Foramina parietalia permagna (FPP) is an autosomal dominant condition characterized by cranial defects of the parietal bones. It can be present as an isolated feature, but it is also one of the characteristics of a contiguous gene syndrome associated with deletions on chromosome 11p11–p12. One of the proteins known to be involved in skull development is the MSX2 homeobox protein. Previously, MSX2 has been shown to be mutated in patients suffering from Boston type craniosynostosis. We have now analyzed the MSX2 gene in five families affected with FPP. An intragenic microsatellite marker did not reveal any recombination and a cumulated LOD score of +3.2 at θ = 0 was obtained. Sequence analysis further showed that in four out of five families an MSX2 mutation was responsible for the skull defect. Moreover, it appears that FPP is caused by haploinsufficiency of the MSX2 gene. This implies that Boston type craniosynostosis and FPP are allelic variants of the same gene, with FPP caused by loss of MSX2 function and craniosynostosis Boston type due to gain of MSX2 function.
Received 28 February 2000; Revised and Accepted 8 March 2000.
The MSX2 gene is a member of the MSX homeobox gene family, a small family of homeobox genes related to the Drosophila gene muscle segment homeobox (msh) (1). At present, two human MSX genes (MSX1 and MSX2) have been isolated (2–4), whereas three Msx homologues (Msx1-3) have already been identified in mouse (5–7). The MSX2 protein appears to play an important role in craniofacial development as is clearly illustrated by expression studies, showing expression at several key sites of the developing skull (1). Moreover, a specific P148H missense mutation in the MSX2 homeodomain has previously been found to cause Boston type craniosynostosis (4). Functional studies have revealed that this mutation affects the DNA binding properties of MSX2 through a stabilizing effect on the DNA–MSX2 binding, leading to a disease-causing gain-of-function (8). This is corroborated by the generation of transgenic mice that overexpress mutated or wild-type MSX2 and show premature suture closure and multiple craniofacial malformations (9,10).
Another example of abnormal skull development is found in patients affected with foramina parietalia permagna (FPP), also referred to as the ‘Catlin mark’. This malformation is characterized by the presence of ossification defects in the parietal bones of the skull (11). These openings, which show great variability among different patients, are mainly located symmetrically to the sagittal suture and their size normally decreases with time (12). Generally, these lesions are of no clinical significance, but on rare occasions the defects can be very large, requiring neurosurgical intervention. FPP can be inherited as an isolated autosomal dominant condition (11,13–16), or it can be present as one of the symptoms in a contiguous gene syndrome associated with microdeletions on the short arm of chromosome 11 (17,18).
We have now analyzed five families with FPP for the presence of mutations in the MSX2 gene to investigate a potential role for this homeobox gene in this disorder.
RESULTS AND DISCUSSION
In this study, we examined five families affected with isolated autosomal dominant FPP (Figs 1 and 2), in order to identify the causal mutation responsible for this malformation. Potential linkage to the MSX2 region on chromosome 5q34–q35 was investigated by the analysis of a previously identified dinucleotide repeat (MSX2GT) in intron 1 of the MSX2 gene (4). No recombinants were observed in any of the five families, and despite the lack of information in some families, a total cumulative LOD score of +3.2 at θ = 0 was obtained.
The MSX2 coding sequence is contained within two exons, and we sequenced the entire coding region in two patients from every family. In two families a nonsense mutation was found in exon 1 (Table 1). In family 5 mutation of cDNA nucleotides 265 and 266 (with the adenine of the start codon designated +1) generates a premature stop codon (Fig. 3). The truncated MSX2 protein lacks the entire C-terminus containing the homeodomain and has therefore lost the function(s) associated with this conserved domain. Similarly, in family 3, a deletion of one cytosine at position 344 or 345 was identified, resulting in a premature termination of translation after 114 amino acids (Fig. 3). The identification of these inactivating mutations supports the hypothesis that FPP is probably caused by loss of function mutations in the MSX2 gene, a theory that is corroborated by a recent finding of Wilkie et al., reporting the identification of a large deletion comprising the entire MSX2 gene in a family affected with FPP (19).
In family 1, a thymine to cytosine missense mutation in exon 2 (Table 1), resulting in a substitution of leucine by proline (L154P), was identified (Fig. 3). A second missense mutation, resulting in the substitution of arginine to histidine (R172H), was detected in family 2 (Fig. 3). The latter mutation is identical to one previously described by Wilkie et al. (19), but whether both families are related is not known. For both mutations, allele-specific restriction analysis showed cosegregation with the disorder in both families and an absence of these substitutions in 100 control chromosomes. The two substituted amino acids, leucine 154 and arginine 172, are located in the homeodomain and are highly conserved between MSX2 of several species from men to Xenopus, and between the MSX1 and MSX2 proteins (4), suggesting that they are crucial for proper functioning of the MSX2 protein. Previously, a P148H mutation in the MSX2 homeodomain has been shown to cause an autosomal dominant form of Boston type craniosynostosis (4). This is most intriguing, as craniosynostosis is due to the premature fusion of one or more sutures, whereas FPP is in fact almost the opposite, with a delayed or incomplete closure of the opening between the frontal and parietal bones. The homeodomains from different genes and organisms show highly conserved amino acid sequences and they form a similar three-dimensional structure with three α-helices and an extended N-terminal arm (20). The P148H missense mutation, which causes Boston type craniosynostosis, is located at position 7 in the N-terminal arm of the MSX2 homeodomain, a region specifically involved in DNA binding (20). As shown by the study of MSX2 transgenic mice (9,10) and analysis of the DNA binding properties of mutant and wild-type MSX2 (8), the P148H mutation causes craniosynostosis through a gain of function by increasing the stability of the mutant MSX2–DNA complex. The missense mutations identified in FPP families 1 and 2, however, are located in helix I and helix II, respectively. These helices are believed to stabilize the folded structure and to help maintain the relative orientation of the N-terminal arm and helices III, which makes direct contact with the DNA. The stabilizing properties of helices I and II depend on the helix structure and the ionic interactions between opposite charged residues of both helices, and therefore substitution in one of these helices can result in conformational and functional loss (21). It is therefore likely that the mutations in our FPP families have such a destabilizing effect on the MSX2 protein that it can result in loss of MSX2 function, although this has to be confirmed by functional analysis of the FPP-associated missense mutations.
No mutation was detected in family 4. Due to the small size of the family with only two affected patients, we can neither exclude nor confirm linkage to the MSX2 region on chromosome 5q. Although MSX2 flanking markers are heterozygous in both patients, we can not exclude a small deletion encompassing the MSX2 gene with the intragenic marker MSX2GT being homozygous or hemizygous in both patients. We could not distinguish this family clinically or radiologically from those with a proven MSX2 mutation, and therefore two possibilities remain open to explain the genetic cause for FPP in this family. The phenotype can be caused by a deletion involving the MSX2 gene or by mutations located outside the MSX2 coding region, which we did not analyze. Secondly, the FPP can be caused by mutations in another gene located elsewhere in the genome. Evidence for genetic heterogeneity of FPP has already been provided since we have previously shown by deletion analysis of patients suffering from DEFECT 11 syndrome, that a gene for FPP must be located on 11p, in close proximity to the EXT2 gene (17,18).
In conclusion, we provide strong evidence that the majority of FPP casesare due to haploinsufficiency of the MSX2 gene. This corroborates the essential role of the MSX2 protein in the skull ossification process, as illustrated previously in transgenic experiments and functional analysis of the P148H mutation causing craniosynostosis (Boston type).
MATERIALS AND METHODS
Five families affected with FPP were included in this study (Fig. 2). Family 3 has been described previously by Preis et al., and part of family 5 has been described by Schmidt-Wittkamp and Christians (14,22). The remaining three families originate from the UK. The diagnosis of all patients was based upon skull X-rays (Fig. 1).
Genetic linkage analysis
Linkage to the MSX2 gene was investigated by PCR amplification of the intragenic MSX2GT repeat (4) using primers MSX2GT1 (5′-TccccTcTcAAcTgAAAgcAc-3′) and M13 reverse-tailed primer MSX2GT2 (5′-ggATAAcAATTTcAcAcAggAcAcATAgTTTTgAcAAAgg-3′). Additional markers (D5S498, D5S2008 and D5S2030) (23) were analyzed in family 4. The PCR–amplification mixture (10 µl) contained dNTPs (10 mM each), 0.5 pmol of two specific primers (0.5 pmol) and an IRD800 labeled M13-reverse primer (5′-GGATAACAATTTCACACAGG-3′), 1× PCR buffer and Taq DNA polymerase. PCR conditions were 5 min at 96°C, 35 cycles of 1 min at 96°C, 45 s at 57°C and 45 s at 72°C, and finally 10 min at 72°C. Amplification products were analyzed on a LI-COR IRD800 detection system. Linkage analysis was performed using the LINKAGE 5.1 program. An autosomal dominant mode with 90% penetrance and a disease frequency of 1/25 000 was assumed.
Exons 1 and 2 of the MSX2 gene were amplified with a GC-rich amplification kit (Clontech, Palo Alto, CA) according to the manufacturer’s recommendations, using intronic primers MSX2ex1a-MSX2ex1b (5′-gcTgccgggTTgccAgcgg-3′ and 5′-ccgcTcccTccAgTAcccc-3′) to amplify exon 1 and MSX2ex2a-MSX2ex2b (5′-gTAAcTTTcTTTTgcTAATccg-3′ and 5′-TcgTggAgAgggAgAggAAAccc-3′) to amplify exon 2. For both amplifications 35 cycles were performed at a Tm of 52°C and GC melt concentration 1×. Amplification products were gel purified with Gel Extraction kit (Qiagen, Valencia, CA), and sequenced using BigDye terminator chemistry (Perkin-Elmer, Foster City, CA) on an ABI 377 automated sequencer.
Allele-specific PCR–restriction digestion
Confirmation of mutations in families 1, 2 and 5 was obtained by restriction analysis of exon 2 with restriction enzymes AluI (family 1), or NlaIII (family 2), and HhaI digestion of exon 1 (family 5). To confirm the W115X mutation in family 3 a modified primer MSX2mod1 (5′-ATcggccgggTTccTggATc-3′) was used, creating a new BamHI restriction site only in the wild-type MSX2ex1a-MSX2mod1 amplification product. The amplification products were digested with BamHI restriction enzyme. All restriction digests were analyzed on 12% acrylamide gels.
We thank patients and families for contributing to this project. W.W. is a postdoctoral researcher for the Fund for Scientific Research, Flanders (FWO). This study was supported by a concerted-action grant from the University of Antwerp to W.V.H.
To whom correspondence should be addressed. Tel: +32 3 820 2630; Fax: +32 3 820 2566; Email: email@example.com
aThe adenine of the start codon designated +1.
‘?’, not known.