Download Document 8573227

Blackwell Publishing LtdOxford, UKNANNeuropathology and Applied Neurobiology0305-1846Blackwell Publishing Ltd, 20062006325469482Original ArticleJNCL serum immunoreactivityM. J. Lim
et al.
Neuropathology and Applied Neurobiology (2006), 32, 469–482
doi: 10.1111/j.1365-2990.2006.00738.x
Distinct patterns of serum immunoreactivity as
evidence for multiple brain-directed autoantibodies
in juvenile neuronal ceroid lipofuscinosis
M. J. Lim*, J. Beake*, E. Bible*, T. M. Curran†, D. Ramirez-Montealegre†, D. A. Pearce†‡§ and
J. D. Cooper*
*Pediatric Storage Disorders Laboratory, Centre for the Cellular Basis of Behaviour, Department of Neuroscience, MRC Social
Genetic and Developmental Psychiatry Centre, Institute of Psychiatry, King’s College London, De Crespigny Park, London,
UK, †Center for Ageing and Developmental Biology, ‡Department of Neurology, and §Department of Biochemistry and
Biophysics, University of Rochester School of Medicine and Dentistry, University of Rochester, Rochester, NY, USA
M. J. Lim, J. Beake, E. Bible, T. M. Curran, D. Ramirez-Montealegre, D. A. Pearce and J. D. Cooper (2006) Neuropathology and Applied Neurobiology 32, 469–482
Distinct patterns of serum immunoreactivity as evidence for multiple brain-directed autoantibodies
in juvenile neuronal ceroid lipofuscinosis
Autoantibodies to glutamic acid decarboxylase (GAD65)
have been reported in sera from the Cln3–/– mouse model of
juvenile neuronal ceroid lipofuscinosis (JNCL), and in individuals with this fatal paediatric neurodegenerative disorder. To investigate the existence of other circulating
autoreactive antibodies, we used sera from patients with
JNCL and other forms of neuronal ceroid lipofuscinosis
(NCL) as primary antisera to stain rat and human central
nervous system sections. JNCL sera displayed characteristic patterns of IgG, but not IgA, IgE or IgM immunoreactivity that was distinct from the other forms of NCL.
Immunoreactivity of JNCL sera was not confined to
GAD65-positive (GABAergic) neurons, but also stained
multiple other cell populations. Preadsorption of JNCL
sera with recombinant GAD65 reduced the intensity of
the immunoreactivity, but did not significantly change its
staining pattern. Moreover, sera from Stiff Person Syndrome and Type I Diabetes, disorders in which GAD65
autoantibodies are present, stained with profiles that were
markedly different from JNCL sera. Collectively, these
studies provide evidence of the presence of autoreactive
antibodies within multiple forms of NCL, and are not
exclusively directed towards GAD65.
Keywords: autoantibodies, autoimmunity, Batten disease, immunoreactivity, neuronal ceroid lipofuscinoses, patientderived sera
Introduction
In recent years, an increasing number of central nervous
system (CNS) disorders have been suggested to have an
Correspondence: Jonathan D. Cooper, Pediatric Storage Disorders
Laboratory, Centre for the Cellular Basis of Behaviour, Department of
Neuroscience, Box P040, MRC Social Genetic and Developmental
Psychiatry Centre, Institute of Psychiatry, King’s College London, De
Crespigny Park, London SE5 8AF, UK. Tel: +44 20 7848 0286; Fax:
+44 20 7848 0273; E-mail: [email protected]
© 2006 Blackwell Publishing Ltd
autoimmune basis. This autoimmune component may
involve the adaptive immune system, via autoantibody
production [1], or abnormal populations of activated T
cells [2]. To demonstrate humoral pathogenicity, the presence of autoantibodies in a target organ, the induction of
disease by passive transfer of autoantibodies and a clinical
response to specific immunomodulatory therapy are all
required [3]. Despite autoantibodies being reported in a
growing number of disorders, a pathogenic role for these
autoantibodies has only been shown convincingly in a
handful of neurological disorders [1,3].
469
470
M. J. Lim et al.
An autoimmune response has also been described
recently in murine and human juvenile neuronal ceroid
lipofuscinosis (JNCL) [4,5], a fatal paediatric storage
disorder. JNCL, also known as Batten disease, is one of
the neuronal ceroid lipofuscinoses (NCLs), a group of at
least eight genetically distinct lysosomal storage disorders (LSDs) [6]. These autosomal recessive disorders
(CLN1–CLN8) commonly present during childhood with
an infantile (INCL), late infantile (LINCL) or juvenile
(JNCL) onset [6–8]. Each form is fatal and clinical signs
include visual failure leading to blindness, an increased
severity of untreatable seizures and neurocognitive
decline; leading up to an inevitable premature death
[6,8].
JNCL is the result of mutations in the CLN3 gene that
codes for a transmembrane protein whose precise function remains unknown [7,9,10]. CLN3-null mutant mice
(Cln3–/–) present with a JNCL-like phenotype, including
the intralysosomal accumulation of autofluorescent storage material and the loss of subpopulations of GABAergic
interneurons [11,12]. Both Cln3–/– mice and individuals
with JNCL have circulating autoantibodies to glutamic
acid decarboxylase (GAD65) that inhibit this enzyme’s
ability to convert glutamic acid to γ-aminobutyric acid
and results in presynaptic elevation of glutamate [4].
Whether these autoantibodies contribute to pathogenesis
remains unclear, but Western blots of whole brain extracts
probed with JNCL sera reveal multiple reactive bands [4],
suggesting the presence of a variety of brain-directed
autoantibodies.
A simple means to begin mapping these other autoantigens in the CNS is to use patient-derived sera as primary
antisera upon tissue sections. This approach has been
used to demonstrate brain-directed autoantibodies in sera
from patients with paraneoplastic neurological diseases
[13], Stiff Person Syndrome (SPS) [14], and as an investigative tool for demonstrating autoimmunity in a wide
range of other CNS and non-CNS disorders [15–17]. Here
we demonstrate characteristic patterns of immunoreactivity on rat and human CNS tissue using JNCL sera. This
immunoreactivity is markedly different from SPS and
Type I Diabetes (TID), two other conditions where GAD65
antibodies are raised. Taken together, these data demonstrate that GAD65 antibodies contribute to only a small
proportion of the JNCL serum immunoreactivity. Moreover, our data provide evidence for the presence of multiple other as yet unidentified autoantibodies in JNCL and in
other forms of NCL.
Materials and methods
Patient-derived serum
Serum samples from patients with NCLs [INCL, n = 2;
LINCL, n = 3; JNCL, n = 59; and vLINCL (CLN6), n = 1]
were provided by (i) volunteers at annual Batten Disease
Support and Research Association (BDSRA) Family Meetings; (ii) Professor P. Santavuori (University of Helsinki,
Finland); and (iii) Dr A. Fensom (Guy’s Hospital NHS
Trust, UK). All JNCL serum samples were confirmed as
GAD65 immunoreactive via binding to recombinant
human GAD65, as described previously [4]. SPS patientderived sera (n = 7) were provided by Dr M. Dalakas
(National Institute of Neurological Disorders and Stroke
(NINDS) (USA) and Professor A. Vincent (University of
Oxford, UK); and TID patient-derived sera (n = 6) were provided by Dr M. Atkinson (University of Florida, USA). For
comparison we used sera from neurologically normal
individuals (n = 12), paraneoplastic controls (n = 2) and a
range of other LSDs provided by Dr D. Wenger (Jefferson
Medical College, USA), and Dr A. Fensom (Guy’s Hospital
NHS Trust, UK) (GM1 gangliosidoses, Tay-Sachs, Sandhoff, Hurler, I-Cell, Niemann-Pick Type A, Galactosialidosis and Gaucher; n = 32). Samples were obtained with
appropriate informed consent and after local ethical
review. The use and storage of these samples for analysis
at the Pediatric Storage Disorders Laboratory (PSDL) was
approved by the Institute of Psychiatry Ethical Committee
(Research) (approval number 181/02). Derivation of
patient sera was performed by centrifuging the samples
for 10 min at 3000 g, with sera subsequently stored at
−70°C.
Tissue preparation
Animals Six-week-old male Sprague-Dawley rats (250–
280 g, Charles River, Margate, UK, n = 9) were housed at
21 ± 1°C under an alternating 12-h light/12-h dark cycle
and perfusion procedures carried out as required by the
UK Animals (Scientific Procedures) Act 1986. Animals
were deeply anaesthetized with sodium pentobarbitone
(100 mg/kg) and transcardially perfused with vascular
rinse (0.8% NaCl in 100 mM NaHPO4) followed by a
freshly made and filtered solution of 4% paraformaldehyde
in 0.1 M sodium phosphate buffer, pH 7.4. Brains were
subsequently removed, postfixed overnight at 4°C in the
© 2006 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 32, 469–482
JNCL serum immunoreactivity
same fixative and cryoprotected at 4°C in a solution of
30% sucrose in Tris buffered saline (TBS: 50 mM Tris,
pH 7.6) containing 0.05% NaN3. To provide sections for
the detailed mapping of serum immunoreactivity, brains
were bisected along the midline and 40-µm frozen coronal
or sagittal sections cut through either hemisphere (Leitz
1321 freezing microtome, Leica Microsystems, Milton
Keynes, UK). Sections were collected, one per well, into 96
well plates containing a cryoprotectant solution (TBS/
30% ethylene glycol/15% sucrose/0.05% sodium azide)
and stored at −40°C prior to histological processing.
Human CNS tissue Paraffin-embedded tissue samples
from neurologically normal subjects (n = 2) were obtained
from the MRC London Neurodegenerative Diseases Brain
Bank. At autopsy, tissues were fixed immediately by
immersion in 4% neutral buffered formaldehyde and subsequently routinely processed and embedded in paraffin
wax. Study protocols for the use of human material were
approved by the Ethical Research Committees of the Institute of Psychiatry (approval numbers 223/00, 181/02).
Eight-µm sections were cut from paraffin-embedded
blocks of temporal lobe using a sledge microtome (Leica
SM2400, Leica Microsystems (UK) (Ltd). Sections were
then floated onto a water bath at 40°C prior to mounting
onto Superfrost-plus glass microscope slides (VWR International, Poole, UK).
Immunohistochemical staining using
patient-derived sera
Rat CNS tissue Human-derived sera were used as primary antisera and incubated with both sagittal and coronal rat CNS sections with no prior knowledge of condition.
For each serum sample, at least two parasagittal sections
were stained, together with coronal sections at the level of
the primary motor cortex, rostral hippocampus, substantia nigra and cerebellum. Following classification into different staining categories (see below), representative
serum samples from each staining category were retested
on every sixth coronal or sagittal sections in order to map
the full extent of serum immunoreactivity.
Briefly, free floating sections were incubated in 1% H2O2
in TBS for 15 min, rinsed in TBS and blocked for 40 min
with TBS/0.3% Triton X-100 (TBST)/15% normal goat
serum before incubation overnight at 4°C in the respective patient serum diluted at 1:500 in TBST. Subsequently
471
sections were rinsed in TBS and incubated for 2 h with
biotinylated goat anti-human IgG (heavy and light chain
BA 3000, Vector Laboratories, Peterborough, UK) at a
dilution of 1 : 1000 in TBST/10% normal goat serum. Following rinsing in TBS, immunoreactivity was detected by
incubation for 2 h in an avidin–biotin–peroxidase complex in TBS (Elite ABC kit, Vectastain, Vector Laboratories). Sections were then rinsed in TBS and staining was
developed by incubation for 12 min in 0.05% DAB
(Sigma, Dorset, UK) and 0.001% H2O2 in TBS. The reaction was stopped by immersion in excess ice-cold TBS. Sections were rinsed again, and then mounted, air-dried,
cleared in xylene and cover slipped with DPX (VWR). In
every staining run relevant positive and negative controls
were used.
Human tissue JNCL patient sera immunoreactivity were
mapped on human temporal lobe sections using paraffinembedded sections with a modified version of previously
published immunoperoxidase methods [18], using patient
sera diluted 1 : 500 followed by biotinylated goat antihuman IgG (heavy and light chain BA 3000, Vector Laboratories, 1 : 200).
Subtype specificity To determine the subtype of immunoglobulins in patient sera that bind to tissue sections, a parallel series of sections were incubated with secondaries
recognizing the various subtypes of immunoglobulins
[biotinylated goat anti-human IgM (µ chain BA-3020);
IgG (γ chain BA 3080); IgG (heavy and light chain BA
3000); IgA (α chain); IgE (ε chain); κ chain and λ chain
(all from kit BAK 1000), Vector Laboratories] at a dilution
of 1 : 1000 in TBST/10% normal goat serum.
Other conditions that raise GAD65
autoantibodies
Comparative studies with other conditions where GAD65
antibodies are found were also performed using serum
samples as primary antisera at optimized dilutions of
1 : 1000 (TID) or 1 : 5000 (SPS), using the protocol
described above. To survey the distribution of GAD65, a
series of sections was stained with a monoclonal antibody
to GAD (Stressgen Biotechnologies, Victoria, BC, Canada,
1 : 500) using a biotinylated goat anti-mouse secondary
(BA-9200 Vector Laboratories, 1 : 500) and ABC reagent
as described above.
© 2006 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 32, 469–482
472
M. J. Lim et al.
Colocalization studies
To study the extent of colocalization of patient-derived sera
with GAD65, rat CNS sections were blocked for 40 min
with TBS/0.3% Triton X-100 (TBST)/15% normal goat
serum before incubation overnight at 4°C in a solution
containing both mouse anti-GAD (Stressgen Biotechnologies) at 1 : 500 dilution and the respective patient serum
diluted at 1 : 500 (NCLs), 1 : 1000 (TID) or 1 : 5000 (SPS)
in TBST/10% normal goat serum. Subsequently sections
were rinsed in TBS and incubated for 4 h in a mixture of
secondary antisera [Alexa Fluor 568 rat-adsorbed Goat
anti-mouse IgG (A11030) 1 : 500 with Alexa Fluor 488
Goat anti-human IgG (A11013) 1 : 1000, or Alexa Fluor
488 rat-adsorbed Goat anti-mouse IgG (A11001) 1 : 500
with Alexa Fluor 568 goat anti-human IgG (A21090)
1 : 1000, Molecular Probes, Eugene, OR, USA]. After rinsing in TBS, sections were mounted and coverslipped with
Vectashield hard-set mounting medium (H 1400, Vector
Laboratories). Immunofluorescence was observed by conventional epifluorescence microscopy (Zeiss Axioskop 2
MOT, Carl Zeiss Ltd, Welwyn Garden City, UK) and
recorded digitally (Zeiss Axiocam, Axiovision 3.0, Carl
Zeiss Ltd).
To determine the extent of colocalization of JNCL sera
with a range of other CNS antigens, we performed similar colocalization studies using patient-derived sera
together with rabbit anti-glutamate (Sigma, 1 : 2000);
rabbit anti-parvalbumin (Swant, Bellinzona, Switzerland 1 : 5000); rabbit anti-calbindin (Swant, 1 : 20000);
rabbit anti-somatostatin (Peninsula Laboratories,
Belmont, CA, USA 1 : 1000) and subsequently with the
appropriate secondary antisera labelled with either
Alexa Fluor 488 (1 : 500) or Alexa Fluor 568 (1 : 1000)
(Molecular Probes). In all colocalization studies, the distribution of antigens was confirmed by switching the fluorochromes used to label the secondary antisera. Similar
results were also obtained with sequential incubation of
primary antisera, rather than as a combined solution
(data not shown).
Preadsorption studies
To assess the relative contribution of anti-GAD antibodies
to immunoreactivity of JNCL sera, we stained sections
with sera that had been preadsorbed with recombinant
GAD protein, either full length or lacking the N-terminus
(Kronus Inc., Boise, ID, USA). Preadsorption was performed by incubating 10 µl of patient sera in the presence
or absence of 3.3 ng recombinant GAD protein overnight
at 4°C with constant gentle agitation. The degree of preadsorption was qualitatively assessed by Western blotting as
described previously [4].
Analysis and review
All sections were reviewed independently by two workers who were blinded and unaware of genotype and
diagnosis. For each sample the distribution of serum
immunoreactivity was detailed and samples that shared
similar patterns of immunoreactivity were grouped
together by consensus between the two observers. These
patterns were categorized as staining Types 1–4, as
described below. For more comprehensive mapping of
JNCL serum immunoreactivity, stained series of coronal
and sagittal sections through the CNS were surveyed in
detail. The distribution of immunoreactive structures
was noted with reference to landmarks in a rat brain
atlas [19].
Results
JNCL patient-derived sera display CNS tissue
immunoreactivity
In an initial pilot study, serum from two individuals genotyped to bear the 1.02-kb major deletion in CLN3 [9], and
neurologically normal controls (n = 2) were used as primary antisera to stain rat CNS sections. The JNCL sera
produced a characteristic pattern of cytoplasmic immunoreactivity within a variety of restricted neuronal subpopulations (defined as Type 1 staining), that was absent
in adjacent series of sections stained with control sera
(Figure 1A). The immunoreactivity of JNCL sera was distributed evenly throughout the cell surface and cytoplasm
with no obvious vesicular or perinuclear concentration
and no distinct staining of axonal processes or white matter tracts. Significantly, this JNCL serum immunoreactivity was not confined to the cell soma, but extended along
the dendrites of multiple cell types. This immunoreactivity provided particularly distinct staining of interneuron
populations in the hippocampal formation, cortical
© 2006 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 32, 469–482
JNCL serum immunoreactivity
473
Figure 1. JNCL serum immunoreactivity in rat CNS sections. ( A) Overview of rat CNS sections stained immunohistochemically using serum
from a neurological normal control (Control) or an individual with JNCL as a primary antiserum. JNCL serum displayed widespread
immunoreactivity within multiple CNS structures in both sagittal and coronal sections. In contrast, control serum displayed little or no
immunoreactivity above background. Scale bar = 2 mm. (B) Examples of JNCL serum immunoreactivity in sagittal sections through the
neocortex, hippocampus and cerebellum. Intensely immunoreactive neurons and apical dendrites were present in lamina V at the border of
primary motor (M1) and somatosensory barrel field cortex (S1BF). Within the hippocampus, dense JNCL serum immunoreactivity labelled
neurons in the hilus (Hi) and interneurons in the stratum oriens (SO) adjacent to layer CA1. Within the cerebellum JNCL serum stained Purkinje
cells (Pk) and their dendrites extending into the molecular layer. Scale bar = 50 µm.
lamina V pyramidal neurons, and cerebellar Golgi and
Purkinje neurons, in addition to a variety of subcortical
nuclei (Figure 1). This staining was clearly visible at dilutions up to 1:2000 (data not shown), and was confirmed
by multiple staining runs. Using sera from the same
patients, we also tested the sera immunoreactivity on
human temporal cortex sections from neurologically normal controls, and obtained a comparable staining of the
soma and dendrites of subgroups of neurons that was not
detected using control sera (Figure 2).
Detailed distribution of Type 1 JNCL serum
immunoreactivity
A systematic survey of a one-in-six series of coronal and
sagittal sections revealed the full extent of JNCL serum
immunoreactivity, staining neuronal subpopulations
throughout the CNS (Figure 1). Within the cortex, lamina
V pyramidal neurons in many regions exhibited dense
immunoreactivity with stained apical dendrites extending
into more superficial laminae where scattered neurons in
laminae II and III were stained (Figure 1B). Immunoreac-
Figure 2. JNCL serum immunoreactivity in human CNS sections.
(A–B) Human temporal cortex stained immunohistochemically
using serum from an individual with JNCL (B) and a neurologically
normal control (A). JNCL serum displayed widespread
immunoreactivity with staining of the soma and dendrites of
subgroups of neurons comparable with staining on rat CNS sections.
In contrast, control serum displayed little or no immunoreactivity
above background. Scale bar = 100 µm.
tivity of this type was particularly evident in the primary
and secondary motor cortex, cingulate cortex and all subdivisions of the somatosensory cortex. Less intense immunoreactivity was present in pyramidal neurons in the
prefrontal, frontal, primary and secondary visual cortices,
© 2006 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 32, 469–482
474
M. J. Lim et al.
the medial entorhinal cortex and all divisions of the insular cortex. In the retrosplenial, granular and agranular
mitral cortices, there was a predominance of stained dendrites. Within the hippocampal formation, immunoreactive neurons with stained dendrites were particularly
prominent in the hilus, but scattered immunoreactive
neurons were also evident in the stratum oriens and in the
dentate gryus and layers CA1–3 with dendrites extending
into the stratum radiatum (Figure 1B).
Throughout the basal ganglia, there was a varying
degree of faintly immunoreactive neurons, with the exception of the lateral globus pallidus which exhibited a more
intense neuronal immunoreactivity with many highly
branched dendrites. Within the thalamus, the majority of
immunoreactivity labelled the dense afferent innervation
supplying sensory relay nuclei including lateral and ventral nuclei, especially the ventral part of the lateral geniculate nucleus. Immunoreactive neurons and terminals
were also present in the reticular thalamic nucleus and in
the anterior dorsal nuclei. Dense immunoreactivity of terminals and occasional neurons extended ventrally into
the zona incerta, subthalamic nuclei and the medial parvicellular and magnocellular and preoptic nuclei of the
hypothalamus. More caudally, densely immunoreactive
neurons were present in the deep layers of the inferior and
superior colliculi, with branched dendrites extending to
more superficial layers.
Intensely immunoreactive neurons were also present in
the red nucleus, the interstitial nucleus of Cajal, nucleus
of Darkschewitsch and oculomotor nucleus, with processes and many immunoreactive fibres running through
the reticular formation. More ventrally, staining was evident in the mammillary nuclei, and substantia nigra pars
reticularis. Within the pons and medulla there was
intense immunoreactivity in the central grey, mesencephalic trigeminal nucleus, cuneiform nucleus, with dense
terminal labelling in the pontine reticular nuclei. Immunoreactive neurons were revealed in pedunculopontine
and dorsal tegmental nuclei, parvicellular reticular
nucleus, medullary reticular nucleus, spinal trigeminal
nucleus, principal sensory trigeminal nucleus, facial
nucleus and all divisions of the vestibular nuclei and the
nucleus of the solitary tract. More caudally, intensely
immunoreactive neurons were also present in the accessory and hypoglossal nuclei and in motor neurons of the
ventral horn of cervical spinal cord, which displayed
many immunoreactive fibres in superficial laminae of the
dorsal horn.
Within the cerebellum particularly intense staining
of Purkinje cells and their dendrites was evident
(Figure 1B), but not equally in all lobules. Scattered
Golgi cell and basket cell immunoreactivity was also
present, but with no immunoreactivity of granule neurons. Neurons of the deep cerebellar nuclei were also
immunoreactive particularly within the medial fastigial
cerebellar nucleus.
JNCL patient-derived sera display distinctive
patterns of immunoreactivity
We next repeated these staining experiments using
extended series of JNCL serum samples (n = 59); other
forms of NCL (INCL, n = 2; LINCL, n = 3; vLINCL, n = 1);
and sera from neurologically normal individuals (n = 12).
These diagnoses were all confirmed by appropriate genotyping and for JNCL samples Western blot analysis to confirm the presence of GAD65 autoantibodies (data not
shown). Sera from patients with other neurological conditions with neurodegeneration (paraneoplastic syndrome, n = 2 and a range of other LSDs, n = 32) were also
used to act as neurological controls.
JNCL serum samples produced two distinctly different
categories of staining (Figure 3); either prominent staining of neuronal soma and dendrites (Type 1 staining, as
depicted in Figure 1) or a more diffuse cytoplasmic immunoreactivity (Type 4). Between these two extremes, there
was a range of staining appearances that exhibited markedly less pronounced staining of neuronal processes
(Types 2 and 3). These latter types of staining could only
be distinguished reliably by the presence (Type 3) or
absence (Type 2) of immunoreactive cells with glial morphology in multiple brain regions, but prominent within
the stratum radiatum of the hippocampal formation
(Figure 3H).
When used as primary antisera, 25% (15 out of 59)
of the JNCL samples precisely reproduced the pattern
of immunoreactivity described in Figure 1 (Type 1,
Figure 3A,F,K). A further 25% stained with many similar characteristics, but with markedly less prominent
staining of dendrites and additional diffuse cytoplasmic
staining of many cell types and increased background
staining of the neuropil (Type 2, Figure 3B,G,L). Taken
together they accounted for more than half of the JNCL
samples tested. A further reduction in dendritic immunoreactivity was apparent in nine of the JNCL samples
© 2006 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 32, 469–482
JNCL serum immunoreactivity
475
Figure 3. Multiple types of JNCL serum immunoreactivity. (A–O) An extended series of JNCL sera displayed immunoreactivity that could be
classified into four subtypes. Sagittal sections through the neocortex (A–E, Ctx), hippocampus (F–J, Hipp) and cerebellum (K–O, Cb) revealed
prominent staining of soma and dendrites as depicted in Figure 1 (Type 1 staining). Type 2 and Type 3 immunoreactivity exhibited successively
less prominent dendritic staining, with additional glial immunoreactivity in the hippocampus of Type 3 staining (inset in H). Type 4 staining
comprised diffuse cytoplasmic immunoreactivity with no dendritic staining. Serum from neurologically normal controls displayed minimal
immunoreactivity above background staining of the neuropil (E, J, O). Scale bar = 50 µm.
(16% of cases), which showed an additional staining of
glial cell populations that was most pronounced in the
stratum radiatum of the hippocampus and the cerebellar cortex (Type 3, Figure 3C,H,M). A further 25% of
cases JNCL cases exhibited a diffuse cytoplasmic stain of
all cell soma, but with no consistent nuclear component
(Type 4, Figure 3D,I,N). The five remaining JNCL cases
could not be fitted into this classification scheme providing inconclusive staining patterns. The staining of JNCL
sera was in marked contrast to control patient sera
that demonstrated little or no immunoreactivity
(Figure 2E,J,O). However, in an extended series (n = 12)
of apparently neurologically normal subjects, a small
subset of individuals produced widespread diffuse cytoplasmic staining resembling Type 4 staining. Sera from
patients with other LSDs and paraneoplastic neurodegeneration, used as neurologically abnormal controls
(n = 32), confirmed the specificity of the above staining
patterns to JNCL sera. Sera from patients with paraneoplastic syndromes demonstrated a pattern of cytoplasmic and nuclear staining that has previously been
reported [20]. Interestingly, LSD patient sera display
their own characteristic staining patterns showing a
variety of intracellular distributions of immunoreactivity with specificity for particular CNS regions and cell
types (data not shown).
The requirement for anonymity of patient serum samples restricted the availability of clinical data to gender and
age. As such, we used patient age as a surrogate marker
for disease severity, but did not find a correlation between
age or gender and the pattern of immunoreactivity. In a
small number of JNCL patients where samples from more
than one time point were available (n = 10), the pattern of
immunoreactivity changed with disease progression with
a tendency to move towards a Type 1 pattern.
NCL patient-derived sera also display CNS tissue
immunoreactivity
Cases from individuals with other forms of NCL produced
patterns of immunoreactivity that either shared the fea-
© 2006 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 32, 469–482
476
M. J. Lim et al.
tures of Type 2 staining (one INCL case, CLN6 case), or
showed little or no distinct immunoreactivity and resembled Type 4 staining (one INCL case, two LINCL cases).
These data suggest the presence of autoreactive antibodies
in these other forms of NCL.
JNCL serum immunoreactivity colocalizes to
multiple antigens
On the basis of distribution and morphology, JNCL
serum immunoreactivity (Type 1) appeared to involve
both inhibitory and excitatory neurons (Figure 1). This
staining pattern could be due to the presence of a single
autoantigen that is expressed in multiple cell types, or
more likely due to the existence of multiple CNS-directed
autoantibodies. To investigate this issue, we undertook
colocalization studies using dual channel indirect immunofluorescence to identify neuronal populations that stain
with JNCL sera. As expected, these studies revealed examples of JNCL serum staining within GAD-positive neurons
(Figure 4A–C) or neuronal populations that stain with
other markers of GABAergic phenotype, such as the
calcium-binding proteins parvalbumin (PV) or calbindin
(CB) (Figure 4D–F) or the neuropeptide somatostatin
(SOM). Within the neocortex, JNCL serum immunoreactivity was clearly also present within populations of
lamina V pyramidal neurons labelled with an antibody
to glutamate (Figure 4G–I). However, many JNCL serum
immunoreactive neurons were not glutamate-, GAD-,
CB-, PV- or SOM-positive, suggesting that JNCL sera recognize only a subset of neurons immunoreactive for each
of these phenotypic markers.
JNCL patient immunoglobulins that stain CNS
sections are IgG subtype specific
To identify the subtype of immunoglobulins in JNCL sera
that bind to tissue sections, we used secondary antibodies that distinguish between heavy and light chains of
IgG (Figure 5) and that specifically recognize IgA, IgE
and IgM immunoglobulins (Figure 6). Sera from individuals with JNCL displayed IgG subtype specificity, with
secondaries recognizing the IgG heavy chain (IgGγ) and
IgG heavy and light chain (IgGH+L) components showing
maximal immunoreactivity (Figure 5). Staining with
IgGγ-specific secondary antisera retained the intensity of
Figure 4. JNCL serum immunoreactivity labels multiple neuronal
populations. (A–I) Dual channel immunofluoresence reveals the
coexpression (yellow) of JNCL serum immunoreactivity (green
channel), together with multiple markers of neuronal phenotype (red
channel). (A–C) Hippocampal interneurons in CA1, with fine
dendritic processes extending into the stratum radiatum, were
stained with JNCL serum (A) and were also immunoreactive for
glutamate decarboxylase (GAD) (B). (D–F) Subpopulations of
calbindin (CB)-positive interneurons in the stratum oriens (E) were
also labelled by JNCL serum (D). (G–I) JNCL serum also stained a
subpopulation of glutamate (Glu)-immunoreactive pyramidal
neurons in the neocortex. Scale bar A–C = 50 µm; D–F = 25 µm;
G–I = 50 µm.
Type 1 JNCL immunoreactivity, but markedly reduced
the amount of background immunoreactivity in the
neuropil (Figure 5).
Secondary antibodies recognizing the λ chain also
detected JNCL serum immunoreactivity (Figure 5), but
with markedly less intense staining of a subset of neurons stained by IgGγ and IgGH+L secondaries. There was
minimal immunoreactivity detectable above background with secondary antisera recognizing κ chains
(Figure 5), or specific for IgM, IgA and IgE (Figure 6).
Similar results were obtained when using these subtypespecific antisera to detect the binding of a representative
series of sera from patients with other forms of NCL (data
not shown).
© 2006 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 32, 469–482
JNCL serum immunoreactivity
477
Figure 5. JNCL serum immunoreactivity demonstrates specificity for IgG subclasses. (A–H) Secondary antisera that distinguish between heavy
and light chains of IgG reveal the subtypes of immunoglobulins that contribute to JNCL serum immunoreactivity in sagittal sections through
the neocortex (Ctx) and cerebellum (Cb). Secondaries recognizing IgG heavy and light chains (IgG H+L) revealed intense staining of neuronal
soma and dendrites (A, E). This pattern was retained using secondaries recognizing the IgG heavy chain (IgGγ), but with decreased background
staining of the neuropil (B, F). Secondary antibodies recognizing the λ chain provided markedly less intense immunoreactivity (C, G),
particularly in the cortex (C). In contrast, JNCL serum immunoreactivity was virtually absent using secondary antisera recognizing κ chains
(D, H). Scale bar = 50 µm.
GAD65 autoantibodies do not contribute
significantly to JNCL patient serum
immunoreactivity
As JNCL patients have circulating autoantibodies to
GAD65 [4,5], we compared JNCL serum immunoreactivity with two other groups of patients where GAD65 antibodies are present, SPS and TID. Both SPS (n = 7) and
TID (n = 6) sera demonstrated a variety of immunoreactive profiles that were clearly different from JNCL serum
immunoreactivity (Figure 7 presents data from the striatum and cerebellum only). Among SPS cases, three out
of seven samples stained with a pattern of immunoreactivity previously reported for this condition (Figure 7C)
[14], and comparable with the immunoreactivity of a
commercially available mouse monoclonal antibody
against GAD (Figure 7D). The remaining SPS and TID
cases stained with novel patterns of immunoreactivity
that were unlike JNCL serum staining, and which are
now being characterized in an expanded series of SPS
and TID cases. These data suggest that although individuals with JNCL, SPS and TID each raise autoantibodies
to GAD65, these three conditions display distinct pat-
terns of serum immunoreactivity, thus questioning how
much of the JNCL immunoreactivity is related to GAD65
autoreactivity.
To evaluate this issue further, we preadsorbed JNCL sera
with recombinant GAD65 protein, hypothesizing that
remaining immunoreactivity would be the result of binding to antigens other than GAD. As expected, unabsorbed
JNCL serum samples reproduced Type 1 staining with
prominently immunoreactive dendrites in the cortex
(Figure 8A), hippocampus (Figure 8B) and cerebellum
(Figure 8C). The same JNCL sample that had been preadsorbed either with full-length recombinant GAD or with
GAD lacking the N-terminal displayed markedly reduced
staining intensity, but with no obvious change in distribution (Figure 8D–F). Although background immunoreactivity was also decreased by preadsorption, the same
pattern of immunoreactivity against neuronal and dendritic targets persisted in the cortex (Figure 8D), hippocampus (Figure 8E) and cerebellum (Figure 8F). Taken
together with the results of SPS and TID serum staining,
these data point towards the existence of other CNSdirected autoantibodies that contribute significantly to
the immunoreactivity of JNCL serum.
© 2006 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 32, 469–482
478
M. J. Lim et al.
Figure 6. IgM, IgA and IgE binding do not significantly contribute
to JNCL serum immunoreactivity. (A–F) Secondary antisera that
recognize IgM, IgA or IgE do not detect prominent JNCL serum
immunoreactivity in sagittal sections through the neocortex (Ctx)
and cerebellum (Cb). IgM-specific antisera reveal only minimal JNCL
serum immunoreactivity (A, D), whereas IgA- (B, E) and IgE- (C, F)
specific secondary antisera displayed no immunoreactivity above
background staining. Scale bar = 50 µm.
Figure 7. JNCL serum immunoreactivity is distinct from other
disorders that raise GAD65 autoantibodies. JNCL serum (B, F)
exhibits immunoreactivity that is unlike staining with sera from
individuals with Stiff Person Syndrome (SPS) (C, G) or Type I diabetes
(TID) (A, E), in sagittal sections through the cerebellum (Cb) or
striatum (St). TID serum produced punctate staining of the neuropil
in the striatum (E) and in the region of the cerebellar Purkinje cell
layer (A). In contrast, both SPS serum and a monoclonal antibody to
glutamate decarboxylase (mGAD) produced intense labelling of
GABAergic terminals in the striatum (G, H) and dense labelling
around the soma and initial segment of the dendritic tree of cerebellar
Purkinje cells (C, D). Scale bar A–D = 50 µm; E–H = 200 µm.
Discussion
This study represents the first demonstration that JNCL
serum binds to tissue sections, and provides a detailed
description of this IgG-mediated immunoreactivity. JNCL
serum stained both inhibitory and excitatory neuronal
populations and displayed immunoreactivity that was distinct from serum from other disorders that raise GAD65
autoantibodies. The pattern of JNCL serum immunoreactivity was not significantly altered by preadsorption with
recombinant GAD protein. Collectively, these data suggest
that GAD65 autoantibodies do not contribute significantly to JNCL serum immunoreactivity. Instead, GAD65
autoantibodies appear to be just one of multiple CNSdirected autoantibodies in JNCL serum, although the
identity and pathological significance of these immunoglobulins remain unclear.
JNCL serum immunoreactivity reveals multiple
CNS targets for autoantibodies
The existence of autoantibodies to antigens within the
CNS has been demonstrated in a number of neurological
conditions by exploring the binding of patient-derived
serum to tissue sections [13–17]. Using this methodology,
our data reveal the potential for widespread CNS
immunoreactivity of JNCL patient-derived sera, which is
IgG-mediated rather than via other subclasses of immu-
Figure 8. Preadsorption of JNCL sera with recombinant GAD does
not abolish its immunoreactivity. Unadsorbed JNCL serum displayed
typical immunoreactivity for pyramidal neurons in the cortex (Ctx)
(A), interneurons in the stratum oriens adjacent to layer CA1 of the
hippocampus (B), and in Purkinje cells of the cerebellum (C).
Preadsorption with recombinant GAD protein did not completely
abolish this immunoreactivity, but reduced its intensity in each of
these regions (D–F). Scale bar = 50 µm.
noglobulins (Figures 5 and 6). We have previously
described the presence of GAD65 autoantibodies in both
human and murine JNCL [4,5]. As such, it was to be
expected that JNCL serum would bind to subpopulations
of GABAergic neurons positive for GAD (Figure 3), or calcium-binding proteins that are expressed by GABAergic
interneurons [21]. However, JNCL serum immunoreactivity was only present in subsets of GABAergic
neurons expressing these markers. More significantly,
JNCL serum also clearly stained multiple populations of
© 2006 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 32, 469–482
JNCL serum immunoreactivity
non-GABAergic neurons, most notably within the neocortex (Figure 4). Taken together with our previous Western
blotting data [4], the immunostaining results presented
here provide further evidence for the presence of multiple
brain-directed autoantibodies in JNCL sera.
The identity of these other autoreactive CNS antigens
remains unknown. However, a direct comparison of JNCL
patient sera immunoreactivity with a mouse anti-GAD65
monoclonal antibody revealed dissimilar staining patterns (Figure 7). Moreover, the inability to abolish the
intensity of immunoreactivity upon adsorption of JNCL
sera with GAD65 (Figure 8) further suggests that GAD65
autoantibodies are just a small component of JNCL sera
immunoreactivity. This is not necessarily surprising as
radioimmunoassays do not consistently detect low levels
of GAD65 autoantibodies in JNCL serum samples [22]. In
contrast, SPS patient sera which display significantly
higher levels of circulating GAD65 autoantibodies measurable by immunoassay [23] demonstrate a pattern of
immunoreactivity similar to that of monoclonal antiGAD65 (Figure 7). The individual patterns of CNS immunoreactivity that we found in these three disorders that
raise antibodies against GAD65 (JNCL, SPS and TID) could
be explained by the differences in the autoantibody specificity to antigenic epitopes of GAD65 in these disorders
[22]. Alternatively, the pattern of JNCL serum immunoreactivity we have described is more likely to reflect the presence of multiple other brain-directed autoantigens in this
disorder.
Production of autoantibodies in JNCL
patient-derived sera
Our immunohistochemical data reinforce evidence for
multiple circulating autoantibodies in JNCL serum [4,5],
but do not reveal the source of these immunoglobulins.
These circulating autoantibodies in JNCL could be
natural autoantibodies (NAAs), may be a direct result
of tissue destruction, or reflect another unidentified
consequence of CLN3 mutation upon antigen presentation and self-recognition. NAAs are antibodies principally of the IgM subtype [24], exhibiting autoreactivity
in normal healthy subjects and are targeted against conserved epitopes [25]. As our data suggest the immunoreactive components in JNCL patient-derived sera are
predominantly IgGs, it seems unlikely that NAAs in
patient sera contribute significantly to the pattern of
JNCL serum immunoreactivity.
479
Autoantibodies found in various neurodegenerative
conditions [26–29] can be generated following tissue
destruction [1]. These autoantibodies may persist due to
inaccurate antigen presentation, or simply abnormal
lysosomal clearance. It remains unclear whether the
brain-directed autoantibodies are generated systemically or within the CNS, but lymphocyte infiltration into
the JNCL CNS is very limited, even at the advanced
stages of disease progression (data not shown). If produced within the CNS, the presence of autoantibodies
may represent a secondary response to the liberation
and vascular dissemination of antigens from dead or
dying neurons. However, these autoantibodies are
present in Cln3–/– mice as early as 1 week of age [4],
whereas these mice do not exhibit significant neurone
loss before 5 months of age [Pontikis, Pearce and Cooper 12, unpub. obs.].
Alternatively, lysosomal dysfunction could play a key
role in the generation or persistence of autoantibodies.
Furthermore, as CLN3 is expressed in the late endosome/
lysosome [10,30,31], a loss of CLN3 function may have an
impact upon antigen presentation. Such a process is
required for the generation of recognition of ‘self ’ antigens
to protect against an autoimmune response. During
dendritic cell maturation, the loading of MHC class 2
complexes is a pH-dependent process [32]. As altered
vacuolar/lysosomal pH has been demonstrated in both
yeast [33] and NCL patient-derived fibroblast cell lines
[34], the influence of pH dysregulation may be a further
contributory factor in the generation and persistence of
autoantibodies, probably as a result of dysfunctional antigen presentation. Indeed, the changes in JNCL serum
immunoreactivity we have described during disease progression may be indicative of an ongoing and ever worsening lysosomal defect.
Potential pathogenicity of autoantibodies
in JNCL
Circulating autoantibodies to GM2 gangliosides and IgG
deposition within the CNS have been reported in patients
and a mouse model of Sandhoff disease, another paediatric storage disorder, and may contribute to pathogenesis
[35]. The presence of multiple autoantibodies in the serum
and cerebrospinal fluid of Cln3–/– mice and patients (data
not shown), their ability to bind to the CNS, and the elevated levels of glutamate and inhibition of GAD activity
demonstrated in brain tissue of Cln3–/– mice [4], collec-
© 2006 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 32, 469–482
480
M. J. Lim et al.
tively raise similar questions for a pathogenic role of
autoantibodies in JNCL. It may be tempting to think that
the presence of GAD65 autoantibodies contributes to the
selective loss of GABAergic interneurons that occurs in
JNCL [7,12,36]. Nevertheless, GAD65 autoantibodies are
not detected in other forms of NCL [5], which also exhibit
significant loss of interneuron populations [7,36]. Furthermore, if the antigenic targets of autoantibodies are
intracellular, such as GAD, this may preclude their pathogenicity [37,38]. As such, it will be important to investigate the identity and potential pathogenic roles of the
other, as yet unidentified autoantibodies present in JNCL
sera.
Antibody-mediated CNS disorders are increasingly
being recognized including conditions where the autoantigen is not identified, for example, in post-streptococcal
CNS disease and neuropsychiatric presentations of systemic lupus erythematosus [1]. Rigorous evidence of
pathogenicity of these autoantibodies in JNCL will still
require the induction of a disease phenotype by passive
transfer of autoantibodies, disease induction with the
autoantigen, or a response to specific immunomodulatory
therapy.
Acknowledgements
NCLs and beyond
References
Preliminary data using sera from individuals with other
forms of NCL also suggest that the presence of autoantibodies is not restricted to JNCL and it will be important to
explore this possibility further in an expanded series of
cases. It would also be equally important to investigate if
these antibodies cross the blood brain barrier and bind to
the CNS in patients. As brain-directed antibodies are
present in NCL patient-derived sera, similar autoimmune
responses may happen in other disorders where lysosomal
dysfunction occurs. Because the lysosome plays an active
role in antigen presentation, a process which is apparently abnormal in mice that model mucopolysaccharidosis VII [39], the LSDs are an interesting group in which to
investigate such responses, as has been done in GM2 [35]
and Gaucher disease [40]. Defining the autoimmune
components of these disorders may lead to the development of immunomodulatory therapeutic strategies.
Nevertheless, it remains imperative that evidence of
pathogenicity of an autoimmune response exists before
clinical trials are initiated, as such treatments are not
without their risks.
We would like to thank the Batten Disease Support and
Research Association and families for providing serum
samples; Professor Pirkko Santavuori, Professor Angela
Vincent, Drs Marinos Dalakas, Mark Atkinson, David
Wenger and Anthony Fensom for kindly providing
patient-derived sera; Stephen Shemilt and Noreen Alexander for their expert advice and the other members of
the PSDL for their valuable contributions; Mavis Kibble
for her skilled technical assistance; Dr John Stephenson
for provision and perfusion of animals; and Drs JeanPierre Lin, Payam Rezaie and Alison Barnwell for constructive comments on the manuscript. These studies
were supported by the Wellchild/Royal College of Paediatric and Child Health (RCPCH) Research Fellowship
(ML), National Institutes of Health (NIH), USA Grants
NS41930 (JDC), NS40580 and NS44310 (DAP), The
Luke and Rachel Batten Foundation (DAP), and grants
to JDC from The Natalie Fund, Batten Disease Support
and Research Association, Batten Disease Family Association, Batten Disease Support and Research Trust and
The Remy Fund.
1 Lang B, Dale RC, Vincent A. New autoantibody mediated
disorders of the central nervous system. Curr Opin Neurol
2003; 16: 351–7
2 Antel JP, Owens T. Immune regulation and CNS autoimmune disease. J Neuroimmunol 1999; 100: 181–9
3 Archelos JJ, Hartung HP. Pathogenetic role of autoantibodies in neurological diseases. Trends Neurosci 2000; 23:
317–27
4 Chattopadhyay S, Ito M, Cooper JD, Brooks AI, Curran
TM, Powers JM, Pearce DA. An autoantibody inhibitory to
glutamic acid decarboxylase in the neurodegenerative
disorder Batten disease. Hum Mol Genet 2002; 11: 1421–
31
5 Chattopadhyay S, Kriscenski-Perry E, Wenger DA, Pearce
DA. An autoantibody to GAD65 in sera of patients with
juvenile neuronal ceroid lipofuscinoses. Neurology 2002;
59: 1816–17
6 Hofmann SL, Peltonen L. The neuronal ceroid lipofuscinoses. In The Metabolic and Molecular Basis of Inherited
Disease, edition8th edn. Eds CR Scriver, AL Beaudet,
WS Sly, D Valle. New York: McGraw-Hill, 2001; 3877–
94
7 Cooper JD. Progress towards understanding the neurobiology of Batten disease or neuronal ceroid lipofuscinosis.
Curr Opin Neurol 2003; 16: 121–8
© 2006 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 32, 469–482
JNCL serum immunoreactivity
8 Goebel HH, Wisniewski KE. Current state of clinical and
morphological features in human NCL. Brain Pathol
2004; 14: 61–9
9 The International Batten Disease Consortium. Isolation
of a novel gene underlying Batten disease, CLN3. The
International Batten Disease Consortium. Cell 1995; 82:
949–57
10 Phillips SN, Benedict JW, Weimer JM, Pearce DA. CLN3,
the protein associated with Batten disease: structure,
function and localization. J Neurosci Res 2005; 79: 573–
83
11 Mitchison HM, Bernard DJ, Greene ND, Cooper JD, Junaid
MA, Pullarkat RK, Vos N, Breuning MH, Owens JW, Mobley WC, Gardiner RM, Lake BD, Taschner PEM, Nussbaum RL. Targeted disruption of the Cln3 gene provides a
mouse model for Batten disease. The Batten Mouse Model
Consortium. Neurobiol Dis 1999; 6: 321–34
12 Pontikis CC, Cella CV, Parihar N, Lim MJ, Chakrabarti S,
Mitchison HM, Mobley WC, Rezaie P, Pearce DA, Cooper
JD. Late onset neurodegeneration in the Cln3-/- mouse
model of juvenile neuronal ceroid lipofuscinosis is preceded by low level glial activation. Brain Res 2004; 1023:
231–42
13 Bataller L, Dalmau JO. Paraneoplastic disorders of the
central nervous system: update on diagnostic criteria and
treatment. Semin Neurol 2004; 24: 461–71
14 Solimena M, Folli F, Denis-Donini S, Comi GC, Pozza G, De
Camilli P, Vicari AM. Autoantibodies to glutamic acid
decarboxylase in a patient with stiff-man syndrome, epilepsy, and type I diabetes mellitus. N Engl J Medical 1988;
318: 1012–20
15 Klushnik TP, Gratchev VV, Belichenko PV. Braindirected autoantibodies levels in the serum of Rett syndrome patients. Brain Dev 2001; 23 (Suppl. 1): S113–
S117
16 Connolly AM, Chez MG, Pestronk A, Arnold ST, Mehta S,
Deuel RK. Serum autoantibodies to brain in LandauKleffner variant, autism, and other neurologic disorders.
J Pediatr 1999; 134: 607–13
17 Bataller L, Wade DF, Fuller GN, Rosenfeld MR, Dalmau
J. Cerebellar degeneration and autoimmunity to zincfinger proteins of the cerebellum. Neurology 2002; 59:
1985–7
18 Tyynelä J, Cooper JD, Khan MN, Shemilt SJ, Haltia M.
Hippocampal pathology in the human neuronal ceroidlipofuscinoses: distinct patterns of storage deposition,
neurodegeneration and glial activation. Brain Pathol
2004; 14: 349–57
19 Paxinos G, Watson C. The Rat Brain in Stereotaxic Coordinates, 4th edn. San Diego: Academic Press, 1997
20 Nagashima T, Mizutani Y, Kawahara H, Maguchi S, Terayama Y, Shinohara T, Orba Y, Chuma T, Mano Y, Itoh T,
Sawa H, Sakai K, Motomura M, Nagashima K. Anti-Hu
paraneoplastic syndrome presenting with brainstemcerebellar symptoms and Lambert-Eaton myasthenic
syndrome. Neuropathology 2003; 23: 230–8
481
21 Freund TF, Buzsaki G. Interneurons of the hippocampus.
Hippocampus 1996; 6: 347–470
22 Ramirez-Montealegre D, Chattopadhyay S, Curran TM,
Wasserfall C, Pritchard L, Schatz D, Petitto J, Hopkins D,
She JX, Rothberg PG, Atkinson M, Pearce DA. Autoimmunity to glutamic acid decarboxylase in the neurodegenerative disorder Batten disease. Neurology 2005; 64:
743–5
23 Lernmark A. Glutamic acid decarboxylase – gene to antigen to disease. J Intern Med 1996; 240: 259–77
24 Guilbert B, Dighiero G, Avrameas S. Naturally occurring
antibodies against nine common antigens in human sera.
I. Detection, isolation and characterization. J Immunol
1982; 128: 2779–87
25 Lacroix-Desmazes S, Kaveri SV, Mouthon L, Ayouba A,
Malanchere E, Coutinho A, Kazatchkine MD. Self-reactive
antibodies (natural autoantibodies) in healthy individuals. J Immunol Meth 1998; 216: 117–37
26 Terryberry JW, Thor G, Peter JB. Autoantibodies in neurodegenerative diseases: antigen-specific frequencies and
intrathecal analysis. Neurobiol Aging 1998; 19: 205–16
27 D’Andrea MR. Evidence linking neuronal cell death to
autoimmunity in Alzheimer’s disease. Brain Res 2003;
982: 19–30
28 Jernigan M, Morcos Y, Lee SM, Dohan FC Jr, Raine C,
Levin MC. IgG in brain correlates with clinicopathological
damage in HTLV-1 associated neurologic disease. Neurology 2003; 60: 1320–7
29 Bushara KO, Nance M, Gomez CM. Antigliadin antibodies
in Huntington’s disease. Neurology 2004; 62: 132–3
30 Jarvela I, Lehtovirta M, Tikkanen R, Kyttala A, Jalanko A.
Defective intracellular transport of CLN3 is the molecular
basis of Batten disease (JNCL). Hum Mol Genet 1999; 8:
1091–8
31 Haskell RE, Carr CJ, Pearce DA, Bennett MJ, Davidson BL.
Batten disease: evaluation of CLN3 mutations on protein
localization and function. Hum Mol Genet 2000; 9: 735–
44
32 Trombetta ES, Ebersold M, Garrett W, Pypaert M, Mellman I. Activation of lysosomal function during dendritic
cell maturation. Science 2003; 299: 1400–3
33 Pearce DA, Ferea T, Nosel SA, Das B, Sherman F. Action
of BTN1, the yeast orthologue of the gene mutated in
Batten disease. Nat Genet 1999; 22: 55–8
34 Holopainen JM, Saarikoski J, Kinnunen PK, Jarvela I.
Elevated lysosomal pH in neuronal ceroid lipofuscinoses
(NCLs). Eur J Biochem 2001; 268: 5851–6
35 Yamaguchi A, Katsuyama K, Nagahama K, Takai T, Aoki
I, Yamanaka S. Possible role of autoantibodies in the
pathophysiology of GM2 gangliosidoses. J Clin Invest
2004; 113: 200–8
36 Mitchison HM, Lim MJ, Cooper JD. Selectivity and types of
cell death in the Neuronal Ceroid Lipofuscinoses (NCLs).
Brain Pathol 2004; 14: 86–96
37 Archelos JJ, Trotter J, Previtali S, Weissbrich B, Toyka KV,
Hartung HP. Isolation and characterization of an oligo-
© 2006 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 32, 469–482
482
M. J. Lim et al.
dendrocyte precursor-derived B-cell epitope in multiple
sclerosis. Ann Neurol 1998; 43: 15–24
38 Dalmau J, Gultekin HS, Posner JB. Paraneoplastic neurologic syndromes: pathogenesis and physiopathology.
Brain Pathol 1999; 9: 275–84
39 Daly TM, Lorenz RG, Sands MS. Abnormal immune function in vivo in a murine model of lysosomal storage disease. Pediatr Res 2000; 47: 757–62
40 Shoenfeld Y, Beresovski A, Zharhary D, Tomer Y, Swissa
M, Sela E, Zimran A, Zevin S, Gilburd B, Blank M. Natural
autoantibodies in sera of patients with Gaucher’s disease.
J Clin Immunol 1995; 15: 363–72
Received 14 September 2005
Accepted after revision 24 January 2006
© 2006 Blackwell Publishing Ltd, Neuropathology and Applied Neurobiology, 32, 469–482