article mouse models for the wolf–hirschhorn deletion syndrome

© 2001 Oxford University Press Human Molecular Genetics, 2001, Vol. 10, No. 2 91–98

ARTICLE

Mouse models for the Wolf–Hirschhorn deletionsyndromeDieter Näf1, Lawriston A. Wilson1, Rebecca A. Bergstrom1, Richard S. Smith1,Neal C. Goodwin1, Annemieke Verkerk2, Gert Jan van Ommen2, Susan L. Ackerman1,Wayne N. Frankel1 and John C. Schimenti1,+

1The Jackson Laboratory, Bar Harbor, ME 04609, USA and 2MGC-Department of Human and Clinical Genetics,Leiden University Medical Center, Leiden, The Netherlands

Received 6 September 2000; Revised and Accepted 7 November 2000

Wolf–Hirschhorn syndrome (WHS) is a deletion syndrome caused by segmental haploidy of chromosome4p16.3. Its hallmark features include a ‘Greek warrior helmet’ facial appearance, mental retardation, variousmidline defects and seizures. The WHS critical region (WHSCR) lies between the Huntington’s disease gene,HD, and FGFR3. In mice, the homologs of these genes map to chromosome 5 in a region of conserved syntenywith human 4p16.3. To derive mouse models of WHS and map genes responsible for subphenotypes of thesyndrome, five mouse lines bearing radiation-induced deletions spanning the WHSCR syntenic region weregenerated and characterized. Similar to WHS patients, these animals were growth-retarded, were susceptibleto seizures and showed midline (palate closure, tail kinks), craniofacial and ocular anomalies (colobomas,corneal opacities). Other phenotypes included cerebellar hypoplasia and a shortened cerebral cortex. Expressionof WHS-like traits was variable and influenced by strain background and deletion size. These mice representthe first animal models for WHS. This collection of nested chromosomal deletions will be useful for mappingand identifying loci responsible for the various subphenotypes of WHS, and provides a paradigm for thedissection of other deletion syndromes using the mouse.

INTRODUCTION

Deletion syndromes [also referred to as segmental aneusomysyndromes or contiguous gene syndromes (CGSs: the abbreviationused here)], such as Cri-du-chat (OMIM 123450), Smith-Magenis (OMIM 182290) and Williams (OMIM 194050), arean interesting class of human disease. Manifested as aconsequence of hemizygosity, CGSs illustrate that diploidy isimportant for proper development. In the case of Prader–Williand Angelman syndromes (OMIM 176270 and 105830,respectively), the phenomenon of genomic imprinting (parent-of-origin-dependent gene silencing) is uncovered. Thus, CGSscan offer insights into such fundamental processes as geneexpression, epigenetics and regulation of development.

Two major difficulties in identifying the genes underlyingcomponent phenotypes of a CGS are that these diseases arequite rare and the deletions are usually large. In the cases ofPrader-Willi, Williams and DiGeorge syndromes, the deletionsoften have common breakpoints, arising from recombinationevents catalyzed by duplicated sequences (1–3). Withoutrandomly staggered breakpoints, it is difficult to preciselydefine the chromosomal regions that underlie aspects of thesesyndromes. Furthermore, the common ‘core’ features associated

with a given CGS can vary in severity between patients, potentiallyconfounding the mapping of syndrome components. Finally,cryptic deletions that do not have clinical consequence goundetected in humans, thereby eliminating a potential geneticmapping resource for excluding regions contributing to a CGS.

Technologies for manipulating the mouse genome haveafforded opportunities to model and genetically dissect aspectsof human CGSs. Individual deletions in mice have been generatedto identify potential roles of genes in the Prader–Willisyndrome (4–6), and to produce heart defects characteristic ofthe DiGeorge syndrome (7). However, there have been noreports of using nested deletion sets to model a CGS or todissect component phenotypes.

Wolf–Hirschhorn syndrome (WHS; OMIM 194190) is a rareCGS characterized by a ‘Greek warrior helmet’ facial appearance(prominent glabella, hypertelorism, widely set eyes), variousmidline closure defects, growth retardation, juvenile seizuredisorders, cataracts, iris colobomas and mental retardation(8,9). Molecular characterization of deletion breakpoints inWHS patients led to the identification of a 165 kb WHS criticalregion (WHSCR), defined by the minimal region of deletionoverlap in two key individuals. The WHSCR is located withina 2 Mb interval flanked by the Huntington’s disease (HD) and

+To whom correspondence should be addressed. Tel: +1 207 288 6402; Fax: +1 207 288 6082; Email: [email protected]

92 Human Molecular Genetics, 2001, Vol. 10, No. 2

FGFR3 genes, a region sequenced during the positionalcloning of HD (10). Candidate WHS genes identified in thisregion include WHSC1, the product of which contains motifspresent in certain developmentally important proteins andexhibits an embryonic expression pattern consistent with structuresaffected in WHS (11), and WHSC2, a ubiquitously expressedgene of unknown function (12).

The mouse orthologs of HD and FGFR3 (Hdh and Fgfr3)map to chromosome 5 in a region of conserved synteny withhuman 4p16.3 (13), and thus presumably demarcate the mouseversion of the WHSCR. To generate potential mouse models ofWHS, a collection of deletions was induced at the Hdh locus ofembryonic stem (ES) cells by irradiation (14). Mice derivedfrom five of these lines were analyzed for phenotypes causedby segmental haploidy. Those with the larger deletionsdisplayed several characteristics of WHS, including craniofacialdysmorphology, seizure susceptibility and midline defects.The combined data suggest that manifestation of the moresevere aspects of WHS requires hemizygosity of loci outsidethe WHSCR, and that the traits are affected by genetic back-ground.

RESULTS

In previous work, a series of nested deletions centered at theHuntington’s locus (Hdh) on chromosome 5 (14) was generatedby the technique of ES cell irradiation (15). This involved threesteps: (i) targeting a herpes simplex virus thymidine kinasegene (tk) into the Hdh locus of F1 hybrid (129S4/SvJae× C57BL/6J) ES cells by homologous recombination;

(ii) subjecting targeted clones to ionizing radiation; and(iii) selecting for deletions of tk by culturing in mediumcontaining 1,2′-deoxy-2′-fluoro-b-D-arabinofuranosyl-5-iodouracil(FIAU). A subset of deletion-bearing ES cell clones wasinjected into blastocysts to create germline chimeric mice. Fivesuch deletions were analyzed in this study (Fig. 1). Somepreliminary studies (on seizure susceptibility; see below) wereperformed with a deletion induced at the Dpp6 locus (Fig. 1).

In matings to C57BL/6J (B6) females, these chimeras siredoffspring that primarily inherited the non-deleted 129S4/SvJae(129) chromosome (Table 1). The few that inherited the largestdeletions (namely Hdhdf2J, Hdhdf4J and Hdhdf6J) were severelygrowth-retarded. Transmission was increased by mating tostrains C3H or CAST/Ei, yet most deletion-bearing progenywere still runted, averaging 50–80% of the weight of same-sexlittermates (Fig. 2). Fitness improved on further backcrossingto C3H (data not shown). When B6(Hdhdf7J/+) N2 animals werebackcrossed to B6, nearly all deletion-bearing pups wereobserved gasping for air immediately after birth, and diedshortly thereafter. Histological examination of these pupsrevealed collapsed lungs (data not shown).

Deletion breakpoints were mapped by exploiting chromo-some 5 microsatellite polymorphisms between B6 and 129 inES cells, or between either of the strains against which thedeletions were placed in trans: C3H and CAST/Ei (14). Tobetter characterize the deletions with respect to the WHSCR,experiments were conducted to ascertain the location ofWhsc1, a candidate WHS gene which maps to the 165 kbWHSCR in humans (11). First, the presumed tight linkage ofWhsc1 to the Hdh locus in mice was evaluated. A single strand

Figure 1. Map of Hdh region deletions. The targeted region of mouse chromosome 5 is represented by a black bar. Genetic map positions (from Mouse GenomeDatabase, http://www.informatics.jax.org/ ) are given in centiMorgans relative to the centromere. Shaded thick boxes indicate the minimum extent of the deletions,and the thinner shaded lines extending from the ends indicate the intervals in which the breakpoints lie. The Dpp6df1J deletion was induced by irradiation of ES cellscontaining a thymide kinase gene inserted at Dpp6, whereas the rest were induced at the Hdh locus (14). The Hdh region is expanded, showing the approximatelocation of the WHSCR (10,12,29). The exact order of mouse genes in the Hdh cluster is not known; however, Hdh and Fgfr3 are clustered together in both species(30), and the conserved orders of flanking loci surrounding Hdh/Fgfr3 imply the arrangement in mouse as shown (13). Positions of the mouse tilted (tlt) locus andsurrounding MIT markers are derived from genetic data (31) and deletion breakpoint analysis (14). LetM1 is a novel gene, putatively encoding a novel Ca2+-bindingprotein; its location is based on the human map position and analyses of mouse bacterial artificial chromosome clones from the Hdh–Fgfr3 region (data not shown).The break in the chromosome between positions 10 and 16 indicates the interval in which there is no synteny to human 4p.

Human Molecular Genetics, 2001, Vol. 10, No. 2 93

conformation polymorphism (SSCP) was used to geneticallymap Whsc1 on the Jackson Laboratory’s BSS interspecific back-cross mapping panel (http://www.jax.org/resources/documents/cmdata/ ). Whsc1 was inseparable from the D5Mit148 andD5Mit149 loci that flank Hdh. Second, a restriction site poly-morphism between strains C3H and B6 in the 3′-untranslatedregion of Whsc1 was identified and exploited to screen the deletionset. Whsc1 was absent in all deletions.

To better characterize Hdhdf9J, which retained heterozygosityat all polymorphic microsatellites near Hdh, FISH analysis ofES cells was performed using mouse PACs corresponding tothe WHSCR syntenic region as probes (Fig. 3E). This revealedthat cells bearing the Hdhdf9J deletion are hemizygous for allgenes in the LetM1–Hdh interval, which spans ∼2 Mb in

humans (the physical size of this region in mice is not known),but retained two copies of the closest flanking microsatellitemarkers (D5Mit149 and D5Mit388) which lie within 1 cM ofHdh according to the Mouse Genome Database. Thus, Hdhdf9J,as well as all the other Hdh deletions used in this study, spanthe entire 165 kb WHSCR of humans (Fig. 1).

Since hemizygosity of the WHSCR is the molecular basis fordiagnosis of WHS in humans, we examined deletion-bearingmice for WHS-like phenotypes. Several animals harboring thedeletions Hdhdf2J, Hdhdf4J, Hdhdf6J and Hdhdf7J in a backgroundcontaining >50% contribution from B6 showed craniofacialfeatures potentially related to the hallmark ‘Greek warriorhelmet’ phenotype of WHS (Fig. 3A). The most prominentanomalies were a domed skull (WHS patients generally have aprotruding nasal bridge), smaller ears (lobeless ears in WHS),widely set eyes (typical of WHS) and, in the most severe cases,a marked facial asymmetry (also typical of WHS) (Fig. 3B and C).Autopsies of two such animals revealed maloccluded lowerincisors (Fig. 3B), misshapened palatine ridges (Fig. 3B), asym-metry of the jaw bones (data not shown) and tail kinks (Fig. 3C),suggesting analogy to midline defects in WHS patients. Tailkinks were commonly present in mice containing the Hdhdf2J,Hdhdf4Jand Hdhdf6J deletions, especially in markedly runtedanimals.

The penetrance and expressivity of marked craniofacialanomalies was clearly background dependent, being mostpronounced in the B6 strain. Attempts to quantitate penetrancewas confounded by the lethality that occurred in crosses of thelargest deletions to this strain (described above). Nevertheless,in cases of individual affected animals that produced multipledeletion-bearing offspring in crosses to B6, the character wasreliably transmitted. For example, all the deletion-bearingoffspring (4 of 21 progeny) of one particular Hdhdf4J malemated to B6 had typical craniofacial defects (domed skull,short ears, wide eyes as in Fig. 3A). However, when the samemale was outcrossed to a non-inbred stock, the deletion-bearing offspring (6/11) had subtle or no observable craniofacialdefects (the phenotype is somewhat subjective). To investigatethe ontogeny of the craniofacial abnormalities, a litterproduced by a mating of this male to a B6 female was sacri-ficed at birth, and bone/cartilage preparations of the skullswere prepared. There were no clear defects observable in theskulls of the four deletion-containing pups in this litter of 10(data not shown), suggesting that the craniofacial dysmorphologiesdevelop during postnatal growth in the mouse.

Two studies found that >85% of WHS patients presentedwith seizure disorders (16,17). In the course of normal mousehusbandry (cage changing; weaning), we observed severalepisodes of spontaneous seizures displayed by deletion-bearing mice. To systematically gauge seizure susceptibility,we tested animals for incidence of electroshock-induced clonicseizures. As detailed in Table 2, only 12% (4/34) of wild-typecontrol littermates suffered induced seizures, whereas >85% ofthe mice heterozygous for Hdhdf6J and Hdhdf7J seized inresponse to the stimulus. The seizure incidence for each deletionwas significantly higher than littermates in both cases (Table 2)and, in general, were qualitatively more severe and sometimesterminal (i.e. maximal tonic-hindlimb seizures). Hdhdf4J

heterozygotes also had more seizures than littermates (60versus 11%, respectively), but the difference was not signifi-cant with the number of animals tested. Notably, half of the

Table 1. Strain-dependent transmission of Hdh deletions from chimeras

aThe fractions denote the number of deletion-bearing offspring sired by thechimeras over total offspring originating from ES cell-derived gametes. In thecase of C57BL/6J dams, only agouti progeny were scored. For C3HeB/FeJdams, progeny were confirmed as ES cell-derived by identifying the presenceof 129/SvJae alleles elsewhere in the genome.bDerived from the one-tailed Fisher’s exact test; compare significance of dif-ferences between transmission to the two indicated strains of dams.

Deletion Strain of dama P valueb

C57BL/6J C3HeB/FeJ

Hdhdf2J 0/14 (0%) 12/21 (57%) <0.0004

Hdhdf4J 1/27 (4%) 4/17 (24%) <0.065

Hdhdf6J 1/22 (5%) 11/42 (26%) 0.032

Hdhdf7J 3/34 (9%) 5/12 (42%) 0.02

Hdhdf9J 3/16 (19%) 6/12 (50%) <0.09

Figure 2. Growth retardation in deletion mice. Mice were weighed between 4and 12 weeks of age. In each litter, the average body weight of wild-type micewas set to 100% and compared with the average weight of sex-matched dele-tion-bearing littermates. Bars represent averages and standard deviations froma minimum of three litters for each deletion.


seizures induced in Hdhdf4J/+ animals were maximal, but nomaximal seizures were observed in controls. The induced seizureincidence in Hdhdf9J/+ mice was similar to that in controls,although occasional Hdhdf9J/+ animals were observed to sufferspontaneous convulsive seizure episodes. It is conceivable thatdeletion of WHSCR alone may be insufficient to confer

electroshock-induced seizure susceptibility, but thatspontaneous seizures might represent a distinct phenotype. It isalso possible, in light of the lower seizure frequency observedin Hdhdf4J/+ mice, that multiple regions contribute to seizuresusceptibility. These issues can be more rigorously addressedif the deletions can be rendered congenic on a defined strainbackground(s). To determine whether hemizygosity of regionsflanking, but not including, the WHSCR cause seizuresusceptibility, mice heterozygous for a deletion centered at theDpp6 locus, Dpp6df1J, were tested by electroshock. Thisdeletion overlaps Hdhdf4J over an ∼5 cM stretch betweenD5Mit348 and D5Mit148 (Fig. 1). None of the four animalstested had an induced seizure (Table 2).

Colobomas (midline closure defects of the optical stalk and/or cup) and cataracts are typical components of the WHSphenotype (18). Several Hdhdf4J/+ and Hdhdf7J/+ animalsexhibited corneal opacities (Fig. 3D). Slit lamp biomicroscopyrevealed eye defects in 23% (9/38) of the Hdhdf7J/+ miceexamined: four animals had iris colobomas (Fig. 3D), four hadcataracts and one had both. No defects were found in Hdhdf9J/+

Figure 3. Phenotypes of Hdh deletions and FISH mapping. (A) Craniofacial features of two Hdhdf4J/+ and wild-type littermates. Note the domed skulls, smallerear lobes and more laterally set eyes of the affected animals. (B) Upper jaw and palate of an Hdhdf2J/+ mouse. An arrow points to the malformed palate. Note alsothe asymmetry of the snout (twisted to the animal’s right) and maloccluded incisors. (C) An example of tail kinks often displayed in deletion-bearing mice. Alsonote the craniofacial assymetry. (D) Eye abnormalities in deletion heterozygotes. (Left) Wild-type; (middle) a coloboma in an Hdhdf4J/+ mouse; (right) a cornealopacity in an Hdhdf7J/+ mouse. (E) FISH analysis of the Hdhdf9J/+ ES cell line. The two red signals correspond to p1-derived artificial chromosome (PAC) probe622E13, which contains the D5Mit76 locus that is just outside the Hdhdf9J deletion region. The green signal corresponds to PAC 571L19, which contains the LetM1,Fgfr3, Whsc1 and Whsc2 genes (Fig. 1). Several nuclei hybridized to these probe pairs gave identical results, demonstrating hemizygosity for the 571L19 PAC.Similar results were obtained with a separate PAC spanning Whsc1 and Hdh.

Table 2. Seizure susceptibility in deletion-bearing mice

aDerived from the one-tailed Fisher’s exact test.

Deletion Del/+ +/+ sibs P valuea

Hdhdf4J 6/10 (60%) 2/11 (18%) 0.063

Hdhdf6J 8/9 (89%) 1/9 (11%) 0.0017

Hdhdf7J 11/13 (85%) 1/6 (16%) <0.01

Hdhdf9J 2/18 (11%) 2/10 (20%) <0.45

Dpp6df1J 0/4 (0%) 0/5 (0%) 1


mice (0/27) of comparable genetic background (∼50% B6).Sporadic colobomas were also observed in seven Hdhdf4J/+mice and two Hdhdf6J/+ animals. Interestingly, no cornealopacities have been observed in mice bearing the Hdhdf2J dele-tion, which is larger than both Hdhdf4J and Hdhdf6J (insufficientanimals were examined for colobomas to draw conclusions).

One of the most defining hallmarks of WHS is mental retar-dation. In a first attempt to identify possible learning deficits,mice bearing the largest deletions were subjected to the MorrisWater Maze test. No differences were noted in comparisonwith controls (data not shown). However, gross examination ofbrains from mice heterozygous for Hdhdf2J, Hdhdf4J, Hdhdf7J andHdhdf9J revealed that the colliculi of mice bearing deletionswere more exposed than those of wild-type controls due tohypoplasia of the cerebral cortex (Fig. 4, arrows). It isunknown whether this anatomical defect has any functionalconsequences.

Although mice bearing the Hdhdf2J and Hdh4J deletionsscored normally in the water maze, we noted that they swamunusually. Furthermore, they displayed an unusual ‘froglike’gait, in which their hindlimbs splayed outward from the bodyaxis as they walked. To investigate, whole mount and sagittalsections through the cerebella of mutant and control animalswere examined. Lobulation defects as a reduction in cerebellarsize was seen in mice heterozygous for Hdhdf4J and Hdhdf2J.Whether these defects underlie the abnormal gait is unclear.No marked cerebellar defects were seen in Hdhdf7J and Hfhdf9J

mice.

DISCUSSION

We have shown in this report that mice carrying large deletionsencompassing the Hdh locus showed several phenotypescharacteristic of WHS. These included seizure susceptibility,iris colobomas, cataracts/corneal opacities, craniofacial

Figure 4. Defects in brain morphology of deletion mice. (Left) Gross analysis of wild-type controls and mice bearing deletions reveal hypoplasia of the cerebralcortex in mutants causing the colliculi (arrows) to be more exposed. (Right) Cerebellar hypoplasia and foliation defects (arrows) were observed in Hdhdf2J mice[and Hdhdf4J (data not shown)]. Mice with these deletions lacked the intraculminate fissure (icul), the posterior superior fissure (ps) and the uvular cus (uvl). Hdhdf9J

mice [and Hdhdf7J (data not shown)] did not lack any of these structures. In all sections, rostral is to the left, dorsal to the top, and the primary fissure (pr) is labelled.


dismorphology, midline defects and growth retardation. Itmust be cautioned that growth retardation is a commonphenotype associated with chromosomal deletions in mice (19)and most other human CGSs. Thus, it is not clear whether thebasis for runting in the Hdh deletion mice is identical to thatwhich underlies intrauterine growth retardation of WHSpatients.

Neonatal mortality, which has also been associated withWHS (18), was especially marked on breeding to the B6genetic background. Interestingly, a similar phenotype wasreported in 129,B6(Ndntm1Stw/+) mice, a model for Prader–Willisyndrome (20). It is possible that strain B6 is intolerant ofhaploid gene expression at these loci or causes certainhemizygous genes to be expressed at subhaploid levels.Another possibility is that, in the B6 background, some genesin deletion regions might undergo randomized uniparentalinactivation, causing a subset of cells to be effectivelynullizygous. Finally, the phenotypic severity of deletion-bearing animals might be related to the genotype of the allelespresent on the non-deleted chromosome. These hypotheses canbe experimentally addressed in the mouse system.

One of the more intriguing findings presented here was thatmice heterozygous for the Hdhdf9J microdeletion, whichremoves the mouse version of the WHSCR but minimalflanking DNA, did not exhibit the prominent WHS-likecharacters. There are several possible explanations. One is thatthe WHS-like phenotypes associated with the larger deletionsare completely unrelated to the genetic causes of WHS inhumans. This possibility would be difficult to prove ordisprove. Another explanation is that hemizygosity of theWHSCR alone may be necessary but not sufficient for the fullmanifestation of WHS. Since there are no known WHSpatients who delete only the WHSCR or subregions thereof, itis currently impossible to evaluate this possibility. However,this hypothesis can be tested in mice. If true, then one predic-tion would be that mice containing deletions that remove DNAflanking, but not including the WHSCR, would fail to displaythe WHS phenotypes reported here. Since mice with therecently isolated Dpp6df1J deletion did not exhibit seizuresusceptibility or other overt WHS-like defects (data notshown), WHSCR-extrinsic genes causing potential autonomousWHS phenocopies would have to reside between D5Mit388and Fgfr3 (the region deleted by Hdhdf2J proximal to theWHSCR) and/or Add1 and D5Mit421 [the region deleted byHdhdf4J and Hdh7J distal to the WHSCR (Fig. 1)]. We arecurrently in the process of analyzing additional deletionscentered at Dpp6 (14) and inducing deletions centered at Qdpr.Analyses of such mice will allow refinement of chromosomalregions containing haploinsufficent genes underlying thephenotypes reported here.

A variation on the hypothesis that both hemizygosity of theWHSCR and deletion of flanking regions is needed to manifestthe core phenotypes of WHS (Greek warrior helmet, seizures,mental retardation) posits that additional phenotypes associ-ated with the syndrome, or modifiers of the syndrome’sseverity, might be related to the extent of DNA deletedflanking the WHSCR. In support of this view, a correlationbetween deletion size and clinical severity has beendocumented for WHS patients in multiple studies (17,21,22).Furthermore, it must be considered that the WHSCR wasdefined on the basis of two patients (CM and LGL7447) who

had deletions (of ∼2.5 and >4 Mb) extending in oppositedirections, overlapping only in the WHSCR (10,23). Sharedfeatures of these patients were seizures, abnormal facies anddevelopmental delay. However, they were not reported to haveother defects typical of WHS, such as colobomas, cleft palateand heart defects. Indeed, analysis of a cadre of patients withchromosome 4p deletions indicated that some anomaliestypical of WHS can be attributable to regions outside theWHSCR (24). In contrast, it was found that certain patientswith Pitt–Rogers–Danks syndrome, which is characterized bymilder WHS-like phenotypes, not only overlap WHS deletionsbut in some cases delete larger stretches of DNA (25,26).

It is evident from examples such as these and others in theliterature that dissecting the genetic basis of phenotypesassociated with WHS, using exclusively clinical data, will beexceedingly difficult or impossible for the following reasons.First, WHS is quite rare, precluding sufficient redundancy bywhich to draw confident conclusions on the phenotypes thatcan be ascribed to a particular chromosomal interval. Second,the variability in the severity and spectrum of abnormalitiesassociated with 4p deletions complicates and exacerbates thisproblem. Third, since subclinical individuals carrying micro-deletions would be identified only rarely, it is difficult to ruleout particular regions from being involved in WHS.

The mouse provides a model system that is suited toovercome the difficulties posed by a disease like WHS. Theability to manipulate the mouse genome in many ways can beexploited to systematically address the genetic basis of WHS.The results presented here lay the groundwork for such studies.Not only do the deletion-bearing mice appear to representrelevant models of WHS, but the severity of the phenotypes ishighly influenced by genetic background, as appears to be thecase in humans. Unlike human patients, the ability to generatean unlimited number of mice with an identical modificationpermits more robust and informative analysis of theconsequences of particular genetic lesions.

Our collection of nested deletion mice, plus others that canbe generated from cryopreserved, deletion-containing ES cellclones centered not only at Hdh but also at flanking loci (14),will be valuable for mapping loci involved in the WHS-likedefects. The current data allow us to assign candidate regionsfor some of the phenotypes examined (the phenotypes of alldeletions are summarized in Table 3). The cerebellarlobulation defect can be localized to either of two intervals:(i) between D5Mit389 and D5Mit388, defined by the proximalbreakpoints of Hdhdf2J, which has the defects, and Hdhdf7J,which does not; or (ii) between D5Mit351 and D5Mit421, inthe event that the distal breakpoint of Hdhdf4J (which has thecerebellar defects) extends more distal to that of Hdhdf7J. Theseizure phenotype lies in a region maximally defined by theHdhdf7J deletion (unless the distal breakpoint of Hdhdf4J is moreproximal), but which excludes the region deleted in Hdhdf9J.The coloboma phenotype can be tentatively localized to theinterval between the proximal breakpoint of Hdhdf6J

(D5Mit388) and the distal breakpoints (D5Mit421) of Hdhdf7J

or Hdhdf4J, all of which have the defect, but not within theregion deleted by Hdhdf9J. The failure to observe corneal opacitiesin Hdhdf2J mice is puzzling. It is possible that this deletion isdiscontinuous, such that an island of DNA containing the locusresponsible for this haploinsufficient defect remains present.Alternatively, this observation may be related to the incomplete


penetrance of the corneal opacities, which are likely due in partto genetic background effects. The Hdhdf2J deletion could notbe transmitted in crosses to C57BL/6J animals (Table 1). It ispossible that those background alleles which predispose tocorneal opacities are incompatible with viability of Hdhdf2J/+ mice.

As regions containing these loci are narrowed, complementationwith individual candidate transgenes or large-insert genomicclones can be attempted to ultimately identify those that areresponsible for phenotypes. Alternatively, knockouts ofindividual WHS candidate genes such as Whsc1 can be generated,but the results here suggest that mice with heterozygousdeficiencies of individual genes in the WHSCR will notreproduce the full clinical spectrum of the disease. Nevertheless,such knockouts might be combined with flanking deletions totest the idea that hemizygosity of genes both within and outsidethe WHSCR might be required for phenotypic manifestation.Finally, further characterization of genetic background effectsmay lead to the identification of loci that participate in thedevelopmental processes that are affected in WHS.

MATERIALS AND METHODS

Induction and characterization of deletions

Targeting into the Hdh gene (C57BL/6J allele), induction ofdeletions in ES cells by irradiation and characterization ofseveral hundred deletion clones with microsatellite markershave been described elsewhere (14). Molecular detection of theWhsc1 locus was assessed by PCR, using primers (whs1, 5′-ATCTAAAATAGAAATGGGAGGG-3′; whs2, 5′-CCAGGAT-GGGCAGTCC-3′) specific for the 3′-untranslated region,followed by digestion of the products (∼445 bp) with TfiI.Whereas the B6 and 129 alleles of Whsc1 contain two TfiIsites, C3H has one.

Tests for seizure susceptibility

Susceptibility to minimal clonic seizures (27) was assessed bya high frequency electroconvulsive threshold test. Briefly,animals were restrained by hand, a drop of saline solutioncontaining the topical anesthetic tetracaine (0.5%) was placedon each eye and 7 mA current was applied via silvertranscorneal electrodes using a electroconvulsive stimulator

(Ugo Basile model 7801). The stimulator was set to producerectangular wave pulses with the following properties: 299 Hz,0.2 s duration, 1.6 ms pulse width. Seven milliamperes issubthreshold for C57BL/6J mice: the CC50 (critical current for50% to have a minimal seizure) in females = 7.92 mA (95%CI: 7.67–8.29); males = 9.35 mA (95% CI: 9.11–9.69).

Eye examinations

A Haag-Streit 900 slitlamp set at a magnification of 40× wasused for ocular examinations. After dilation of the pupil with1% cyclopentolate, the fundus was examined using a binocularindirect ophthalmoscope.

FISH analysis

ES cells treated with 0.1 M KCl were fixed with 3:1methanol:acetic acid, dropped on glass slides and air-driedovernight. PACs were labeled by nick translation using eitherbiotin-11-dUTP or digoxigenin-11-dUTP. Probes were pre-annealed for 1–2 h in the presence of 2 µg of mouse Cot1DNA, then hybridized to the slides overnight. Dual-colorfluorescence techniques were performed essentially asdescribed (28). Signals were visualized in three detection stepsusing either Streptavidin Alexa594 (1:100; Molecular Probes),goat anti-StreptavidinBio (1:100; Vector) and StreptavidineAlexa594 (1:100) or mouse anti-DIG–FITC (1:250; Sigma),rabbit anti-mouse FITC (1:1000; Sigma) and goat anti-rabbitFITC (1:1000). The slides were mounted in Vectashield(Vector) containing DAPI (100 ng/ml; Roche).

Histology and whole brain mounts

Adult mice were deeply anesthetized with avertin and transcardiallyperfused with Bouin’s solution. Skulls were postfixed over-night and brains were removed for examination under adissecting stereomicroscope (Wild M10; Leica) prior toparaffin embedding. Brains were sectioned sagittally and 7 µmsections were stained with hematoxylin and eosin. Sectionswere examined using a Leica DMRXE microscope. High resolutiondigital images were captured by a Spot 1.1 camera (DiagnosticInstruments). The samples shown in Figure 4 are representative ofthe following numbers of individuals examined: two Hdhdf2J,three Hdhdf4J, one Hdhdf7, two Hdhdf9J and three control (non-deletion) littermates.

ACKNOWLEDGEMENTS

We are grateful to M. MacDonald for Hdh plasmids and phageclones; K. Hirschhorn, M. Altherr, T. Vogt, M. Bucan, M.MacDonald, S. White and the members of the Schimentilaboratory for valuable advice and discussion; T. Gridley, J.Naggert and J. den Dunnen for critical reviews of themanuscript; A. Costa and K. Walsh for conducting the Morriswater maze tests; E. Gutteling for technical assistance withFISH; C. Mooser for technical assistance; and Johan denDunnen for continuous support of A.V. This work wassupported by NIH grant HD35984 to J.C.S., NIH grantsNS31348 and NS40246 to W.N.F., Cancer Center grantCA34196 to the Jackson Laboratory, and the Leiden UniversityMedical Center. N.C.G. was supported by a training grant inDevelopmental Biology, HD07065.

Table 3. Summary of phenotypes in deletion mice

+, defect was observed; –, defect was not observed; ND, not determined.aNo corneal opacities were seen, but a determination of colobomas was notmade.

Deletion Facialdefects

Seizures Eye defects Brain defects

Cortex Cerebellum

Hdhdf2J + ND –a + +

Hdhdf6J + + + ND ND

Hdhdf7J + + + + –

Hdhdf4J + + + + +

Hdhdf9J – – – + –

Dpp6df1J ND – ND ND ND


REFERENCES

1. Urban, Z., Helms, C., Fekete, G., Csiszar, K., Bonnet, D., Munnich, A.,Donis-Keller, H. and Boyd, C.D. (1996) 7q11.23 deletions in Williamssyndrome arise as a consequence of unequal meiotic crossover. Am. J.Hum. Genet., 59, 958–962.

2. Dutly, F. and Schinzel, A. (1996) Unequal interchromosomal rearrangementsmay result in elastin gene deletions causing the Williams–Beuren syndrome.Hum. Mol. Genet., 5, 1893–1898.

3. Amos-Landgraf, J.M., Ji, Y., Gottlieb, W., Depinet, T., Wandstrat, A.E.,Cassidy, S.B., Driscoll, D.J., Rogan, P.K., Schwartz, S. and Nicholls, R.D.(1999) Chromosome breakage in the Prader-Willi and Angelman syndromesinvolves recombination between large, transcribed repeats at proximal anddistal breakpoints. Am. J. Hum. Genet., 65, 370–386.

4. Gabriel, J.M., Merchant, M., Ohta, T., Ji, Y., Caldwell, R.G., Ramsey, M.J.,Tucker, J.D., Longnecker, R. and Nicholls, R.D. (1999) A transgene insertioncreating a heritable chromosome deletion mouse model of Prader-Willi andAngelman syndromes. Proc. Natl Acad. Sci. USA, 96, 9258–9263.

5. Tsai, T.F., Jiang, Y.H., Bressler, J., Armstrong, D. and Beaudet, A.L. (1999)Paternal deletion from Snrpn to Ube3a in the mouse causes hypotonia, growthretardation and partial lethality and provides evidence for a gene contributingto Prader–Willi syndrome. Hum. Mol. Genet., 8, 1357–1364.

6. Yang, T., Adamson, T.E., Resnick, J.L., Leff, S., Wevrick, R., Francke,U., Jenkins, N.A., Copeland, N.G. and Brannan, C.I. (1998) A mousemodel for Prader-Willi syndrome imprinting-centre mutations. NatureGenet., 19, 25–31.

7. Lindsay, E.A., Botta, A., Jurecic, V., Carattini-Rivera, S., Cheah, Y.C.,Rosenblatt, H.M., Bradley, A. and Baldini, A. (1999) Congenital heart diseasein mice deficient for the DiGeorge syndrome region. Nature, 401, 379–383.

8. Lurie, I., Lazjuk, G., Ussova, Y., Presman, E. and Gurevich, D. (1980)The Wolf-Hirschhorn syndrome. Clin. Genet., 17, 375–384.

9. Wilson, M., Towner, J., Coffin, G., Ebbin, A., Siris, E. and Brager, P. (1981)Genetic and clinical studies in 13 patients with the Wolf-Hirschhornsyndrome [del(4p)]. Hum. Genet., 59, 297–307.

10. Wright, T.J., Ricke, D.O., Denison, K., Abmayr, S., Cotter, P.D.,Hirschhorn, K., Keinanen, M., McDonald-McGinn, D., Somer, M.,Spinner, N. et al. (1997) A transcript map of the newly defined 165 kbWolf–Hirschhorn syndrome critical region. Hum. Mol. Genet., 6, 317–324.

11. Stec, I., Wright, T.J., van Ommen, G.J., de Boer, P.A., van Haeringen, A.,Moorman, A.F., Altherr, M.R. and den Dunnen, J.T. (1998) WHSC1, a 90kb SET domain-containing gene, expressed in early development andhomologous to a Drosophila dysmorphy gene maps in the Wolf–Hirschhornsyndrome critical region and is fused to IgH in t(4;14) multiple myeloma.Hum. Mol. Genet., 7, 1071–1082.

12. Wright, T.J., Costa, J.L., Naranjo, C., Francis-West, P. and Altherr, M.R.(1999) Comparative analysis of a novel gene from the Wolf-Hirschhorn/Pitt-Rogers-Danks syndrome critical region. Genomics, 59, 203–212.

13. Grosson, C.L., MacDonald, M.E., Duyao, M.P., Ambrose, C.M., Roffler-Tarlov, S. and Gusella, J.F. (1994) Synteny conservation of the Hunting-ton’s disease gene and surrounding loci on mouse Chromosome 5. Mamm.Genome, 5, 424–428.

14. Schimenti, J., Libby, B., Bergstrom, R., Wilson, L., Naf, D., Tarantino, L.,Alavizadeh, A., Lengeling, A. and Bucan, M. (2000) Interdigitated dele-tion complexes on mouse chromosome 5 induced by irradiation of embry-onic stem cells. Genome Res., 10, 1043–1050.

15. You, Y., Bergstrom, R., Klemm, M., Lederman, B., Nelson, H., Ticknor, C.,Jaenisch, R. and Schimenti, J. (1997) Chromosomal deletion complexes inmice by radiation of embryonic stem cells. Nature Genet., 15, 285–288.

16. Battaglia, A., Carey, J.C., Cederholm, P., Viskochil, D.H., Brothman,A.R. and Galasso, C. (1999) Natural history of Wolf-Hirschhorn syn-drome: experience with 15 cases. Pediatrics, 103, 830–836.

17. Wieczorek, D., Krause, M., Majewski, F., Albrecht, B., Horn, D., Riess,O. and Gillessen-Kaesbach, G. (2000) Effect of the size of the deletion andclinical manifestation in Wolf-Hirschhorn syndrome: analysis of 13patients with a de novo deletion. Eur. J. Hum. Genet., 8, 519–526.

18. Lazjuk, G.I., Lurie, I.W., Ostrowskaja, T.I., Kirillova, I.A., Nedzved,M.K., Cherstvoy, E.D. and Silyaeva, N.F. (1980) The Wolf-Hirschhornsyndrome. II. Pathologic anatomy. Clin. Genet., 18, 6–12.

19. Cattanach, B.M., Burtenshaw, M.D., Rasberry, C. and Evans, E.P. (1993)Large deletions and other gross forms of chromosome imbalance compat-ible with viability and fertility in the mouse. Nature Genet., 3, 56–61.

20. Gerard, M., Hernandez, L., Wevrick, R. and Stewart, C.L. (1999) Disrup-tion of the mouse necdin gene results in early post-natal lethality. NatureGenet., 23, 199–202.

21. Zollino, M., Wright, T.J., Di Stefano, C., Tosolini, A., Battaglia, A., Altherr,M.R. and Neri, G. (1999) ‘Tandem’ duplication of 4p16.1p16.3 chromosomeregion associated with 4p16.3pter molecular deletion resulting in Wolf-Hirschhorn syndrome phenotype. Am. J. Med. Genet., 82, 371–375.

22. Zollino, M., Di Stefano, C., Zampino, G., Mastroiacovo, P., Wright, T.J.,Sorge, G., Selicorni, A., Tenconi, R., Zappala, A., Battaglia, A. et al.(2000) Genotype–phenotype correlations and clinical diagnostic criteriain Wolf-Hirschhorn syndrome. Am. J. Med. Genet., 94, 254–261.

23. Gandelman, K.Y., Gibson, L., Meyn, M.S. and Yang-Feng, T.L. (1992)Molecular definition of the smallest region of deletion overlap in theWolf-Hirschhorn syndrome. Am. J. Hum. Genet., 51, 571–578.

24. Estabrooks, L.L., Rao, K.W., Driscoll, D.A., Crandall, B.F., Dean, J.C.,Ikonen, E., Korf, B. and Aylsworth, A.S. (1995) Preliminary phenotypic mapof chromosome 4p16 based on 4p deletions. Am. J. Med. Genet., 57, 581–586.

25. Wright, T.J., Clemens, M., Quarrell, O. and Altherr, M.R. (1998) Wolf-Hirschhorn and Pitt-Rogers-Danks syndromes caused by overlapping 4pdeletions. Am. J. Med. Genet., 75, 345–350.

26. Kant, S.G., Van Haeringen, A., Bakker, E., Stec, I., Donnai, D.,Mollevanger, P., Beverstock, G.C., Lindeman-Kusse, M.C. and VanOmmen, G.J. (1997) Pitt-Rogers-Danks syndrome and Wolf-Hirschhornsyndrome are caused by a deletion in the same region on chromosome 4p16.3. J. Med. Genet., 34, 569–572.

27. Peterson, S.L. (1998) Electroshock. In Peterson, S. and Albertson, T.(eds), Neuropharmacology Methods in Epilepsy Research. CRC Press,Boca Raton, FL, pp. 1–26.

28. Wiegant, J., Galjart, N.J., Raap, A.K. and d’Azzo, A. (1991) The geneencoding human protective protein (PPGB) is on chromosome 20.Genomics, 10, 345–349.

29. Baxendale, S., MacDonald, M.E., Mott, R., Francis, F., Lin, C., Kirby, S.F.,James, M., Zehetner, G., Hummerich, H., Valdes, J. et al. (1993) A cosmidcontig and high resolution restriction map of the 2 megabase region contain-ing the Huntington’s disease gene. Nature Genet., 4, 181–186.

30. Cheng, S.V., Martin, G.R., Nadeau, J.H., Haines, J.L., Bucan, M., Kozak,C.A., MacDonald, M.E., Lockyer, J.L., Ledley, F.D., Woo, S.L. et al.(1989) Synteny on mouse chromosome 5 of homologs for human DNAloci linked to the Huntington disease gene. Genomics, 4, 419–426.

31. Ying, H.C., Hurle, B., Wang, Y., Bohne, B.A., Wuerffel, M.K. and Ornitz,D.M. (1999) High-resolution mapping of tlt, a mouse mutant lacking oto-conia. Mamm. Genome, 10, 544–548.

article mouse models for the wolf–hirschhorn deletion syndrome

Documents