Download Intraventricular Enzyme Replacement Improves Infantile Neuronal Ceroid Lipofuscinosis

© The American Society of Gene Therapy
original article
Intraventricular Enzyme Replacement Improves
Disease Phenotypes in a Mouse Model of Late
Infantile Neuronal Ceroid Lipofuscinosis
Michael Chang1, Jonathan D Cooper2, David E Sleat3,4, Seng H Cheng5, James C Dodge5,
Marco A Passini5, Peter Lobel3,4 and Beverly L Davidson1,6,7
Department of Molecular Physiology and Biophysics, University of Iowa, Iowa City, Iowa, USA; 2Pediatric Storage Disorders Laboratory, Department
of Neuroscience, Institute of Psychiatry, King’s College London, London, UK; 3Center for Advanced Biotechnology and Medicine, Robert Wood Johnson
Medical School, Piscataway, New Jersey, USA; 4Department of Pharmacology, Robert Wood Johnson Medical School, Piscataway, New Jersey, USA;
5
Genzyme Corporation, Framingham, Massachusetts, USA; 6Department of Internal Medicine, University of Iowa, Iowa City, Iowa, USA;
7
Department of Neurology, University of Iowa, Iowa City, Iowa, USA
1
Late infantile neuronal ceroid lipofuscinosis (LINCL) is an
autosomal recessive neurodegenerative disease caused by
mutations in CLN2, which encodes the lysosomal protease tripeptidyl peptidase 1 (TPP1). LINCL is characterized
clinically by progressive motor and cognitive decline, and
premature death. Enzyme-replacement therapy (ERT) is
currently available for lysosomal storage diseases affecting
peripheral tissues, but has not been used in patients with
central nervous system (CNS) involvement. Enzyme delivery through the cerebrospinal fluid is a potential alternative route to the CNS, but has not been studied for LINCL.
In this study, we identified relevant neuropathological
and behavioral hallmarks of disease in a mouse model
of LINCL and correlated those findings with tissues from
LINCL patients. Subsequently, we tested if intraventricular
delivery of TPP1 to the LINCL mouse was efficacious. We
found that infusion of recombinant human TPP1 through
an intraventricular cannula led to enzyme distribution in
several regions of the brain of treated mice. In vitro activity assays confirm increased TPP1 activity throughout the
rostral–caudal extent of the brain. Importantly, treated
mice showed attenuated neuropathology, and decreased
resting tremor relative to vehicle-treated mice. This data
demonstrates that intraventricular enzyme delivery to the
CNS is feasible and may be of therapeutic value.
Received 25 October 2007; accepted 7 January 2008; published online
12 February 2008. doi:10.1038/mt.2008.9
Introduction
Late infantile neuronal ceroid lipofuscinosis (LINCL) is one
of numerous autosomal recessive neurodegenerative diseases
caused by lysosomal protein mutations and characterized clinically by progressive motor and cognitive decline, vision loss,
seizures, and premature death.1 These disorders are categorized
as lysosomal storage diseases based on their characteristic intracellular protein aggregates, and as a group, have a collective
world-wide incidence of ~1 in 12,500 (ref. 2). Currently available treatments for these disorders are merely supportive and
do not alter disease progression, supporting investigations into
novel treatment approaches.
LINCL is caused by mutations in the CLN2 gene (chromosome 11p15, MIM#204500), which encodes the lysosomal protease tripeptidyl peptidase 1 (TPP1).3 The lack of TPP1 activity
results in central nervous system (CNS) degeneration.4–6 A mouse
model of LINCL has been generated by targeting CLN2, resulting in undetectable levels of TPP1, and the recapitulation of many
characteristics of the human disease including motor dysfunction,
CNS pathology, and reduced lifespan.7 In this work, we present
additional characterization of neuropathology and behavior in
the TPP1-deficient mouse, as well as neuropathology in human
LINCL tissues.
Gene transfer studies in TPP1-deficient mice suggest that restoration of enzyme activity can improve the disease phenotype.8–10
While some may consider gene transfer as the ultimate cure for
treating recessive genetic disorders such as lysosomal storage diseases, enzyme replacement therapy (ERT) has already shown clinical benefit for some of these diseases, such as non-neuronopathic
Gaucher disease, Fabry disease, Pompe disease, and several of the
mucopolysaccharidoses.11–16 Of note, intravenous ERT has thus far
been effective only for those storage diseases that affect primarily
peripheral tissues, and has not been effective for CNS disease in
patients due to inefficient passage of enzymes across the blood–
brain barrier. Thus, intravenous administration of recombinant
TPP1 is unlikely to have any therapeutic value for CNS disease in
LINCL. Although it is possible to permeabilize the blood–brain
barrier with osmotic or other agents experimentally,17–19 this is not
a viable option in a clinical setting where repeated enzyme administration may be necessary.
An alternative delivery strategy is to bypass the blood–brain
barrier completely with direct injections to the brain parenchyma
or ventricular space. Studies in nonhuman primates demonstrate the feasibility of glucocerebrosidase delivery directly into
the brain, and that enzyme spread can be further enhanced via
Correspondence: Beverly L. Davidson, University of Iowa, 200 EMRB, Iowa City, Iowa 52242, USA. E-mail: [email protected]
Molecular Therapy vol. 16 no. 4, 649–656 april 2008
649
© The American Society of Gene Therapy
Enzyme Replacement for LINCL
Results
Neuropathology of TPP1-deficient mouse and LINCL
patient tissue
The TPP1-deficient mouse has been previously reported to
exhibit widespread neuropathology.7 Although sparce loss of
­cerebellar Purkinje cells was originally reported, it was surprising that overt neuron loss within the CNS was not observed
given the pronounced motor deficits in this mouse, and the
extent of neurodegeneration in LINCL patients. Here we present
more detailed characterization and quantitation of neuropathology in the TPP1-deficient mouse, including the identification of
several degenerating neuronal populations that may be significant in the disease progression in this mouse, and of relevance
to LINCL patients.
Reactive astrocytosis is a well-described pathological process
that generally occurs in response to neuron injury or dysfunction
in the CNS. To determine the extent of astrocytosis in the TPP1deficient mouse CNS, we immunostained brain sections from
mice of various ages for glial fibrillary acidic protein (GFAP).
TPP1-deficient mice exhibit progressive reactive astrocytosis in
several regions of the brain, including the motor cortex, hippocampus, striatum, and cerebellum (Figure 1a–f), in contrast to
heterozygous control mice which show little to no astrocytosis.
Thresholding analysis with the public domain software Image J22
was used to determine the relative extent of GFAP immunoreactivity. This analysis revealed a significant degree of astrocytosis in
the primary motor cortex as early as at 63 days of age, which progresses aggressively to include most CNS structures by 99 days
of age, and continues to intensify until the end stage of disease at
148 days of age (Figure 1g).
Reactive astrocytosis is observed in the cerebellum of TPP1deficient mice but not in the heterozygous controls (Figure 2a
and b). Severe astrocytosis occurs in the granule cell layer and
appears to be homogeneous throughout this layer, whereas the
molecular layer exhibits sporadic patches of gliosis corresponding to reactive Bergmann glia. No apparent pattern was observed
as to which lobules of the cerebellum were more susceptible to
this event, but interestingly, the patches of Bergmann gliosis
occur in regions where Purkinje cells are lost, as demonstrated
650
a
35 days
b
63 days
c
99 days
d
148 days
e
g
Hippocampus
Striatum
f
12
10
Area fraction (%)
a technique termed ‘convection-enhanced delivery’.20 However,
the invasiveness of intraparenchymal injections is a major limitation in a clinical setting. Intrathecal enzyme delivery is also
feasible, and potentially much less invasive than intraparenchymal injection.21 Intrathecal delivery takes advantage of the large
surface area of the brain that comes in contact with cerebrospinal fluid, either through ependyma-lined ventricular surfaces,
or perivascular (Virchow–Robbins) spaces. In the canine model
of mucopolysaccharidosis type I, intrathecal injection of human
α-l-iduronidase led to reductions in brain and meningeal glycosaminoglycan levels, as well as histological improvements in
lysosomal storage,21 suggesting that exogenous enzymes can be
endocytosed and trafficked correctly to lysosomes. In this study,
we tested the therapeutic efficacy of intraventricular TPP1 infusion in the TPP1-deficient mouse, and demonstrated that several
behavioral and neuropathological phenotypes in this mouse are
attenuated following enzyme delivery.
8
6
4
2
0
+/−
35
63
Age (days)
99
148
Figure 1 Progressive reactive astrocytosis in the CLN2–/– mouse.
Coronal brain sections from 35-, 63-, 99-, and 148-days-old CLN2–/–
mice were immunostained for glial fibrillary acidic protein (GFAP).
(a–d) Images from primary motor cortex, (e) hippocampus, (f) striatum.
­(g) ­Thresholding analysis via Image J software indicates the fraction of
total area that is positive for GFAP staining in primary motor cortex (n = 3
mice per timepoint, three sections per mouse). Data expressed as mean ±
SEM. Scale bars for (a–d) is 100 µm, while for (e and f) it is 200 µm.
by co-immunostaining for GFAP and the Purkinje cell marker
calbindin (Figure 2c and d).
Cerebellar neuron loss in TPP1-deficient mice is not confined
to Purkinje cells. Neurons of the deep cerebellar nuclei (DCNs), as
visualized by Nissl stain or Neu-N immunostaining, also degenerate in these mice, which exhibit a near complete loss of DCN
neurons by 150 days of age, whereas those of heterozygote mice
remain intact (Figure 2e and f). Neurons in the lateral medullary
reticular formation are also lost in the TPP1-deficient mouse but
remain intact in heterozygous controls (Figure 2g and h).
Tissue specimens from LINCL patient brains exhibit pathological findings similar to those noted in TPP1-deficient mice.
Sections through the primary motor cortex show fewer neurons
and a prominent astroglial response (Figure 3a and b). In some
cases, the hypertrophied astrocytes encircle the remaining cortical neurons. Clusters of what appear to be microglia are also
evident (Figure 3b). In human LINCL cerebellum, the molecular layer is atrophic with pronounced granule neuron loss and
an almost complete absence of Purkinje cells (Figure 3c and d).
As with neurons in the cortex, the few Purkinje cells remaining
in the LINCL cerebellum are distended with cytoplasmic ­storage
www.moleculartherapy.org vol. 16 no. 4 april 2008
© The American Society of Gene Therapy
a
Enzyme Replacement for LINCL
b
ML
ML
GL
GL
c
d
e
f
g
h
Figure 2 Cerebellar and brainstem pathology. Sections through (a–f)
cerebellum and (g and h) brainstem of 150-days-old CLN2–/– mice
(b–d,f,h) or age-matched CLN2+/– control littermates (a,e,g). (a and b)
Immunostain for glial fibrillary acidic protein (GFAP) reveals gliosis in granular layer (GL) and Bergmann gliosis in molecular layer (ML, arrow) of
CLN2–/– mice (b) but not CLN2+/– controls (a). (c and d) Co-­immunostain
for GFAP (red) and the Purkinje cell marker calbindin (green). (e and f)
Immunostain for Neu-N in cerebellum. (g and h) Nissl-stained ­coronal sections through the brainstem at the level of the lateral medullary ­reticular
formation (arrows). Scale bars for panels (a and b) is 200 µm, while for
panels (c and d) it is 100 µm, and for panels (e–h) is 500 µm.
material (Figure 3d). Within persisting neurons, autofluorescent storage is observed and interestingly, aggregates are also
localized within GFAP-positive astrocytes (Figure 3e–h). GFAP
immunostaining in the LINCL cerebellum also reveals prominent astrocytosis in the granular layer, and Bergmann gliosis in
the molecular layer (Figure 3i and j), similar to our findings in
the TPP1-­deficient mouse (Figure 2). The severe neurodegeneration in the LINCL cerebellum is accompanied by a drastic loss of
axons in the cerebellar white matter as observed by Bielschowsky
silver staining (Figure 3k and l).
Intraventricular enzyme delivery
Several studies suggest that the cerebrospinal fluid can mediate the
distribution of lysosomal enzymes and their subsequent uptake
into the brain parenchyma.21,23 To test if TPP1 can be delivered in
this manner, we infused recombinant human TPP1 proenzyme
Molecular Therapy vol. 16 no. 4 april 2008
(1 mg/ml) directly into the lateral ventricles of TPP1-deficient
mice. Osmotic pumps each delivered 100 µl enzyme through a
cannula implanted in the right lateral ventricle. Correct cannula
placement was confirmed visually in serial coronal brain sections
(Figure 4a). Delivery of TPP1 in this manner results in enzyme
spread to regions of the brain in both hemispheres, particularly
the meninges, corpus callosum, and periventricular areas of the
cortex and hippocampus (Figure 4b–d). Interestingly, TPP1
shows a strong propensity to traffic along the corpus callosum
(Figure 4c). Western blots for TPP1 from these brain homogenates show an accumulation of the mature TPP1 enzyme
(Figure 4e), indicating correct trafficking of the proenzyme to
the lysosome and appropriate processing to the mature enzyme.
In vitro fluorometric TPP1 activity assays with brain homogenate
from serial bilateral coronal sections show that ERT-treated mice
exhibit elevated TPP1 ­activity levels which, in all cases, exceeded
that of heterozygous mice (Figure 4f).
Effects of enzyme replacement on tremor phenotype
We have shown that intraventricular infusion of recombinant
TPP1 can lead to broad distribution of the enzyme in the brain,
and appropriate trafficking of the enzyme to the lysosome within
the cells of the CNS. We next tested if intraventricular ERT has
an effect on the gross motor deficits of TPP1-deficient mice,
which have previously been characterized via accelerating and
rocking rotarod.7 The accelerating rotorod, however, does not
detect a significant difference in performance until 16 weeks of
age, near the end stage of disease in the mouse. Here we developed a behavioral assay to quantify the tremor that is readily
observable by eye in the diseased mouse. In this assay, the mouse
is placed on a platform, which is attached to a force transducer
that continuously detects tremor amplitudes during the testing
period.
We initiated enzyme delivery at 60 days of age and assayed
for mean tremor amplitude in treated and control mice every
5 days until 105 days of age. At 60 days of age, no statistical difference was detected between CLN2+/– control mice, CLN2–/– mice
mock-treated with saline, and CLN2–/– mice treated with ERT.
However, by 80 days of age the tremor amplitude of CLN2+/– control mice decreased, presumably due to habituation to the testing
chamber. Both treated and mock-treated mice maintained their
starting tremor amplitude until 90 days of age. Between 95 and
105 days of age, however, the tremor amplitude of mock-treated
mice increased progressively, whereas that of ERT-treated mice
never increased beyond the levels detected at 80 days (Figure 5a).
The mean tremor amplitude of enzyme-treated mice were significantly lower than that of mock-treated mice at 100 and 105 days
of age (Figure 5a, P < 0.05).
Effect of enzyme replacement on neuropathology
Enzyme-treated TPP1-deficient mice also showed improvements
in neuropathological measures. To determine the effect of enzyme
delivery on the reactive gliosis observed in the CLN2–/– mouse,
we immunostained for GFAP in brain sections from CLN2+/–,
CLN2–/– mock-treated, and CLN2–/– ERT-treated mice. Staining
in the primary motor cortex was markedly decreased in ERTtreated mice compared to mock-treated mice (Figure 5b–d).
651
© The American Society of Gene Therapy
Enzyme Replacement for LINCL
a
Con GFAP
b
LINCL GFAP
c
Con
d
LINCL
*
*
*
*
e
*
*
g
f
405 nm AF
i
Con GFAP
h
546 nm AF
j
LINCL GFAP
ML
ML
GL
GL
Merge
GFAP
k
Con silver
l
LINCL silver
Figure 3 Neuropathology in human late infantile neuronal ceroid lipofuscinosis (LINCL). (a and b) Immunohistochemical staining for the astrocytic marker glial fibrillary acidic protein (GFAP) in human brain sections from the primary motor cortex of (a) control and (b) LINCL brains. In (b),
arrowheads denote hypertrophied astrocytes and GFAP+ processes encircling persisting distended neurons (denoted by *). Arrows denote microglia
clusters. (c and d) Representative hematoxylin-stained sections from the cerebellar vermis from (c) control and (d) LINCL brains. Purkinje neurons in
control tissue (arrows, c), are diminished in LINCL cerebellar sections (d, arrowhead) and if present are distended. (e–g) Multichannel confocal microscopy to assess autofluorescence and GFAP immunoreactivity in LINCL brain. (e) Blue 405 nm; (f) red 546 nm; (g) GFAP immunoreactivity, 488 nm;
(h) merged image. Autofluorescence is seen within GFAP-positive astrocytes (arrowheads) and neurons (arrow). (i and j) GFAP-stained sections from
(i) control and (j) LINCL cerebellum. Arrowheads in (j) depict hypertrophied astrocytes in the granular layer (GL) and molecular layer (ML); *denotes
Bergmann gliosis in the molecular layer. (k and l) Bielschowsky silverstain of axons in cerebellar white matter tract in sections from (k) ­control and
(l) LINCL brain. Scale bar for panels a and b is 50 µm; for panels c and d is 500 µm (C–D); for panels e–h is 10 µm; for panels k–l 100 µm.
Thresholding analysis confirms a significant difference exists
between ERT and mock-treated groups (Figure 5e, P < 0.005).
A similar analysis for autofluorescent storage indicates a significant decrease in autofluorescence in the primary motor cortex of
enzyme-treated mice (Figure 5f, P < 0.005).
The DCNs of CLN2+/– mice contain large neurons that stain
robustly for Neu-N (Figure 6a). Analysis of sections from treated
and mock-treated LINCL mice revealed that enzyme-treated mice
consistently maintained some integrity of the DCNs (Figure 6b)
whereas some mock-treated mice exhibited total loss of Neu-N
positive cells (Figure 6c). To quantify this phenotype, blinded
grading of cerebellar sections immunostained for the neuronal
marker Neu-N was done by a neuropathologist. For this, a scale of
1–3 was devised where 1 shows few remaining intact neurons; 2
shows intermediate degree of neuron loss; 3 shows no discernable
neuron loss. As expected, all heterozygous mice received a grade
of 3. Mock-treated mice, however, received grades of 2 and 1.
652
Importantly, ERT-treated LINCL mice consistently received an
intermediate score of 2.
Discussion
Neuropathology: Although TPP1-deficient mice exhibit severe
tremor and ataxia, and CNS pathology,7 the precise time course of
neurodegenerative and reactive changes has not previously been
defined. In this study, we establish a time course over which reactive gliosis occurs, and identify several neuron populations that
degenerate in this mouse model. Reactive gliosis is first observed
in the primary motor cortex at 63 days of age, ~1–2 weeks prior
to onset of the tremor phenotype. This timing suggests that cortical dysfunction may contribute to the motor deficits observed in
this mouse. In addition, neurons of the DCNs and regions of the
brainstem reticular formation degenerate almost entirely in the
TPP1-deficient mouse. These neurons mediate a variety of physiological functions ranging from motor coordination to ­respiration,
www.moleculartherapy.org vol. 16 no. 4 april 2008
© The American Society of Gene Therapy
Enzyme Replacement for LINCL
a
a
*
c
b
b
5
60
70
80
+/−
e
d
75
c
−/− Mock
90
95
Age (days)
e
f
ERT
Mock
d
FU · min−1 · µg−1
+/−
1,750
−/− Mock
1,000
600
−/− ERT
−/− ERT
Area fraction (%)
PE
200
0
1
2
3
4
5
6
Section #
7
8
9
10
Figure 4 Tripeptidyl peptidase 1 (TPP1) enzyme distribution following intraventricular infusion. (a) Nissl-stained serial sections through
cerebral cortex. Entrance of cannula into ventricular space (arrow). (b–d)
Immunofluorescent staining for TPP1 in periventricular parenchyma
including the (b) hippocampus (*), (c) corpus callosum, and (e) meninges. At these camera settings, autofluorescence from endogenous storage material in the late infantile neuronal ceroid lipofuscinosis (LINCL)
brain is not detectable. (e) Representative Western blot for TPP1 in
brain homogenate from enzyme-treated [enzyme replacement therapy
(ERT)] and mock-treated (mock) CLN2–/– mice. PE denotes purified TPP1
proenzyme. Only the mature enzyme is detectable in ERT-treated mice
indicating uptake and autoactivation in vivo. (f) In vitro fluorometric
TPP1 enzyme activity assay with brain homogenates from serial bilateral coronal sections of enzyme-treated CLN2–/–, mock-treated CLN2–/–,
and CLN2+/– mice. Data presented as mean ± SEM. Scale bars for panels
b and d is 500 µm, while for panel c it is 100 µm.
and therefore degeneration of these neuron populations may contribute to the motor deficits displayed by this mouse, and perhaps
ultimately its death.
Severe Purkinje cell degeneration occurs in LINCL, and was
expected to occur in the TPP1-deficient mouse, especially in
light of its ataxic phenotype.7 However, we found that only small
patches of Purkinje cells are lost. Initially these appear to be
randomly disbursed patches, but our observations suggest that
Molecular Therapy vol. 16 no. 4 april 2008
105
GFAP
15
10
*
5
+/−
−/− Mock
−/− ERT
Autofluorescence
4
3
*
2
1
0
400
100
20
0
50
35
−/− Mock
−/− ERT
0
*
+/−
*
10
Area fraction (%)
Mean amplitude
15
+/−
−/− Mock
−/− ERT
Figure 5 Effect of enzyme replacement on tremor phenotype and
gliosis of motor cortex. (a) Mean tremor amplitude, expressed in millivolts, in enzyme-treated CLN2–/–, mock-treated CLN2–/–, and CLN2+/–
mice from 60 days of age to 105 days of age. (b–d) Images from motor
cortex of (b) CLN2+/–, (c) mock-treated CLN2–/–, and (d) enzymetreated CLN2–/– mice immunostained for glial fibrillary acidic protein
(GFAP). (e and f) Thresholding analysis via Image J software for (e) positive GFAP staining and (f) autofluorescence in primary motor cortex.
Enzyme-treated group was compared to the mock-treated group at each
time point and statistical significance was determined by student’s t-test.
P ≤ 0.05 are indicated by *. All data are expressed as mean ± SEM. Scale
bars for panels b–d is 100 µm. ERT, enzyme replacement therapy.
areas of Purkinje cell loss are co-localized with discrete areas of
Bergmann gliosis. The precise relationship between these events
is currently unclear, but an early activation of astrocytes prior to
the onset of neuron loss is a consistent feature of multiple forms
of NCL.24–26 We show that glial cells accumulate lysosomal storage material like other cells, so it is conceivable that Purkinje
cell degeneration may be a secondary result of Bergmann glial
dysfunction.
LINCL patient brain tissues exhibit extensive autofluorescent
storage and reactive gliosis in the motor cortex and cerebellum,
similar to our observations in TPP1-deficient mice. However, cellular degeneration was pronounced in human LINCL tissue, with
profound loss of neurons in the motor cortex, cerebellar atrophy, and a dramatic loss of cerebellar Purkinje cells. Consistent
with findings in other forms of NCL,27 the few persisting cortical
neurons were frequently encircled by GFAP-positive processes,
a characteristic appearance not seen in other lysosomal storage disorders. Within the human LINCL cerebellum, there was
also marked astrogliosis. This was accompanied by Bergmann
gliosis that involved the entire molecular layer rather than the
653
© The American Society of Gene Therapy
Enzyme Replacement for LINCL
a
b
−/− ERT
+/−
d
Histological grade
c
−/−
3
2
1
0
+/−
−/− Mock −/− ERT
Figure 6 Effect of enzyme replacement on survival of neurons in
deep cerebellar nuclei (DCNs). (a–c) Sagittal cerebellar sections immunostained for Neu-N from (a) CLN2+/–, (b) enzyme-treated CLN2–/–,
and (c) mock-treated CLN2–/– mice. (d) Scatter plot of results from a
blinded neuropathological grading of DCNs (1 represents few remaining
intact neurons; 2 represents intermediate degree of neuron loss; 3 represents no discernable neuron loss). Scale bars for panels a–c is 500 µm.
ERT, enzyme replacement therapy.
­ iscrete patches seen in TPP1-deficient mice. As expected, neud
ronal degeneration in human LINCL cerebellum was accompanied by loss of axons in cerebellar white matter tracts. Although
it is unclear why severe cell loss was not observed in the TPP1d­eficient mouse, the extensive autofluorescence and reactive gliosis observed in both mouse and human LINCL tissues suggest
those may be relevant neuropathological features to monitor in
a preclinical study.
Enzyme replacement: Patients afflicted with LINCL suffer and
ultimately succumb to symptoms related to CNS dysfunction.1 As
such, it is reasonable to postulate that effective therapies for LINCL
will require CNS targeting. Furthermore, the neurological deficits
do not appear to be localized to a particular CNS structure, region,
or system,4–6 suggesting that a large percentage of the CNS volume
must be treated. We tested the hypothesis that the cerebrospinal
fluid can be exploited to distribute the recombinant TPP1 enzyme
globally within the brain. Most lysosomal hydrolases, including TPP1, are known to acquire mannose-6-phosphate residues
as they traffic through the golgi apparatus after translation. This
post-translational modification mediates the direct intracellular
targeting of lysosomal proteins to the lysosome, but importantly,
also allows mannose-6-phosphate-­mediated endocytosis and
lysosomal targeting of extracellular proteins. The purified recombinant TPP1 proenzyme used in our studies contains mannose6-phosphate and can be readily endocytosed by cultured cells.28
Importantly, TPP1 proenzyme is known to be extremely stable,
and once endocytosed, still has an half-life of 11–12 days.28
After TPP1 infusion into the lateral ventricle, immunostaining for TPP1 in brain sections shows bilateral enzyme distribution to widespread regions of the brain. In vitro TPP1 activity
assays show that TPP1 activity in treated mice ranged from
2-fold to 18-fold that of heterozygous mice throughout the
entire rostral–caudal extent of the brain, which is comparable
to results obtained from similar studies performed in a canine
654
model of ­ mucopolysaccharidose I.21 Importantly, heterozygous
enzyme activity levels are sufficient to prevent disease, and
LINCL leukocytes and fibroblasts typically exhibit undetectable
levels of TPP1 activity.29 Western blots of brain homogenate from
enzyme-treated mice show an accumulation of the mature TPP1,
indicating that the proenzyme was appropriately trafficked to,
and processed in, the lysosomal compartment.
Intravenous ERT has proven to be beneficial for numerous
lysosomal storage disorders that affect primarily peripheral tissues,11–16 indicating that exogenous enzymes can be functional
upon uptake and sorting to the lysosome. We first demonstrated
that recombinant TPP1 can be taken up into the brain parenchyma from the cerebrospinal fluid. We next tested the effects
of intraventricular ERT on the behavioral and neuropathological phenotypes in the TPP1-deficient mouse. Although both
treated and mock-treated mice had tremor amplitudes greater
than that of heterozygous controls, ERT-treated mice exhibited
attenuated tremor amplitudes relative to mock-treated mice,
suggesting that TPP1 replacement may be slowing disease progression. Enzyme-treated mice also showed improvements in
neuropathological findings, including decreased gliosis in the
motor cortex, decreased autofluorescence, and a partial rescue
of the DCNs. These findings are consistent with that of previous enzyme replacement studies in that the exogenous enzyme
can be functionally utilized by diseased cells. In this study,
TPP1 delivery was confined to a 2-week period due to limitations of the pump used. Nevertheless, the significant phenotypic
improvements after such a short duration of enzyme replacement provide encouragement that this approach may be more
efficacious if the enzyme delivery can be maintained or readministered. Our study was also limited by the chronic implantation
of the intraventricular cannula, which resulted in the development of a glial scar within the brain, and predisposed the TPP1deficient mice (both enzyme-treated and mock-treated) to fatal
seizures ~100–110 days of age (data not shown). Heterozygous
mice receiving the implant were never observed to have seizures. Although it is interesting that TPP1-deficient mice are
more susceptible to seizures than heterozygous littermates, this
confounding ­factor precluded the use of survival as an outcome
measure in our study. Similar studies in large animal models can
be done with intrathecal cannulas that do not require passage
through the brain parenchyma, thus minimizing the risk of seizures due to glial scar formation.
An important concern accompanying the administration of
any exogenous protein to patients is the possibility of immune
responses against the protein. In this study, TPP1-deficient mice
appeared to tolerate the enzyme replacement well, with no immediately observable responses to enzyme delivery. Anti-TPP1
antibodies were generated in response to the enzyme (data not
shown), but this is not unexpected given that enzyme replacement studies in Gaucher, Fabry, and mucopolysaccharidoses I,
II, and VI all reported development of immunoglobulin G antibodies against the administered enzyme even when patients
retained residual enzyme activity. In some cases these antibodies
can neutralize the activity of the administered enzyme.14,15,30,31
However, tolerization regimens to overcome this limitation exist
and continue to be developed and improved.32,33
www.moleculartherapy.org vol. 16 no. 4 april 2008
© The American Society of Gene Therapy
Evidence from this and our prior work34 suggests that the cerebrospinal fluid can be exploited for delivery and distribution of
lysosomal enzymes. The results of the current study suggest that
TPP1 can also be delivered via this route, and they warrant additional studies in large animal models. Interestingly, TPP1 activity can be detected in serum following infusion into the lateral
ventricles, suggesting that central administration of TPP1 may
also indirectly rescue enzyme activity in the periphery. Despite
the apparently CNS-specific clinical symptoms in NCL, lysosomal
storage is also observed in peripheral tissues, suggesting that
peripheral enzyme replacement may be necessary if CNS disease
is rescued or delayed.
Materials And Methods
Experimental animals. CLN2–/– mice have been described previously,7
and were backcrossed with C57BL/6 mice at the University of Iowa animal facility. All animal maintenance conditions and experimental protocols were approved by the University of Iowa Animal Care and Use
Committee.
Human tissue specimens. Paraffin-embedded tissue samples were
obtained from the archives of (i) Human Brain and Spinal Fluid Resource
Center; (ii) NICHD Brain and Tissue Bank for Developmental Disorders;
(iii) The MRC London Neurodegenerative Diseases Brain Bank, Institute
of Psychiatry, King’s College London. These specimens were obtained
at routine autopsies of patients with a diagnosis of LINCL confirmed
by mutational analysis, TPP1 enzyme assays or ultrastructural criteria
(n = 4; mean age at death 11.5 years) or controls (n = 3; aged 2, 25, and
26 years) with informed written consent from their families. At autopsy,
tissues were fixed immediately by immersion in 4% neutral-buffered formaldehyde, routinely embedded in paraffin wax, and then cut at 7 µm on a
sledge microtome (Leica Microsystems, Welwyn Garden City, UK). 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).
Recombinant TPP1 expression and purification. Chinese hamster ovary
cells expressing human TPP1 described previously28 were seeded in
850-cm2 roller bottles in 200 ml phenol red–free Dulbecco’s modified
Eagle’s medium/F12 (Gibco, Carlsbad, CA, cat. no. 11039-021) containing 10% fetal bovine serum. Once confluent, the medium was changed
to Dulbecco’s modified Eagle’s medium/F12 without fetal bovine serum.
After 7 days of incubation, the conditioned medium was centrifuged to
remove cells and sterile-filtered through a 0.2-µm filter.
The purification process was done as described previously,28 with
minor modifications. Conditioned medium was adjusted to 20 mmol/l Tris
pH 8.0 and 2.4 mol/l NaCl, then refiltered. Seventy five column volumes
(CVs) of the adjusted medium were loaded at 300 cm/hour on a Butyl
Sepharose 4 FF column pre-equilibrated in 20 mmol/l Tris, 2.4 mol/l
NaCl, pH 8.0. The column was washed with 3 CV 20 mmol/l Tris, 2.4 mol/l
glycine, pH 8.0 and eluted with an 8 CV gradient of 2.4–0 mol/l glycine in
20 mmol/l Tris pH 8.0. Fractions containing TPP1 were pooled, adjusted
to 40 mmol/l NaCl, then loaded at 300 cm/hour on a 12-ml UnoQ12
column pre-equilibrated in 20 mmol/l Tris, 40 mmol/l NaCl, pH 8.0. The
column was then washed with 3 CV equilibration buffer and eluted by
using an 11 CV gradient of 40–550 mmol/l NaCl in 20 mmol/l Tris pH 8.0.
Fractions containing TPP1 were pooled for maximal purity and recovery.
This pool was diluted with an equal volume of 20 mmol/l Tris, 40 mmol/l
NaCl, pH 8.0 and applied to a Blue Sepharose column equilibrated in the
same buffer. The unbound TPP1 was concentrated and buffer exchanged to
phosphate-buffered saline pH 7.2 in a stirred cell using a 10,000 molecular
weight cut-off membrane. The final purified TPP1 was sterile-filtered and
stored frozen at –80 °C until use.
Molecular Therapy vol. 16 no. 4 april 2008
Enzyme Replacement for LINCL
Surgical implantation of osmotic pumps. CLN2–/– mice (60 days old) were
anesthetized with ketamine (100–125 mg/kg) and xylazine (10–12.5 mg/
kg). Cannulas, (0.31-mm diameter, Brain Infusion Kit III; Alzet Cupertino,
CA) were stereotaxically implanted into the lateral ventricles (A/P –0.4mm caudal, M/L 1.0 mm, D/V –2.0 mm). The cannula was connected via
a catheter tube to an osmotic pump (Model 1002, Alzet, Cupertino, CA),
which was subcutaneously implanted in the left flank of each mouse. Each
pump was filled with 100 µl purified TPP1 proenzyme (1 mg/ml). Pumps
were excised at the end of the study to confirm delivery of the enzyme.
Quantitation of tremor in TPP1-deficient mice. Mean tremor amplitude
was assessed with a tremor monitor (SD Instruments, San Diego, CA).
Enzyme-treated CLN2–/– mice, saline mock-treated CLN2–/– mice, and
age-matched CLN2+/– control littermates (n = 5–11 per group) were each
allowed a 1-minute habituation period on the testing platform, after which
20 samples of tremor amplitude were taken per mouse. Samples were taken
every 5 seconds, and each sample consisted of the mean amplitude over a
1-second interval. Mice were assayed every 5 days beginning at 55 days
of age until kill at 105 days of age. Statistical difference between groups at
each timepoint was calculated by students t-test.
Mouse tissue processing and histology. Animals were deeply anesthetized
with intraperitoneal ketamine (100–125 mg/kg) and xylazine (10–12.5 mg/kg),
then transcardially perfused with normal saline (20 ml) followed by 4% paraformaldehyde in normal saline (20 ml). Brains were then extracted and postfixed in 4% paraformaldehyde for 24 hours at 4 °C. 40 µmol/l thick sections
were cut on a freezing microtome and collected free floating in cryoprotectant solution (30% ethylene glycol, 15% sucrose, 0.05% sodium azide, in Trisbuffered saline) for storage at –20 °C. Immunohistochemistry was done with
the following antibodies: mouse-anti-GFAP (Sigma, St. Louis, MO; 1:500),
mouse anti-Neu-N (Chemicon, Temecula, CA; 1:100), mouse anti-calbindin
D-28K (Sigma, St. Louis, MO; 1:3,000). All antibodies were diluted in Trisbuffered saline with 2% bovine serum albumin, 0.1% NaN3, and 0.05% Tween
20. After incubation with primary antibody overnight at 4 °C, tissue sections
were rinsed with Tris-buffered saline and incubated with the appropriate secondary antibody:biotinylated goat anti-mouse (Jackson, Immuno Research,
West Grove, PA; 1:5,000), Alexa 488–­conjugated goat anti-rabbit (Invitrogen,
Carlsbad, CA; 1:1,000). Stains were developed with 3,3′-­diaminobenzidine
peroxidase substrate kit (Vector Laboratories, Burlingame, CA).
Human tissue processing and histology. As described previously, paraffin
wax sections of human tissue sections were stained to reveal the presence of
GFAP immunoreactive astrocytes (polyclonal rabbit anti-GFAP, 1:10,000;
Dako Glostrup, Denmark), and counterstained with hematoxylin to reveal
neuronal cytoarchitecture.27 Multichannel confocal microscopy (Leica SP5,
Leica Microsystems, UK) was used to assess the distribution of autofluorescent storage material in tissue that had been immunofluorescently
stained to reveal GFAP immunoreactive astrocytes (polyclonal rabbit antiGFAP, followed by Alexa 488–labeled goat anti-rabbit immunoglobulin
G, 1:500; Molecular Probes, Eugene, OR). A standard Bielschowsky silverstain method was used to assess axonal degeneration.
Thresholding analysis with Image J software. Digital images (×40 magni-
fication) of immunostained brain sections were taken (Images from three
sections per mouse per region of interest, n = 3 mice per group in Figure 1,
n = 5–9 mice per group in Figure 5). Once opened in Image J, images were
converted to 8-bit grayscale images. The thresholding function was then
used to set a black and white threshold corresponding to the immunostain.
Once a threshold is set, the “Analyze Particles” function was used to sum
up the total area of positive staining and calculate the fraction of the total
area that is positive for the stain.
TPP1 enzyme activity assay in mouse brain homogenate. TPP1 activity
was assayed via a modification of the method described previously.29 Mice
were transcardially perfused with ice-cold normal saline, and extracted
655
Enzyme Replacement for LINCL
brains were cut into 10 coronal sections through both hemispheres using a
brain matrix. Brain sections were homogenized with a laboratory homogenizer (P200; Pro Scientific, Oxford, CT) in 300 µl ice-cold homogenization
buffer—0.1% Triton X-100 in normal saline, with complete protease inhibitor cocktail (Roche, Mannheim, Germany). Insoluble material was removed
from the homogenate by centrifugation at 21 × 103 rcf at 4 °C × 15 minutes,
and protein content in the supernatant was quantified by DC Protein Assay
(Biorad, Hercules, CA). Protein (10 µg) was added to wells of a 96-well
black wall plate containing 50 µl sodium citrate buffer (pH 4.0). The enzyme
reaction was then initiated by addition of substrate (250 µmol/l Ala-AlaPhe 7-amido-4-methylcoumarin in sodium citrate buffer, pH 4.0). Plates
were then read in a FLUOstar microplate reader (BMG Labtechnologies,
Durham, NC) at 37 °C with a 460-nm emission filter.
Acknowledgments
The authors thank Steve Moore (University of Iowa), and Ming J Lim,
Noreen Alexander, and Andrew Wong (Institute of Psychiatry), for
­expert assistance with pathology and human LINCL tissue analysis.
This work was supported by NS037918 (to P.L.), HD 33531, and NS
34568 (to B.L.D.), F31 NS054462 (to M.C.), and the Batten Disease
Support and Research Association.
References
1. Goebel, HH and Wisniewski, KE (2004). Current state of clinical and morphological
features in human NCL. Brain Pathol 14: 61–69.
2. Rider, JA and Rider, DL (1988). Batten disease: past, present, and future. Am J Med
Genet Suppl 5: 21–26.
3. Sleat, DE, Donnelly, RJ, Lackland, H, Liu, CG, Sohar, I, Pullarkat, RK et al. (1997).
Association of mutations in a lysosomal protein with classical late-infantile neuronal
ceroid lipofuscinosis. Science 277: 1802–1805.
4. Autti, T, Raininko, R, Santavuori, P, Vanhanen, SL, Poutanen, VP and Haltia, M (1997).
MRI of neuronal ceroid lipofuscinosis. II. Postmortem MRI and histopathological
study of the brain in 16 cases of neuronal ceroid lipofuscinosis of juvenile or late
infantile type. Neuroradiology 39: 371–377.
5. Lee, CW, Bang, H, Oh, YG, Yun, HS, Kim, JD and Coe, CJ (2003). A case of late
infantile neuronal ceroid lipofuscinosis. Yonsei Med J 44: 331–335.
6. Petersen, B, Handwerker, M and Huppertz, HI (1996). Neuroradiological
findings in classical late infantile neuronal ceroid-lipofuscinosis. Pediatr Neurol
15: 344–347.
7. Sleat, DE, Wiseman, JA, El-Banna, M, Kim, KH, Mao, Q, Price, S et al. (2004).
A mouse model of classical late-infantile neuronal ceroid lipofuscinosis based on
targeted disruption of the CLN2 gene results in a loss of tripeptidyl-peptidase I
activity and progressive neurodegeneration. J Neurosci 24: 9117–9126.
8. Cabrera-Salazar, MA, Roskelley, EM, Bu, J, Hodges, BL, Yew, N, Dodge, JC et al.
(2007). Timing of therapeutic intervention determines functional and survival
outcomes in a mouse model of late infantile batten disease. Mol Ther
15: 1782–1788.
9. Passini, MA, Dodge, JC, Bu, J, Yang, W, Zhao, Q, Sondhi, D et al. (2006). Intracranial
delivery of CLN2 reduces brain pathology in a mouse model of classical late infantile
neuronal ceroid lipofuscinosis. J Neurosci 26: 1334–1342.
10. Sondhi, D, Hackett, NR, Peterson, DA, Stratton, J, Baad, M, Travis, KM et al. (2007).
Enhanced survival of the LINCL mouse following CLN2 gene transfer using
the rh.10 rhesus macaque-derived adeno-associated virus vector. Mol Ther
15: 481–491.
11. Amalfitano, A, Bengur, AR, Morse, RP, Majure, JM, Case, LE, Veerling, DL et al. (2001).
Recombinant human acid α-glucosidase enzyme therapy for infantile glycogen
storage disease type II: results of a phase I/II clinical trial. Genet Med 3: 132–138.
12. Barton, NW, Brady, RO, Dambrosia, JM, Di Bisceglie, AM, Doppelt, SH, Hill, SC et al.
(1991). Replacement therapy for inherited enzyme deficiency—macrophage-targeted
glucocerebrosidase for Gaucher’s disease. N Engl J Med 324: 1464–1470.
656
© The American Society of Gene Therapy
13. Barton, NW, Furbish, FS, Murray, GJ, Garfield, M and Brady, RO (1990).
Therapeutic response to intravenous infusions of glucocerebrosidase in a patient
with Gaucher disease. Proc Natl Acad Sci USA 87: 1913–1916.
14. Harmatz, P, Whitley, CB, Waber, L, Pais, R, Steiner, R, Plecko, B et al. (2004).
Enzyme replacement therapy in mucopolysaccharidosis VI (Maroteaux-Lamy
syndrome). J Pediatr 144: 574–580.
15. Kakkis, ED, Muenzer, J, Tiller, GE, Waber, L, Belmont, J, Passage, M et al. (2001).
Enzyme-replacement therapy in mucopolysaccharidosis I. N Engl J Med
344: 182–188.
16. Schiffmann, R, Kopp, JB, Austin, HA 3rd, Sabnis, S, Moore, DF, Weibel, T et al. (2001).
Enzyme replacement therapy in Fabry disease: a randomized controlled trial. JAMA
285: 2743–2749.
17. Barranger, JA, Rapoport, SI, Fredericks, WR, Pentchev, PG, MacDermot, KD,
Steusing, JK et al. (1979). Modification of the blood-brain barrier: increased
concentration and fate of enzymes entering the brain. Proc Natl Acad Sci USA
76: 481–485.
18. Neuwelt, EA, Barranger, JA, Brady, RO, Pagel, M, Furbish, FS, Quirk, JM et al. (1981).
Delivery of hexosaminidase A to the cerebrum after osmotic modification of the
blood-brain barrier. Proc Natl Acad Sci USA 78: 5838–5841.
19. Young, PP, Fantz, CR and Sands, MS (2004). VEGF disrupts the neonatal blood-brain
barrier and increases life span after non-ablative BMT in a murine model of
congenital neurodegeneration caused by a lysosomal enzyme deficiency. Exp Neurol
188: 104–114.
20. Lonser, RR, Walbridge, S, Murray, GJ, Aizenberg, MR, Vortmeyer, AO, Aerts, JM et al.
(2005). Convection perfusion of glucocerebrosidase for neuronopathic Gaucher’s
disease. Ann Neurol 57: 542–548.
21. Kakkis, E, McEntee, M, Vogler, C, Le, S, Levy, B, Belichenko, P et al. (2004).
Intrathecal enzyme replacement therapy reduces lysosomal storage in the brain
and meninges of the canine model of MPS I. Mol Genet Metab 83: 163–174.
22. Rasband, WS. Image J. NUS National Institues of Health 1997–2007.
23. Dickson, P, McEntee, M, Vogler, C, Le, S, Levy, B, Peinovich, M et al. (2007).
Intrathecal enzyme replacement therapy: successful treatment of brain disease via the
cerebrospinal fluid. Mol Genet Metab 91: 61–68.
24. Kielar, C, Maddox, L, Bible, E, Pontikis, CC, Macauley, SL, Griffey, MA et al. (2007).
Successive neuron loss in the thalamus and cortex in a mouse model of infantile
neuronal ceroid lipofuscinosis. Neurobiol Dis 25: 150–162.
25. Oswald, MJ, Palmer, DN, Kay, GW, Shemilt, SJ, Rezaie, P and Cooper, JD (2005).
Glial activation spreads from specific cerebral foci and precedes neurodegeneration
in presymptomatic ovine neuronal ceroid lipofuscinosis (CLN6). Neurobiol Dis
20: 49–63.
26. Pontikis, CC, Cella, CV, Parihar, N, Lim, MJ, Chakrabarti, S, Mitchison, HM et al.
(2004). Late onset neurodegeneration in the Cln3–/– mouse model of juvenile
neuronal ceroid lipofuscinosis is preceded by low level glial activation. Brain Res
1023: 231–242.
27. Tyynelä, J, Cooper, JD, Khan, MN, Shemilts, SJ and Haltia, M (2004).
Hippocampal pathology in the human neuronal ceroid-lipofuscinoses: distinct
patterns of storage deposition, neurodegeneration and glial activation. Brain Pathol
14: 349–357.
28. Lin, L and Lobel, P (2001). Production and characterization of recombinant human
CLN2 protein for enzyme-replacement therapy in late infantile neuronal ceroid
lipofuscinosis. Biochem J 357: 49–55.
29. Sohar, I, Lin, L and Lobel, P (2000). Enzyme-based diagnosis of classical late
infantile neuronal ceroid lipofuscinosis: comparison of tripeptidyl peptidase I and
pepstatin-insensitive protease assays. Clin Chem 46: 1005–1008.
30. Linthorst, GE, Hollak, CE, Donker-Koopman, WE, Strijland, A and Aerts, JM (2004).
Enzyme therapy for Fabry disease: neutralizing antibodies toward agalsidase α and β.
Kidney Int 66: 1589–1595.
31. Richards, SM, Olson, TA and McPherson, JM (1993). Antibody response in
patients with Gaucher disease after repeated infusion with macrophage-targeted
glucocerebrosidase. Blood 82: 1402–1409.
32. Kakkis, E, Lester, T, Yang, R, Tanaka, C, Anand, V, Lemontt, J et al. (2004).
Successful induction of immune tolerance to enzyme replacement therapy in
canine mucopolysaccharidosis I. Proc Natl Acad Sci USA 101: 829–834.
33. Raben, N, Nagaraju, K, Lee, A, Lu, N, Rivera, Y, Jatkar, T et al. (2003) Induction of
tolerance to a recombinant human enzyme, acid α-glucosidase, in enzyme deficient
knockout mice. Transgenic Res 12: 171–178.
34. Liu, G, Martins, I, Wemmie, JA, Chiorini, JA and Davidson, BL (2005).
Functional correction of CNS phenotypes in a lysosomal storage disease model
using adeno-associated virus type 4 vectors. J Neurosci 25: 9321–9327.
www.moleculartherapy.org vol. 16 no. 4 april 2008