The Neuronal Migration Defect in Mice with Zellweger Syndrome (Pex5 Knockout) is not Caused by the Inactivity of Peroxisomal -Oxidation

Transcription

1 Journal of Neuropathology and Experimental Neurology Vol. 61, No. 4 Copyright 2002 by the American Association of Neuropathologists April, 2002 pp The Neuronal Migration Defect in Mice with Zellweger Syndrome (Pex5 Knockout) is not Caused by the Inactivity of Peroxisomal -Oxidation M. BAES, PHD, P. GRESSENS, MD, PHD, S. HUYGHE, PHARM., K. DE NYS, MD, C. QI, PHD, Y. JIA, PHD, G. P. MANNAERTS, MD, PHD, P. EVRARD, MD, PHD, P. P. VAN VELDHOVEN, PHD, P. E. DECLERCQ, PHD, AND J. K. REDDY, MD Abstract. The purpose of this study was to investigate whether deficient peroxisomal -oxidation is causally involved in the neuronal migration defect observed in Pex5 knockout mice. These mice are models for Zellweger syndrome, a peroxisome biogenesis disorder. Neocortical development was evaluated in mice carrying a partial or complete defect of peroxisomal oxidation at the level of the second enzyme of the pathway, namely, the hydratase-dehydrogenase multifunctional/bifunctional enzymes MFP1/L-PBE and MFP2/D-PBE. In contrast to patients with multifunctional protein 2 deficiency who present with neocortical dysgenesis, impairment of neuronal migration was not observed in the single MFP2 or in the double MFP1/MFP2 knockout mice. At birth, the double knockout pups displayed variable growth retardation and about one half of them were severely hypotonic, whereas the single MFP2 knockout animals were all normal in the perinatal period. These results indicate that in the mouse, defective peroxisomal -oxidation does not cause neuronal migration defects by itself. This does not exclude that the inactivity of this metabolic pathway contributes to the brain pathology in mice and patients with complete absence of functional peroxisomes. Key Words: -Oxidation; Hypotonia; Multifunctional protein; Neuronal migration; Peroxisome; Very long chain fatty acid; Zellweger syndrome. INTRODUCTION Zellweger syndrome is a peroxisome biogenesis disorder that is caused by the dysfunction of the import system for peroxisomal matrix proteins (1). As a result, most peroxisomal metabolic pathways, of which the -oxidation and ether phospholipid synthesis pathways are the best-known examples, are inactive. The neurodevelopmental abnormalities in Zellweger syndrome constitute some of the most important characteristics of the disease (1). These include a partial impediment of neuronal migration in the neocortex, Purkinje cell heterotopias in cerebellum, and abnormal formation of the inferior olive. In addition, abnormalities in the white matter have been observed, which might be related to the demyelination seen in X-linked adrenoleukodystrophy (1). Zellweger patients present with severe hypotonia and convulsions in the neonatal period and usually only survive a few months. The metabolic alterations in Zellweger patients and in aborted fetuses (14 22 wk old) include accumulation of very long chain fatty acids (VLCFAs), a depletion of plasmalogens, which are From the Laboratory of Clinical Chemistry (MB, SH, PED), Faculty of Pharmaceutical Sciences, K.U. Leuven, Leuven, Belgium; Laboratoire de Neurologie de Développement (PG, PE), Hôpital Robert-Debré, Paris, France; Department of Pathology (CQ, YJ, JKR), Northwestern University Medical School, Chicago, Illinois; Department of Pharmacology (KDN, GPM, PPVV), Faculty of Medicine, K.U. Leuven, Leuven, Belgium. Correspondence to: Dr. Myriam Baes, Laboratory of Clinical Chemistry, K.U. Leuven, Herestraat 49 O/N, B 3000, Leuven, Belgium. This work was supported by grants from Fonds Wetenschappelijk Onderzoek Vlaanderen (G ) and Geconcerteerde Onderzoeksacties (GOA/99/09). the most important class of ether phospholipids (1 3), and a deficit of the polyunsaturated fatty acid, docosahexaenoic acid (4). At present, the molecular basis of the brain dysgenesis has not been elucidated. The possibility that the defect in peroxisomal -oxidation is involved in the pathogenesis is strongly supported by some cases of patients with a deficient multifunctional protein 2 (MFP2) in whom similar neuronal migration defects were observed (5, 6). MFP2, also called D-bifunctional protein, is an enzyme with enoyl-coa hydration and 3-hydroxy-acyl CoA dehydrogenation activities, which is essential for the progression of the peroxisomal -oxidation of bile acid intermediates and pristanic acid. It was recently shown that MFP2 is also engaged in the breakdown of VLCFAs such as lignoceric (C24:0) and cerotic acid (C26:0) (7). The increased level of VLCFAs, the common indicator of peroxisomal -oxidation dysfunction, has often been suggested as a pathogenic factor contributing to the central nervous system defects (3, 8). It has also been hypothesized that the accumulation of VLCFA in biomembranes changes their structure and fluidity and, consequently, could alter signal transduction pathways (9). Recently, 2 mouse models for peroxisome biogenesis disorders were generated, i.e. Pex5 (10) and Pex2 (11) knockout mice. In both mouse strains, a neuronal migration defect was observed that closely mimics the abnormalities in human Zellweger brain. The metabolic alterations in these peroxisome-deficient mice were also comparable with those reported in man (10, 11). The purpose of the present study was to investigate the contribution of peroxisomal -oxidation dysfunction to 368

2 NEURONAL MIGRATION AND PEROXISOME DYSFUNCTION 369 the neuronal migration disorder in Pex5 knockout mice by examining brain development in mice with selective inactivation of this pathway. It is now well known that peroxisomal -oxidation consists of 4 enzymatic steps, each of which can be catalyzed by at least 2 different enzymes that display particular but sometimes overlapping substrate specificities (12). Consequently, in order to generate peroxisomal -oxidation-deficient mice, the simultaneous inactivation of 2 enzymes needed to be achieved. We chose to inactivate peroxisomal -oxidation at the level of the hydratase/dehydrogenase enzymes MFP1 and MFP2 because both the MFP1 (13) and MFP2 (7) knockout mouse lines were already available. In addition, because MFP2-deficient patients are known to have neuronal migration defects, we wanted to examine whether this would be mimicked in mice carrying the same enzymatic defect. MATERIALS AND METHODS Mouse Breeding The generation of MFP1 (C57B6 background) (13) and MFP2 (Swiss/129sv background) (7) knockout mice has been described. MFP1 / mice were mated with MFP2 / mice to obtain MFP1 / MFP2 / mice. These were intercrossed to generate MFP1 / MFP2 / and MFP1 / MFP2 / mice. Each of these 2 mouse lines were further maintained to generate MFP1 / MFP2 / and MFP1 / MFP2 / mice. For embryonic analysis, the morning of detection a vaginal plug was considered as embryonic day 0.5 (E0.5). All mice were genotyped by Southern or PCR analysis on tail DNA, as previously described (7, 13). All animal experiments were approved by the Ethical Committee for Animal Experimentation of the K.U. Leuven and conform to the Declaration of Helsinki. Neuronal Migration Analysis For routine histological analysis, brains of E18.5 or newborn P0.5 pups were fixed in 4% paraformaldehyde (Sigma, Bornem, Belgium) and embedded in paraffin. Coronal sections of 7- m thickness were stained with cresyl violet. Eight to 15 brains were analyzed in each experimental group. In order to trace the position of migrating neurons, pregnant dams were intraperitoneally injected with 50 g bromodeoxyuridine (BrdU) (Sigma) at E13.5. Fetuses were killed at E18.5 by decapitation, heads were fixed in 70% ethanol, and brains were embedded in paraffin. Seven- m-thick coronal sections were used for immunohistochemical detection of BrdU. Based on previous studies performed in Pex5 knockout mice (10, 14), the density of BrdU-stained cells in the intermediate zone (prospective white matter) 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 (SI), were treated simultaneously. Counts of BrdUpositive cells were performed in a 500- m 2 area of the intermediate zone (prospective white matter); for each experimental group, cells were counted in 10 different fields (5 brains from 3 different litters, 2 non-adjacent sections of the right hemisphere per brain). Results are shown as means SEM. -Oxidation in Fibroblast Cultures Fibroblast cultures were established from the skin of E15.5 or E16.5 embryos. The -oxidation of [1-14 C]-lignoceric acid, 2-methyl-[1-14 C]-hexadecanoic acid and [1-14 C]-palmitic acid in these cultures was performed as previously described (7, 15). VLCFA Analysis in Brain Lipid Extraction and Separation of Lipid Fractions: After decapitation of the newborn pups, brain tissue was quickly removed, snap frozen in liquid N 2, and stored at 80 C. Whole brains (approximately 85 mg) were homogenized in 3 ml CH 3 OH/CHCl 3 (2:1) 0.7 ml water using a Polytron homogenizer. The total lipid fraction was isolated as described (16), dried, dissolved in 0.5 ml of CHCl 3, and separated into neutral lipids, fatty acids, and phospholipids by means of solid phase extraction (Bond-Elut NH2-column [500 mg], Varian, Zaventem, Belgium) (17), but phospholipids were eluted with 4 ml of CHCl 3 :CH 3 OH:5 M NH 4 Cl (30:60:10) instead of methanol in order to also elute acidic phospholipids under conditions where plasmalogens remained intact. After adding 1.2 ml chloroform and 1.8 ml 1 M NH 4 Cl, the phases separated by centrifugation and phospholipids were quantitatively recovered in the chloroform phase (PP Van Veldhoven, unpublished data). An aliquot of the latter fraction was used for wet washing, followed by determination of the phosphate content (16). All solvents were of the highest quality commercially available (Biosolve, Valkenswaard, The Netherlands) and were supplemented with 0.05% butylhydroxytoluene (Acros, Geel, Belgium) to prevent lipid peroxidation. Fatty Acid Extraction The fatty acids were released from the phospholipid fraction by using a modification of the method described by Vreken et al for the quantification of fatty acids in plasma (18). A 0.6-ml aliquot of the phospholipid fraction was dried and 100 l of internal standard solution was added (2.5 M 2 H 3 -pristanic acid, 2.5 M 2 H 3 -phytanic acid, 25 M 2 H 4 -C22:0, 25 M 2 H 4 -C24: 0 and 25 M 2 H 4 -C26:0 in toluene) (standards purchased from Dr. HJ ten Brink, Free University Hospital, Amsterdam, The Netherlands). The samples were hydrolyzed for 45 min at 110 C in 2.0 ml 0.5 M HCl in acetonitrile. Subsequently, 2.0 ml 1.0 M NaOH in CH 3 OH was added and the samples were further hydrolyzed at 110 C for 45 min. After cooling, the ph was lowered by adding 1.0 ml of 10 M HCl, and the fatty acids were extracted with 4.0 ml hexane. The extract was brominated (25 l Br 2, 0.1% in CHCl 3 for 20 min), dried, and derivatized with 100 l (MTBSTFA 1% TBDMCS) (Fluka-Sigma-Aldrich, Bornem, Belgium) at 80 C for 30 min. The fatty acids were analyzed on a GC-MS system (GCQ-Finnigan MAT, San Jose, CA), with a DB-5MS capillary column from J&W Scientific (Alltech, Lokeren, Belgium). The chromatography temperature program was taken from Vreken and the source was at 200 C. Concentrations of fatty acids were calculated using the internal standards and normalized to the phospholipid content.

3 370 BAES ET AL TABLE 1 Peroxisomal -Oxidation in Fibroblast Cultures Genotype C24 2-methyl C16 C16 Oxidation Oxidation/ Esterification Oxidation Oxidation/ Esterification Oxidation Oxidation/ Esterification Wild type (n 4) MFP1 KO (n 6) MFP2 KO (n 6) MFP1/MFP2 KO (n 4) Pex5 KO (n 4) Fibroblast cultures of the indicated genotypes were incubated with 14C-carboxy-labeled lignoceric acid (C24), 2-methylhexadecanoic acid (2-methyl-C16), or palmitic acid (C16). The combined production of CO2 and acid water soluble products (expressed in pmol/24 h and per mg protein) is taken as a measure for -oxidation, while incorporation of label into triglycerides and phospholipids is taken as measure of esterification (mean SEM). RESULTS Breeding and Macroscopic Analysis of Newborn MFP1 / MFP2 / and MFP1 / MFP2 / Pups Because of the reduced fertility of MFP2 knockout mice (7), heterozygous MFP2 / mice were inbred with MFP1 / mice (13). In this mixed genetic background, 2 mouse lines were maintained: MFP1 / MFP2 / and MFP1 / MFP2 / mice that were used to generate, respectively, MFP1 / MFP2 / and MFP1 / MFP2 / mice. Macroscopic and further biochemical and brain developmental analysis did not reveal any differences between MFP2 knockouts previously generated in the 129sv/Swiss (7) and those from the mixed 129sv/Swiss/ C57B6 backgrounds. As previously described (7), mating of MFP2 / mice yielded MFP2 knockouts with the expected Mendelian inheritance frequency of 25%. At birth, the MFP2 / mice were indistinguishable from their heterozygous and wild type littermates. In contrast, genotyping during the first postnatal week identified only 18% double knockouts (43 of 244 pups) in litters born from MFP1 / MFP2 / parents. However, one third of these MFP1 / MFP2 / mice died within 24 hours (h) after birth, suggesting that the missing pups might have been born but were eaten by the mother. Based on the genotype distribution at E18.5 (41 double knockouts of 185 pups [22%]), embryonic lethality was indeed unlikely. In contrast to MFP2 knockouts, the MFP1/MFP2 double knockouts displayed variable growth retardation at birth, with body weights ranging between 66% and 90% of the MFP1 / MFP2 / and MFP1 / MFP2 / pups. The double knockout pups that died in the perinatal period were severely hypotonic. These pups were less active and were unable to support themselves on their paws and to feed, as evidenced by the lack of milk in the stomach. The precise cause of their early death is not clear at this point. The stronger double knockout pups fed sufficiently, but weight gain was minimal and almost all of these mice died within 1 or 2 wk. The latter pathology, which might be related to impaired bile acid formation and steatorrhea, requires further investigation. Only a few MFP1 / MFP2 / mice were still alive at the time of weaning. Peroxisomal -Oxidation In order to document the impairment of peroxisomal -oxidation in mice with the different genotypes, embryonic skin fibroblast cultures were incubated with radiolabeled substrates for peroxisomal -oxidation: the very long chain [1-14 C]-lignoceric acid and the branched chain 2-methyl-[1-14 C]-hexadecanoic acid. As shown in Table 1, the degradation of lignoceric acid was reduced in fibroblasts derived from MFP2 and from MFP1/MFP2 knockout mice. Also in the MFP1 / fibroblasts, a decreased oxidation was noticed, but this compound was also esterified at a lower rate in these cells, indicating a lower degree of uptake and/or activation of this fatty acid. This resulted in an unaltered oxidation/esterification ratio of lignoceric acid in MFP1 knockout fibroblasts as compared to controls. These data confirm our previous observations that MFP2 plays a more important role in the degradation of VLCFAs in mice than MFP1 (7). In order to further evaluate the -oxidation defect in the various fibroblast cultures, 2-methylhexadecanoic acid, known to be mainly -oxidized in peroxisomes (15, 19), was used as a substrate. The -oxidation of this branched chain fatty acid was unaltered in MFP1-deficient fibroblasts as compared to wild type cells, but negligible rates were found in MFP2, MFP1/MFP2, and in Pex5-deficient cultures. On the contrary, the oxidation of [1-14 C]-palmitic acid, which primarily occurs via the mitochondrial -oxidation pathway, was similar in the different cell lines. Together, these data indicate that in MFP1/MFP2 knockout mice, peroxisomal -oxidation is completely blocked and that mitochondrial -oxidation is unaltered. Nevertheless, the residual capacity (20%) to -oxidize the C24:0 substrate in the double KO fibroblasts and in Pex5

4 NEURONAL MIGRATION AND PEROXISOME DYSFUNCTION 371 TABLE 2 Accumulation of C26:0 in Brain Phospholipids Genotype C26:0 Wild type (n 6) MFP1 KO (n 4) MFP2 KO (n 3) MFP1/MFP2 KO (n 6) Pex5 KO (n 3) Values are expressed as nmol C26:0/100 nmol phospholipids. KO mice is quite substantial. Hence, it seems likely that the breakdown of C24:0 results from extraperoxisomal catabolism (e.g. from mitochondrial -oxidation). Accumulation of Very Long Chain Fatty Acids In view of the brain phenotype that we intended to examine and because VLCFA are most abundantly present in the central nervous system, C26:0 levels were analyzed in brains of newborn pups. In agreement with the outcome of the oxidation experiments in fibroblasts, no accumulation of C26:0 was found in MFP1-deficient brains, whereas a 3- to 4-fold accumulation of C26:0 was found in brain phospholipids in MFP2 and MFP1/MFP2 double knockout mice (Table 2). These increases are similar to the 4-fold increase of C26:0 in brain phospholipids in Pex5 knockout mice (Table 2 and previous results [20]). As expected, no accumulation of C26:0 was found in MFP1-deficient brains. In order to exclude that other peroxisomal metabolic pathways were affected in the oxidation-deficient mice, we determined plasmalogen levels in brain of mice with the different genotypes. No differences in plasmalogen levels were found in all the mice tested, indicating that ether phospholipid synthesis remained intact (data not shown). Cortical Neuronal Migration Neuronal migration was evaluated at E18.5 and at P0.5 by examining the layering of the neocortex in cresyl violet-stained coronal sections. More detailed analysis was done in E18.5 pups by quantifying in the intermediate zone (prospective white matter) neurons labeled with BrdU at E13.5. Neither technique revealed any significant abnormality in, or difference between, wild type, MFP2- deficient, and MFP1/MFP2-deficient mice (Fig. 1A, C E). Several growth-retarded as well as less affected MFP1/MFP2-deficient pups were examined. In contrast, Pex5 knockout mice displayed an increased density of cells in the intermediate zone, confirming the migration defect in these animals (Fig. 1B E). DISCUSSION Since the description of the peroxisome biogenesis disorders, it is well known that intact peroxisomal function is required for the normal formation of the brain (1). In the absence of functional peroxisomes, a characteristic defect of neuronal migration manifests in man (1), as well as in transgenic mice (10, 11). The pathogenesis of this defect is still not understood, but based on correlations between biochemical and morphological changes in patients with peroxisomal -oxidation defects on one hand, and peroxisome biogenesis disorders on the other hand, a pathogenic role of elevated VLCFA levels has been proposed. This assumption is also supported by the findings that VLCFAs already accumulate during neuronal migration in Zellweger patients (3) and in peroxisomedeficient mice (21). However, in the present study we demonstrated that defective peroxisomal -oxidation by itself has no adverse effect on the neuronal migration process, at least in the mouse. In brain of MFP2 knockout mice carrying a partial inactivation of peroxisomal -oxidation, the level of C26:0 was elevated to the same extent as in peroxisomedeficient Pex5 knockout mice. However, no defect in neuronal migration could be found in the MFP2 knockout mice, making it unlikely that increased levels of C26:0 on their own induce brain abnormalities. Two other transgenic mouse lines containing elevated C26:0 levels in plasma and/or tissues were previously generated. In accordance with the present findings, no brain pathology was observed in X-ALD knockout mice (22 24) and no obvious signs of neurological defects were observed in the acyl-coa oxidase knockout mice (25, 26) that survived at least 15 months (JK Reddy, unpublished observations). In order to examine whether other substrates that depend on peroxisomal -oxidation might be involved in the pathogenesis, a complete block of peroxisomal -oxidation was established in mice by inactivating 2 enzymes that act in parallel in the -oxidation pathway, i.e. MFP1 and MFP2. Since no abnormalities in brain structure were observed at the time of birth in those mice, it seems unlikely that the inactivity of peroxisomal -oxidation is directly involved in the neuronal migration impairment in the mouse model for Zellweger syndrome. The absence of neuronal migration anomalies in MFP2 knockout mice was unexpected in light of the reports from patients with mutations in MFP2. Indeed, MFP2 deficiency is the only other single enzyme peroxisomal disorder that is accompanied by neuronal migration defects in man (5). In 15 of 17 patients with MFP2 deficiency in whom a more or less detailed brain analysis was performed, a disturbed neuronal migration was described. In 1 MFP2-deficient patient a more extensive study was performed, revealing that the pattern of brain malformation resembled that in Zellweger syndrome (6) with dysgenesis of the centrosylvian cortex, the dentate and inferior olive and heterotopias in the neocortical white matter, and of Purkinje cells in the cerebellum.

5 372 BAES ET AL Fig. 1. A D: Cresyl violet-stained coronal sections of P0.5 brains showing wild type (A), Pex5 / (B), MFP2 / (C), or MFP1 / /MFP2 / (D) pups. E: Counts of BrdU-labeled cells at E18.5 in the intermediate zone after injection of BrdU into pregnant animals at E13.5. ***, statistically different from wild type embryos (ANOVA with Bonferroni s post-test, p 0.001).

6 NEURONAL MIGRATION AND PEROXISOME DYSFUNCTION 373 However, the overall severity of the malformations appeared to be less in the MFP2-deficient patient than in Zellweger patients. This suggests that in man, the -oxidation defect does account for some but not all the brain anomalies seen in Zellweger syndrome. It should be added that in mice, neuronal migration defects and minor heterotopias are more difficult to determine than in man, because of the small white matter area through which the neurons have to migrate. A more extensive description of the brain malformations in additional MFP2-deficient patients seems required in order to interpret the differences between mice and man. The present results seem to indicate that the inactive peroxisomal -oxidation is less harmful in the mouse than in man. It is possible that substrates that depend on MFP2 for their degradation in man can be disposed of by non-peroxisomal pathways in the mouse. Alternatively, for some unknown reason, these substances might be less toxic to the mouse brain. Besides its catabolic function, peroxisomal -oxidation could also be responsible for the generation of a compound that is essential for normal neuronal migration. One important brain compound that is known to be generated through peroxisomal -oxidation is the polyunsaturated fatty acid, docosahexaenoic acid (27 29). However, we previously demonstrated that the neuronal migration was not improved when Pex5-deficient fetuses were supplemented with docosahexaenoic acid, making this molecule unlikely to be involved in the pathogenesis (20). It may be argued that the lack of a migration phenotype in the peroxisomal -oxidation-deficient mice undermines the usefulness of mouse models to study the neuronal migration defect of human peroxisomal diseases. However, it remains crucial that a clearcut migration defect causing malformation of precisely the same areas as those affected in Zellweger patients (neocortex, inferior olive, and cerebellum) was seen in Pex5 and Pex2 knockout mice (10, 11, 30). The unaffected neuronal migration in the MFP2 and MFP1/MFP2 knockout mice is somewhat reminiscent of the absence of brain defects in the X-ALD knockout mouse models. The latter mice also carry a partial defect of peroxisomal -oxidation leading to increased VLCFA levels, but the neurodegeneration seen in X-ALD patients was not observed (22 24). Whereas no signs of neuronal migration defects were found in any of the -oxidation-deficient mice, a partially penetrant hypotonia was noticed in the double knockout mice. Neonatal hypotonia is a constant feature of the disorders of the Zellweger spectrum (Zellweger syndrome, neonatal adrenoleukodystrophy, and Infantile Refsum disease). Also in other peroxisomal diseases (i.e. in nearly all MFP2-deficient patients and in several patients with rhizomelic chondrodysplasia punctata) carrying a defect in both ether phospholipid synthesis and -oxidation, hypotonia is characteristic (5). As previously reported, all MFP1 / MFP2 / mice displayed a normal tonus in the neonatal period (7), whereas MFP1 / mice did not show a pathological phenotype throughout life (13). In contrast, approximately one half of the MFP1 / MFP2 / mice were severely hypotonic and died within 24 h, suggesting that the inactivity of peroxisomal -oxidation contributes to the severe hypotonia in peroxisome-deficient mice. No straightforward candidate substrate of the peroxisomal oxidation pathway can be proposed to underlie this defect, however, it appears that it can be redundantly degraded by MFP1 or MFP2. The partial penetration of this hypotonic phenotype as well as the variable degree of growth retardation observed in the double knockout mice might be related to the mixed genetic background of this line. Future studies with other transgenic mice should clarify whether defective ether phospholipid synthesis and -oxidation also cause hypotonia. In any case, the hypotonia does not seem to be directly related to the neuronal migration defect, and thus the pathogenic mechanisms must be different. In summary, a block in peroxisomal -oxidation at the level of the MFP enzymes in mice does not induce a neuronal migration defect, but it causes hypotonia and intrauterine growth retardation to a variable extent. ACKNOWLEDGMENTS The guidance of Dr. K. Desmet in using the GC-MS and the excellent technical assistance of Benno Das, Lies Pauwels and Leslie Schwendimann are gratefully acknowledged. REFERENCES 1. Gould SJ, Raymond GV, Valle D. The peroxisome biogenesis disorders. In: Scriver BVS, Beaudet, Valle, Sly, eds. The metabolic and molecular bases of inherited disease. New York: McGraw-Hill, 2001: Moser AB, Kreiter N, Bezman L, Lu S, Raymond GV, Naidu S, Moser HW. Plasma very long chain fatty acids in 3,000 peroxisome disease patients and 29,000 controls. Ann Neurol 1999;45: Powers JM, Tummons RC, Caviness VS, Moser AB, Moser HW. Structural and chemical alterations in the cerebral maldevelopment of fetal cerebro-hepato-renal (Zellweger) syndrome. J Neuropathol Exp Neurol 1989;48: Martinez M. Polyunsaturated fatty acids in the developing human brain, erythrocytes and plasma in peroxisomal disease: Therapeutic implications. J Inherit Metab Dis 1995;18: Wanders RJA, Barth PG, Heymans HHA. Single peroxisomal enzyme deficiencies. In: Scriver, Beaudet, Valle, Sly, eds. The metabolic and molecular bases of inherited disease. New York: Mc- Graw-Hill, 2001: Kaufmann WE, Theda C, Naidu S, Watkins PA, Moser AB, Moser HW. Neuronal migration abnormality in peroxisomal bifunctional enzyme defect. Ann Neurol 1996;39: Baes M, Huyghe S, Carmeliet P, et al. Inactivation of the peroxisomal multifunctional protein-2 in mice impedes the degradation of not only 2-methyl branched fatty acids and bile acid intermediates but also of very long chain fatty acids. J Biol Chem 2000;275: Powers JM. The pathology of peroxisomal disorders with pathogenetic considerations. J Neuropathol Exp Neurol 1995;54:710 19

7 374 BAES ET AL 9. Ho JK, Moser H, Kishimoto Y, Hamilton JA. Interactions of a very long chain fatty acid with model membranes and serum albumin. J Clin Invest 1995;96: Baes M, Gressens P, Baumgart E, et al. A mouse model for Zellweger syndrome. Nat Genet 1997;17: Faust PL, Hatten ME. Targeted deletion of the PEX2 peroxisome assembly gene in mice provides a model for Zellweger syndrome, a human neuronal migration disorder. J Cell Biol 1997;139: Mannaerts GP, Van Veldhoven PP, Casteels M. Peroxisomal lipid degradation via - and -oxidation in mammals. Cell Biochem Biophys 2000;32: Qi C, Zhu Y, Pan J, Usuda N, et al. Absence of spontaneous peroxisome proliferation in enoyl-coa hydratase/l-3-hydroxyacyl- CoA dehydrogenase-deficient mouse liver. J Biol Chem 1999;274: Gressens P, Baes M, Leroux P, et al. Neuronal migration disorder in Zellweger syndrome is secondary to glutamate receptor dysfunction. Anal Neurol 2000;48: Van Veldhoven PP, Huang S, Eyssen HJ, Mannaerts GP. The deficient degradation of synthetic 2- and 3-methyl-branched fatty acids in fibroblasts from patients with peroxisomal disorders. J Inherit Metab Dis 1993;16: Van Veldhoven P, Bell RM. Effect of harvesting methods, growth conditions and growth phase on diacylglycerol levels in cultured human adherent cells. Biochim Biophys Acta 1988;959: Kaluzny MA, Duncan LA, Merritt MV, Epps DE. Rapid separation of lipid classes in high yield and purity using bonded phase columns. J Lipid Res 1985;26: Vreken P, van Lint AEM, Bootsma AH, Overmars H, Wanders RJA, van Gennip AH. Rapid stable isotope dilution analysis of verylong-chain fatty acids, pristanic acid and phytanic acid using gas chromatography-electron impact mass spectrometry. J Chromatogr 1998;713: Vanhove G, Van Veldhoven PP, Vanhoutte F, Parmentier G, Eyssen HJ, Mannaerts GP. Mitochondrial and peroxisomal -oxidation of the branched chain fatty acid 2-methylpalmitate in rat liver. J Biol Chem 1991;266: Janssen A, Baes M, Gressens P, Mannaerts GP, Declercq P, Van Veldhoven PP. Docosahexaenoic acid deficit is not a major pathogenic factor in peroxisome-deficient mice. Lab Invest 2000;80: Huyghe S, Casteels M, Janssen A, et al. Prenatal and postnatal development of peroxisomal lipid-metabolizing pathways in the mouse. Biochem J 2001;353: Forss-Petter S, Werner H, Berger J, et al. Targeted inactivation of the X-linked adrenoleukodystrophy gene in mice. J Neurosci Res 1997;50: Lu J-F, Lawler AM, Watkins PA, et al. A mouse model for X-linked adrenoleukodystrophy. Proc Nat Acad Sci USA 1997;94: Kobayashi T, Shinnoh N, Kondo A, Yamada T. Adrenoleukodystrophy protein-deficient mice represent abnormality of very long chain fatty acid metabolism. Biochem Biophys Res Comm 1997;232: Fan C-Y, Pan J, Chu R, et al. Hepatocellular and hepatic peroxisomal alterations in mice with a disrupted peroxisomal fatty acylcoenzyme A oxidase gene. J Biol Chem 1996;271: Fan C-Y, Pan J, Usuda N, Yeldandi AV, Rao MS, Reddy JK. Steatohepatitis, spontaneous peroxisome proliferation and liver tumors in mice lacking peroxisomal fatty acyl-coa oxidase. Implications for peroxisome proliferator-activated receptor alpha natural ligand metabolism. J Biol Chem 1998;273: Voss A, Reinhart M, Sankarappa S, Sprecher H. The metabolism of 7,10,13,16,19-docosapentaenoic acid to 4,7,10,13,16,19-docosahexaenoic acid in rat liver is independent of a 4-desaturase. J Biol Chem 1991;266: Martinez M. Abnormal profiles of polyunsaturated fatty acids in the brain, liver, kidney and retina of patients with peroxisomal disorders. Brain Res 1992;583: Su HM, Moser AB, Moser HW, Watkins PA. Peroxisomal straightchain acyl-coa oxidasee and d-bifunctional protein are essential for the retroconversion step in docosahexaenoic acid synthesis. J Biol Chem 2001;276: Faust PL, Su H-M, Moser A, Moser HW. The peroxisome deficient Pex2 Zellweger mouse. Pathologic and biochemical correlates of lipid dysfunction. J Mol Neurosci 2001;16: Received November 1, 2001 Revision received December 19, 2001 and January 22, 2002 Accepted January 23, 2002