9732 The Journal of Neuroscience, October 29, 2003 23(30):97329741

Development/Plasticity/Repair

Neuronal Migration Depends on Intact Peroxisomal Function in Brain and in Extraneuronal Tissues
Anneleen Janssen,1 Pierre Gressens,3 Markus Grabenbauer,4 Eveline Baumgart,4 Arno Schad,4 Ilse Vanhorebeek,1 Annelies Brouwers,1 Peter E. Declercq,1 Dariush Fahimi,4 Philippe Evrard,3 Luc Schoonjans,5 De sire Collen,6 6 2 2 1 Peter Carmeliet, Guy Mannaerts, Paul Van Veldhoven, and Myriam Baes
Laboratory of Clinical Chemistry and 2Department of Pharmacology, K. U. Leuven, 3000 Leuven, Belgium, 3Laboratory of Developmental Neurology, Hospital Robert-Debre , 75019 Paris, France, 4Department of Anatomy and Cell Biology II, University of Heidelberg, 69117 Heidelberg, Germany, 5ThrombX, 3000 Leuven, Belgium, and 6Center for Transgene Technology and Gene Therapy, Flanders Interuniversity Institute for Biotechnology, 3000 Leuven, Belgium
1

Functional peroxisome deficiency, as encountered in Zellweger syndrome, causes a specific impairment of neuronal migration. Although the molecular mechanisms underlying the neuronal migration defect are at present unknown, the excess of very long chain fatty acids in brain, a consequence of peroxisomal -oxidation deficiency, has often been hypothesized to play a major role. The purpose of the present study was to investigate the contribution of peroxisomal dysfunction in brain as opposed to peroxisomal dysfunction in extraneuronal tissues to the migration defect. Peroxisomes were selectively reconstituted either in brain or liver of Pex5 knock-out mice, a model for Zellweger syndrome, by tissue-selective overexpression of Pex5p. We found that both rescue strains exhibited a significant correction of the neuronal migration defect despite an incomplete reconstitution of peroxisomal function in the targeted tissue. Animals with a simultaneous rescue of peroxisomes in both tissues displayed a pattern of neuronal migration indistinguishable from that of wild-type animals on the basis of cresyl violet staining and 5,3-bromo-2-deoxyuridine birth-dating analysis. These data suggest that peroxisomal metabolism in brain but also in extraneuronal tissues affects the normal development of the mouse neocortex. In liver-rescued mice, the improvement of the neuronal migration was not accompanied by changes in very long chain fatty acid, docosahexaenoic acid, or plasmalogen levels in brain, indicating that other metabolic factors can influence the neuronal migration process. Key words: neuronal migration; peroxisome; Zellweger syndrome; VLCFA; plasmalogen; DHA

Introduction
The laminar organization of neocortex depends on the radial migration of numerous neurons from the subventricular zone toward the surface using radial glial cells as a scaffold (Hatten, 1999; Gressens, 2000). Although the molecular mechanisms controlling the migration process are not fully understood, recent studies have implicated a number of neuronal and glial ligand receptor systems and cytoskeleton-interacting proteins in this directed migration, including adhesion molecules, extracellular matrix proteins, cell surface receptors, neurotransmitters, and calcium influx (Komuro and Rakic, 1992; Rakic et al., 1994; Behar et al., 1996; Walsh, 1999).
Received May 1, 2003; revised Aug. 14, 2003; accepted Aug. 28, 2003. This work was funded by European Community Grant Biomed BMH4-98-3569, European Commission Grant QLG1-CT2001-01277, Fonds Wetenschappelijk Onderzock Vlaanderen Grant G.0280.97, and Geconcerteerde Onderzocksacties Grant 99/09. A.J. was an IWT fellow. We thank Dr. R. McKay [National Institutes of Health (NIH), Bethesda, MD], Dr. M. Brenner (NIH), and Dr. S. Tilghman (Princeton University, Princeton, NJ) for providing plasmids, Dr. S. Subramani (University of California, San Diego, CA) and Dr. G. Dodt ( Ruhr University, Bochum, Germany) for providing antisera, and L. Kiekens, S. Danloy, V. Kreemers, E. Meyhi, B. Das, L. Pauwels, I. Frommer, and G. Kra mer for technical help. Correspondence should be addressed to Dr. Myriam Baes, Laboratory of Clinical Chemistry, K. U. Leuven, Herestraat 49, Onderwijs and Navorsing, B-3000 Leuven, Belgium. E-mail: [Link]@[Link]. E. Baumgarts present address: Institute for Anatomy and Cell Biology, University of Giessen, 35390 Giessen, Germany. Copyright 2003 Society for Neuroscience 0270-6474/03/239732-10$15.00/0

Much less attention has been paid to metabolic factors controlling neuronal migration. Nonetheless, it is well known that peroxisome deficiency, as encountered in the rare inherited syndrome of Zellweger, is associated with a very characteristic impairment of neuronal migration (Evrard et al., 1978), resulting in gyral abnormalities of the cerebral cortex and heterotopias in neocortex, cerebellum, and inferior olivary complex. Patients suffering from Zellweger syndrome present with extreme hypotonia, neonatal seizures, and severe mental retardation and die usually within the first year of life. Zellweger syndrome is the prototype and most severe form of the peroxisome biogenesis disorders, a heterogeneous group of autosomal recessive diseases caused by the defective import of peroxisomal matrix or membrane proteins (Subramani et al., 2000; Terlecky and Fransen, 2000; Gould et al., 2001; Purdue and Lazarow, 2001). The defective import of peroxisomal proteins results in the loss of multiple peroxisomal functions attributable to inactivity of most unimported enzymes in the cytosol. The major metabolic consequences include increased levels of very long chain fatty acids (VLCFAs), bile acid intermediates, and 2- and 3-methylbranched chain fatty acids, a depletion of ether phospholipids, including plasmalogens and platelet-activating factor, and reduced levels of the polyunsaturated fatty acid docosahexaenoic acid (DHA, C22:6n-3). Despite extensive knowledge of the met-

Janssen et al. Neuronal Migration and Peroxisomes

J. Neurosci., October 29, 2003 23(30):97329741 9733

abolic role of peroxisomes, to date it remains unclarified how defective peroxisomal function influences neuronal mobility during the migration process (Powers, 1995). Animal models of peroxisome biogenesis disorders have been generated by targeted inactivation of either the Pex5 gene (Baes et al., 1997), encoding the import receptor for most peroxisomal matrix proteins, or the Pex2 gene (Faust and Hatten, 1997), encoding a peroxisomal membrane protein. Both knock-out mouse models displayed intrauterine growth retardation, were severely hypotonic at birth, died within 72 hr, and exhibited the known metabolic abnormalities of Zellweger patients. Neurodevelopmental analysis of peroxisome-deficient fetuses revealed a neuronal migration defect in neocortex, a delay in neuronal maturation, significantly increased apoptosis in the cortical plate, and structural abnormalities in the inferior olivary nucleus, all typical of Zellweger syndrome (Baes et al., 1997; Faust and Hatten, 1997). The purpose of the present study was to investigate the contribution of peroxisomal dysfunction in brain as opposed to peroxisomal dysfunction in extraneuronal tissues to the neuronal migration impairment in Pex5 knock-out mice. To this end, transgenic mice were generated with selective expression of peroxisomes in either brain or liver, an organ with abundant peroxisomes, by tissue-selective overexpression of the Pex5 protein (Pex5p) in the knock-out mice.

Materials and Methods

Transgenic DNA constructs
The full-length coding region of the long form of mouse Pex5 cDNA was generated by reverse transcription (RT)-PCR using Expand reverse transcriptase (Roche, Brussels, Belgium) on mouse liver RNA and Expand high-fidelity PCR enzyme mix (Roche) using the primers 5-TACTACGGCGCGCCATGGCAATGCGGGAGCTGGTGGAG-3 (forward) and 5-GTGGTGTCTAGATCACTGGGGCAGGCCAAACATAGC-3 (reverse). The Pex5L cDNA and the myc9E10 epitope preceded by the Kozak consensus sequence were subcloned in a modified pNEB vector (New England Biolabs, Leusden, The Netherlands). For the brain-selective reconstitution of Pex5p, the nestin second intron was subcloned from the p401ZgII plasmid provided by Dr. R. McKay [National Institutes of Health (NIH), Bethesda, MD] (Zimmerman et al., 1994). The thymidine kinase (TK) minimal promoter was amplified by PCR, and the 0.6 kb mouse protamine 1 gene (mP1) segment containing an intron, a 3 untranslated region and, a polyadenylation signal was taken from the pGfa2lac-1 plasmid obtained from Dr. M. Brenner (NIH). On both sides of the nestin/TK-mycPex5-mp1 construct, an 800 bp insulator sequence was subcloned, derived from the human -globin 3-DNase I-hypersensitive site (Fleenor and Kaufman, 1993), and obtained by PCR on HeLa cell genomic DNA. For the liver-selective reconstitution of Pex5p, the 7.6 kb 5 flanking region of the -fetoprotein (AFP) gene (Godbout et al., 1986; Hammer et al., 1987) was subcloned from the pAFP vector provided by S. Tilghman (Princeton University, Princeton, NJ). The AFP-mycPex5-mP1 construct was inserted into the pLA39 vector in between two identical insulator sequences, each consisting of two copies of a 1.2 kb fragment of the chicken -globin gene (Pikaart et al., 1998; Potts et al., 2000). More details on the assembly of the vectors destined for the brain- or liver-selective overexpression of Pex5p are available on request. A diagram of these constructs is shown in Figure 1.

Figure 1. Transgene constructs and identification of Pex5 rescue mice. A, B, Diagrams of the transgene constructs designed for brain- and liver-selective expression of mycPex5p, respectively. iso, Isolator; Nes/TK, nestin second intron fused to thymidine kinase promoter. C, Southern blot analysis of BamHI-digested tail DNA identifying the wild-type (WT) and recombinant (HR) Pex5 allele as well as the inserted transgene (TR). Lanes 1, 4, Genomic DNA of LR:Pex5/ and BR:Pex5/ mice; lanes 2, 3, DNA of Pex5/ and Pex5/ mice.

Generation, identification, and breeding of transgenic mice

The nestin-TK-mycPex5 transgene was liberated by PacI digestion, whereas the plasmid containing the AFP-mycPex5 transgene was linearized with PvuI. Transgenes were gel-purified, diluted to a final concentration of 5 ng/l, and used for microinjection into the male pronucleus of fertilized eggs of the FVB/N strain. The incorporation of the transgene was examined in 3-week-old mice by Southern blot analysis of tail DNA using BamHI digestion and a 5 genomic Pex5 probe, previously devel-

oped to distinguish the wild-type from the recombinant allele in Pex5/ mice (Baes et al., 1997) (Fig. 1C). Alternatively, PCR analysis was performed to demonstrate the transgene using as forward primers 5CTCACTGAAGGTTACTAGTTAACAGGC-3 (AFP promoter) and 5TACTACAAGCTTGGCCCCGCCCAGCGTCTT-3 (TK promoter) and the previously mentioned Pex5 reverse primer. On the basis of Southern blot analysis, it was estimated that the copy number of the inserted transgene was 10 copies in the nestin-TK-mycPex5 transgenic line and two copies in the AFP-mycPex5 transgenic line. Founder mice were crossed with Pex5/ mice (inbred in the Tac:[Sw]fBR strain), and the nestinPex5:Pex5/ mice identified in the offspring were again mated with Pex5/ mice to generate nestin-Pex5:Pex5/ mice (further denoted BR:Pex5/ mice, for brain-rescued Pex5 knock-out mice). In a similar way, AFP-Pex5:Pex5/ mice (further denoted LR:Pex5/ mice, for liver-rescued Pex5 knock-out mice) were obtained. In this breeding scheme, the rescue mice and Pex5/ mice are generated at an equal frequency (one per eight) allowing the direct comparison of their phenotypes. LR:Pex5/ and BR:Pex5/ mice were further intercrossed to obtain double-rescue mice denoted LBR:Pex5/ mice (for liver- and brain-rescued Pex5 knock-out mice). Pex5/ or Pex5/ littermates with or without transgene incorporation were used as controls in all studies. Pregnancies were staged as 0.5 d at 12 PM after detection of a vaginal plug. The mice were bred in the animal housing facility of the University of Leuven under conventional conditions. They had unlimited access to standard rodent food chow and water and were kept on a 12 hr light/dark cycle. All animal experiments were approved by the Institutional Animal Ethical Committee of the University of Leuven.

Biochemical analysis
For biochemical analysis of mouse fetuses [embryonic day 14.5 (E14.5) and E18.5], pregnant females were killed by cervical dislocation; the fe-

9734 J. Neurosci., October 29, 2003 23(30):97329741

Janssen et al. Neuronal Migration and Peroxisomes

tuses were removed; and tissues were immediately frozen in liquid nitrogen. RNA analysis. RNA extraction from tissues (20 100 mg) and Northern blot analysis were performed as previously described (Huyghe et al., 2001). To remove residual genomic DNA for the evaluation of transgene expression by RT-PCR, 10 g of total RNA was treated for 15 min with 20 U of DNase I (Roche) at 37C in a buffer consisting of (in mM): 25 Tris-HCl, pH 8.3, 75 KCl, and 3 MgCl2. Superscript reverse transcriptase (Invitrogen, Merelbeke, Belgium) and 0.5 g of oligo-dT1218 was used to generate cDNA, followed by PCR with the Pex5 forward and reverse primers previously mentioned. Western blot analysis. Western blot analysis on homogenates from entire liver, brain, or neocortex was performed as previously described (Huyghe et al., 2001). Rabbit polyclonal antibodies directed to the human Pex5p (a gift from G. Dodt, Ruhr University, Bochum, Germany), the 52 kDa subunit of the rat acyl-CoA oxidase (Van Veldhoven et al., 1994), and the 41 kDa rat 3-ketoacyl-CoA thiolase (Antonenkov et al., 1997) were used. Determination of dihydroxyacetonephosphate acyl transferase activity. Twenty five microliters of tissue homogenate (1:5, w/v) prepared in 5 mM 4-morpholinepropanesulfonic acid, pH 7.2, 1 mM EDTA, and 0.25 M sucrose were used in the assay previously described (Jones and Hajra, 1994). Determination of hexacosanoic acid, C22: 6n-3, and plasmalogen levels. These lipids were quantified according to previously published protocols (Janssen et al., 2000).

Analysis of peroxisomes
Electron microscopic analysis of peroxisomes in liver. For the electron microscopic analysis of liver, newborn mice were perfused transcardially with 4% (w/v) depolymerized paraformaldehyde and 0.05% (v/v) glutaraldehyde in 1 PBS, pH 7.4, and 2% (w/v) sucrose. For detection of peroxisomes, the peroxidatic activity of catalase was demonstrated by the alkaline 3,3diaminobenzidine procedure (Fahimi, 1969) with subsequent postfixation in reduced osmium and Epon 812 embedding. Immunocytochemical and immunohistochemical analysis of peroxisomes in brain. Fetuses were removed and immersion-fixed (E14.5) or perfused intracardially (E18.5) with and functionality of mycPex5 transgene expression. Analyses of BR:Pex5/ mice are shown on the left; 4% freshly depolymerized paraformaldehyde Figure 2. Selectivity / analyses of LR:Pex5 mice are shown on the right. A, B, RT-PCR analysis of the indicated tissues derived from E18.5 Pex5 rescue in PBS, pH 7.4, and embedded in paraffin. For mice. Sc, Spinal cord; B, brain; L, liver; S, stomach; C, colon; K, kidney; H, heart; Lu, lung; M, skeletal muscle; Bo, bone. C, D, immunofluorescent detection of catalase, the antigen retrieval process consisted of treatment Immunoblots of brain (E14.5; C), neocortex (E18.5; C), and liver homogenates (E14.5, E18.5; D) using an antibody specific for with 0.1% trypsin for 5 min, followed by 15 min Pex5p. The full-length protein is indicated by an arrow; a shorter form appearing in the rescue tissues is indicated by an arrowhead. of microwaving (850 W) in citrate buffer, pH 6 KO, Knock-out; CT, control. E, F, Immunoblots of neocortex (E18.5; E) and liver homogenates (E18.5; F ) using an antibody specific (Grabenbauer et al., 2001). After incubation for the 52 kDa subunit of ACO. The positions of the unprocessed 71 kDa and the processed 52 kDa bands are indicated. G, H, with a monospecific antibody to catalase Restoration of DHAPAT activity in brain and liver of E14.5 and E18.5 Pex5 rescue mice. Values are expressed as percentages of (Baumgart et al., 1989) the antigen-binding activities in control mice of the same age. sites were detected by an Alexa 488-conjugated secondary antibody (Molecular Probes, Eugene, OR). The nuclei were counterstained vided by M. Fransen, K. U. Leuven, Leuven, Belgium; results not shown with propidium iodide, and the sections of neocortex were analyzed by a in figures). Antigen retrieval was achieved by boiling sections for 20 min confocal laser scanning microscope (CLSM; Leica, Heidelberg, Germain citrate buffer, pH 6, in a steam cooker. After blocking of endogenous ny). MycPex5p expression was detected by anti-Pex5p (antibody properoxidase with 3% H2O2 and blocking of endogenous biotin with an avidin biotin blocking kit (PerkinElmer Life Sciences, Boston, MA), vided by G. Dodt) or anti-myc immunohistochemistry (antibody pro-

Janssen et al. Neuronal Migration and Peroxisomes

J. Neurosci., October 29, 2003 23(30):97329741 9735

studies performed in Pex5 knock-out mice (Baes et al., 1997; Gressens et al., 2000), the density of BrdU-stained cells was measured in the ventricular zone, intermediate zone (prospective white matter), and neocortical plate and was used as an index of the severity of the neuronal migration disorder. To avoid regional and experimental variations in labeling, sections from the different experimental groups including comparable anatomic regions in the frontoparietal area were treated simultaneously. Counts of BrdU-positive cells were performed by hand, using a counting grid under a 40 magnification objective. The counts were performed by an observer blind to the different experimental groups in a sector measuring 500 m in length in the coronal plane within the frontoparietal cerebral wall (see Fig. 6 F). This sector was divided into a ventricular zone (average surface area, 0.03 mm 2), an intermediate zone (prospective white matter; average surface area, 0.05 mm 2), and a neocortical plate (average surface area, 0.1 mm 2) (Takahashi et al., 1992). For each experimental group, cells were counted in 10 different fields (five brains from three different litters, two nonadjacent sections of the right hemisphere per brain). Both the intensely and the weakly labeled cells were counted. To further confirm these quantitative data on a larger sample, we performed extensive counts for the most important data set, which comprised BrdU-positive cells in the intermediate zone after injection of BrdU at E13.5. In this subset of analysis, cells were counted in 40 different fields (five brains from three different litters, eight nonadjacent sections of the right hemisphere per brain, each section being separated from the next one by a fixed 14 m distance; see supplementary material, available at [Link]).

Neural cell death analysis

Figure 3. Reconstitution of peroxisomes in the cortex. AD, Immunohistochemical detection of the mycPex5 fusion protein using a Pex5p antibody. Pex5p is not detectable in neocortex of E14.5 knock-out mice ( A) but is strongly expressed in E14.5 ( B) Pex5 BR mice. Pex5 staining, expressed strongly in all layers of E18.5 wild-type mice ( C), is markedly diminished in the outer cortical layers of E18.5 BR:Pex5/ mice but is still present in the germinative zone ( D). EH, High-resolution CLSM analysis of catalase distribution in E14.5 and E18.5 neocortex. The selected neocortex areas for CLSM analysis are marked by asterisks in B and D. In all cells of wild-type (WT) mice ( E), catalase is present in a punctate pattern (green dots), representing peroxisomes. In Pex5 knock-out mice ( G), immunofluorescent detection of catalase results in diffuse green fluorescent staining in the cytoplasm and light blue staining of the nuclei of all cells. In BR:Pex5/ mice (F, H ), the staining pattern of catalase is punctate in neural cells but diffusely cytosolic and nuclear in endothelial cells (EC). Scale bars: AD, 100 m; EH, 10 m. To document neocortical cell death, sections immediately adjacent to those used for BrdU staining were used for immunodetection of cleaved caspase 3 (rabbit polyclonal antibody; Cell Signaling, Beverly, MA) and for terminal deoxynucleotidyl transferase-mediated biotinylated UTP nick end-labeling (TUNEL) staining (in situ cell death detection kit; Roche, Meylan, France). Cleaved caspase 3-labeled cells and TUNEL-labeled nuclei were counted in a 1 mm 2 area in the neocortical plate at the level of the frontoparietal cerebral wall. For each experimental group, 10 different fields (five brains from three different litters, two nonadjacent sections of the right hemisphere per brain) were analyzed. Quantitative data are expressed as mean SEM for each treatment group. Results were compared using ANOVA with Bonferronis multiple comparison of means test. Differences were considered statistically significant at p 0.05.

antigen-binding sites were detected with a peroxidase-coupled biotin avidin system (rabbit ExtraVidin kit; Sigma, Munich, Germany) and visualized by histochemical staining for peroxidase using Novared (Vector Laboratories, Burlingame, CA) as a substrate. Nuclei were counterstained with hematoxylin.

Statistical analysis

Neuronal migration analysis

To quantify the defects of neuronal migration, pulse chase experiments with 5,3-bromo-2-deoxyuridine (BrdU) were performed. Pregnant dams carrying transgenic and wild-type mice were injected intraperitoneally at E13.5 or E15.5 with 50 mg/kg BrdU (Sigma). Mothers were killed at E18.5 by decapitation; fetuses were removed; and brains were fixed by immersion fixation in 70% ethanol. After paraffin embedding, 7 m sections were cut coronally and either stained with cresyl violet or used for immunohistochemical detection of BrdU (mouse monoclonal antibody; Becton Dickinson, San Jose, CA). On the basis of previous

Results

Generation of transgenic mice with brain- or liver-selective expression of Pex5p To reconstitute peroxisomes in fetal Pex5 knock-out mice in a tissue-selective way, transgenic mice were generated expressing

9736 J. Neurosci., October 29, 2003 23(30):97329741

Janssen et al. Neuronal Migration and Peroxisomes

myc-tagged Pex5p under the control of a tissue-selective promoter. The mycPex5 fusion protein was able to restore the import of peroxisomal matrix proteins in Pex5/ fibroblasts (data not shown). For the brain-selective reconstitution of Pex5p, the second intron of the nestin intermediate filament gene fused to the thymidine kinase minimal promoter was used (Zimmerman et al., 1994; Yamaguchi et al., 2000). This promoter restricts transgene expression to neuroepithelial cells that are precursors for the neuronal and glial lineage. The 7.6 kb 5 flanking region of the -fetoprotein gene was used to overexpress Pex5p in hepatocytes (Godbout et al., 1986; Hammer et al., 1987; Potts et al., 2000) (Fig. 1 A,B). Pronuclear zygote injection yielded three and two founder mice that had inserted the nestin-mycPex5 and the AFP-mycPex5 constructs, respectively. After crossbreeding with Pex5/ mice, one of the AFP-mycPex5 founders apthe liver of LR:Pex5/ mice. Electron micrographs of hepatocytes of newborn peared to be infertile, and from one of the Figure 4. Reconstitution of peroxisomes in / ( CE) mice are shown. Glycogen deposits (Gly), peroxisomes (PO), and lipid nestin-mycPex5 founders, no pups with control ( A), Pex5 knock-out ( B), and LR:Pex5 droplets (Lip) are marked. Crystalline deposits presumably consisting of VLCFAs are indicated by arrowheads in E. D, E, Hepatocytes the desired genotype were obtained. Furof LR:Pex5/ mice with reconstituted peroxisomes (1) and with absence of peroxisomes and abnormal mitochondrial structure ther phenotypic analysis was performed (2; asterisks) are shown. BC, Bile canaliculus; Nuc, nucleus. on one mouse line expressing mycPex5p in liver (denoted LR:Pex5 mice) and on one of the two mouse lines expressing mycPex5p in brain (denoted BR:Pex5 mice). Characterization of transgenic mice with brain- or liverselective expression of mycPex5p As illustrated in Figures 2 and 3, the selectivity and functionality of transgene expression were evaluated in the relevant tissues of LR:Pex5/ and BR:Pex5/ mice. RT-PCR analysis on a variety of tissues of E18.5 fetuses revealed that mycPex5 cDNA was only demonstrable in brain and spinal cord of a BR:Pex5/ mouse and in liver and kidney of a LR:Pex5/ mouse, in correspondence with the promoter activity present in the constructs (Fig. 2 A,B). To assess the expression of the Pex5 protein, immunoblotting and immunocytochemical experiments were performed. In BR:Pex5/ mice, mycPex5p was detectable in E14.5 brain homogenates and in E18.5 neocortex but was present at lower levels than the endogenous Pex5p in a wild-type mouse (Fig. 2C). In liver homogenates of LR:Pex5/ mice, mycPex5p was detectable at both E14.5 and E18.5 (Fig. 2 D). As expected, no Pex5p immunoreactivity was observed in homogenates prepared from brain or liver of a generalized Pex5 knock-out mouse. In liver as well as brain, the transgene gave rise to the full-length Pex5 protein as well as to a shorter protein (Fig. 2C,D, arrowheads). Because this immunoreactive band was not detected with an antibody directed to the N terminus of Pex5p, we assume that this is a shorter Pex5 product that is generated by using an alternative translation start site. These data were confirmed by strong immunocytochemical staining of Pex5p in all layers of the cortex of E14.5 BR:Pex5/ mice whereas no staining was observed in knock-out animals (Fig. 3A,B). In E18.5 BR:Pex5/ mice, Pex5p staining was strongly reduced in the outer cortical layers but was still very intense in the germinative zone (Fig. 3D). Subsequently, we examined whether MycPex5p expressed in the rescue mice was capable of restoring peroxisomal matrix im-

Figure 5. Peroxisomal metabolic parameters in brain of BR:Pex5/, LR:Pex5/, and LBR:Pex5/ mice. The levels of C26:0, DHA, and plasmalogens were measured in brain of E14.5 control (CT), complete Pex5 knock-out (KO), and BR:Pex5/ mice (left plots) and E18.5 control, complete knock-out, BR:Pex5/, and LR:Pex5/ mice (right plots). Plasmalogen levels were also measured in E18.5 LBR:Pex5/ mice. Values are expressed as percentages of levels in age-matched controls.

port. Therefore, the processing of acyl-CoA oxidase, a protein with a C-terminal peroxisomal-targeting signal 1 (PTS1), and peroxisomal 3-ketoacyl-CoA thiolase, a protein with an N-terminal cleavable targeting signal 2 (PTS2), was monitored in liver and brain homogenates by Western blot analysis. It is indeed

Janssen et al. Neuronal Migration and Peroxisomes

J. Neurosci., October 29, 2003 23(30):97329741 9737

known to be inactive in Zellweger patient fibroblasts (Datta et al., 1984). DHAPAT activity was not detectable in liver or brain homogenates of the generalized Pex5 knock-out mice at E14.5 and at E18.5, confirming previous results (Baes et al., 1997). In liver of LR:Pex5/ mice, DHAPAT activity was restored to 30% of wild-type levels at both E14.5 and E18.5 (Fig. 2 H). This is in good agreement with the partial restoration of plasmalogen levels (6.4 pmol/nmol phospholipids in knock-out mice, 15.4 pmol/nmol in LR: Pex5/ mice, and 28.2 pmol/nmol in wild-type mice) and of urate oxidase activity (25 4% of wild type activity in LR: Pex5/ mice, not measurable in Pex5 knock-out mice). In brain of BR:Pex5/ mice, DHAPAT activity was restored to 70% of wild-type levels at E14.5 but was decreased again to 50% of normal levels in brain of E18.5 pups (Fig. 2G). To demonstrate the presence of structurally intact peroxisomes in the cortex of BR:Pex5/ mice, immunocytochemical staining of the peroxisomal marker enzyme catalase was performed. As expected, a punctate fluorescence pattern was observed in all cell types of wild-type mice (Fig. 3E), and diffuse cytoplasmic staining was observed in both neural and endothelial cells of Pex5/ animals (Fig. 3G). In the absence of peroxisomes, catalase also resided in the nucleus, as previously observed (Baes et al., 1997). Importantly, in the cortex of E14.5 and E18.5 / mice (Fig. 3F,H ), catalase imFigure 6. Neuronal migration analysis in BR:Pex5/, LR:Pex5/ and LBR:Pex5/ mice. AE, Cresyl violet-stained coronal BR:Pex5 sections of E18.5 mice with the indicated genotypes. mz, Marginal zone; cp, cortical plate; iz, intermediate zone; vz, ventricular munoreactivity was found as a punctate zone. F, Frontoparietal region of the murine cerebral wall at E18.5. Nuclei in S-phase are immunostained (dark deposit) with pattern in all cells except endothelial cells. anti-BrdU antibody, and the tissue is counterstained with cresyl violet. BrdU (50 mg/kg) was injected on E13.5. The 500-m-wide In comparison with wild-type mice, fewer sector on which the quantitative analyses are based is enclosed within the vertical arrows. Labeled nuclei were counted (see Fig. peroxisomes were found in neural cells of 7) within these two vertical lines, in the vz, iz, and cp. E14.5 mice, and even fewer were found in E18.5 BR:Pex5/ mice (Fig. 3E,F,H ). known that the proteolytic cleavage of these enzymes depends on However, in the latter mice, no sign of cytosolic or nuclear localtheir import into peroxisomes. As shown in Figure 2, E and F, the ization of catalase was found in any of the cortical layers, indicatprocessed form of acyl-CoA oxidase was not detectable in liver or ing that even the diminished levels of Pex5p in the outer cortical brain of generalized Pex5 knock-out mice. In contrast, the 52 kDa layers were sufficient to allow import of peroxisomal proteins. In subunit of acyl-CoA oxidase was found in neocortex of E18.5 liver of three LR:Pex5/ mice, examined by electron microBR:Pex5/ mice and in liver of E18.5 LR:Pex5/ mice, as also scopic analysis after cytochemical staining for catalase, structurobserved in the respective wild-type tissues. Similarly, ally intact peroxisomes were found in a fraction of hepatocytes 3-ketoacyl-CoA thiolase was processed in the target tissue of the (10 40%) but not in Kupffer or endothelial cells (Fig. 4CE). rescue mice (data not shown), suggesting that the import of PTS1 Interestingly, in the hepatocytes lacking catalase-positive peroxiand PTS2 proteins was restored in the two tissue-selective Pex5 somes, mitochondria with abnormal cristae were found (Fig. rescue strains. In the kidney of LR:Pex5/ mice, no processing of 4 D,E, asterisk) as previously observed in liver of generalized Pex5 acyl-CoA oxidase was observed on Western blots, indicating that knock-out mice (Baumgart et al., 2001). the Pex5 transcripts found in this organ were insufficient to generate functional Pex5 protein (data not shown). Phenotypic analysis of transgenic mice with brain- or liverBecause the neuronal migration defects in Pex5 knock-out selective reconstitution of peroxisomes mice predominantly manifest between E14.5 and E18.5, it is cruMacroscopic evaluation cial that peroxisomal function resumes in the targeted tissue beOn macroscopic examination at birth, LR:Pex5/ and BR: Pex5/ mice were indistinguishable from the generalized Pex5 fore this stage. This was investigated by measuring the activity of knock-out mice (Baes et al., 1997), whereas they could easily be dihydroxyacetonephosphate acyl transferase (DHAPAT), an endiscriminated from wild-type littermates. They displayed intrazyme of the ether phospholipid synthesis pathway, which is

9738 J. Neurosci., October 29, 2003 23(30):97329741

Janssen et al. Neuronal Migration and Peroxisomes

uterine growth retardation with an average 30% body weight reduction at birth compared with control littermates. Newborn LR:Pex5/, BR:Pex5/, and generalized Pex5/ pups were severely hypotonic, unable to support their body weight on their legs and to feed themselves, and they kept a contracted posture. Most died within 24 hr after birth, and none survived 48 hr. Double-rescue mice (LBR:Pex5/), obtained by intercrossing LR:Pex5/ and BR: Pex5/ mice, exhibited the same macroscopic phenotype and had the same life span as the single rescue or generalized Pex5 knock-out mice. Biochemical analysis In peroxisome deficiency disorders, the accumulation of VLCFA and the depletion of plasmalogens and C22:6n-3 have been hypothesized to be related to the neurodevelopmental disturbances. To correlate the levels of these compounds in the BR:Pex5/ and the LR:Pex5/ mice with the brain phenotype, their concentrations were determined in phospholipids extracted from whole brain of E14.5 and E18.5 fetuses. In brain of E14.5 and E18.5 Pex5/ fetuses, hexacosanoic acid (C26:0) levels were twofold to threefold elevated, in agreement with previous reports (Baes et al., 1997; Janssen et al., 2000). As shown in Figure 5, A and B, the C26:0 concentration was normalized in brain of BR:Pex5/ mice at both fetal ages. Also, the concentration of DHA in brain, which was 30% reduced in generalized Pex5 knock-out mice at E18.5 or birth (Janssen et al., 2000), was restored to wild-type values in BR:Pex5/ mice (Fig. 5E,F ). Finally, the severe depletion of plasmalogens in brain of Pex5/ mice was fully normalized in E14.5 BR:Pex5/ fetuses (Fig. 5C). However, in brain of E18.5 BR:Pex5/ pups (Fig. 5D), plasmalogen levels were only 60% of wild-type values, which is in agreement with the decline of DHAPAT activity in brain of BR:Pex5/ mice at the end of Figure 7. Quantitative analysis of BrdU-labeled cells in BR:Pex5/, LR:Pex5/, and LBR:Pex5/ mice. Counts of BrdUgestation. labeled cells at E18.5 in the germinative zone, intermediate zone, and neocortical plate after injection of BrdU into pregnant These peroxisomal metabolic parame- animals at E13.5 or E15.5 are shown. The average counts SEM of five different mice of each genotype are represented. ters were also monitored in brain of mice Statistically significant differences from black bars (*) or hatched bars (): p 0.05; **,p 0.01; ***,p 0.001, with reconstituted peroxisomal function ANOVA with Bonferronis multiple-comparison test. in liver. C26:0 levels were elevated and with a delay of neuronal migration (Fig. 6 A,B). Interestingly, in plasmalogens and C22:6n-3 were depleted to the same extent in both BR:Pex5/ and LR:Pex5/ mice, a significant improvebrain of LR:Pex5/ mice (Fig. 5B,D,F ) as in generalized Pex5 / knock-out mice. In the double-rescue LBR:Pex5 mice, levels of ment of the neuronal migration process was observed. In both plasmalogens in brain were comparable with the levels found in cases, fewer neurons were located in the intermediate zone comBR:Pex5/ mice (Fig. 5D). pared with Pex5 knock-out mice (Fig. 6C,D). Finally, in doublerescue LBR:Pex5/ mice, neuronal migration appeared to be Neocortical neuronal migration normalized because the density of cells in the intermediate zone In agreement with previous reports, examination of cresyl violet/ in the double-rescue mice was indistinguishable from that in stained coronal sections of E18.5 Pex5 pups revealed altered wild-type mice (Fig. 6 E). cell densities in the cortical plate and white matter, consistent

Janssen et al. Neuronal Migration and Peroxisomes

J. Neurosci., October 29, 2003 23(30):97329741 9739

To quantify the neuronal migration phenotype, BrdU pulse chase experiments were performed. Pregnant animals were given a single BrdU injection at E13.5 or at E15.5 of gestation, and the numbers of BrdU-labeled nuclei present in the germinative zone, intermediate zone, and neocortical plate were monitored at E18.5 (Fig. 6 F). As shown in Figure 7, a 2.4-fold increase of BrdUlabeled nuclei was observed in the intermediate zone of Pex5 knock-out brains compared with wild-type littermates, whereas a 48% reduction of BrdU-labeled nuclei was observed in the neocortical plate (Fig. 7E,F ), and no change was detectable in the germinative zone (Fig. 7A,B), confirming previous reports (Baes et al., 1997). In LR:Pex5/ and in BR:Pex5/ mice, this accumulation in the intermediate zone was significantly reduced, with 1.8- and 1.4-fold more nuclei in the intermediate zone in comparison with wild-type mice (Fig. 7C,D). Similarly, the paucity of labeled cells in the neocortical plate was significantly improved in LR:Pex5/ and BR:Pex5/ mice, with 23 and 29% reductions, respectively (Fig. 7E,F ). Finally, double-rescue LBR:Pex5/ mice exhibited BrdU counts that were not significantly different from wild-type values, in both the intermediate zone and neocortical plate (Fig. 7C,E). More extensive counts (eight instead of two sections per brain) of BrdU-positive cells in the intermediate zone yielded essentially the same results as the data reported above (see supplementary material, available at [Link]. org). Only the slight difference between the BrdU counts in the intermediate zone of the LR versus the BR mice (Fig. 7C) became smaller, indicating that there is no important difference in restoration of the migration defect between the two rescue strains. Neocortical neural cell death Confirming previous results (Baes et al., 1997), increased neural cell death was found in the neocortical plate of Pex5 / mice when compared with wild-type animals using both immunodetection of cleaved caspase-3 (Fig. 8 AE) and TUNEL staining (Fig. 8 F). In both BR:Pex5/ and LR:Pex5/ mice, a significant reduction of neural cell death was observed (Fig. 8 E,F ). In double-rescue LBR:Pex5/ mice, cleaved caspase-3 counts were close to wild-type values, whereas counts of TUNEL-labeled nuclei were greatly reduced when compared with Pex5/ mice but still significantly higher than in Pex5/ mice (Fig. 8 E,F ).

Figure 8. Cell death analysis in BR:Pex5/, LR:Pex5/ and LBR:Pex5/ mice. Shown are caspase-3 staining in wild-type ( A), Pex5 knock-out ( B), BR:Pex5/ ( C), and LR:Pex5/ ( D) neocortex and counts of cleaved caspase-3-positive ( E) or TUNEL-labeled ( F) cells at E18.5 in the neocortical plate. Statistically significant differences from black bars (*) or hatched bars (): **p 0.01; ***,p 0.001, ANOVA with Bonferronis multiple-comparison test.

Discussion
Although the adverse effects of defective peroxisomal function on the formation of cortex, cerebellum, and inferior olivary nucleus have been well documented, the precise molecular mechanisms causing these morphogenic defects are still enigmatic. The observation that the neuronal migration impairment also occurs in peroxisome-deficient transgenic mice (Baes et al., 1997; Faust and Hatten, 1997) opened new possibilities to study the links between peroxisomal function and neuronal migration. In the present study, we show that peroxisomal metabolism in brain as well as in extraneuronal tissues can influence the neuronal migration process in the mouse. In addition, by correlating peroxisomal metabolic parameters in brain to the migration phenotype, it seems unlikely that the accumulation of very long chain fatty acids, the depletion of plasmalogens, or the reduced content of DHA by themselves cause the migration defect in peroxisome deficiency disorders. Intact peroxisomal function in neuronal and extraneuronal tissues is essential for normal neuronal migration Both brain- and liver-selective reconstitution of peroxisomes in Pex5 knock-out mice resulted in a partial improvement of the

neuronal migration process in comparison with the defect observed in generalized Pex5 knock-out mice. Concomitantly, the neural cell death was less severe in the two rescue strains than in the Pex5 knock-out mice. In the Pex5 brain rescue strain, the reconstitution of peroxisomes appeared to be virtually complete during the critical period of neuronal migration, as shown by the presence of catalasepositive peroxisomes, the import of peroxisomal enzymes, and normalized levels of DHA and C26:0, at both E14.5 and E18.5. Plasmalogen content was completely restored in the brain at E14.5 but was again reduced to 60% of normal levels at E18.5, coincident with the lower activity of DHAPAT in E18.5 compared with E14.5 pups. Although we do not have a clear-cut explanation for this decline, it may be related to the temporary activity of the nestin enhancer, which is known to be switched off in postmitotic neurons (Zimmerman et al., 1994). As a result, mycPex5p levels may be lower in certain areas of the brain in the prenatal period. This did not seem to affect the function of peroxisomal oxidation but might have been limiting for the import of ether phospholipid-synthesizing enzymes. Immunocytochemical analysis of catalase in the cortex of E18.5 pups confirmed that peroxisomes were still present in the different cortical layers at that age. Taken together, we cannot exclude the possibility that the partial restoration of neuronal migration in the BR:Pex5/ mice is attributable to the incomplete reconstitution of peroxisomal function in the prenatal period. However, we favor the possibility that normal cortical development depends on functional peroxisomes in other organs as well. This idea was corroborated by the finding that a significant improvement of neuronal migration was observed in Pex5 liver rescue mice, even though the restora-

9740 J. Neurosci., October 29, 2003 23(30):97329741

Janssen et al. Neuronal Migration and Peroxisomes

tion of peroxisomal function in liver was only 30%. In addition, simultaneous peroxisomal reconstitution in both liver and brain resulted in a pattern of neuronal migration indistinguishable from that of wild-type mice with the currently used techniques and a level of apoptotic neural cell death intermediate between BR:Pex5/ and wild-type mice. The partial restoration of peroxisomal function in liver of LR:Pex5/ mice raises the question of whether complete restoration of hepatic peroxisomal function might normalize neuronal migration. An additional question is whether the impact of liver peroxisomes on brain development is specific for hepatocytes or whether it could also be exerted by peroxisomes in other tissues. Such questions might be resolved by generating mice with liver- and brain-selective depletion of peroxisomes using Pex5loxP mice (Baes et al., 2002a) and appropriate Cre-expressing mice. Pathogenic factors causing the neuronal migration defect in peroxisome deficiency disorders Increased levels of VLCFA have often been considered as the cause of the neuronal migration defect in Zellweger patients (Moser and Moser, 1996). The availability of transgenic mouse models allows the analysis of neuronal migration in concert with the measurement of VLCFA levels in brain. An important improvement of the neuronal migration process was observed in Pex5 liver rescue mice despite the fact that in brain, C26:0 was accumulating to the same extent as in the Pex5-deficient mice. Conversely, normalization of the C26:0 levels in brain of brain rescue mice did not lead to a full restoration of neuronal migration. These data do not support a role for VLCFA as the single cause of the neuronal migration defect in peroxisome deficiency and are in line with results in other mouse models. Indeed, abnormalities in the cortical lamination were seen in Pex11 knockout mice, which exhibit only minor changes in brain C26:0 levels (Li et al., 2002). On the other hand, in mice with selective peroxisomal oxidation defects (acyl-CoA oxidase and X-linked adrenoleukodystrophy knock-out mice), C26:0 accumulations were found in brain, but no neurodevelopmental abnormalities were reported (Fan et al., 1996; Forss-Petter et al., 1997; Kobayashi et al., 1997; Lu et al., 1997). Furthermore, in MFP-2 knock-out mice, which accumulate C26:0 in brain to the same extent as Pex5 knock-out mice at the time of birth, no signs of disturbed neuronal migration were found by using the same brain analysis procedures as for the Pex5 knock-out mice (Baes et al., 2002b). Thus, on the basis of the present and earlier experimental evidence, it seems unlikely that the excess of VLCFA on its own causes the neuronal migration defect in Pex5 knock-out mice, although it cannot be excluded that it is a contributing factor. In the brain of liver-rescued mice, besides the high levels of C26:0, a depletion of plasmalogens and a reduction of C22:6n-3 were also found. Therefore, the significant improvement of the neuronal migration defect in these mice seems to be mediated by the correction of other metabolic factors. After a landmark analysis of a Zellweger brain, Evrard et al. (1978) speculated that there is some type of toxic insult to both migrating neurons and radial glial cells. Increased levels of a neurotoxic compound are indeed among the likely pathogenic mechanisms that fit well with the beneficial role of hepatic peroxisomes on brain development documented in this study. We do not know which metabolic factors underlie the improved neuronal migration after local restoration of peroxisomal function in brain. Given the corrected levels of C26:0, the improved levels of plasmalogens, and the normalization of C22:

6n-3, it is possible that a combination of these and potentially other factors could be involved. In conclusion, the present findings are consistent with the notion that multiple metabolic disturbances are etiological to the migration disturbance in conditions of peroxisome deficiency. To further decipher the role of peroxisomes in the neuronal migration process, it will be necessary to conduct a more extensive metabolic investigation of the affected brain and to document possible alterations in the signaling molecules that are starting to be uncovered as key players in the control of the neuronal migration process.
Antonenkov VD, Van Veldhoven PP, Waelkens E, Mannaerts GP (1997) Substrate specificities of 3-oxoacyl-CoA thiolase A and sterol carrier protein 2/3-oxoacyl-CoA thiolase purified from normal rat liver peroxisomes. J Biol Chem 272:2602326031. Baes M, Gressens P, Baumgart E, Carmeliet P, Casteels M, Fransen M, Evrard P, Fahimi D, Declercq PE, Collen D, Van Veldhoven PP, Mannaerts GP (1997) A mouse model for Zellweger syndrome. Nat Genet 17:49 57. Baes M, Dewerchin M, Janssen A, Collen D, Carmeliet P (2002a) Generation of Pex5-IoxP mice allowing the conditional elimination of peroxisomes. Genesis 32:177178. Baes M, Gressens P, Huyghe S, De Nys K, Qi C, Jia Y, Mannaerts GP, Evrard P, Van Veldhoven PP, Declercq PE, Reddy JK (2002b) The neuronal migration defect in mice with Zellweger syndrome (Pex5 knockout) is not caused by the inactivity of peroxisomal -oxidation. J Neuropathol Exp Neurol 61:368 374. Baumgart E, Vo lkl A, Hashimoto T, Fahimi HD (1989) Biogenesis of peroxisomes: immunocytochemical investigation of peroxisomal membrane proteins in proliferating rat liver peroxisomes and in catalase-negative membrane loops. J Cell Biol 108:22212231. Baumgart E, Vanhorebeek I, Grabenbauer M, Borgers M, Declercq P, Fahimi HD, Baes M (2001) Mitochondrial alterations caused by defective peroxisomal biogenesis in a mouse model for Zellweger syndrome (PEX5 knockout mouse). Am J Pathol 159:14771494. Behar TN, Li Y-X, Tran HT, Ma W, Dunlap V, Scott C, Barker JL (1996) GABA stimulates chemotaxis and chemokinesis of embryonic cortical neurons via calcium-dependent mechanisms. J Neurosci 16:1808 1818. Datta NS, Wilson GN, Hajra AK (1984) Deficiency of enzymes catalyzing the biosynthesis of glycerol-ether lipids in Zellweger syndrome. N Engl J Med 311:1080 1083. Evrard P, Caviness VS, Prats-Vinas J, Lyon G (1978) The mechanism of arrest of neuronal migration in the Zellweger malformation: an hypothesis based upon cytoarchitectonic analysis. Acta Neuropathol (Berl) 41:109 117. Fahimi HD (1969) Cytochemical localization of peroxidatic activity of catalase in rat hepatic microbodies (peroxisomes). J Cell Biol 43:275288. Fan C-Y, Pan J, Chu R, Lee D, Kluckman KD, Usuda N, Singh I, Yeldandi AV, Rao MS, Maeda N, Reddy JK (1996) Hepatocellular and hepatic peroxisomal alterations in mice with a disrupted peroxisomal fatty acylcoenzyme A oxidase gene. J Biol Chem 271:24698 24710. Faust PL, Hatten ME (1997) Targeted deletion of the PEX2 peroxisome assembly gene in mice provides a model for Zellweger syndrome, a human neuronal migration disorder. J Cell Biol 139:12931305. Fleenor DE, Kaufman RE (1993) Characterization of the DNase I hypersensitive site 3 of the human globin gene domain. Blood 81:27812790. Forss-Petter S, Werner H, Berger J, Lassmann H, Molzer B, Schwab MH, Bernheimer H, Zimmermann F, Nave K-A (1997) Targeted inactivation of the X-linked adrenoleukodystrophy gene in mice. J Neurosci Res 50:829 843. Godbout R, Ingram R, Tilghman SM (1986) Multiple regulatory elements in the intergenic region between the -fetoprotein and albumin genes. Mol Cell Biol 6:477 487. Gould SJ, Raymond GV, Valle D (2001) The peroxisome biogenesis disorders. In: The metabolic and molecular bases of inherited disease (Scriver CR, Beaudet AL, Valle D, Sly WS, eds), pp 31813217. New York: McGraw-Hill. Grabenbauer M, Fahimi HD, Baumgart E (2001) Detection of peroxisomal proteins and their mRNAs in serial sections of fetal and newborn mouse organs. J Histochem Cytochem 49:155164.

References

Janssen et al. Neuronal Migration and Peroxisomes Gressens P (2000) Mechanisms and disturbances of neuronal migration. Pediatr Res 48:725730. Gressens P, Baes M, Leroux P, Lombet A, Van Veldhoven P, Janssen A, Vamecq J, Marret S, Evrard P (2000) Neuronal migration disorder in Zellweger mice is secondary to glutamate receptor dysfunction. Ann Neurol 48:336 343. Hammer RE, Krumlauf R, Camper SA, Brinster RL, Tilghman SM (1987) Diversity of alpha-fetoprotein gene expression in mice is generated by a combination of separate enhancer elements. Science 235:5358. Hatten ME (1999) Central nervous system neuronal migration. Annu Rev Neurosci 22:511539. Huyghe S, Casteels M, Janssen A, Meulders L, Mannaerts GP, Declercq PE, Van Veldhoven PP, Baes M (2001) Prenatal and postnatal development of peroxisomal lipid-metabolizing pathways in the mouse. Biochem J 353:673 680. Janssen A, Baes M, Gressens P, Mannaerts GP, Declercq P, Van Veldhoven PP (2000) Docosahexaenoic acid deficit is not a major pathogenic factor in peroxisome-deficient mice. Lab Invest 80:3135. Jones KM, Hajra AK (1994) Assay of dihydroxyacetone phosphate acyltransferase with 32P-labeled substrate. Clin Chem 40:946 947. Kobayashi T, Shinnoh N, Kondo A, Yamada T (1997) Adrenoleukodystrophy protein-deficient mice represent abnormality of very long chain fatty acid metabolism. Biochem Biophys Res Commun 232:631 636. Komuro H, Rakic P (1992) Selective role of N-type calcium channels in neuronal migration. Science 257:806 809. Li X, Baumgart E, Morrell JC, Jimenez-Sanchez G, Valle D, Gould SJ (2002) PEX11 deficiency is lethal and impairs neuronal migration but does not abrogate peroxisome function. Mol Cell Biol 22:4358 4365. Lu J-F, Lawler AM, Watkins PA, Powers JM, Moser AB, Moser HW, Smith KD (1997) A mouse model for X-linked adrenoleukodystrophy. Proc Natl Acad Sci USA 94:9366 9371. Moser HW, Moser AB (1996) Very long-chain fatty acids in diagnosis, pathogenesis, and therapy of peroxisomal disorders. Lipids 31:141144.

J. Neurosci., October 29, 2003 23(30):97329741 9741 Pikaart MJ, Recillas-Targa F, Felsenfeld G (1998) Loss of transcriptional activity of a transgene is accompanied by DNA methylation and histone deacetylation and is prevented by insulators. Genes Dev 12:28522862. Potts W, Tucker D, Wood H, Martin C (2000) Chicken -globin 5HS4 insulators function to reduce variability in transgenic founder mice. Biochem Biophys Res Commun 273:10151018. Powers JM (1995) The pathology of peroxisomal disorders with pathogenic considerations. J Neuropath Exp Neurol 54:710 719. Purdue PE, Lazarow PB (2001) Peroxisome biogenesis. Annu Rev Cell Dev Biol 17:701752. Rakic P, Cameron RS, Komuro H (1994) Recognition, adhesion, transmembrane signaling and cell motility in guided neuronal migration. Neurobiology 4:63 69. Subramani S, Koller A, Snyder WB (2000) Import of peroxisomal matrix and membrane proteins. Annu Rev Biochem 69:399 418. Takahashi T, Nowakowski RS, Caviness VS (1992) BrdU as an S-phase marker for quantitative studies of cytokinetic behaviour in the murine cerebral ventricular zone. J Neurocytol 21:185197. Terlecky SR, Fransen M (2000) How peroxisomes arise. Traffic 1:465 473. Van Veldhoven PP, Van Rompuy P, Fransen M, de Be thune B, Mannaerts GP (1994) Large-scale purification and further characterization of ratpristanoyl-CoA oxidase. Eur J Biochem 222:795 801. Walsh CA (1999) Genetic malformations of the human cerebral cortex. Neuron 23:19 29. Yamaguchi M, Saito H, Suzuki M, Mori K (2000) Visualization of neurogenesis in the central nervous system using nestin promoter-GFP transgenic mice. Dev Neurosci 11:19911996. Zimmerman L, Lendahl U, Cunningham M, McKay R, Parr B, Gavin B, Mann J, Vassileva G, McMahon A (1994) Independent regulatory elements in the nestin gene direct transgene expression to neural stem cells or muscle precursors. Neuron 12:1124.