Download Spatial and temporal correlation between neuron neuronopathic Gaucher disease

Human Molecular Genetics, 2011, Vol. 20, No. 7
doi:10.1093/hmg/ddr019
Advance Access published on January 20, 2011
1375–1386
Spatial and temporal correlation between neuron
loss and neuroinflammation in a mouse model of
neuronopathic Gaucher disease
Tamar Farfel-Becker 1, Einat B. Vitner 1, Sarah N.R. Pressey 3, Raya Eilam 2, Jonathan D. Cooper 3
and Anthony H. Futerman 1,∗
1
Received December 29, 2010; Revised and Accepted January 12, 2011
Gaucher disease (GD), the most common lysosomal storage disorder, is caused by a deficiency in the lysosomal enzyme glucocerebrosidase (GlcCerase), which results in intracellular accumulation of glucosylceramide (GlcCer). The rare neuronopathic forms of GD are characterized by profound neurological impairment
and neuronal cell death, but little is known about the neuropathological changes that underlie these
events. We now systematically examine the onset and progression of various neuropathological changes
(including microglial activation, astrogliosis and neuron loss) in a mouse model of neuronopathic GD, and
document the brain areas that are first affected, which may reflect vulnerability of these areas to
GlcCerase deficiency. We also identify neuropathological changes in several brain areas and pathways,
such as the substantia nigra reticulata, reticulotegmental nucleus of the pons, cochlear nucleus and the
somatosensory system, which could be responsible for some of the neurological manifestations of the
human disease. In addition, we establish that microglial activation and astrogliosis are spatially and temporally correlated with selective neuron loss.
INTRODUCTION
Gaucher disease (GD), the most common lysosomal storage
disorder (LSD), is caused by mutations in the gene encoding
the lysosomal hydrolase, glucocerebrosidase (GlcCerase),
which results in accumulation of glucosylceramide (GlcCer).
Patients with GD are usually classified into three types,
based on the presence or absence of neurological manifestations and their rate of progression. Types 2 and 3, which
comprise only a small percent of GD patients, are collectively
referred to as neuronopathic GD (nGD), with both displaying
central nervous system (CNS) involvement in addition to systemic disease.
Patients with type 2 GD typically present either prenatally
or during infancy and usually die before 3 years of age (1).
Type 2 patients fail to thrive, and display severe and rapidly
progressive brainstem degeneration (2). The most frequent
initial clinical signs are hyperextension of the neck, swallowing impairment and strabismus (2). The most common cause
of death is prolonged spontaneous apnea which occurs with
increased frequency in the later stages of the disease (2,3).
Type 3 patients present with similar signs but with a later
onset and decreased severity, and these patients usually
survive until adolescence or adulthood (3).
Eye movement abnormalities are common in nGD (4) and
their detection is diagnostic of this disorder. In type 3 nGD,
oculomotor signs may precede the appearance of overt
neurological signs by many years. The main sign is
severe difficulty (in type 3) (5), or a total inability (in
type 2) (6) to generate saccades (ocular motor apraxia).
Auditory brainstem response (ABR) abnormalities are also
an early neurological sign in nGD. These findings may be
isolated, or appear together with developmental delay and
seizures (3,5).
∗
To whom correspondence should be addressed. Tel: +972 89342704; Fax: +972 89344112; Email: [email protected]
# The Author 2011. Published by Oxford University Press. All rights reserved.
For Permissions, please email: [email protected]
Downloaded from hmg.oxfordjournals.org at King's College London on March 7, 2011
Department of Biological Chemistry and 2Department of Veterinary Resources, Weizmann Institute of Science,
Rehovot 76100, Israel and 3Pediatric Storage Disorders Laboratory, Department of Neuroscience, Centre for the
Cellular Basis of Behavior, MRC Centre for Neurodegeneration Research, Institute of Psychiatry, King’s College
London, London, UK
1376
Human Molecular Genetics, 2011, Vol. 20, No. 7
RESULTS
Microglial activation and astrogliosis are spatially and
temporally correlated
To determine the time course of neuropathological changes in
nGD, we used a mouse model in which GlcCerase deficiency
is restricted to neurons and macroglia, with normal GlcCerase
activity in microglia (the Gbaflox/flox; Nestin-Cre mouse, hereafter referred to as the 2/2 mouse) (8). These mice exhibit
rapid motor dysfunction including rigidity of limbs and abnormal gait, leading to seizures and paralysis by 21 days of age, at
which time mice exhibit massive microglial activation, astrocytosis and neuron loss (8). To accurately define disease onset
and progression, we monitored mouse body weight over a
21-day period. There was no difference between 2/2 and
+/2 mice between 10 and 16 days (Fig. 1); however, at 18
days, 2/2 mice started to loose weight and by 21 days of
age, 2/2 mice weighed 73% of +/2 mice. A similar
pattern was obtained for brain weight, with no difference at
12 days, and a significant reduction of brain weight at 21
days (378 + 12 mg for +/2 mouse brains and 324 + 9 mg
for 2/2 brains). We thus classified the nGD mice as presymptomatic at 12 days, early symptomatic at 16 days and
late symptomatic at 21 days, the age at which they were euthanized.
Neuroinflammation, including microglial activation and
astrocytosis, is a common feature of neurodegenerative diseases (9), which may serve as a marker of ongoing neuronal
damage or dysfunction (10). To determine the age and the
brain areas where neuropathological changes first occur, we
stained sections throughout the rostrocaudal extent of brains
with the microglial marker, Mac2 (11). While +/2 brains
showed no immunoreactivity for Mac2, faint but detectable
Mac2 labeling was observed as early as 12 days in 2/2
mice (i.e. at the pre-symptomatic stage) (Fig. 2A). By 16
days, intense and localized microglial activation was evident
Figure 1. Progressive reduction in body weight in a mouse model of nGD.
∗
P , 0.05, ∗∗ P , 0.01, ∗∗∗ P , 0.001, n ¼ 13–15 + SEM.
in many brain areas, spreading to additional brain areas by
21 days (Fig. 2A). A similar pattern was observed for the
astrocyte marker glial fibrillary acidic protein (GFAP)
(Fig. 2B), with astrocytosis apparent at the same age and in
the same brain areas as microglial activation (Fig. 2A).
Indeed, astrocytosis was invariably observed with a similar
pattern to microgliosis, but since the latter gave a more localized and distinct labeling pattern, most of the data below
document microgliosis. Although the identity and sequence
in which areas displayed microglial activation and astrocytosis
was maintained between 2/2 mice, there was considerable
variation in the time of onset and rate of progression of
these reactive phenotypes. Nevertheless, the most striking
microglial activation and astrocytosis were seen in the thalamocortical area, which prompted us to investigate the relationship between glial activation and neuron loss in these brain
regions.
Progressive glial activation and neuron loss in the nGD
cortex
Within the cortical mantle, astrocytosis and glial activation
displayed both regional and laminar specificity. As previously
reported (7,8), layer V of the cortex showed the highest degree
of microglial activation. Occasional Mac2 immunoreactive
microglia were present in this layer of the somatosensory
and motor cortex as early as 12 days (data not shown). By
16 days, many more intensely stained Mac2-positive microglia
were evident, but were confined specifically to layer V
(Figs 3A and 4A), and were more frequent in motor and somatosensory areas than more caudal cortical regions, such as the
visual cortex (Figs 2A and 3A). This distribution of Mac2
staining was closely paralleled by a band of GFAP staining
in layer V of the same cortical regions (Figs 2B and 3A).
By 21 days, Mac2 immunoreactivity had spread ventrally
into layer VI, as well as in more superficial cortical layers
(Fig. 4A). A similar subfield specificity was also revealed by
cortical thickness measurements, with motor and somatosensory subfields displaying greater atrophy than the visual
cortex (Fig. 3B). No cortical atrophy was detected in
16-day-old 2/2 mice (Fig. 3B). Unbiased optical fractionator
estimates of the number of Nissl-stained layer V neurons in
S1BF revealed that these reactive changes were accompanied
by progressive neuron loss. There was no difference in the
Downloaded from hmg.oxfordjournals.org at King's College London on March 7, 2011
Little is known about the neuropathological events that
cause these neurological abnormalities. A systematic neuropathological analysis of human GD brain (7) highlighted
specific patterns of astrogliosis and neuron loss, in addition
to non-specific gray and white matter gliosis. However,
since this study (7) was performed on autopsy material, it
was not possible to determine anything about the temporal
sequence of pathogenesis. To address this issue, we have systematically examined the onset and progression of neuropathological events (microglial activation, astrogliosis and neuron
loss) in a mouse nGD model (8). We identify several brain
areas and pathways that may be responsible for some of the
neurological manifestations in nGD, and show that microglial
activation and astrogliosis are spatially and temporally correlated with selective neuron loss. This study may pave the
way for the identification of disease markers for use in
animal studies, since investigation of the areas affected at
different stages of disease progression will allow determination of the efficacy of experimental therapies. In addition,
it will direct further investigations into the molecular mechanisms by which GlcCer accumulation leads to the death of
specific neuronal populations.
Human Molecular Genetics, 2011, Vol. 20, No. 7
1377
nucleus [ventral posteromedial/posterolateral (VPM/VPL)],
which relays somatosensory information to the regions of
the sensory cortex that were most severely affected. Microglial
activation in the thalamus progressed rapidly, particularly in
the VPM/VPL, such that it was densely stained by 16 days
and was the most affected brain area at this age (Fig. 5A).
Optical fractionator estimates of neuron number revealed
that the robust microglial activation in the VPM/VPL was
accompanied by massive and rapid neuron loss, with a 57%
reduction at 16 days and 78% at 21 days (Fig. 5B). Other thalamic nuclei also exhibited localized microglial activation from
16 days onwards (Table 1), including the auditory relay
nucleus [medial geniculate (MGN)] (Fig. 5A), whereas no
immunoreactive microglia were found in the visual relay
nucleus (dorsal lateral geniculate) up to 21 days.
Figure 2. Progressive glial activation in nGD mice. Low magnification of
coronal sections of 12-, 16- and 21-day-old nGD mice reveals onset and progression of immunohistochemical staining for markers of microglial activation
(Mac2, A) or astrocytosis (GFAP, B). Pronounced and localized microglial
activation and astrocytosis were evident in multiple brain regions, occurring
at the same stage of disease progression and becoming more intense with
increased age. In contrast, no staining was observed in control brains (B).
Schematic diagrams indicate the rostrocaudal level of each section. Scale
bar ¼ 1 mm.
number of these neurons at 12 days, but there was a loss (30%
reduction) of layer V S1BF neurons in 2/2 mice at 16 days,
which became more pronounced (58% reduction) at 21 days
(Fig. 4B).
Progressive glial activation and neuron loss in the
thalamus
Late symptomatic 2/2 mice displayed localized and progressive microglial activation in many individual thalamic
nuclei (Figs 2A and 5A). Microglial activation began in the
reticular thalamic nucleus (Rt), which provides inhibitory
input to the rest of the thalamus, and in the ventral posterior
We next used Mac2 immunostaining to survey other areas of
the CNS. The basal ganglia, the nuclei globus pallidus and
the substantia nigra started showing faint Mac2 immunoreactivity from 12 days (Figs 5A and 6A), which became more
intense and widespread with increased age. Microglial activation was evident in both pars compacta [substantia nigra
compacta (SNC)] and reticularis [substantia nigra reticulata
(SNR)] of the substantia nigra of 2/2 mice, spreading from
the dorsolateral parts to the whole region between days 16
and 21 (Fig. 6A), the time at which neuronal loss was first
detected in SNR (Fig. 6B). In contrast, microglial activation
was far less prominent in the striatum, another major component of the basal ganglia, with only a few scattered Mac2stained microglia at 21 days in the dorsolateral part of the
caudate putamen (data not shown).
The SNR is involved in the initiation of saccadic eye movements (12,13), suggesting that the pathology seen in this area
maybe of clinical significance in nGD. Another key area of
the oculomotor circuits which regulate horizontal saccades is
the reticulotegmental nucleus of the pons (RtTg, nucleus Bechterew) (14), which contained many intensely stained Mac2positive microglia at 16 days (Fig. 7A). Other areas that may
be related to neurological manifestations in nGD are the
cochlear nucleus and the inferior colliculus, which may contribute to brainstem auditory abnormalities (15). Microglial activation was first detected in the cochlear nucleus of 2/2 mice
at 12 days (data not shown), with very little increase in activation
versus time (Fig. 7B). Microglial activation also occurred in the
inferior colliculus at 16 days (Fig. 7C), but was not pronounced
and did not increase with age.
Several cerebellar-related nuclei also displayed specific
microglial activation, among them the red nucleus, pontine
nucleus and external cuneate nucleus (Fig. 7D– F). While
the red nucleus displayed microglial activation only at
16 days (Fig. 7D), the pontine nucleus and external cuneate
nucleus displayed Mac2-immunoreactive microglia at
12 days (data not shown), which became more pronounced
with disease progression (Fig. 7E and F). Additional cerebellarrelated structures that displayed Mac2-positive microglia
were the lateral reticular nucleus (LRt) (Fig. 8), vestibular
nucleus and spinal cord (data not shown). Microglial activation in the LRt began at 12 days, became more pronounced
Downloaded from hmg.oxfordjournals.org at King's College London on March 7, 2011
Midbrain and brainstem pathology
1378
Human Molecular Genetics, 2011, Vol. 20, No. 7
with disease progression (Fig. 8) and was accompanied by
severe neuronal loss, such that this structure, readily detectable
in 12-day-old 2/2 mice, was barely detectable by Nissl staining at 16 days (data not shown) and was completely absent at
21 days (Fig. 8). The facial nucleus, an adjacent nucleus to the
LRt, was clearly detected by Nissl staining in late symptomatic 21-day-old 2/2 mice (Fig. 8) and no microglial activation was detected (data not shown), further emphasizing
the correlation between microglial activation and neuron loss
and the specificity of neuronal loss within the brainstem.
Another severely affected brainstem structure was the
sensory spinal trigeminal nucleus (SP5). Microglial activation
was detected as early as 12 days and rapidly progressed, with
numerous intensely stained Mac2-positive microglia evident at
21 days (Fig. 9A). The increase in microglial activation was
accompanied by abnormal development of the SP5 in 2/2
mice, which did not show the progressive increase in the
number of Nissl-stained SP5 neurons that was evident in
control mice (Fig. 9B). The changes within individual brainstem nuclei were not accompanied by an overall reduction
in brainstem volume (Fig. 9C, P ¼ 0.13).
Figure 4. Loss of cortical layer V S1BF neurons in nGD mice. (A) Double
immunofluorescence labeling for NeuN (red) and Mac2 (green) reveals the
loss of neurons in layer V, together with pronounced staining for Mac2 in
this layer, which was absent in age-matched controls. Cortical layers are indicated. Scale bar ¼ 100 mm. (B) Unbiased optical fractionator estimates of the
number of Nissl-stained neurons in layer V of the S1BF. ∗∗ P , 0.01, n ¼ 3–
4 + SEM.
Additional brain areas displaying microglial activation
A unique and distinct pattern of Mac2 immunoreactivity was
observed in the olfactory bulb at 12 days, which was highly
specific to the mitral cell layer and progressed to some
extent with age (Fig. 10A). Another area that showed early
Downloaded from hmg.oxfordjournals.org at King's College London on March 7, 2011
Figure 3. Cortical subfield specificity of glial activation and cortical atrophy in nGD mice. (A) Immunohistochemical staining for Mac2 and GFAP reveals
activated microglia and astrocytosis in the primary motor cortex (M1), somatosensory barrel field cortex (S1BF) and primary visual cortex (V1) of nGD
mice, with an earlier and more pronounced staining in M1 and S1BF versus V1. (B) Cortical thickness measurements reveal late onset thinning of the cortical
mantle to a larger extent within M1 and S1BF in late symptomatic animals. ∗∗ P , 0.01, ∗∗∗ P , 0.001, n ¼ 5 + SEM. Scale bar ¼ 100 mm.
Human Molecular Genetics, 2011, Vol. 20, No. 7
1379
and robust microglial activation was the caudomedial
entothinal cortex (CEnt) (Fig. 10B). The hippocampus
displayed a similar pattern of pathology as reported in
human patients (7) [and in a previous study on the nGD
mouse (8)], with microglial activation in subfields CA2 –
CA3 and in the hippocampal hilus (CA4) (Fig. 10C). Microglial activation in the hippocampus started at 16 days and
was most pronounced in the caudal hippocampus, progressing
to only a mild extent by 21 days at which time it had also
spread to CA1 (Fig. 10C). A summary of the progression of
microglial activation versus time in various brain areas is
given in Table 1.
colliculus (Fig. 11A) and the periaquaductal gray (PAG)
(Fig. 11A), displaying little or no Mac2 immunoreactivity.
Likewise there was specificity in which brainstem nuclei displayed microglial activation, with no activation even at the
end-stage of the disease in nuclei, such the facial nucleus
and the motor trigeminal nucleus. This lack of microglial activation suggested that there was no neuronal loss, which was
confirmed by optical fractionator analysis of the number of
PAG neurons (Fig. 11B). Together, our data strongly suggest
a direct correlation between the extent of microglial activation
and neuronal loss.
DISCUSSION
Several brain areas are resistant to GlcCerase deficiency
Even in late symptomatic 2/2 mice, microglial activation
remained localized, with many brain regions, such as the
hypothalamus [with the exception of the mammillary
nucleus, which showed some microglial activation at 16
days (data not shown)], superficial layers of the superior
In this study, we provide the first detailed systematic description of the temporal and spatial progression of neuropathological changes in a mouse model of nGD. We have identified a
number of brain areas that display pronounced microglial activation which were not previously known to be involved in
nGD pathology, among them regions that may be of
Downloaded from hmg.oxfordjournals.org at King's College London on March 7, 2011
Figure 5. Thalamic pathology in nGD. (A) Immunohistochemical labeling for Mac2 shows faint, but detectable staining (arrows) in the reticular thalamic
nucleus (Rt), globus pallidus (GP) and ventral posterior nucleus (VPM/VPL) in 12-day-old 2/2 mice which increased with age, while staining in the MGN
was only apparent from 16 days onwards. No staining was observed in control brains (data not shown). Scale bar ¼ 200 mm. (B) Unbiased optical fractionator
estimates of the number of Nissl stained neurons in VPM/VPL. ∗ P , 0.05, ∗∗ P , 0.01, n ¼ 3 –5 + SEM.
1380
Human Molecular Genetics, 2011, Vol. 20, No. 7
Table 1. Distribution of Mac2-immunoreactive microglia in brain regions
during disease progression
Brain area
++
++
++
2/+
2
+
+
2
2
2
2
+
+++
++
+++
+++
+
++
++
++
++++
++++
++++
++++
++++
++
+++
+++
+++
++++
2
2
2/+
+/++
+
2
++/+++
++
+/++
2
+++
++/+++
2
2
2
2
2
2
++
++
2
+++
++
+++
++/+++
2
++++
2
2
2
2
2
2
2
2
2
++/+++
2
2
2
2
+++
2
2
2
2/+
+
++/+++
+++
2
++/+++
++++
++++
++++
2
+++
++++
2
2
2
2
2
++/+++
2
++
++
2
+++
2
+++
+++
++
2
++/+++
+++
2
2
++
2
++
2
2
2
2
2
2
++
++
2
+++
2
++++
++++
+
+
2
2
2
+
+
2
++++
+/++
2
2
++
++/+++
+++
2
++++
++
2
2
++/+++
+++
++++
2
2
2
2
++++
2
++++
Continued
Brain area
Principal sensory trigeminal nucleus
(Pr5)
Sp5
Dorsal motor nucleus of vagus (10N)
Hypoglossal nucleus (12N)
Spinal cord
Intensity of Mac2 labelinga
12 days 16 days
21 days
2
++
+++
+
2
2
+
+++
2
2
++
++++
2
2
+++
a
The number of plus signs indicates the extent of Mac2 immunoreactivity,
with + indicating a few scattered Mac2-positive cells and ++++ indicating a
high density of Mac2-positive cells. A minus sign indicates no Mac2
immunoreactivity.
significance for understanding the clinical neurological manifestations observed in type 2 and 3 GD patients.
In the nestin-Cre Gbaflox/flox mice, only neurons and macroglia were engineered to have defective GlcCerase activity,
with microglia unaltered (8). Nestin-Cre Gbaflox/flox mice
develop similar neurological signs and brain pathology to
another nGD mouse, which does contain defective GlcCerase
activity in microglia (the K14-Cre Gbalnl/lnl mouse) (8).
Similar regions appear to be affected in both models, although
the Nestin-Cre Gbaflox/flox mice show a higher extent of microglial activation than K14-Cre Gbalnl/lnl mice (8). The main
advantage of the nestin-Cre Gbaflox/flox mouse is that it lives
considerably longer than the K14-Cre Gbalnl/lnl mouse (8),
allowing more in-depth analysis of the neuropathological
changes over a longer period of time.
Our data suggest that the affected brain areas that we have
identified in nGD mice may also be affected in human GD
patients. Human nGD brain shows a selective and specific
loss of neurons in cortical layers III and V (7), consistent
with our data that cortical layer V is severely affected in
nGD mice. However, we were also able to detect a regional
specificity, with motor and somatosensory cortex more
affected than caudal cortical regions, such as the visual
cortex. Human nGD patients also show selective loss of pyramidal neurons in CA2 – CA4, but not in the CA1 regions of the
hippocampus (7), and a similar pattern was observed in nGD
mice with respect to microglial activation, although this
occurred late in disease progression and was not as pronounced as in other brain areas. Finally, both substantia
nigra and red nucleus pathology were observed in our study
and also previously in the human brain (7).
Considering the similarities in the pathological changes
between the mouse model used in this study and in the
human disease, important lessons about the progress of the
human disease can be derived, along with lessons concerning the time of possible therapeutic intervention. Thus, neuroinflammation was detected well before noticeable
symptoms in the mouse [although it was not present in
the brains of younger [10-day-old) nGD mice (data not
shown)], which was followed by neuronal loss in defined
areas prior to symptom onset. This suggests that early therapeutic intervention, even before the appearance of clinical
symptoms, will be necessary once such interventions
become available. An additional implication of our study
Downloaded from hmg.oxfordjournals.org at King's College London on March 7, 2011
Olfactory bulb
Mitral cell layer
Cortex
Frontal association (FrA)
Orbital
Motor
Sensory
Piriform
Parietal
Auditory
Visual
CEnt
Basal ganglia
Caudate putamen
Nucleus accumbens
Globus pallidus (GP)
Substantia nigra (SN)
Hippocampus
CA1
CA2/CA3
Hilus (CA4)
Dentate gyrus (DG)
Dorsal subiculum (DS)
Hypothalamus
Anterior hypothalamic area (AHA)
Lateral nucleus (LA)
Medial preoptic nucleus (MPO)
Ventromedial nucleus (VMH)
Mammillary nucleus
Thalamic nuclei
Anteroventral (AV)
Anterodorsal (AD)
Paraventricular, anterior part (PVA)
Reticular (Rt)
Ventral posteromedial/posterolateral
(VPM/VPL)
Mediodorsal (MD)
Habenular
Latrodorsal (LD)
Posterior (Po)
Lateral geniculate, dorsal part
(LGND)
MGN
Midbrain
Red nucleus
Superficial area of superior
colliculus (SC)
Intermediate and deep layers of SC
Periaqueductal gray (PAG)
Inferior colliculus (IC)
Pons and medulla
Reticulotegmental nucleus of the
pons (RtTg)
Pontine nucleus
Ventral cochlear nucleus (VCA)
8 nerve
Facial nucleus (7N)
Vestibular nucleus (Ve)
External cuneate (ECu)
LRt
Parvicellular reticular nucleus
(PCRt)
Motor trigeminal nucleus
Motor root of the trigeminal nerve
(m5)
Intensity of Mac2 labelinga
12 days 16 days
21 days
Table 1. Continued
Human Molecular Genetics, 2011, Vol. 20, No. 7
1381
Figure 7. Midbrain and brainstem pathology in nGD mice. (A– F) Immunohistochemical staining for Mac2 in 16-day-old 2/2 mice. No staining was observed
in control brains (data not shown). Sections in (A) and (D) were counterstained with Nissl (purple). Scale bar ¼ 100 mm for A, B, D and F and 200 mm for
C and E.
Figure 8. LRt pathology in nGD mice. Immunohistochemical staining for Mac2 revealed pronounced microglial activation in 21-day-old 2/2 mice, which was
accompanied by the loss of neurons, evident by Nissl staining. In contrast, the adjacent facial nucleus was unaltered. Scale bar ¼ 200 mm.
Downloaded from hmg.oxfordjournals.org at King's College London on March 7, 2011
Figure 6. Substantia nigra pathology in nGD mice. (A) Nissl (purple) stained sections labeled for Mac2 (brown) revealed microglial activation in the substantia
nigra of 12-day-old 2/2 mice (arrows), which increased with age. Borders between the SNC (C) and the reticulata (R) are indicated. No staining was observed
in the control brain (data not shown). Scale bar ¼ 200 mm. (B) Unbiased optical fractionator estimates of the number of Nissl-stained neurons in SNR. ∗ P ,
0.05, n ¼ 3 + SEM.
1382
Human Molecular Genetics, 2011, Vol. 20, No. 7
Figure 10. Additional brain regions affected in nGD mice. Immunohistochemical staining for Mac2 revealed localized microglial activation in (A) the mitral cell
layer of the olfactory bulb and in (B) the CEnt of 12-day-old 2/2 mice. (C) Microglial activation in CA2–CA3 and in the hippocampal hilus (CA4) was apparent in 2/2 mice at 16 days, and had spread to CA1 by 21 days. No staining was observed in control brain (data not shown). Scale bar ¼ 200 mm.
concerns the efficacy of a stem cell therapy approach in
treating neuronal forms of GD, as well as LSDs in
general. Such a therapeutic approach is based on GlcCerase
secretion by healthy (stem) cells, and its subsequent uptake
by unhealthy Gaucher cells (cross-correction). However,
since the mouse model used in the current study contains
normal GlcCerase activity in microglia, our data would
suggest that secretion of GlcCerase by microglia is not sufficient to rescue Gaucher neurons, as massive neuronal
death occurs after microglial proliferation, raising questions
concerning the suitability of stem cell therapy in treating
acute forms of nGD.
Downloaded from hmg.oxfordjournals.org at King's College London on March 7, 2011
Figure 9. Sensory trigeminal nucleus pathology in nGD mice. (A) Immunohistochemical staining for Mac2 revealed microglial activation (arrows) in the SP5 of
12-day-old 2/2 mice, which became more pronounced with age. No staining was observed in control brains (data not shown). Scale bar ¼ 200 mm.
(B) Unbiased optical fractionator estimates of the number of Nissl-stained neurons in the interpolar part of SP5 (SP5I). ∗ P , 0.05, ∗∗ P , 0.01, n ¼ 3 +
SEM. (C) Unbiased Cavalieri estimates of brainstem volume. n ¼ 4– 5 + SEM.
Human Molecular Genetics, 2011, Vol. 20, No. 7
1383
Figure 11. Not all brain areas exhibit microglial activation and neuron loss at
the end-stage of the disease. (A) Immunohistochemical staining for Mac2
demonstrated that in contrast to adjacent areas of the midbrain reticular formation, the superficial layers of the superior colliculus (SC) and the PAG displayed very little microglial activation in 21-day-old 2/2 mice. Scale bar ¼
500 mm. (B) Unbiased optical fractionator estimates of the number of Nissl
stained neurons in PAG. n ¼ 2 + SEM.
One of the most important findings from our study is identification of pathological changes within several brain structures
that may explain early clinical signs in nGD patients. The most
consistent and well characterized of these are abnormalities in
eye movements, specifically a failure to initiate saccadic eye
movements, slow saccades and an abnormality in horizontal
saccades, which can also be followed by abnormality in vertical saccades (3). We have demonstrated early microglial activation, astrocytosis and neuron loss in the SNR, an area that
plays a vital role in the initiation of saccades (12,13). We
also show pronounced and localized microglial activation in
the RtTg, another area involved in horizontal saccades (14).
Abnormality in ABRs is also an early neurological sign in
nGD (16– 18), and these abnormalities could result from
damage to nuclei in the brainstem auditory pathway, i.e. the
eighth cranial nerve (vestibulocochlear nerve), cochlear
nucleus, superior olivary complex, lateral lemniscus or inferior
colliculus (15). Although we found that the vestibulocochlear
nerve is apparently unaffected, the cochlear nucleus displayed
microglial activation and astrocytosis at the pre-symptomatic
stage, in agreement with a study in one nGD patient (19).
The inferior colliculus, another component of the brainstem
auditory pathway, also displayed microglial activation at the
mid-stage of disease progression.
A key feature of our study was the early and pronounced
targeting of the somatosensory system, which relays information about touch, pain, temperature and body position
from somatosensory receptors (20). Many components of
this pathway showed significant early pathology in nGD
mice (Fig. 12), including the sensory trigeminal nucleus
SP5, which receives somatosensory information from the
face and sends it to the thalamic nucleus VPM (20). Both
SP5 and VPM/VPL displayed intense microglial activation
and a significant reduction in neuron number in nGD mice
by the mid-stage of disease progression. While the VPM
receives somatosensory information via the sensory trigeminal
nuclei, the VPL receives somatosensory information from the
limbs and trunk via the dorsal column nuclei and spinal cord
(20). One of these dorsal column nuclei is the external
cuneate nucleus, which also showed microglial activation at
early stages. The VPM/VPL sends its major output to the
primary somatosensory cortex, a region in which pronounced
microglial activation, astrogliosis and neuron loss were all
detected, although these phenomena were not confined to the
somatosensory cortex. Pathological targeting of the VPM/
VPL and the cortical layers that receive projections from it,
or project to it, is a consistent feature in mouse models of
LSDs (21 – 23). In contrast, in nGD mice, inflammation in
the cortex was largely confined to layer V, but not layer IV,
which receives direct input from VPM/VPL (20). Interestingly, nGD patients show a greatly increased amplitude of
somatosensory evoked potentials (24).
Another anatomical characteristic of the structures that
display microglial activation in nGD mice is that many of
them send mossy fibers to the cerebellum. Cerebellar mossy
fibers form large glutamatergic synapses that contact numerous granule cells in the cerebellum (25). These structures
include the pontine nuclei [which receives its major input
from cortical layer V (26)], the RtTg, the vestibular nuclei,
the LRt, the dorsal column nuclei, the trigeminal nuclei and
the spinal cord (26), which all displayed early microglial activation.
The fact that only specific sub-populations of neurons were
lost upon GlcCerase deficiency further strengthens the need to
carefully choose the correct cellular model for nGD research.
We previously demonstrated that conclusions based on study
of fibroblasts were not relevant for neuronal models (27),
and we now suggest that further research on nGD should
Downloaded from hmg.oxfordjournals.org at King's College London on March 7, 2011
Figure 12. A summary of somatosensory system components that displayed
early and pronounced microglial activation in nGD mice.
1384
Human Molecular Genetics, 2011, Vol. 20, No. 7
MATERIALS AND METHODS
Nestin-flox/flox mice
Nestin-Cre Gbaflox/flox mice were used as a model of nGD (8).
Flox/flox Cre 2/2 mice were crossed with wild-type/flox Cre
+/2 mice to generate flox/flox Cre +/2 mice (referred to as
2/2 mice) and wild-type/flox Cre +/2 mice (referred to as
+/2 mice), which served as healthy controls. Genotyping was
performed by PCR using genomic DNA extracted from mouse
tails (27). The colony was maintained in the Experimental
Animal Center of the Weizmann Institute of Science.
Histological analysis
Brains were removed and immersion fixed in 2.5% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4 for 2 days,
then cryoprotected in 30% sucrose in 50 mM Tris-buffered
saline (TBS), pH 7.6, for several days and kept frozen at
2208C prior to sectioning. Brains were bisected along the
midline, and 40 mm frozen coronal sections (through the rostrocaudal extent of the hemi-brain) were collected in 96-well
plates (one section per well) containing a cryoprotectant
(30% ethylene glycol, 15% sucrose, 0.05% sodium azide in
TBS).
Nissl staining
A series of every sixth section through each brain was
mounted and stained with cresyl violet (23). Slides were incubated with 0.05% cresyl fast violet (Merck, Germany), 0.05%
acetic acid in water for 45 min at 608C, rinsed in deionized
water, differentiated through an ascending series of alcohols,
before clearing in xylene, and cover slipped with DPX
(VWR, West Chester, PA, USA).
Immunohistochemistry
A series of every sixth section was collected in six-well plates,
and endogenous peroxidase activity was quenched with 1%
hydrogen peroxidase in TBS for 30 min. Sections were
rinsed in TBS and blocked with 15% normal horse serum
(Vector, Burlingame, CA, USA) in TBS with 0.3% Triton X
(TBS-T) for 30 min, and then incubated with primary antibodies: rabbit anti-GFAP (1:4000, AbD Serotec, Oxford,
UK) or rat anti-Mac2 (1:1000, Cedarlane, Ontario, Canada)
in 10% normal horse serum in TBS-T, overnight at 48C. Sections were rinsed in TBS and incubated with the appropriate
biotinylated secondary antibodies: biotinylated donkey antirabbit (1:200, Jackson Immunoresearch, West Grove, PA,
USA) or biotinylated donkey anti-rat (1:200, Jackson) in
10% normal horse serum in TBS-T for 2 h at room temperature. Sections were rinsed in TBS followed by incubation
with avidin – biotin – peroxidase (Vectastain Elite ABC kit,
Vector, 1:1000) in TBS for 2 h and rinsed in TBS. To visualize
immunoreactivity, sections were incubated with 0.05%
3,3′ -diaminobenzidine tetrahydrochloride containing 0.001%
hydrogen peroxide in TBS for 10 min, and rinsed in ice-cold
TBS. Sections were mounted on gelatin – chrome-coated
Superfrost microscope slides, air-dried overnight, cleared in
xylene and cover slipped with DPX. In some cases, Nissl
Downloaded from hmg.oxfordjournals.org at King's College London on March 7, 2011
focus on the same specific neuronal sub-populations that show
the most acute response to GlcCerase deficiency, such as cortical layer V, VPM/VPL and neurons which project mossy
fibers.
A major remaining challenge is to determine the mechanistic basis for the sensitivity of specific neuronal subpopulations to a loss of GlcCerase. One plausible explanation
may be the relative level of GlcCerase activity/expression in
these neuron populations. It might be predicted that neurons
which normally contain high levels of GlcCerase would be
most vulnerable when this enzyme is deficient. Unfortunately,
there is no information available about GlcCerase expression
(or of GlcCer levels) at the resolution of the specific brain
areas identified in our study. Most mouse neurons were
reported to show high GlcCerase mRNA expression at 2
days and at 4 weeks (28), but this contrasts with another
study in 56-day-old mice that revealed significant heterogeneity between different brain regions (29). Indeed, some of
the pathological findings that we describe match data from
this latter study. For instance, the mitral cell layer of the olfactory bulb, CEnt, red nucleus of the midbrain, RtTg and pontine
nucleus, which each display distinct microglial activation, all
have relatively high GlcCerase expression (29). Distinct patterns of GlcCerase expression are also apparent in cortical
layer V and to a lesser extent in layers II/III, in addition to
high levels throughout the hippocampus (29). In the human
brain, strong GlcCerase immunoreactivity was observed in
hippocampal regions CA2 – CA4, in cortical layer V and to a
lesser extent in layer 3. However, not all brain regions that displayed pathology in the nGD mouse had high GlcCerase
expression levels in the adult mouse (http://mouse.bra
in-map.org), including the thalamus and the trigeminal
sensory nuclei. Levels of expression during development
may be of more relevance for disease progression, or pathology in these areas may be a secondary process, or expression
levels of GlcCerase mRNA may not correspond with GlcCerase activity.
Another area that displays distinct and relatively high
levels of GlcCerase expression is the SNC (http://mouse.
brain-map.org). The loss of dopaminergic neurons in this
area causes Parkinson disease (PD), which may be of particular importance since mutations in the GlcCerase gene
were found to be a risk factor for PD (30). Another area
that may be involved in both disorders is the olfactory
bulb, since a marked decrease in olfaction is highly characteristic of PD (31). The olfactory bulb mitral cell layer has
relatively high GlcCerase expression (22), and in nGD mice
it presented with early onset and specific microglial activation and astrocytosis.
Finally, neuroinflammation has emerged as a common
feature in LSDs with CNS pathology (32), and antiinflammatory drugs were shown to benefit mice with various
LSDs, including Sandhoff disease (33). Our study demonstrates that neuroinflammation, i.e. microglial activation and
astrocytosis, occurs in the same brain regions where the
neuron loss takes place. Presently, we do not know whether
this inflammation precedes neuron death or is concomitant
with it. Irrespective of the precise time course, we suggest
that the use of anti-inflammatory drugs may be beneficial to
alleviate neurological manifestations in nGD.
Human Molecular Genetics, 2011, Vol. 20, No. 7
staining was also performed on the same sections to allow
better identification of brain structures.
A standard immunofluorescence protocol was used for double
labeling of neurons and activated microglia. Twenty micrometer
frozen coronal sections were blocked using 20% normal horse
serum and 0.2% Triton X-100 in PBS for 2 h, and then incubated
with a mixture of mouse anti-NeuN (1:50, Chemicon, Temecula, CA, USA) and rat anti-Mac2 (1:200, Cedarlane) in 2%
normal horse serum containing 0.5% Triton X-100 in PBS overnight at room temperature. For detection of NeuN, sections were
incubated with biotinylated donkey anti-mouse secondary antibody (1:200, Jackson) for 90 min followed by Cy3-conjugated
streptavidin (1:200, Jackson) for 60 min. Mac2 was detected
using a Cy2-conjugated donkey anti-rat antibody (1:200,
Jackson) (2 h).
Unbiased stereological estimates of regional volume or neuron
number were obtained using Stereo Investigator software
(Microbrightfield Inc., Williston, VT, USA) with regions of
interest defined by referring to the neuroanatomical landmarks
described by Paxinos and Franklin (34). All measurements
were made blind (with respect to genotype) using Nisslstained sections. Cavalieri estimates of volume were obtained
using a sampling grid superimposed over the brainstem region
and the number of points covering the relevant area was
assessed. Brainstem volume was estimated from one in three
series of sections through the brainstem using a 200 mm
sampling grid.
To measure cortical thickness, the length of perpendicular
lines extending from the white matter to the pial surface of
the cortex was measured in M1, S1BF and V1 cortical subfields with Stereo Investigator software, using 10 lines
placed on three consecutive sections spanning each cortical
region. Results are expressed as the mean cortical thickness
(mm per region).
The design-based optical fractionator probe was used to
estimate cell number (35) in Nissl-stained sections. Nisslstained cells were only counted if they had a neuronal morphology and a clearly identifiable nucleolus. A line was
traced around the boundary of the region of interest, a grid
was superimposed and cells were counted using a 100× objective within a series of dissector frames placed according to
intersections of the sampling grid. Different grid and dissector
sizes were determined according to each brain region using a
coefficient of error (CE) value of ,0.1 to indicate sampling
efficiency. Layer V of the S1BF grid was 200 × 200 mm,
frame 60 × 40 mm; the VPM/VPL grid was 175 × 175 mm,
frame 70 × 42 mm; the SNR grid was 150 × 150 mm, frame
73 × 44 mm; the SP5I grid (extending from immediately
caudal to the facial nucleus to the first appearance of area postrema) was 220 × 220 mm, frame 65 × 35 mm for 21- and
16-day-old animals, and 190 × 190 mm, frame 65 × 35 mm
for 12-day-old animals; the PAG grid (extending from
immediately caudal to mammillary bodies to the first appearance of the inferior colliculus) was 175 × 173 mm, frame
65 × 35 mm. A 1:3 series was used for SP5I analysis, and a
1:6 series was used for all other brain regions.
Statistical analysis
The mean CE for all individual optical fractionator and Cavalieri estimates was calculated according to the method of Gundersen and Jensen (36) and was ,0.1 in the majority of cases,
except for few 21-day-old 2/2 brains, in which neurons were
largely absent. P-values were calculated using a one-tailed,
two-independent sample Student’s t-test. A P-value ≤ 0.05
was considered statistically significant.
ACKNOWLEDGEMENTS
We would like to thank Drs Andrew Wong and Zvi Farfel for
helpful discussions, and Dr Stefan Karlsson for supplying the
Gbaflox/flox;Nestin-Cre mice. A.H. Futerman is the Joseph
Meyerhoff Professor of Biochemistry at the Weizmann Institute of Science.
Conflict of Interest statement. None declared.
FUNDING
This work was supported by the Children’s Gaucher Research
Fund (A.H.F.), the Batten Disease Support and Research
Association, Batten Disease Family Association and the
Natalie Fund (J.D.C.).
REFERENCES
1. Gupta, N., Oppenheim, I.M., Kauvar, E.F., Tayebi, N. and Sidransky, E.
(2011) Type 2 Gaucher disease: phenotypic variation and genotypic
heterogeneity. Blood Cells Mol. Dis., 46, 75– 84.
2. Mignot, C., Doummar, D., Maire, I. and De Villemeur, T.B. (2006) Type
2 Gaucher disease: 15 new cases and review of the literature. Brain Dev.,
28, 39–48.
3. Schiffmann, R. and Vellodi, A. (2007) Neuronopathic Gaucher disease. In
Futerman, A.H. and Zimran, A. (eds), Gaucher Disease, CRC Press,
Taylor & Francis Group, pp. 175–196.
4. Harris, C.M., Taylor, D.S. and Vellodi, A. (1999) Ocular motor
abnormalities in Gaucher disease. Neuropediatrics, 30, 289–293.
5. Patterson, M.C., Horowitz, M., Abel, R.B., Currie, J.N., Yu, K.T.,
Kaneski, C., Higgins, J.J., O’Neill, R.R., Fedio, P., Pikus, A. et al. (1993)
Isolated horizontal supranuclear gaze palsy as a marker of severe systemic
involvement in Gaucher’s disease. Neurology, 43, 1993– 1997.
6. Altarescu, G., Hill, S., Wiggs, E., Jeffries, N., Kreps, C., Parker, C.C.,
Brady, R.O., Barton, N.W. and Schiffmann, R. (2001) The efficacy of
enzyme replacement therapy in patients with chronic neuronopathic
Gaucher’s disease. J. Pediatr., 138, 539–547.
7. Wong, K., Sidransky, E., Verma, A., Mixon, T., Sandberg, G.D.,
Wakefield, L.K., Morrison, A., Lwin, A., Colegial, C., Allman, J.M. et al.
(2004) Neuropathology provides clues to the pathophysiology of Gaucher
disease. Mol. Genet. Metab., 82, 192– 207.
8. Enquist, I.B., Lo Bianco, C., Ooka, A., Nilsson, E., Mansson, J.E.,
Ehinger, M., Richter, J., Brady, R.O., Kirik, D. and Karlsson, S. (2007)
Murine models of acute neuronopathic Gaucher disease. Proc. Natl Acad.
Sci. USA, 104, 17483–17488.
9. Amor, S., Puentes, F., Baker, D. and van der Valk, P. (2010) Inflammation
in neurodegenerative diseases. Immunology, 129, 154– 169.
10. Raivich, G., Bohatschek, M., Kloss, C.U., Werner, A., Jones, L.L. and
Kreutzberg, G.W. (1999) Neuroglial activation repertoire in the injured
brain: graded response, molecular mechanisms and cues to physiological
function. Brain Res. Brain Res. Rev., 30, 77–105.
11. Rotshenker, S. (2009) The role of Galectin-3/MAC-2 in the activation of
the innate-immune function of phagocytosis in microglia in injury and
disease. J. Mol. Neurosci., 39, 99–103.
Downloaded from hmg.oxfordjournals.org at King's College London on March 7, 2011
Regional volume, cortical thickness and cell number
estimation
1385
1386
Human Molecular Genetics, 2011, Vol. 20, No. 7
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
marker of disease burden in type 3 Gaucher disease. Neurology, 56, 391–
394.
Voogd, J. (2004) Cerebellum. In Paxinos, G. (ed.), The Rat Nervous
System. Elsevier, pp. 205– 242.
Ruigrok, T.J.H. (2004) Precerebellar nuclei and red nucleus. In Paxinos,
G. (ed.), The Rat Nervous System. Elsevier, pp. 167– 204.
Farfel-Becker, T., Vitner, E., Dekel, H., Leshem, N., Enquist, I.B.,
Karlsson, S. and Futerman, A.H. (2009) No evidence for activation of the
unfolded protein response in neuronopathic models of Gaucher disease.
Hum. Mol. Genet., 18, 1482– 1488.
Ponce, E., Witte, D.P., Hung, A. and Grabowski, G.A. (2001) Temporal
and spatial expression of murine acid beta-glucosidase mRNA. Mol.
Genet. Metab., 74, 426–434.
Lein, E.S., Hawrylycz, M.J., Ao, N., Ayres, M., Bensinger, A., Bernard,
A., Boe, A.F., Boguski, M.S., Brockway, K.S., Byrnes, E.J. et al. (2007)
Genome-wide atlas of gene expression in the adult mouse brain. Nature,
445, 168–176.
Sidransky, E., Nalls, M.A., Aasly, J.O., Aharon-Peretz, J., Annesi, G.,
Barbosa, E.R., Bar-Shira, A., Berg, D., Bras, J., Brice, A. et al. (2009)
Multicenter analysis of glucocerebrosidase mutations in Parkinson’s
disease. N. Engl. J. Med., 361, 1651–1661.
Goker-Alpan, O., Lopez, G., Vithayathil, J., Davis, J., Hallett, M. and
Sidransky, E. (2008) The spectrum of parkinsonian manifestations
associated with glucocerebrosidase mutations. Arch. Neurol., 65, 1353–
1357.
Vitner, E.B., Platt, F.M. and Futerman, A.H. (2010) Common and
uncommon pathogenic cascades in lysosomal storage diseases. J. Biol.
Chem., 285, 20423– 20427.
Jeyakumar, M., Smith, D.A., Williams, I.M., Borja, M.C., Neville,
D.C., Butters, T.D., Dwek, R.A. and Platt, F.M. (2004) NSAIDs
increase survival in the Sandhoff disease mouse: synergy with
N-butyldeoxynojirimycin. Ann. Neurol., 56, 642– 649.
Franklin, K.B.J. and Paxinos, G. (2008) The Mouse Brain in Stereotaxic
Coordinates. Academic Press.
West, M.J., Slomianka, L. and Gundersen, H.J. (1991) Unbiased
stereological estimation of the total number of neurons in the subdivisions
of the rat hippocampus using the optical fractionator. Anat. Rec., 231,
482–497.
Gundersen, H.J. and Jensen, E.B. (1987) The efficiency of systematic
sampling in stereology and its prediction. J. Microsc., 147, 229– 263.
Downloaded from hmg.oxfordjournals.org at King's College London on March 7, 2011
12. Hikosaka, O. and Wurtz, R.H. (1983) Visual and oculomotor functions of
monkey substantia nigra pars reticulata. I. Relation of visual and auditory
responses to saccades. J. Neurophysiol., 49, 1230– 1253.
13. Hikosaka, O. and Wurtz, R.H. (1985) Modification of saccadic eye
movements by GABA-related substances. II. Effects of muscimol in
monkey substantia nigra pars reticulata. J. Neurophysiol., 53, 292– 308.
14. Buttner-Ennever, J.A. and Horn, A.K. (1997) Anatomical substrates of
oculomotor control. Curr. Opin. Neurobiol., 7, 872 –879.
15. Campbell, P.E., Harris, C.M. and Vellodi, A. (2004) Deterioration of the
auditory brainstem response in children with type 3 Gaucher disease.
Neurology, 63, 385– 387.
16. Kaga, M., Azuma, C., Imamura, T., Murakami, T. and Kaga, K. (1982)
Auditory brainstem response (ABR) in infantile Gaucher’s disease.
Neuropediatrics, 13, 207–210.
17. Bamiou, D.E., Campbell, P., Liasis, A., Page, J., Sirimanna, T., Boyd, S.,
Vellodi, A. and Harris, C. (2001) Audiometric abnormalities in children
with Gaucher disease type 3. Neuropediatrics, 32, 136– 141.
18. Campbell, P.E., Harris, C.M., Sirimanna, T. and Vellodi, A. (2003) A
model of neuronopathic Gaucher disease. J. Inherit. Metab. Dis., 26, 629–
639.
19. Lacey, D.J. and Terplan, K. (1984) Correlating auditory evoked and
brainstem histologic abnormalities in infantile Gaucher’s disease.
Neurology, 34, 539– 541.
20. Tracey, D. (2004) Somatosensory system. In Paxinos, G. (ed.), The Rat
Nervous System. Elsevier, pp. 797 –815.
21. Pressey, S.N., O’Donnell, K.J., Stauber, T., Fuhrmann, J.C., Tyynela, J.,
Jentsch, T.J. and Cooper, J.D. (2010) Distinct neuropathologic phenotypes
after disrupting the chloride transport proteins ClC-6 or ClC-7/Ostm1.
J. Neuropathol. Exp. Neurol., 69, 1228– 1246.
22. Partanen, S., Haapanen, A., Kielar, C., Pontikis, C., Alexander, N.,
Inkinen, T., Saftig, P., Gillingwater, T.H., Cooper, J.D. and Tyynela, J.
(2008) Synaptic changes in the thalamocortical system of cathepsin
D-deficient mice: a model of human congenital neuronal
ceroid-lipofuscinosis. J. Neuropathol. Exp. Neurol., 67, 16– 29.
23. Kielar, C., Maddox, L., Bible, E., Pontikis, C.C., Macauley, S.L., Griffey,
M.A., Wong, M., Sands, M.S. and Cooper, J.D. (2007) Successive neuron
loss in the thalamus and cortex in a mouse model of infantile neuronal
ceroid lipofuscinosis. Neurobiol. Dis., 25, 150–162.
24. Garvey, M.A., Toro, C., Goldstein, S., Altarescu, G., Wiggs, E.A., Hallett,
M. and Schiffmann, R. (2001) Somatosensory evoked potentials as a