this hypothesis, jugular bulb abnormalities would become symptomatic only when associated with raised intracranial pressure. Excluding this last condition, the above-mentioned causes of venous tinnitus can be treated by occlusion of the IJV. It is difficult, however, to propose an intervention that carries a risk of increasing the intracranial pressure and definitively suppresses one of the two brain-draining veins with so little evidence of a clinical benefit. In our patient, the aneurysm was a well-circumscribed lesion, distinct from the normal sinus. Therefore, a selective treatment that did not affect the cerebral drainage was possible. The fact that the aneurysm was responsible for the tinnitus was clearly demonstrated by its disappearance after selective embolization. Treatment was decided on because of the debilitat-

ing nature of the symptom and the ease of selective endovascular sac occlusion. Detachable platinum coils have been used for several years in the treatment of intracranial berry aneurysms,11 the shape of which is identical to that of this sinus aneurysm. Our unique concern was the potential extension of the thrombosis from the aneurysm to the normal sinus. This is a well-known, although rare, complication of intracranial arterial aneurysm embolization. To prevent it, aspirin treatment was given before and continued 1 month after embolization. The origin of this venous aneurysm remains debatable. The accepted mechanism of intracranial sacciform arterial aneurysm formation is herniation of the intima through a medial defect under the effect of arterial turbulences. This mechanism is unlikely here. Indeed, blood flow turbulences in a venous sinus are weak, and the dura mater is a very sturdy membrane. Moreover, the hernia developed through the bone and not toward the brain matter, which is less resistant. There was no history of trauma, eliminating the possibility of a hernia through a bone fracture. Thus, the origin of this aneurysm remains unknown. In conclusion, this newly described cause of pulsatile tinnitus, although probably rare, should be suspected in front of a venous tinnitus. Diagnosis is easily made by temporal bone CT. Endovascular treatment of such a lesion seems easy and safe. We thank Prof Jean-Paul Monteil, who referred the patient to us for diagnosis and treatment.

References 1. Waldvogel D, Mattle HP, Sturzenegger M, Schroth G. Pulsatile tinnitus: a review of 84 patients. J Neurol 1998;245:137–142 2. Sismanis A. Pulsatile tinnitus: a 15 year experience. Am J Otol 1998;19:472– 477 3. Curforth R, Wiseman J, Sutherland RD. The genesis of the cervical venous hum. Am Heart J 1970;80:488 – 492 4. Hardison JE, Smith RB, Crowley IS, et al. Self-heard venous hums. JAMA 1981;245:1146 –1147 5. Adler JR, Ropper AH. Self-audible venous bruits and high jugular bulb. Arch Neurol 1986;43:257–259 6. Sismanis A, Butts FM, Hughes GB. Objective tinnitus in benign intracranial hypertension: an update. Laryngoscope 1990; 100:33–36 7. Dietz RR, Davis WL, Harnsberger HR, et al. MR imaging and MR angiography in the evaluation of pulsatile tinnitus. Am J Neuroradiol 1994;15:879 – 889 8. Lambert PR, Cantrell RW. Objective tinnitus in association with an abnormal posterior condylar emissary vein. Am J Otolaryngol 1986;7:204 –207 9. Forte V, Turner A, Liu P. Objective tinnitus associated with abnormal mastoid emissary vein. J Otolaryngol 1989;18:232–235 10. Meador KJ, Swift TR. Tinnitus from intracranial hypertension. Neurology 1984;34:1258 –1261 11. Vinuela F, Duckwiler G, Mawad M, et al. Guglielmi detachable coil embolization of acute intracranial aneurysm: perioperative anatomical and clinical outcome in 403 patients. J Neurosurg 1997;86:3, 475– 482

Peripheral Neuropathy with Hypomyelination, Chronic Intestinal PseudoObstruction and Deafness: A Developmental “Neural Crest Syndrome” Related to a SOX10 Mutation Ve´ronique Pingault, PhD,* Anne Guiochon-Mantel, MD, PhD,† Nade`ge Bondurand, PhD,* Christophe Faure, MD,‡ Catherine Lacroix, MD,§ Stanislas Lyonnet, MD, PhD,㛳 Michel Goossens, MD,* and Pierre Landrieu, MD§¶

We describe the case of a girl with an unusual congenital phenotype, combining peculiar peripheral nerve lesions with hypomyelination, chronic intestinal pseudoobstruction, and deafness. She was found to have a de novo heterozygous frameshift mutation in the gene encoding the SOX10 transcription factor. The likely role of SOX10 in determining the fate of Schwann cells during early embryogenesis may explain the peripheral nervous system developmental disorder observed in this patient. Pingault V, Guiochon-Mantel A, Bondurand N, Faure C, Lacroix C, Lyonnet S, Goossens M, Landrieu P. Peripheral neuropathy with hypomyelination, chronic intestinal pseudoobstruction and deafness: a developmental “neural crest syndrome” related to a SOX10 mutation. Ann Neurol 2000;48:671– 676

SOX10, an SRY-related transcription factor, is expressed early in the neural crest, and then preferentially in Schwann cells and oligodendrocytes. SOX10 interacts in vitro with several other transcription factors that modulate Schwann cells activity during myelination or premyelination (EGR2, PAX3, POU3F1).1 SOX10 is mutated in an animal model of Hirschsprung disease

From *Ge´ne´tique Mole´culaire et Physiopathologie, Inserm U468, and Laboratoire de Biochimie et Ge´ne´tique, Assistance Publique Hoˆpitaux de Paris, Hoˆpital Henri Mondor, Cre´teil; †Laboratoire d’Hormonologie et Biologie Mole´culaire, ¶Service de Neurologie Pe´diatrique, §Laboratoire de Neuropathologie, Hoˆpital Biceˆtre, Le Kremlin Biceˆtre; and ‡Service de Gastroenterologie Pe´diatrique, Hoˆpital Robert Debre´, and 㛳Service de Ge´ne´tique et Inserm U393, Hoˆpital des Enfants Malades, Paris, France. Received Mar 30, 2000, and in revised form May 26. Accepted for publication May 30, 2000. Address correspondence to Dr Goossens, Laboratoire de Biochimie, Hoˆpital Henri Mondor, 94010 Cre´teil Cedex, France.

Copyright © 2000 by the American Neurological Association

671

(HD), the dominant megacolon (Dom) mouse,2 and SOX10 mutations have been found in families with Waardenburg syndrome type 4, characterized by intestinal aganglionosis, pigmentation disorders and deafness (WS4, OMIM 277580).3 Here we report an association between a SOX10 mutation and an unusual phenotype, combining peculiar nerve lesions, chronic intestinal pseudoobstruction and deafness, supporting the role of SOX10 in the peripheral nervous system (PNS) development.

aided at 15 months and to walk at 25 months. As a result of hyperkinetism between 5 and 7 years, she had fractures of the tibia and of the femur, which had a normal course. Urinary continence was achieved at age 6 years. Hypolacrimation and absence of sweating were present since birth without complications. Reaction to pain was normal. The patient entered an educational program for deaf children at age 22 months, special kindergarten at 4 years, and primary school at 8 years. She communicates abundantly with sign language and is described as a girl with authoritarian behavior and a peculiar sense of humor, with a lownormal IQ. No pigmentary abnormality was noted.

Case Report The girl was born to a 38-year-old woman and a 36-year-old man, after a 40-week pregnancy marked by a moderate excess of amniotic fluid. The mensurations and Apgar score were normal. Intestinal occlusion with dilated colon led to laparotomy and left colostomy at day 19. Colostomy was closed after 1 year with Duhamel’s procedure. Shortly afterward, recurrent subocclusive episodes developed, while esophageal and antroduodenal manometries were compatible with visceral neuropathy. Home-based parenteral nutrition was used until age 27 months, then oral feeding was gradually reintroduced for 3 years. At age 6 years, a third surgical procedure resulting in ileostomy was necessary. For 1 year, parenteral nutrition was associated with progressive oral feeding. At age 8 years, weight is 27 kg and height 127 cm. Neonatal hypotonia, weak spontaneous or reflex movements, and no reaction to sounds contrasted with good social contact. Large ocular saccades resolved after a few months. Deep tendon reflexes were present from the sixth month. Delayed motor development enabled the patient to sit un-

Physiological Investigations Electromyographic studies showed a regular improvement of all parameters during the 8 years of follow-up (Table). At 4 years, the polysynaptic response of the femoral biceps muscle to calibrated noxious stimuli of the sural nerve showed a normal threshold of 9 mA but an abnormal latency (314 msec; normal, ⬍120 msec). Auditory evoked potentials were flat at 6 months. Biopsy specimens of the superficial peroneal nerve and of the peroneus brevis muscle at age 1.5 months were processed for conventional morphological studies. Muscle fibers showed slight caliber variation. On 0.5-␮m nerve section, the transverse organization exhibited unusual micropolyfasciculation, each nerve fascicle itself being subdivided into subfascicles (Fig 1A). In each fascicle, the density of myelinated fibers was slightly diminished. In most myelinated fibers, the myelin sheath was thin or occasionally absent but showed no active myelin breakdown (see Fig 1C), including on ultrastructure. Basement membranes were nor-

Table. Course of the Electromyographic Data over the 8-Year Follow-Up Period Age Nerve

Data

6 wk

3 yr

7.5 yr

Median (motor)

NCV (msec) DL (msec) Pot (mV) F-M (msec) NCV (msec) Pot (␮V) NCV (msec) DL (msec) Pot (mV) F-M (msec)

8 11 1.9 42.6 5.3 0.5 32.5 6.4 0.6 50

33.5 3.5 5.2 25.4 34 9.1

38 3.5 7.2 30.2 37.5 11.2

Median (sensory) Lateral popliteal (motor)

Potentials were recorded through implanted electrodes. NCV ⫽ nerve conduction velocity; DL ⫽ distal latency; Pot ⫽ potential; F-M ⫽ F-M latency.

Fig 1. (A) A 0.5-␮m cross-section of the cutaneous branch of the lateral popliteal nerve. Nerve fascicles are increased in number and many are reduced to a small population of fibers. In each fascicle, the density of myelinated nerve fibers is reduced. The axonal diameters of myelinated fibers and their unimodal distribution are normal for age. (B) Control biopsy showing the usual fascicular organization of the same nerve at the same level (biopsy from a 1-month-old boy with Werdnig-Hoffmann disease, in which sensory nerves are considered to develop in the near-normal range). (C) Many nerve fascicles show an arrangement in microsubfascicles. The myelin sheath is reduced in many fibers (arrows). (D) Ultrastructure. Marked reduction of myelin in a large fiber. (Original magnification: A and B, ⫻100; C, ⫻630; D, ⫻5,000. Bar ⫽ 1 ␮m.)

672

Annals of Neurology

Vol 48

No 4

October 2000

‹

Brief Communication: Pingault et al: SOX10 Mutation and Peripheral Neuropathy

673

mal (see Fig 1D). Myelin interperiodic lines frequently showed discontinuity in fibers with no fixation artifacts otherwise. Amyelinic fibers had a normal density. Intestinal morphology on the colostomy specimen at day 19 and on two colon specimens from the Duhamel procedure ruled out HD: the plexi of Meissner and Auerbach exhibited a regular disposition, numerous ganglionic cells, and normal-sized nerve processes. Staining for acetylcholinesterase and immunohistochemistry of neuron-specific enolase and PS100 gave normal results, although nerve processes were slightly less abundant in the muscle lamina than in the submucosal lamina. Myelination indexes at age 5 months on brain magnetic resonance imaging were normal. Ophthalmoscopy, visual acuity, and cardiovascular parameters were normal.

Genetic Investigations Karyotype was normal (46,XX). Search for the 17p 1.5-Mb duplication/deletion was performed by fluorescent quantitative polymerase chain reaction (PCR) of microsatellites RM11GT, D17S921, D17S839, and D17S955.4,5 The two coding exons of EGR2 were analyzed with 7 overlapping sets

of primers6 and sequenced as previously described7 with slight modifications. The RET gene was screened as described.8 The SOX10 coding sequence was sequenced along both strands using the following primers: 1. 5⬘CTGTGCCCACGTCCTGTCTC3⬘/5⬘TCGCCGTCCTGCTGCTCCTT3⬘ 2. 5⬘GCGAGCTGGGCAAGGTCAAG3⬘/5⬘GAATCCACCCGAAGCTAGAG3⬘ 3. 5⬘GGAGTGCTCTGGCATTCACG3⬘/5⬘CTTGCCCCACCCTCAGCTCT3⬘ 4. 5⬘GGAAGTTCACGTGCGCCCAC3⬘/5⬘GCGGCAGGTACTGGTCCAAC3⬘ 5. 5⬘GGTAATGTCCAACATGGAGACC3⬘/5⬘GTAGGCGATCTGTGAGGTGG3⬘ 6. 5⬘CCACACTACACCGACCAGCC3⬘/5⬘GGGTGGTGGCGACAGGGC3⬘

PCR conditions will be published elsewhere. The mutation was verified by cloning the PCR product (TOPO Cloning Kit; Invitrogen, Groningen, The Netherlands).

Fig 2. The 795delG mutation. (A) Sequencing of the SOX10 gene coding region revealed a de novo heterozygous mutation consisting of a 1-bp deletion. (B) This frameshift mutation generates a new carboxy tail of 19 amino acids (in gray) and a premature stop codon. (C) The normal SOX10 factor is represented, with its functional and structural domains. The 795delG mutation lies in the middle of a domain which is conserved in subgroup E of SOX factors, comprising SOX8, SOX9, and SOX10. The mutation results in truncation of this domain and loss of the transactivation domain.

674

Annals of Neurology

Vol 48

No 4

October 2000

Paternity was checked using 10 microsatellites from the ABI PRISM Linkage Mapping Set. Results No 17p11.2-12 rearrangement and no mutations in the RET and EGR2 genes coding sequences were found. Sequencing of SOX10 revealed a heterozygous deletion of 1 bp (795delG) in the last coding exon, which results in a frameshift generating a new carboxyterminal tail and premature truncation (Fig 2). Studies of the parents revealed the mutation had occurred de novo, in keeping with the sporadic occurrence of the syndrome in the family. Discussion The present phenotype corresponds to a non–wellestablished syndrome, but each element pertains to a known category of disorders that result from abnormal neural crest development. The peripheral neuropathy was peculiar: the clinical manifestations, the marked slowing of nerve conduction velocity, the depopulation of myelinated fibers, and the subnormal or absent myelination of the remaining large fibers situate the disorder within the congenital hypomyelinating neuropathies.9 However, the unusual multiplication of nerve fascicles, each reduced to a small population of fibers, suggests that the transversal organization of the nerve trunks was impaired by the SOX10 mutation. The absence of active demyelination and the favorable course of PNS maturation, both clinically and neurophysiologically suggest that the developmental dysregulating process predominated over degenerative phenomena. No defect of peripheral myelination has been registered in Dom mutant heterozygous mice10 or in WS4 patients carrying SOX10 mutations, but no detailed PNS studies have been performed. In a 4-year-old girl with a 1400del12 SOX10 mutation, whose picture was dominated by severe brain atrophy, NCVs were moderately slowed, but the peripheral nerves were submitted neither to serial tests nor to morphological studies.11 In 3 other WS4 patients with SOX10 truncating mutations, the phenotype included not only CNS impairment but also dysautonomic features similar to those of our patient, which suggests that PNS impairment was present.12 Thus, the present neuropathy contrasts with other early-onset progressive demyelinating neuropathies, in which mutations affecting the major structural myelin proteins are candidates and occasionally are found.13 Chronic intestinal pseudoobstruction is a rare syndrome in which impaired motility does not result from a mechanical obstruction. Whether primary or secondary to extragastrointestinal diseases, neuropathic, myopathic, and unclassified forms have been distinguished on histopathological basis.14 The ill-defined forms are delineated from HD mainly by the persistence of ganglionic cells and nervous plexi in the submucosal com-

partments.15 The present case, together with reports of WS4 patients with SOX10 mutations3 and the variable phenotypes observed in some HD families with at least one typical case,16 suggest that severe colon hypomotility at birth may cover a variety of neurogenic defects, ranging from typical aganglionosis to hypoganglionosis and forms with no clear quantitative lesions. SOX10 mutations in WS4 patients have been suggested to result in haploinsufficiency.3 Accordingly, the heterozygous mutation described here leads to premature truncation and loss of the transactivation domain (see Fig 2B) and probably results in a loss of protein function. Its occurrence de novo is in keeping with a major role of SOX10 in the observed phenotype. The abnormal myelination described by Inoue and colleagues was suggested to result from a dominant negative effect, given the large number of prolines in the new carboxy terminus of the tail generated by the 1400del12 mutation.11 In contrast, the new tail generated by the present 795delG mutation is very short and contains no proline. The P0 myelin protein, which is involved in the demyelinating peripheral neuropathy CMT-1b, is expressed in glial and neuronal progenitors long before myelin synthesis begins and could play a role in the developmental specification of these cells.17 The P0 gene was shown to be a target for the transactivation factor SOX10.18 The different morphological features and clinical courses of neuropathies resulting from SOX10 and P0 mutations suggest that interactions between the two genes could vary during development and involve several other players. Together with cellular biology and transgene experiments, neurophysiological and morphological studies of peripheral nerve in SOX10-deficient patients should throw further light on the role of SOX10 in the PNS development. This work was supported by Biomed (contract BMH4-CT97-2107) and grants from PHRC (AOA 94060) and the Association pour la Recherche sur le Cancer. N.B. is supported by a “Claude Bernard” fellowship from the Ligue Nationale contre le Cancer. We are indebted to S. Metral for neurophysiological studies; M. Fabre, G. Fromont, and M. Peuchmaur for pathological studies; A. Lababidi and P. de Lagausie for surgical management; and A. Saillant and C. Crumiere for pediatric care. We also thank Isabelle Deval-Boucly and Josette Bacci for technical assistance.

References 1. Kuhlbrodt K, Herbarth B, Sock E, et al. Sox10, a novel transcriptional modulator in glial cells. J Neurosci 1998;18:237–250 2. Southard-Smith EM, Kos L, Pavan WJ. Sox10 mutation disrupts neural crest development in Dom Hirschsprung mouse model. Nat Genet 1998;18:60 – 64 3. Pingault V, Bondurand N, Kuhlbrodt K, et al. SOX10 mutations in patients with Waardenburg-Hirschsprung disease. Nat Genet 1998;18:171–173 4. Lupski JR, de Oca-Luna RM, Slaugenhaupt S, et al. DNA du-

Brief Communication: Pingault et al: SOX10 Mutation and Peripheral Neuropathy

675

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

plication associated with Charcot-Marie-Tooth disease type 1A. Cell 1991;66:219 –232 Cudrey C, Chevillard C, Le Paslier D, et al. Assignment of microsatellite sequences to the region duplicated in CMT1A (17p12): a useful tool for diagnosis. J Med Genet 1995;32: 231–233 Timmerman V, De Jonghe P, Ceuterick C, et al. Novel missense mutation in the early growth response 2 gene associated with Dejerine-Sottas syndrome phenotype. Neurology 1999;52: 1827–1832 Warner LE, Mancias P, Butler IJ, et al. Mutations in the early growth response 2 (EGR2) gene are associated with hereditary myelinopathies. Nat Genet 1998;18:382–384 Attie T, Pelet A, Edery P, et al. Diversity of RET protooncogene mutations in familial and sporadic Hirschsprung disease. Hum Mol Genet 1995;4:1381–1386 Guzzetta F, Ferriere G, Lyon G. Congenital hypomyelination polyneuropathy. Pathological findings compared with polyneuropathies starting later in life. Brain 1982;105:395– 416 Lane PW, Liu HM. Association of megacolon with a new dominant spotting gene (Dom) in the mouse. J Hered 1984;75:435– 439 Inoue K, Tanabe Y, Lupski JR. Myelin deficiencies in both the central and the peripheral nervous systems associated with a SOX10 mutation. Ann Neurol 1999;46:313–318 Touraine RL, Attie-Bitach T, Manceau E, et al. Neurological phenotype in Waardenburg syndrome type 4 correlates with novel SOX10 truncating mutations and expression in developing brain. Am J Hum Genet 2000;66:1496 –1503 Warner LE, Hilz MJ, Appel SH, et al. Clinical phenotypes of different MPZ (P0) mutations may include Charcot-MarieTooth type 1B, Dejerine-Sottas, and congenital hypomyelination. Neuron 1996;17:451– 460 Faure C, Goulet O, Ategbo S, et al. Chronic intestinal pseudoobstruction syndrome: clinical analysis, outcome, and prognosis in 105 children. French-Speaking Group of Pediatric Gastroenterology. Dig Dis Sci 1999;44:953–959 Navarro J, Sonsino E, Boige N, et al. Visceral neuropathies responsible for chronic intestinal pseudo-obstruction syndrome in pediatric practice: analysis of 26 cases. J Pediatr Gastroenterol Nutr 1990;11:179 –195 Badner JA, Sieber WK, Garver KL, Chakravarti A. A genetic study of Hirschsprung disease. Am J Hum Genet 1990;46: 568 –580 Hagedorn L, Suter U, Sommer L. P0 and PMP22 mark a multipotent neural crest–derived cell type that displays community effects in response to TGF-beta family factors. Development 1999;126:3781–3794 Peirano RI, Goerich DE, Riethmacher D, Wegner M. Protein zero gene expression is regulated by the glial transcription factor Sox10. Mol Cell Biol 2000;20:3198 –3209

Impaired Activation of Oxygen Consumption and Blood Flow in Visual Cortex of Patients with Mitochondrial Encephalomyopathy M. S. Vafaee, PhD,*† E. Meyer, PhD,* and A. Gjedde, MD, DM*†

A current hypothesis claims that an increase of blood flow is required for oxygen consumption to rise during neuronal excitation (activation). Chronic progressive external ophthalmoplegia is a mitochondrial disease associated with deletions of mtDNA or by point mutation of tRNA genes. We tested the hypothesis that the cerebral metabolic rate of oxygen (CMRO2) may not rise in this disorder if the accompanying cerebral blood flow increase is insufficient. Two patients with progressive external ophthalmoplegia were visually stimulated with a colored checkerboard pattern reversing as different frequencies. When stimulated, Patient 1 had a small increase of cerebral blood flow, while Patient 2 had no cerebral blood flow increase. In the visually active state, the patients had no significant change of CMRO2, while healthy subjects had a pronounced increase of CMRO2 in the pericalcarine visual cortex at 4 Hz and a further slight increase at 8 Hz during activation. Vafaee MS, Meyer E, Gjedde A. Impaired activation of oxygen consumption and blood flow in visual cortex of patients with mitochondrial encephalomyopathy. Ann Neurol 2000;48:676 – 679

Mitochondria supply adenosine triphosphate for cellular work by means of oxidative phosphorylation. About half of patients with progressive external ophthalmoplegia (PEO)1,2 have detectable mitochondrial DNA (mtDNA) deletions.2 A significant number of the remaining patients negative for deletions carry the 3243 point mutation in tRNAleu(UUR) that is commonly associated with mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes (MELAS).3

From the *McConnell Brain Imaging Centre, Montreal Neurological Institute, McGill University, Montreal, Quebec, Canada; and †PET Center, Aarhus University Hospital, Aarhus, Denmark. Received Aug 23, 1999, and in revised form May 9 and June 8, 2000. Accepted for publication Jun 8, 2000. Address correspondence to Dr Vafaee, PET Center, Aarhus University Hospital, Norrebrogade 44, Aarhus C, 8000 Denmark.

676

Copyright © 2000 by the American Neurological Association

Theoretically, the brain should be able to increase its oxygen consumption without requiring an increase in cerebral blood flow (CBF), as there is a substantial reserve in arterial blood, but in practice an increase in physiological activity raises metabolism and releases vasodilators, which in turn raise blood flow. A current hypothesis claims that the blood flow increase is required for oxygen consumption to rise during neuronal excitation.4 – 6 According to this hypothesis, deficient mitochondrial use of oxygen may be related to insufficient increase of blood flow on stimulation. The goal of the study was to test this hypothesis by measuring CBF and the cerebral metabolic rate of oxygen (CMRO2) at rest and during physiological stimulation of 2 patients with chronic progressive external ophthalmoplegia (PEO), compared with a series of normal subjects studied by the same method.6

Patients and Methods Of the patients with CPEO, 1 was a 47-year-old woman with a mtDNA deletion. The other was a 59-year-old man presenting with proximal limb muscle weakness and chronic progressive external ophthalmoplegia,7 in whom a mtDNA point mutation was detected in the tRNAleu gene (G to A at position 12315). PET measurements were made with the ECAT EXACT HR⫹ (CTI/Siemens) whole-body tomograph, operating in a 3D acquisition mode, as described previously.6,8 Each patient also underwent a magnetic resonance imaging (MRI) examination for image registration purpose. The patients also underwent a visually evoked potential test to ensure an intact visual system and absent photosensitivity to the stimulus. Both had normal vision and visual reactivity. The stimulus was a yellow-blue annular checkerboard with a visual angle of about 17°.9 In the baseline condition, the patients fixated a cross-hair in the center of the screen 30 seconds before the onset of, and throughout, the subsequent 3-minute tomography. In two successive activation sessions, the patients were shown the checkerboard reversing its contrast at frequencies of 4 or 8 Hz. Stimulation began 4 minutes before the start of the dynamic PET scan and continued throughout the following 3-minute scan for a total of 7 minutes. MRIs were transformed into stereotaxic coordinates10 by means of an automatic registration algorithm.11 The reconstructed PET images were co-registered with the subjects’ MRI scans using an automatic registration program based on the automatic image registration algorithm.12 The globally averaged CMRO2 and CBF values were then determined for each patient by averaging the values of all intracerebral voxels. The regionally averaged CMRO2 and CBF values of each patient (pericalcarine visual cortex) were determined in manually drawn regions of interests on PET images coregistered to corresponding MRIs transfered to a standard orthogonal coordinate system.10 Both global and regional CMRO2 and

CBF values were then compared with average values obtained from normal subjects.6

Results The global CBF values of the patients were normal at baseline and in the activation conditions shown in part A of the Figure. In the activation states, Patient 1 showed a small increase of CBF (7% at 4 Hz, 20% at 8 Hz), but Patient 2 had no CBF increase (see Fig, B), unlike normal subjects who have pronounced increases of CBF in the pericalcarine visual cortex when stimulated (38% at 4 Hz, 42% at 8 Hz). The globally averaged CMRO2 value of Patient 1 was normal, whereas that of Patient 2 was below normal (see Fig, C). Although the globally averaged CMRO2 values of the patients did not change significantly with activation, the globally averaged CMRO2 values of Patient 2 were significantly lower in both states than in normal subjects. In the activation, normal subjects have a pronounced increase of CMRO2 in the pericalcarine visual cortex at 4 Hz (15%) and a slight increase at 8 Hz (5%) compared with baseline. In contrast, no patient had a significant change of CMRO2 during activation (see Fig, D). Discussion PEO is a sporadically appearing syndrome with abnormal ocular motility. About 50% of patients with PEO have mtDNA deletions.2 The syndrome is characterized by ophthalmoplegia, ptosis, and proximal limb and respiratory muscle weakness, cataracts, and hearing loss. Muscle tissue from these patients shows several defects of respiratory chain enzymes.13 The two patients of the current study had a diagnosis of chronic PEO. Patient 1 had a 4-kb deletion in mtDNA. Patient 2 demonstrated severe cytochrome oxidase deficiency, presumably due to translation defects which affected all of the respiratory chain complexes encoded by mtDNA. The observation that both patients consistently demonstrated low CMRO2 values implies a disturbance of the normal coupling between perfusion and energy metabolism in these patients. This observation is consistent with the fact that oxidative phosphorylation depends on enzyme complexes assembled from subunits derived from both mtDNA and nuclear DNA. Defective mtDNA could therefore result in an impairment of energy metabolism. On the other hand, if oxidative phosphorylation is impaired as the result of a primary defect of mitochondrial function, excess ADP or NADH⫹ would be excited to stimulate glycolysis via the anaerobic pathway, leading to excess production of lactate in the visual cortex and other areas. The serum lactate level of Patient 1 was normal (1.7 mM) but that of Patient 2 was high (2.4 mM).

Brief Communication: Vafaee et al: CBF and CMRO2 in Mitochondrial Disorders by PET

677

Fig. Global cerebral blood flow (CBF) values (A), occipital CBF values (B), global CMRO2 values (C), and occipital CMRO2 values (D) in the baseline and activation states (4 and 8 Hz) in 2 patients and 12 controls.

Although both patients had been diagnosed with chronic PEO, there were striking differences between their test results. Patient 2 consistently showed low CBF values in comparison with Patient 1, without demonstrating any sensitivity to the stimulus frequency. Patient 1 also had normal global CMRO2 values, while those of Patient 2 were significantly lower. We attribute the differences between the two patients to the different severity and duration of their diseases (25 years for Patient 1, 40 years for Patient 2). The high lactate level of Patient 2 is consistent with the speculation that longer-lasting mitochondrial dysfunction may cause the patients to rely more on anaerobic glycolysis, but MRS would be required to test this claim. Previous studies showed a marked decline of oxygen metabolism in patients with mitochondrial disorders.14,15 Based on these results, we speculate that the capacity of brain cells for increased oxygen consumption is reduced in patients with a mitochondrial disorder compared with normal individuals. The low occipital CMRO2 value of Patient 2 suggests that the cells in this region work close to the maximum rate of oxygen consumption. On the other hand, the normal global but lower occipital CMRO2 value of Patient 1 678

Annals of Neurology

Vol 48

No 4

October 2000

suggests that the oxidative metabolism of this patient is active at a level somewhat below the maximum. The CBF and CMRO2 values of the 2 patients are consistent with the hypothesis that blood flow and oxygen consumption are uncoupled in these patients. The results show that PET can be used to measure pathological changes of the coupling between CBF, CMRO2, and the cerebral metabolic rate of glucose, which may help to establish the stage and severity of disorders of oxidative phosphorylation. The impaired response of the CBF and CMRO2 regulatory mechanisms to physiological stimulation is consistent with the hypothesis that increased blood flow, possibly influenced by a signal from neurons with normal mitochondrial function, supports the increased oxygen consumption.4,6

Supported by Medical Research Council (Canada) grant SP-30, the Isaac Walton Killam Fellowship Fund of the Montreal Neurological Institute, the McDonnell-Pew Program in Cognitive Neuroscience, and Medical Research Council (Denmark) grants 9305246, 9305427, and 9601888.

References 1. Holt IJ, Harding AE, Cooper JM, et al. Mitochondrial myopathies: clinical and biochemical features of 30 patients with major deletions of muscle mitochondrial DNA. Ann Neurol 1989;26:699 –708 2. Moraes CT, DiMauro S, Zeviani M, et al. Mitochondrial DNA deletions in progressive external ophthalmoplegia and KearnsSayer syndrome. N Engl J Med 1989;320:1293–1299 3. Ciacci F, Moraes CT, Silvestri S, et al. The ‘MELAS-3243’ mutation in mtDNA is found in many patients with progressive external ophthalmoplegia (PEO). Neurology 1992;42(Suppl 3): 417 (Abstract) 4. Gjedde A. The relation between brain function and cerebral blood flow and metabolism. In: Hunt Batjer H, ed. Cerebrovascular disease. Philadelphia: Lippincott-Raven, 1997:23– 40 5. Buxton R, Frank LR. A model for the coupling between cerebral blood flow and oxygen metabolism during neuronal stimulation. J Cereb Blood Flow Metab 1997;17:64 –72 6. Vafaee MS, Gjedde A. Model of blood-brain transfer of oxygen explains nonlinear flow-metabolism coupling during stimulation of visual cortex. J Cereb Blood Flow Metab 2000;20:747– 754 7. Fu K, Hartlen R, Johns T, et al. A novel heteroplasmic tRNALeu(CUN) mtDNA point mutation in a sporadic patient with mitochondrial encephalomyopathy segregates rapidly in skeletal muscle and suggests an approach to therapy. Hum Mol Genet 1996;5:1835–1940 8. Vafaee MS, Murase K, Gjedde A, Meyer E. Dispersion correction for automatic sampling of O-15 labeled H2O and red blood cells. In: Myers R, Cunningham VJ, Bailey DL, Jones T, eds. Quantification of brain function using PET. San Diego: Academic Press, 1996:72–75 9. Vafaee MS, Meyer E, Marrett S, et al. Frequency-dependent changes in cerebral metabolic rate of oxygen during activation of human visual cortex. J Cereb Blood Flow Metab 1999;19: 272–277 10. Talairach J, Tournoux P. Co-planar stereotactic atlas of the human brain: 3-dimensional proportional system—an approach to cerebral imaging. Stuttgart: George Thieme Verlag, 1988 11. Collins DL, Neelin P, Peters TM, Evans AC. Automatic 3D intersubject registration of MR volumetric data in standardized Talairach space. J Comput Assist Tomogr 1994;18:192–205 12. Woods RP, Mazziotta JC, Cherry SR. MRI-PET registration with automated algorithm. J Comput Assist Tomogr 1993;17: 536 –546 13. Servidei S, Zeviani M, Manfredi G, et al. Dominantly inherited mitochondrial myopathy with multiple deletions of mitochondrial DNA: clinical, morphological, and biochemical studies. Neurology 1991;41:1053–1059 14. Frackowiak RSJ, Herold S, Petty RKH, Morgan-Hughes JA. The cerebral metabolism of glucose and oxygen measured with positron emission tomography in patients with mitochondrial diseases. Brain 1988;111:1009 –1024 15. Berkovic SF, Carpenter S, Evans A, et al. Myoclonus epilepsy and ragged-red fibers (MERRF). 1. A clinical, pathological, biochemical, magnetic resonance spectrographic and positron emission tomographic study. Brain 1989;112:1231–1260

Muscle Fibers in Inflammatory Myopathies and Cultured Myoblasts Express the Nonclassical Major Histocompatibility Antigen HLA-G Heinz Wiendl, MD,* Lueder Behrens, PhD,* Sabine Maier, PhD,† Margaret A. Johnson, MD,‡ Elisabeth H. Weiss, PhD,† and Reinhard Hohlfeld, MD*§

We demonstrate that HLA-G, a nonclassical major histocompatibility complex class I antigen, is expressed in muscle fibers in various inflammatory myopathies. Further, interferon-␥ induces surface expression and upregulation of mRNA transcripts corresponding to different isoforms of HLA-G in myoblasts cultured from control subjects and patients. HLA-G may have important immunological functions in inflammatory myopathies and other local immune reactions as they occur during vaccination, myoblast transplantation, and gene therapy. Wiendl H, Behrens L, Maier S, Johnson MA, Weiss EH, Hohlfeld R. Muscle fibers in inflammatory myopathies and cultured myoblasts express the nonclassical major histocompatibility antigen HLA-G. Ann Neurol 2000;48:679 – 684

Muscle fibers normally do not express detectable amounts of major histocompatibility (MHC) class I antigens. However, classical MHC-I is strongly upregulated in pathological conditions, especially inflammatory myopathies.1,2 Indeed, upregulation of MHC-I on muscle fibers is essential for the interaction with CD8⫹ cytotoxic T cells observed in polymyositis (PM) and inclusion body myositis (IBM).3,4 The classical human MHC-I (Ia) antigens (HLA-A, -B, and -C) are highly polymorphic molecules that bind short antigenic peptides and “present” them at the cell surface to CD8⫹ T cells. The antigenic pep-

From the *Department of Neuroimmunology, Max-Planck-Institute for Neurobiology, Martinsried, and †Department of Anthropology and Human Genetics, Ludwig Maximilians University, and §Institute for Clinical Neuroimmunology, Klinikum Grosshadern, Munich, Germany; and ‡Muscular Dystrophy Group Research Laboratories, Newcastle General Hospital, Newcastle upon Tyne, UK. Received Mar 6, 2000, and in revised form Jun 8. Accepted for publication Jun 8, 2000. Address correspondence to Dr Hohlfeld, Department of Neuroimmunology, Max-Planck Institute of Neurobiology; D-82152 Martinsried, Germany.

Copyright © 2000 by the American Neurological Association

679

tides are usually derived from pathogens that replicate in the cytosol of infected cells, typically viruses and some bacteria. Because classical MHC-I molecules are widely expressed in most tissues, almost every infected cell can be recognized in this way. In contrast, HLA-G is a nonpolymorphic “nonclassical” MHC-I (Ib) molecule mainly expressed in cytotrophoblasts of the placenta.5,6 The function of HLA-G remains elusive. It is thought that, by unknown mechanisms, HLA-G prevents maternal lymphocytes from attacking fetal tissue. Like classical HLA class I molecules, HLA-G can bind CD8 and antigenic peptides. Receptors capable of sensing HLA-G have been identified recently that are expressed on natural killer (NK) and myeloid cells.5,6 Materials and Methods Clinical Material

separation and cultured as previously described.7 Myoblasts used in our experiments stained over 95% positive for the neural cell adhesion molecule (NCAM) by FACS analysis. For cytokine induction, myoblasts were cultured in the presence of 500 U/ml IFN-␥ or 250 U/ml TNF-␣. Surface expression of HLA-G and other molecules was assessed by flow cytometry on a FACSort (see Table).

Reverse Transcription and PCR Total RNA was reverse-transcribed. One microliter of cDNA was used to amplify HLA-G mRNA in a 20-␮l standard polymerase chain reaction (PCR) as described.8,9 ␤-Actin was amplified as a control. Specificity of amplification was confirmed by direct DNA sequencing of extracted bands after reamplification.

DNA Probes for Hybridization Analysis

Diagnostic muscle biopsy specimens were obtained from patients with inflammatory myopathy (polymyositis [PM], n ⫽ 6; dermatomyositis [DM], n ⫽ 6; and IBM, n ⫽ 5), degenerative muscle disease (Duchenne muscular dystrophy [DMD], n ⫽ 3) and nonmyopathic controls (n ⫽ 3).

Immunohistochemical Studies Flash-frozen muscle biopsy specimens were cut into 8- to 10-␮m cryostat sections and analyzed by immunohistochemistry. Acetone-fixed, air-dried sections were blocked and incubated with primary antibodies or corresponding nonimmune IgG isotype controls, and diluted in phosphate buffered saline containing 2% BSA for 45 minutes at the concentrations indicated in the Table. Antibody binding was visualized by immunofluorescence microscopy and confocal laser microscopy (CY3- or FITC-labeled secondary antibody).

Myoblast Culture, Cytokine Induction, and FACS Analysis Myoblasts were isolated from normal subjects and patients with inflammatory myopathy, purified by magnetic bead

An HLA-G cDNA probe was generated by PCR with the HLA-G4 cDNA as template using the PCR DIG Probe Synthesis Kit (Roche, Basel, Switzerland). The following 3⬘-end– labeled oligonucleotides (DIG oligonucleotide 3⬘-end labeling Kit; Roche) were used for discrimination of the HLA-G isoforms: G3-5 (5⬘-GAAGACTGCTCCGCGCGCT-3⬘) for HLA-G4, and G2-3⬘ (5⬘-GGGGGGGTTGGCCTCGCT3⬘) for HLA-G2. Separated PCR products were hybridized as described elsewhere.10

Results Immunohistochemical Analysis of Biopsy Specimens Consistent with previous reports,1,2 we found that normal muscle fibers do not express detectable levels of classical MHC-I antigens. In contrast, in muscle of patients with inflammatory myopathies, many muscle fibers and inflammatory cells strongly express MHC-I molecules, as demonstrated by staining with the monoclonal antibody (mAb) W6/32 (Fig 1).1,2 To detect HLA-G, we stained cryostat muscle sections with the

Table. Primary Monoclonal Antibodies Used in the Study Antigen

Immunogen

Clone (Isotype)

Source

Concentration/ Dilution

HLA-G

Peptide (aa 61–83) of HLA-G ␣-domain L-h␤2m/HLA-G murine transfectants Human tonsil lymphocytes cell membranes Human lymphoblastoid B-cell line RPMI 8866 KG1a cell line Purified human monocytes Peripheral blood monocytes HUVEC cells Human HLA-A2 cytotoxic T-cell clone L-cell CD40-transfectants

4H84 (IgG1)

M. McMaster

1:75

87G (IgG2b)

D. Geraghty

10 ␮g/ml

W6/32 (IgG2a)

Dako

2 ␮g/ml

L243 (IgG2a)

ATCC

10 ␮g/ml

MY31 (IgG1) BEAR1 (IgG1) M⌽P9 (IgG2b)

Becton-Dickinson Immunotech Becton-Dickinson

1:100 10 ␮g/ml 10 ␮g/ml

1.G11B1 (IgG1) B9.11 (IgG1)

Novocastra Immunotech

10 ␮g/ml 1:8

EA-5 (IgG1)

Natutec

10 ␮g/ml

HLA-G MHC class I (HLA-ABC) MHC class II (HLA-DR) NCAM (Leu-19, CD56) CD11b (Mac-1, CR3) CD14 (LPS-R) CD106 (VCAM-1) CD8 CD40

L ⫽ mouse L cells; h␤2m ⫽ human ␤2-microglobulin; aa ⫽ amino acids; ATCC ⫽ American Type Culture Collection.

680

Annals of Neurology

Vol 48

No 4

October 2000

HLA-G–specific mAbs 87G and 4H84.11,12 HLA-G was undetectable in normal muscle (see Fig 1) and in specimens from patients with DMD (not shown). However, HLA-G was detectable in 5 of 6 PM, 5 or 6 DM, and all of 5 IBM cases (Fig 1 shows representa-

tive sections). The distribution of HLA-G closely paralleled the pattern of classical MHC class I assessed in serial sections, although the intensity of HLA-G was weaker than that of classical MHC class I. The MHCI–positive fibers were also positive for HLA-G. Con-

Fig 1. Expression of classical and nonclassical HLA class I proteins in muscle. (a–f ) Consecutive cryosections of muscle from patients with PM (a, b), DM (c, d), and nonmyopathic controls (e, f ) were stained with the HLA-G-specific mAb 87G (a, c, e), or the anti-HLA class I mAb W6/32 (b, d, f ). Many muscle fibers in PM and DM are double-positive for both antibodies. Capillaries and inflammatory cells also stain positive for HLA-G. In the nonmyopathic control specimens, muscle fibers did not stain for HLA-G or HLA class I. However, capillaries are positive for HLA-class I as described previously.1,2 (g) In a section from another patient with PM, stained with the anti–HLA-G-specific mAb 4H84, HLA-G is localized in the cytoplasm of a regenerating fiber (arrowhead) and in an invaded muscle fiber (arrow). The invading mononuclear cells are also HLA-G⫹ (arrow). (h) The autoinvasive cells (arrow in g) were analyzed by confocal laser microscopy in a consecutive section. The infiltrate contains CD3⫹ T cells (green) and CD40⫹ macrophages (red). (Original magnifications: a, b: ⫻250; c, d: ⫻500; e, f: ⫻250; g: ⫻500; h: ⫻800; all before 20% reduction. Same exposure time for all photographs.)

Brief Communication: Wiendl et al: HLA-G in Muscle

681

versely, we could not find any HLA-G–positive fibers that were negative for MHC-I. Many inflammatory cells stained positive for HLA-G as well as for classical MHC-I. HLA-G Expression in Cultured Myoblasts Analyzed by FACS To confirm that muscle cells can express HLA-G and to identify potential regulatory cytokines, we analyzed several lines of cultured myoblasts for HLA-G expression. In all experiments, myoblasts were over 95% positive for NCAM, which serves as a reliable marker of myoblasts (Fig 2A).13 As reported previously, myoblasts constitutively express MHC-I but not MHCII.7,14 Further, myoblasts do not constitutively express HLA-G (see Fig 2A). Next, we investigated the effects of IFN-␥ and TNF-␣ on HLA-G expression. Myoblasts were incubated for 24, 48, or 72 hours with IFN-␥, TNF-␣, or both. As reported previously,14,15 MHC-II was induced on myoblasts by treatment with IFN-␥. Furthermore, FACS analysis showed that IFN-␥ induces a low level of HLA-G on the surface of myoblasts from different donors (see Fig 2A). TNF-␣ failed to induce HLA-G (not shown). HLA-G mRNA in Cultured Myoblasts Analyzed by PCR and Hybridization HLA-G differs from other HLA class I genes by extensive alternative splicing of HLA-G transcripts, resulting in shorter mRNA molecules lacking one or two consecutive exons of the full-length transcript HLA-G1. To further corroborate these results and to identify the isoforms expressed in myoblasts, we tested RNA for HLA-G transcripts by RT-PCR using primers that allow the amplification of all alternatively spliced HLA-G transcripts and their differentiation by size. The following isoforms can be distinguished: HLA-G1 encodes a heavy chain consisting of three extracellular domains (␣1, ␣2, and ␣3), a transmembrane anchor, and a very short cytoplasmic tail.16 The products encoded by HLA-G2, which do not contain the exon 3

sequence, differ from the full-length polypeptides by the absence of the ␣2-domain, and HLA-G3 encodes only one extracellular domain ␣1. The RT-PCR product of HLA-G4, lacking exon 4 coding for the ␣3 domain, cannot be distinguished by size from the HLA-G2 fragment. Direct visualization of the PCR products did not show any HLA-G transcripts in unstimulated myoblasts (see Fig 2B). After stimulation with IFN-␥, the G1 transcript was clearly detectable. The other isoforms yielded only very weak bands or were undetectable. The HLA-G1 band was also dominant in JEG-3, a trophoblast-derived cell line known to express all HLA-G transcripts. After reamplifying individual PCR fragments, we directly sequenced the PCR products obtained from IFN-␥–stimulated myoblasts and obtained sequences corresponding to G1, G2, and G3. To better visualize the additional isoforms, the PCR products were blotted and hybridized with a probe detecting all HLA-G isoforms (see Fig 2C). In noninduced myoblasts, a weak band corresponding to HLA-G1 was visible. In IFN␥–stimulated myoblasts, HLA-G1 transcripts also dominate, but bands corresponding to G3 and G2/G4 are now clearly visible (compare Fig 2C and B). To differentiate between G2 and G4, we hybridized two additional membranes with oligonucleotides specific for G2 and G4. Only the G2-specific probe yielded a positive signal (not shown). Discussion We demonstrate that HLA-G is expressed in muscle of patients with inflammatory myopathy and in IFN-␥– stimulated myoblasts. HLA-G is a nonclassical HLA class I molecule expressed mainly in cytotrophoblasts of the placenta.5,6 Like the classical HLA class I molecules HLA-A, -B, and -C, HLA-G associates with ␤2microglobulin and the transporter associated with antigen processing (TAP). Because HLA-G binds antigenic peptides and CD8, it should be capable of presenting antigenic peptides to T cells in a way similar to the classical HLA class I molecules.5,6

Fig 2. (A) FACS analysis of HLA-G surface expression on cultured myoblasts. Purified unstimulated NCAM ⫹ myoblasts express MHC-I, but are negative for MHC-II and HLA-G (mAb 87G). Weak surface expression of HLA-G is detectable after 48 hours of induction with IFN-␥ (arrows), as shown for myoblast cultures from two different donors (labeled I and II). Histograms show staining with the designated antibodies (shaded) underlaid with isotype controls for HLA-G mAb 87G (IgG2b). (B) RT-PCR of HLA-G transcripts in unstimulated and stimulated myoblasts. Direct visualization of the RT-PCR products from two different myoblast cultures shows no HLA-G transcripts in unstimulated myoblasts (Myobl. I, II). After stimulation with IFN-␥, the HLA-G1 band is visible (Myobl. ⫹ IFN-␥ I, II). As positive controls, cDNA from JEG-3 cells was amplified. Again, only the G1 band is clearly visible, although this cell line is known to express all isoforms of HLA-G. ␤-actin was used as an internal control for the quality of the cDNA (30 cycles). H2O marks the control PCR-reaction without a cDNA template. (C) Hybridization of amplified RT-PCR HLA-G-transcripts. To increase the sensitivity for detection of additional transcripts, PCR products were hybridized with a cDNA probe detecting all isoforms of HLA-G (the three lanes shown in C correspond to lanes 3–5 in B). In stimulated myoblasts, transcripts for HLA-G1, -G2/4, and -G3 are detectable after hybridization. In unstimulated myoblasts (Myobl. I), a weak G1 band is visible.

682

Annals of Neurology

Vol 48

No 4

October 2000

‹

Furthermore, HLA-G has recently been shown to interact with different receptors expressed on different types of lymphocytes, macrophages/monocytes, and dendritic cells, consistent with an immunoregulatory function. However, the precise role of HLA-G remains unknown.

In normal muscle, HLA-G is undetectable by immunohistochemistry. However, in three inflammatory muscle diseases—PM, DM, and IBM—muscle fibers express HLA-G in a distribution closely resembling classical HLA class I, that is, in muscle fibers, inflammatory cells, and capillaries. This is consistent with

Brief Communication: Wiendl et al: HLA-G in Muscle

683

previous reports that lymphocytes17 and monocytes18 can express HLA-G, especially after stimulation with IFN-␥.18 In PM and sporadic IBM, CD8⫹ T cells and macrophages invade initially nonnecrotic muscle fibers. All of the invaded muscle fibers and some noninvaded fibers express classical HLA class I.3,4 This has been taken as evidence for an HLA class I–restricted cytotoxic T cell–mediated response against surface antigens expressed on muscle fibers. Our present observation that muscle fibers coexpress HLA-G raises the intriguing possibility that some of the autoaggressive T cells recognize their antigenic peptides in the molecular context of HLA-G rather than the classical MHC-I molecules. Another, not necessarily exclusive, possibility is that HLA-G protects muscle fibers from NK cell– mediated injury, as has been postulated for fetal cells in the placenta.5,6 Locally accumulating inflammatory cells produce a plethora of proinflammatory cytokines, including TNF-␣ and IFN-␥. Because these cytokines are known to induce HLA expression in many cell types, it seems likely that the HLA-G expression noted in inflammatory myopathies at least partly results from stimulation by locally produced cytokines. We found that IFN-␥, but not TNF-␣ alone, induces HLA-G in cultured myoblasts. HLA-G transcripts are upregulated on the mRNA level. After cytokine treatment, myoblasts coexpress the full-length transcript HLA-G1 and, at a reduced level, the shorter transcripts representing HLA-G2 and HLA-G3. The functional significance of the proteins encoded by the alternatively spliced transcripts is as yet unknown. It is conceivable that HLA-G plays a role in many other immune reactions that occur in muscle.3 In muscle infections, for example, HLA-G might present bacterial or viral peptides to cytotoxic T cells. After vaccination or intramuscular injection of vectors for gene therapy, HLA-G may present peptides of the immunizing antigen or vector-encoded peptides. After therapeutic transfer of myoblasts, HLA-G may be involved in allograft rejection. Note Added in Proof After submission of our paper, expression of HLA-G was demonstrated in human transplanted heart during rejection.19 Supported by the Deutsche Forschungsgemeinschaft, SFB 217, projects C13 and A1; and European Community Grant BMH4CT96-0893. Dr Wiendl holds a postdoctoral fellowship from the Deutsche Forschungsgemeinschaft (Wi 1722/1-1). The Institute for Clinical Neuroimmunology is supported by the Hermann and Lilly Schilling foundation. We thank Drs M. McMaster (University of California, San Francisco) and D. Geraghty (Fred Hutchinson Cancer Research Center,

684

Annals of Neurology

Vol 48

No 4

October 2000

Seattle) for kindly providing anti–HLA-G mAbs, Dr N. Goebels for help with histochemistry, and Ingrid Eiglmeier for excellent technical assistance.

References 1. Karpati G, Pouliot Y, Carpenter S. Expression of immunoreactive major histocompatibility complex products in human skeletal muscles. Ann Neurol 1988;23:64 –72 2. Emslie-Smith AM, Arahata K, Engel AG. Major histocompatibility complex class I antigen expression, immunolocalization of interferon subtypes, and T cell–mediated cytotoxicity in myopathies. Hum Pathol 1989;20:224 –231 3. Hohlfeld R, Engel AG. The immunobiology of muscle. Immunol Today 1994;15:269 –274 4. Dalakas MC. Immunopathogenesis of inflammatory myopathies. Ann Neurol 1995;37:74 – 86 5. Braud VM, Allan DSJ, McMichael AJ. Functions of nonclassical MHC and non–MHC-encoded class I molecules. Curr Opin Immunol 1999;11:100 –108 6. Le Bouteiller P, Blaschitz A. The functionality of HLA-G is emerging. Immunol Rev 1999;167:233–244 7. Goebels N, Michaelis D, Wekerle H, Hohlfeld R. Human myoblasts as antigen presenting cells. J Immunol 1992;149: 661– 667 8. Ulbrecht M, Rehberger B, Strobel I, et al. HLA-G: expression in human keratinocytes in vitro and in human skin in vivo. Eur J Immunol 1994;24:176 –180 9. Maier S, Geraghty DE, Weiss EH. Expression and regulation of HLA-G in human glioma cell lines. Transplant Proc 1999;31: 1849 –1853 10. Engler-Blum G, Meier M, Frank J, Mueller GJ. Reduction of background problems in nonradioactive Northern and Southern blot analyses enables higher sensitivity than 32P based hybridizations. Anal Biochem 1993;210:235–244 11. Lee N, Malacko AR, Ishitani A, et al. The membrane-bound and soluble forms of HLA-G bind identical sets of endogenous peptides but differ with respect to TAP association. Immunity 1995;3:591– 600 12. Crisa L, McMaster MT, Ishii JK, et al. Identification of a thymic epithelial cell subset sharing expression of the class Ib HLA-G molecule with fetal trophoblasts. J Exp Med 1997;186: 289 –298 13. Illa I, Leon-Monzon M, Dalakas MC. Regenerating and denervated human muscle fibers and satellite cells express neural cell adhesion molecule recognized by monoclonal antibodies and natural killer cells. Ann Neurol 1992;31:46 –52 14. Hohlfeld R, Engel AG. Induction of HLA-DR expression on human myoblasts with interferon-␥. Am J Pathol 1990;136: 503–508 15. Mantegazza R, Hughes SM, Mitchell D, et al. Modulation of MHC class II antigen expression in human myoblasts after treatment with IFN-␥. Neurology 1991;41:1128 –1132 16. Carosella ED, Dausset J, Kirszenbaum M. HLA-G revisited. Immunol Today 1996;17:407– 409 17. Amiot L, Onno M, Dre´nou B, et al. Distribution of HLA-G alternative mRNAs including soluble forms in normal lymphocytes and in lymphoid cell derived leukemia. Eur J Immunogenet 1996;23:311–320 18. Yang Y, Chu W, Geraghty DE, Hunt JS. Expression of HLA-G in human mononuclear phagocytes and selective induction by interferon-␥. J Immunol 1996;156:4224 – 4231 19. Lila N, Carpentier A, Amrein C, et al. Implication of HLA-G molecule in heart-graft acceptance. Lancet 2000;355:2138