review generalised and conditional inactivation of pex ... · review generalised and conditional...

Biochimica et Biophysica Acta 1763 (2006) 1785–1793www.elsevier.com/locate/bbamcr

Review

Generalised and conditional inactivation of Pex genes in mice

Myriam Baes a,⁎, Paul P. Van Veldhoven b

a Laboratory for Cell Metabolism, Campus Gasthuisberg Onderwijs en Navorsing II, bus 823 Herestraat 49 B-3000,Department of Pharmaceutical Sciences, Katholieke Universiteit Leuven, Leuven, Belgium

b Department of Molecular Cell Biology, Division Pharmacology, Katholieke Universiteit Leuven, Leuven, Belgium

Received 4 May 2006; received in revised form 17 August 2006; accepted 18 August 2006Available online 25 August 2006

Abstract

During the past 10 years, several Pex genes have been knocked out in the mouse with the purpose to generate models to study the pathogenesisof peroxisome biogenesis disorders and/or to investigate the physiological importance of the Pex proteins. More recently, mice with selectiveinactivation of a Pex gene in particular cell types were created. The metabolic abnormalities in peroxisome deficient mice paralleled to a largeextent those of Zellweger patients. Several but not all of the clinical and histological features reported in patients also occurred in peroxisomedeficient mice as for example hypotonia, cortical and cerebellar malformations, endochondral ossification defects, hepatomegaly, liver fibrosis andultrastructural abnormalities of mitochondria in hepatocytes. Although the molecular origins of the observed pathologies have not yet beenresolved, several new insights on the importance of peroxisomes in different tissues have emerged.© 2006 Elsevier B.V. All rights reserved.

Keywords: Pex gene; Peroxisome; Knockout; Mouse model; Zellweger syndrome; Conditional gene inactivation

1. Introduction

The metabolic role of peroxisomes has been extensivelystudied during the last 30 years mostly by using the rat liver as asource of peroxisomes [1]. The identification of human diseasesthat are due to peroxisomal dysfunction underscored thephysiological importance of peroxisomal metabolism in differ-ent tissues. However, the molecular mechanisms leading toanomalies such as neuronal migration defects, hypotonia, dys-and demyelination, kidney, eye and testicular defects, abnor-malities in bone development and facial dysmorphism remainunresolved. Since naturally occurring peroxisomal disease mo-dels are lacking, it was necessary to generate such models bygene manipulation to study the pathogenesis of these diseases.The newly discovered Pex genes and the novel insights in the

Abbreviations: Acox1, acyl-CoA oxidase 1; ALD, adrenoleukodystrophy;DHAPAT, dihydroxyacetonephosphate acyltransferase; DHAP, dihydroxyace-tonephosphate; DHA, docosahexaenoic acid; MFP, multifunctional protein;NMDA, N-methyl-D-aspartate; PTS, peroxisome targeting signal; PUFA,polyunsaturated fatty acid; SCP, sterol carrier protein; VLCFA, very longchain fatty acid⁎ Corresponding author. Tel.: +32 16 34 72 83; fax: +32 16 34 72 91.E-mail address: [email protected] (M. Baes).

0167-4889/$ - see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.bbamcr.2006.08.018

mechanism of peroxisomal matrix protein import offered aunique opportunity to apply gene targeting techniques. De-pending on the mutated gene, mice were created in whicheither PTS1- and PTS2-dependent protein import (Pex5, Pex2,Pex13), or only PTS2-dependent import (Pex7), or peroxisomeproliferation (Pex11) were affected. More recently, mice withconditional inactivation of peroxisome biogenesis in specificcell types became available i.e. mice with selective inactivationin hepatocytes and in Sertoli cells.

2. Generalised inactivation of Pex genes involved inPTS1- and PTS2-dependent protein import: models forZellweger syndrome?

2.1. Phenotype of the mice

Three different Pex genes were inactivated in the mouse i.e.Pex5, Pex2 and Pex13 [2–4]. In view of the complete block ofPTS1- and PTS2-dependent matrix enzyme import in each case,the three knockout lines were expected to constitute a model forZellweger syndrome, the most severe peroxisome biogenesisdisorder [5]. All knockout mice displayed severe hypotonia andgrowth retardation at birth and they died within a few days. A

mailto:[email protected]

http://dx.doi.org/10.1016/j.bbamcr.2006.08.018

1786 M. Baes, P.P. Van Veldhoven / Biochimica et Biophysica Acta 1763 (2006) 1785–1793

notable exception were the Pex2 knockout mice bred in aspecific genetic background i.e. Swiss Webster×129Sv Ev afraction of which was less hypotonic and survived 1, 2 or rarely 3weeks [6]. As discussed below, all knockout strains displayedabnormalities in the formation of the brain. However, othertypical hallmarks present in fetal Zellweger patients were notfound, including renal cysts and facial dysmorphism. At birth themice did not display major liver pathology, a feature that onlydevelops in the postnatal period in patients.

At the subcellular level, embryonic fibroblast cultures de-rived from Pex5, Pex2 and Pex13 knockout mice containedperoxisomal remnants which were reduced in number butincreased in size as compared to regular peroxisomes [2–4]. Thisis very similar to findings in Zellweger patient fibroblasts.Recently, it was demonstrated that in cultured Pex13 null mousefibroblasts (like in Pex1 deficient human fibroblasts), remnantperoxisomes displayed an altered distribution as compared tonormal peroxisomes in wild type fibroblasts (Fig. 1A, B) [7].The peroxisomal ghosts were only present in the cell centre butnot in the periphery [7]. However, the peroxisomal ghosts werestill associated with microtubuli, as previously shown for normalperoxisomes in different cell lines like HepG2 and Cos cells [8].Likewise, in cultured Pex13 null neurons and astrocytesperoxisomes were clustered in the cell soma.

Ultrastructural examination of the liver of the three Pexknockout models revealed severe abnormalities of the mito-chondria in hepatocytes, in particular at the level of the innermitochondrial membrane [2,4,9,10] (see also 5.1). Suchmitochondrial anomalies in hepatocytes were also described insome early papers on Zellweger patients [11–13].

Overall, it can be concluded that several but not all features ofhuman Zellweger patients are mimicked in the mouse models.This might be related to the shorter timeframe of murine ascompared to human fetal development. It is in fact quite remark-able that within the short period in which the mouse brain isformed, the absence of functional peroxisomes has such animportant impact, as will be further discussed below.

2.2. Metabolic changes

Most of the major metabolic changes observed in Zellwegerpatients were recapitulated in the mouse models. An importantadvantage of the latter models is that metabolite levels can bemeasured not only in body fluids but also in the diseased tissuesat several stages. It was striking that C26:0 was already 2-foldincreased at E16.5 in Pex5 knockout mouse brain and liverindicating that peroxisomal β-oxidation was already active atthis stage in wildtype mice [14]. In newborn Pex knockout mice,this very long chain fatty acid (VLCFA) was about 10-fold

Fig. 1. Pathology in generalised Pex knockout mice. (A–B) Immunofluorescentperoxisomal membranes in cultured fibroblasts. In wild type fibroblasts (A) peroxisfibroblasts (B) peroxisomes are clustered in the cell centre but still aligned with microwild type (C) and Pex5−/− mice (D) reveals increased cell densities in the intermediGZ=germinative zone. (E–F) Immunofluorescent staining of Purkinje cells in the cerD28K antibody. In the knockout, these cells remain polydendritic and the size of the dmount skeletal staining of cartilage (blue) and bone (red) of newborn wild type (Ghindpaws of the knockout is visible (arrows). Reprinted from [7,2,6,29] with permi

increased in plasma (Pex2) [3]) in liver and brain tissue lipids(Pex13) [4] and in liver and brain phospholipids (Pex5) [2].More detailed investigations documented that C26:0 was also10-fold increased in brain sphingolipids (ceramide and sphin-gomyelin) in Pex5 knockout mice [15]. The concentration of thepolyunsaturated fatty acid, docosahexaenoic acid (DHA) whichdepends on peroxisomal β-oxidation for its synthesis wasreduced by 30% in newborn Pex5 knockout mice [16]. Thebranched chain fatty acids, phytanic and pristanic acid, and bileacids have thus far not been measured in newborn Pex knockoutmice. Plasmalogens, a subclass of ether phospholipids, were20- to 100-fold reduced in liver and brain (Pex5, Pex13) anderythrocytes (Pex2) of newborn mice due to the inactivity ofthe peroxisomal enzymes DHAPAT and alkyl DHAP synthase.

Newborn Pex5 knockout mice contained normal levels ofcholesterol in plasma, liver and brain [17] and the activity ofcholesterol synthesizing enzymes was either normal or slightlyincreased [18]. This is not consistent with the 40% reduction ofplasma cholesterol described in newborn Pex2 knockout mice[10]. In 10-day-old Pex2 knockout mice, cholesterol levels ofplasma and liver were also 40% reduced but they were normal inbrain and other tissues. The reduced cholesterol levels wereaccompanied by a concerted increase in expression and activityof cholesterol synthesizing enzymes in liver homogenatesprobably mediated by the induction of SREBP2. Because it iswell known that the brain is self sufficient with regard tocholesterol synthesis, the normal cholesterol content in brain ofPex5 and Pex2 knockout mice strongly suggests that the absenceof import competent peroxisomes has no direct negative effectson the cholesterogenesis. The aberrant cholesterol homeostasisin liver of Pex2 knockout mice might be a consequence of themetabolic perturbations that occur in the liver as exemplified bythe development of steatosis and cholestasis [10].

2.3. Cortical neuronal migration

The disturbed lamination of the cortical plate associated withmedial pachygyria and lateral polymicrogyria is a major charac-teristic of Zellweger patients [19–21]. It is ascribed to a uniquedefect of the neuronal migration process that is clearlydistinguishable from other migration disorders such as thoseseen in lissencephaly or double cortex syndrome. In all three Pexknockout models an altered distribution of cortical neurons wasobserved [2–4] (Fig. 1C, D). For Pex5 and Pex2 knockout mice,a migration disorder was demonstrated by performing 5′,3′-bromo-2′-deoxyuridine (BrdU) birthdating experiments. Fur-thermore, a delay in the differentiation of neurons, increasedapoptotic cell death but a normal distribution of radial glial cellswere reported for Pex5 knockout pups.

staining of α-tubulin (red) and Pex14p (green) to visualise microtubuli andomes are associated with microtubuli throughout the cell whereas in Pex13−/−

tubuli. (C–D) Cresyl violet staining of frontal sections of the cortex of newbornate zone (IZ) of the knockout, the prospective white matter. CP=cortical plate,ebellum of 7-day-old wild type (E) and Pex2−/− mice (F) using the anti-calbindinendritic arbor and the degree of branching are markedly reduced. (G–H) Whole) and Pex7−/− pups (H). A lack of ossification in the middle phalanges of thession from the respective publishers.

1787M. Baes, P.P. Van Veldhoven / Biochimica et Biophysica Acta 1763 (2006) 1785–1793


It was further investigated whether reduced levels of DHA orincreased levels of VLCFA could play a causative role in thesedefects. Supplementation of pregnant dams with DHA normal-ised the levels of this PUFA in brain but did neither improvehypotonia, nor the migration disorder [16]. In mice with aselective defect of peroxisomal β-oxidation by inactivatingmultifunctional protein 2 (MFP-2), levels of C26:0 were in-creased to the same extent as in Pex5 knockout mice but nomigration defect could be observed [22]. This is at variance withthe human situation where MFP-2 deficiency is often associatedwith brain malformations which are very similar to those presentin Zellweger syndrome [23]. These observations rule out thatalterations in fatty acid levels on their own induce cytoarchitec-tonic abnormalities in mouse brain.

The possible involvement of glutamatergic neurotransmis-sion via the NMDA receptor, known to control the speed ofmigration was also examined [24]. Treatment of Pex5−/−

embryos with NMDA antagonists induced embryonic deathwhereas NMDA agonists partially reversed the migration defect[25]. A deficit in NMDA signal transduction was demonstratedin neuronal cultures derived from Pex5 knockout mice bymonitoring calcium influx in response to NMDA. Pex5−/− cellswere less sensitive to NMDA than wild type cells but sensitivitycould be restored by preincubation with the ether phospholipidplatelet activating factor (PAF) [25]. Attempts to prove that theconcentration of PAF is reduced in brain of Pex5−/− mice did notsucceed probably due to the very low levels and instability ofPAF but the content of the degradation product lysoPAF was3-fold reduced (H. Van Overloop, M. Baes and P. VanVeldhoven, unpublished observations). Although PAF is a wellknown regulator of neuronal migration [24] and thus a very goodcandidate to link peroxisome deficiency to migration, theabsence of clearcut cortical migration defects in RhizomelicChondrodysplasia Punctata (RCDP) patients, who have aselective and severe deficiency of ether lipids, contradicts sucha primary role of PAF. To further document a potential involve-ment of PAF in the neuronal migration defect associated withperoxisome deficiency, it will be of importance to investigatecortical lamination in DHAPAT knockout mice [26], a modelwith a selective depletion of ether lipids.

A major question is whether the brain malformations ofperoxisome deficient mice are caused by the local absence ofperoxisomes in the brain or by extraneural deletion of peroxi-somal metabolic activity. Pex5−/− mice with a selective recon-stitution of peroxisomal function in brain or in liver bothexhibited a significant correction of the neuronal migrationdefect despite an incomplete reconstitution of peroxisomalfunction in the targeted tissue [27]. These data suggest thatperoxisomal metabolism in brain but also in extraneural tissuesis necessary for the normal development of the mouse neocortex.Interestingly, despite the improvements of the neuronalmigration, both Pex5 rescue models were as hypotonic as thegeneralised Pex5 knockout mice and died on the day of birth.

Overall, these investigations with the Pex knockout mousemodels were important to prove that peroxisome deficiency leadsto cortical malformations in another species besides humans. Anumber of metabolites were excluded as single causative factors

but there is still no clear view on the molecular mechanismunderlying the migration defect.

2.4. Formation of cerebellum

In contrast to the cortex, the formation of the cerebellumextends into the postnatal period of the mouse. Therefore, theearly death of Pex5 and Pex13 knockout mice precluded theanalysis of cerebellar development but the longer survival ofsome Pex2 knockout mice allowed an extensive analysis ofcerebellar histogenesis [6]. Abnormalities were reported at thelevel of granule cells with slower migration from the externalgranular layer (EGL) to the internal granular layer (IGL) andincreased cell death in the EGL. These migratory abnormalitiesin the postnatally developing cerebellum were more severe thanthose seen in the prenatally developing cortex [6]. This wasattributed to postnatal malnutrition problems, increasing hepaticor renal dysfunction after birth and/or to the fact that maternalmetabolism could clear toxic substances from the embryoniccirculation [6]. The majority of Purkinje cells aligned under thecerebellar surface and only rare cells mislocalised in the IGLwhich is very different from the extensive Purkinje cell hetero-topias in human Zellweger patients [19,20]. However, thePurkinje cells displayed stunted dendritic trees with abnormalarborization (Fig. 1E, F). A delayed maturation of the olivaryclimbing fibers and their defective translocation from theperisomatic to the dendritic compartment of Purkinje cells re-sulted in numerous spines on the soma and dendrites of Purkinjecells. Preliminary studies by Faust et al. with mixed cerebellarcultures further suggested that the abnormal maturation ofPurkinje cells might be due to the inactivity of intrinsic pero-xisomal metabolism as well as to the elimination of hepaticperoxisomal metabolism [28]. Indeed, also in cultured Pex2knockout Purkinje cells a deficient branching was observedwhereas on the other hand, supplementation of pups with maturebile acids not only improved survival of the mice but alsoarborization of Purkinje cells.

3. Inactivation of Pex7 involved in PTS-2 dependentprotein import: a model for RCDP type 1?

In mammals, the task of Pex7p is restricted to the importof a few PTS2 containing proteins. Thus far only alkylDHAPsynthase, phytanoyl CoA hydroxylase and 3-oxoacyl-CoA thiolase were unequivocally shown to possess such asignal but they take part in three major peroxisomal pathwaysi.e. ether phospholipid synthesis, α-oxidation and β-oxidation,respectively.

In Pex7 knockout mice [29] the inactivity of mislocalisedalkyl DHAP synthase causes the expected depletion of plasma-logens in brain and erythrocytes. Due to the mislocalisation ofphytanoyl-CoA hydroxylase a slight accumulation of phytanicacid was seen in newborn mice. Supplementation of adult Pex7knockout mice with phytol triggered a massive increase inplasma and tissue levels of this branched chain fatty acid. Basedon data of RCDP type 1 patients in which straight chain pero-xisomal β-oxidation substrates are not accumulating [5,30],


probably because of the alternative thiolytic activity exerted bySCPx, it was surprising to find that C26:0 levels were elevated inplasma and tissues of newborn Pex7 knockout mice.

Given this impaired functioning of all major peroxisomalpathways, it is not surprising that the phenotype of Pex7 knock-out mice is very similar to that of Pex5 knockout mice althoughthe penetrance is very variable [29]. About 50% of the Pex7knockout mice were very hypotonic at birth and died postnatally,20% died before weaning and the others survived past 18months. All knockout mice displayed a neuronal migrationdefect that was less severe than in Pex5 knockout mice. At birthseveral cartilage based structures that depend on endochondralossification for their formation were not ossified or showed amarked delay in ossification e.g. distal bone elements of thelimbs, parts of the skull and the vertebrae (Fig. 1G, H) [29].

Since abnormalities in bone formation are a major hallmarkof RCDP patients, this mouse model should allow for the furtherinvestigation of the precise role of peroxisomes in the ossifi-cation process. Other features observed in Pex7 knockout micethat need further study are cataract and male infertility [29].

4. Inactivation of Pex11α and β

In contrast to the other knocked out peroxins, Pex11 proteins,encoded by three different genes in mammals, do not have afunction in the import of peroxisomal matrix proteins [31]. Theyare peroxisomal membrane proteins that seem to be involved inthe division of peroxisomes since overexpression of Pex11βleads to peroxisome proliferation whereas inactivation causesreduced peroxisome abundance [32,33]. Pex11αp seems to beresponsible for peroxisome proliferation in response to externalstimuli whereas Pex11βp is required for constitutive peroxisomebiogenesis [31]. A recently reported function of Pex11βp is torecruit the dynamin like protein DLP1 to the peroxisomalmembrane [34].

Pex11α knockout mice did not have an obvious phenotypesuggesting that its function can be taken over by other Pex11proteins [35]. In tissues of Pex11β knockout mice only minormetabolic alterations were found i.e. a 20% depletion of plasma-logen levels in brain and a 1.4-fold elevation of C26:0 in liver[36]. Moreover, studies in fibroblasts documented normalimport of PTS1 and PTS2 containing proteins. It was thereforeunexpected that the phenotype of Pex11β deficient mice wasvery similar to that of Pex5, Pex2 and Pex13 knockout mice [36]i.e. these mice were all very hypotonic and growth retarded atbirth, they displayed a neuronal migration defect and died in thepostnatal period. However, no alterations of mitochondrialstructure in hepatocytes were noted. Although it could beenvisaged that the migration defect and hypotonia are caused bymetabolic abnormalities different from peroxisomal β-oxidationor ether phospholipid synthesis defects, it seems inexplicablethat such a severe phenotype is present despite intact import ofPTS1 and PTS2 proteins. In order to reconcile these discre-pancies an alternative model of pathogenesis was proposed inwhich the disease phenotype correlates with the abundance anddistribution of peroxisomes rather than with their metabolicactivity [7]. Indeed, Pex11β deficient fibroblasts and hepato-

cytes have a reduced peroxisome abundance and increasedclustering and elongation of peroxisomes in common withPex13 knockout fibroblasts and neurons [7]. For Pex13knockout fibroblasts and neurons it was shown that the remnantperoxisomes were associated only with centrally located micro-tubuli, and they were apparently not able to travel to the cellperiphery (see also 2.1). It is not unlikely that the abnormaldistribution of peroxisomes on these microtubuli hinders thedynamic activity of the cytoskeleton which is necessary fornormal migration to occur [37,38].

5. Conditional Pex5 KO mice

Because of their early postnatal death, the generalised Pex2,Pex5 and Pex13 knockout mice do not allow to study thefunctional importance of peroxisomes in adult tissues. Fortu-nately, thanks to a second wave of technological developments agene can now be conditionally inactivated in the mouse. Thetechnique is based on the use of two mouse lines, one in whichtwo LoxP sites are introduced by homologous recombination inintrons flanking an essential part of the gene of interest andanother line expressing Cre recombinase. In the presence of Crerecombinase, the gene fragment encompassed by the two loxPsites will be deleted, given that these sites have the sameorientation [39]. By mating the loxP containing mice with miceexpressing Cre recombinase under the control of a cell typespecific promotor, the target gene will be inactivated from themoment that the promotor driving Cre expression becomesactive in the target cells. At this point two mouse lines withfloxed Pex genes are available i.e. Pex5-loxP [40] and Pex13-loxP mice [41].

5.1. Mice with selective inactivation of peroxisomes inhepatocytes in the postnatal period

As already mentioned, the only defect observed in liver ofnewborn Pex2, Pex5 and Pex13 knockout mice was a change inthe ultrastructure of the mitochondrial inner membrane. In orderto investigate hepatic changes developing at later ages, Pex5-loxPmice were bred with mice expressing Cre under the controlof the albumin promotor [42]. The resulting mice with hepa-tocyte specific elimination of peroxisomes will be denotedfurther as L-Pex5 knockout mice [43].

Functional peroxisomes were eliminated from the liverbetween the first and second postnatal week, as shown by anumber of biochemical parameters. Electronmicroscopic anal-ysis of 10-week-old mice further confirmed the virtual absenceof catalase containing peroxisomes from hepatocytes but notfrom endothelial cells and Kuppfer cells [43]. Only a few smallclusters of hepatocytes with catalase-positive peroxisomes werefound. Surprisingly, plasmalogen and C26:0 levels were normalin the peroxisome deficient livers suggesting that peroxisomespresent in other tissues provided precursors for plasmalogensand degraded C26:0. Other parameters i.e. the concentration ofbranched chain fatty acids phytanic acid and pristanic acid andthe ratio of C27/C24 bile acids were increased in L-Pex5 knock-out livers, as expected.


Two features that are common in postnatal Zellweger patientsalso occurred in L-Pex5 knockout mice. First, they displayed asevere hepatomegaly which was due to hypertrophic andhyperplastic hepatocytes, particularly in the pericentral region(Fig. 2A, B). Furthermore, fibrosis developed in these liversfrom the age of 20 weeks on (Fig. 2C, D). Throughout thelifetime of the mice (>15 months), they looked healthy, werefertile and hepatocellular integrity was unaffected as judged bynormal plasma values of alanine and aspartate aminotransferase.However, all the L-Pex5 knockout mice developed extensiveliver tumours from 12 months on [43].

Severe changes in mitochondrial ultrastructure were ob-served in 60–70% of the mitochondria in peroxisome deficienthepatocytes. The most common finding was the proliferation ofpleomorphic mitochondria with rarefication of cristae or withabnormally curled or stacked cristae. Strikingly, mitochondriawere normal in the few hepatocytes in which catalase-positiveperoxisomes were present, indicating that this is a cell autono-mous phenomenon. Additional ultrastructural changes inhepatocytes lacking perixosomes were the proliferation ofsmooth endoplasmic reticulum (sER) and the appearance oflipid droplets and large groups of lysosomes with electron-densedeposits around dilated bile canaliculi. The ER proliferationwas associated with the induction of the cytochrome P450ω-hydroxylation enzyme CYP4A1 [43].

Analyses of mitochondrial functions demonstrated that thecomplexes I, III and Vof the respiratory chain that are embeddedin the inner membrane were severely inactivated whereas severalmatrix enzymes including citrate synthase and mitochondrialβ-oxidation were induced. Given the importance of the com-plexes to generate ATP, it was surprising that no significantdepletion of ATP levels were found [43]. However, a compen-satory increase in glycolysis as an alternative source of ATP wasdemonstrated. Another consequence of the impaired functioningof the respiratory chain was a collapse of the inner mitochondrialmembrane potential (Fig. 2E, F). Because dysfunction of theelectron transport chain is often accompanied by the generationof oxygen radicals that in turn can cause mitochondrial damage,the hypothesis that reactive oxygen species were increased in theperoxisome deficient livers was investigated. However, neitheroxidative damage to proteins or lipids, nor increased peroxideproduction in cultured hepatocytes or elevation of oxidativestress defence mechanisms were found, indicating that themitochondrial alterations were not related to an excessive pro-duction of oxidative radicals [43].

Altogether, it can be concluded that peroxisomes are essentialfor the maintenance of other subcellular compartments in adulthepatocytes although their absence is not detrimental for thefunctioning of the cell. The molecular interrelationship betweenperoxisomes and mitochondria remains to be investigated.

5.2. Mice with selective inactivation of peroxisomes in Sertolicells

In view of the early postnatal death of patients with pero-xisome biogenesis disorders, not much is known about theimportance of peroxisomes for testicular function and fertility.

Testicular abnormalities were reported in patients with otherperoxisomal diseases with longer survival such as X-linkedadrenoleukodystrophy [44,45] or X-linked adrenomyeloneuro-pathy [46]. Male infertility was reported in three different mousemodels with single peroxisomal enzyme deficiencies i.e. acyl-CoA oxidase [47], MFP-2 [48] and DHAPAT [26] knockoutmice. In acyl-CoA oxidase deficient mice a reduction in theLeydig cell population and hypospermatogenesis were seenwhereas inDHAPAT knockout mice an arrest of spermatogenesisbefore the stage of round spermatids resulting in the completeabsence of spermatozoa was described. In contrast, infertility inMFP-2 knockout mice [48] was associated with the accumula-tion of huge lipid droplets in Sertoli cells followed by a completefatty degeneration of the tubuli seminiferi while Leydig cellfunction was preserved. In fact, until recently it was thought thatperoxisomes were present only in the interstitial cells of Leydigbut according to new data [48,49] they occur also in sperma-togonia and in Sertoli cells, both located in the basal compart-ment of the seminiferous epithelium.

The Cre-loxP technology was recently applied to demon-strate the crucial role of peroxisomes in Sertoli cells. Therefore,Pex5-loxP mice were bred with mice expressing Cre under thecontrol of the Anti-Mullerian hormone (AMH) promoter[48,50]. In the testis extensive accumulations of neutral lipidswere observed in Sertoli cells (Fig. 2G, H), beginning in pre-pubertal mice and evolving in complete testicular atrophy by theage of 4 months. Spermatogenesis was already severely affectedat the age of 7 weeks and pre- and postmeiotic germ cellsgradually disappeared from the tubuli seminiferi. The AMH-Pex5 knockout mice were completely infertile. Together withanalogous Sertoli cell abnormalities in MFP-2 knockout mice[48], these data strongly indicate that peroxisomalβ-oxidation isessential to maintain the lipid balance in Sertoli cells which inturn is crucial for male fertility. It needs to be further investigatedhow the fatty degeneration of the testis relates to the loss ofperoxisomal function in Sertoli cells.

6. Conclusions

The phenotypic analyses of mouse models with peroxisomebiogenesis disorders have taught us that several but not allchanges overlap with those occurring in human diseases. Thediscrepancies may relate to the huge time difference in develop-mental period between mice and men and to differences indietary composition.

In order to understand the importance of peroxisomes for thefunctioning of different tissues, these animal models still need tobe investigated in more detail, including gene expression andmore extensive lipidomic analyses. The latter will requiresensitive techniques because several of the substrates or reactionproducts of the peroxisomal pathways are present in very lowamounts. It will be further instrumental to compare the metabolicand histological abnormalities of the Pex knockouts with thoseof single peroxisomal enzyme or transporter knockouts that arealready available or being generated (Acox1 [47], MFP-1[51],MFP-2 [48,52,53], SCPx [54], phytanoyl-CoA hydroxylase, α-methylacyl-CoA racemase [55], DHAPAT [26], ALD [56–58],

Fig. 2. Pathology in tissue selective Pex5 knockout mice. (A–B) Hematoxylin–eosin staining of livers of 10-week-old control (A) and L-Pex5 knockout mice (B)reveals hypertrophic and hyperplastic hepatocytes in the knockout, in particular in the pericentral region. (C–D) Sirius red stains collagen fibers in 20-week-old L-Pex5knockout mice (D), compatible with fibrosis, but not in control mice (C). (E–F) Hepatocyte cultures were incubated with the mitotracker JC-1. Mitochondria fromcontrol mice (E) fluoresce orange corresponding with a normal inner mitochondrial membrane potential whereas those of L-Pex5 knockout mice fluoresce greenindicative of a reduced potential, n=nucleus. (G–H)Oil redO stains neutral lipid droplets in the outer layer of tubuli seminiferi of 9-week-oldAMH-Pex5 knockoutmice(H) (arrowheads) but not in control mice (G). This is compatible with a Sertoli cell localisation. Leydig cells (arrows) stain in both genotypes because of their high steroidcontent. (A–F) Reprinted from [43] with permission from the publisher.



ALDR [59], PMP34,…). The possibility that some of the ob-served changes are not related to metabolic disturbances but toan altered distribution of peroxisomes or peroxisomal remnantswill also need to be further substantiated.

Finally, the breeding of Pex-loxP mice with appropriate Creexpressing mice may reveal more unsuspected roles of peroxi-somes in certain cell types as already exemplified by the Sertolicell selective inactivation of Pex5.

Acknowledgements

This work was supported by grants from Fonds Wetenschap-pelijk Onderzoek Vlaanderen (G.0235.01 and G.0385.05),Geconcerteerde Onderzoeksacties (99/09 and 2004/08) and theEuropean Union (“MMPD”, QLG1-CT2001-01277, FP5 and“Peroxisomes”, LSHG-CT-2004-512018, FP6). The authorswish to thank Prof. G. Mannaerts for critically reading themanuscript.

References

[1] R.J.A. Wanders, H.R. Waterham, Biochemistry of mammalian peroxi-somes revisited, Annu. Rev. Biochem. 75 (2006) 295–332.

[2] M. Baes, P. Gressens, E. Baumgart, P. Carmeliet, M. Casteels, M. Fransen,P. Evrard, D. Fahimi, P.E. Declercq, D. Collen, P.P. Van Veldhoven, G.P. Mannaerts, A mouse model for Zellweger syndrome, Nat. Genet. 17(1997) 49–57.

[3] P.L. Faust, M.E. Hatten, Targeted deletion of the PEX2 peroxisomeassembly gene in mice provides a model for Zellweger syndrome, a humanneuronal migration disorder, J. Cell Biol. 139 (1997) 1293–1305.

[4] M. Maxwell, J. Bjorkman, T. Nguyen, P. Sharp, J. Finnie, C. Paterson, I.Tonks, B.C. Paton, G.F. Kay, D.I. Crane, Pex13 inactivation in the mousedisrupts peroxisome biogenesis and leads to a Zellweger syndrome pheno-type, Mol. Cell. Biol. 23 (2003) 5947–5957.

[5] S.J. Gould, G.V. Raymond, D. Valle, The peroxisome biogenesis disorders,in: C.R. Scriver, A.L. Beaudet, D. Valle, W.S. Sly (Eds.), The Metabolicand Molecular Bases of Inherited Disease, McGraw-Hill, New York, 2001,pp. 3181–3217.

[6] P.L. Faust, Abnormal cerebellar histogenesis in Pex2 Zellweger micereflects multiple neuronal defects induced by peroxisome deficiency,J. Comp. Neurol. 461 (2003) 394–413.

[7] T. Nguyen, J. Bjorkman, B.C. Paton, D.I. Crane, Failure of microtubule-mediated peroxisome division and trafficking in disorders with reducedperoxisome abundance, J. Cell Sci. 119 (2006) 636–645.

[8] M. Schrader, M. Thiemann, H.D. Fahimi, Peroxisomal motility andinteraction with microtubules, Microsc. Res. Tech. 61 (2003) 171–178.

[9] E. Baumgart, I. Vanhorebeek, M. Grabenbauer, M. Borgers, P. Declercq,H.D. Fahimi, M. Baes, Mitochondrial alterations caused by defectiveperoxisomal biogenesis in a mouse model for Zellweger syndrome (PEX5knockout mouse), Am. J. Pathol. 159 (2001) 1477–1494.

[10] W.J. Kovacs, J.E. Shackelford, K.N. Tape, M.J. Richards, P.L. Faust, S.J.Fliesler, S.K. Krisans, Disturbed cholesterol homeostasis in a peroxi-some-deficient PEX2 knockout mouse model, Mol. Cell. Biol. 24 (2004)1–13.

[11] S. Goldfischer, C.L. Moore, A.B. Johnson, A.J. Spiro, M.P. Valsamis, H.K.Wisniewski, R.H. Ritch, W.T. Norton, I. Rapin, L.M. Gartner, Peroxisomaland mitochondrial defects in the cerebro-hepato-renal syndrome, Science182 (1973) 62–64.

[12] J.M.F .Trijbels, J.A. Berden, L.A.H. Monnens, J.L. Willems, A.J.M.Janssen, R.B.H. Schutgens, M. Van den Broek-Van Essen, Biochemicalstudies in the liver and muscle of patients with Zellweger syndrome,Pediatr. Res. 17 (1983) 514–517.

[13] R.K. Mathis, J.B. Watkins, P. Szczepanik-Van Leeuwen, I.T. Lott, Liver in

the cerebro-hepato-renal syndrome: defective bile acid synthesis andabnormal mitochondria, Gastroenterology 79 (1980) 1311–1317.

[14] S. Huyghe, M. Casteels, A. Janssen, L. Meulders, G.P. Mannaerts, P.E.Declercq, P.P. Van Veldhoven, M. Baes, Prenatal and postnatal develop-ment of peroxisomal lipid-metabolizing pathways in the mouse, Biochem.J. 353 (2001) 673–680.

[15] B.J. Pettus, M. Baes, M. Busman, Y.A. Hannun, P.P. Van Veldhoven, Massspectrometric analysis of ceramide perturbations in brain and fibroblasts ofmice and human patients with peroxisomal disorders, Rapid Commun.Mass. Spectrom. 18 (2004) 1569–1574.

[16] A. Janssen, M. Baes, P. Gressens, G.P. Mannaerts, P. Declercq, P.P.Van Veldhoven, Docosahexaenoic acid deficit is not a major pathogenicfactor in peroxisome-deficient mice, Lab. Invest. 80 (2000) 31–35.

[17] I. Vanhorebeek, M. Baes, P.E. Declercq, Isoprenoid biosynthesis is notcompromised in a Zellweger syndrome mouse model, Biochim. Biophys.Acta 1532 (2001) 28–36.

[18] S. Hogenboom, G.J. Romeijn, S.M. Houten, M. Baes, R.J.A. Wanders,H.R. Waterham, Absence of functional peroxisomes does not lead todeficiency of enzymes involved in cholesterol biosynthesis, J. Lipid Res.43 (2002) 90–98.

[19] P. Evrard, V.S. Caviness, J. Prats-Vinas, G. Lyon, The mechanism of arrestof neuronal migration in the Zellweger malformation: an hypothesis basedupon cytoarchitectonic analysis, Acta Neuropathol. 41 (1978) 109–117.

[20] J.J. Volpe, R.D. Adams, Cerebro-hepato-renal syndrome of Zellweger: aninherited disorder of neuronal migration, Acta Neuropathol. 20 (1972)175–198.

[21] J.M. Powers, H.W. Moser, Peroxisomal disorders: genotype, phenotype,major neuropathologic lesions, and pathogenesis, Brain Pathol. 8 (1998)101–120.

[22] M. Baes, P. Gressens, S. Huyghe, K. De Nys, C. Qi, Y. Jia, G.P. Mannaerts,P. Evrard, P.P. Van Veldhoven, P.E. Declercq, J.K. Reddy, The neuronalmigration defect in mice with Zellweger syndrome (Pex5 knockout) is notcaused by the inactivity of peroxisomal b-oxidation, J. Neuropathol. Exp.Neurol. 61 (2002) 368–374.

[23] S. Ferdinandusse, S. Denis, P.A.W. Mooyer, C. Dekker, M. Duran, R.J.Soorani-Lunsing, E. Boltshauser, A. Macaya, J. Gärtner, C.B.L.M.Majoie,P.G. Barth, R.J.A. Wanders, B.T. Poll, The clinical and biochemicalspectrum of D-bifunctional protein deficiency, Ann. Neurol. 59 (2006)92–104.

[24] G.J. Bix, G.D. Clark, Platelet-activating factor receptor stimulationdisrupts neuronal migration in vitro, J. Neurosci. 18 (1998) 307–318.

[25] P. Gressens, M. Baes, P. Leroux, A. Lombet, P. Van Veldhoven, A. Janssen,J. Vamecq, S. Marret, P. Evrard, Neuronal migration disorder in Zellwegermice is secondary to glutamate receptor dysfunction, Ann. Neurol. 48(2000) 336–343.

[26] C. Rodemer, T.P. Thai, B. Brugger, T. Kaercher, H. Werner, K.-A. Nave, F.Wieland, K. Gorgas, W.W. Just, Inactivation of ether lipid biosynthesiscauses male infertility, defects in eye development and optic nervehypoplasia in mice, Hum. Mol. Genet. 12 (2003) 1881–1895.

[27] A. Janssen, P. Gressens, M. Grabenbauer, E. Baumgart, A. Schad, I.Vanhorebeek, A. Brouwers, P.E. Declercq, D. Fahimi, P. Evrard, L.Schoonjans, D. Collen, P. Carmeliet, G. Mannaerts, P. Van Veldhoven, M.Baes, Neuronal migration depends on intact peroxisomal function in brainand in extraneuronal tissues, J. Neurosci. 23 (2003) 9732–9741.

[28] P.L. Faust, D. Banka, R. Siriratsivawong, V.G. Ng, T.M. Wikander,Peroxisome biogenesis disorders: the role of peroxisomes and metabolicdysfunction in developing brain, J. Inherit. Metab. Dis. 28 (2005) 369–383.

[29] P. Brites, A.M. Motley, P. Gressens, P.A.W. Mooyer, I. Ploegaert, V.Everts, P. Evrard, P. Carmeliet, M. Dewerchin, M. Duran, H.R. Waterham,R.J.A. Wanders, M. Baes, Impaired neuronal migration and endochondralossification in PEX7 knockout mice: a model for rhizomelic chondrodys-plasia punctata, Hum. Mol. Genet. 12 (2003) 2255–2267.

[30] P.E. Purdue, M. Skoneczny, X. Yang, J.W. Zhang, P.B. Lazarow,Rhizomelic chondrodysplasia punctata, a peroxisomal biogenesis disordercaused by defects in Pex7p, a peroxisomal protein import receptor: aminireview, Neurochem. Res. 24 (1999) 581–586.

[31] S. Thoms, R. Erdmann, Dynamin-related proteins and Pex11 proteins inperoxisome division and proliferation, FEBS J. 272 (2005) 5169–5181.


[32] X. Li, S.J. Gould, PEX11 promotes peroxisome division independently ofperoxisome metabolism, J. Cell Biol. 156 (2002) 643–651.

[33] M. Schrader, B.E. Reuber, J.C. Morrell, G. Jimenez-Sanchez, C. Obie,T.A. Stroh, D. Valle, T.A. Schroer, S.J. Gould, Expression of PEX11βmediates peroxisome proliferation in the absence of extracellularstimuli, J. Biol. Chem. 273 (1998) 29607–29614.

[34] X. Li, S.J. Gould, The dynamin-like GTPase DLP1 is essential forperoxisome division and is recruited to peroxisomes in part by PEX11,J. Biol. Chem. 278 (2003) 17012–17020.

[35] X. Li, E. Baumgart, G.-X. Dong, J.C. Morrell, G. Jimenez-Sanchez, D.Valle, K.D. Smith, S.J. Gould, PEX11α is required for peroxisomeproliferation in response to 4-phenylbutyrate but is dispensable forperoxisome proliferator-activated receptor alpha-mediated peroxisomeproliferation, Mol. Cell. Biol. 22 (2002) 8226–8240.

[36] X. Li, E. Baumgart, J.C. Morrell, G. Jimenez-Sanchez, D. Valle, S.J.Gould, PEX11β deficiency is lethal and impairs neuronal migration butdoes not abrogate peroxisome function, Mol. Cell. Biol. 22 (2002)4358–4365.

[37] S. Bielas, H. Higginbotham, H. Koizumi, T. Tanaka, J.G. Gleeson, Corticalneuronal migration mutants suggest separate but intersecting pathways,Annu. Rev. Cell Dev. Biol. 20 (2004) 593–618.

[38] P. Gressens, Pathogenesis of migration disorders, Curr. Opin. Neurol. 19(2006) 135–140.

[39] K.-M. Kwan, Conditional alleles in mice: practical considerations fortissue-specific knockouts, Genesis 32 (2002) 49–62.

[40] M. Baes, M. Dewerchin, A. Janssen, D. Collen, P. Carmeliet, Generation ofPex5-IoxP mice allowing the conditional elimination of peroxisomes,Genesis 32 (2002) 177–178.

[41] J. Bjorkman, I. Tonks, M.A. Maxwell, C. Paterson, G.F. Kay, D.I. Crane,Conditional inactivation of the peroxisome biogenesis Pex13 gene byCre-IoxP excision, Genesis 32 (2002) 179–180.

[42] C. Postic, M. Shiota, K.D. Niswender, T.L. Jetton, Y. Chen, J.M. Moates,K.D. Shelton, J. Lindner, A.D. Cherrington, M.A. Magnuson, Dual rolesfor glucokinase in glucose homeostasis as determined by liver andpancreatic β cell-specific gene knock-outs using Cre recombinase, J. Biol.Chem. 274 (1999) 305–315.

[43] R. Dirkx, I. Vanhorebeek, K. Martens, A. Schad, M. Grabenbauer, D.Fahimi, P. Declercq, P.P. Van Veldhoven, M. Baes, Absence of peroxi-somes in hepatocytes causes mitochondrial and ER abnormalities,Hepatology 41 (2005) 868–878.

[44] J.M. Powers, H.H. Schaumburg, The testis in adreno-leukodystrophy, Am.J. Pathol. 102 (1981) 90–98.

[45] H.W. Moser, K.D. Smith, P.A. Watkins, J.M. Powers, A.B. Moser,X-linked adrenoleukodystrophy, in: C.R. Scriver, A.L. Beaudet, D. Valle,W.S. Sly (Eds.), The Metabolic and Molecular Bases of Inherited Disease,McGraw-Hill, New York, 2001, pp. 3257–3301.

[46] A. Aversa, S. Palleschi, G. Cruccu, L. Silverstroni, A. Isidori, A. Fabbri,Rapid decline of fertility in a case of adrenoleukodystrophy, Hum. Reprod.13 (1998) 2474–2479.

[47] C.-Y. Fan, J. Pan, R. Chu, D. Lee, K.D. Kluckman, N. Usuda, I.Singh, A.V. Yeldandi, M.S. Rao, N. Maeda, J.K. Reddy, Hepatocellularand hepatic peroxisomal alterations in mice with a disrupted

peroxisomal fatty acyl-coenzyme A oxidase gene, J. Biol. Chem. 271(1996) 24698–24710.

[48] S. Huyghe, H. Schmalbruch, K. De Gendt, G. Verhoeven, F. Guillou, P.P.Van Veldhoven, M. Baes, Peroxisomal multifunctional protein 2 isessential for lipid homeostasis in Sertoli cells and for male fertility in mice,Endocrinology 147 (2006) 2228–2236.

[49] G. Lüers, S. Thiele, A. Schad, A. Völkl, S. Yokota, J. Seitz, Peroxisomesare present in murine spermatogonia and disappear during the course ofspermatogenesis, Histochem. Cell Biol. 30 (2005) 1–11.

[50] C. Lecureuil, I. Fontaine, P. Crepieux, F. Guillou, Sertoli and granulosacell-specific Cre recombinase activity in transgenic mice, Genesis 33(2002) 114–118.

[51] C. Qi, Y. Zhu, J. Pan, N. Usuda, N. Maeda, A.V. Yeldandi, M.S. Rao, T.Hashimoto, J.K. Reddy, Absence of spontaneous peroxisome proliferationin enoyl-CoA hydratase/L-3-hydroxyacyl-CoA dehydrogenase-deficientmouse liver, J. Biol. Chem. 274 (1999) 15775–15780.

[52] M. Baes, S. Huyghe, P. Carmeliet, P.E. Declercq, D. Collen, G.P.Mannaerts, P.P. Van Veldhoven, Inactivation of the peroxisomalmultifunctional protein-2 in mice impedes the degradation of not only2-methyl branched fatty acids and bile acid intermediates but also of verylong chain fatty acids, J. Biol. Chem. 275 (2000) 16329–16336.

[53] S. Huyghe, H. Schmalbruch, L. Hulshagen, P.P. Van Veldhoven, M. Baes,D. Hartmann, Peroxisomal multifunctional protein-2 deficiency causesmotor deficits and glial lesions in the adult CNS, Am. J. Pathol. 168 (2006)1321–1334.

[54] U. Seedorf, M. Raabe, P. Ellinghaus, F. Kannenberg, M. Fobker, T. Engel,S. Denis, F. Wouters, K.W.A. Wirtz, R.J.A. Wanders, N. Maeda, G.Assmann, Defective peroxisomal catabolism of branched fatty acylcoenzyme A in mice lacking the sterol carrier protein-2/sterol carrierprotein-x gene function, Genes Dev. 12 (1998) 1189–1201.

[55] K. Savolainen, T.J. Kotti, W. Schmitz, T.I. Savolainen, R.T. Sormunen, M.Ilves, S.J. Vainio, E. Conzelmann, J.K. Hiltunen, A mouse model for a-methylacyl-CoA racemase deficiency: adjustment of bile acid synthesisand intolerance to dietay methyl-branched lipids, Hum. Mol. Genet. 13(2004) 955–965.

[56] S. Forss-Petter, H. Werner, J. Berger, H. Lassmann, B. Molzer, M.H.Schwab, H. Bernheimer, F. Zimmermann, K.-A. Nave, Targeted inactiva-tion of the X-linked adrenoleukodystrophy gene in mice, J. Neurosci. Res.50 (1997) 829–843.

[57] T. Kobayashi, N. Shinnoh, A. Kondo, T. Yamada, Adrenoleukodystrophyprotein-deficient mice represent abnormality of very long chain fatty acidmetabolism, Biochem. Biophys. Res. Commun. 232 (1997) 631–636.

[58] A. Pujol, C. Hindelang, N. Callizot, U. Bartsch, M. Schachner, J.L.Mandel, Late onset neurological phenotype of the X-ALD geneinactivation in mice: a mouse model for adrenomyeloneuropathy, Hum.Mol. Genet. 11 (2002) 499–505.

[59] I. Ferrer, J.P. Kapfhammer, C. Hindelang, S. Kemp, N. Troffer-Charlier,V. Broccoli, N. Callyzot, P. Mooyer, J. Selhorst, P. Vreken, R.J.A.Wanders, J.-L. Mandel, A. Pujol, Inactivation of the peroxisomal ABCD2transporter in the mouse leads to late-onset ataxia involving mitochondria,Golgi and endoplasmic reticulum damage, Hum. Mol. Genet. 14 (2005)3565–3577.

review generalised and conditional inactivation of pex ... · review generalised and conditional...

Documents