Knock-out mouse for Canavan disease: a model for gene transfer to the central nervous system

RESEARCH ARTICLE

Knock-out mouse for Canavan disease: a modelfor gene transfer to the central nervous system

Reuben Matalon1,6*

Peter L. Rady1

Kenneth A. Platt2

Henry B. Skinner2

Michael J. Quast3

Gerald A. Campbell4

Kimberlee Matalon1,5

Jeffrey D. Ceci6

Stephen K. Tyring1

Michael Nehls2

Sankar Surendran1

Jingna Wei3

Ed L. Ezell3

Sylvia Szucs1

1Department of Pediatrics, Children'sHospital, UTMB Galveston, TX, USA2Lexicon Genetics Incorporated,The Woodlands, TX, USA3Anatomy & Neurosciences,Galveston, TX, USA4Department of Pathology, Galveston,TX, USA5Department of Human Developmentand CS, University of Houston,Houston TX, USA6Human Biological Chemistry &Genetics, UTMB Galveston, TX, USA

*Correspondence to: R. Matalon,Department of Pediatrics, UTMBChildren's Hospital, 301 UniversityBoulevard, Galveston,TX 77555-0359, USA.E-mail: [email protected]

Received: 14 January 2000

Revised: 13 March 2000

Accepted: 17 March 2000

Published online: 20 March 2000

Abstract

Background Canavan disease (CD) is an autosomal recessive leukodystro-phy characterized by de®ciency of aspartoacylase (ASPA) and increased levelsof N-acetylaspartic acid (NAA) in brain and body ¯uids, severe mentalretardation and early death. Gene therapy has been attempted in a number ofchildren with CD. The lack of an animal model has been a limiting factor indeveloping vectors for the treatment of CD. This paper reports the successfulcreation of a knock-out mouse for Canavan disease that can be used for genetransfer.

Methods Genomic library l knock-out shuttle (lKOS) was screened and aspeci®c pKOS/Aspa clone was isolated and used to create a plasmid with 10base pair (bp) deletion of exon four of the murine aspa. Followinglinearization, the plasmid was electroporated to ES cells. Correctly targetedES clones were identi®ed following positive and negative selection andcon®rmed by Southern analysis. Chimeras were generated by injection of EScells to blastocysts. Germ line transmission was achieved by the birth ofheterozygous mice as con®rmed by Southern analysis.

Results Heterozygous mice born following these experiments have no overtphenotype. The homozygous mice display neurological impairment, macro-cephaly, generalized white matter disease, de®cient ASPA activity and highlevels of NAA in urine. Magnetic resonance imaging (MRI) and spectroscopy(MRS) of the brain of the homozygous mice show white matter changescharacteristic of Canavan disease and elevated NAA levels.

Conclusion The newly created ASPA de®cient mouse establishes animportant animal model of Canavan disease. This model should be usefulfor developing gene transfer vectors to treat Canavan disease. Vectors for thecentral nervous system (CNS) and modulation of NAA levels in the brainshould further add to the understanding of the pathophysiology of Canavandisease. Data generated from this animal model will be useful for developingstrategies for gene therapy in other neurodegenerative diseases. Copyright #2000 John Wiley & Sons, Ltd.

Keywords Canavan disease; aspartoacylase de®ciency; N-acetylasparticacid (NAA); knock-out mouse; spongy degeneration of the brain; centralnervous system model.

Introduction

Canavan disease (CD) is an autosomal recessive leukodystrophy associatedwith spongy degeneration of the white matter of the brain, leading to mentalretardation, megalencephaly, inability to attain milestones, hypotonia and

THE JOURNAL OF GENE MEDICINEJ Gene Med 2000; 2: 165±175.

Copyright # 2000 John Wiley & Sons, Ltd.

early death [1]. Brain histology in CD shows characteristicspongy degeneration of the white matter and astrocyticswelling, while neurons are spared [2±4]. Matalon et al.[5] showed that aspartoacylase (ASPA) de®ciency is thebasic defect in CD. ASPA is one of two amino acylases,aminoacylase I hydrolyzes acetate from other aminoacids, aspartocylase (aminoacylase II) is speci®c forN-acetylaspartic acid (NAA), which is hydrolyzed toaspartate and acetate [6]. NAA acid is abundantlysynthesized in human and other mammalian brains,while its role in brain metabolism remains an enigma[7±12]. The de®ciency in the hydrolysis of NAA leads tospongy degeneration and the phenotypic characteristicsof CD, although the mechanism is unclear. None the less,the hydrolysis of NAA is crucial for intact white matter.Studies have described that the synthesis of NAA occurs inthe gray matter, while ASPA is localized in the myelintracts of the white matter [13,14]. This difference inlocalization suggests biochemical compartmentation ofthe enzyme and substrate.

The gene for ASPA has been cloned [15,16] andmutations leading to ASPA de®ciency have been studiedin patients with CD [15±22]. There have been more than30 mutations described [23]. Interestingly, only twomutations are predominant among 98% of affectedAshkenazi Jews; Glu285Ala and Tyr231X. In non-Jewish individuals of European ancestry the mostcommon mutation is Ala305Glu. The carrier frequencyof the two common mutations in the Jewish population is1/37 to 1/40 [24,25]. This high frequency of carriersamong Jews has led to recommended carrier testing inthis population [26]. The severity of the disease, poorprognosis, and lack of speci®c treatment have led toattempts to develop an effective gene therapy. CD is theonly neurodegenerative disease in which gene therapyhas been tried by injecting a liposome containing humancDNA of ASPA into the brains of such children [27].Results of these experiments are not yet available in themedical literature.

We describe a genetically engineered animal model forCD by inactivating the murine aspa gene by homologousrecombination in embryonic stem cells using our recentlydescribed knock-out shuttle vector system (lKOS) [28].This model is immediately useful for two main purposes:(1) to study the metabolic role of NAA in brain, and todetermine if levels of NAA can be modulated, and (2) touse the mutant mice as a model for the exploration anddevelopment of gene therapy approaches, direct enzymedelivery and other therapeutic modalities of CD inchildren.

Materials and methods

Cloning and analysis of the murineaspa locus

We screened an arrayed lKOS genomic library (LexiconGenetics, The Woodlands, TX, USA) by PCR analysis with

two primers aspa-3, and aspa-4 (Figure 1). This primerpair ampli®ed 108 bp of exon four of the murine aspagene that was isolated earlier [16, Matalon et al.,unpublished data]. After identi®cation of seven positivephage pools, two consecutive rounds of hybridization,using the aspa3/aspa4 PCR product as a probe, resulted inthe isolation of three independent lvKOS phage clonesthat were converted into pKOS plasmids by Cre-mediatedrecombination as described [29]. The pKOS clones weresequenced and the partial genomic structure of the aspalocus determined as shown in Figure 1.

Generation of mutant mice

Clone pKOS/Aspa/62 (Figure 1) was used to create a10 bp deletion in exon four by yeast-mediated recombi-nation as described previously [28]. The yeast selectionmarker was replaced with an IRESbgeo (pA)/PGKpuro(pA) expression cassette by S®I- cleavage and ligation[28]. The resulting plasmid was linearized with NotI andchaperone oligonucleotide linkers were ligated onto thesticky ends to form closed hairpins [29]. Passage 6 Lex1ES cells (Lexicon Genetics) were electroporated and threeof 20 ES clones that survived positive selection (G418 andpuromycin) as well as negative selection with ®alouridine(FIAU) were correctly targeted. The correctly targetedclones were identi®ed by nested PCR analysis usingvector-speci®c primers Puro3, and Puro4, in combinationwith external genomic primers ASPA-14, and ASPA-15, asindicated in Figure 1. The clones were con®rmed bySouthern analysis using external probes. ES cell genomicDNA was analyzed with a 5k probe generated by PCRusing primer ASPA-13, and primer ASPA-12, after BglIdigestion and with a 3k probe generated by PCR usingprimer aspa-29, and primer aspa-31, after EcoRI diges-tion. Chimeras were generated by injection of targeted EScells into C57Bl/6-Tyrc-Brd blastocysts [30]. One clonetransmitted the mutant allele, and heterozygous andhomozygous mutants were generated and identi®ed bySouthern analysis using external probes (Figure 1).

Biochemical studies

ASPA activity in kidneys from wild-type, heterozygousand homozygous knock-out mice of CD were assayed.Tissues were homogenized in water, freeze thawed andassayed as described earlier [5]. NAA acid in the urinewas assayed by gas chromatography mass spectrometryfollowing acidi®cation and solvent extraction accordingto the method of Goodman and Markey [31]. Finalquantitation was kindly determined by S. Goodman usingstable isotope dilution.

Magnetic resonance imaging andspectroscopy

All MRI/MRS experiments were performed using a 4.7 T,33 cm diameter bore system (Varian Inova, Palo Alto, CA,USA) equipped with a 10 gauss/cm (max.), 12.8 cm

166 R. Matalon et al.

Copyright # 2000 John Wiley & Sons, Ltd. J Gene Med 2000; 2: 165±175.

diameter gradient insert. A single-turn, 1.7 cm surfacecoil, tuned to 200 MHz was used to acquire MRI and MRSdata. Mice were anesthetized with a 40 mg/kg dose of2,2,2-tribromoethanol (Avertin) for MRI/MRS studies.Proton MRI was performed using a gradient echo pulsesequence in order to set the voxel position for localizedspectroscopy in the brain. The voxel was of thedimensions 4 mmr3 mmr3 mm. The voxel was cen-tered in the thalamus but included some brain stem andcortex as well. Spectra were acquired using the pointresolved spectroscopy (PRESS) pulse sequence [32]. Each

of the two echo times was 40 ms, the repetition time was2.5 s. Each voxel was electronically homogenized(shimmed) using the unsuppressed water peak in orderto minimize spectral line widths. Subsequently, water

suppression was achieved using three successive chemicalshift selective 90-degree pulses (centered on the waterpeak) each pulse followed by a magnetic ®eld gradient(crusher gradient) to destroy the residual water signal.Water suppressed spectra were acquired with between160 and 400 signal averages, resulting in total spectralacquisition times of between 6.7 and 16.7 min. Theresulting raw nuclear magnetic resonance (NMR) signals(free induction decays) were processed by Fouriertransformation with a 5 Hz exponential line broadening®lter in order to improve signal-to-noise ratio. Baseline

correction (included in the MRI system software package)was applied in order to eliminate baseline roll whichappears in some spectra. Spectral peak integrals forcreatine/phosphocreatine (Cr, 3.05 ppm) and NAA

Figure 1. Targeted disruption of the murine aspartoacylase (aspa) gene. (A) Map of the murine aspa gene, the targeting vector(pKOS/Aspa/62) restriction enzyme cutting sites. The position of a 400 bp sized probe 5k from the targeting vector and theposition of 600 bp sized probe 3k from the targeting vector are shown. The probes were generated by PCR using 12/13 and29/31 primer pairs. The numbered boxes denote exons, however exon six is not shown in the Figure. A targeting vector fordeleting 10 bp from exon four of the mouse aspa gene was engineered with the following important features shown. An IRES-bGeo cassette was constructed. This is a fusion gene of the b-galactosidase and the neomycin resistance gene which acted as apromoter trap. The fused gene was inserted at the 5k direction to a phosphoglycerate kinase (PGK) promoter-Puromycin cas-sette. This construct allowed for the selection of correctly targeted ES clones by expressing resistance to neomycin and/or pur-omycin. Puro3, Puro4 sense primers; Aspa14, Aspa15 antisense primers were used for PCR screening of the selected clones.The position of the PCR fragment generated by Puro3±4/aspa14±15 primer sets are indicated. The diagram indicates a BglI cut-ting site within the IRES-bGeo/PGK-Puro fusion complex which created a 23 kb mutant allelle and 29 kb wild-type. The dia-gram shows EcoRI restriction sites of the DNA of the wild-type allele that results in a 6.5 kb DNA fragment. The EcoRI sites ofthe mutant allele are indicated which give a 8.7 kb DNA fragment upon digestion. (B) Southern blot analysis of BglI digestedgenomic DNA from F1 intercross progeny using the PCR generated 5k probe. (C) Southern blot analysis of EcoRI digested geno-mic DNA from F1 intercross progeny using the PCR generated 3k probe. Expected size of normal and mutated fragments areindicated. Wild-type genotype (+/+) lane 5, heterozygous genotypes (+/x) lanes 1 and 4 and homozygous mutant genotypes(x/x) lanes 2,3,6,7 can be seen, respectively. KOS, knock-out shuttle; IRES, internal ribosome entry sequence; bGeo, b-galac-tosidase-neomycin resistance fusion gene; PGK, phosphoglycerate kinase; Puro, puromycin

Knock-out Mouse for Canavan Disease 167


(2.02 ppm) were measured. Experimental groups forNMR spectroscopy included knock-out (x/x, n=6)heterozygote (+/x, n=4) and wild-type (+/+, n=2)mice. Since heterozygous mice did not express CD, theywere pooled with wild-type mice. In this study the Cr peakserved as the internal standard as reported previously[33]. Statistical comparison of the ratio of the NAA/Crpeak intensities was made using a t-test for differencesbetween sham+control vs Canavan expressors.

Multislice T2-weighted spin echo images were per-formed in two of the wild-type mice and two of the CDexpressing mice. 14 contiguous 1 mm slices wereacquired with a TR of 3.0 s, a TE of 50 ms, a ®eld ofview of 3 cmr3 cm, 256 phase encode steps and twosignal averages per step.

Histologic processing of brain tissue

Mice were killed by cervical dislocation and the brainswere rapidly removed and placed in cold (4uC) physio-logic phosphate-buffered saline (PBS). Brain slices werecut to 1 mm thickness and placed in ®xative (4%paraformaldehyde, 0.1% glutaraldehyde, in PBS) over-night at 4uC. The slices were then submitted for standardparaf®n processing (dehydration through graded alco-hols, clearing with xylene and paraf®n in®ltration),embedded in paraf®n and sectioned at 7 mm. Sequentialsections were stained using standard histologic proce-dures for hematoxylin and eosin (H&E) or luxol fast bluemyelin stain with H&E counterstain.

Results

Generation of mice with aspatruncation

An arrayed murine genomic lKOS library (129 Sv/Ev)was screened by PCR ampli®cation using primers speci®cfor exon four of the murine aspa gene. Severalindependent lambda clones were isolated as describedpreviously [28] and sequenced in order to establish thegenomic structure of the murine aspa locus. A 10 bpdeletion in exon four was generated by yeast-mediatedrecombination using clone pKOS/Aspa/62, and thetargeting vector was completed by replacing the yeastprototrophic cassette with an IRES geo (pA)/PGK puro(pA) expression cassette (Figure 1). This targeting vectordesign afforded the advantage of providing a promotertrap (G418 selection) in addition to standard positiveselection (puromycin) following electroporation of theconstruct into ES cells. Three targeted ES cell clones wereidenti®ed by PCR analysis using primers Puro3 and Puro4in combination with primers Aspa-14 and Aspa-15(Figure 1). The results from the PCR analysis werecon®rmed by Southern analysis using external probes.Following microinjection of targeted ES cells into hostblastocysts and implantation into host pseudo-pregnantfemales one ES cell line generated chimeric animals that

transmitted the mutant allele through the mouse germ-line. Intercrossing the heterozygotes resulted in amendelian recessive distribution of wild-type (+/+),heterozygous (+/x) and homozygous knock-out (x/x) mutants that were identi®ed by Southern analysisusing external probes as shown in Figure 1. Threeheterozygote breeding pairs produced 122 mice withthe expected distribution of the genotypes for anautosomal recessive trait, 32 wild-type (+/+; 26.2%),55 heterozygous (+/x; 45.1%) and 35 homozygous forthe knock-out gene (x/x; 28.7%).

Phenotypic characterization of theknock-out mouse for Canavan disease

Physical growthHomozygous knock-out mice were clearly distinguishablefrom normal mice. The homozygous knock-out micefailed to thrive, a clear sign of a sick mouse (Figure 2).The weight of mice homozygous for the knock-out wassigni®cantly lower than their littermates. There was nodifference in the weight of the heterozygous and wild-type mice and they served as the control group. Eightmice from each group were matched for litter, age andgender. The mean weight at weaning of the eight mice inthe control group was 14.06 g (t0.84). The eighthomozygous knock-out (x/x) mice had a meanweight of 7.23 g (t0.93) with p<0.001. Four micedied shortly after weaning: the four were homozygousx/x mice. Other homozygous mice have survived from1.5 to 9 months.

Craniofacial abnormalitiesSince macrocephaly is a characteristic feature of childrenwith CD, it is interesting that macrocephaly also occurredin the knock-out mice. The normal depression of thefronto-nasal junction in the wild-type mice is differentfrom the bulging of that region which is clearly seen in theknock-out mouse (Figure 2).

Neurological featuresNeurologically, the homozygous mice had tremors,walked with splayed legs and stood with a wide base(Figure 3). Their pace was slower, shaky and they wereless mobile in their cages than their normal or hetero-zygous littermates. They seemed lethargic and insensitiveto pain when their tails were cut for genotyping. Some ofthe homozygous knock-out mice developed seizures atabout 6 months of age with clonic jerks and a post ictalstate after the seizure that lasted for several minutes.

Ataxia is a prominent feature of the knock-out mousefor CD. The ability of the mice to maintain balance wasmeasured using a rotorod, a sensitive screening methodto identify ataxia. Homozygous mice for the knock-outwere able to sustain their balance on the rotorod for avery short period of time [34]. 18 x/x mice maintainedbalance for a mean of 1.16t1.69 s compared to 35 +/+and +/x mice that maintained balance for a mean of44.9t21.4 s.



(A)

(B)

Figure 2. Failure to thrive of the knock-out mice. (A) Four week old littermates, showing the small size of the knock-outmouse (x/x) on the right. The prominence of the forehead of the knock-out mouse (x/x) can be seen compared to thedepression of the fronto-nasal junction of the wild-type mouse (+/+) on the left. (B) A pro®le of a mouse with Canavandisease showing the prominence of the forehead



Biochemical characterizationThe urine was quantitatively examined for NAA inhomozygous knock-out mice (x/x), heterozygote mice(+/x) and wild-type (+/+), using gas chromatographymass spectrometry. The levels of NAA in urine of thex/x mice were ninefold higher than the NAA in thenormal or heterozygote mice. There was no difference inthe NAA levels between the +/x or +/+ mice(Table 1). ASPA activity was determined in kidneyextracts since mammalian kidneys, as well as brain,show robust aspartoacylase activity [13]. The resultsshow a de®ciency of ASPA activity in homozygous knock-out (x/x) mice with approximately 2.7% residualenzyme activity in kidney and 3.4% residual enzymeactivity in brain (Table 1).

Magnetic resonance imaging (MRI) andmagnetic resonance spectroscopy(MRS) of mouse brains

Wild-type, heterozygous and homozygous knock-outmice were examined by T2-weighted MRI of the brain.The brains of the wild-type and heterozygous mice wereindistinguishable. In contrast, brains from homozygousknock-out mice show high signal intensity in thesubcortical regions of the brain. The T2-weightedimages of brains from normal and homozygous knock-out mice are shown in Figure 4. These experiments wereperformed on six control (4 +/x and 2 +/+) and sixhomozygous knock-out (x/x) mice. The MRI signalintensity is abnormally high in the brain stem (mainlymidbrain) and diencephalon of the homozygous knock-out mice, indicative of high water content. This type ofsignal intensity is typical of MRI ®ndings of the whitematter of patients with CD [35].

MRS was performed on a voxel 4 mmr3 mmr3 mmcentered in the thalamus. Spectral peak integrals forcreatine/phosphocreatine (Cr, 3.05 ppm) and NAA (2.02)were measured [33]. Experimental groups for NMRspectroscopy consisted of six homozygous knock-outmice, six controls (four heterozygous mice and twowild-type mice). These spectra are presented in Figure 5.The integrated values of NAA/creatine are shown inFigure 6 where the difference is statistically signi®cant

Figure 3. Neurological unsteadiness of the knock-out mouse (x/x). The mouse with CD (x/x) stands with a wide base andsplayed legs, indicative of ataxia. The small size of the knock-out (x/x) mouse can be seen

Table 1. Aspartoacylase activity in mouse kidney, brain andNAA levels in urine

Mouse

ASPA(mU/mg protein)

NAA(ng/mg creatinine)

Kidney Brain Urine(n) meantSD (n) mean (n) mean

Homozygous (x/x) (4) 0.013t0.004 (2) 0.007 (2) 1541% Residual activity 2.7% 3.4%Heterozygous(+/x) (2) 0.205 (1) 0.062 (1) 184% Residual activity 42% 31%Normal (+/+) (3) 0.487t0.12 (2) 0.202 (1) 170



( p<0.003) and indicate elevated NAA levels in the brains

of homozygous knock-out mice.As a con®rmatory measure for these MRS results, and

because of the broad in vivo peaks in the NAA region

(Figure 5), perchloric acid extracts from the subcortical

region of brains of homozygous knock-out and control

mice (+/x or +/+) were prepared and high-

resolution (400 MHz) NMR spectra were performed

(Figure 7). The resulting 1H spectra show the sharp,

distinct peak of NAA. As in the in vivo spectra, the x/x

mice have a high concentration of NAA in the perchloric

acid extracts compared with controls. The multiple peaks

at 2.10 ppm, which are adjacent to the sharp NAA

singlet, correspond to the protons on the C-3 of

glutamate, which may have caused the broad resonance

of the NAA peak in the in vivo spectra. The ratio of the

integrated peaks of NAA/creatine in the brain extract of

the homozygous knock-out mice was 1.92 while in the

controls the ratio was 0.83. These data are similar to the

in vivo results.

Figure 4. Mouse brain MRI. T2-weighted spin echo magnetic resonance images showing two contiguous images (each slice1 mm thick) in the region of the midbrain (left) to the thalamus (right). The top row is a wild-type mouse and the bottomrow is a knock-out mouse (x/x)

Figure 5. Mouse brain MRS. MRS of mouse brain of knock-out homozygous (x/x), wild-type (+/+) and heterozygous(+/x). A voxel 4r3r3 mm was centered in the thalamus (right frame)



Neuropathology

The histology of the brain of the homozygous knock-out

mice revealed a striking resemblance to human CD

pathology. Sections from the retinas of control and

homozygous knock-out mice were examined. The

affected mice showed vacuolation in ganglion and

nerve ®ber layers (Figure 8a,b). In the homozygous

knock-out mice the subcortical white matter throughout

was affected showing extensive vacuolation of the white

matter in the deep cortex and white matter bundles in the

corpus striatum as compared with the wild-type mouse

brain (Figure 8c,d). The hippocampus of the homozygous

knock-out mice show vacuolation in the pyramidal area

and sparing of the dentate gyrus (Figure 8e,f). Sections

from the control cerebellum showed the molecular layer,

Purkinje cell layer, granular cell layer and white matter.

The cerebellum of the homozygous knock-out mice show

the most severe changes in the Purkinje cell layer and less

severe changes in the other layers (Figure 8g,h). Histio-

logical analysis of the brains from heterozygous mice

were indistinguishable from wild-type controls.b-Galactosidase in our targeting vector is driven by the

endogenous promoter of aspa. It serves to localize the

cellular site of expression of aspa. The reporter gene b-

galactosidase was stained by immunohistochemical reac-

tion. There was no stain in the neurons while cells around

vacuoles were positive for b-galactosidase (data not

shown).

Discussion

A knock-out mouse for CD has been genetically engi-neered for the ®rst time using a newly developed vector,generated with lKOS system [28]. This mouse modelresults in the truncation of ASPA protein caused by adeletion of 10 bp in exon four and the concomitant frameshift. The phenotype of the knock-out mouse, thebiochemical, histological and MRI and MRS ®ndingsindicate that this mouse is an excellent model for CD andthat it will be a useful tool for studying the role of NAAand ASPA in normal brain and the maintenance of normalwhite matter. The ®eld of biotechnology with genetherapy as the major outcome is rapidly growing. The

Figure 6. Integrated values of NAA in mouse brain. Integratedvalues of NAA in mouse brains. The NAA/Cr peak integralratio was signi®cantly higher ( p<0.003) in Canavan-expressing mice (2.28t0.42 mean/SEM) versus control mice(0.79t0.20)

Figure 7. MRS of NAA extracted from mouse brain. The sub-cortical region of wild-type and knock-out mouse brains wereextracted with perchloric acid. (A) is from (+/+) and (B)from (x/x) brain

Figure 8. Neuropathology of mouse brain using H&E stain. Comparison of central nervous system tissues in control (a, c, e, g)and affected (b, d, f, h) mice: (a) normal retina; (b) retina from affected mouse demonstrating vacuolation in the ganglioncell and nerve ®ber layers (arrowheads); (c) control cortex (right), white matter and corpus striatum (left); (d) similar areafrom affected mouse with severe vacuolation in the deep cortex and white matter bundles in corpus striatum and less severechanges in the super®cial cortex and deep white matter; (e) control hippocampus showing pyramidal cell layer (PL) and den-tate gyrus (DG); (f) hippocampus of affected mouse with vacuolation of PL and sparing of DG; (g) control cerebellum showingthe molecular layer (ML), Purkinje cell layer (PCL), granular cell layer (GCL) and white matter (WM); (h) cerebellum ofaffected mouse with severe vacuolation in PCL and WM but not as severe in ML and GCL



treatment of CD will be enhanced by the use of the animalmodel for developing targeting vectors to determinesafety, timing and ef®cacy of therapy.

Animal models that mimic human diseases have beenimportant in the study of the pathophysiology of diseases.Homozygous knock-out mice frequently lack an apparentphenotype or are lethal, rendering them unsuitable fortherapeutic experiments [36]. There have been reports ofCanavan-like animals models that never had ASPAactivity determined and such animals are no longeravailable [37±40]. Our genetically engineered Canavanmouse has the clinical features of neurodegenerativedisease and can be systematically studied for genetargeting to the central nervous system.

The striking similarity in the morphological ®ndings ofthe white matter of the knock-out mouse brain and thebrains of patients with CD, as well as the increase in NAAlevels, indicate the similarity between the Canavanpatient and the knock-out mouse. Sager et al., usingcultured neurons that contained 3±5% astrocytes, suggestthe transport mechanism of NAA is expressed in glial cells,and that the transport of NAA occurs by a novel sodium-dependent carrier that is present only in glial cells [41].The mouse model can also be used for the study of thetransport of NAA to the glial compartment in the whitematter.

Reduced levels of NAA have been reported in manyother situations including Huntington's disease [42], HIV[43,44], stroke [45±47], the myodystrophic mouse[48,49], the mouse model for scrapie [50], Alzheimer's[51], Parkinson's [52], drug use [53] and malignant braintumors [54]. Increased level of NAA in the brain is speci®cfor CD. Our mouse model for CD and the increased levelof NAA in the mouse brain can be used to study the role ofNAA in the metabolism of white matter. There have beensuggestions that NAA may serve as an osmolite [55], ormay provide acetate for the synthesis of lipid componentsof myelin [56,57]. These suggestions involve differentmechanisms. Lowering the NAA level in the brain mayhelp CD if NAA is an osmolite. Hydrolysis of NAA at thesite of myelin synthesis in the glial compartment maysuggest an active metabolic role for NAA, which willrequire a different strategy for vector targeting. Themouse model could be used to determine the role of NAAin the synthesis of the neurotransmitter N-acetyl-aspartyl-glutamate (NAAG) and whether this has any relation tothe pathogenesis of CD. Thus, the knock-out mouse willprovide an excellent model for basic neuroscienceresearch.

The advantage of the newly developed knock-outmouse for CD is that the feasibility of gene therapy canbe tried prior to human experimentation. The mousemodel should provide urgently needed answers on genetargeting to the CNS.

Acknowledgements

We thank Dr K. Balamuragan for his advice and Stephen I.

Goodman for analysis of NAA in the urine of the mice. This

work was supported in part by research grants from the John

Sealy Foundation Endowment Grant 2504-98 and NIH-RO1

NS38562-01.

References

1. Matalon R, Michals-Matalon K. Molecular basis of CanavanDisease. Eur J Pediatr Neurol 1998; 2: 69±76.

2. Globus JH, Strauss I. Progressive degenerative subcorticalencephalopathy (Schilder's disease). Arch Neurol Psychiatr1928; 28: 1190±1228.

3. Canavan MM. Schilder's Encephalitis Periaxialis Diffusa. ArchNeurol Psychiatr 1931; 25: 299±308.

4. Van Bogaert L, Bertrand I. Sur une idiotie familiale avecdegerescense sponglieuse de neuraxe (note preliminaire). ActaNeurol Belg 1949; 49: 572±587.

5. Matalon R, Michals K, Sebasta D, Deanching M, Gashkoff P,Casanova J. Aspartoacylase de®ciency and N-acetylasparticaciduria in patients with Canavan Disease. Am J Med Genet1988; 29: 463±471.

6. Birnbaum SM, Levintow L, Kingsley RB, Greenstein JP.Speci®city of amino acid acylases. J Biol Chem 1952; 194:455±462.

7. Tallan HH, Moore S, Stein WH. N-Acetyl-L-aspartic acid in brain.J Biol Chem 1956; 219: 257±264.

8. Goldstein FB. Biosynthesis of N-acetyl-L-aspartic acid. J BiolChem 1959; 234: 2702±2706.

9. Goldstein FB. The enzymatic synthesis of N-acetyl-L-aspartic acidby subcellular preparations of rat brain. J Biol Chem 1969; 244:4257±4260.

10. Miyake M, Kakimoto Y, Sorimachi M. A gas chromatographicmethod for the determination of N-acetyl-L-aspartic acid, N-acetyl-aspartylglutamic acid and beta-citryl-L-glutamic acid andtheir distributions in the brain and other organs of variousspecies of animals. J Neurochem 1980; 36: 804±810.

11. Jacobson KB. Studies on the role of N-acetylaspartic acid inmammalian brain. J Gen Physiol 1957; 43: 323±333.

12. McIntosh JM, Cooper JR. Studies on the function of N-acetylaspartic acid in the brain. J Neurochem 1965; 12: 825±835.

13. Kaul RK, Casanova J, Johnson A, Tang P, Matalon R.Puri®cation, characterization and localization of aspartoacylasefrom bovine brain. J Neurochem 1991; 56: 129±135.

14. Johnson A, Kaul R, Casanova J, Matalon R. Aspartoacylase, thede®cient enzyme in spongy degeneration (Canavan disease) is amyelin-associated enzyme. J Neuropathol Exp Neurol 1989; 48:348

15. Kaul R, Gao GP, Balamurugan K, Matalon R. Human asparto-acylase cDNA and mis-sense mutation in Canavan disease. NatGenet 1993; 5: 118±123.

16. Kaul R, Balamurugan K, Gao GP, Matalon R. Canavan disease:Genomic organization and localization of Human ASPA to17p13-ter and conservation of the ASPA gene during evolution.Genomics 1994; 21: 364±370.

17. Kaul R, Gao GP, Aloya M, Balamurugan K, Petrosky A, Michals K,Matalon R. Canavan disease: mutations among Jewish and non-Jewish patients. Am J Hum Genet 1994; 55: 34±41.

18. Kaul R, Gao GP, Michals K, Whelan DT, Levin S, Matalon R. Anovel (cys 152>arg) missense mutation in Arab patient withCanavan disease. Hum Mutat 1995; 5: 269±271.

19. Kaul R, Gao GP, Matalon R, Aloya M, Su Q, Jin M, Johnson AB,Schutgens RB, Clarke JT. Identi®cation and expression of eightnovel mutations among non-Jewish patients with Canavandisease. Am J Hum Genet 1996; 59: 95±102.

20. Shaag A, Anikster Y, Christensen E, et al. The molecular basis ofCanavan (aspartoacylase de®ciency) disease in European non-Jewish patients. Am J Hum Genet 1995; 57: 572±580.

21. Kobayashi K, Tsujino S, Ezoe T, Hamaguchi H, Nihei K,Sakuragawa N. A missense mutation (I143T) in a Japanesepatient with Canavan disease. Hum Mutat 1998; 1: 308S±309S.

22. Rady PL, Vargas T, Tyring SK, Matalon RM, Langenbeck U.Novel missense mutation (Y231C) in a Turkish patient withCanavan disease. Am J Med Genet 1999; 87: 273±275.

23. Matalon R, Michals-Matalon K. Recent advances in Canavandisease. Adv Pediatr 1999; 46: 493±506.

24. Matalon R, Kaul R, Michals K. Carrier rate of Canavan disease



among Ashkenazi Jewish individuals. Am J Hum Genet 1994; 55:908A

25. Kronn D, Oddoux C, Phillips J, Ostrer H. Prevalence of Canavandisease heterozygotes in the New York metropolitan AshkenaziJewish population. Am J Hum Genet 1995; 57: 1250±1252.

26. ACOG Committee Opinion. Screening for Canavan Disease.Committee on Genetics, American College of Obstetricians andGynecologists. Int J Gynaecol Obstet 1999; 65: 91±92.

27. During MJ. Gene trial in New Zealand. Lancet 1996; 348: 61828. Wattler S, Kelly M, Nehls M. Construction of gene targeting

vectors from lKOS genomic libraries. BioTechniques 1999; 26:1150±1160.

29. Nehls M, Kyewski B, Messerle M, Waldschutz R, Schuddekopf K,Smith AJ, Boehm T. Two genetically separable steps in thedifferentiation of thymic epithelium. Science 1996; 272:886±889.

30. Zheng B, Sage M, Cai WW, Thompson DM, Tavsanli BC, CheahYC, Bradley A. Engineering a mouse balancer chromosome. NatGenet 1999; 4: 375±378.

31. Goodman SI, Markey SP. Diagnosis of Organic Acidemias by GasChromatography-Mass Spectrometry. Alan R. Liss, Inc.: NewYork, 1981; 1±43.

32. Bottomley PA. Spatial localization in NMR spectroscopy in vivo.Ann N Y Acad Sci 1987; 508: 333±348.

33. Pioro EP, Wang Y, Moore JK, Ng TC, Trapp BD, Klinkosz B,Mitsumoto H. Neuronal pathology in the wobbler mouse brainrevealed by in vivo proton magnetic resonance spectroscopy andimmunocytochemistry. Neuroreport 1998; 14: 3041±3046.

34. Ingram DK. Toward the behavioral assessment of biologicalaging in the laboratory mouse: Concepts, terminology, andobjectives. Exp Aging Res 1983; 9: 225±238.

35. Matalon R, Kaul R, Michals K. Canavan Disease. In The Molecularand Genetic Basis of Neurological Disease. Rosenberg RN,Prusiner SB, Di Mauro S, Barchi RL (eds). Butterworth-Heinemann: Boston, MA, 1997; 493±502.

36. Smithies O. Animal models of human genetic diseases. TrendsGenet 1993; 9: 112±116.

37. Hagan G, Bjerkas I. Spongy degeneration of white matter in thecentral nervous system of silver foxes (Vulpes vulpes). Vet Pathol1990; 27: 187±193.

38. Zachary JF, O'Brien CP. Spongy degeneration of the centralnervous system in two canine litter mates. Vet Pathol 1985; 22:561±567.

39. Mason RW, Hartley WJ, Randall M. Spongiform generation ofthe white matter in a Saymoyed pup. Aust Vet Pract 1979; 9:11±13.

40. Azzam NA, Bready JV, Vinters HV, Camcilla PA. Spontaneousspongy degeneration of the mouse brain. J Neuropathol ExpNeurol 1984; 43: 118±130.

41. Sager TN, Thomsen C, Laursen JS, Hansen AJ. Astroglia containa speci®c transport mechanism for N-acetyl-L-aspartate.J Neurochem 1999; 2: 807±811.

42. Dunlop DS, McHale DM, Lajtha A. Decreased brain N-acetyl-aspartate in Huntington's disease. Brain Res 1992; 580: 44

43. Meyerhoff DJ, MacKay S, Bahman L, Polle N, Dillon WP, WeinerMW, Fein G. Reduced brain N-acetylaspartate suggests neuronalloss in cognitively impaired human immunode®ciency virus-

seropositive individuals: in vivo 1H magnetic resonance spectro-scopic imaging. Neurology 1993; 43: 509±515.

44. Meyerhoff DJ, Bloomer C, Cardenas V, Norman D, Weiner MW,Fein G. Elevated subcortical choline metabolites in cognitivelyand clinically asymptomatic HIV+ patients. Neurology 1999; 5:995±1003.

45. Gideon P, Henrikson O, Sperling B, Christiansen P, Olsen TS,Jorgense HS, Arlien-Soborg P. Early time course of N-acetylaspartate, creatine and phosphocreatine, and compoundscontaining choline in the brain after acute stroke: a protonmagnetic resonance spectroscopy study. Stroke 1992; 23:1566±1572.

46. Pereira AC, Saunders DE, Doyle VL, Bland JM, Howe FA,Grif®ths JR, Brown MM. Measurement of initial N-acetylaspartate concentration by magnetic resonance spectroscopyand initial infarct volume by MRI predicts outcome in patientswith middle cerebral artery territory infarction. Stroke 1999; 8:1577±1582.

47. Stevens H, Jakobs C, de Jager AE, Cunningham RT, Korf J.Neuron-speci®c enolase and N-acetyl-aspartate as potentialperipheral markers of ischaemic stroke. Eur J Clin Invest 1999;1: 6±11.

48. Marcucci F, Colombo L, De Ponte G, Mussimi E. Decrease in N-acetylaspartic acid in brain of myodystrophic mice. J Neurochem1984; 43: 1484±1486.

49. Akiguchi I, Nakano S, Shiino A, Kimura R, Inubushi T, Handa J,Nakamura M, Tanaka M, Oka N, Kimura J. Brain protonmagnetic resonance spectroscopy and brain atrophy in myotonicdystrophy. Arch Neurol 1999; 3: 325±330.

50. Bell JD, Cox IJ, Williams SC, Belton PS, McConnell I, Hope J. Invivo detection of metabolic changes in a mouse model of scrapieusing nuclear magnetic resonance spectroscopy. J Gen Virol1991; 72: 2419±2423.

51. Jaarsma D, Veenma-van der Duin L, Korf J. N-Acetylaspartateand N-acetyl aspartyl-glutamate levels in Alzheimer's diseasepost-mortem brain tissue. J Neurol Sci 1994; 127: 230

52. Taylor-Robinson SD, Turjanski N, Bhattacharya S, Seery JP,Sargentoni J, Brooks DJ, Bryant DJ, Cox IJ. A proton magneticresonance spectroscopy study of the striatum and cerebral cortexin Parkinson's disease. Metab Brain Dis 1999; 14: 45±55.

53. Li SJ, Wang Y, Pankiewicz J, Stein EA. Neurochemicaladaptation to cocaine abuse: reduction of N-acetyl aspartatein thalamus of human cocaine abusers. Biol Psychiatry 1999; 11:1481±1487.

54. McConnell JR. Reduced levels of N-acetylaspartate in malignantbrain tumors. CA Cancer J Clin 1999; 3: 192

55. Baslow MH. Molecular water pumps and the aetiology ofCanavan disease: a case of the sorcerer's apprentice. J InheritMetab Dis 1999; 2: 99±101.

56. McIntosh JM, Cooper JR. Studies on the function of N-acetylaspartic acid in the brain. J Neurochem 1965; 12: 825±835.

57. Shigematsu H, Okamura N, Shimeno H, Kishimoto Y, Kan L,Feuselean C. Puri®cation and characterization of the heat stablefactors essential for conversion of lignoceric acid to cerebronicacid and glutamic acid: identi®cation of N-acetyl-L-aspartic acid.J Neurochem 1983; 40: 814±820.



Knock-out mouse for Canavan disease: a model for gene transfer to the central nervous system

Documents

Transcript of Knock-out mouse for Canavan disease: a model for gene transfer to the central nervous system