Triphalangeal thumb-polysyndactyly syndrome (TPT-PS; OMIM 174500), also called preaxial polydactyly type II (PPD-II), is a well defined autosomal dominant disorder that has been described in several families. The phenotype comprises, as indicated by the name, triphalangeal thumbs, pre- and postaxial polydactyly, complex polysyndactyly, and isolated syndactyly. Within families the phenotype is variable, ranging from mild to severe, with the individual features occurring in combination or as single entities. In the past, linkage analysis of large pedigrees with TPT-PS1 2 and related phenotypes with hand and foot involvement, such as syndactyly type IV3 or acropectoral (F-) syndrome,4 revealed linkage to chromosome 7q36. Subsequently, point mutations in a non-coding region 5’ of Sonic hedgehog (SHH) were shown to result in TPT-PS.5 Further studies identified this non-coding sequence as a long range cis-regulatory element of SHH located about 1 Mb upstream of the target gene in an intronic sequence of another gene, LMBR1 (Limb region 1 homolog, OMIM 605522). The region comprises a multi-species conserved sequence (MCS) in vertebrates with limbs or limb-like structures such as wings or fins, but is absent in limbless species such as snakes. Other descriptions of mutations within this regulatory sequence associated with TPT-PS supported the functional importance of this region. To date, seven different single nucleotide exchanges are known in humans.5^–7 In addition, mutations in the murine region homologous to human intron 5 of LMBR1 were shown to cause limb phenotypes in several mouse mutants: Hemimelic extra toes (Hx), Sasquatch (Ssq), M101116 and M100081.5 8 9

Shh is a known key regulator in defining the limb anterior–posterior axis in early embryogenesis. During development of the limb Shh expression is restricted to a region called zone of polarising activity (ZPA) in the posterior part of the limb bud from where it sets up a morphogen gradient which patterns the limb. Expression of Shh in the ZPA is controlled by a region designated as ZPA regulating sequence (ZRS).5 10 11 In the mouse mutants Hx and Ssq an ectopic expression of Shh in the anterior margin of the limb bud results in an additional ZPA,12 13 a mechanism which has also been shown in chick limb to result in supernumerary digits.14 These results suggest that similar mechanisms produce polysyndactyly and triphalangeal thumbs in humans. The here described microduplication in 7q36.3 containing the limb SHH regulator region results in TPT-PS, thus imitating TPT-PS caused by point mutations in the ZRS.

CLINICAL REPORT

We investigated a large family with a phenotype comprising triphalangeal thumbs, preaxial and/or postaxial synpolydactyly, and cutaneous/osseous syndactyly of fingers III–V or IV/V leading to the diagnosis of typical TPT-PS. The phenotype varies among affected individuals and the latter features occurred either in isolation or in combination. In general, the feet were less severely affected than the hands. Further organ anomalies or dysmorphic features were not present in the affected individuals. The disorder follows an autosomal dominant inheritance as can be deduced from the pedigree (fig 1A). Some typical examples of the hand phenotypes are shown in fig 1B. In total, 15 members of this pedigree were clinically investigated. An overview of the hand phenotype is given in table 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Pedigree and typical examples of the triphalangeal thumb and polysyndactyly (TPT-PS) phenotype. (A) Pedigree of the family. Affected individuals are indicated by black symbols. Symbols with crossed lines indicate individuals from whom material was available for testing. (B) Examples of phenotypes observed in affected individuals. Numbers refer to individuals in pedigree 1A. Photographs, III1: note broadened, triphalangeal thumb. Radiographs, IV1, left image, thumb: note small extra phalanx between the proximal and distal phalanges (minimal variant of a triphalangeal thumb), IV1, right image: postaxial hexadactyly at finger V; IV2: note syndactyly involving fingers III, IV and V and additional postaxial hexadactyly.

PATIENTS AND METHODS

Molecular analysis

Informed consent was obtained for genetic analyses from all patients or their legal guardians. Molecular testing was performed on purified genomic DNA obtained from blood samples or buccal swabs. Symbols with crosses in fig 1A indicate the individuals from whom material could be obtained—that is, nine clinically affected and six unaffected individuals. Mutation screening of the highly conserved SHH regulatory sequence consisting of about 800 bp located in intron 5 of LMBR1 was carried out.10 Purified polymerase chain reaction (PCR) products were sequenced in both directions. PCR primers were also used as sequencing primers and the sequencing reaction was performed with the ABI Prism BigDye terminator cycle sequencing reaction kit (Applied Biosystems, Foster City, California, USA). Primer pairs and PCR conditions are available on request. Products were evaluated on an automated capillary sequencer (Applied Biosystems).

Locus specific linkage analysis was performed with several microsatellite markers in the region located close to SHH and the ZRS in 7q36.3. Microsatellite primers used and PCR conditions are available upon request. PCR products were analysed by capillary automated genotyping. Primers already designed for quantitative real-time PCR were used for the breakpoint analysis and identification of a junction fragment to confirm our hypothesis of a tandem duplication (primers E-forward/C-reverse).

Cytogenetics and microarray based comparative genomic hybridisation (array CGH)

Karyotyping of GTG-banded chromosomes from lymphocytes at 500 bands resolution was performed according to standard procedures. Array comparative genomic hybridisation (array CGH) was carried out using a submegabase whole human genome tiling path bacterial artificial chromosome (BAC) array as previously described.15 Images were scanned using GenePix 4100A and analysed by GenePix Pro 6.0 software (Axon Instruments, Foster City, California, USA). Further analyses and visualisation were performed with the BlueFuse software (Bluegnome, Cambridge, UK). Raw data were normalised by “Subgrid LOWESS”. Copy number gains and losses were determined by a conservative log2ratio threshold of 0.3 and −0.3, respectively. Profile deviations consisting of three or more neighbouring BAC clones are considered as genomic aberrations and are further evaluated by quantitative real-time PCR unless they coincided with published copy number polymorphisms Database of Genomic Variants (http://projects.tcag.ca/variation/; version August, 2007).

Quantitative real-time PCR (qPCR)

Genomic DNA samples were obtained from EDTA blood and buccal swabs of 15 blood relatives and of four in-laws used as control samples (see supplementary table 1). The first subset of primers was evenly distributed over the duplicated region and the flanking sequences as proposed by the array CGH experiments (mean distance 50 kb), which allowed screening for the breakpoint regions of the duplication. Subsequent primer pairs covered the respective proximal and distal breakpoint regions in more detail. Primer sequences can be obtained upon request.

qPCR was performed on ABI Prism 7500 Sequence Detection System in a total volume of 24 μl in each well containing 12 μl of SYBR-Green PCR Master Mix (ABI SYBR Green PCR Master Mix), 20 ng of genomic DNA (10 μl) and 2 μl primers (0.2 μmol each). Samples were run in triplicates in separate tubes to permit the quantification of the target sequences normalised to albumin (ALB). PCR conditions were according to manufacturer’s protocol, and consisted of an initial denaturation step of 95°C for 8 min followed by 40 cycles with denaturation at 95°C for 15 s and a combined annealing/elongation step at 60°C for 1 min. By using calibrator samples of normal control genomic DNA the gene copy number was estimated based on the ddCt method. In addition, we performed an identification of the individuals’ genders calculating the coagulation factor VIII (F8, Xq28) relative to the two-copy-control ALB to assure its reliability.

RESULTS

A microduplication in 7q36 identified by array CGH results in TPT-PS

Karyotyping of GTG-banded chromosomes in an affected family member (III1) did not reveal any chromosomal abnormalities. One patient with TPT-PS was screened for point mutations in the SHH regulatory sequence as described previously, but no mutation was identified. However, locus specific linkage analysis using microsatellite markers showed linkage to 7q36. Because microduplications or microdeletions were already shown to result in isolated hand and foot malformations, we subsequently performed array CGH analysis to screen for submicroscopic chromosomal aberrations in an affected family member (II3). Using this technique, a duplication on chromosome 7 ranging from BAC clone RP11-691F21 to BAC clone CTD-2008L01 was detected (fig 2, arrow). According to the array CGH data the duplication extends from 155.85 Mb to 156.31 Mb on 7q36.3. The duplication is most likely causative since deletions/duplications in this genomic region and of these BAC clones have not been observed in more than 700 individuals and are not described as a DNA copy number variant in the genomic variants database.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 Array comparative genomic hybridisation (array CGH) profile of chromosome 7q36 revealed a duplication on 7q36.3. The microduplication of 7q36.3 is indicated by the arrow. Each spot represents one bacterial artificial chromosome (BAC) clone. Vertical lines mark the log2ratio thresholds of −0.3 (deletion) and 0.3 (duplication).

Co-segregation of the microduplication confirmed by qPCR

Using qPCR, the duplication was confirmed in all affected family members and was not detected in any of the clinically asymptomatic individuals, confirming a co-segregation of the aberration with TPT-PS and suggesting full penetrance of the disorder in the described family. By qPCR analysis the breakpoints could be narrowed down to a region of about 1.7 kB between primers B and C (centromeric breakpoint) and 1.1 kB between primers E and F (telomeric breakpoint), as indicated in table 2 and shown in fig 3.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3 Microduplication on 7q36.3 confirmed by quantitative real-time polymerase chain reaction (qPCR). The mean values for relative quantification (RQ) were exported from the 7500 SDS software. For affected and non-affected individuals mean values and SDs (error bars) relative to albumin (ALB) on 4q11 as autosomal two-copy reference gene were calculated for each target primer. For gender determination mean values and SDs for coagulation factor VIII (F8, Xq28) were calculated relative to ALB. Results were calibrated to the mean value determined for a healthy female control. A–F refer to primers shown in table 2. One duplicated allele plus one normal allele results in three copies for the amplicons of primer pairs C, D and E in the affected individuals and therefore in a ratio of 1.5 relative to the two copies of the healthy female control.

Breakpoint identification revealed a tandem-duplication of 588,819 bp

Analysis by PCR with primers E-forward and C-reverse (table 2) on genomic DNA level allowed the identification of a junction fragment that included the transition site between the telomeric and centromeric breakpoints in affected family members’ DNA. This fragment of approximately 330 bp was undetectable in the non-affected individuals (data not shown). By direct sequencing (fig 4) we could identify the centromeric breakpoint of the microduplication to be between nucleotide 155,836,146 and 155,836,147 on chromosome 7 and the telomeric between nucleotide 156,424,965 and 156,424,966, respectively (positions according to Ensembl release 46, August 2007). The tandem orientation of the duplicated fragment of exactly 588,819 bp could be confirmed. A summary of performed investigations for each family member is listed in supplementary table 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4 Breakpoint identification by sequence analysis. The top line displays the telomeric, the second line the centromeric reference sequence as given by Ensembl release 46, August 2007. Below are the electropherograms of the affected individuals II3, IV1 and IV2. The yellow colour indicates the recognition site for topoisomerase I (5’-CATT-3’). The arrow indicates the cleavage site of topoisomerase I followed by three homologous nucleotides (5’-TTC-3’) surrounded by the blue box. Thus, the three homologous nucleotides of the junction fragment belong to the centromeric reference sequence (bold letters).

DISCUSSION

Long range regulators are currently known for several genes with functions in developmental processes such as signalling molecules. Regulatory elements can be dispersed in regions over hundreds of kilobases upstream or downstream of the gene itself. These regulators modulate the program of transcription and gene expression in specific tissues and control their expression at defined temporal processes in embryonic development. Genes with a highly tissue-specific expression appear to require multiple regulatory elements. The interacting regulatory mechanisms seem to be very complex and are poorly understood to date. Deletions of regulatory elements or chromosomal rearrangements such as inversions or balanced translocations involving regulating regions are known to be the underlying cause for several disorders,16 17 but duplications of regulatory elements have so far not been described to be associated with human disease.

The identified microduplication encompassing a region from 155.836 to 156.424 Mb begins about 540 kb upstream of SHH (fig 5). With the exception of one hypothetical gene (Q6ZVM9_human), the 5’ region of SHH is a gene desert extending over more than 800 kb. Gene deserts located in adjacent regions of developmental control genes argue for the presence of non-coding regulator elements as shown for other genes—that is, SHOX, PAX6, and SOX9.18^–20 The identified duplication comprises about 300 kb of this gene-free region, the further distally located genes NM_032625, C7orf13, RNF32 and the entire LMBR1 including the ZRS.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5 Schematic representation of the critical region on 7q36.3 and the identified duplication. Sonic hedgehog (SHH) and its known limb regulator element ZRS are shown. The ZRS lies in intron 5 of LMBR1 (156.166 – 156.378 Mb) and is thus located about 1 Mb upstream of SHH. Genes located between the ZRS and SHH are indicated by black bars. Small triangular arrows under the genes denote the direction of transcription. The duplicated region (grey box) in the here described family comprises 588.819 bp beginning at nucleotide 155,836,147 (centromeric breakpoint) and ending at nucleotide 156,424,965 (telomeric breakpoint), thus including LMBR1 and the ZRS. The detailed view into LMBR1 indicates already known mutations in the critical regions. * Seven different point mutations in humans have been reported to be associated with triphalangeal thumb-polysyndactyly syndrome (TPT-PS)6^–8 that are clustered within a ∼800 bp highly conserved region of the ZRS (156.276 Mb). In addition, three point mutations in mice lead to the similar phenotype.5 9 The fasciated box indicates the transgene insertion and duplication of ZRS causing the identical phenotype in the mouse mutant Ssq.8 Double underlined is the homozygous deletion of exon 4 causing acheiropody in humans.29 Jeong et al22 identified Shh enhancers for the ventral forebrain in mice being located in Bacs 447L17 and 265M1. In the middle, the human homologous regions of murine Bac clones 447L17 (155.638 and 155.817 Mb) 265M1 (155.769 and 155.952 Mb) and 265M10 (156.228 and 156.406 Mb) are shown. The duplication begins at 155.836 Mb and is thus lying after Bac 447L17 and in the middle of Bac 265M1 when compared with the murine genomic sequence is lying within Bac 265M1.

Genomic duplications can either be caused by homologous or non-homologous recombination. Since the breakpoint intervals of the duplication identified by qPCR contained two AluSx elements on the same strand, homologous SINE mediated recombination was initially considered as a possible underlying mechanism. However, the exact breakpoints shown by sequence analysis of the junction fragment were not located within those AluSx elements, making this mechanism rather implausible. Instead, non-homologous recombination mediated by topoisomerases may have caused this event based on the identification of a hot spot sequence for illegitimate recombination as described by Zhu and Schiestl.21 Topoisomerase I creates a nick immediately next to the 3’ end of the 5’-(G/C)(A/T)T-3’ recognition sequence. In case of illegitimate recombination the cleavage site is followed by two to six homologous nucleotides. After the recognition site 5’-CAT-3’ identified here three nucleotides (5’-TTC-3’) are homologous as shown in fig. 4.

The functions of Shh are manifold in different tissues and developmental stages, thus multiple regulators are needed to control Shh transcription. In addition to the cis-acting limb specific regulator, further long range enhancer elements were shown to regulate Shh expression in the ventral forebrain and neural tube.22 A disruption of these enhancers leads to a 50% reduction of Shh expression in the ventral forebrain. In humans this results in disorders of the holoprosencephaly (HPE) spectrum similar to loss of function mutations in the coding region of SHH.23 It was shown in mice22 that ventral forebrain enhancers for Shh are located within Bac clones 447L17 and 265M1 on murine chromosome 5 corresponding to highly conserved regions upstream of SHH in humans (155.638–155.817 Mb and 155.769–155.952 Mb, respectively). The centromeric breakpoint of the here described microduplication is located at 155.836 Mb. Since all affected individuals of the investigated family do not show signs of HPE we hypothesise that in humans the regulatory elements for SHH expression in the brain are located within the ∼200 kB interval between 155.638 and 155.836 Mb (fig 5).

Investigation of the gene desert upstream of SHH in more detail reveals other highly conserved sequences around the ZRS, but their significance is unknown.24 These conserved elements as well as the ZRS are located within the here described tandem duplication since the telomeric breakpoint is located upstream of LMBR1. Our observation that in this family only the limbs are affected indicates that just the function of the limb-specific regulators is disturbed by the duplication. This conclusion is underlined by the data of Jeong et al22 who were able to show that distal limb specific regulation of Shh in mice is restricted to the region confined by Bac 265M10 corresponding to the interval between 156.228 and 156.406 Mb on human chromosome 7 that is entirely included in the microduplication (fig 5).

The precise mechanisms of the pathogenic effect caused by the microduplication described here are unclear, thus different functional possibilities have to be taken into consideration. Since SHH belongs to the group of developmental master genes, its function needs to be tightly regulated with respect to spatiotemporal expression pattern in specific tissues. This regulation requires enhancer elements (cis-regulatory elements) that are located within or around the gene either in gene deserts or within the introns of neighbouring genes.8 25 Impaired position effects between cis-acting regulators and target genes could be an explanation for the pathogenic consequences. Such interactions can be disturbed by physical dissociation of the regulator from the target gene—for example, by a translocation or an inversion. Tolhuis et al26 proposed the term “active chromatin hub” (ACH) for a distinct spatial unit of regulatory elements. Physical interaction between cis-acting sequences (enhancers) and the promoter mediated by the presence of protein factors (for example, transcription factors) leads to the formation of the ACH and initiation of gene expression. Thus, the repression or activation of gene transcription is contingent upon the structure of the ACH—that is, the orientation and nature of its elements.27 It is conceivable that at the SHH locus tissue-specific ACH formation between the relevant tissue-specific cis-acting elements (that is, forebrain and limb enhancers) and the promoter mediate spatiotemporal regulation of SHH expression during development.

The role of the ZRS in limb development has been interpreted as a positive regulator driving Shh expression in the posterior limb bud as well as a repressor that silences the anterior expression. Accordingly, mice with a deleted ZRS display a severe truncation of all distal skeletal elements due to a specific loss of SHH expression in the limb bud.24 This phenotype is also present in mice with inactivated Shh alleles and imitates the human autosomal recessive disorder acheiropody which is characterised by a symmetrical distal aplasia of the limbs.28 This condition is caused by homozygous deletions of exon 4 and flanking intronic sequences of LMBR1.29 The deletion does not include the ZRS but the similarity of phenotypes makes it highly likely that SHH regulation is disturbed by an up to date unknown mechanism resulting in a loss of function of the limb specific SHH enhancer. TPT-PS is a much milder condition with a different pathogenic pattern that cannot be explained by loss of SHH function in the limb bud. Alterations of the ZRS lead to misexpression of Shh in the anterior part of the limb bud as already shown in the Ssq and Hx mouse models.9 13 Maas and Fallon demonstrated by an in vivo reporter assay that point mutations within the limb specific Shh enhancer are sufficient to drive ectopic Shh expression in the anterior part in addition to the typical Shh expression domain in the posterior region of the limb bud.30 Interestingly, the same effect was observed in the Ssq mouse in which a 20 kb region in intron 5 of Lmbr1 is duplicated.8 The duplication described here is likely to function through similar mechanisms.

One can speculate that a protein that could be either activating or repressing binds to the ZRS and thus modulates the regulation. The duplication of the ZRS results in more binding sites and thus more bound protein, possibly leading to an augmented and/or erroneous SHH expression in the limb bud. The TPT-PS phenotype in the family described here is compatible with an ectopic expression of SHH at the anterior side of the limb bud. The presence of postaxial polydactyly and syndactyly argues for an additional increased posterior expression of SHH. To date, the precise molecular mechanism as to how a duplicated ZRS region or single nucleotide alterations within the ZRS lead to SHH misexpression is unresolved. One possible explanation is a gain of function mechanism leading to an augmented SHH expression in the limb bud. However, a specific loss of repressing function due to failed binding of inhibitory transcription factors resulting in an increased SHH expression cannot be ruled out.

In summary, this work describes a genomic duplication containing a limb-specific SHH enhancer as a cause of TPT-PS. The phenotype associated with this duplication is similar to those caused by single nucleotide alterations in the SHH regulatory element ZRS. Disruption of regulatory elements by cryptic chromosomal rearrangements might account for developmental defects of yet unknown cause.