Reduced penetrance of craniofacial anomalies - Human Molecular

Reduced penetrance of craniofacial anomalies as afunction of deletion size and genetic background ina chromosome engineered partial mouse model forSmith–Magenis syndrome

Jiong Yan1, Victoria W. Keener1, Weimin Bi1, Katherina Walz1, Allan Bradley4,

Monica J. Justice1 and James R. Lupski1,2,3,*

1Department of Molecular and Human Genetics and 2Department of Pediatrics, Baylor College of Medicine, One

Baylor Plaza, Houston, TX 77030, USA, 3Texas Children0s Hospital, Houston, TX, USA and 4The Wellcome Trust

Sanger Institute, Hinxton, UK

Received May 25, 2004; Revised September 4, 2004; Accepted September 10, 2004

Smith–Magenis syndrome (SMS) is a multiple congenital anomaly/mental retardation syndrome associatedwith del(17)(p11.2p11.2). The phenotype is variable even in patients with deletions of the same size. RAI1has been recently suggested as a major gene for majority of the SMS phenotypes, but its role in the full spec-trum of the phenotype remains unclear. Df(11)17/1 mice contain a heterozygous deletion in the mouse regionsyntenic to the SMS common deletion, and exhibit craniofacial abnormalities, seizures and marked obesity,partially reproducing the SMS phenotype. To further study the genetic basis for the phenotype, we con-structed three lines of mice with smaller deletions [Df(11)17-1, Df(11)17-2 and Df(11)17-3 ] using retrovirus-mediated chromosome engineering to create nested deletions. Both craniofacial abnormalities and obesityhave been observed, but the penetrance of the craniofacial phenotype was markedly reduced when com-pared with Df(11)17/1 mice. Overt seizures were not observed. Phenotypic variation has been observed inmice with the same deletion size in the same and in different genetic backgrounds, which may reflect the vari-ation documented in the patients. These results indicate that the smaller deletions contain the gene(s), mostlikely Rai1, causing craniofacial abnormalities and obesity. However, genes or regulatory elements in thelarger deletion, which are not located in the smaller deletions, as well as genes located elsewhere, also influ-ence penetrance and expressivity of the phenotype. Our mouse models refined the genomic region importantfor a portion of the SMS phenotype and provided a basis for further molecular analysis of genes associatedwith SMS.

INTRODUCTION

Clinical manifestations of the Smith–Magenis syndrome(SMS) include craniofacial and skeletal anomalies, neurobeha-vioral abnormalities such as sleep disturbance and seizures,ophthalmic anomalies, otolaryngologic anomalies, cardiacand renal anomalies (1). Majority of the patients (75–90%)have an �4 Mb deletion on chromosome 17p11.2 (2–4).Patients with smaller deletions have enabled the refinementof SMS critical region (SMCR), which contains about 20genes (5). The SMS phenotype is variable even in patients

with deletions of the same size (3). Recently, point muta-tions in the RAI1 gene were identified in five patients withcharacteristic SMS features but without fluorescence in situhybridization (FISH)-detectable deletions (6,7). Obviously,RAI1 is an important gene for SMS, but the small number ofpatients and the phenotypic differences among the patientsmake it hard to determine whether it is the only haploinsuffi-cient gene required for the phenotype.

Mouse models have proven to be powerful tools for study-ing contiguous gene syndromes, such as DiGeorge syndrome(DGS) (8–10). Tbx1, a gene responsible for the cardiac

Human Molecular Genetics, Vol. 13, No. 21 # Oxford University Press 2004; all rights reserved

*To whom correspondence should be addressed at: Department of Molecular and Human Genetics, Baylor College of Medicine, Room 604B, OneBaylor Plaza, Houston, TX 77030, USA. Tel: þ1 7137986530; Fax: þ1 7137985073; Email: [email protected]

Human Molecular Genetics, 2004, Vol. 13, No. 21 2613–2624doi:10.1093/hmg/ddh288Advance Access published on September 30, 2004

Dow

nloaded from https://academ

ic.oup.com/hm

g/article/13/21/2613/587478 by guest on 06 February 2022

defects in DGS, was identified through mouse models. The32–34 cM region of mouse chromosome 11 is syntenic to thehuman genomic interval deleted in patients with SMS (5).Gene order and number are highly conserved, especiallyin the region syntenic to the SMS critical interval. Nineteengenes in that region have the same order and orientation inmice and humans. Using chromosome engineering, micehave been created that contain an �2 Mb deletion[Df(11)17 ] from the Csn3 gene to the Zfp179 gene. Thisgenomic interval contains majority of the mouse region synte-nic to the human SMS common deletion (11). The Df(11)17/þmice manifest craniofacial abnormalities, seizures, markedobesity and abnormal circadian rhythm (11,12), which par-tially recapitulates the SMS phenotype.

To further refine the genomic interval and potential gene(s)responsible for the SMS phenotypic features observed inDf(11)17, we have created mice containing smaller deletionsusing a retrovirus-mediated nested deletion strategy (13). Inthis strategy, a loxP site is introduced by gene targeting at aselected locus as an anchoring point. Subsequently, a retro-viral vector is used to introduce another loxP site. After theintroduction of Cre, recombination occurs between twodirectly oriented loxP sites. A series of deletions around theselected locus can be constructed rapidly because of therandom insertion of the retrovirus. Furthermore, the desireddeletion clones can be identified by selectable markers. Usingthis approach, we engineered three deletions [Df(11)17-1,Df(11)17-2 and Df(11)17-3 ] of the mouse region syntenic toSMCR. Df(11)17-1 consists of a 590 kb deletion, whereas theremaining two deletions are 595 kb in size. Heterozygousmice with these deletions exhibit craniofacial anomalies andobesity, whereas overt seizures were not observed. Interest-ingly, the severity and penetrance of the craniofacial phenotypewas reduced when compared with Df(11)17/þ mice. Further-more, the phenotype varies in the same and in differentgenetic backgrounds. These data indicate that genes for thecraniofacial anomalies and obesity are located in the refinedgenomic interval. However, genes or regulatory regions in thelarger deletion that are not located in the smaller deletion, aswell as genes located elsewhere, also influence the penetranceand expressivity of the phenotype.

RESULTS

Generation of small deletions in the mouse regionsyntenic to the human SMCR

Csn3 was chosen as the fixed anchoring point to constructsmall deletions because its human homolog mapped to oneend of the SMCR region (5), and this locus was used as oneend point in the chromosome-engineered SMS mouse model(11) (Fig. 1). Cre induced site-specific recombinationbetween the loxP site anchored in the Csn3 gene by an inser-tional vector (14) and the second loxP site randomly insertedvia the retrovirus (Fig. 2A). The ES cells with deletions wereselected from other types of events by drug selection followedby screening for the markers contained in the insertional andretrovirus vectors (13).

FISH analysis was carried out for 11 selected clones to con-firm the deletion and to estimate the size of the rearrangement.

Four BAC clones (RP23-438J17, 157J1, 278I21 and 40J4) thatmapped inside the mouse region sytenic to SMCR, and oneclone (RP23-480F3) that mapped outside this region, but onthe chromosome 11 as a control probe, were used (Fig. 1)(5, and unpublished data). In majority of the ES cell clones,including clone C12, fluorescence signals were detected by438J17 and 157J1 on only one chromosome 11 homolog,whereas 278I21 and 40J4 detected both chromosome 11homologs (Figs 1 and 2B; and data not shown). Two clones,D10 and H12, seemed to be deleted for part of the regiondetected by 157J1, because cells with either one or twosignals from 157J1 were observed.

Proviral/host junction fragments were detected by genomicSouthern analysis using Xba I digestion, which cuts once in theprovirus, and a probe from 30 Hprt contained in the V15 retro-virus (Fig. 2A and B; and data not shown) (13). The size of thejunction fragments enabled us to determine whether the dele-tion clones had the same or different insertion sites. The sizeof the Xba I fragment represents the distance of the virusfrom the nearest genomic Xba I site. Junction fragmentswere detected in all the deletion clones examined. ClonesC12 and H12 had the same size of the junction fragment,�2.8 kb, and clone D10 had a fragment �3.3 kb in size.

Clones C12, D10 and H12 were injected for germline trans-mission, because they had deletions covering half of themouse region syntenic to SMCR and it seemed that they hadat least two different-sized deletions. The precise insertionsites for these clones were analyzed using inverse PCR(IPCR) and virus insertion site amplification PCR (VISA-PCR) (Fig. 2). BLASTN analysis was performed for theDNA sequences from the PCR products against the assembledgenomic sequence in the mouse syntenic region (5). Theretrovirus insertion point for clone D10 was 590 kb telomericto Csn3 in the third intron of the annotated gene4933439F18Rik (NCBI: AK017120) from BAC RP23-181C17 (5), and C12 and H12 had the same insertion site,which was 595 kb away from Csn3 in the fifth intron of thesame gene 4933439F18Rik (Fig. 1). These results were con-sistent with both FISH and proviral junction fragment ana-lyses. Primers were synthesized to amplify from the V15 tothe flanking genomic regions to further confirm the insertionpoints (Fig. 2A and C). Eleven genes, including Rai1, weredeleted in each of the engineered smaller deletions (Fig. 1).

Chimeras from each ES cell clone were mated to C57BL/6Tyrc-Brd (B6) mice, and the F1 progeny were backcrossed toB6 mice to obtain the N2 generation. Alleles from ES cellclones D10, C12 and H12 were referred to as Df(11)17-1,Df(11)17-2 and Df(11)17-3, respectively. The deletion micewere identified with two pairs of PCR primers, and confirmedby FISH carried out on the fibroblast cells from the tails of ran-domly picked deletion mice (Fig. 2A and C; and data not shown).

The mice with smaller deletions are overweight

Marked obesity was observed in the Df(11)17/þ mice at F1and N2 generations (11), and some SMS patients are over-weight (6). Weights were measured for both male andfemale mice with either the 590 kb [Df(11)17-1 ] or 595 kb[both Df(11)17-2 and Df(11)17-3 ] deletions at F1 and N2mice from 3 to 30 weeks of age (Fig. 3). Overall, mice with

2614 Human Molecular Genetics, 2004, Vol. 13, No. 21

Dow


ic.oup.com/hm


deletions were significantly heavier than the wild-type litter-mates in all the measurements after about 10 weeks of age.No significant difference was observed between the F1 andN2 generations. There was no difference between the micewith either the 590 or 595 kb deletion, which might be anti-cipated since the sizes of these deletions were similar and theinsertion points were in the same gene. In fact, we did notobserve any significant difference between mice with either ofthese two deletion sizes in any of the phenotypic analyses.

Craniofacial abnormalities

Craniofacial features of SMS include midface hypoplasia, bra-chycephaly, broad nasal bridge, abnormal ears, down-turnedupper lip, frontal bossing and prognathia (1). Df(11)17/þmice had obviously shorter, concave and/or curved snouts anda broader distance between the eyes (hypertelorism) whencompared visually with the wild-type littermates (11). Thisphenotype was observed in .90% of the deletion mice inthe F1 generation, and 70–80% of the N2 generation, but nosuch craniofacial anomalies were observed in the pure 129background by the age of 5 months (N ¼ 30). The distancebetween the eyes and the length of the snouts were quite vari-able among the mice with smaller deletions [Df(11)17-1,Df(11)17-2 and Df(11)17-3 ] and wild-type littermates, whichsometimes made it hard to determine whether these distanceswere abnormal in the deletion mice. However, the snouts insome deletion mice were flatter or slightly concave, and most

of them were curved to the left or right. These mice also hadmore discernable short snouts and hypertelorism (Fig. 4A–C).The percentage of the small deletion mice with discernableabnormalities was much less than observed in the Df(11)17/þmice and was different at both F1 (4–10%) and N2 (29–37%) generations (Table 1).

In order to more objectively analyze the craniofacial abnorm-alities than by visual inspection, skeletal preparations were per-formed on several litters of F1 male mice at age 7–11 months(N ¼ 8 for the wild-type; N ¼ 7 for the deletion mice) and N2male mice at age 11–12 months (N ¼ 16 for the wild-type;N ¼ 18 for the deletion mice) (Fig. 4D–L). The observationsof skulls and live mice were consistent. Short, concave andcurved nasal parts were observed in two out of seven skullsfrom F1 deletion mice and in nine out of 18 skulls from N2 dele-tion mice (Fig. 4J–L), and these deformed skulls turned out tobe from the mice with discernable abnormalities observed inlive mice. The skull sutures of these mice appeared normal,which indicates that the curved nasal bone resulted from asym-metrical bone growth.

Several landmarks were selected to take objective measure-ments (Fig. 4G, H and M–O). In N2 mice with smaller dele-tions, the length of the nasal bones (distances from a to b andc to b) and the height of the nasal bones at the tip (distancefrom h to i) were statistically shorter than the wild-type litter-mates, whereas the distance between the eyes (from a to c)was broader. Other measurements (distances from b to d, d toe and f to g) had no significant difference. In N2 mice with

Figure 1. Human SMS deletion region and the mouse syntenic region. The human deletion region is on chromosome 17p11.2 (HSA17p) and the mouse syntenicregion is on chromosome 11 (MMU11). The SMS common deletion and the critical interval (SMCR) are indicated on top with horizontal bold lines. Shownbelow are BACs used for FISH analysis, and the deletion regions in Df(11)17 (from Csn3 to Zfp179 ), Df(11)17-1, Df(11)17-2 and Df(11)17-3 (bold dashedlines). Note the mouse region syntenic to the SMCR is highly conserved.

Human Molecular Genetics, 2004, Vol. 13, No. 21 2615

Dow


ic.oup.com/hm


Df(11)17, all the distances measured are significantly differentfrom that seen in the wild-type littermates (11), which indicatesthat overall the craniofacial phenotype in the mice with thelarger deletion was more severe than the mice with smaller del-etions, although the severity of each mouse was different.

The distances varied in both the N2 mice with smaller del-etions and the wild-type littermates, and the ranges overlapped

for these two groups (Fig. 4N and O). Since the skulls withobservable abnormalities had more extreme measurementsthan the rest of the skulls, we did a separate analysis withthese skulls. The length and height of nasal bones and thedistances between the eyes were significantly different fromthat seen in the wild-type mice, whereas the rest of the mea-surements were not (data not shown). It seems that the mice

Figure 2. Generation of smaller deletions. (A) Strategy for generating smaller deletions. Csn3 was used as the fixed anchoring point and was targeted with avector containing a loxP site, 50 Hprt and the neomycin resistance gene (Neo ). V15 retrovirus containing another loxP site, 30 Hprt and the puromycin resistancegene (Puro ) randomly integrated into the genome. The insertion of V15 in two representative places is depicted. The genomic regions are represented as a, b andc. Cre induced recombination between two loxP sites, which resulted in the deletion of the intervening genomic regions and the Neo and Puro genes and thereconstruction of the Hprt gene utilized for selection. The deletion ES cell clones were identified as neomycin and puromycin sensitive (Neos and Puros) andHAT resistant (HATr) clones. The restriction enzyme site, XbaI, for the proviral junction fragment analysis is indicated, and the probe is shown as a solid hori-zontal bar. All the primers used are depicted as arrows. The primers without label and the primer 3031 were used for IPCR. (B) Characterization of deletions.Shown on the left are representative FISH results for one ES cell clone with the probes 438J17 and 278I21 (red signals). The green signals are from the controlprobe 480F3. The result of Southern analysis of proviral junction fragments is on the right. (C) Confirmation of deletions with PCR. PCR results with primers 50

HPRT 2TyF and 50 HPRT 2TyR, which amplified from Csn3 genomic region into the targeting vector, and 3033–2 and C12IN-R, which amplified from V15 intothe flanking genomic region, are shown. C12, H12, D10: genomic DNA from ES cell clones; C12F1: genomic DNAs from Df(11)17-2/þ F1 mice; WT:wild-type; minus: negative control.


Dow


ic.oup.com/hm


with a larger deletion had an even more severe phenotype thanthis group of mice. From both the observed and the objectivelymeasured distances, the height and the length of the nasal boneshad the most obvious change in the deletion mice. No obviousabnormalities were observed in other skeletal structures.

For the F1 mice with smaller deletions, overall no measure-ment was significantly different from that of the wild-typelittermates (Fig. 4M), but two skulls with discernable abnorm-alities had significantly underdeveloped nasal bones asobserved in N2 generation mice (data not shown).

Three-dimensional craniofacial scan

A protocol for soft-tissue three-dimensional surface scan analy-sis on mice has been recently created and tested (V.W. Keeneret al., manuscript in preparation). We implemented this methodin deficiency mice with the larger [Df(11)17/þ] or smallerdeletions [Df(11)17-1/þ, Df(11)17-2/þ and Df(11)17-3/þ]and their wild-type littermates from both F1 and N2 males at5–9 months of age (Fig. 5A–C). The three-dimensionalimages confirmed the observations in live mice.

Several landmarks were selected to obtain objectivemeasurements (Fig. 5A and C). In F1 mice with smaller dele-tions, the en–en and n–sn distances were significantly differ-ent from that seen in the wild-type (N ¼ 13 for the deletionmice; N ¼ 15 for the wild-type). In N2 (N ¼ 13 for the dele-tion mice; N ¼ 8 for the wild-type), only en–en was signifi-cantly larger. The n–sn was shorter but not statisticallysignificant. Three distances (en–en, n–sn and sn–t) of N2Df(11)17/þ mice were statistically different from the wild-type mice (N ¼ 5 for the deletion mice; N ¼ 6 for the wild-type). The range of distances of the smaller deletion micelargely overlapped with that of the wild-type (Fig. 6A), butthe range of Df(11)17/þ mice did not (Fig. 6B), which

indicates that they had a more severe phenotype than themice with smaller deletions.

A cluster analysis enables a pair-wise comparison of one setof pooled distance data to determine which samples are stat-istically more similar to each other (V.W. Keener et al., manu-script in preparation). The distance data were obtained byEuclidean distance matrix analysis (EDMA), which performeddistance measurements on every possible combination of pairsof selected landmarks. The F1 and N2 mice with smaller del-etions largely mingled with the wild-type mice, indicatingthey were not distinctly different regarding the distances com-pared (Fig. 6C). The cluster of the Df(11)17/þ mice appearedmore robust (Fig. 6D). Three out of the five mice were discri-minated after only two iterations, which indicates they weremore distinct than the mice with smaller deletions.

Overall, we observed similar craniofacial abnormalities inthe mice with smaller deletions [Df(11)17-1, Df(11)17-2 andDf(11)17-3 ] as in Df(11)17/þ mice, including broader dis-tance between eyes, shorter nasal bridge and abnormal nasalbone shapes, but the penetrance and severity of the phenotypewere less. In addition, the phenotype varied in mice with eithersmall or large deletions in the same genetic background (F1 orN2) or in the different genetic backgrounds (F1, N2, 129).

Other phenotypic analyses

As was observed for the Df(11)17/þ mice, the mice with het-erozygous smaller deletions were viable, and gross necropsyrevealed no obvious organ abnormalities. Overt seizureswere observed in 20% of the Df(11)17/þ mice (11),whereas seizures were not observed in any mice (N ¼ 152)with the smaller deletions. Another phenotype observed inthe Df(11)17/þ mice was reduced fertility in males whenthey were maintained under less favorable environmental con-ditions (11). In a pathogen-free and quiet environment, thisphenotype was not present (unpublished data). No reducedfertility was observed in the Df(11)17-1/þ, Df(11)17-2/þ orDf(11)17-3/þ mice with the smaller deletions.

DISCUSSION

Df(11)17/þ mice, which harbor an �2 Mb heterozygous dele-tion in the mouse region syntenic to the human SMS commondeletion, reproduced some of the SMS features including cra-niofacial abnormalities, abnormal circadian rhythm, seizuresand obesity (11,12). Most of the abnormalities, including thecraniofacial features and obesity, have been proved to resultfrom the reduced dosage of gene(s) inside the deletion (11).Using a retrovirus to construct nested deletions, we generatedmice containing either a 590 kb [Df(11)17-1/þ] or a 595 kb[Df(11)17-2/þ and Df(11)17-3/þ] deletion covering half ofthe mouse region syntenic to the SMCR that is locatedwithin the Df(11)17 deletion. Both craniofacial abnormalitiesand obesity have been observed, which indicates that gene(s)responsible for these features are located in this 590 kbregion. There are 11 annotated genes that map inthis region, including Rai1 (Fig. 1) (5).

Recently, mutations have been found in the RAI1 genein five patients with characteristic SMS features including

Figure 3. Weight analysis of smaller deletion mice. Weight curves for maleand female wild-type and Df(11)17-1/þ F1 mice are shown. Values representmean + SEM. The asterisk indicates the age at which the deletion mice beganto be significantly heavier than the wild-type mice (P-value ,0.05). N ¼ 22for wild-type male; N ¼ 17 for wild-type female; N ¼ 14 for deletion male;N ¼ 21 for deletion female.


Dow


ic.oup.com/hm



Dow


ic.oup.com/hm


craniofacial anomalies and obesity, but without FISH-detectable deletions (6,7). Rai1 was first cloned in themouse embryonic carcinoma cell line P19 as a retinoic acid-induced gene (15). It is expressed in many tissues with highexpression in brain. Retinoic acid was found to be able toinduce asymmetric craniofacial growth in the mouse fetus(16). It is highly likely that Rai1 caused both the craniofacialanomalies and the obesity observed in mouse models withengineered chromosome 11 deletions.

The phenotype is quite variable in SMS patients withcommon deletions (3), and the number of patients with RAI1mutations is too small to make definitive statementsabout the nature of the RAI1 mutant phenotype in humans,which makes it difficult to determine whether RAI1 is theonly gene responsible for the complete SMS phenotype. Thecontribution of other genes within the smaller deletion tothe craniofacial and obesity phenotype is still possible. Pemt(17), Srebp1 (18) and Csn3 (14), the anchoring point for thesmaller deletions, have been inactivated by insertional muta-genesis. Heterozygous mice show no apparent phenotype,which makes these genes less likely to contribute to thephenotypes observed in mice heterozygous for the engineereddeletions. Rasd1 encodes a G protein (19). Its expression isstrongly and rapidly induced by dexamethasone, whichsuggests the possibility of a role in glucocorticoid action andweight control. The remaining genes are each expressed inmultiple tissues (5). Targeted disruption of these genes orBAC complementation analysis in the deletion mice willhelp unravel their potential contribution to the phenotypes.

The fraction of animals with obvious craniofacial abnorma-lities was less in the mice with the smaller deletions than in theDf(11)17/þ mice, and the phenotype was less severe. Sincethe mice with smaller deletions and Df(11)17 were analyzedin the same genetic background and the same externalenvironment, these differences most likely result from the

difference in the deletion size. Deletion size may influencethe phenotypes through a position effect by removal or reten-tion of the control elements for the phenotype-causing gene(s),or by juxtaposing flanking gene(s) to a different genomicenvironment (20). In either case, the phenotypic differencemay result from an altered expression level of a phenotype-causing gene such as Rai1. However, such a hypotheticalchange in Rai1 expression might occur only in selectedtissues during a specific developmental time interval,perhaps via a craniofacial skeletal tissue specific enhancer.It will be interesting to determine whether retinoic acid sup-plementation during pregnancy or early life will reduce thepenetrance of the Df(11)17/þ craniofacial anomalies.

Nevertheless, it is distinctly possible that other gene(s) inthe Df(11)17 but outside the small deletion contribute to thephenotypic differences. Since there is no difference of thecraniofacial abnormalities observed in patients with commondeletion or SMCR deletion, genes in the SMCR are morelikely to contribute (3). There are nine genes in that region(5). Myo15 can be excluded since individuals carrying hetero-zygous mutations in this gene show no features of SMS (21).Fliih and Top3a have been knocked out, and the heterozygousmice show no phenotype (22,23). The potential contribution ofthe remaining six genes needs to be further investigated.

About 20% of the Df(11)17/þ mice show overt seizures (11),whereas this was not observed in any of the mice with smallerdeletions. One out of three patients with heterozygous trun-cating mutations in RAI1 has seizures; this patient also hasan Arnold-Chiari malformation, which is likely the cause ofthe seizures instead of the RAI1 mutation (6). Seizures canhave numerous etiologies including structure abnormalities ofthe brain that result from either trauma or developmentalabnormalities, and mutations in ion-channels (24). In theregion outside the small deletion and inside the Df(11)17, anygene involved in brain development or ion-channels couldpotentially cause seizures. Seizures were not reported in indi-viduals with heterozygous MYO15 mutations (21) and in Fliihand Top3a heterozygous mice (22,23). Llglh is highly expressedin brain and is proposed to be important for brain developmentin mice (25). Homozygous mutation in its Drosophila homologcauses neoplastic transformation of neuroblasts in imaginaldiscs and the presumptive adult optic centers of the larvalbrain. Kcnj12 encodes for a subunit of an inward rectifierpotassium channel (26), which is located outside the humanSMS deletion region (5). Although it could be a gene that con-tributes to the seizures observed in Df(11)17/þ mice, thephenotype present in mice seems quite similar to that observedin humans, suggesting that a gene other than Kcnj12 is respon-sible for the seizures (11). It is also possible that several genes

Table 1. Penetrance of craniofacial abnormalities in deletion mice

Strain No. ofdeletion mice

No. of micewith phenotype

%

F1 generationDf(11)17-1/þ 49 5 10.2Df(11)17-2/þ 21 2 9.5Df(11)17-3/þ 28 1 4Df(11)17/þ .50 .90

N2 generationDf(11)17-1/þ 30 11 36.7Df(11)17-2/þ 24 7 29Df(11)17/þ .50 70–80

Figure 4. Craniofacial abnormalities. Images of mouse heads and the side and top views of skeleton preparations of deletion mice (B, C, G–L) and their wild-type littermates (A, D–F) are presented. The obvious craniofacial abnormalities were observed only in some of the live deletion mice (B, C) and in some of theskeleton preparations (G–L). Note the shorter and flatter snouts in the live deletion mouse (C), and the shorter, concave and curved nasal bone in the skeletonpreparation (arrow heads in J–L). The greatly underdeveloped nasal bone resulted in a concave shape at the bone and cartilage junction (arrows in J, L). Statisticalanalysis results of some measurements are shown at the bottom. (M) The measurement results for both F1 (DEL N ¼ 7, WT N ¼ 8) and N2 (DEL N ¼ 18, WTN ¼ 16) generations. The mean + SEM values are presented. Landmarks used for the measurements are labeled (G, H): a and c, anterior notch on frontal processsituated laterally in relation to the infraorbital fissure; b, nasale; d, intersection of parietal and interparietal bones; e, intersection of interparietal and occipitalbones at the midline; f, bregma; g, intersection of zygoma with zygomatic process of temporal, superior aspect; h, nasale; i, anterior-most point at intersectionof premaxillae and nasal bones (36). DEL, deletion; WT, wild-type; �P , 0.05. (N) Scatter plot of the hi distance for N2 mice. Each mark represents one mouse.The mice with discernable abnormalities are in black. Note the greatly shorter distances in the mice with discernable abnormalities. (O) Scatter plot of the bcdistance (x-axis) and the ac distance (y-axis) for N2 mice. Note more severe changes in mice with discernable abnormalities (black dots).


Dow


ic.oup.com/hm


Figure 5. Three-dimensional craniofacial scans. Top and side views of the three-dimensional scan of mice with smaller deletions (A) and with Df(11)17 deletion(B). The images of one wild-type mouse (right), one deletion mouse without discernable abnormalities (middle) and one deletion mouse with discernableabnormalities (left) are shown for smaller deletions, and one wild-type (right) and one deletion mouse (left) are shown for Df(11)17. (C) Statistical analysisof the distances between some landmarks of mice with smaller deletions (For F1: DEL N ¼ 13, WT N ¼ 15; For N2: DEL N ¼ 13, WT N ¼ 8) andDf(11)17 deletion (DEL N ¼ 5; WT N ¼ 6). The mean + SEM values are presented. Landmarks are as follows: tragion (t), endocanthion (en), exocanthion(ex), vertex (v), subnasale (sn), nasion (n), gnathion (gn) and tip of the ear (sa). DEL, deletion mice; WT, wild-type; �P , 0.05.


Dow


ic.oup.com/hm


together cause the seizure phenotype or a position effect (20)may be responsible for seizures in the deletion mice.

Phenotype modification by modifier genes has been increa-singly recognized in both humans and mice (27). Mice havebeen a powerful tool for dissecting such genes, whichcan facilitate the unraveling of the biological pathways. Thephenotypes of SMS are variable even in patients with thesame deletion size (3). In both the Df(11)17/þ mice andmice with smaller deletions, we have observed significantdifferences in the craniofacial abnormalities in differentgenetic backgrounds, which indicates there are potential modi-fier gene(s). Congenic mice will be created to further study thegenetic modification. The identification of such gene(s) willprovide further insights into craniofacial development. Inaddition, the craniofacial abnormalities were variable in F1mice with the same size deletion. F1 mice have 50% of 129and 50% of the B6 backgrounds. Both 129 and B6 are

inbred strains that are essentially homozygous at all geneticloci except for occasional spontaneous mutations (28). Thedifference in F1 mice may be potentially explained byenvironmental or statistical factors.

In the craniofacial analysis, we used a newly developedsoft-tissue three-dimensional scan method in addition to theskeleton analysis (V.W. Keener et al., manuscript in pre-paration). It provides a quick way to quantitatively analyzeanimal models without sacrificing them and a better wayto compare to the analysis in humans, since the latter has tobe performed in vivo. Limitations of this method are that theexact position of some landmarks can be difficult to determinebecause of the presence of fur and a limited number of land-marks can be used. Because of this, the underdevelopednasal bone was not identified in our three-dimensional scananalysis. This problem can probably be solved by performingsurface analysis, which can use many more landmarks (29).

Figure 6. Scatter plots of n–sn and en–en distances (A, B) and cluster analysis results of three-dimensional craniofacial scans for F1 and N2 mice with smallerdeletions (C) and N2 mice with Df(11)17 (D). Black dots are for the deletion mice with discernable abnormalities.


Dow


ic.oup.com/hm


The analysis of the craniofacial changes of our mouse modelsis preliminary. Detailed analysis will need to be performed tosee how the changes in mouse models correspond to the cra-niofacial phenotype in SMS patients.

Our mice with engineered deletions have refined the regionfor the craniofacial abnormalities and obesity to a 590 kbinterval. Although RAI1 mutations cause SMS in humans,our results indicate that RAI1 may not be the only gene thatcontributes to the SMS phenotype and that there are potentialmodifier genes located outside the SMS region. Our mice withsmaller deletions provide an important reagent to further studythe etiology of SMS phenotypic features.

MATERIALS AND METHODS

Generation of deletions in embryonic stem (ES) cells

AB2.2 ES cells derived from Hprt-deficient 129/SvEvBrd (129)mice were used for all manipulations (30). ES cell culture, drugselection and electroporation were performed as described(30,31). Csn3 gene targeting, Southern blot confirmation oftargeting and germline transmission were the same as pre-viously reported (14). Instead of the pWY1-50 vector previouslyused, pWY2-50 was used, which is the same as pWY1-50 exceptthat the genomic fragment is in the opposite orientation. Thesevectors are insertional vectors containing a loxP site, half of the50 Hprt gene and the neomycin resistance gene (Neo ) (32). TheCsn3 gene-targeted ES cells were infected with V15 retrovirus,which introduces a loxP site, 30 Hprt and the puromycin resist-ance gene into the genome randomly, and selected in puromy-cin for viral infection events (13). Infected cells were pooled,transfected with a Cre expression plasmid and selected inHAT (hypoxanthine, aminopterine and thymidine) medium(33). HAT resistant clones were selected and screened withmedium containing G418 and puromycin. The desired deletionclones are HAT resistant and G418 and puromycin sensitive.The predicted deletions were confirmed and further character-ized by FISH, Southern analysis of proviral junction fragmentand viral insertion point analysis.

Mouse strains and genotyping

Three deletion ES cell clones (C12, D10, H12) were injectedinto C57BL/6 Tyrc-Brd blastocysts. Chimeras were mated toC57BL/6 Tyrc-Brd (B6) mice, and F1 progeny were back-crossed to B6 mice to obtain the N2 generation. All thestrains were maintained by backcrossing to B6 wild-typemice. Animals were treated in compliance with relevantanimal welfare policies.

Deletion mice were identified using two pairs of primers.One pair was made to amplify from the flanking Csn3genomic region into the pWY2-50 vector, which resulted ina 600 bp product [50 HPRT 2Ty (tyrosinase end) F 50-CTGGGA GAA AAC ATA TTT TGA GAG A-30; 50 HPRT 2Ty(tyrosinase end) R 50-TTC CTG TTT GGG GTA GAA TGTACT-30]. Another PCR primer pair amplified from the V15retrovirus into the flanking genomic region. This is describedin the viral insertion point analysis section. Fibroblastcells were cultured from tails of mice selected randomly andsubjected to FISH analysis to confirm the deletion.

Fluorescence in situ hybridization (FISH)

BACs from the mouse RPCI-23 library (BACPAC Resources)were used as probes. FISH was carried out on ES cell or fibro-blast cell spreads using a standard protocol (34). The cells werevisualized under a Zeiss Axioplan2 fluorescence microscope.

Southern analysis of proviral junction fragments

Genomic DNA was digested with Xba I. A 753 bp Xmn I andXho I fragment from pPGKhprtmini5 (13) was gel purified(Qiagen, gel extraction kit) and used as the probe. The deletionclones were expected to have one proviral junction fragment.

Viral insertion point analysis

IPCR and VISA-PCR were performed to obtain the nucleotidesequence at the virus insertion sites.

For IPCR, genomic DNA (10 mg) was digested with Xba Iand electrophoresed. The genomic fragment correspondingto the size of the proviral junction fragment was gel extractedand half of it was circularized in 500 ml volume at 168C over-night. The ligated DNA was purified using a PCR purificationkit (Qiagen), and half of it was used for three rounds of nestedPCR (13).

For VISA-PCR, two rounds of PCR were performed (35). Inthe first round, one primer from the V15 retrovirus (3034-150-AGT GTT ACG TTG AGA AAG AA-30) and a degenera-tive primer that consisted of an M13 forward primer, 8 Nsand a Hind III site (50-GGG TTT TCC CAG TCA CGACNN NNN NNN AAG CTT-30), were used. For PCR the25 ml reaction mixture consisted of 100 ng of genomicDNA, 1� PCR buffer II (Roche), 1.5 mM MgCl2, 0.2 mM

dNTPs, 2.5 pmol 3034-1, 25 pmol degenerative primer, 1 M

betaine and 0.25 U Taq. The PCR conditions were 948C3 min, 948C 30 s, 508C 2 min, 728C 2 min, 29 more times tostep 2 and 728C 7 min. In the second round of PCR, primer3034-2 (50-TCA GGA ACA GAT GGA ACA GC-30), whichis inside the primer 3034-1, and the M13 forward primer(50-GGG TTT TCC CAG TCA CGA C-30) were used. The reac-tion mix was similar to the first round PCR, except that 1 mlPCR product from the first round of PCR was used, insteadof genomic DNA, and 25 pmol 3034-2 and M13 were used.The program for PCR reaction was also the same, except thatthe annealing temperature was changed to 558C.

Both the IPCR and VISA-PCR products were gel purifiedand the DNA sequence was determined using an ABI 377and standard methods. The primers for sequencing were3031 50-GAC GCG CCG CTG TAA AGT GTT ACG TTGAG-30 for ES cell clone C12; HPRT END 50-GCT GAACAA GTA CCA AAC ATG TAA A-30 for clone D10.

Sequences obtained were analyzed to find the V15 sequenceand restriction enzyme sites using Sequencer 4.1 software(Gene Codes). After dispensing with the V15 sequence, the restof the sequences were compared to the assembled genomicsequence in the mouse region syntenic to the SMCR usingBLASTN analysis (NCBI).

Primers were synthesized to confirm the viral insertion sites:primers 3033-2 50-TCC CGA TCA AGG TCA GGA ACA-30

and C12IN-R 50-TCC CTT TTC AAT GTG AAA CCA-30


Dow


ic.oup.com/hm


producing a 490 bp fragment were for clone C12 and H12,which have the same insertion point; primers 3033-2 andD10IN-R 50-CTG AAC TGC AGC AGA GAT GC-30 werefor clone D10, which yield a 570 bp product. These primerswere also used for mice genotyping.

Three-dimensional craniofacial scans

Mice were scanned on a Cyberware Desktop 3D Scanner, usingthe included Cyberware CyDir/CyScan software. First, micewere anesthetized using Avertin, and their fur was paintedwith a cornstarch and water mixture to create a white reflectivesurface for scanning. Once the mixture was dry, the mice wereplaced in a specifically machined chair that allows the head tobe positioned upright for the most complete scanning. A 3608scanning required �10 s once initiated. Once scanned, thefiles were edited and converted to .3ds extensions using theCyberware Mtool software. The .3ds files were opened onfree VIScam Solid Viewer software made by Marcam Engi-neering, and facial landmark points were marked and theirthree-dimensional coordinates were recorded. EDMA and sub-sequent statistical analyses were performed on the coordinatesusing downloadable PAST version 1.15 statistical analysis soft-ware (V.W. Keener et al., manuscript in preparation).

Skeleton preparations of mice

Mice were sacrificed with inhalation of isoflurane, skinned,eviscerated and fixed in 95% ethanol (EtOH). After the stainingof cartilage with 0.05% alcian blue 8GX solution, mice wererinsed in 95% EtOH and transferred to 2% KOH to denudesoft tissues. Alizarin red (0.015% in 1% KOH) was then usedto stain the skeletons. Images were obtained with Image ProPlus software (Media Cybernetics), and analyzed in AdobePhotoshop version 5.5 (Adobe systems, Incorporated).

Statistical analysis

Statistical analysis was performed using Student’s t test func-tion in Microsoft Excel (Microsoft office 2001, Redford, WA,USA). Values are presented as mean + SEM. It was con-sidered statistically significant when the P-value was ,0.05.

ACKNOWLEDGEMENTS

We thank Dr Hong Su for providing the V15 retrovirus. Thiswork was supported in part by grants from the National CancerInstitute (PO1 CA75719) to A.B. and National Institute forDental and Craniofacial Research (RO1 DEO15210) to J.R.L.

REFERENCES

1. Chen, K.-S., Potocki, L. and Lupski, J.R. (1996) The Smith–Magenissyndrome [del(17)p11.2]: clinical review and molecular advances. Ment.Retard. Dev. Disabil. Res. Rev., 2, 122–129.

2. Chen, K.-S., Manian, P., Koeuth, T., Potocki, L., Zhao, Q., Chinault, A.C.,Lee, C.C. and Lupski, J.R. (1997) Homologous recombination of aflanking repeat gene cluster is a mechanism for a common contiguousgene deletion syndrome. Nat. Genet., 17, 154–163.

3. Potocki, L., Shaw, C.J., Stankiewicz, P. and Lupski, J.R. (2003) Variabilityin clinical phenotype despite common chromosomal deletion in Smith–Magenis syndrome [del(17)(p11.2p11.2)]. Gene Med., 5, 430–434.

4. Vlangos, C.N., Yim, D.K. and Elsea, S.H. (2003) Refinement of theSmith–Magenis syndrome critical region to approximately 950 kb andassessment of 17p11.2 deletions. Are all deletions created equally? Mol.

Genet. Metab., 79, 134–141.

5. Bi, W., Yan, J., Stankiewicz, P., Park, S.-S., Walz, K., Boerkoel, C.F,Potocki, L., Shaffer, L.G., Devriendt, K., Nowaczyk, M.J.M., Inoue, K.and Lupski, J.R. (2002) Genes in a refined Smith–Magenis syndromecritical deletion interval on chromosome 17p11.2 and the syntenic regionof the mouse. Genome Res., 12, 713–728.

6. Slager, R.E., Newton, T.L., Vlangos, C.N., Finucane, B. and Elsea, S.H.(2003) Mutations in RAI1 associated with Smith–Magenis syndrome. Nat.Genet., 33, 466–468.

7. Bi, W., Saifi, G.M., Shaw, C.J., Walz, K., Fonseca, P., Wilson, M.,Potocki, L. and Lupski, J.R. (2004) Mutations of RAI1, a PHD-containingprotein, in nondeletion patients with Smith–Magenis syndrome. Hum.Genet., doi:10.1007/S00439-004-1187-6.

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

9. Lindsay, E.A., Vitelli, F., Su, H., Morishima, M., Huynh, T., Pramparo,T., Jurecic, V., Ogunrinu, G., Sutherland, H.F., Scambler, P.J., Bradley, A.and Baldini, A. (2001) Tbx1 haploinsufficieny in the DiGeorge syndromeregion causes aortic arch defects in mice. Nature, 410, 97–101.

10. Merscher, S., Funke, B., Epstein, J.A., Heyer, J., Puech, A., Lu, M.M.,Xavier, R.J., Demay, M.B., Russell, R.G., Factor, S. et al. (2001) TBX1 isresponsible for cardiovascular defects in velo-cardio-facial/DiGeorgesyndrome. Cell, 104, 619–629.

11. Walz, K., Caratini-Rivera, S., Bi, W., Fonseca, P., Mansouri, D.L., Lynch,J., Vogel, H., Noebels, J.L., Bradley, A. and Lupski, J.R. (2003) Modelingdel(17)(p11.2p11.2) and dup(17)(p11.2p11.2) contiguous gene syndromesby chromosome engineering in mice:phenotypic consequences of genedosage imbalance. Mol. Cell. Biol., 23, 3646–3655.

12. Walz, K., Spence, C., Kaasik, K., Lee, C.C., Lupski, J.R. andPaylor, R. (2004) Behavioral characterization of mouse models forSmith–Magenis syndrome and dup(17)(p11.2p11.2). Hum. Mol. Genet.,13, 367–378.

13. Su, H., Wang, X. and Bradley, A. (2000) Nested chromosomal deletionsinduced with retroviral vectors in mice. Nat. Genet., 24, 92–95.

14. Yan, J., Walz, K., Nakamura, H., Carattini-Rivera, S., Zhao, Q.,Vogel, H., Wei, N., Justice, M.J., Bradley, A. and Lupski, J.R. (2003)COP9 signalosome subunit 3 is essential for maintenance of cellproliferation in the mouse embryonic epiblast. Mol. Cell. Biol., 23,6798–6808.

15. Imai, Y., Suzuki, Y., Matsui, T., Tohyama, M., Wanaka, A. and Takagi, T.(1995) Cloning of a retinoic acid-induced gene, GT1, in the embryonalcarcinoma cell line P19: neuron-specific expression in the mouse brain.Mol. Brain Res., 31, 1–9.

16. Padmanabhan, R. and Ahmed, I. (1997) Retinoic acid-induced asymmetriccraniofacial growth and cleft palate in the TO mouse fetus. Reprod.Toxicol., 11, 843–860.

17. Walkey, C.J., Donohue, L.R., Bronson, R., Agellon, L.B. andVance, D.E. (1997) Disruption of the murine gene encodingphosphatidylethanolamine N-methyltransferase. Proc. Natl Acad. Sci.USA, 94, 12880–12885.

18. Shimano, H., Shimomura, I., Hammer, R.E., Herz, J., Goldstein, J.L.,Brown, M.S. and Horton, J.D. (1997) Elevated levels of SREBP-2 andcholesterol synthesis in livers of mice homozygous for a targeteddisruption of the SREBP-1 gene. J. Clin. Invest., 100, 2115–2124.

19. Kemppainen, R.J. and Behrend, E.N. (1998) Dexamethasone rapidlyinduces a novel ras superfamily member-related gene in AtT-20 cells.J. Biol. Chem., 273, 3129–3131.

20. Kleinjan, D.-J. and van Heyningen, V. (1998) Position effect in humangenetic disease. Hum. Mol. Genet., 7, 1611–1618.

21. Liburd, N., Ghosh, M., Riazuddin, S., Naz, S., Khan, S., Ahmed, Z.,Riazuddin, S., Liang, Y., Menon, P.S.N. et al. (2001) Novel mutations ofMYO15A associated with profound deafness in consanguineous familiesand moderately severe hearing loss in a patient with Smith–Magenissyndrome. Hum. Genet., 109, 535–541.


Dow


ic.oup.com/hm


22. Li, W. and Wang, J.C. (1998) Mammalian DNA topoisomerase IIIa isessential in early embryogenesis. Proc. Natl Acad. Sci. USA, 95,1010–1013.

23. Campbell, H.D., Fountain, S., McLennan, I.S., Berven, L.A., Crouch, M.F.,Davy, D.A., Hooper, J.A., Waterford, K., Chen, K.-S., Lupski, J.R. et al.(2002) Fliih, a gelsolin-related cytoskeletal regulator essential for earlymammalian embryonic development. Mol. Cell. Biol., 22, 3518–3526.

24. Chang, B.S. and Lowenstein, D.H. (2003) Epilepsy. N. Engl. J. Med., 349,1257–1266.

25. Tomotsune, D., Shoji, H., Wakamatsu, Y., Kondoh, H. and Takahashi, N.(1993) A mouse homologue of the Drosophila tumour-suppressor genel(2)gl controlled by Hox-C8 in vivo. Nature, 365, 69–72.

26. Takahashi, N., Morishige, K.-I., Jahangir, A., Yamada, M., Findlay, I.,Koyama, H. and Kurachi, Y. (1994) Molecular cloning and functionalexpression of cDNA encoding a second class of inward rectifier potassiumchannels in the mouse brain. J. Biol. Chem., 269, 23274–23279.

27. Nadeau, J.H. (2001) Modifier genes in mice and humans. Nat. Rev.Genet., 2, 165–174.

28. Hogan, B., Beddington, R., Costantini, F. and Lacy, E. (1994)Manipulating the Mouse Embryo: A Laboratory Manual, 2nd ed. ColdSpring Harbor Laboratory Press, New York, NY, p. 7.

29. Hammond, P., Hutton, T.J., Alloson, J.E., Campbell, L.E., Hennekam,R.C.M., Holden, S., Patton, M.A., Shaw, A., Temple, I.K., Trotter, M.

et al. (2004) 3D analysis of facial morphology. Am. J. Med. Genet., 126A,339–348.

30. Matzuk, M.M., Finegold, M.J., Su, J.-G.J., Hsueh, A.J.W. and Bradley, A.(1992) a-inhibin is a tumour-suppressor gene with gonadal specificity inmice. Nature, 360, 313–319.

31. Ramı́rez-Solis, R., Davis, A.C. and Bradley, A. (1993) Gene targeting inembryonic stem cells. Methods Enzymol., 225, 855–878.

32. Adams, D.J., Biggs, P.J., Cox, T., Davies, R., Van Der Weyden, L.,Jonkers, J., Smith, J., Plumb, B., Taylor, R., Nishijima, I. et al. (2004)Mutagenic Insertion and Chromosome Engineering Resource (MICER).Nat. Genet., 36, 867–871.

33. Liu, P., Zhang, H., McLellan, A., Vogel, H. and Bradley, A. (1998)Embryonic lethality and tumorigenesis caused by segmental aneuploidyon mouse chromosome 11. Genetics, 150, 1155–1168.

34. Baldini, A. and Lindsay, E.A. (1994) Mapping human YAC clones byfluorescence in situ hybridization using Alu-PCR from single yeastcolonies. Methods Mol. Biol., 33, 75–84.

35. Hansen, G.M., Skapura, D. and Justice, M.J. (2000) Genetic profile ofinsertion mutations in mouse leukemias and lymphomas. Genome Res.,10, 237–243.

36. Richtsmeier, J.T., Baxter, L.L. and Reeves, R.H. (2000) Parallels ofcraniofacial maldevelopment in Down syndrome and Ts65Dn mice. Dev.Dyn., 217, 137–145.


Dow


ic.oup.com/hm


Documents

Reduced penetrance of craniofacial anomalies - Human Molecular