вход по аккаунту


Anti-Hu antibodies in Merkel cell carcinoma.

код для вставкиСкачать
Adenine Nucleotide
Translocator 1
Deficiency Associated with
Sengers Syndrome
Eric Z. Jordens, PhD,1 Luigi Palmieri, PhD,2
Marjan Huizing, PhD,1
Lambert P. van den Heuvel, PhD,1
Rob C. A. Sengers, MD, PhD,1 Andrea Dörner, PhD,3
Wim Ruitenbeek, PhD,1 Frans J. Trijbels, PhD,1
Jullius Valsson, MD,4 Gunnlaugur Sigfusson, MD,4
Ferdinando Palmieri, MD,2
and Jan A. M. Smeitink, MD, PhD1
Sengers syndrome is characterized by congenital cataracts, hypertrophic cardiomyopathy, mitochondrial myopathy, and lactic acidosis, but no abnormalities have
been found with routine mitochondrial biochemical diagnostics (the determination of pyruvate oxidation rates
and enzyme measurements). In immunoblot analysis, the
protein content of the mitochondrial adenine nucleotide
translocator 1 (ANT1) was found to be strongly reduced
in the muscle tissues of two unrelated patients with Sengers syndrome. In addition, low residual adenine nucleotide translocator activity was detected upon the reconstitution of detergent-solubilized mitochondrial extracts
from the patients’ skeletal or heart muscle into liposomes. Sequence analysis and linkage analysis showed
that ANT1 was not the primary genetic cause of Sengers
syndrome. We propose that transcriptional, translational,
or posttranslational events are responsible for the ANT1
deficiency associated with the syndrome.
Ann Neurol 2002;52:95–99
In many patients with structurally or functionally abnormal mitochondria, the underlying cause of their
condition is unknown. Some of these patients have
Sengers syndrome, which is characterized by congenital
cataracts, hypertrophic cardiomyopathy, mitochondrial
From the 1Department of Pediatrics, University Medical Center
Nijmegen, Nijmegen, the Netherlands; 2Department of Pharmacobiology, Laboratory of Biochemistry and Molecular Biology, University of Bari, Bari, Italy; 3Benjamin Franklin Hospital, Free University, Berlin, Germany; and 4Barnaspitali Hringsins,
Landspitalinn, Reykjavik, Iceland.
Received Sep 14, 2001, and in revised form Feb 6, 2002. Accepted
for publication Feb 14, 2002.
Published online May 31, 2002, in Wiley InterScience
( DOI: 10.1002/ana.10214
Address correspondence to Dr Smeitink, Nijmegen Center for Mitochondrial Disorders, Department of Pediatrics, University Medical
Center Nijmegen, P.O. Box 9101, NL-6500 HB Nijmegen, the
Netherlands. E-mail:
myopathy, and lactic acidosis1 (OMIM no. 103220).
To date, this inherited autosomal recessive syndrome
has been described in approximately 20 patients.2–5 It
is associated with histopathological abnormalities that
include the abnormal arrangement and loss of mitochondrial cristae, crystals within the mitochondrial matrix, deposits of lipids and glycogen, and an increase in
the volume density of mitochondria located beneath
the sarcolemma.1,3
Despite the observed structural abnormalities of mitochondria and lactic acidemia, normal substrate oxidation rates were demonstrated in 600g supernatants of
skeletal muscle of investigated patients.5 Recently, we
focused our attention on transmembrane transport proteins as possible pathogenic candidates in unresolved
mitochondrial cytopathies.6 When we investigated
transmembrane carriers in two unrelated patients with
Sengers syndrome, we found that the amount of adenine nucleotide translocator 1 (ANT1) was specifically
reduced in the muscle tissues of both patients.
Patients and Methods
The patients, both male, fulfilled the criteria for Sengers syndrome originally described by Sengers and colleagues.1 They
came from Iceland (Patient 1) and the Netherlands (Patient
2).5 There was a family relation five and six generations back
between the parents of Patient 1.2 Patient 1 had the adult
form of the syndrome, whereas Patient 2 had the neonatal
form. Muscle tissue from Patient 1 was obtained when the
patient was 17 years old. Muscle tissue from Patient 2 was
obtained postmortem when the patient was 7 days old.
Samples from Patient 1 were collected during life and
used or stored according to established protocols.7 Some tissues from Patient 2 were collected within the first hour after
death and partly used for extensive diagnostic studies. The
remaining tissues were immediately deep-frozen in liquid nitrogen and stored at ⫺70°C until usage.
Immunochemical Studies
Western blot analysis was conducted as described previously.8
Blots were incubated sequentially with an isoform-specific antiserum against amino acids 40 to 54 of human ANT19 and a
polyclonal antiserum against human cytochrome c oxidase 4.
Solubilization and Functional Reconstitution of
Adenine Nucleotide Translocator
Mitochondrial proteins were solubilized in a buffer containing 1.8% Triton X-114 (Aldrich, Milwaukee, WI), (wt/vol)
4mg/ml cardiolipin (Sigma, St. Louis, MO), 10mM Na2SO4,
0.5mM ethylenediaminetetraacetic acid, and 5mM N-2hydroxyethylpiperazine-N-2-ethanesulfonic acid (HEPES;
pH 7.5) at a final concentration of 180␮g of protein per
milliliter. After incubation for 20 minutes at 4°C, the mixture was spun at 138,000g for 20 minutes. The supernatant
(Triton-solubilized extract) immediately was incorporated
into phospholipid vesicles by cyclic removal of the detergent
with a hydrophobic column10 as described in this article.
© 2002 Wiley-Liss, Inc.
The protein content of Triton-solubilized extracts was determined by the Lowry method, which was modified as described previously.11 The composition of the initial mixture
used for reconstitution was as follows: 100␮l of Tritonsolubilized extract (approximately 15␮g of protein), 57␮l of
10% Triton X-114, 100␮l of 10% phospholipids in the
form of sonicated liposomes, 20mM adenosine triphosphate
(ATP), 10mM HEPES (pH 7.5), 0.3mg of cardiolipin, and
water for a final volume of 700␮l. These components were
mixed thoroughly, and the mixture was recycled 13 times
through an Amberlite column (3.2 ⫻ 0.5cm; Superchrom,
Milan, Italy) preequilibrated with a buffer containing 20mM
ATP and 10mM HEPES (pH 7.5). All operations were performed at 4°C, except the passages through the Amberlite
column, which were performed at room temperature.
Transport Studies
The external substrate was removed from proteoliposomes on
a Sephadex G-75 column (Amersham Pharmacia Biotech,
Uppsala, Sweden) preequilibrated with a buffer containing
50mM NaCl and 10mM HEPES (pH 7.5). Transport at
25°C was started by the addition of 0.1mM [14C]-ATP to
the proteoliposomes and was terminated at predetermined
time intervals by the addition of 30mM pyridoxal 5⬘phosphate and 10mM bathophenanthroline (the inhibitorstop method9). The external radioactivity was removed on
the Sephadex G-75 column, and the internal radioactivity
was measured. In control experiments, transport was inhibited by the prior addition of 5␮M bongkrekate and 10␮M
carboxyatractylate. The transport activity was calculated from
the difference between experimental and inhibited experiments. The initial rate of transport was calculated from the
time course of isotope equilibration10 evaluated with a firstorder equation with a computer program (DeltaGraph 4.0
by SPSS, Chicago, IL).
Mutation Analyses and Linkage Studies
Primers (Table 1) were designed on the basis of the human
ANT1 gene sequence.9 The polymerase chain reactions for
linkage analysis were conducted with a Weber mix with the
markers (Fig 1A) according to established procedures.
Measurements of the activities of the respiratory chain
enzymes in muscle samples from Patient 1 were normal. Only complex I activity was slightly decreased (re-
sidual activity was 93% of the lowest control value;
data not shown). Detailed biochemical studies of the
mitochondrial energy generating system, including
measurements of substrate oxidation rates, the ATP
and creatine phosphate production rate from pyruvate,
and the activities of the respiratory chain enzymes, all
of which indicated normal results, have been reported
previously for Patient 2.5
Immunochemical analyses of skeletal muscle specimens from Patients 1 and 2 and heart muscle specimens from Patient 2 indicated a severe decrease in
ANT1 protein (Fig 2). Patient 2 exhibited some aspecificity, which was also found in controls (data not
For a direct test of the functional capacity of adenine
nucleotide translocator (ANT) from different specimens, ANT was solubilized from the 600g supernatants of skeletal or heart muscle homogenates and reconstituted into phospholipid vesicles to determine the
actual transport rates. The ATP/ATP homoexchange
was measured because it is indifferent to membrane
potential. This transport was nearly completely
(⬎95%) abolished by the preincubation with carboxyatractylate and bongkrekic acid. Only very low residual ATP uptake was measured with proteoliposomes
reconstituted with the Triton extract from skeletal
muscle mitochondria of Patient 1. Proteoliposomes reconstituted with an extract from an ANT-deficient
heart specimen (Patient 2) showed a strongly reduced
exchange rate (3–5%) in comparison with a control
sample from the same tissue (C1H; Table 2).
In an attempt to identify the genetic defect underlying the ANT1 deficiency in Sengers syndrome, the
ANT1 exons were amplified by polymerase chain reaction and sequenced from genomic DNA or complementary DNA samples from the patients and a control
subject who did not have mitochondrial myopathy. Although complementary DNA could be used for determining the level of ANT1 expression, the obtained
complementary DNA was just sufficient for sequencing. Only a single prominent band of the expected size
was observed upon polymerase chain reaction amplifi-
Table 1. Polymerase Chain Reaction Primers for ANT1 Exons and cDNA Amplification
cDNA (exons 1–4)
Exon 2
Exon 3
Exon 4
PCR ⫽ polymerase chain reaction; ANT1 ⫽ adenine nucleotide translocator 1.
Annals of Neurology
Vol 52
No 1
July 2002
mained obscure. In this report, we show that in unresolved cases of Sengers syndrome4 ANT1 was deficient.
The extent to which ANT1 antibodies are capable of
cross-reacting with ANT2 or ANT3 was not determined in this study. However, because the predominant ANT isoform in muscle tissue is ANT1,13 our
results clearly indicate a deficiency of ANT1 as a possible cause of the pathological condition of our patients, who did not have mental disorders but had affected heart and skeletal muscle.
Our transport measurements provided direct evidence
Fig 1. Linkage study in the family of Patient 1. (A) Location
of the markers used in relation to the adenine nucleotide
translocator 1 (ANT1) gene. The marker D4S1554 was located 8.5cM centromeric of the ANT1 gene. The markers
ANT1 and D4S171 were located 0.0cM telomeric of the
ANT1 gene. (B) Haplotypes of the ANT1 loci in the parents
and two siblings of Patient 1.
Fig 2. Immunoblot analysis of muscle tissues. Samples (5mU
of cytochrome c oxidase) were separated by sodium dodecyl
sulfate polyacrylamide gel electrophoresis, transferred to a nitrocellulose membrane, and sequentially incubated with antibodies against adenine nucleotide translocator 1 (ANT1) and
cytochrome c oxidase 4 (COX4). (A) Enriched (14,000g) mitochondrial fractions of skeletal muscle homogenates from Patients 1 and 2 (P1 and P2), control infantile muscle (C1;
biopsied at 5 years and stored for 2 years at ⫺70°C), and
control neonatal muscle (C2; biopsied at 1 year and stored for
2 years at ⫺70°C). (B) Enriched (14,000g) mitochondrial
fractions of heart muscle homogenate from Patient 2 (P2),
control heart muscle (C1H), and control skeletal muscle (C3).
Lane M shows a low-range biotinylated molecular weight
marker from Bio-Rad (Richmond, CA). Patient 2 exhibited
some aspecificity, which was also found in controls.
cation, indicating that no aberrant splicing had occurred. Whatever the template, no sequence variations
among the patients, the control, or the Genebank sequence9 could be found.
A deficiency in ANT1 protein could have been
caused by a mutation in the muscle-specific regulatory
sites of the ANT1 promoter. We found that the haplotypes for the ANT1 locus of an affected sister of Patient 1 were heterozygous and identical to those of an
unaffected brother (see Fig 1). Because of the consanguinity of the parents (five and six generations back),2
a linkage between Sengers syndrome and the ANT1
locus is highly unlikely.
In the past 2 decades, substantial progress has been
made in understanding the clinical, morphological,
biochemical, and genetic aspects of mitochondrial cytopathies.5,12 However, the underlying causes of various clinical syndromes associated with mitochondrial
cytopathies, exemplified by Sengers syndrome, have re-
Jordens et al: ANT1 Deficiency and Sengers Syndrome
Table 2. Adenine Nucleotide Translocator Activity Assays
[14C]-ATP uptake
Skeletal muscle
Patient 1
Heart muscle
Patient 2
ATP/min/g protein
0.07 ⫾ 0.09
5.39 ⫾ 1.32
11.7 ⫾ 2.49
5.90 ⫾ 1.72
2.16 ⫾ 2.82
209 ⫾ 51
179 ⫾ 38
181 ⫾ 53
2.49 ⫾ 0.60
53.1 ⫾ 13.8
12.4 ⫾ 2.99
393 ⫾ 102
[14C]-ATP uptake by proteoliposomes reconstituted with Triton extracts of the indicated skeletal or heart muscle samples and preloaded
with 20mM ATP. C1–3 and C1H are from different control subjects
(see text). Transport rates were determined as described in articles. In
controls, 5␮M bongkrekate and 10␮M carboxyatractylate were added
2 minutes before the labeled substrate. Values are given as means ⫾
standard errors of at least five independent experiments.
ATP ⫽ adenosine triphosphate; COX ⫽ cytochrome c oxidase.
of impaired adenine nucleotide transport across the inner membranes of mitochondria from either skeletal or
heart muscle of patients affected by Sengers syndrome.
This defect can be explained by a decrease of ANT1 in
the mitochondrial inner membrane. The higher residual
transport level observed in the heart mitochondria most
likely mirrors the higher level of expression of ANT2 in
the heart than in the skeletal muscle.13
The ANT1 knockout mice reported by Graham and
colleagues14 represent the first animal models for mitochondrial myopathy and cardiomyopathy. The mice,
like the patients with the adult form of Sengers syndrome, showed normal growth and development, lactic
acidemia, hypertrophic cardiomyopathy, and exercise
intolerance. The morphological and enzyme histochemical investigations of the affected sister of Patient
1, who was evaluated at 12 years of age,2 and Patient
23 agreed with the findings described in the 4 to
6-month-old mice.14
Despite the observed ANT1 deficiency, the biochemical investigations pointed to normal functioning
of the respiratory chain in both patients in our study.
In contrast, skeletal muscle mitochondria from the 4 to
6-month old mice showed a clear disturbance of state 3
respiration rates.14 These discrepancies might have resulted from differences between mouse and human
ANT expression patterns, as well as the different conditions under which these biochemical investigations
were performed.
Linkage studies have argued against the involvement
of the ANT1 locus as the primary cause of the pathogenesis of Sengers syndrome. This is consistent with mutation analyses. Progressive external ophthalmoplegia was
not a clinical sign in our patients.15 Impairment of
ANT1 transcription may explain the faulty functioning
Annals of Neurology
Vol 52
No 1
July 2002
of heart and skeletal muscle. Another possibility is defective targeting of the ANT1 protein to the mitochondrial inner membrane by chaperones, the translocases of
outer mitochondrial membranes/translocases of inner
mitochondrial membranes complex, or both.16,17
Regardless of the genetic defect underlying Sengers
syndrome, the ANT1 protein deficiency demonstrated
in this study appears to be causally related to the clinical symptoms, tissue specificity, and characteristic
morphological features. To our knowledge, this is the
first time that a specific mitochondrial carrier protein,
but not its gene, has been demonstrated to be crucially
involved in an inherited mitochondrial disease.
This study was supported by the Prinses Beatrix Fonds (98-0204,
F.J.T., L.P.H., and J.A.M.S.) and Fondazione Telethon-Italy
(E.0958, F.P.).
We are grateful to Antoon Janssen for technical assistance. We also
thank E. Holme, A. Oldfors, M. Tulinius, B. Steinmann, and H. P.
Schultheiss for fruitful discussions and B. Kadenbach for a supply of
a polyclonal antiserum against human cytochrome c oxidase 4.
1. Sengers RC, Trijbels JM, Willems JL, et al. Congenital cataract
and mitochondrial myopathy of skeletal and heart muscle associated with lactic acidosis after exercise. J Pediatr 1975;86:
873– 880.
2. Valsson J, Laxdal T, Jonsson A, et al. Congenital cardiomyopathy and cataracts with lactic acidosis. Am J Cardiol 1988;61:
3. Ruitenbeek W, Huizing M, DePinto, V, et al. Defects of mitochondrial membrane-bound transport proteins in human mitochondrial myopathies: a biochemical approach. In: Palmieri
F, editor. Progress in cell research. Vol 5. Amsterdam: Elsevier
Science, 1995:225–229.
4. van Ekeren GJ, Stadhouders AM, Smeitink JA, Sengers RC. A
retrospective study of patients with the hereditary syndrome of
congenital cataract, mitochondrial myopathy of heart and skeletal
muscle and lactic acidosis. Eur J Pediatr 1993;152:255–259.
5. Smeitink JA, Sengers RC, Trijbels JM, et al. Fatal neonatal cardiomyopathy associated with cataract and mitochondrial myopathy. Eur J Pediatr 1989;148:656 – 659.
6. Trijbels FJ, Ruitenbeek W, Huizing M, et al. Defects in the
mitochondrial energy metabolism outside the respiratory chain
and the pyruvate dehydrogenase complex. Mol Cell Biochem
7. Tulinius MH, Holme E, Kristiansson B, et al. Mitochondrial
encephalomyopathies in childhood. I. Biochemical and morphologic investigations. J Pediatr 1991;119:242–250.
8. Huizing M, DePinto V, Ruitenbeek W, et al. Importance of
mitochondrial transmembrane processes in human mitochondriopathies. J Bioenerg Biomembr 1996;28:109 –114.
9. Li K, Warner CK, Hodge JA, et al. A human muscle adenine
nucleotide translocator gene has four exons, is located on chromosome 4, and is differentially expressed. J Biol Chem 1989;
264:13998 –14004.
10. Palmieri F, Indiveri C, Bisaccia F, Iacobazzi V. Mitochondrial
metabolite carrier proteins: purification, reconstitution, and
transport studies. Methods Enzymol 1995;260:349 –369.
11. Capobianco L, Bisaccia F, Mazzeo M, Palmieri F. The mitochondrial oxoglutarate carrier: sulphydryl reagents bind to
cysteine-184, and this interaction is enhanced by substrate
binding. Biochemistry 1996;35:8974 – 8980.
12. Zeviani M, Fernandez-Silva P, Tiranti V. Disorders of mitochondria and related metabolism. Curr Opin Neurol 1997;10:160–167.
13. Doerner A, Pauschinger M, Badorff A, et al. Tissue-specific
transcription pattern of the adenine nucleotide translocase isoforms in humans. FEBS Lett 1997;414:258 –262.
14. Graham BH, Waymire KG, Cottrell B, et al. A mouse model
for mitochondrial myopathy and cardiomyopathy resulting
from a deficiency in the heart/muscle isoform of the adenine
nucleotide translocator. Nat Genet 1997;16:226 –234.
15. Kaukonen J, Juselius JK, Tiranti V, et al. Role of adenine nucleotide translocator 1 in mtDNA maintenance. Science 2000;
16. Leuenberger D, Bally NA, Schatz G, Koehler CM. Different import pathways through the mitochondrial intermembrane space
for inner membrane proteins. EMBO J 1999;18:4816 – 4822.
17. Grad LI, Descheneau AT, Neupert W, et al. Inactivation of the
Neurospora crassa mitochondrial outer membrane protein
TOM70 by repeat-induced point mutation (RIP) causes defects
in mitochondrial protein import and morphology. Curr Genet
Angiostrongylus cantonensis
Infection Mimicking a
Spinal Cord Tumor
Supinda Petjom, MD,1 Benjaporn Chaiwun, MD,1
Jongkolnee Settakorn, MD,1 Pannee Visrutaratna, MD,2
Samreung Rangdaeng, MD,1
and Paul S. Thorner, MD, PhD3
Angiostrongylus cantonensis is the most common cause of
eosinophilic meningitis and meningoencephalitis. Almost
all cases are self-limiting and are diagnosed by cerebrospinal fluid eosinophilia and enzyme-linked immunosorbent assay; pathology reports are restricted to postmortem samples from lethal cases. We report on what we
believe is the first case of A. cantonensis infection diagnosed by biopsy in a living patient. The spinal cord was
biopsied because of the unusual clinical presentation of a
myelopathy without meningeal symptoms, together with
a mass lesion that was clinically and radiologically diagnosed as a spinal cord tumor.
Ann Neurol 2002;52:99 –102
From the Departments of 1Pathology and 2Radiology, Faculty of
Medicine, Chiang Mai University, Chiang Mai, Thailand; and 3Department of Laboratory Medicine and Pathobiology, Hospital for
Sick Children and University of Toronto, Toronto, Canada.
Received Nov 27, 2001, and in revised form Feb 13, 2002. Accepted for publication Feb 14, 2002.
Published online Jun 18, 2002, in Wiley InterScience
( DOI: 10.1002/ana.10215
Address correspondence to Dr Chaiwun, Department of Pathology,
Faculty of Medicine, Chiang Mai University, Chiang Mai 50200,
Thailand. E-mail:
Angiostrongylus cantonensis is the most common cause
of eosinophilic meningitis and meningoencephalitis in
the world and is endemic in Southeast Asia.1– 4 The
causative agent is the rat lung worm A. cantonensis, primarily a parasite of rodents. Infective third-stage larvae
develop in slugs and snails. Humans are infected primarily in the central nervous system after ingesting an
infected intermediate host,2 such as a land snail or a
crustacean.5,6 The disease usually manifests as headache, neck stiffness, paresthesia, and generalized
weakness.1–3,6,7 The clinical course in most cases is
self-limited; however, fatal cases have been reported.5,6 Virtually all tissue diagnoses of this disease have
been postmortem; therefore, most diagnoses are presumptive, based on clinical findings, cerebrospinal
fluid (CSF) eosinophilia, and positive serological testing.
We report on an unusual case of A. cantonensis infection in a Thai male who presented with a myelopathy rather than meningoencephalitis. The working diagnosis was an intramedullary tumor of the spinal
cord, which resulted in the patient’s undergoing surgery with biopsy of the spinal cord. The worm was
identified in the biopsy. Not only is the clinical presentation very atypical, but this case provides what we
believe is the first tissue diagnosis of this infection before death.
Case Report
A 26-year-old Thai man was admitted with a history of paresthesia of the left lower back and left leg for 5 months,
associated with anorexia and weight loss. By the time of admission, he had weakness and paresthesia of both legs. He
denied ingestion of snails or prawns, but he did eat uncooked vegetables. On physical examination, he was alert
and afebrile. There was no neck stiffness, and cranial nerve
function was normal. Neurological assessment of the upper
extremities was normal. However, there was marked weakness of the lower extremities associated with hyperreflexia
and a left extensor plantar response. Decreased sensation in
response to pain and light touching was noted below the
spinal level of T10. Laboratory testing indicated a normal
white blood cell count of 8,900/mm3 with mild eosinophilia
of 6.4% (normal ⫽ 1–3%) and normal blood chemistry and
urinalysis. A spinal tap failed. Chest and lumbosacral X-rays
were normal, but magnetic resonance imaging showed enlargement of the spinal cord by a mass lesion with a heterogeneous signal intensity, extending from T7 to T10 (Fig 1).
The interpretation favored an intramedullary spinal cord tumor. The patient underwent a T8 to T9 laminectomy; the
thoracic cord was found to be necrotic and was biopsied.
Postoperatively, the patient developed progressive weakness
of his lower limbs. Over the next 6 months, there was some
improvement in neurological function; he was lost to
follow-up thereafter.
© 2002 Wiley-Liss, Inc.
Fig 1. Magnetic resonance imaging indicates an intramedullary spinal cord lesion extending from T7 to T10 with heterogeneous signal intensity.
The spinal cord biopsy contained multiple sections of
partially degraded parasitic larvae surrounded by necrotic tissue and inflammatory cells. Most were lymphocytes and plasma cells, but eosinophils were readily
identified (Fig 2). Also present were poorly formed
granulomata and rare multinucleated giant cells. The
cross sections of the worm showed a smooth cuticle
with lateral projections (cords) along the interior aspect
(Fig 3). The internal organ structure was degenerated.
The average cross-sectional diameter was about 200 to
250␮m. The appearance, size, and presence of lateral
cords allowed the worm to be identified as A. cantonensis.8 The final diagnosis was A. cantonensis myelitis.
The classic clinical picture of A. cantonensis infection
is meningoencephalitis with headache, neck stiffness,
paresthesia, and generalized weakness without focal
neurological signs.1–3,6,7 In up to 25% of cases, there
is visual impairment and extraocular muscular paralysis.1,9,10 Fever is variable.1,9 Myelitis has only rarely
been reported.11–14 The first 3 cases had myelitis in
combination with meningoencephalitis, whereas the
fourth case, like our patient, presented purely with spinal cord deficits, a clinical picture that is distinctly unusual.
Most patients with A. cantonensis infection give a
history of having consumed raw or undercooked Pila
and Achatina snails, both intermediate hosts of A. cantonensis. About one third of Achatina snails in northeastern Thailand are infected with this parasite.15 Our
patient denied consuming snails, but he did eat raw
vegetables, and it has been reported that mucus left
behind by snails on plants can also contain the parasite.16 The number of parasites required for infection
may not be very large. Low doses of larvae from rats
were still found to result in infection.17
The diagnosis is usually made clinically, in conjunction with blood and CSF eosinophilia1,7,9,10,13 and
specific serological testing by enzyme-linked immunosorbent assay when available.3,7,10,13,14 Before the
development of enzyme-linked immunosorbent assay,
almost all diagnoses were presumptive, based on the
finding of CSF eosinophilia.6 Meningoencephalitis was
presumed to be due to A. cantonensis,1,18 whereas myeloencephalitis was presumed to be due to Gnathostoma
spinigerum, another roundworm.19 It was, therefore,
Fig 2. Microscopic appearance of the spinal cord biopsy. (A) The spinal cord shows
areas of necrosis with an inflammatory
reaction in association with cross sections
of the parasite (arrows) (hematoxylin and
eosin; ⫻65 before 17% reduction). (B)
The inflammation consists of lymphocytes
and plasma cells with some eosinophils
(arrows) and granuloma formation (hematoxylin and eosin; ⫻125 before 17%
Annals of Neurology
Vol 52
No 1
July 2002
Fig 3. Higher magnification of the parasite shows a diameter of 200 to 250␮m
with a smooth cuticle and the presence of
a lateral cord (LC) typical of Angiostrongylus. The internal organ structure is degenerated (hematoxylin and eosin;
⫻250 before 16% reduction).
impossible to determine the true incidence of myelitis
in A. cantonensis infections from the literature we reviewed.
The diagnosis remains presumptive unless the parasite can actually be identified, and only occasionally
is this the case. In living patients, there are reports of
the larvae being identified in CSF,10,18,20 including
one series that found the parasite in 42% of spinal
taps.9 There is a single report of a worm seen on fundoscopy in a patient who later recovered.21 Virtually
all demonstrations of the worm in tissue have been in
fatal cases at autopsy.11–13,18 To the best of our
knowledge, a tissue diagnosis of A. cantonensis infection in a living patient has not previously been reported.
The larvae of A. cantonensis invade the central nervous system of mammalian hosts as part of their life
cycle.22 Studies of experimentally infected rats have
shown that the maximum number of larvae are found
in the cerebral hemispheres and medulla.17 Pathophysiologically, there are several possible mechanisms
for the neurological damage in cases of angiostrongyliasis: (1) penetrating damage from motile
worms; (2) inflammatory response to the live worm,
degenerated larval remains, or both; (3) release of
toxic worm metabolites; and (4) vascular insufficiency
caused by vascular invasion by worms.3,11 Our case
provides one of the rare demonstrations of the tissue
changes caused by this parasite in the human central
nervous system. Degenerated worms were associated
with tissue necrosis and a lymphoplasmacytic inflammatory infiltrate, rather than a predominance of eosinophils. This finding correlates with the mild degree
of eosinophilia in the peripheral blood and indicates
that marked peripheral eosinophilia, although a helpful clue toward making the diagnosis, is not always
present. Other postmortem reports describe eosinophilic inflammation with abscess formation and multiple tracks and cavities containing larvae.11–13 In
these fatal cases, both the brain and spinal cord contained the larvae, suggesting that spinal cord involvement in A. cantonensis infection is seen only in the
most severe cases.
In conclusion, we present a case of A. cantonensis
infection that is unusual in its clinical presentation of
myelitis rather than meningoencephalitis. This, together with the magnetic resonance imaging results, led
to the mistaken diagnosis of an intramedullary spinal
cord tumor. In more typical cases, the diagnosis would
be made by CSF examination and enzyme-linked immunosorbent assay, and a biopsy would not be performed. In our case, surgical biopsy at laminectomy established the diagnosis of A. cantonensis myelitis, and
we believe this is the first time that this infection has
been diagnosed by biopsy in a living patient. The neurological findings were directly attributable to spinal
cord invasion by the worm and the associated necrosis
and inflammatory response. This case illustrates that
the clinical spectrum of this parasitic infection is more
varied than usually thought. A. cantonensis infection
should be considered in the differential diagnosis of a
central nervous system mass lesion, especially in endemic areas, but perhaps outside these areas as well,
given that the geographical distribution of this parasite
has now expanded beyond Southeast Asia into India, Africa, the Caribbean, and the southeastern United States.4
Petjom et al: Angiostrongylus cantonensis Myelitis
1. Punyagupta S, Juttijudata P, Bunnag T. Eosinophilic meningitis in Thailand. Clinical studies of 484 typical cases probably
caused by Angiostrongylus cantonensis. Am J Trop Med Hyg
2. Jindrak K. Angiostrongyliasis cantonensis (eosinophilic meningitis, Alicata’s disease). Contemp Neurol Ser 1975;12:133–164.
3. Koo J, Pien F, Kliks M. Angiostrongylus (Parastrongylus) eosinophilic meningitis. Rev Infect Dis 1988;10:1155–1162.
4. Prociv P, Spratt D, Carlisle M. Neuro-angiostrongyliasis: unresolved issues. Int J Parasitol 2000;30:1295–1303.
5. Vejjajiva A. Parasitic diseases of the nervous system in Thailand.
Clin Exp Neurol 1978;15:92–97.
6. Kuberski T, Wallace G. Clinical manifestations of eosinophilic
meningitis due to Angiostrongylus cantonensis. Neurology
1979;29:1566 –1570.
7. Schmutzhard E, Boongird P, Vejjajiva A. Eosinophilic meningitis and radiculomyelitis in Thailand, caused by CNS invasion
of Gnathostoma spinigerum and Angiostrongylus cantonensis.
J Neurol Neurosurg Psychiatry 1988;51:80 – 87.
8. Dooley J, Neafie R. Angiostrongylus cantonensis infections. In:
Binford C, Connor D, eds. Pathology of tropical and extraordinary diseases. Washington, DC: Armed Forces Institute of
Pathology, 1976:446 – 451.
9. Hwang K, Chen E. Clinical studies on angiostrongyliasis cantonensis among children in Taiwan. Southeast Asian J Trop
Med Public Health 1991;22(suppl):194 –199.
10. Tsai H, Liu Y, Kunin C, et al. Eosinophilic meningitis caused
by Angiostrongylus cantonensis: report of 17 cases. Am J Med
2001;111:109 –114.
11. Cooke-Yarborough C, Kornberg A, Hogg G, et al. A fatal case
of angiostrongyliasis in an 11-month-old infant. Med J Aust
12. Witoonpanich R, Chuahirun S, Soranastaporn S, Rojanasunan
P. Eosinophilic myelomeningoencephalitis caused by Angiostrongylus cantonensis: a report of three cases. Southeast
Asian J Trop Med Public Health 1991;22:262–267.
13. Kliks M, Kroenke K, Hardman J. Eosinophilic
radiculomyeloencephalitis: an angiostrongyliasis outbreak in
American Samoa related to ingestion of Achatina fulica snails.
Am J Trop Med Hyg 1982;31:1114 –1122.
14. Wood G, Delamont S, Whitby M, Boyle R. Spinal sensory radiculopathy due to Angiostrongylus cantonensis infection. Postgrad Med J 1991;67:70 –72.
15. Pipitgool V, Sithithaworn P, Pongmuttasaya P, Hinz E. Angiostrongylus infections in rats and snails in northeast Thailand.
Southeast Asian J Trop Med Public Health 1997;28(suppl 1):
190 –193.
16. Waisberg J, Corsi C, Rebelo M, et al. Jejunal perforation
caused by abdominal angiostrongyliasis. Rev Inst Med Trop
Sao Paulo 1999;41:325–328.
17. Limaye L, Bhopale M, Renapurkar D, Sharma K. The distribution of Angiostrongylus cantonensis (Chen) in the central
nervous system of laboratory rats. Folia Parasitol (Praha) 1983;
18. Yii C. Clinical observations on eosinophilic meningitis and meningoencephalitis caused by Angiostrongylus cantonensis on
Taiwan. Am J Trop Med Hyg 1976;25:233–249.
19. Punyagupta S, Bunnag T, Juttijudata P. Eosinophilic meningitis in Thailand. Clinical and epidemiological characteristics of
162 patients with myeloencephalitis probably caused by Gnathostoma spinigerum. J Neurol Sci 1990;96:241–256.
20. Kuberski T, Bart R, Briley J, Rosen L. Recovery of Angiostrongylus cantonensis from cerebrospinal fluid of a child
with eosinophilic meningitis. J Clin Microbiol 1979;9:
629 – 631.
© 2002 Wiley-Liss, Inc.
21. Nelson R, Warren R, Scotti F, et al. Ocular angiostrongyliasis
in Japan: a case report. Am J Trop Med Hyg 1988;38:
130 –132.
22. Nishimura K, Hung T. Current views on geographic distribution and modes of infection of neurohelminthic diseases.
J Neurol Sci 1997;145:5–14.
Microscopic R2* Mapping
of Reduced Brain Iron in
the Belgrade Rat
Holly A. Zywicke, BS,1 Peter van Gelderen, PhD,2
James R. Connor, PhD,3 Joseph R. Burdo, BS,3
Michael D. Garrick, PhD,4 Kevin G. Dolan, BS,4
Joseph A. Frank, MD,1 and Jeff W. M. Bulte, PhD1
R2* mapping has recently been used to detect iron overload in patients with movement disorders. We demonstrate here that this technique can also be used to detect
reduced brain iron, as in the case of a missense mutation
in the iron-transporting protein divalent metal transporter 1. Surprisingly, we found that the same brain regions are affected (ie, the globus pallidus, substantia
nigra, and cerebellar dentate nucleus); this suggests a
much more extensive role for these structures in regulating overall brain iron homeostasis. Therefore, for the
clinical monitoring of movement disorders for which
normal brain iron homeostasis (either overload or reduction) may be implicated, R2* mapping appears to be
Ann Neurol 2002;52:102–105
In the brain, iron is an essential element for basic cellular processes such as myelination, neurotransmitter
production, and adenosine triphosphate synthesis. Evidence has now accumulated1 for an association be-
From the 1Laboratory of Diagnostic Radiology Research, Clinical
Center, and 2Laboratory of Advanced Magnetic Resonance Imaging,
National Institute of Neurological Diseases and Stroke, National Institutes of Health, Bethesda, MD; 3Department of Neuroscience
and Anatomy, Pennsylvania State University, Hershey, PA; and
Department of Biochemistry, State University of New York, Buffalo, NY.
Received Oct 30, 2001, and in revised form Feb 15, 2002. Accepted
for publication Feb 15, 2002.
Published online May 31, 2002, in Wiley InterScience
( DOI: 10.1002/ana.10216
Address correspondence to Dr Bulte, Department of Radiology,
Johns Hopkins University School of Medicine, 217 Traylor Building, 720 Rutland Avenue, Baltimore, MD 21205-2195.
tween excess brain iron and a number of common neurodegenerative disorders, such as Parkinson’s disease,
Huntington’s disease, Friedreich’s ataxia, Alzheimer’s
disease, epilepsy, Hallervorden-Spatz disease,2 and neuroferritinopathy.3 Elucidating the exact biochemical
pathways of brain iron delivery, uptake, and intracerebral transport is of paramount importance for a better
understanding of the role of iron in the underlying pathologies of these disorders.
It is currently believed that iron transport to the
brain is mediated by transferrin and the transferrin receptor. Upon the endocytosis of transferrin-irontransferrin receptor and the release of iron in the acidic
endosomes, iron is translocated to the cellular cytoplasm by divalent metal transporter 1 (DMT1).4 At
present, very little is known about intracerebral intercellular iron transport beyond the endothelium. The
Belgrade (b/b) rat5 has a G185R mutation in DMT1,
resulting in hypochromic, microcytic anemia with the
presence of mitochondria that contain less iron.6 Less
iron gets to mitochondria because of a defect in the
exit of iron from endosomes during the transferrin cycle.7 The defect also affects transferrin-independent
iron uptake in the intestinal tract8 and elsewhere.9 The
b/b rat model has, therefore, been used to determine
the effect of inactive DMT1 on cellular iron distribution in the brain.10 It has been determined that Belgrade rats have fewer and less dense iron-positive pyramidal neurons and oligodendrocytes, indicating the
importance of DMT1 in maintaining normal brain
iron homeostasis and suggesting a role for brain iron
Although these and other reports rely on traditional
histochemical staining techniques for the detection of
iron (ie, diaminobenzidine-enhanced Perls’ reaction),
recent advances in magnetic resonance imaging have
shown that it is possible to detect iron noninvasively.
The antiferromagnetic Fe(III) oxyhydroxide cores of
the protein ferritin form local microscopic field inhomogeneities that enable faster T2 and T2* proton relaxation. On standard spin-echo and gradient-echo
magnetic resonance imaging sequences, the net result is
an area of hypointensity for the iron-containing tissue,
but in general these methods are prone to artifacts, lack
specificity for iron (other tissue components may cause
hypointensity), and are semiquantitative at best. The
rate of the transverse relaxation, R2, is proportional to
the local iron concentration and increases linearly with
the field. Therefore, the field-dependent R2 increase
may be used as a specific parameter for the quantification of tissue iron,11,12 but its clinical use requires at
least two instruments. R2* refers to T2 relaxation without pulse compensation for proton dephasing that occurs in the vicinity of microscopic field gradients, and
it is most sensitive to the presence of iron. Although
other tissue components may contribute to R2*, R2*
mapping has recently been used to detect brain iron
overload in the basal ganglia of patients with Parkinson’s disease13 and in the cerebellar dentate nucleus in
Friedreich’s ataxia.14 Although this technique has been
proven to be able to detect excess brain iron, its suitability for detecting diminished brain iron (eg, for the
Belgrade rat) remains unproven. We report here that
R2* mapping may also be used to provide uniquely
sensitive and quantitative information about reduced
brain iron.
Materials and Methods
Brains from b/b Fischer/Wistar rats, ages 26 to 28 months
(n ⫽ 3), and age-matched ⫹/b littermate wild-type controls
(n ⫽ 3) were perfused with 4% paraformaldehyde and placed
in 12ml syringes filled with a perfluoropolyether (devoid of
proton signals). Three-dimensional, multigradient-echo magnetic resonance images were obtained at a 104␮m isotropic
resolution with a 4.7T Varian INOVA NMR spectrometer
(Palo Alto, CA) and a DOTY Litz coil 25mm in diameter.
The scan parameters were as follows: field of view, 30 ⫻
20 ⫻ 20mm; matrix, 288 ⫻ 192 ⫻ 192; number of excitations, 12; repetition time, 100msec; echo time, 6msec; number of echoes, 6; and flip angle, 15 degrees. Interactive Data
Language (IDL) processing software was then used to create
images from the three-dimensional data set. From the amplitude images of each different echo, R2* maps were calculated
by the fitting of an exponential decay to every voxel with
sufficient intensity.
For each region of interest in both hemispheres, the average R2* from three adjacent slices was calculated with a fixed
number (n) of pixels. This was done for the globus pallidus
(n ⫽ 425), substantia nigra (n ⫽ 200), dentate nucleus (n ⫽
225), normal cortical gray matter (n ⫽ 325), and normal
white matter (ventral hippocampal commissure, n ⫽ 250).
The mean R2* values for both hemispheres were used because the difference between the two sites was less than or
equal to 5%. R2* data were analyzed for statistical significance with a two-tailed Student t test.
Figure 1 shows the R2* maps of mutant animals versus
wild-type animals. For histopathological correlation, a
side-by-side comparison is shown for deparaffinated
10␮m tissue sections stained for ferric iron with a
diaminobenzidine-enhanced Perls’ reaction. R2* maps
are shown as intensity maps; that is, areas with higher
R2* values (more iron) display higher signal intensity.
To obtain quantitative information on the relative
amount of iron, we calculated the average R2* for each
region of interest. Figure 2 shows a striking and significant ( p ⬍ 0.001) difference between the two groups
for the globus pallidus, substantia nigra, and dentate
nucleus, whereas normal gray and white matter were
not affected ( p ⬎ 0.05).
We were surprised to observe that the same brain regions commonly reported to be involved in iron over-
Zywicke et al: DMT1 and Reduced Brain Iron
Fig 1. Microscopic R2* maps and diaminobenzidine (DAB)-enhanced Prussian blue stains of ⫹/ b wild-type control and b/ b rats.
Arrows indicate the outline of the globus pallidus (GP), substantia nigra (SN), and cerebellar dentate nucleus (CDN).
load (the globus pallidus, substantia nigra, and dentate
nucleus) are apparently also specifically implicated in
iron reduction due to impaired DMT1. This parallel
implicates these regions in a much more extensive role
in regulating overall brain iron homeostasis, serving not
only as a well-established depot for storing excess iron
but also as centers that become depleted when intracellular (endosomal) DMT1-dependent iron trafficking is
blocked. It is at present unclear whether this is the direct result of reduced iron uptake in these regions; it
could also result from an increased demand for iron
from other areas of the brain, which could accelerate
depletion. In Belgrade rats, there is a decrease in imFig 2. Calculated average R2* values for the globus pallidus
(GP), substantia nigra (SN), and cerebellar dentate nucleus
(CDN) of ⫹/ b and b/ b rats. The R2* values for the unaffected areas of normal white matter (NWM) and gray matter
(NGM) are shown as internal controls.
Annals of Neurology
Vol 52
No 1
July 2002
munodetectable DMT1 on endothelial cells in comparison with normal rats.15
It is noteworthy that R2* mapping of patients with
restless legs syndrome recently indicated significantly
decreased R2* values in the substantia nigra and to a
lesser extent in the putamen, which was speculated to
be the result of brain iron insufficiency.16 This study
shows a direct, specific correlation between reduced
brain iron content and R2* values. Of further significance is our observation that, contrary to the results of
a clinical restless legs syndrome study,16 there was a
significant reduction in R2* values of the dentate nucleus in Belgrade rats. Our studies were carried out at a
higher field strength (4.7T) than that of most clinical
scanners (3.0 and 1.5T), and because 1/T2 of ferritin
increases linearly with the field,17 this higher field
strength may be more sensitive in detecting the presence of iron. Our data clearly indicate that Belgrade
rats should be carefully observed for any symptoms associated with restless legs syndrome, and that it should
also be determined whether DMT1 expression and
iron transport are compromised in restless legs syndrome patients.18 Therefore, we believe that this rat
model may serve as a basis for further study of this
prevalent movement disorder, which affects between 5
and 15% of adults.19,20
In summary, these results demonstrate that magnetic
resonance imaging in conjunction with R2* mapping
can be used to pinpoint, with microscopic precision,
tissue areas that are most affected by a missense mutation in the iron-transporting protein DMT1. In addition to detecting iron overload, R2* mapping may,
therefore, also be used to visualize reduced brain iron,
opening the door to noninvasive monitoring of brain
iron homeostasis at regional and cellular levels. Because
this can be done noninvasively and serially (repeatedly
over time), this technique appears well suited for the
clinical monitoring of movement disorders for which
normal brain iron homeostasis may be implicated, be it
an excess or insufficiency.
1. Thompson KJ, Shoham S, Connor JR. Iron and neurodegenerative diseases. Brain Res Bull 2001;55:155–164.
2. Zhou B, Westaway SK, Levinson B, et al. A novel pantothenate
kinase gene (PNK2) is defective in Hallervorden-Spatz syndrome. Nat Genet 2001;28:345–349.
3. Curtis ARJ, Fey C, Morris CM, et al. Mutation in the gene
encoding ferritin light polypeptide causes dominant adult-onset
basal ganglia disease. Nat Genet 2001;28:350 –354.
4. Fleming MD, Romano MA, Su MA, et al. Nramp2 is mutated
in the anemic Belgrade (b) rat: evidence of a role for Nramp2
in endosomal iron transport. Proc Natl Acad Sci U S A 1998;
95:1148 –1153.
5. Sladic-Simic D, Zivkovic N, Pavic D, et al. Hereditary hypochromic microcytic anemia in the laboratory rat. Genetics
1966;53:1079 –1089.
6. Garrick LM, Gniecko K, Liu Y, et al. Iron distribution in Belgrade rat reticulocytes after inhibition of heme synthesis with
succinylacetone. Blood 1993;81:3414 –3421.
7. Garrick MD, Gniecko K, Liu Y, et al. Transferrin and the
transferrin cycle in Belgrade rat reticulocytes. J Biol Chem
8. Oates PS, Morgan EH. Defective iron uptake by the duodenum of Belgrade rats fed diets of different iron contents. Am J
Physiol Gastrointest Liver Physiol 1996;270:G826 –G832.
9. Garrick LM, Dolan KG, Romano MA, Garrick MD. Nontransferrin-bound iron uptake in Belgrade and normal rat erythroid cells. J Cell Physiol 1999;178:349 –358.
10. Burdo JR, Martin J, Menzies SL, et al. Cellular distribution of
iron in the brain of the Belgrade rat. Neuroscience 1999;93:
1189 –1196.
11. Bartzokis G, Aravagiri M, Oldendorf WH, et al. Fielddependent transverse relaxation rate increase may be a specific
measure of tissue iron stores. Magn Reson Med 1993;29:
459 – 464.
12. Bulte JWM, Miller GM, Vymazal J, et al. Hepatic hemosiderosis in non-human primates: quantification of liver iron using
different field strengths. Magn Reson Med 1997;37:530 –536.
13. Ye FQ, Martin WRW, Allen PS. Estimation of brain iron in
vivo by means of the interecho time dependence of image contrast. Magn Reson Med 1996;36:153–158.
14. Waldvogel D, van Gelderen P, Hallett M. Increased iron in the
dentate nucleus of patients with Friedreich’s ataxia. Ann Neurol
15. Burdo JR, Menzies SL, Simpson IA, et al. Distribution of divalent metal transporter 1 and metal transport protein 1 in the
normal and Belgrade rat. J Neurosci Res 2001;66:1198 –1207.
16. Allen RP, Barker PB, Wehrli F, et al. MRI measurement of
brain iron in patients with restless legs syndrome. Neurology
17. Vymazal J, Brooks RA, Zak O, et al. T1 and T2 of ferritin at
different field strengths: effect on MRI. Magn Reson Med
1992;27:368 –374.
18. Earley CJ, Allen RP, Beard JL, Connor JR. Insight into the
pathophysiology of restless legs syndrome. J Neurosci Res 2000;
62:623– 628.
19. Lavigne GJ, Montplaisir JY. Restless legs syndrome and sleep
bruxism: prevalence and association among Canadians. Sleep
1994;17;739 –743.
20. Phillips B, Young T, Finn L, et al. Epidemiology of restless legs
symptoms in adults. Arch Intern Med 2000;160:2137–2141.
Preferential Loss of Paternal
19q, but Not 1p, Alleles in
Marc Sanson, MD, PhD,1,2 Pascal Leuraud, MSc,2
Yannick Marie, MSc,2 Jean-Yves Delattre, MD,1,2
and Khê Hoang-Xuan, MD1,2
Portions of chromosomes 1p and 19q, which are frequently deleted in oligodendrogliomas, are subject to
genomic imprinting, suggesting that the putative tumor
suppressor genes could be monoallelically expressed. The
parental origins of 1p and 19q allele losses were determined in 6 cases of pure oligodendroglioma. An equilibrated parental loss (3 maternal and 3 paternal) was
found for 1p deletions. In contrast, 19q deletions always
occurred on the paternal copy ( p ⴝ 0.015). In this setting, a cloning strategy based on a search for homozygous
deletion or mutation of the remaining allele would be
appropriate for identifying the tumor suppressor gene located on 1p but inappropriate for identifying the presumably monoallelically expressed tumor suppressor gene
located on 19q.
Ann Neurol 2002;52:105–107
Combined losses of 1p and 19q chromosomes occur in
50 to 70% of patients with oligodendroglioma. This
finding is important, not only as a diagnostic marker of
oligodendroglioma but also as an indicator of higher
chemosensitivity and prolonged survival.1 Strategies for
cloning the putative tumor suppressor genes located on
these chromosomes are being developed, based on a
search for homozygous deletion or mutation of the remaining allele. However, this approach may not be
From 1Fédération Neurologique Mazarin and 2Unité INSERM
U495, Hôpital de la Salpêtrière, Université Pierre et Marie Curie,
Paris, France.
Received Jan 2, 2002, and in revised form Feb 15. Accepted for
publication Feb 15, 2002.
Published online May 31, 2002, in Wiley InterScience
( DOI: 10.1002/ana.10217
Address correspondence to Dr Sanson, Fédération Neurologique
Mazarin, Hôpital de la Salpêtrière, 47 Bd de l’Hôpital, 75013 Paris,
France. E-mail:
© 2002 Wiley-Liss, Inc.
valid for tumor suppressor genes monoallelically expressed because of genomic imprinting, a feature that is
suspected when a nonrandom parental origin of the allele loss is observed.2 Because portions of 1p and 19q
chromosomes contain monoallelically expressed, putative tumor suppressor genes that could be involved in
oligodendrogliomas,3–5 a study of the parental origins
of 1p and 19q losses was undertaken with a series of
these tumors.
Materials and Methods
Blood and tumor DNA from 32 patients with pure oligodendrogliomas (10 World Health Organization grade II and
22 grade III) were typed with a panel of microsatellites located on 1p (D1S468, D1S214, D1S450, D1S2667,
D1S234, D1S255, D1S2797, D1S2890, D1S206) and 19q
(D19S219, D19S888, D19S412, D19S418) with fluorescent
primers (Perkin Elmer, Norwalk, CT) and a gene scan program (Abi-Prism, Perkin Elmer). When a loss of 1p and 19q
was found, a blood sample from one or both parents was
collected when possible, after informed consent, to identify
the parental origins of the deleted chromosomes according to
the microsatellite allelic profile.
Loss of heterozygosity (LOH) on chromosomes 1p and
19q was detected in 14 of 32 tumors (44%) and 15 of
32 tumors (47%), respectively. The two alterations
were closely associated: 13 of 14 tumors with LOH on
1p demonstrated concomitant LOH on 19q; conversely, 13 of 15 tumors with LOH on 19q also
showed LOH on 1p ( p ⬍ 0.00001, ␹2 test). LOH on
1p and 19q was found in a significantly higher percentage of grade II oligodendrogliomas (OII) than
grade III oligodendrogliomas (OIII): LOH on 1p and
19q was observed in 7 of 10 OII (70%) versus 7 of 22
(32%) and 8 of 22 (36%) OIII, respectively ( p ⫽
0.04, ␹2 test), suggesting that the loss of 1p and 19q is
an early event.
Blood samples from one or both parents could be
obtained in 6 cases of oligodendroglioma with 1p and
19q loss. The parental origins of the 1p and 19q chromosome losses could be identified in all 6 cases (Fig).
The results showed that paternal loss involved all of 6
cases for 19q alleles ( p ⫽ 0.015, binomial test) and 3
of 6 cases for 1p alleles, suggesting that the 19q, but
not the 1p, putative tumor suppressor gene locus was
maternally imprinted (Table).
Fig. (A) Patient 2 (World Health Organization [WHO]
grade III oligodendroglioma): genotyping with the D1S252
microsatellite of constitutional DNA, tumor DNA, and paternal and maternal DNA shows loss of the paternal allele in the
tumor. (B) Patient 4 (WHO grade III oligodendroglioma):
genotyping with the D19S412 microsatellite of constitutional
DNA, tumor DNA, and maternal DNA demonstrates loss of
the paternal allele in the tumor. Paternal alleles are indicated
by a solid arrow, and maternal alleles are indicated by a
blank arrow. LOH ⫽ loss of heterozygosity.
Table. Parental Origin of the Loss of Heterozygocity on 1p
and 19q for the 6 Cases Investigated
Case No.
The frequent, closely associated 1p and 19q losses that
we and others1 found in low-grade oligodendrogliomas
suggest that putative tumor suppressor genes located at
1p and 19q are involved in a common pathway and are
inactivated at an early stage during tumorigenesis. At
this stage, the probability of combining four somatic
genetic events appears very low, given the classic two-
Annals of Neurology
Vol 52
No 1
July 2002
Loss of 1p
Loss of 19q
OII ⫽ oligodendroglioma (World Health Organization [WHO]
grade II); OIII ⫽ anaplastic oligodendroglioma (WHO grade III).
hit model.6 In the absence of genetic instability, this
combination is more likely to occur when the genes are
imprinted. In this situation, only one copy of the paternally or maternally inherited allele is active, and a
single hit may, therefore, suppress its function. Although imprinted tumor suppressor genes have been
implicated in various cancers,2,7–12 the possible role of
genomic imprinting in oligodendrogliomas has not yet
been evaluated, as far as we know.
Interestingly, we found exclusive and significant paternal loss of 19q alleles (all of 6) in oligodendrogliomas. This finding may indicate the involvement of a
paternally expressed tumor suppressor gene. Recently,
an imprinted tumor suppressor gene, PEG3 (paternally
expressed gene 3), which encodes a zinc finger protein,
was mapped to the 19q13.4 region. There are data suggesting that the PEG3 gene could be involved in oligodendrogliomas: PEG3 is strongly expressed in adult
brain cells, in both neurons and glial cells,5,13 whereas
expression is very low in gliomas because of either deletion or hypermethylation of the active allele.14 Moreover, transfection of glioma cell lines with PEG3 has
been shown to result in loss of tumorigenicity.13 On
the basis of deletion mapping, the putative locus involved in oligodendroglioma tumorigenesis has been
mapped to the 19q13.3 region, just proximal to
PEG3,15 but this result needs confirmation. Alternatively, large imprinted domains are frequently physically clustered into chromosomal regions16 and may
contain several other genes, such as the numerous zinc
finger protein-containing genes close to PEG3.5
Imprinting on 1p36 was also suggested because of
the preferential loss of maternal alleles in neuroblastoma.11,17–18 This region contains P73, a maternally
expressed tumor suppressor gene that shares considerable homology with P53.3 The finding that only 3 of
our 6 patients had a loss of the maternal copy of 1p
does not support a role for P73 during oligodendroglioma tumorigenesis; this conclusion is in agreement
with a previous study showing no mutation of P73 in
this tumor type.19 Similarly, the involvement of NOEY2,
a maternally imprinted ras-related tumor suppressor
gene located on 1p 31, was ruled out by our results.4
If confirmed in a larger series, 19q paternal loss will
indicate the involvement of a maternally imprinted tumor suppressor gene. As a result, the search for homozygous deletions or inactivating mutations of this
gene may be fruitless because the remaining maternal
allele may be inactivated by an epigenetic mechanism.
In contrast, equilibrated parental 1p loss does not support genomic imprinting, and a search for homozygous
deletion on 1p, therefore, remains a valid approach to
identifying the critical locus.
This work was supported by the Association pour la Recherche contre le Cancer (ARC 5892, K.H.X.), the Ligue Nationale contre le
Cancer (75/01-RS/42, K.H.X.), and the Délégation à la Recherche
Clinique (CRIC 99042, K.H.X.).
We are indebted to the patients and their families who agreed to
participate in this study.
1. Cairncross JG, Ueki K, Zlatescu MC, et al. Specific genetic
predictors of chemotherapeutic response and survival in patients
with anaplastic oligodendrogliomas. J Natl Cancer Inst 1998;
2. Falls JG, Pulford DJ, Wylie AA, Jirtle RL. Genomic
imprinting: implications for human disease. Am J Pathol 1999;
154:635– 647.
3. Kaghad M, Bonnet H, Yang A, et al. Monoallelically expressed
gene related to p53 at 1p36, a region frequently deleted in neuroblastoma and other human cancers. Cell 1997;90:809 – 819.
4. Yu Y, Xu F, Peng H, et al. NOEY2 (ARH1), an imprinted
putative tumor suppressor gene in ovarian and breast carcinomas. Proc Natl Acad Sci U S A 1999;96:214 –219.
5. Kim J, Ashworth L, Branscomb E, Stubbs L. The human homologue of a mouse-imprinted gene, Peg3, maps to a zinc finger gene-rich region of human chromosome 19q13.4. Genome
Res 1997;7:532–540.
6. Knudson AG. Mutation and cancer: statistical study of retinoblastoma. Proc Natl Acad Sci U S A 1971;68:820 – 823.
7. Jirtle RL. Genomic imprinting and cancer. Exp Cell Res 1999;
248:18 –24.
8. Toguchida J, Ishizaki K, Sasaki MS, et al. Preferential mutation
of paternally derived RB gene as the initial event in sporadic
osteosarcoma. Nature 1989;338:156 –158.
9. Williams JC, Brown KW, Mott MG, Maitland NJ. Maternal
allele loss in Wilms’ tumor. Lancet 1989;1:283–284.
10. Scrable H, Cavenee W, Ghavimi F, et al. A model for embryonal rhabdomyosarcoma tumorigenesis that involves genome
imprinting. Proc Natl Acad Sci U S A 1990;86:7480 –7484.
11. Caron H, van Sluis P, van Hoeve M, et al. Allelic loss of chromosome 1p36 neuroblastoma is of preferential maternal origin
and correlates with N-myc amplification. Nat Genet 1993;4:
12. Katz F, Webb D, Gibbons, et al. Possible evidence for genomic
imprinting in childhood acute myeloblastic leukaemia associated with monosomy for chromosome 7. Br J Haematol 1992;
13. Kohda T, Asai A, Kuroiwa Y, et al. Tumour suppressor activity
of human imprinted gene PEG3 in a glioma cell line. Genes
Cells 2001;6:237–247.
14. Maegawa S, Yoshioka H, Itaba N, et al. Epigenetic silencing of
PEG3 gene expression in human glioma cell lines. Mol Carcinog 2001;31:1–9.
15. Smith JS, Alderete B, Minn Y, et al. Localization of common
deletion regions on 1p and 19q in human gliomas and their
association with histological subtype. Oncogene 1999;18:
4144 – 4152.
16. Reik W, Walter J. Genomic imprinting: parental influence on
the genome. Nat Rev Genet 2001;2:21–32.
17. Caron H, Peter M, van Sluis P, et al. Evidence for two tumour
suppressor loci on chromosome bands 1p35–36 involved in
neuroblastoma: one probably imprinted, another associated
with N-myc amplification. Hum Mol Genet 1995;4:535–539.
18. Hogarty MD, Maris JM, White PS, et al. Analysis of genomic
imprinting at 1p35–36 in neuroblastoma. Med Pediatr Oncol
19. Mai M, Huang H, Reed C, et al. Genomic organization and
mutation analysis of p73 in oligodendrogliomas with chromosome 1p-arm deletion. Genomics 1998;51:359 –363.
Casanova et al: Minicolumnar Pathology in Dyslexia
Minicolumnar Pathology
in Dyslexia
Manuel F. Casanova, MD,1 Daniel P. Buxhoeveden, PhD,2
Morris Cohen, EdD,3 Andrew E. Switala,2
and Emil L. Roy, PhD2
The minicolumn is an anatomical and functional unit of
the brain whose genesis accrues from germinal cell divisions in the ventricular zone of the brain. Disturbances
in the morphometry of minicolumns have been demonstrated recently for both autism and Down’s syndrome.
We report minicolumnar abnormalities in the brain of a
dyslexic patient. The corresponding developmental disturbance (ie, large minicolumns) could account for the
perceptual errors observed in dyslexia.
Ann Neurol 2002;52:108 –110
Dyslexia is classified as a disturbance of language. Affected individuals have difficulties with reading, spelling, and, sometimes, solving mathematical problems.
This learning disability usually is manifested in grade
school and continues throughout the life of the individual. Because of the symptoms, as well as their
chronic nature, researchers have argued that a part of
the brain dealing with language is altered. Magnetic
resonance imaging studies detailing abnormalities in
the sylvian region support this contention.1,2 Unfortunately, few researchers have conducted postmortem
studies of dyslexia, and purported planum temporale
findings remain subjective and in need of corroboration. Therefore, although the presence of ectopias and
micropolygyria in the sylvian region and other cortical
areas supports a developmental onset for dyslexia, it
fails to explain abnormalities pertaining to cerebral
Recently, our group described interhemispheric
changes in the morphometry of minicolumns that
could provide for the speciation of hominids.4 Because
of its location in area 22 (part of Wernicke’s region),
the morphometric difference may play a role both in
the development of human language and in its disor-
From the 1Department of Psychiatry and Neurology, Medical College of Georgia; 2Downtown Veterans Affairs Medical Center; and
Department of Neurology, Medical College of Georgia, Augusta,
Received Nov 5, 2001, and in revised form Feb 11, 2002. Accepted
for publication Feb 20, 2002.
Published online June 18, 2002,
in Wiley InterScience
( DOI: 10.1002/ana.10226
Address correspondence to Dr Casanova, 3B-121, Downtown Veterans Affairs Medical Center (24), One Freedom Way, Augusta, GA
30904-6285. E-mail:
© 2002 Wiley-Liss, Inc.
ders. Our study looked for the presence of abnormalities of minicolumnar organization and lateralization in
area 22 of the brain of a dyslexic patient.
Standard morphometrics provide static information
such as cell number, area, and orientation. However,
significant advantages accrue from examining the minicolumn rather than individual cells. Not only is the
functionality of a cell much more limited in scope than
that of the minicolumn, but individual cells are organized around a vertical stratigraphical model. This linear arrangement of cells is the earliest anatomical organization of neurons within the cortical plate and a
template on which later environmental influences act.
Because of their simplicity and the availability of computer techniques, state-of-the-art computerized imaging
can easily analyze this linear structure.
Case Report
The history of the patient and preliminary neuropathological
findings were published in this journal.5 The patient was a
20-year-old man who died from an accidental fall. His reading difficulties were diagnosed at an early age (first grade);
his communication aptitudes were lower than expected for
his intellectual level, sociocultural opportunities, and educational experience. Both his brothers and father were slow
readers. Serial sections showed an area of polymicrogyria in
the left temporal speech region and scattered cortical dysplasias on the same hemisphere. For this study, the posterior
portion of area 22 or Tpt (part of Wernicke’s region) and
area 9 (taken as a control) were available bilaterally in Nisslstained, 35␮m-thick sections.
Computerized analysis segments the image into foreground (cell somas) and background (neuropil) with a
threshold, which is determined automatically for each image.
Each foreground object is further classified according to its
area; that is, large neurons are those with an area greater than
30␮m2, small neurons have areas between 10 and 30␮m2,
and objects smaller than 10␮m2 are excluded from further
computation. Local cell density is calculated as follows: Obtain the mean neuronal spacing d equal to the square root of
the image area divided by the number of large neurons. Divide the image into horizontal bands 2d high. Within each
band, smooth the observed cell density as a function of horizontal position with a Gaussian kernel of width 1⁄2d. Local
maxima of cell density indicate minicolumn cores, and local
minima indicate peripheral neuropil space. Maxima in adjacent strips, along with the minima on either side, are linked
to form chains. The neurons and neuropil enclosed within a
chain form a minicolumn segment. The algorithm is described in detail elsewhere.6
Several types of information can be obtained from minicolumn segments. Columnar width is the mean distance between the left and right boundaries of the segment. Neuropil
space is the width of the cell-poor space on both sides of the
column core. Mean cell spacing is the mean distance between neighboring neurons. The neighborhood of a neuron
is defined with the minimum spanning tree: that set of
edges, that is, line segments connecting all the neurons,
which has the minimum total length of all such sets. A neu-
ron’s neighborhood comprises those other neurons sharing
with it an edge of the minimum spanning tree. Although
neuropil space measures the neuropil space in the column
periphery, mean cell spacing measures the neuropil space
within the minicolumn core. The relative dispersion ratio
(RDR) is the ratio of the second moments of the distribution
of cells within a segment divided by the ratio of second moments of the shape of the segment. RDR is a measure of
compactness. The least compact arrangement, a uniform distribution of cells within a segment, would result in an RDR
of 1. The more the neurons cluster about the column core,
the greater the RDR value is.
Normal limits for the morphometric parameters were
based on a population of 38 controls between 3 and 98
years of age, with a mean age of 47 years. Samples
from the control cases were mixed for hemisphere and
cortical area (Table 1). Statistical testing involved a
multivariate analysis of variance including diagnosis,
hemisphere, and area as fixed factors and age as a covariate. Dependent variables in the analysis included
columnar width, mean cell spacing, neuropil space, and
RDR. The overall multivariate test (Wilks ␭) places the
dyslexic case well outside the normal range ( p ⬍
0.001). Columnar width, mean cell spacing, and peripheral neuropil space were all larger than normal in
the dyslexic patient (all p ⬍ 0.001; Table 2 and Fig).
RDR was within normal limits ( p ⫽ 0.667). The effect
of cortical area was significant ( p ⬍ 0.001), and significant diagnosis ⫻ area and diagnosis ⫻ hemisphere ⫻ area interactions emerged (both p ⬍ 0.001);
however, the effect of cortical area and the interactions
was not significant in the follow-up univariate tests on
columnar width, mean cell spacing, and neuropil space.
Nevertheless, Table 2 breaks down the results by side
and cortical area. There was no significant dependency
on age ( p ⫽ 0.514). The study indicates a significant
abnormality of minicolumnar morphometry in the patient studied. Overall, minicolumns were significantly
wider and had more peripheral neuropil space than
normal. Contrary to our original expectations, findings
were confined neither to one hemisphere nor to brain
region 22.
Table 2. Results, Corrected for Age
Area 9 (␮m)
Diagnosis Parameter
Area Tpt
CW ⫽ columnar width; MCS ⫽ mean cell spacing; NS ⫽ neuropil
mammalian class.7 However, researchers have found
evidence of species and task specificity with respect to
minicolumn size in homologous regions of cortex.8 –10
This is best exemplified in the primary visual cortex, in
which column size apparently is related to functional
requirements. The visual system of monkeys is more
sophisticated than those of cats or rabbits and processes
information differently. Although the brains of monkeys tested were larger, the minicolumns in lamina V1
were much smaller, approximately 31␮m versus 56␮m
for cats and rabbits.10 –12 Variations in minicolumn
size in homologous regions also were found between
human and nonhuman primate Tpt. Therefore, similar
variations probably exist in other cortical regions as
Changing the column size and configuration alters
the interaction between pathways and other columns.
Fig. The dyslexic case falls clearly outside normal prediction
limits in this plot of columnar width versus mean cell spacing
(area Tpt). The overall mean parameter value for each brain
is plotted, with averaging over hemisphere and area. MCS ⫽
mean cell spacing; CW ⫽ columnar width.
Minicolumns generally are considered to be essentially
the same size in brains of all members of the diverse
Table 1. Sample Size
Area 9
Area Tpt
Casanova et al: Minicolumnar Pathology in Dyslexia
In effect, Seldon13 showed that changes in the width of
minicolumns as small as 15% could seriously affect the
organization of a region, such as its input and output
relationships. If this is the case, then variations in the
number, size, and maturation of minicolumns also can
lead to pathology. This occurs in two conditions recently described as exhibiting minicolumnar abnormalities: Down’s syndrome and autism.
In the small brains of patients with Down’s syndrome, minicolumns are of normal width; however,
these radial structures attain adult minicolumnar size
much earlier than normal.14 The changes are consistent
with the accelerated aging observed in this condition.
In contrast, the brains of autistic individuals exhibit
smaller and more numerous minicolumns.15 Because
the brains of autistic children are also generally larger
than normal, this difference magnifies the presence of
smaller columns; it also results in a relative and absolute increase in processing units.
Computer modeling of small minicolumns shows excess signaling and too little inhibition within these
structures.16 Unsurprisingly, approximately one third
of autistic patients develop seizures, 17 another condition in which signaling and inhibition is out of balance. In fact, case reports in selected populations of
epileptic patients have shown that anticonvulsant medications have had a beneficial effect on some autistic
traits.18 –20 This case report now suggests a different
type of pathology in dyslexia. This pathology is virtually the exact opposite of that previously reported in
autism and involves the presence of significantly enlarged minicolumns. Therefore, it appears that minicolumns exist within a phenotypic spectrum that intertwines the inhibitory/excitatory flow of neocortical
information with a tweaking of the signal-to-noise ratio
relevant to feature extraction. This yin-yang phenomenon may offer important clinicopathological correlates
to a host of conditions characterized by minicolumnar
This study was funded by grants from the Theodore and Vada Stanley Foundation (M.F.C.) and the Veterans Affairs Merit Review
Board (M.F.C.).
1. Green RL, Hutsler JJ, Loftus WC, et al. The caudal infrasylvian
surface in dyslexia: novel magnetic resonance imaging-based
findings. Neurology 1999;53:974 –981.
2. Morgan AE, Hynd GW. Dyslexia, neurolinguistic ability, and
anatomical variation of the planum temporale. Neuropsychol
Rev 1998;8:79 –93.
3. Humphreys P, Kaufmann WE, Galaburda AM. Developmental
dyslexia in women: neuropathological findings in three patients.
Ann Neurol 1990;28:727–738.
4. Buxhoeveden D, Switala A, Roy E, et al. Lateralization in human planum temporale is absent in nonhuman primates. Brain
Behav Evol 2001;57:349 –358.
Annals of Neurology
Vol 52
No 1
July 2002
5. Galaburda AM, Kemper TL. Cytoarchitectonic abnormalities in
developmental dyslexia: a case study. Ann Neurol 1979;6:
94 –100.
6. Buxhoeveden D, Switala AE, Roy E, Casanova MF. Quantitative analysis of cell columns in the cerebral cortex. J Neurosci
Methods 2000;97:7–17.
7. Mountcastle VB. The columnar organization of the neocortex.
Brain 1997;120:701–722.
8. Buxhoeveden DP, Lefkowitz W, Loats P, Armstrong E. The
linear organization of cell columns in human and nonhuman
anthropoid Tpt cortex. Anat Embryol 1996;194:23–36.
9. Buxhoeveden D, Switala A, Roy E, et al. Morphological differences between minicolumns in human and nonhuman primate
cortex. Am J Phys Anthropol 2001;115:361–371.
10. Peters A, Sethares C. Myelinated axons and the pyramidal cell
modules in monkey primary visual cortex. J Comp Neurol
11. Orban GA. Neuronal operation in the visual cortex. Berlin:
Springer, 1984. Studies in brain function; vol 2.
12. Peters A, Yilmaz E. Neuronal organization in area 17 of cat
visual cortex. Cereb Cortex 1993;3:49 – 68.
13. Seldon HL. Structure of human auditory cortex I: cytoarchitectonics and dendritic distributions. Brain Res 1981;229:
14. Buxhoeveden D, Fobbs A, Roy E, Casanova MF. Quantitative
comparison of radial cell columns in Down’s syndrome and
controls. J Intellect Disabil Res 2002;46:76 – 81.
15. Casanova MF, Buxhoeveden D, Switala A, Roy E. Minicolumnar pathology in autism. Neurology 2002;58:428 – 432.
16. Buxhoeveden D, Casanova MF. Computer modeling of
excitation-inhibition defects in autism. Proceedings Summary
of the American Psychiatric Association Annual Meeting, 2001;
San Diego, CA, NR 728.
17. Volkmar FR, Nelson DS. Seizure disorders in autism. J Am
Acad Child Adolesc Psychiatry 1995;29:127–129.
18. Childs JA, Blair JL. Valproic acid treatment of epilepsy in autistic twins. J Neurosci Nurs 1997;29:244 –248.
19. Jambaque I, Chiron C, Dumas C, et al. Mental and behavioural outcome of infantile epilepsy treated by vigabatrin in tuberous sclerosis patients. Epilepsy Res 2000;38:151–160.
20. Plioplys AV. Autism: electroencephalogram abnormalities and
clinical improvement with valproic acid. Arch Pediatr Adolesc
Med 1994;148:220 –222.
Anti-Hu Antibodies in
Merkel Cell Carcinoma
John E. Greenlee, MD,1,2 John D. Steffens, MD,2
Susan A. Clawson, BS,2 Kenneth Hill, BS,1
and Josep Dalmau, MD3
Anti-Hu antibody is an antineuronal autoantibody found
in a subset of patients with paraneoplastic neurological
disease. The antibody was first associated with small cell
carcinoma of the lung and is most often used as a marker
for this neoplasm in patients presenting with suspected
paraneoplastic syndromes. Here we report a patient with
a multifaceted neurological disorder in the setting of
Merkel cell carcinoma. The patient’s serum contained antibodies against the Hu antigen, and the expression of the
Hu antigen was demonstrated in the patient’s tumor.
Ann Neurol 2002;52:111–115
The anti-Hu antibody is an autoantibody reactive with
nuclei and cytoplasm of neurons throughout the central and peripheral nervous systems. Anti-Hu antibody
response was first identified in patients with paraneoplastic encephalomyelitis in the setting of small cell
cancer of the lung and has been used as a marker for
small cell lung cancer in patients presenting with suspected paraneoplastic neurological syndromes.1–3 In
rare patients, anti-Hu antibody has been associated
with other neoplasms. These have included non-small
cell lung neoplasms, extrapulmonary small cell carcinomas, testicular carcinoma, neuroblastomas, breast cancer, chondroid myxosarcoma, and neuroendocrine neoplasms at other sites.4 – 6 Anti-Hu antibody in
association with Merkel cell carcinoma has not been
described previously. Here we report a patient with a
multifaceted neurological disorder in the setting of
Merkel cell carcinoma. The patient’s serum contained
antibodies against the Hu antigen, and the expression
of the Hu antigen was demonstrated in the patient’s
From the 1Neurology Service, Veterans Affairs Medical Center, and
Department of Neurology, University of Utah School of Medicine,
Salt Lake City, UT; and 3Department of Neurology, University of
Arkansas for Medical Sciences, Little Rock, AR.
Received Dec 17, 2001, and in revised form Feb 21, 2002. Accepted for publication Feb 21, 2002.
Published online May 31, 2002, in Wiley InterScience
( DOI: 10.1002/ana.10225
Address correspondence to Dr Greenlee, Neurology Service (127),
Veterans Affairs Medical Center, 500 Foothill Drive, Salt Lake City,
UT 84148. E-mail:
This is a US Government work and, as such, is in the public domain in the United States of America.
Case Report
Five years before admission, this 77-year-old woman developed a left forearm lesion that was diagnosed as Merkel cell
carcinoma. The lesion was treated by excision and radiation,
with no clinical evidence of recurrence. One year before admission, the patient developed paresthesias of the second
digit of her left hand, which ascended to the midforearm
over the next month. She subsequently developed a similar
pattern of paresthesias in her right hand and within 2
months developed stocking distribution paresthesias.
Eight months before admission, she developed nausea and
dysequilibrium, requiring intravenous hydration. During
hospitalization at an outside facility, she developed a left peripheral facial palsy, right lower extremity paresis, constipation with rare fecal and urinary incontinence, and vertical
nystagmus. Lumbar punctures showed a protein concentration of 85mg/dl, lymphocytic pleocytosis, normal glucose,
and no oligoclonal bands. Chest and abdominal computed
tomography scans; esophagogastroduodenoscopy with gastric
biopsy; cranial, cervical, and thoracic magnetic resonance imaging; human immunodeficiency virus testing; and multiple
serological studies were normal except for a strongly positive
antinuclear antibody and an erythrocyte sedimentation rate
of 55 to 60mm/hr. Her physicians diagnosed a variant of
Guillain-Barré syndrome and treated her with intravenous
immunoglobulin G (IgG) and high-dose corticosteroids. She
improved slightly, but symptoms recurred as steroids were
tapered, and she developed rapid cognitive decline and hallucinations. She was admitted to the University of Utah
Health Sciences Center with increasing weakness, worsening
hallucinations, and behavioral disturbance.
General physical examination was negative except for epigastric tenderness. The patient was alert but actively hallucinating. She exhibited a peripheral left facial weakness, hyperactive gag reflex, and impersistence of gaze. Motor
examination showed proximal greater than distal left-sided
weakness and increase in tone, as well as distal more than
proximal right leg weakness. The patient had a left-sided tactile neglect and stocking-glove loss of proprioception, vibration, and, to a lesser extent, light touch and pinprick. She
stood on a widened base with assistance but could not ambulate. She had a mild left hyperreflexia, with distal loss of
reflexes. Babinski’s sign could not be tested, but the patient
had bilateral palmomental reflexes and prominent snout and
palmar grasp reflexes.
Hematological studies, blood chemistries, and rheumatological studies were normal except for an antinuclear antibody of 1:10,240. Magnetic resonance imaging showed nonspecific white matter changes and atrophy. Electromyogram
demonstrated lumbar plexopathy and peripheral neuropathy,
without myopathic changes. A muscle biopsy showed predominantly neurogenic atrophy with minimal myopathic
changes and without evidence of vasculitis. Abdominal computed tomography showed a 17cm intraabdominal mass, and
a biopsy documented Merkel cell carcinoma. At the family’s
request, the patient was discharged from the hospital, and
she died 3 weeks later. An autopsy demonstrated intraabdominal Merkel cell carcinoma, with invasion of the adrenal
gland, liver, and pancreas. No lung or other neoplasms were
identified despite an extensive search. Sections of the spinal
Published 2002 by Wiley-Liss, Inc.
cord exhibited bilateral posterior column degeneration. The
brain was without evidence of metastasis or infarction. There
was loss of myelinated fibers in the deep white matter, with
rare neurofibrillary tangles in the hippocampi. There was no
evidence of perivascular inflammation. Sections of peripheral
nerve showed neurogenic edema and the loss of myelinated
fibers with rare digestion chambers suggesting ongoing atrophy. Dorsal root ganglia were not examined.
Samples of serum and formalin-fixed autopsy sections of the
tumor were obtained under appropriate institutional review
board guidelines.
Immunofluorescence Labeling of Human Cerebellar
and Brainstem Sections
Sections of human cerebellum and paraformaldehyde-fixed
vibratome sections of rat cerebellum were reacted with patient or control serum as previously described and were examined with confocal microscopy.7 Controls included sera
from normal individuals and from patients with nonparaneoplastic neurological diseases.
Gel Electrophoresis and Western
Immunoblot Methods
Lysates of human Purkinje cells and liver were subjected to
Western blot analysis with patient and control sera.8,9 Patient serum also was reacted against the HuD recombinant
protein, which was cloned by Szabo and colleagues9 and is
routinely used as a marker for anti-Hu antibody activity.
Immunoperoxidase Labeling of Tumor Tissue for the
Expression of Hu Antigens
Sections of paraffin-embedded tissue were deparaffinized; microwaved; incubated with 10% normal human serum in
phosphate-buffered saline for 20 minutes, followed by biotinylated serum IgG from the patient or from individuals
with documented anti-Hu antibody response; and labeled
with avidin-biotin-peroxidase complex and diaminobenzidine
as previously described.10 Tissue sections incubated with biotinylated serum IgG from a normal individual were used as
Immunofluorescence Studies
Sera from the patient produced bright immunofluorescence labeling of neuronal nuclei throughout the brain,
as well as less intense, ground-glass cytoplasmic staining (Fig 1). Nuclear staining was not abolished by the
absorption of patient serum with human liver powder.
Western Blot Analysis
The patient’s serum reacted with 35 to 42kDa (molecular mass) neuronal proteins and the HuD recombinant protein, confirming the presence of the anti-Hu
antibody response. The patient’s serum did not label
corresponding proteins in blots of liver, indicating
specificity of the immune response for neural antigens
(Fig 2).
Fig 1. Thick section of rat cerebellum reacted with patient serum at a dilution of 1 to 80, stained with immunofluorescence methods, and examined with confocal microscopy. Neurons throughout the cerebellum show bright nuclear fluorescence with nucleolar
sparing and ground-glass cytoplasmic staining, which is typical of anti-Hu antibody response. Immunofluorescence staining was not
abolished by the absorption of serum with human liver (⫻200).
Annals of Neurology
Vol 52
No 1
July 2002
anti-Hu IgG immunoreacted with the tumor tissue; biotinylated IgG from a normal individual did not show
Fig 2. (A) Western blot of patient’s serum reacted with lysates
of cerebellar neurons (lanes 1– 4) or liver cells (lanes 5–7).
There is specific labeling of proteins of approximately 35 to
42kDa (molecular mass) in blots of cerebellum characteristic
of the anti-Hu antibody response. These proteins are not labeled in lysates of liver. (Lane 1) Control serum from a patient without cancer or neurological disease. (Lane 2) Serum
from a patient with the anti-Hu antibody. (Lane 3) Serum
from a Merkel cell patient. (Lane 4) Serum from a Merkel
cell patient, absorbed with human liver, showing no loss of
reactivity. (Lane 5) Normal serum reacted with liver lysate.
(Lane 6) Anti-Hu serum reacted with liver. (Lane 7) Serum
from a Merkel cell patient reacted with liver. (B) Western
blot showing the labeling of the HuD recombinant protein,
used as a marker of the anti-Hu antibody response, by patient
and anti-Hu sera but not by normal serum. (Lanes 1, 3, 5)
Serum from the Merkel cell patient. (Lane 2) Anti-Hu serum.
(Lane 4) Serum from a normal control. (Lane 6) Anti-Hu
Analysis of the Patient’s Tumor for the Expression of
the Hu Antigen
The expression of the Hu antigen was detected in tumor sections (Fig 3). The specificity of staining for the
Hu antigen was confirmed with several samples of biotinylated anti-Hu IgG from different patients. All
Merkel cell carcinoma (trabecular carcinoma) is a primary neuroendocrine tumor of the skin first described
as a distinctive neoplasm in 1972. The tumor characteristically arises in the dermis or subcutaneous tissues
of elderly individuals and most frequently involves the
head or neck, followed by the arms, legs, and
trunk.11,12 The tumor spreads by both local extension
and distant metastasis and is poorly responsive to therapy. Pathologically, the tumor consists of solid masses,
nests, or sheets of neoplastic cells.11 The immunohistochemical analysis of Merkel cell tumors has shown
the expression of several neuronal marker proteins, although not all of these proteins are present in all cases:
approximately 50% of cases have been found to stain
for neuron-specific enolase, 40% stain for neurofilament proteins, up to 64% stain for 44 to 54kDa (molecular mass) cytokeratins, and 32% stain for intestinal
vasoactive polypeptide.12,13 The tumor is differentiated
clinically from small cell carcinoma of the lung by
much higher survival rates (40 – 60%) and immunohistologically by the expression of cytokeratin 20.14 Dalmau and colleagues10 and Gultekin and colleagues15
have detected the expression of the Hu antigen in random Merkel cell tumors. However, the incidence of
expression of the Hu antigen is not known, nor has the
tumor previously been associated with an anti-Hu antibody response.
Merkel cell carcinoma may metastasize to the central
nervous system.16,17 In addition, the neoplasm has
been associated with several paraneoplastic neurological
syndromes, including encephalomyelitis and LambertEaton myasthenic syndrome.17–19 Only a few investigators, however, have published studies characterizing
the immune response in these patients. Cher and colleagues described antibodies to three groups of proteins
between 46 and 80kDa (molecular mass) in a patient
with brainstem encephalitis in the setting of Merkel
cell carcinoma,17 and Eggers and colleagues, in extensive studies of a patient with Merkel cell carcinoma
and Lambert-Eaton myasthenic syndrome, described
antibodies previously associated with Lambert-Eaton
myasthenic syndrome: P/Q-type calcium channel antibodies, muscle acetylcholine receptor binding antibodies, acetylcholine receptor modulating antibodies, and
striational antibodies.18
The patient reported here presented with Merkel cell
carcinoma, progressive sensorimotor neuropathy, autonomic neuropathy with gastroparesis, and encephalopathy with cognitive decline. Her sera contained unequivocal anti-Hu antibodies, and the expression of Hu
proteins was clearly demonstrated in her tumor.
Greenlee et al: Anti-HU Antibodies
Fig 3. (A) Section of the patient’s tumor reacted with the patient’s biotinylated immunoglobulin G (IgG), visualized with avidinbiotin complex and diaminobenzidine, and counterstained with hematoxylin. Cells within the tumor react with anti-Hu IgG, indicating the expression of Hu proteins (⫻250). Anti-Hu IgG from other patients showed a similar type of reactivity (data not
shown). (B) Biotinylated IgG from a normal individual did not show reactivity with the patient’s tumor.
Anti-Hu antibodies are most commonly associated
with small cell carcinoma of the lung but also have
been described in small cell carcinomas of other sites,
certain other neuroendocrine tumors, testicular carcinomas, and neuroblastomas. The role of anti-Hu autoantibodies in this patient’s clinical syndrome is unknown. However, this study demonstrates that patients
with Merkel cell carcinoma may develop antibodies to
the Hu antigen and may, in this setting, develop paraneoplastic neurological disease. Patients presenting with
progressive neurological syndromes and anti-Hu antibody response first should be suspected of having small
cell carcinoma involving lung or, rarely, extrapulmonary sites. Merkel cell carcinoma, however, should be
considered as a diagnostic possibility in such patients if
pulmonary neoplasia is not detected.
This work was supported by a Merit Review research award from
the United States Department of Veterans Affairs (J.E.G.).
1. Graus F, Cordon-Cardo C, Posner JB. Neuronal antinuclear
antibody in sensory neuronopathy from lung cancer. Neurology
1985;35:538 –543.
2. Greenlee JE, Lipton HK. Anticerebellar antibodies in serum
and cerebrospinal fluid of a patient with oat cell carcinoma of
the lung and paraneoplastic cerebellar degeneration. Ann Neurol 1986;19:82– 85.
Annals of Neurology
Vol 52
No 1
July 2002
3. Moll JW, Antoine JC, Brashear HR, et al. Guidelines on the
detection of paraneoplastic anti-neuronal-specific antibodies: report from the workshop to the fourth meeting of the International Society of Neuro-Immunology on paraneoplastic neurological disease, held October 22–23, 1994, in Rotterdam, The
Netherlands. Neurology 1995;45:1937–1941.
4. Graus F, Keime-Guibert F, Rene R, et al. Anti-Hu-associated
paraneoplastic encephalomyelitis: analysis of 200 patients. Brain
2001;124:1138 –1148.
5. Fisher PG, Wechsler DS, Singer HS. Anti-Hu antibody in a
neuroblastoma-associated paraneoplastic syndrome. Pediatr
Neurol 1994;10:309 –312.
6. Verschuuren J, Twijnstra A, De Baets M, et al. Hu antigens
and anti-Hu antibodies in a patient with myxoid chondrosarcoma. Neurology 1994;44:1551–1552.
7. Greenlee JE, Boyden JW, Pingree M, et al. Antibody types and
IgG subclasses in paraneoplastic neurological syndromes. J Neurol Sci 2001;184:131–137.
8. Greenlee JE, Parks TN, Jaeckle KA. Type IIa (anti-Hu) antineuronal antibodies produce destruction of rat cerebellar granule neurons in vitro. Neurology 1993;43:2049 –2054.
9. Szabo A, Dalmau J, Manley G, et al. HuD, a paraneoplastic
encephalomyelitis antigen, contains RNA-binding domains and
is homologous to Elav and Sex-Lethal. Cell 1991;67:325–333.
10. Dalmau J, Furneaux HM, Cordon Cardo C, Posner JB. The
expression of the Hu (paraneoplastic encephalomyelitis/sensory
neuronopathy) antigen in human normal and tumor tissues.
Am J Pathol 1992;141:881– 886.
11. Sibley RK, Dehner LP, Rosai J. Primary neuroendocrine (Merkel cell?) carcinoma of the skin. I. A clinicopathological and
ultrastructural study of 43 cases. Am J Surg Pathol 1985;9:
12. Sibley RK, Dahl D. Primary neuroendocrine (Merkel cell?) carcinoma of the skin. II. An immunocytochemical study of 21
cases. Am J Surg Pathol 1985;9:109 –116.
13. Schmidt U, Müller U, Metz A, Leder L-D. Cytokeratin and
neruofilament protein staining in Merkel cell carcinoma of the
small cell type and small cell carcinoma of the lung. Am J Dermatopathol 1998;20:346 –351.
14. Chan JK, Suster S, Wenig BM, et al. Cytokeratin 20 immunoreactivity distinguishes Merkel cell (primary cutaneous neuroendocrine) carcinomas and salivary gland small cell carcinomas
from small cell carcinomas of various sites. Am J Surg Pathol
1997;21:226 –234.
15. Gultekin SH, Rosai J, Demopoulos A, et al. Hu immunolabeling as a marker of neural and neuroendocrine differentiation in
normal and neoplastic human tissues: assessment using a recombinant anti-Hu Fab fragment. Int J Surg Pathol 2000;8:
109 –117.
16. Snodgrass SM, Landy H, Markoe AM, Feun L. Neurologic
complications of Merkel cell carcinoma. J Neurooncol 1994;22:
17. Cher LM, Hochberg FH, Teruya J, Nitschke M, Valenzuela
RF, Schmahmann JD, et al. Therapy for paraneoplastic neurologic syndromes in six patients with protein A column immunoadsorption. Cancer 1995;75:1678 –1683.
18. Eggers SZ, Salomao D, Dinapoli RP, Vernino S. Paraneoplastic
and metastatic neurological complications of Merkel cell carcinoma. Mayo Clin Proc 2001;76:327–330.
19. Batchelor TT, Platten M, Hochberg FH. Immunoadsorption
therapy for paraneoplastic syndromes. J Neurooncol 1998;40:
Creutzfeldt-Jakob Disease
Cluster in an Australian
Rural City
Steven Collins, MD,1,2
Alison Boyd, Post Grad Dip Gen Coun,1,2
Ashley Fletcher, BSc,1,2 John Kaldor, PhD,3
Andrew Hill, PhD,2 Stephen Farish, MEd,4
Catriona McLean, MD,1,2 Zahid Ansari, MPH,5
Margaret Smith, BSc,2,6 and Colin L. Masters, MD1,2
Through the Australian National Creutzfeldt-Jakob Disease Registry, 6 pathologically confirmed sporadic cases
were recognized over a 13-year period in persons who
had been long-term residents of a moderate-sized rural
city, whereas the expected number was 0.923. An extensive investigation could not find any point-source or
case-to-case transmission links. This occurrence is highly
statistically significant (p ⴝ 0.0027) when viewed in isolation and remains significant (p < 0.02) when only the
cases that arose after the cluster was recognized were
taken into account. However, a more conservative statistical analysis suggests that such a grouping could have
arisen by chance in at least one population group of this
size when the whole country is taken into consideration.
Ann Neurol 2002;52:115–118
Transmissible spongiform encephalopathies (TSEs) constitute a group of invariably fatal neurodegenerative disorders that affect both humans and animals.1 The European epidemic of bovine spongiform encephalopathy
and the more recently recognized variant CreutzfeldtJakob disease (CJD; first reported in the United Kingdom in 19962) have drawn considerable attention to this
group of diseases. In contrast with variant CJD, classic
CJD typically presents as a rapidly progressive dementia
in the elderly associated with ataxia and myoclonus and
is the most common human TSE phenotype. Systematic
From the 1Australian National Creutzfeldt-Jakob Disease Registry
and 2Department of Pathology, University of Melbourne, Victoria;
National Centre in HIV Epidemiology and Clinical Research, University of New South Wales, Sydney, New South Wales; 4Epidemiology and Biostatistics Unit, School of Population Health, University of Melbourne, Melbourne; 5Health Outcomes Section, Public
Health Division, Department of Human Services, Melbourne; and
Molecular Biology Laboratory, Melbourne Health Shared Pathology
Services, Royal Melbourne Hospital, Parkville, Victoria, Australia.
Received Dec 12, 2001, and in revised form Feb 13, 2002. Accepted for publication Feb 23, 2002.
Published online May 31, 2002, in Wiley InterScience
( DOI: 10.1002/ana.10224
Address correspondence to Dr Collins, Australian National
Creutzfeldt-Jakob Disease Registry, Level 5, Department of Pathology, University of Melbourne, Parkville, Victoria, Australia 3010.
© 2002 Wiley-Liss, Inc.
prospective genotyping detects likely causal mutations in
the prion protein gene (PRNP) in up to 13 to 14% of
CJD patients,3 and iatrogenic transmission appears to
explain only a few additional cases.4 Recognized sources
of iatrogenic transmission are diverse, including contaminated neurosurgical instruments, and long incubation
periods of 15 or more years are reported.1,4 For most
cases of CJD, there is no identifiable cause. Although de
novo somatic mutations in PRNP or spontaneous deleterious conformational changes in the normal prion protein are favored to underlie sporadic CJD,1 other postulated mechanisms include unrecognized iatrogenic or
natural transmission events.5,6 Recent reports of spatiotemporal groupings of variant CJD have rekindled interest in undertaking detailed analyses of such occurrences,
with the outcome that covert or uncertain transmission
mechanisms may be elucidated.7
During the course of systematic Australian National
Creutzfeldt-Jakob Disease Registry surveillance activities, it was observed that 7 deaths from TSE had been
reported from 1988 to 2000 in persons who had been
long-term residents within or immediately neighboring
a single moderate-sized rural city. We present the findings of our detailed investigation.
Patients and Methods
After potential CJD cases are ascertained, a thorough medicodemographic profile is collected from medical records and
family members, including the completion of a detailed questionnaire largely aimed at a priori risk factors for CJD. PRNP
analysis (as described previously8) and autopsy are actively facilitated and undertaken whenever possible. Final classification
is made once all data are collected and is consistent with internationally recognized diagnostic criteria, with pathological
verification required for definite CJD.9 The national CJD
mortality data for 1988 to 2000 were used to obtain Australian mortality rates (adjusted for age and gender) and the expected CJD occurrence rate among residents for the rural city
over this period (0.923 cases) and after the cluster was first
suspected in 1996 (0.284 cases). Formal semiannual analyses
are performed and reports are generated to monitor patterns
and trends of CJD occurrence, including geographical ones.
The statistical significance of the cluster was assessed by
estimation of relative risk through comparisons of the expected (following age and gender adjustment) and observed
CJD death rates for the rural city. The comparison was made
first for the entire observation period and then separately for
the period from 1997 after the diagnosis of the earlier cases
had given rise to the suspicion of a cluster. In addition, Poisson probabilities were used to assess the likelihood of similar
clusters of CJD cases occurring in any equivalent-sized individual cohort (of approximately 73,000) with the entire Australian population divided equally into 241 groups. p values
less than 0.05 were considered significant.
Australian National Creutzfeldt-Jakob Disease Registry surveillance methods have been reported in detail previously.5
In brief, ascertainment methods include reviews of morbidity
separation coding data from all university-affiliated tertiary
referral hospitals in Australia, as well as the centralized databases of state and territory health departments; regular national death certificate searches; and semiannual mail-out
questionnaires to all neurologists and pathologists within
Australia. Simultaneously with comprehensive prospective
monitoring since September 1993, all Australian cases have
been retrospectively sought to January 1, 1970. Through the
combination of ascertainment measures, it is believed that
essentially all human cases of TSE within Australia have been
detected, particularly since the late 1980s.
During the course of systematic surveillance activities,
it was observed that 5 deaths from pathologically confirmed TSE had been reported from 1988 to 2000 in
persons who were residents of a single moderate-sized
rural city (population in 1991 census, 72,976). An additional death from Gerstmann-Sträussler-Scheinker
syndrome associated with the proline102valine
(P102L) mutation was confirmed in a person who resided just outside the city. Further investigation, therefore, concentrated on the other five patients (Tables 1
and 2; Fig). Patients 1 to 4 were all long-term residents
Table 1. Selected Demographic Features of Sporadic Creutzfeldt-Jakob Disease Cases
Year of
Age at
Total Duration
of Residence in
City (yr)
Radial Distance
from City Center
Mutations/codon 129b
Last city residential address.
Archival retrieval of DNA from formalin fixed brain permitted reliable nucleotide sequencing for codons 1 to 133 only in Case 2, and from
1 to 135 only in Case 4.
Over 3 separate periods spanning 22 years.
ND ⫽ not determined.
Annals of Neurology
Vol 52
No 1
July 2002
Table 2. Salient Clinical Features of Sporadic Creutzfeldt-Jakob Disease Cases
Illness Duration
Family History of
Similar Illness
CSF ⫽ cerebrospinal fluid; EEG ⫽ electroencephalogram; ND ⫽ not done.
of the rural city (30 – 65 years) and died there. The
most recent case (Patient 5) was a person who lived
only the last 3.5 years of his life in the rural city. In
addition to these 5 deaths among residents, the Australian National Creutzfeldt-Jakob Disease Registry ascertained an additional sporadic case (Patient 6; see Tables 1 and 2; see Fig) who died while a resident of the
state capital but who previously had been a long-term
resident of the rural city (a total of 13 years over a span
of 22 years). Patient 6 spent an additional 2 years living on a small farm 27km from the city center from
which he commuted weekdays to his office located in
the central city. Total residence in the rural city constituted the longest by far in any specific location over
the patient’s lifetime. To explore possible linkages
among the cases, we included this case in the investigation. All 6 cases resided within an approximately
11km radius of the city center, with the shortest distance between any of their last residences being approximately 3.5km. A family history of similar neurodegenerative disorders was absent in all patients, and the
clinical phenotype of the six patients was typical of
Fig. Temporal summary of rural city residence, illness onset,
and death of all patients: (small vertical line) death, (greater
than symbol) onset, (horizontal line) time periods in rural
city, and (asterisk) patient with Gerstmann-SträusslerScheinker syndrome.
sporadic CJD (as outlined in Table 2), with PRNP
genotyping excluding mutations in the five patients in
whom it could be undertaken. The average age at
death in the six patients was 63.0 years, which is similar to the national mean of 65.2 years.
Exhaustive medicodemographic analysis, including a
communal meeting with family members of 4 of the
sporadic cases, could not determine any point-source
or case-to-case transmission link between the 7 persons
(including Patient 7 with Gerstmann-SträusslerScheinker syndrome). In particular, no orthodox medical, alternative medical, or dental practitioner was
shared by all nor were invasive procedures or operations performed on these patients at a single institution. The only association detected was for Patients 2,
3, and 4 (see Tables 1 and 2), who shared the same
ophthalmologist for 11 years, with all 3 first examined
during a 7-month period commencing September
1983. Aside from a single flourescein angiogram in Patient 2, the consultations were for routine clinical assessments, which included application tonometry
(Goldman) on two or three separate occasions each,
with the shortest interpatient interval being approximately 9 weeks.
The simplest assessment of the statistical significance
of the apparent cluster employed the hypothesis that by
chance alone the number of sporadic cases observed in
long-term residents from this city was greater than the
number expected for this period. For the cases dying in
the rural city (Patients 1–5), the relative risk was 5.4
( p ⫽ 0.0027). However, this approach can be questioned on the grounds that the hypothesis was generated from the data rather than being stated a priori. A
more valid test of significance, therefore, applied the
same method to cases arising after the cluster was suspected (represented by the third patient dying of sporadic CJD in 1996). On this basis, the observed rate of
an additional 2 sporadic cases during 4 years developing in residents, compared with the expected (0.284),
still represented a significant excess ( p ⫽ 0.02).
More conservative statistical analysis, however, assessed the likelihood that somewhere in Australia, in a
Collins et al: Creutzfeldt-Jakob Disease Cluster
population unit of the same size as this city, 5 or more
cases of sporadic CJD could occur when 0.923 cases
were expected. With the Australian population divided
equally, there are 241 units of approximately 73,000
people. Under Poisson probability assumptions, the occurrence of a cluster of this size or greater in at least 1
of these 241 population groups quite easily could arise
by chance alone ( p ⫽ 0.49), even if the nonresident
case (Patient 6) is included to make a cluster of 6 ( p ⫽
Spatiotemporal groupings of sporadic CJD have been
reported previously,10 –12 including in the context of
comprehensive epidemiological studies.6,7,13 Generally,
clusters detected during systematic national surveillance
studies are construed as more valid for detailed analysis
by obviating potential selection and detection biases
that may arise when the cluster itself prompts further
It is acknowledged that the relatively short duration
of rural city residence of the most recent case (Patient
5) appears inconsistent with disease acquisition during
this period given the incubation periods determined to
date for recognized peripheral iatrogenic transmission
events.4 Nevertheless, their inclusion was determined
to be important to facilitate the investigation of potentially novel transmission mechanisms. Despite the high
relative risk and extensive investigation of individual
cases, we were unable to identify a putative transmission pathway. Three patients infrequently attended the
same ophthalmologist, but no convincing mechanism
of transmission was uncovered, with the role of routine
clinical tonometry being highly contentious.14 –17 Furthermore, although this cluster is apparently of high
significance when viewed in isolation, more conservative statistical analysis demonstrates that such a grouping of 5 or 6 cases in at least 1 population unit the
same size as the rural city within Australia would not
be unexpected by chance alone over a 13-year period.
Given the cumulative evidence to date, apparent clustering of CJD above background or expected average
national incidence rates is not rare, and each instance
requires careful epidemiological and statistical assessment to determine its true significance.
The Australian National Creutzfeldt-Jakob Disease Registry is
funded by the Commonwealth Department of Health and Aging.
The authors thank Drs M. Pilbeam and G. Hunter for their generous support in facilitating the clinicopathological investigation of
the patients in this study and A. Bell for his technical assistance.
1. Prusiner SB. Prions. Proc Natl Acad Sci U S A 1998;95:
Annals of Neurology
Vol 52
No 1
July 2002
2. Will RG, Ironside JW, Zeidler M, et al. A new variant of
Creutzfeldt-Jakob disease. Lancet 1996;347:921–925.
3. Windl O, Dempster M, Estibeiro JP, et al. Genetic basis of
Creutzfeldt-Jakob disease in the United Kingdom: a systematic
analysis of predisposing mutations and allelic variations in the
PRNP gene. Hum Genet 1996;98:259 –264.
4. Brown P, Preece M, Brandel J-P, et al. Iatrogenic CreutzfeldtJakob disease at the millennium. Neurology 2000;55:
5. Collins S, Law M, Fletcher A, et al. Surgical treatment and risk
of sporadic Creutzfeldt-Jakob disease: a case control study. Lancet 1999;353:693– 697.
6. Matthews WB. Epidemiology of Creutzfeldt-Jakob disease in
England and Wales. J Neurol Neurosurg Psychiatr 1975;38:
210 –213.
7. Cousens S, Smith PG, Ward H, et al. Geographical distribution of variant Creutzfeldt-Jakob disease in Great Britain, 1994 –
2000. Lancet 2001;357:1002–1007.
8. Collins S, Boyd A, Fletcher A, et al. Novel prion protein gene
mutation in an octogenerian with Creutzfeldt-Jakob disease.
Arch Neurol 2000;57:1058 –1063.
9. Masters C, Harris J, Gajdusek D, et al. Creutzfeldt-Jakob
disease: patterns of worldwide occurrence and the significance
of familial and sporadic clustering. Ann Neurol 1979;5:
10. Will RG, Matthews WB. Evidence for case-to-case transmission
of Creutzfeldt-Jakob disease. J Neurol Neurosurg Psychiatr
11. Farmer PF, Kane WC, Hollenberg-Sher J. Incidence of
Creutzfeldt-Jakob disease in Brooklyn and Staten Island.
N Engl J Med 1978;298:283–284.
12. Arakawa K, Nagara H, Itoyama Y, et al. Clustering of three
cases of Creutzfeldt-Jakob disease near Fukuoka City, Japan.
Acta Neurol Scand 1991;84:445– 447.
13. Raubertas RF, Brown P, Cathala F, Brown I. The question of
clustering of Creutzfeldt-Jakob disease. Am J Epidemiol 1989;
129:146 –154.
14. Davanipour Z, Alter M, Sobel E, et al. Creutzfeldt-Jakob
disease: possible medical risk factors. Neurology 1985;35:
15. Alter M. How is Creutfeldt-Jakob disease acquired? Neuroepidemiology 2000;19:55– 61.
16. van Duijn CM, Delasnerie-Laupretre N, Masullo C, et al.
Case-control study of risk factors of Creutzfeldt-Jakob disease
in Europe during 1993–1995. Lancet 1998;351:1081–1085.
17. Zerr I, Brandel J-P, Masullo C, et al. European surveillance on
Creutzfeldt-Jakob disease: a case-control study for medical risk
factors. J Clin Epidemiol 2000;53:747–754.
In Vivo Study Indicating
Loss of Intracortical
Inhibition in TumorAssociated Epilepsy
Kerstin Irlbacher, MD, Stephan A. Brandt, MD,
and Bernd-Ulrich Meyer, MD†
Transcranial magnetic stimulation was performed in 2
patients with focal motor seizures in the right hand
caused by a circumscribed tumor process affecting the
left precentral gyrus. In both cases, paired-pulse transcranial magnetic stimulation showed a loss of intracortical
inhibition for interstimulus intervals of 2 to 4msec that
was replaced by an enormous facilitation in the lesioned
hand motor cortex. The uniform impairment of inhibitory mechanisms in epileptogenic tumors with different
histologies suggests a common, nonspecific cause of
tumor-related epileptogenesis.
Ann Neurol 2002;52:119 –122
The pathophysiology of tumor-associated epilepsy is
not well understood. Histological and immunohistochemical studies suggest that the balance between intracortical inhibitory and excitatory mechanisms appears to be shifted toward excitation in epileptogenic
brain tumors.1 By applying focal transcranial magnetic
stimulation (TMS) in 2 patients with circumscribed
histologically different tumors infiltrating the hand
motor cortex and provoking focal motor seizures, we
had the rare opportunity to study noninvasively excitatory and inhibitory neuronal mechanisms within a cortex area generating epileptic fits. In both patients, motor function of the affected hand was only moderately
impaired, so that it could be expected that many corticospinal motor neurons were still functioning. This
fact made the epileptogenic hand motor cortex ideally
accessible for the application of TMS. A paired-pulse
From the Department of Neurology, Charité, HumboldtUniversity, Berlin, Germany.
Dr Bernd-Ulrich Meyer died in an air accident near Zürich, Switzerland on November 24, 2001.
Received Oct 30, 2001, and in revised form Jan 28, 2002. Accepted
for publication Feb 25, 2002.
Published online Jun 18, 2002, in Wiley InterScience
( DOI: 10.1002/ana.10229
Address correspondence to Dr Irlbacher, Unit for Motor Disturbances and Cortex Function, Department of Neurology, Charité,
Campus Virchow-Klinikum, Humboldt-University, Augustenburger
Platz 1, D-13353 Berlin, Germany.
paradigm2 was applied to test intracortical inhibition,
presumably related to GABA(A)ergic mechanisms,3
and intracortical facilitation, presumably related to glutamatergic neuronal mechanisms.4 Furthermore, we
measured the duration of the cortical silent period
(CSP) in muscles contralateral to cortex stimulation
probably reflecting, at least in its later part, activity
in gamma-aminobutyric acid (GABA)ergic cortical inhibitory circuits, mediated by GABA(B) receptors.5,6
Using this approach, we found a selective breakdown
of intracortical inhibition in the affected epileptogenic
motor cortex in both patients.
Patients and Methods
Patient 1
A 52-year-old man had three focal myoclonic seizures of the
right hand within 2 weeks preceding our study. Fist grip of
the right hand was moderately paretic (Medical Research
Council grade 4), tapping frequency was slightly reduced
(4.8 vs 6.0Hz), and sequential and complex finger movements were slow. Sensory functions of the hand and median
nerve somatosensory evoked potentials were normal. Interictal electroencephalogram showed slow-wave activity in the
left centroparietal and temporal region. T1-weighted
gadolinium-enhanced magnetic resonance imaging showed a
small lesion (1.0 ⫻ 0.6 ⫻ 1.6 cm) within the hand knob of
the precentral gyrus7 with homogeneous solid contrast enhancement (Figs 1a and b) and a local space-occupying
edematous effect (see Fig 1c). One month later and under
dexamethasone therapy, a contrast-enhancing tumor with
early central necrosis had grown (3.2 ⫻ 3.0 ⫻ 3.2 cm; see
Fig 1d). Another 5 months later and after radiation with 70
grays, a subtotal resection of the tumor was performed. Histopathological examination showed a glioblastoma multiforme (World Health Organization grade 4).
Patient 2
A 62-year-old woman, with a known adenocarcinoma of the
lung, had three focal myoclonic seizures of the right hand 2
days before our study. Fist grip of the right hand was paretic
(Medical Research Council grade 3– 4), tapping frequency
was reduced (2.1 vs 4.2Hz), but sensory functions were unremarkable. Under dexamethasone therapy, another magnetic
resonance image showed a local space-occupying, round metastasis (2.1 ⫻ 2.1 ⫻ 1.9 cm), with irregular contrast enhancement located in the hand knob of the left precentral
gyrus (Fig 2).
With written informed consent and approval of the local
ethics committee, TMS studies were performed in both patients 1 day after acquisition of the magnetic resonance image (see Figs 1 and 2) and before anticonvulsive medication
was started. Both patients had no further seizures during hospitalization. Focal TMS (figure-of-eight–shaped coil, 8.5cm
outside diameter of half-coil; posteroanterior direction of induced currents; Magstim 200; Magstim, Dyfed, United
Kingdom) was consecutively performed over both hand-
© 2002 Wiley-Liss, Inc.
Fig 1. Histologically proven glioma located
within the hand knob of the precentral
gyrus in Patient 1 (time to inversion
[T1]-weighted gadolinium-enhanced magnetic resonance imaging, repetition time
[TR] 560msec/echo time [TE] 14msec)
with a homogeneous solid contrast enhancement (a, b) and local space-occupying
edematous effect (c) (TR 6000/TE
100msec). Under medication with dexamethasone, a follow-up 1 month later
showed a grown contrast-enhancing tumor
with beginning central necrosis (d) (TR
6000/TE 150, TI 2000).
associated motor cortices, with the coil centered over a position at which maximal responses could be elicited. Surface
electromyographic signals were recorded from the first dorsal
interosseous muscles of both hands and then fed into an
analogue-digital converter.
Single-pulse TMS was used to measure (1) resting and
active motor thresholds (percentage of maximum stimulator
output), (2) peak-to-peak amplitudes (in millivolts; stimulus
intensity: 1.5 the motor threshold), and (3) duration of the
CSP (measured as the interval between onset of the motor
response and the end of the complete suppression of tonic
voluntary electromyographic contraction) at stimulus intensities increasing in steps of 5% of the maximum stimulator
output (only Patient 1). Intracortical inhibition and facilitation were studied with the previously described paired-pulse
technique.2 The intensity of the conditioning stimulus was
set to 95% of the active motor threshold, and the intensity
of the test stimulus was set so that the test stimulus alone
produced a response of 0.5 to 1mV peak-to-peak size.
Stimulation effects obtained in the patients were compared intraindividually between affected and unaffected sides
and with 12 right-handed, healthy volunteers (mean plus or
minus standard deviation, 53.0 ⫾ 6.6 years).
Single-pulse Stimulation
The affected motor cortex showed elevated resting
thresholds (Patient 1, 60 vs 42%; Patient 2, 61 vs
42%; control, 39.6 ⫾ 6.6%; right-left difference,
3.3 ⫾ 3.4) and reduced amplitudes (Patient 1, 1.3 vs
Annals of Neurology
Vol 52
No 1
July 2002
7.5mV; Patient 2, 2.3 vs 4.6mV; control, 7.5 ⫾
1.9mV; right-left difference, 1.7 ⫾ 1.2mV). The active
thresholds were lower, and the side-to-side difference
was reduced (Patient 1, 35 vs 30%; Patient 2, 40 vs
38%). Response latencies were normal (Patient 1, 21.6
vs 21.0msec; Patient 2, 22.0 vs 20.8msec; control,
20.7 ⫾ 1.9msec; right-left difference, 0.7 ⫾ 0.1).
With increasing stimulus intensities, the duration of
the CSP was the same on both sides (Fig 3) and lay
within normal range in both patients (Patient 2, 177 vs
140msec; control, 176 ⫾ 38msec; stimulus intensity,
80% maximum).
Paired-pulse Stimulation
In normal subjects and in the patients’ unaffected
hands, inhibition of test motor responses occurred at
interstimulus intervals (ISIs) of 1 to 5msec (values ⬍
100%), whereas facilitation occurred at ISIs of 7, 10,
and 15msec (values ⬎ 100%; see Fig 3). For the affected hand, motor cortex inhibition was present only
for an ISI of 1msec and was absent for ISIs of 2 to
5msec in both patients. The latter was replaced by an
enormous facilitation of 250 to 500% (see Fig 3) for
ISIs of 2 to 7msec.
In 2 patients with focal motor seizures of the right
hand caused by a circumscribed glioblastoma or metas-
facilitation of 250 to 500% for ISIs of 2 to 7msec in
the affected motor cortex area. Intracortical inhibition
occurring at intervals of 2 to 5msec has been enhanced
by GABA(A) receptor agonists and by N-methyl-Daspartate receptor blockers.3,4,8 Facilitation occurring at
short ISIs could be due to an overwhelming activity of
glutamatergic excitatory influences and simultaneous
lack or reduction of GABAergic inhibition because of
either reduced excitability or loss of inhibitory neurons
or changes in GABAergic receptor function. EnhanceFig 3. (A) Intracortical inhibition and facilitation as assessed
by the transcranial magnetic stimulation paired-pulse paradigm in normal subjects and in 2 patients with epileptogenic
processes unilaterally affecting the hand motor cortex. Note the
loss of intracortical inhibition for interstimulus intervals (ISIs)
of 2 to 5msec in both patients that is replaced by an enormous facilitation for ISIs of 2 to 7msec. (B) Normal duration
of the cortical silent period generated in the affected and nonaffected motor cortex of Patient 1 at different stimulus intensities. ms ⫽ milliseconds; Pat. ⫽ patient.
Fig 2. Cerebral metastasis of a lung adenocarcinoma in Patient 2 located within the hand knob of the left precentral
gyrus with local space-occupying effect (T1-weighted
gadolinium-enhanced magnetic resonance imaging, TR 12/TE
5msec; under medication with dexamethasone).
tasis, TMS of the affected hand motor cortex was used
to assess the function of corticospinal neurons and neurons with influence on them. In both cases, a decreased
corticospinal output from the affected cortex area is
suggested by increased thresholds and reduced amplitudes of the cortically elicited hand motor responses.
This may be because of a reduced density and number
of corticospinal neurons and is in agreement with the
clinical finding of slight weakness and impaired finger
movements. The most significant finding was that in
both patients intracortical inhibition was lacking for
ISIs of 2 to 5msec and was replaced by an enormous
Irlbacher et al: Loss of Intracortical Inhibition
ment of excitability of glutamatergic neurons is less
likely because facilitation at ISIs of 10 and 15msec was
less marked. The finding of normal inhibition at an ISI
of 1msec in the paired-pulse paradigm is compatible
with the assumption that this type of inhibition is
caused by axonal refractoriness9 and, therefore, is not
influenced by the tumor-related epileptogenetic processes. The conditioning stimulus appears to activate
some of the excitatory inputs, being subthreshold for
discharging pyramidal tract neurons. If there was no
superimposed inhibition, then a test stimulus at an ISI
greater than 1msec would sum with this excitation and
produce the facilitation observed. Therefore, the facilitation appears to reflect an active excitatory process
rather than a simple lack of inhibition. Because loss of
inhibition was present in both patients, epileptogenesis
may be independent of the histology of the tumor.
The CSP generated in the affected motor cortices of
both patients had a normal duration. The silent period
beyond 50msec clearly can be attributed to cortical
mechanisms10 and, in particular, to GABA(B)mediated inhibitory neuronal circuits.5,6 Therefore, the
aforementioned loss of paired-pulse inhibition in the
presence of a preserved CSP suggests a relatively selective impairment of local GABA(A)-mediated inhibition
within the border zone of the glioblastoma or the motor cortex infiltrated by the metastasis, probably as a
common, nonspecific phenomena of tumor-related epileptogenesis.
The CSP was elicited by a large stimulus that probably recruits inhibitory inputs from areas outside the
tumor. These inputs appear to work normally and produce a normal silent period.
This noninvasive study reports alterations of cortical
excitability in epileptogenic brain tumors located
within the motor cortex with TMS. In future studies,
this stimulation technique may be useful to assess the
potency of anticonvulsive drugs to control cortical hyperexcitabilty in brain tumors.
6. Siebner HR, Dressnandt J, Auer C, Conrad B. Continuous intrathecal baclophen infusions induced a marked increase of the
transcranially evoked silent period in a patient with generalized
dystonia. Muscle Nerve 1998;21:1209 –1212.
7. Yousry TA, Schmid DU, Alkadhi H, et al. Localization of the
motor hand area to a knob on the precentral gyrus. A new
landmark. Brain 1997;120:141–157.
8. Di Lazzaro V, Oliviero A, Meglio M, et al. Direct demonstration of the effect of lorazepam on the excitability of the human
motor cortex. Clin Neurophysiol 2000;111:794 –799.
9. Nakamura Y, Rothwell J, Mondugno N. The micro-physiology
of intracortical inhibition studied using paired-pulse transcranial magnetic stimulation (TMS) in intact humans. Clin Neurophysiol 2000;111(suppl 1):S24.
10. Meyer BU, Kühn A, Röricht S, Kupsch A. Direct activation of
corticospinal fibres at the level of the internal capsule in man.
Clin Neurophysiol 2001;112:S7.
Germline Mosaicism of a
Novel Mutation in
Membrane Protein-2
Deficiency (Danon Disease)
Maki Takahashi, MD,1 Ayaka Yamamoto, MD,2
Kyoko Takano, MD,1 Akira Sudo, MD,1
Takahito Wada, MD, PhD,1 Yu-ichi Goto, MD, PhD,3
Ichizo Nishino, MD, PhD,2,4
and Shinji Saitoh, MD, PhD1
We identified a family with lysosome-associated membrane
protein-2 deficiency (Danon disease) associated with a
novel 883 ins-T mutation in the lysosome-associated membrane protein-2 gene located at Xq24. Although the affected son and daughter carried the same mutation, it was
not detected in their mother’s peripheral blood or buccal
cells; this indicated germline mosaicism. This is the first
molecular evidence for germline mosaicism in Danon disease and has important implications for genetic counseling.
Ann Neurol 2002;52:122–125
1. Beaumont A, Whittle IR. The pathogenesis of tumour associated epilepsy. Acta Neurochir (Wien) 2000;142:1–15.
2. Kujirai L, Caramia MD, Rothwell JC, et al. Corticocortical inhibition in human motor cortex. J Physiol (Lond) 1993;471:
3. Ziemann U, Lönnecker S, Steinhoff BJ, Paulus B. The effect of
lorazepam on motor cortical excitability in man. Exp Brain Res
4. Ziemann U, Chen R, Cohen LG, Hallett M. Dextromethorphan decreases excitability of the human motor cortex. Neurology 1998;51:1320 –1324.
5. Werhahn KJ, Kunesch E, Nochtar S, et al. Differential effects
on motorcortical inhibition induced by blockade of GABA uptake in humans. J Physiol (Lond) 1999;517:591–597.
© 2002 Wiley-Liss, Inc.
From the 1Department of Pediatrics, Hokkaido University School
of Medicine, Sapporo; and the Departments of 2Ultrastractural
Research, 3Mental Retardation and Birth Defect Research, and
Neuromuscular Research, National Institute of Neuroscience, National Center of Neurology and Psychiatry, Kodaira, Japan.
Received Nov 17, 2001, and in revised form Mar 4, 2002. Accepted
for publication Mar 4, 2002.
Published online May 31, 2002, in Wiley InterScience
( DOI: 10.1002/ana.10235
Address correspondence to Dr Saitoh, Department of Pediatrics,
Hokkaido University School of Medicine, N-15, W-7, Kita-ku,
Sapporo 060-8638, Japan. E-mail:
Danon and colleagues1 first reported in 1981 on 2 unrelated boys with cardiomyopathy, proximal myopathy,
and mental retardation as lysosomal glycogen storage
disease with normal acid maltase. LAMP-2, which encodes lysosome-associated membrane protein-2 (LAMP2), has been identified as the causative gene, and it is
located in Xq24.2– 4 Because lysosomal glycogen storage
disease with normal acid maltase encompasses heterogeneous disorders, the names LAMP-2 deficiency and
Danon disease (DD) have been proposed for the disease
caused by primary LAMP-2 deficiency.2 Female DD patients often manifest milder phenotypes, usually with
cardiac involvement, suggesting X-linked dominant inheritance with significant phenotypic differences between sexes.5,6 Because sudden death has been reported
in DD, it is crucial to make the diagnosis early to provide genetic counseling and to optimize clinical management, with particular attention to the cardiac manifestations.6 Nevertheless, the diagnosis of DD has
been difficult to achieve because the characteristic tiny
vacuoles in muscle biopsies are easily overlooked unless
samples are stained histochemically for nonspecific esterase or acetylcholinesterase. Furthermore, the disease
is rarely suspected in female patients. Here we report a
family of DD with 2 affected siblings carrying a novel
LAMP-2 mutation. The application of anti–LAMP-2
antibody to muscle biopsies and molecular genetic
analyses will facilitate the early diagnosis, genetic counseling, and treatment of DD.2
Patients and Methods
The patient was a 7-year-old boy with normal early motor
and cognitive development who was referred to us because
he had an abnormal electrocardiogram (EKG) and elevated
blood creatine kinase (CK). EKG indicated left ventricular
hypertrophy, and a subsequent ultrasonic cardiogram (UCG)
examination indicated a hypertrophic cardiomyopathy. A
physical examination did not indicate a cardiac arrhythmia,
heart murmur, or signs of cardiac failure. A neurological examination disclosed normal mentation and mild weakness of
neck muscle. He attended regular school; however, his running capacity and motor skills were far below the average.
Blood tests indicated an elevated CK level of 2,333IU/L
(normal range: male, 40 –165IU/L; female, 30 –165IU/L). A
muscle biopsy of his left biceps brachii muscle was performed after informed consent was obtained from his parents.
The patient had two siblings. His 6-year-old sister had
developed normally and had an unremarkable clinical examination. However, an EKG examination indicated left ventricular hypertrophy, and a UCG examination demonstrated
hypertrophic cardiomyopathy. Surprisingly, the thickness of
her ventricle wall was the same as that seen in the patient.
Her serum CK level was almost normal (171IU/L). The patient’s 2-year-old brother had a normal clinical examination,
although his serum CK level was slightly elevated (265IU/L).
EKG and UCG examinations were normal. The 30-year-old
mother was normal on clinical examination. EKG and UCG
tests did not show any cardiac abnormalities. She had a normal level of serum CK (82IU/L).
Muscle Pathology
Serial sections of a muscle specimen were stained with hematoxylin and eosin, modified Gomori trichrome, and a battery of histochemical methods and were immunostained with
the monoclonal antibodies against dystrophin (NCL-DYS1,
2, and 3; Novocastra, Newcastle upon Tyne, United Kingdom), merosin (5H2; Chemicon, Temecula, CA), and
LAMP-2 (H4B4; Developmental Studies Hybridoma Bank,
University of Iowa, Iowa City, IA), as reported.2
LAMP-2 Mutation Detection
Genomic DNA was extracted from peripheral blood leukocytes from the patient, younger sister, younger brother, and
the mother after informed consent was obtained. DNA was
also extracted from buccal cells of the mother. Coding sequences of LAMP-2 were amplified by polymerase chain reaction and sequenced as reported.2
A pathological examination of the muscle from the patient demonstrated moderate variation in fiber size and
tiny vacuoles scattered in cytoplasm (Fig 1A). No necrotic or regenerating fibers were seen. The vacuolar
membrane had high activities of acetylcholinesterase (see
Fig 1B) and nonspecific esterase and was decorated by
antibodies to dystrophin and merosin (data not shown).
These pathological findings are hallmarks of DD. Immunostaining with the anti–LAMP-2 monoclonal antibody demonstrated the complete absence of LAMP-2
(see Fig 1C), confirming the diagnosis of DD.
A hemizygous insertion of T at nucleotide 883 (883
ins-T) was identified in the patient (Fig 2). This previously unreported mutation is predicted to cause a
reading frameshift and to create a new stop codon
13bp from the insertion. In blood DNA, the mutation
was present in a heterozygous state in the sister but was
absent in the brother and mother (see Fig 2). The mutation was also not detected in the mother’s buccal
smear DNA (see Fig 2).
We have described an unusual family with DD. The
patient, a 7-year-old boy, demonstrated typical features, including hypertrophic cardiomyopathy, mild
proximal muscle weakness, and elevated CK level but
no mental retardation. The affected sister demonstrated
a milder phenotype. Her cardiac involvement was only
detectable by EKG and UCG. According to earlier reports, female patients with DD develop cardiac symptoms during middle age.1,5,6 However, our findings
suggest that subclinical cardiac involvement can be
demonstrated by EKG or UCG as early as 6 years of
Takahashi et al: Germline Mosaicism in Danon Disease
Fig 1. Histological examination of the muscle biopsy specimen from the patient. (A) On hematoxylin and eosin, tiny vacuoles often
look like basophilic granules (arrows). (B) Histochemical staining for acetylcholinesterase shows the enzyme activity in the vacuolar
membrane (arrow). (C) Immunostaining for LAMP-2 demonstrates the complete absence of LAMP-2 in the skeletal muscle from
the patient. (D) LAMP-2 is present in a granular pattern in control muscle.
age. We initially thought that the mother would be a
heterozygous patient and so evaluated her cardiac function. However, EKG and UCG examinations were
normal. These findings were explained by germline
mosaicism of the mutation in the mother.
Although conventional histochemical stainings, in-
cluding nonspecific esterase and acetylcholinesterase, of
muscle biopsies are useful for identifying autophagic
vacuolar myopathies, immunostaining for LAMP-2 is
crucial for making the specific diagnosis of DD because
infantile autophagic vacuolar myopathy and X-linked
myopathy with excessive autophagy show similar intra-
Fig 2. Mutation detection in the family.
An insertion of T at nt.-883 (arrow) was
demonstrated in the patient (hemizygous)
and in the sister (heterozygous) but was
not present in leukocyte DNA from the
brother or in DNA from the mother’s leukocytes or buccal cells.
Annals of Neurology
Vol 52
No 1
July 2002
cytoplasmic vacuoles.7–10 In the patient’s muscle,
LAMP-2 was completely absent, thereby confirming
that the patient had DD.
We have identified a novel mutation of 883 ins-T in
the patient and his affected sister. This mutation, located in exon 7, causes a frameshift and introduces a
stop codon 13bp from the insertion.2 All eight reported mutations are predicted to cause truncation of
LAMP-2. No missense mutations have been found in
DD. Therefore, complete loss of function of LAMP-2
might be necessary to develop typical DD. In our patient, the absence of LAMP-2 protein by immunostaining highly suggests the complete loss-of-function nature of the mutation. This result suggests that a slight
increase of LAMP-2 expression in muscle, for example,
by gene therapy might improve the disease.
The mutation was not detected in peripheral blood
or buccal cells from the mother. Therefore, the 883
ins-T mutation was most likely introduced into the
maternal germline as a de novo mutation. This is the
first report of germline mosaicism in DD. Our results
have significant implications for genetic counseling.
The possibility of germline mosaicism must be considered in genetic counseling for DD when mutations are
not detected in the mothers of patients.
This work was supported in part by a grant from the Japanese Ministry of Education, Science, and Culture (S.S.).
We thank Drs Kunihiko Kobayashi, Richard J. Gibbons, and Michio Hirano for discussions and critical comments on the manu-
script and Dr Takeo Kubota, Kumiko Murayama, and Fumie Igarashi for their technical assistance.
1. Danon MJ, Oh SJ, DiMauro S, et al. Lysosomal glycogen storage disease with normal acid maltase. Neurology 1981;31:
2. Nishino I, Fu J, Tanji K, et al. Primary LAMP-2 deficiency
causes X-linked vacuolar cardiomyopathy and myopathy
(Danon disease). Nature 2000;406:906 –910.
3. Tanaka Y, Guhde G, Suter A, et al. Accumulation of autophagic vacuoles and cardiomyopathy in LAMP-2 deficiency mice.
Nature 2000;406:902–906.
4. Mattei MG, Matterson J, Chen JW, et al. Two human lysosomal membrane glycoproteins, h-lamp-1 and h-lamp-2, are encoded by genes localized to chromosome 13q34 and chromosome Xq24 –25, respectively. J Biol Chem 1990;265:
7548 –7551.
5. Byrne E, Dennett X, Crotty B, et al. Dominantly inherited cardioskeletal myopathy with lysosomal glycogen storage disease
and normal acid maltase levels. Brain 1986;109:523–536.
6. Dworzak F, Casazza F, Mora M, et al. Lysosomal glycogen storage disease with normal acid maltase: a familial study with successful heart transplant. Neuromusc Disord 1994;4:243–247.
7. Kalimo H, Savontaus ML, Lang H, et al. X-linked myopathy
with excessive autophagy; a new hereditary muscle disease. Ann
Neurol 1988;23:258 –265.
8. Villanova M, Louboutin JP, Chateau D, et al. X-linked vacuolated myopathy; complement membrane attack complex on surface membrane of injured muscle fibers. Ann Neurol 1995;37:
637– 645.
9. Auranen M, Villanova M, Muntoni F, et al. X-linked vacuolar
myopathies: two separate loci and refined genetic mapping.
Ann Neurol 2000;47:666 – 669.
10. Yamamoto A, Morisawa Y, Verloes A, et al. Infantile autophagic vacuolar myopathy is distinct from Danon disease. Neurology 2001;57:903–905.
Takahashi et al: Germline Mosaicism in Danon Disease
Без категории
Размер файла
3 019 Кб
carcinoma, antibodies, merkel, anti, cells
Пожаловаться на содержимое документа