Download Cln3 Targeted Disruption of the Gene Provides a Mouse Model for Batten Disease

Neurobiology of Disease 6, 321–334 (1999)
Article ID nbdi.1999.0267, available online at http://www.idealibrary.com on
Targeted Disruption of the Cln3 Gene Provides
a Mouse Model for Batten Disease
Hannah M. Mitchison,* David J. Bernard,†
Nicholas D. E. Greene,* Jonathan D. Cooper,‡
Mohammed A. Junaid,§ Raju K. Pullarkat,§ Nanneke de Vos,¶
Martijn H. Breuning,¶ Jennie W. Owens,\ William C. Mobley,‡
R. Mark Gardiner,* Brian D. Lake,** Peter E. M. Taschner,¶
and Robert L. Nussbaum†
*Department of Paediatrics, Royal Free and University College London Medical School, Rayne
Institute, University Street, London, WC1E 6JJ, United Kingdom; †Genetic Disease Research
Branch, National Human Genome Research Institute, National Institutes of Health, Bethesda,
Maryland 20892-4472; ‡Department of Neurology and Neurological Sciences and the Program
in Neuroscience, Stanford University, Stanford, California 94305-5489; §New York State
Institute for Basic Research in Developmental Disabilities, Staten Island 10314, New York;
¶Department of Genetics, Section of Human Genetics, Sylvius Laboratories, Leiden University,
P.O. Box 9503, 2300 RA Leiden, The Netherlands; \Veterinary Resource Program, Office of
Research Services, National Institutes of Health, Bethesda, Maryland 20892-4472;
**Department of Histopathology, Great Ormond Street Hospital for Children, London
WC1N 3JH, United Kingdom
Received August 23, 1999; revised September 15, 1999; accepted for publication September 15, 1999
Batten disease, a degenerative neurological disorder with juvenile onset, is the most common form of
the neuronal ceroid lipofuscinoses. Mutations in the CLN3 gene cause Batten disease. To facilitate
studies of Batten disease pathogenesis and treatment, a murine model was created by targeted
disruption of the Cln3 gene. Mice homozygous for the disrupted Cln3 allele had a neuronal storage
disorder resembling that seen in Batten disease patients: there was widespread and progressive
intracellular accumulation of autofluorescent material that by EM displayed a multilamellar rectilinear/
fingerprint appearance. Inclusions contained subunit c of mitochondrial ATP synthase. Mutant animals
also showed neuropathological abnormalities with loss of certain cortical interneurons and hypertrophy of many interneuron populations in the hippocampus. Finally, as is true in Batten disease patients,
there was increased activity in the brain of the lysosomal protease Cln2/TPP-1. Our findings are
evidence that the Cln3-deficient mouse provides a valuable model for studying Batten disease. r 1999
Academic Press
Goebel et al., 1999). Widespread loss of neurons in the
neural retina and central nervous system (CNS), predominantly the cortex and cerebellum, account for the
neurological symptoms. No other organs appear to be
clinically affected. The juvenile onset form of NCL,
referred to as Batten disease (JNCL), is caused by
mutations in the CLN3 gene (International Batten
Disease Consortium, 1995). The CLN3 gene encodes a
438 amino acid predicted transmembrane protein (International Batten Disease Consortium, 1995; Janes et
INTRODUCTION
The neuronal ceroid lipofuscinoses (NCL) are a
group of autosomal recessive disorders that comprise
the most common neurodegenerative diseases of childhood, with an incidence of up to 1:12,500 (Santavuori,
1998). The clinical course is progressive and is marked
by blindness, seizures, psychomotor deficits and dementia leading to a vegetative state and death in early
adulthood (reviewed in Santavuori, 1998; Rapola, 1993;
0969-9961/99 $30.00
Copyright r 1999 by Academic Press
All rights of reproduction in any form reserved.
321
322
al., 1996) whose function is unknown. On the basis of
structural predictions, it has been suggested that CLN3
could be a transporter (Pearce et al., 1999), and there is
potential functional homology to nucleotide-sugar
transporters at residues 193–214.
The pathogenesis of Batten disease is unknown, but
a compelling case can be made for the involvement of
lysosomes. First, a key diagnostic feature of Batten
disease is the intracellular accumulation of autofluorescent material within lysosomes. Stored material is
present in neurons as well as in other cell types
(Rapola, 1993; Goebel et al., 1999; Lake, 1997). This
material resembles lipofuscin, a substance found in
normal aging brains of humans and mice (Mann et al.,
1978; Sekhon & Maxwell, 1974). However, the deposits
in JNCL are unique in that they have a characteristic
multilamellar ‘‘fingerprint’’ ultrastructure and contain
subunit c of the mitochondrial ATP synthase as their
major protein component (Hall et al., 1991; Palmer et
al., 1992). Second, potential dileucine and tyrosinebased (residues 52–55) lysosomal targeting signals are
present in the CLN3 protein (White et al., 1998; Hunziker & Fumey, 1999). The presence of such signals is
consistent with the recently reported localization of
CLN3 in lysosomes (Järvelä et al., 1998, 1999). Third,
the clinical features and composition of storage material in Batten disease are similar to another disorder,
classical late infantile NCL (CLN2), which is known to
be caused by mutations in a lysosomal pepstatininsensitive tripeptidyl peptidase I (CLN2 protease/
TPP-1) (Sleat et al., 1997; Vines & Warburton, 1999).
Fourth, the brains of Batten disease patients show a
significant increase in the level of glycoproteins containing mannose-6-phosphate, a modification that directs
newly synthesised proteins to the lysosome (Sleat et al.,
1998; Junaid & Pullarkat, 1999). This increase in the
level of lysosomal proteins is associated with a general
elevation in the activity of lysosomal enzymes, which
may reflect increased synthesis, a common feature of
other lysosomal storage diseases in which one enzyme
is defective (Burditt et al., 1980; Young et al., 1997).
Finally, treatment with inhibitors of lysosomal proteases recreates certain features of the pathology of
Batten disease, resulting in accumulation of lipofuscinlike dense bodies in neurons (Ivy et al., 1984). This
evidence suggests that dysfunction of one or more
aspects of lysosomal biology contributes to Batten
disease pathogenesis.
The development of an animal model that recapitulates the clinical and pathological features of Batten
Copyright r 1999 by Academic Press
All rights of reproduction in any form reserved.
Mitchison et al.
disease is an important step toward elucidating underlying disease mechanism and testing potential treatment strategies. Although NCLs occur naturally in a
range of animal species (Goebel et al., 1999; Jolly, 1995),
there is no model in which the disease is caused by
mutations in CLN3. By targeted disruption of the
mouse Cln3 gene in embryonic stem cells, we generated mice with a Cln3 null allele. Progressive accumulation of autofluorescent material with the staining and
ultrastructural characteristics of the material stored in
Batten disease patients was detected in mice homozygous for the disrupted Cln3 allele. In addition, there
were neuropathological abnormalities involving the
cortex and hippocampus. Cln2 protease activity in the
brain was significantly elevated, implying that a correlation with human disease also exists at the biochemical level. We conclude that because Cln3-deficient mice
recapitulate several neuropathological features seen in
human patients, they provide a model for the study of
Batten disease.
METHODS
Construction of Targeting Vector
The Cln3-targeting vector was constructed as outlined previously (Greene et al., 1999). The insert from
mouse Cln3 cDNA clone pMNCL (Taschner et al., 1997)
was tested by Southern blot hybridization to verify it
was single copy sequence. This fragment was used to
probe a 129/Sv mouse genomic cosmid library (van
Ree et al., 1994). One cosmid, mcos6, was found to
contain the complete Cln3 gene by Southern hybridization, using CLN3 cDNA fragments as probes, by PCR
using exon-specific primers, and by partial sequence
analysis. A 9-kb genomic HindIII fragment containing
exons 1–8 of Cln3 was isolated from mcos6 and cloned
into the pBluescript II SK⫹ HindIII site (Stratagene,
CA). A SacII–EcoRI fragment of this clone containing
Cln3 exon 1 (including the start codon) and exons 2–6
was replaced with a 1.9-kb HindIII fragment containing a PGK promoter–neomycin resistance–PGK-polyA
signal cassette cloned in the reverse orientation. This
was achieved by end-filling the SacII, EcoRI, and
HindIII sites, followed by blunt-end ligation. A 2740-bp
cassette containing the herpesvirus thymidine kinase
gene under control of the PGK promoter was digested
from the pPNT vector (Tybulewicz et al., 1991) with
EcoRI and HindIII and endfilled. This fragment was
inserted at an endfilled HindIII site at the 58 end of the
9-kb Cln3 fragment for negative selection. An XhoI site
323
A Mouse Model for Batten Disease
in the pBluescript polylinker creates a unique vector
linearization site. The targeting vector comprised 4 kb
of genomic sequence upstream and 2 kb downstream
of the PGKneo cassette.
Gene Targeting in Embryonic Stem Cells
The targeting construct was linearized with XhoI
and electroporated into 129/Sv TC1 ES cells as
previously described (Deng et al., 1996). Targeting
was identified in two of 140 G418- and FIAUresistant colonies by EcoRV Southern blot analysis
using a probe located outside the targeted region of
homologous recombination (probe C). Positive cell
lines were confirmed by EcoRV genomic Southern blot
using external probe A and internal probe B (see
Fig. 1).
Generation of Cln3 Null Mice
Two correctly targeted ES cell lines were used to
generate chimeras by injection into C57BL/6 blastocysts followed by transfer into pseudopregnant recipient NIH Black Swiss females, using standard techniques (Hogan et al., 1994). One ES cell line produced
chimeras that showed germline transmission as indicated by the agouti coat color of offspring following
breeding to NIH Black Swiss females. F1 offspring
were genotyped by Southern analysis and heterozygotes were interbred to generate Cln3⫺/⫺ homozygotes. All animal procedures were approved by the
National Human Genome Research Institute’s Animal
Care and Use Committee under National Institutes of
Health guidelines, ‘‘Using Animals in Intramural Research.’’
RT-PCR
Brain and kidney samples were collected into liquid
nitrogen and total RNA isolated with RNAzol (Biogenesis Ltd., UK). RT-PCR of total RNA was performed using the Superscript kit (Life Technologies,
Inc.), according to the manufacturer’s instructions
with Neomycin primers 1–58 (58-AGAGGCTATTCGGCTATGACTG-38) and 2–38 (58-TTCGTCCAGATCATCCTGATC-38) or Cln3 primers M7F
(58-ACCCACCTGTCCCAGACTTT-38) and M6R (58CACTCCGACTATCCAACCGA-38). M7F and M6R amplify within the coding region of Cln3, across an intron,
such that DNA contamination can be excluded based
on size difference of the relevant PCR products.
Autofluorescent and Immunohistochemical
Analysis
Brains were collected and placed in 20 vol of
10% phosphate-buffered formalin, embedded in paraffin wax, and sectioned by standard methods. For
detection of autofluorescence, 5-µm sections were
mounted on slides, air-dried, coverslipped with
DPX-mounting medium (BDH), and then photographed under epi-illumination with broad excitation
and a barrier filter at 410 nm. For confocal microscopy,
40-µm unstained sections through the hippocampal
formation were prepared, as described previously
(Cooper et al., 1999), prior to examination with a Nikon
PCM 2000 confocal microscope. For each region of
interest a series of 35 images were collected at 0.5-µm
intervals on the z-axis and combined to form a single
image. Five-micrometer sections were also stained
with Luxol fast blue and Sudan black by standard
methods (Luna, 1992), or with antisera to subunit c of
the mitochondrial ATP synthase, using a horseradish
peroxidase detection system (Elleder et al., 1997), and
were examined and photographed under light microscopy.
Electron Microscopy
Anesthetized mice were fixed by perfusion with 2%
paraformaldehyde/2.5% glutaraldehyde, and hippocampus and brain stem samples were dissected. These
were diced into 2-mm blocks, fixed overnight in
perfusate, and postfixed in 1% osmium tetroxide.
Tissues were dehydrated through alcohol and propylene oxide and embedded in Eponate 12 resin (Ted Pella
Inc.). Thin sections were prepared with an ultramicrotome (Leica Ultracut-uct), contrasted with uranyl acetate and lead citrate and examined using a Philips 201
electron microscope.
Analysis of Neuronal Populations
Mice of either sex (n ⫽ 3 of each genotype) were
anesthetized with pentobarbitone and transcardially
perfused with vascular rinse followed by a filtered
solution of 4% paraformaldehyde, 0.2% picric acid in
100 mM phosphate buffer, pH 7.4. The mass of each
perfused brain was determined at the time of removal.
After postfixation and cryoprotection, brains were
frozen on dry ice and 40-µm coronal cryosections cut.
Sections were processed for Nissl or immunohistochemical staining to reveal the presence of interneu-
Copyright r 1999 by Academic Press
All rights of reproduction in any form reserved.
324
rons in the hippocampus and cortex as described
previously (Cooper et al., 1999).
Unbiased Estimates of Regional Volume
Unbiased estimates of the volume of the hippocampus and adjacent cortical areas were made using
Cavalieri’s method (Cavalieri, 1966), by counting the
number of points of a randomly superimposed sampling grid which fell over each structure in a one-in-six
series of Nissl-stained coronal sections as described
previously (Cooper et al., 1999).
Measurements of Detectable Neuronal Number
and Cross-Sectional Area
A one-in-six series of sections through the hippocampal formation and cortical mantle of each brain was
stained to reveal the presence of neurons expressing
parvalbumin (PV), calbindin (Cb), or somatostatin-14
(SOM). Counts were made of detectable PV-expressing
neurons in layer II and IV of the entorhinal cortex and
PV-, Cb-, and SOM-positive neurons in the hippocampal formation. Counts were expressed as the number
of detectable neurons per section and corrected by the
method of Abercrombie (Abercrombie, 1946). Measurements of neuronal cross-sectional area were made in
the same sections. All methods used as described
previously (Cooper et al., 1999).
Cln2 Protease Activity
Brains were collected in liquid nitrogen and frozen
until use. The Cln2 protease enzyme activity was
measured as described previously (Junaid & Pullarkat,
1999; Junaid et al., 1999).
RESULTS
Gene Targeting of Cln3
To inactivate Cln3, exons 2–6 and most of exon 1,
including the start codon, were deleted in TC1 embryonic stem cells (Figs. 1A and 1B) by replacement with a
neomycin resistance gene transcribed in reverse orientation from a mouse PGK promoter, as previously
described (Greene et al., 1999). Chimeras established
from these cells were used to generate both 129/Sv
inbred and 129/Sv ⫻ Black Swiss outbred lines; the
outbred line was used for subsequent analysis (Fig.
1B). Intercrosses between F1 heterozygotes produce
Copyright r 1999 by Academic Press
All rights of reproduction in any form reserved.
Mitchison et al.
viable F2 offspring in normal Mendelian ratios:
wild-type (n ⫽ 27, 25%), heterozygote (n ⫽ 59, 55%),
and homozygous offspring (n ⫽ 22, 20%). RTPCR confirmed the absence of Cln3 transcripts in
homozygous mutant mice (Fig. 1C). Therefore, although the gene is known to exhibit widespread
low-level expression during embryonic development
(N. D. E. Greene and H. M. Mitchison, unpublished
data), Cln3 function is not essential for development of
the embryo. Cln3 ⫺/⫺ mice were viable and fertile
and by 12 months of age did not exhibit obvious
clinical signs.
Cln3 ⴚ/ⴚ Mice Exhibited Accumulation of
Intracellular Storage Material
In spite of the absence of evident clinical disease,
morphological studies showed that Cln3 ⫺/⫺ mice
exhibited several of the features of Batten disease. We
observed a marked accumulation of storage material
in neuronal cell bodies in the brains of Cln3 ⫺/⫺ (Fig.
2). Autofluorescent cytoplasmic inclusions were detected in Cln3 ⫺/⫺ mice at 3 months of age and
analysis up to 12 months of age showed that there was
an increasing amount of material, indicating that the
storage was progressive (Figs. 2A–2D). Storage levels
at 5 months of age have also previously been described
(Greene et al., 1999). Stored material fluoresced over a
wide range of excitatory and barrier filter combinations. It appeared to be typically granular and was
distributed widely throughout the cytoplasm, always
sparing the nucleus. The composition of the stored
material also appeared to be similar to that in human
patients, because it stained positively with the lipid
stains Luxol fast blue (Figs. 2E and 2F) and Sudan black
(data not shown). Significantly, the material stored in
Cln3⫺/⫺ mice did contain subunit c of the mitochondrial
ATP synthase, a major component of storage bodies in
Batten disease patients (Figs. 2G and 2H).
Autofluorescent material was distributed widely in
the brain and was found in both neurons and glial
cells. Storage was particularly prominent in neurons of
the cortex, hippocampus, basal ganglia, and reticular
formation of the brainstem. Within the hippocampus
the distribution was very similar to that seen in the
mnd mouse, which is a model of another form of NCL
(Cooper et al., 1999; Ranta et al., 1999), with prominent
accumulation in interneurons of the hilar formation,
stratum oriens, stratum radiatum, and in interneurons
dispersed among the principal cell layers (CA1-3). A
representative comparison of the hippocampus of a
wild-type and a mutant animal at 7 months of age is
shown in (Figs. 3A and 3B). The superior resolution
A Mouse Model for Batten Disease
325
FIG. 1. Generation of Cln3-deficient mice by homologous recombination. (A) Structure of the wild-type allele, the targeting vector, and the
predicted targeted Cln3 locus. The external and internal probes used for detection of targeted ES cells are indicated below the wild-type allele
(probe A, B, C) and the relevant restriction sites are shown (R1, EcoRI; RV, EcoRV; H, HindIII; S, SacII). Figure drawn to scale. (B) Southern blot of
EcoRV-digested genomic DNA. Homologous recombination results in the reduction of a wild-type 9.5-kb allele (lane 1) to a 5.5-kb allele (lane 2)
in targeted ES cells (probe C). Representative samples from the offspring of an intercross between Cln3 heterozygous mice are included in lanes
3–5. Hybridization with probe B indicates heterozygote (lane 3), homozygous ⫺/⫺ (lane 4), and wild-type (lane 5) genotypes. (C) mRNA
expression analysis of kidney (lanes 1 and 2) and brain (lanes 3 and 4) by RT-PCR indicates absence of Cln3 exons 1 and 2 and expression of the
neo cassette in homozygous Cln3 ⫺/⫺ mice (lanes 2 and 4). Wild-type controls in lanes 1 and 3.
afforded by confocal microscopy permitted visualization of sparse deposits of lipopigment in the cytoplasm
of granule neurons of the dentate gyrus (Fig. 3B).
Heterozygous mice were indistinguishable from wildtype mice (not shown).
Under electron microscopy, storage bodies showed a
multilamellar ultrastructure that is very reminiscent of
the rectilinear/fingerprint profile seen in Batten disease (Rapola, 1993; Goebel et al., 1999; Lake, 1997). The
number of individual storage bodies within cells inCopyright r 1999 by Academic Press
All rights of reproduction in any form reserved.
326
Copyright r 1999 by Academic Press
All rights of reproduction in any form reserved.
Mitchison et al.
A Mouse Model for Batten Disease
327
FIG. 3. Confocal and electron microscopy of storage material. (A, B) Unstained sections of hippocampus from 7-month-old mice visualized by
UV confocal microscopy shows autofluorescent intracellular inclusions in the dentate gyrus (dg) and hilus (hi) in mutant (B) compared to
wild-type (A). Magnification ⫻300. (C, D) Representative images of storage material from 12-month-old mutants visualized by electron
microscopy illustrate the multilamellar electron-dense ultrastructure of the inclusion bodies. Magnification, ⫻22,400 (C) and ⫻33,570 (D).
FIG. 2. Histological and immunochemical analysis of Cln3 ⫺/⫺ mice. (A–D) Under UV light, autofluorescent inclusions (arrows in B, D) are
detected in unstained sections of cortex from homozygous mutant (B, D) mice at 5 (A, B) and 12 (C, D) months of age compared to a low
background autofluorescence in age-matched controls (A, C). Storage material in 8-month-old mutants (F, H) also stains positively with Luxol
fast blue (E, F) and antisera to subunit c of the mitochondrial ATP synthase (G, H), compared to background levels in age-matched controls (E,
G). Representative sections from hippocampus are shown; storage bodies are indicated by arrows. Magnification, ⫻1000 (A–D), ⫻1200 (E–H).
Copyright r 1999 by Academic Press
All rights of reproduction in any form reserved.
328
Mitchison et al.
creased with age and these usually had a compact
multi-layered structure containing a number of more
electron-dense areas (Fig. 3C). Later in the disease
course, larger and less tightly packed bodies were
observed occupying the majority of the cell soma (Fig.
3D). In the wild-type littermate controls, an occasional
deposit of autofluorescent lipofuscin resulted in a
small increase in the background levels of autofluorescence (Also seen in Figs. 2A and 2D). However, these
naturally occurring deposits were simple dense structures in secondary lysosomes (not shown) (Mann et al.,
1978; Sekhon & Maxwell, 1974). They were rare and
readily distinguished from the inclusion bodies seen in
Cln3⫺/⫺ mice.
Disruption of Cln3 Caused Hypertrophy and Loss
of Specific Populations of Interneurons
We examined the brains of Cln3 ⫺/⫺ mice at 7
months of age in order to determine, whether accumulation of autofluorescent material was associated with
other neuropathological changes. There was a small
reduction in total brain mass in mice homozygous for
the disrupted Cln3 allele (Table 1). However, the
change was not statistically significant. Unbiased volumetric analysis of the neocortex revealed a decrease
that approached statistically significance. There was
no reduction in the volume of the hippocampal formation in Cln3 ⫺/⫺ mice (Table 1).
We examined subpopulations of neurons containing
autofluorescent accumulated storage material. Apparent loss and hypertrophy of remaining interneurons
has been reported in the mnd mouse (Cooper et al.,
1999), suggesting that these neuronal populations may
also be susceptible in Cln3 ⫺/⫺ animals. Therefore, we
stained sections through the hippocampal formation
TABLE 1
Genotype
Brain
mass (g)
Cortex volume
Hippocampus
volume
⫹/⫹
⫹/⫺
⫺/⫺
0.51 ⫾ 0.01
0.47 ⫾ 0.01
0.49 ⫾ 0.00
32.19 ⫾ 1.57
30.31 ⫾ 1.35
27.11 ⫾ 0.95*
10.54 ⫾ 0.23
10.35 ⫾ 0.54
10.19 ⫾ 0.57
Note. Brain mass and unbiased estimates of regional volumes
made by stereological point counting through the entire rostrocaudal extent of the CNS. Comparisons between 7-month-old mice
(n ⫽ 3 of each genotype) indicate slight reduction in reference
volume of the neocortical mantle in Cln3⫺/⫺ (*P ⫽ 0.051, wild-type
vs homozygous mutant), while the hippocampal formation is unaffected.
Copyright r 1999 by Academic Press
All rights of reproduction in any form reserved.
and cortex for immunohistochemical markers of interneuronal phenotype. We examined interneuronal populations that stained positive for the calcium-binding
protein parvalbumin (PV) in the entorhinal cortex, and
subpopulations that stained positive for PV or another
calcium-binding protein calbindin (Cb) or the modulatory neuropeptide somatostatin (SOM) in the hippocampal formation.
Significantly fewer (P ⬍ 0.05) PV-positive interneurons were detected in layers II and IV of the cortex in
Cln3 ⫺/⫺ mice as compared with controls (Fig. 4A).
Measurements of cross-sectional area indicated that
persisting PV-positive neurons in the entorhinal cortex
of these mice were of the same size as controls (Figs. 4B
and 4C). Though we found mild atrophy of these
PV-positive interneurons in the entorhinal cortex of
heterozygotes, the change was quite small (Figs. 4B
and 4C).
The most significant change in the hippocampus
was the hypertrophy of most interneuronal populations examined. The changes were most marked for
SOM- and Cb-positive interneurons; in these populations the increases were all statistically significant and
in some cases the neuronal areas were increased by
almost 25% over control values (Fig. 5B). The population-wide nature of changes in cell size in Cln3 ⫺/⫺
mice is clearly seen by comparison of cell-size distribution histograms across different genotypes. Representative examples (Figs. 5C and 5D) show the typical shift
of the whole population of cells towards an increase in
area. While in heterozygous mice atrophy of PVpositive interneurons was observed, the changes were
relatively small (Fig. 5B). Analysis of hippocampal
interneuronal number showed a consistent, although
not statistically significant, trend towards reduced
number of SOM- and Cb-positive neurons in most
subregions of Cln3 ⫺/⫺ (Fig. 5A). There were no
significant changes in the number of PV-positive
interneurons. Taken together, our findings point to
specific and significant changes in the number and
size of certain interneuron populations in Cln3 ⫺/⫺
mice.
Cln2 Protease Activity Is Elevated in
Cln3 ⴚ/ⴚ Mice
To further compare the phenotype of Cln3 ⫺/⫺ mice
with Batten disease patients we measured the
activity of murine Cln2 protease, one of the lysosomal enzymes known to have elevated activity in
Batten disease patients (Sleat et al., 1998; Junaid &
329
A Mouse Model for Batten Disease
Pullarkat, 1999). In parallel with human patients, Cln2
protease activity was significantly elevated in the
brains of Cln3 ⫺/⫺ mice compared to wild-type
littermates, while heterozygotes are unaffected (Fig. 6).
At 5 months of age the activity in Cln3 ⫺/⫺ animals
was approximately doubled compared to controls. Similar increases have reported in Batten
disease patients. As with accumulation of storage material,
this phenotype appeared to be progressive in that it
was more pronounced at 5 months than 2 months of
age. No increase was seen in heterozygotes (Fig. 6).
DISCUSSION
FIG. 4. Loss of detectable parvalbumin (PV)-positive interneurons
in the entorhinal cortex. (A) Histogram of Abercrombie-corrected
number of PV-positive interneurons per section of entorhinal cortex
in mice at 7 months of age (5 sections per animal). Fewer neurons
were detected in Cln3 homozygous mutant compared to controls (*t
test, P ⬍ 0.05 vs wild-type). (B) Mean cross-sectional area of PVpositive interneurons (n ⫽ 300 neurons per genotype). Measurements indicate atrophy of PV-positive interneurons in heterozygotes
(*P ⬍ 0.03 vs wild-type). (C) Plot of cell size distribution.
Given that the Cln3 gene has 82% sequence identity and
85% amino acid similarity with human CLN3 (Lee et al.,
1996), we hypothesized that Cln3⫺/⫺ mice would show
many of the pathological and biochemical changes found
in human Batten disease patients. We have created mice
deficient in the Cln3 gene by gene targeting. Histological
analysis reveals progressive cellular abnormalities that are
characteristic of Batten disease pathology. There is widespread accumulation of autofluorescent material contained in storage bodies that stain positively for subunit c
of the mitochondrial ATP synthase and have a multilamellar ultrastructure. Moreover, Cln3⫺/⫺ mice demonstrate an elevated activity of Cln2 protease as is
observed in Batten disease patients (Sleat et al., 1998;
Junaid & Pullarkat, 1999).
Despite the cellular pathology that develops in
Cln3⫺/⫺ mice, they did not develop obvious clinical
symptoms by 12 months of age. The lack of overt
neurological signs is likely to be due to a pathologic
process that requires a longer interval to produce
symptoms or due to the lack of involvement of neuronal populations whose dysfunction is readily detected
or both. Continued observation over time will determine whether the development of neuronal pathology
is sufficient to cause clinical symptoms prior to agerelated death. It is noteworthy that Cln3⫺/⫺ mice are
less severely affected than the autosomal recessive
NCL mouse models, mnd (motor neuron degeneration)
and nclf (neuronal ceroid lipofuscinosis), which exhibit
progressive motor dysfunction from 4 and 8 months of
age, respectively (Messer & Flaherty, 1986; Bronson et
al., 1993, 1998; Pardo et al., 1994). The motor deficits in
mnd and nclf are preceded by accumulation of NCLlike storage material from just a few weeks of age
followed by retinal atrophy from approximately 4
months of age (Bronson et al., 1998; Chang et al., 1994).
These observations suggest that the presence of stor-
Copyright r 1999 by Academic Press
All rights of reproduction in any form reserved.
330
Mitchison et al.
FIG. 5. Neuronal number and area in hippocampus of mice at 7 months of age. (A) Abercrombie-corrected counts of detectable interneuronal
number per coronal section through hippocampus (5 sections per animal). (B) Mean cross-sectional area of interneuron subpopulations (number
of neurons, n ⫽ 120–150, except SOM- and PV-positive radiatum, n ⫽ 60–73. *t test, P ⬍ 0.0005; **P ⬍ 0.05 vs wild-type. #P ⬍ 0.0005; ##P ⬍ 0.05
vs Cln3⫺/⫺ ). (C, D) Representative plots of cell size distribution for neuronal areas of (C) SOM-positive neurons in oriens and (D) Cb-positive
neurons in radiatum.
Copyright r 1999 by Academic Press
All rights of reproduction in any form reserved.
A Mouse Model for Batten Disease
FIG. 6. Cln2 protease activity (nmol/h/mg protein) in brains of
Cln3 ⫺/⫺ mice. Cln2 protease activity is significantly elevated (*t
test, P ⬍ 0.001) in brains of ⫺/⫺ compared to ⫹/⫺ or ⫹/⫹ mice at 5
months of age. Carriers are indistinguishable from wild-type. Protease activity is also elevated (**t test, P ⬍ 0.005) in ⫺/⫺ mice at 2
months of age but to a lesser extent.
age bodies may substantially precede the onset of neuronal dysfunction in aging Cln3 ⫺/⫺ mice. If so, studies of
cortical and hippocampal physiology may identify subclinical abnormalities in animals at 12 months of age. Similarly
in humans, storage material is known to be present
before birth (Munroe et al., 1996) although clinical
symptoms do not develop until 4–7 years of age.
The disease-causing genes nclf and mnd are known
to be genetically distinct from Cln3 based on chromosomal map location (Lee et al., 1996; Bronson et al.,
1998; Messer et al., 1992). The mnd gene is orthologous
to CLN8, which was recently recognized as a new NCL
subtype (Goebel et al., 1999; Ranta et al., 1999). The nclf
gene is located on mouse chromosome 9 and human
synteny suggests it is likely to represent a naturally
occurring model for a variant late infantile/early
juvenile form of NCL, CLN6 (Bronson et al., 1998;
Sharp et al., 1997). In accordance with the milder
phenotype, the CNS of Cln3 ⫺/⫺ mice at 7 months of
age exhibits less profound morphological changes
than those seen in either mnd or nclf mice (Cooper et al.,
1999; Bronson et al., 1993, 1998; Pardo et al., 1994). It
should be noted however that genetic background
could also influence the penetrance of the Cln3 mutation; such an effect has been reported for the mnd
331
mouse (Messer et al., 1999). An alternative, but not
exclusive, explanation of the milder phenotype of the
Cln3⫺/⫺ mouse is that disruption of Cln3/CLN3 influences neurons whose dysfunction is less evident clinically than those affected CLN6 or CLN8.
Though the cellular changes in Cln3 ⫺/⫺ mice were
less severe than in mnd and nclf mice, we did observe
significant loss of one subpopulation of neurons (PVpositive) in the entorhinal cortex, which is also reflected in the hippocampus, in addition to a strong
trend toward reduced numbers of SOM-positive interneurons in the hippocampus. It is not yet clear whether
the apparent loss of interneurons is due to cell death or
the down-regulation of normally expressed phenotypic markers (Cooper et al., 1999). There was also
significant hypertrophy of many subpopulations of
interneurons, indicating dysregulation of cell volume.
In contrast, cortical interneurons, although reduced in
number in Cln3 ⫺/⫺ mice, did not exhibit hypertrophy. Distinct effects of Cln3 disruption on different
interneuronal populations raise the possibility that
different mechanisms may be operative. Since there is
progressive cortical and cerebellar atrophy and selective loss of neurons in Batten disease, it will be
important to examine the CNS of aged mice for evidence
of the progressive development of more significant
pathologic changes. Studies of retinal function are
currently ongoing in the Cln3 ⫺/⫺ mice (S. Nusinowitz and H. M. Mitchison, unpublished data).
Histological examination of the CNS of Cln3 ⫺/⫺
mice revealed several features in common with the
other NCL mouse models. In mnd mice, GABAergic
interneurons of the hippocampus and neocortex also
exhibit dense accumulations of autofluorescent material in interneurons and loss of staining for certain phenotypic markers together with pronounced hypertrophy of
remaining detectable interneurons (Cooper et al., 1999).
Further evidence for the pathologic involvement of interneurons has also been found in the nclf mouse (J. D.
Cooper, unpublished data). Although strain differences do
not permit direct comparisons, it is apparent that GABAergic interneurons are affected in each of these three mouse
models of NCL. Inhibitory interneurons in the hippocampus and cortex exert a powerful influence upon excitatory transmission in these brain regions (Freund &
Buzsaki, 1996; Singer, 1996). It will be important to
examine whether morphological abnormalities in Cln3
⫺/⫺ mice are reflected by alterations in excitatory
transmission or thresholds for seizure activation.
Recent progress in understanding the possible function of the CLN3 protein has come from studies of a S.
Copyright r 1999 by Academic Press
All rights of reproduction in any form reserved.
332
cerevisiae ortholog, BTN1. BTN1 is not required for
viability or for degradation of the mitochondrial ATP
synthase (Pearce & Sherman, 1997). The orthologous
protein Btn1p localizes to the vacuole, the yeast
lysosome equivalent, which is responsible for lysosomal function as well as osmotic regulation (Pearce et
al., 1999; Croopnick et al., 1998). Yeast carrying
a null allele of BTN1, strain (btn-1⌬), are resistant to
ANP (D-(-)-threo-2-amino-1-[ p-nitrophenyl]-1,3- propanediol) due to an increased ability to acidify the
growth medium. This phenotype is associated with a
lowered pH in the mutant vacuole early in the growth
course and is rescued by introduction of the human
CLN3 gene (Pearce & Sherman, 1998), implicating
involvement of CLN3 in cellular pH regulation (Pearce
et al., 1999). DNA microarray analysis shows that a
yeast gene with homology to human and drosophila
HOOK1, known to be involved in endocytosis (Kramer
& Phistry, 1996, 1999), is upregulated in the btn-1⌬
strain, indicating possible impairment of protein
trafficking in mutant cells (Pearce et al., 1999). The
results in yeast suggest possible pathogenetic
mechanisms for Batten disease in mice and humans.
Altered lysosomal pH regulation as a result of Cln3
disruption may reduce the activity of lysosomal proteases, with profound effects on the degradation of
proteins such as mitochondrial ATP-synthase subunit c within the lysosome, cause abnormal aggregation of proteins, or interfere with intracellular protein
trafficking. In addition, loss of Cln3 function may
interfere with the regulation of cellular volume. Little
is known about the mechanisms underlying this process, but controlled shifts in lysosomal activity
represent a means to regulate neuronal cytoplasmic
volume.
The targeted disruption of Cln3 provides an animal
model for a known NCL-associated gene in humans.
Further studies are required to assess subtle behavioral
abnormalities and to determine whether clinical symptoms develop in aging mice. However, many of the
characteristic features of Batten disease are recapitulated in the Cln3 ⫺/⫺ mouse, indicating that it will
provide a valuable model system for analyzing disease
mechanisms and for evaluating potential therapeutic
interventions.
ACKNOWLEDGMENTS
We thank Lisa Garrett, Amy Chen, and Theresa Hernandez for
help and advice with gene targeting and mouse colony manage-
Copyright r 1999 by Academic Press
All rights of reproduction in any form reserved.
Mitchison et al.
ment. We thank Michael Eckhaus, Sharon Suchy, Marlene Dressman,
Elina Hellsten, Steve Nussinowitz, and members of the Gardiner lab
for helpful discussion and suggestions. We also thank Sara Mole and
Kit-Yi Leung for comments on the manuscript. We thank E.
Kominami, Juntendo University, Tokyo, Japan, for the generous gift
of antisera to subunit c. This work was supported by the Division of
Intramural Research, National Institutes of Health (U.S.A.) (H.M.M.,
D.J.B., R.L.N., J.W.O.), the Medical Research Council (UK) (N.D.E.G.,
H.M.M.), the Remy Foundation (J.D.C.); the Natalie Fund and the
Batten Disease Support and Research foundation (W.C.M.); NIH
Grants NS 30147 (M.A.J., R.K.P.) and NS 30152 (M.H.B., P.E.M.T.). We
also thank the Children’s Brain Diseases Foundation (U.S.A.) for
support.
REFERENCES
Abercrombie, M. (1946) Estimation of nuclear populations from
microtome sections. Anat. Record 94, 239–247.
Bronson, R. T., Lake, B. D., Cook, S., Taylor, S., & Davisson, M. T.
(1993) Motor neuron degeneration of mice is a model of neuronal
ceroid lipofuscinosis (Batten’s disease). Ann. Neurol. 33, 381–
385.
Bronson, R. T., Donahue, L. R., Johnson, K. R., Tanner, A., Lane, P. W.,
& Faust, J. R. (1998) Neuronal ceroid lipofuscinosis (nclf ), a new
disorder of the mouse linked to chromosome 9. Am. J. Med. Genet.
77, 289–297.
Burditt, L. J., Chotal, K., Hirani, S., Nugent, P. G., & Winchester, B. G.
(1980) Biochemical studies on a case of feline mannosidosis.
Biochem. J. 189, 467–473.
Cavalieri, B. (1966) Geometria degli indivisibili. Torino: Unione
Tipografico, Editrice.
Chang, B., Bronson, R. T., Hawes, N. L., Roderick, T. H., Peng, C.,
Hageman, G. S., & Heckenlively, J. R. (1994) Retinal degeneration
in motor neuron degeneration: A mouse model of ceroid lipofuscinosis. Invest. Ophthalmol. Vis. Sci. 35, 1071–1076.
Cooper, J. D., Messer, A., Feng, A. K., Chua-Couzens, J., & Mobley,
W. C. (1999) Apparent loss and hypertrophy of interneurons in a
mouse model of neuronal ceroid lipofuscinosis: Evidence for
partial response to insulin-like growth factor-1 treatment. J.
Neurosci. 19, 2556–2567.
Croopnick, J. B., Choi, H. C., & Mueller, D. M. (1998) The subcellular
location of the yeast Saccharomyces cerevisiae homologue of the
protein defective in the juvenile form of Batten disease. Biochem.
Biophys. Res. Commun. 250, 335–341.
Deng, C., Wynshaw-Boris, A., Zhou, F., Kuo, A., & Leder, P. (1996)
Fibroblast growth factor receptor 3 is a negative regulator of bone
growth. Cell 84, 911–921.
Elleder, M., Sokolova, J., & Hrebicek, M. (1997) Follow-up study of
subunit c of mitochondrial ATP synthase (SCMAS) in Batten
disease and in unrelated lysosomal disorders. Acta Neuropathol.
(Berlin) 93, 379–390.
Freund, T. F., & Buzsaki, G. (1996) Interneurons of the hippocampus.
Hippocampus 6, 347–470.
Goebel, H. H., Mole, S. E., & Lake, B. D. (Eds.) (1999) The Neuronal
Ceroid Lipofuscinoses (Batten Disease). Biomedical and Health Res.,
V33. IOS Press, Netherlands.
Greene, N. D. E., Bernard, D. L., Taschner, P. E. M., Lake, B. D., de
Vos, N., Breuning, M. H., Gardiner, R. M., Mole, S. E., Nussbaum,
R. L., & Mitchison, H. M. (1999) A murine model for juvenile
NCL: Gene targeting of mouse Cln3. Mol. Genet. Metab. 66, 309–313.
A Mouse Model for Batten Disease
Hall, N. A., Lake, B. D., Dewji, N. N., & Patrick, A. D.
(1991) Lysosomal storage of subunit c of mitochondrial ATP
synthase in Batten’s disease (ceroid lipofuscinosis). Biochem. J. 275,
269–272.
Hogan, B., Beddington, R., Constantini, F., & Lacy, E. (Eds.) (1994)
Manipulating the Mouse Embryo. Cold Spring Harbor Laboratory
Press, Cold Spring Harbor, NY.
Hunziker, W., & Fumey, C. (1994) A di-leucine motif mediates
endocytosis and basolateral sorting of macrophage IgG Fc receptors in MDCK cells. EMBO J. 13, 2963–2967.
International Batten Disease Consortium. (1995) Isolation of a novel
gene underlying Batten disease (CLN3). Cell 82, 949–957.
Ivy, G. O., Schottler, F., Wenzel, J., Baudry, M., & Lynch, G. (1984)
Inhibitors of lysosomal enzymes: Accumulation of lipofuscin-like
dense bodies in the brain. Science 226, 985–7.
Janes, R. W., Munroe, P. B., Mitchison, H. M., Gardiner, R. M., Mole,
S. E., & Wallace, B. A. (1996) A model for Batten disease protein
CLN3: Functional implications from homology and mutations.
FEBS Lett. 399, 75–77.
Järvelä, I., Sainio, M., Rantamaki, T., Olkkonen, V. M., Carpén, O.,
Peltonen, L., & Jalanko, A. (1998) Biosynthesis and intracellular
targeting of the CLN3 protein defective in Batten disease. Hum.
Mol. Genet. 7, 85–90.
Järvelä, I., Lehtovirta, M., Tikkanen, R., Kyttala, A., & Jalanko,
A. (1999) Defective intracellular transport of CLN3 is the molecular basis of Batten disease (JNCL). Hum. Mol. Genet. 8, 1091–
1098.
Jolly, R. D. (1995) Comparative biology of the neuronal ceroid
lipofuscinoses (NCL): An overview. Am. J. Med. Genet. 57, 307–
311.
Junaid, M. A., & Pullarkat, R. K. (1999) Increased brain lysosomal
pepstatin-insensitive proteinase activity in patients with neurodegenerative diseases. Neurosci. Lett. 264, 157–160.
Junaid, M. A., Brooks, S. S., Wisniewski, K. E., & Pullarkat, R. K.
(1999) A novel assay for lysosomal pepstatin-insensitive proteinase and its application for the diagnosis of late-infantile neuronal
ceroid lipofuscinosis. Clin. Chim. Acta. 281, 169–176.
Kramer, H., & Phistry, M. (1996) Mutations in the Drosophila hook
gene inhibit endocytosis of the boss transmembrane ligand into
multivesicular bodies. J. Cell Biol. 133, 1205–1215.
Kramer, H., & Phistry, M. (1999) Genetic analysis of hook, a gene
required for endocytic trafficking in drosophila. Genetics 151,
675–684.
Lake, B. D. (1997) Lysosomal and peroxisomal disorders. In Greenfields Neuropathology, 6th ed., pp. 658–753. Edward Arnold, London.
Lee, R. L., Johnson, K. R., & Lerner, T. J. (1996) Isolation and
chromosomal mapping of a mouse homolog of the Batten disease
gene CLN3. Genomics 35, 617–619.
Luna, L. G. (1992) In Histopathological Methods and Color Atlas of
Special Stains and Tissue. American Histolabs, Inc., Publications
Division, Gaithersburg, MD.
Mann, D. M., Yates, P. O., & Stamp, J. E. (1978) The relationship
between lipofuscin pigment and aging in the human nervous
system. J. Neurol. Sci. 37, 83–93.
Messer, A., & Flaherty, L. (1986) Autosomal dominance in a lateonset motor neuron disease in the mouse. J. Neurogenet. 3, 345–
355.
Messer, A., Plummer, J., Maskin, P., Coffin, J. M., & Frankel, W. N.
(1992) Mapping of the motor neuron degeneration (Mnd) gene, a
333
mouse model of amyotrophic lateral sclerosis (ALS). Genomics 13,
797–802.
Messer, A., Manley, K., & Plummer, J. A. (1999) An early-onset
congenic strain of the motor neuron degeneration (mnd) mouse.
Mol. Genet. Metab. 66, 393–397.
Munroe, P. B., Rapola, J., Mitchison, H. M., Mustonen, A., Mole, S. E.,
Gardiner, R. M., & Jarvela, I. (1996) Prenatal diagnosis of Batten’s
disease. Lancet 347, 1014–1015.
Palmer, D. N., Fearnley, I. M., Walker, J. E., Hall, N. A., Lake,
B. D., Wolfe, L. S., Haltia, M., Martinus, R. D., & Jolly, R. D.
(1992) Mitochondrial ATP synthase subunit c storage in the
ceroid-lipofuscinoses (Batten disease). Am. J. Med. Genet. 42,
561–567.
Pardo, C. A., Rabin, B. A., Palmer, D. N., & Price, D. L. (1994)
Accumulation of the adenosine triphosphate synthase subunit C
in the mnd mutant mouse. A model for neuronal ceroid lipofuscinosis. Am. J. Pathol. 144, 829–835.
Pearce, D. A., & Sherman, F. (1997) BTN1, a yeast gene corresponding to the human gene responsible for Batten’s disease, is not
essential for viability, mitochondrial function, or degradation of
mitochindrial ATP synthase. Yeast 13, 691–697.
Pearce, D. A., & Sherman, F. (1998) A yeast model for the study of
Batten disease. Proc. Natl. Acad. Sci. USA 95, 6915–6918.
Pearce, D. A., Ferea, T., Nosel, S. A., Das, B., & Sherman, F. (1999)
Action of BTN1, the yeast orthologue of the gene mutated in
Batten disease. Nature Genet. 22, 55–58.
Ranta, S., Zhang, Y., Ross, B., Lonka, L., Takkunen, E., Messer, A.,
Sharp, J., Wheeler, R., Kusumi, K., Mole, S., Liu, W., Bento Soares,
M., de Fatima Bonaldo, M., Hirvasniemi, A., de la Chapelle, A.,
Gilliam, T. C., Lehesjoki, A-E. (1999) The neuronal ceroid lipofuscinoses human CLN8 and murine mnd are associated with mutations n orthologous novel genes encoding putative transmembrane proteins. Nature Genet., in press.
Rapola, J. (1993) Neuronal Ceroid-Lipofuscinoses in Childhood.
Perspect. Pediatr. Pathol. 17, 7–44.
Santavuori, P. (1988) Neuronal ceroid lipofuscinosis in childhood.
Brain Dev. 10, 80–83.
Sekhon, S. S., & Maxwell, D. S. (1974) Ultrastructural changes in
neurons of the spinal anterior horn of aging mice with particular
reference to the accumulation of lipofuscin pigment. J. Neurocytol.
3, 59–72.
Sharp, J. D., Wheeler, R. B., Lake, B. D., Savukowski, M., Järvelä, I. E.,
Peltonen, L., Gardiner, R. M., & Williams, R. E. (1997) Loci for
classical and a variant late infantile neuronal ceroid lipfuscinoses
map to chromosomes 11p15 and 15q21-23. Hum. Mol. Genet. 6,
591–596.
Singer, W. (1996) Neurophysiology: The changing face of inhibition.
Curr. Biol. 6, 395–397.
Sleat, D. E., Donnelly, R. J., Lackland, H., Liu, C.-G., Sohar, I.,
Pullarkat, R. K., & Lobel, P. (1997) Association of mutations in a
lysosomal protein with classical late-infantile neuronal ceroid
lipofuscinosis. Science 277, 1802–1805.
Sleat, D. E., Sohar, I., Pullarkat, P. S., Lobel, P., & Pullarkat, R. K.
(1998) Specific alterations in levels of mannose 6-phosphorylated
glycoproteins in different neuronal ceroid lipofuscinoses. Biochem.
J. 334, 547–551.
Taschner, P. E. M., de Vos, N., & Breuning, M. H. (1997) Cross-species
homology of the CLN3 gene. Neuropediatrics 28, 18–20.
Tybulewicz, V. L., Crawford, C. E., Jackson, P. K., Bronson, R. T., &
Mulligan, R. C. (1991) Neonatal lethality and lymphopenia in mice
Copyright r 1999 by Academic Press
All rights of reproduction in any form reserved.
334
with a homozygous disruption of the c-abl proto-oncogene. Cell 65,
1153–1163.
van Ree, J. H., van den Broek, W. J., Dahlmans, V. E., Grrot, P. H.,
Vidgeon-Hart, M., Frants, R. R., Wieringa, B., Havekes, L. M., &
Hofker, M. H. (1994) Diet-induced hypercholesterolemia and
atherosclerosis in heterozygous apolipoprotein E-deficient mice.
Atherosclerosis 111, 25–37.
Vines, D. J., & Warburton, M. J. (1999) Classical late infantile neuronal ceroid lipofuscinosis fibroblasts are deficient
Copyright r 1999 by Academic Press
All rights of reproduction in any form reserved.
Mitchison et al.
in lysosomal tripeptidyl peptidase I. FEBS Lett. 443, 131–
135.
White, S., Hatton, S. R., Siddiqui, M. A., Parker, C. D., Trowbridge,
I. S., & Collawn, J. F. (1998) Analysis of the structural requirements
for lysosomal membrane targeting using transferrin receptor
chimeras. J. Biol. Chem. 273, 14355–14362.
Young, E., Chatterton, C., Vellodi, A., & Winchester, B. (1997) Plasma
chitotriosidase activity in Gaucher disease patients who have been
treated either by bone marrow transplantation or by enzyme replacement therapy with alglucerase. J. Inher. Metab. Dis. 20, 595–602.