How to cite this article: Gurnett CA, Bowcock AM, Dietz FR, Morcuende JA, Murray JC, Dobbs MB. 2007. Two novel point mutations in the long-range SHH enhancer in three families with triphalangeal thumb and preaxial polydactyly. Am J Med Genet Part A 143A:27–32.
Triphalangeal thumb is a fingerlike digit containing three phalanges that occurs on the radial (preaxial) aspect of the hand and is often found in association with preaxial poyldactyly. Both disorders may occur in isolation, but familial cases of preaxial hand abnormalities most often demonstrate a range of phenotypes [Qazi and Kassner, 1988] including triphalangeal thumb, polydactyly, syndactyly, duplicated thumbs, duplicated halluces, and rare instances of more severe forms of long bone aplasia [Kantaputra and Chalidapong, 2000]. Most cases described have autosomal dominant transmission with complete penetrance [Radhakrishna et al., 1996; Zguricas et al., 1999] although one family with duplicated halluces was described with incomplete penetrance [Ray, 1987]. Although most families demonstrate variable expressivity, families have been described with more restricted phenotypic expression, including an Indian family with predominantly duplicated halluces [Ray, 1987], a Cuban family that also has radial and tibial aplasia [Zguricas et al., 1999] and a North American Caucasian family we previously described with isolated triphalangeal thumb [Dobbs et al., 2000].
Normal asymmetric hand development results from early posterior localization of the zone of polarizing activity (ZPA). The critical factor within the ZPA that orchestrates limb axis patterning is sonic hedgehog (Shh), whose gradient appears to specify normal preaxial development [Echelard et al., 1993; Riddle et al., 1993]. Further evidence in support of a dominant role for Shh in limb fate comes from mouse models of polydactyly (Ssq and Hx) that demonstrate additional ectopic anterior Shh expression [Sharpe et al., 1999].
Linkage mapping of several large families with triphalangeal thumb and preaxial polydactyly resulted in the identification of a PPD locus within a 450-kb region on chromosome 7q36, located approximately 1 Mb upstream of SHH [Heutink et al., 1994]. Mapping of a spontaneous balanced translocation breakpoint in a girl with preaxial polydactyly and a transgenic insertion in the Ssq mouse implicated the same region [Lettice et al., 2002], and more specifically narrowed the region to the fifth intron of the LMBR1 gene. Comparative genomic analysis of four vertebrate species revealed several highly conserved domains within the region, including 1.3-kb that is also conserved in Fugu rubripes (pufferfish). The highly conserved region was designated ZRS (ZRA regulatory sequence), and point mutations within this region were described in the hemimelic extra toes mouse, M100081 (a polydactylous mouse generated by large-scale ENU-mutagenesis) [Sagai et al., 2004], and four families with preaxial polydactyly [Lettice et al., 2003]. Analysis of the ZRS sequence did not reveal any known regulatory domains, and therefore identification of additional mutations and characterization of the resultant limb abnormalities may provide insight into ZRS functional domains.
MATERIALS AND METHODS
Four families with autosomal dominant triphalangeal thumb and preaxial polydactyly were ascertained (Fig. 1); two families (A and B) were described previously [Dobbs et al., 2000]. Interval history of family A, previously described as manifesting only isolated triphalangeal thumb, revealed that 1 of the 19 affected individuals also had a duplicated thumb. Notably, four family members were carriers of the disorder, but did not have evidence of limb deformity on examination or by report. All other affected individuals had short, opposable bilateral triphalangeal thumbs with delta phalanx of both thumbs on radiographs (Fig. 2).
Detailed clinical description of the nine affected members of family B was not given in the previous report and is given here. Four affected members had isolated bilateral triphalangeal thumb. Two had bilateral triphalangeal thumb and polydactyly of the right hand (duplicated thumb), one had bilateral triphalangeal thumb and polydactyly of the left hand. A single individual had supranumerary digits involving the right great toe as well as bilateral triphalangeal thumb. Another individual died early in life as a consequence of multiple deformities, including severe poorly described limb abnormalities. Family C includes two affected members with triphalangeal thumb and a single individual bilateral triphalangeal thumb and polydactyly. Family D consists of an affected mother and son with triphalangeal thumb. All families were from Northern European descent and no common ancestor could be determined by family history, although all lived in the Midwestern United States. Linkage to chromosome 7q36 was previously shown for families A and B [Dobbs et al., 2000].
Control samples are from the HGDP-CEPH human genome diversity cell line panel [Cann et al., 2002] and represents 1,051 unrelated individuals from 51 different world populations.
Consent was obtained and DNA was collected from cheek swabs as described previously [Dobbs et al., 2000]. Institutional review board approval (according to 45CFR46) was obtained for this research. DNA was obtained from 57 individuals, including 20 affected individuals. Sequencing of the most highly conserved 774 bp region of the ZRS was performed with primers as specified in Lettice et al. 2003 for all affected and unaffected family members. Direct sequencing performed by the Genome Sequencing Center (Washington University). Data was analyzed with Sequencher software.
Four nonsynonymous SNPs were identified within GLI3 (rs2079451, rs929387, rs10259802, rs846266) using dbSNP. SNP allele determination was performed for all individuals in whom DNA was available using direct sequencing.
Mutations within ZRS were identified in three of four families with isolated triphalangeal thumb and preaxial polydactyly. In pedigree A and C, a novel A/G mutation was identified near the 5′ end of ZRS at bp 739 that segregated with disease or carrier status (Fig. 3). A novel C/G mutation at bp 621 was identified in pedigree B that segregated with disease status in seven affected individuals, including four who manifested both preaxial polydactyly and triphalangeal thumb. Neither mutation was identified in more than 600 control individuals (1200 chromosomes). A C/G polymorphism, reported previously at bp 3 [Lettice et al., 2003] occurred in four unaffected individuals and did not segregate with disease status. No other polymorphisms were identified.
GLI3 SNP allele determination was performed for all individuals for 4 nonsynonymous coding SNPs in order to identify the potential role of this gene as a modifier of penetrance or phenotype. Two GLI3 SNPs (rs2079451, rs10259802) were uninformative as the minor alleles were not present in any individual. Two GLI3 SNPs (rs929387, rs846266) were informative in these families with high heterozygosity. There was no apparent correlation between allele frequency of these two GLI3 SNPs and nonpenetrance in family A (data not shown) or between allele frequency and presence of polydactyly (in addition to triphalangeal thumb) in family B.
Analysis of conserved putative transcription factor bindings sites within the ZRS between human and mouse revealed conserved consensus binding sites for three transcription factors known to be involved in limb development: Cdx, Meis1, and Sox9 (Fig. 4). There were 18 conserved Cdx binding sites; 3 of the 6 known human mutations causing preaxial polydactyly (A739G) and 2 described in Lettice et al. 2003 disrupt conserved Cdx binding sites. One mutation described here (C621G) disrupts a putative Cdx site in human sequence that is not conserved in mouse. The Meis1 and Sox9 consensus binding sites are also conserved in pufferfish (Takifugu rubripes), although none of the Cdx sites are found.
Triphalangeal thumb is a distinct hand abnormality that appears to be the minimal expression of ZPA dysregulation. Individuals of family A manifest opposable, triphalangeal thumb of relatively short length (containing a delta phalanx) consistently throughout five generations, with only a single individual who also had preaxial polydactyly in the form of a duplicated thumb. The ZRS point mutation identified in this family (A739G) lies at the most distal end of the conserved region, suggesting the possibility that this mutation minimally alters the ZRS and hence results in limited expression of the phenotype: triphalangeal thumb. Furthermore, the disorder in family A is not fully penetrant, as four carriers of this mutation had no evidence of hand abnormalities but passed on the disorder to their children. As familial preaxial polydactyly is typically fully penetrant [Qazi and Kassner, 1988; Heutink et al., 1994; Radhakrishna et al., 1996], nonpenetrance also suggests that the mutation carried by this family may be less disruptive to the ZRS. However, it is also possible that modifier genes are also common to most members of this family and therefore result in a similar phenotype.
GLI3 is expressed in an anterior domain complementary to SHH [Buscher et al., 1997] and is known to negatively regulate the expression of SHH and its target genes [Wang et al., 2000]. GLI3 mutations are found in patients with Greig cephalopolysyndactyly and Pallzister Hall syndrome [Johnston et al., 2005] and result in a variety of hand abnormalities, including preaxial polydactyly. SHH and GLI3 have both been shown in mouse models to regulate digit number and identity [Litingtung et al., 2002]. For these reasons, GLI3 appeared to be an excellent candidate as a modifier gene for preaxial polydactyly. However, we did not find a correlation between the alleles of two GLI3 nonsynonymous SNPs and nonpenetrance or polydactyly in our families with preaxial polydactyly and triphalangeal thumb. A more complete evaluation of modifier genes may be considered when the full complement of genes involved in anterior–posterior limb specification are known, but will likely require extremely large families in order to identify modifier genes of small effect.
Interestingly, several ZRS mutations, including two described here, disrupt consensus Cdx transcription factor binding sites. Cdx likely plays a role in limb development as demonstrated by the presence of “split” first digits seen in some double Cdx1 and Cdx2 hypomorphic mice [van den Akker et al., 2002]. However, inconsistent effects of ZRS mutations were seen with in vitro transcription factor assays, although the ZRS did bind both cdx1 and cdx2 in vitro (Gurnett, data not shown). Within the ZRS, there are numerous Cdx consensus transcription factor binding sites, possibly as a result of the extreme degeneracy of its consensus binding sequence (A-A/T-T-A/T-A-T-A/G) [Margalit et al., 1993]. It is unknown whether any of the identified Cdx sites within ZRS are functional in vivo. Certainly, alterations of Cdx binding do not explain the pathogenesis of preaxial polydactyly entirely, as several of the previously identified mutations do not appear to directly alter Cdx consensus binding sites [Lettice et al., 2003].
Mutations in the ZRS appear to be frequent causes of familial preaxial polydactyly and isolated triphalangeal thumb, but do not account for the disease in all families. We identified two novel ZRS mutations that segregate with disease in three of four families. Lettice et al. identified four different mutations in four of seven families with preaxial polydactyly (two Belgian, one Cuban, and one Dutch). Of note, there are at least three other highly conserved domains located within the critical region [Lettice et al., 2003] [Sagai et al., 2005] that may also contribute to the ZRS and be responsible for disease in these families where the mutation is not known. Mutational analysis and clinical descriptions of additional families with preaxial polydactyly and triphalangeal thumb will be required to better understand the minimal domains required for normal anteriorposterior limb formation.
MBD is supported by the Orthopaedic Research and Education Foundation, the Aircast Foundation, and the Shriners Hospital for Children. CAG is supported by the NINDS K12 (NS01690).