17p13.3 genomic rearrangement in a Chinese family with split-hand/foot malformation with long bone deficiency: report of a complicated duplication with marked variation in phenotype

Split-hand/foot malformation, affecting 1 in 8500–25,000 newborns, is a developmental limb malformation characterized by median clefts of the hands and feet, syndactyly, and aplasia and/or hypoplasia of the phalanges, metacarpals, and metatarsals [1]. However, the phenotype varies greatly between families, within members of one family, and even among different extremities of a single individual, ranging from slight syndactyly and/or ectrodactyly to severe monodactyly. Currently, SHFM is categorized as either non-syndromic or syndromic, in terms of whether there are extra-limb manifestations. Non-syndromic SHFM includes isolated SHFM; and SHFM in association with other limb manifestations, such as tibia or femur aplasia, is called SHFM with long bone deficiency (SHFLD), and others involving fibula defects are designated as fibular agenesis with ectrodactyly (OMIM 113310).

Until now, six genes/loci (SHFM1–6) have been identified as responsible for isolated SHFM. These include deletions, duplications, or rearrangements on 7q21.​2-q21.​3, heterozygous/homozygous mutations in DLX5 or heterozygous mutations in DLX6 (SHFM1, OMIM 183600), heterozygous duplications on 10q24 (SHFM3, OMIM 246560), heterozygous mutations in TP63 (SHFM4, OMIM 605289), heterozygous deletions on 2q31 (SHFM5, OMIM 606708) and biallelic mutations in WNT10B (SHFM6, OMIM 225300); in addition, SHFM2 has been mapped to Xq26.3 by linkage analysis. Causative genes in isolated SHFM are also responsible for some syndromic SHFM. For example, dominant TP63 mutations can also result in ectrodactyly, ectodermal dysplasia, and cleft lip/palate syndrome 3 (EEC3, OMIM 604292).

SHFLD is genetically distinct from isolated SHFM. SHFLD is usually inherited in an autosomal dominant manner with variable expressivity and incomplete penetrance, and in some cases, autosomal recessive inheritance is presented, and digenic inheritance can also be possible. In general, SHFLD is associated with three different loci: 1q42.2-q43 for SHFLD1 (OMIM 119100); 6q14.1 for SHFLD2 (OMIM 610685); and 17p13.3 for SHFLD3 (OMIM 612576). The first two SHFLD susceptibility loci, 1q42.2-q43 and 6q14.1, were identified in the same large consanguineous family by genome-wide linkage analysis and multipoint parametric linkage analysis [2, 3], raising the possibility of digenic inheritance. For SHFLD3, 17p13.3 duplications have been identified in several cases and families, and the smallest duplication region only includes BHLHA9 [4‐10]. BHLHA9 plays a crucial role in limb development during apical ectodermal ridge (AER) formation, and its dosage has been implicated in SHFM and SHFLD [11]. SHFLD with 17p13.3 duplication is less than 50% penetrant and shows markedly variable expressivity. The correlation between a specific phenotype and the 17p13.3 duplications remain unclear. Although BHLHA9 duplication is highly associated with long-bone deficiency, it is insufficient to explain the variable expressivity and reduced penetrance in SHFM and SHFLD.

Here, we present a genetic analysis of a Chinese family with SHFM and SHFLD. A ~ 966 kb genomic rearrangement in 17p13.3 is associated with the observed malformations in the family. Variable expressivity and incomplete penetrance are demonstrated in the family.

In the present family, a total of nine members were affected with limb malformation, including four fetuses. Most patients manifested as SHFM with unilateral or bilateral hand(s) involved, while lower limb involvement was present in one third of the patients. One male patient (II1) presented a severe SHFLD phenotype with long-bone deficiency in three of four limbs, which included a hypoplastic right forearm, and bilateral absence of lower legs, in addition to a split left hand. According to previous reports, only 43% of SHFM patients with 17p13.3 duplication are affected with long bone deficiency, with tibia involved in most cases, but also with radial hypoplasia, femur hypoplasia, and a reduction defect of upper and/or lower limbs [9]. To our knowledge, patient II1 of this study is the most severe SHFLD case in all reported 17p13.3 genomic rearrangement carriers. The most severe phenotypes previously reported were reduction defects of the right upper and lower limbs with a split left hand in a fetus [9], while all four limbs of patient II1 herein were involved, of which three of four limbs were seriously dysplasic or aplasic. Whether there is a sex bias in the incidence and severity of SHFLD is still controversial [9, 10, 12]. In Caucasian patients, a clear sex bias has been observed, with more males affected than females [12], while no significant sex difference has been observed in Japanese populations [10]. In this study, one severe SHFLD patient is observed out of four males, one suspicious case with bilateral leg flexion deformity out of six females, and one SHFLD case with unknown gender in the whole family. Two asymptomatic carriers, one male and one female, were also observed. Limited by the small sample size, the relationship between gender and the severity and incidence of long-bone deficiency is speculated, but not confirmed, in this family.

A ~ 966 kb genomic rearrangement was identified at 17p13.3 in a four-generation Chinese family with nine individuals affected with SHFM or SHFLD. Within this large region, there was one part containing ABR1 and BHLHA9 presenting triplication, which showed a complicated structural variation in this region. The phenotypes associated with the 17p13.3 duplications included neurological dysfunction, cleft lip/palate, SHFLD, and connective tissue phenotypes [6]. To our knowledge, YWHAE duplication is responsible for the neurological dysfunction phenotype [13], and a duplication disrupting ABR appears to result in cleft lip/palate [6], while the SHFLD phenotype has been associated with duplication of BHLHA9 [5, 12]. The identified duplication in our study does not contain YWHAE, and it encompasses ABR, along with BHLHA9. Accordingly, patients in this family were only affected with limb malformations, but no neurological disorders or cleft lip/palate were observed. These observations are in accordance with the genotype-phenotype correlations summarized by Curry et al. [6]. Increased BHLHA9 copy numbers have been reported to be highly correlated with the long bone deficiency phenotype. According to Nagata et al., the incidence of long bone defect is significantly higher in families with BHLHA9 triplicates than with duplicates [10]. In this study, the increased copy number of 17p13.3 containing BHLHA9 was confirmed by qPCR assay, and it was considered to be responsible for the phenotypes observed in this Chinese family. In addition to previous studies in Caucasian and Japanese patients, the same genomic variations were observed in Chinese cases for the first time, indicating that increased BHLHA9 copy number is associated with SHFLD in different populations.

The ~ 966 kb genomic rearrangement identified in this family is the largest SHFLD associated 17p13.3 structural variation identified to date, and the long bone deficiency phenotype in one family member is extremely severe. Previously, the largest duplication identified in SHFLD patients was 594 kb, and the minimum critical region only contained a single BHLHA9 [12]. Compared with the previously identified 594 kb duplication, the complicated duplication described here encompassed five additional genes, RFLNB, VPS53, FAM57A, GEMIN4, and DBIL5P (Fig. 3a). No involvement in limb development of VPS53, FAM57A, GEMIN4, and DBIL5P has been reported up to now, thus the severe phenotype in this family could not be explained simply by the duplication of these four genes. Although RFLNB has been implicated in regulating differentiation and proliferation of chondrocytes, the Rflnb knock-out mice are phenotypically normal [14], thus the role of RFLNB in limb malformation remains unknown. It is currently believed that overexpression of BHLHA9 is the major associated factor of SHFLD. BHLHA9 plays a significant role in limb development. It is expressed in the distal limb bud mesenchyme underlying the AER in mouse embryos [12]. BHLHA9 loss-of-function mutation results in limb malformation in human and animal models. Knocking down Bhlha9 in zebrafish leads to truncated pectoral fins [12], and Bhlha9-knockout mice show syndactyly [11]. Moreover, loss-of-function mutations in the BHLHA9 DNA-binding domain can cause autosomal-recessive inherited mesoaxial synostotic syndactyly (MSSD, OMIM 609432) [15]. No mutation was observed in the present family by sequencing the entire BHLHA9 coding region, and the inheritance pattern shown here was autosomal dominant, thus the severe phenotype was not caused by BHLHA9 loss of function. The BHLHA9 dosage is highly correlated to the long bone defect phenotype [10]. We identified a BHLHA9-containing triplication, while the phenotypes in the present family were more severe than a BHLHA9-containing triplication in Japanese populations [10]. Therefore, the severe phenotype in this family could not be simply explained by the increased copy number of BHLHA9 as well. Gene expression dosage can be regulated by multiple factors, such as a long-range cis-acting regulatory element. Disruption of long-range cis-acting regulatory elements has been demonstrated to be responsible for several limb malformations. For example, microduplications and mutations in the ZRS enhancer, a well-known long-range cis-element of Sonic hedgehog (SHH), is associated with triphalangeal thumb-polysyndactyly syndrome and preaxial polydactyly [16]. It is noteworthy that the ZRS enhancer is about 980 kb upstream of SHH, which is similar to the genomic rearrangement size found in this study. Therefore, it is possible that the genomic rearrangement region identified here contains a long-range cis-acting regulatory element of BHLHA9. Since the smaller duplication only includes BHLHA9 and nearby regulatory elements, it is assumed that gene expression is influenced mainly by the increased copy number of BHLHA9 and the nearby promoter and enhancers. However, since the large genomic rearrangement includes both long-range regulatory elements and BHLHA9, gene expression is also influenced by enhanced long-range regulation, and, as a result, the expression level is much greater and the phenotype is more severe. The long-range regulatory element may not be included in the previously reported duplicated regions, which could explain why Petit et al. and Klopocki et al. did not observe a correlation between the duplication size and phenotype severity [7, 12]. Further evidence should be provided to support this assumption, including additional severe SHFLD patients with large genomic rearrangements along with the direct identification of a long-range cis-acting regulatory element of the BHLHA9 gene. In addition, incomplete penetrance and phenotypic variability are often observed in SHFLD patients, such as observed in the family in this study. Particularly, a patient-to-carrier-to-patient transmission pattern was observed in the present family, suggesting the involvement of additional factors from the complicated genetic and environmental backgrounds. The highly variable expressivity and incomplete penetrance, particularly the asymptomatic carriers, imply that an individual with a 17p13.3 genomic rearrangement does not certainly present a malformation phenotype, which perplex the prenatal diagnosis and genetic counseling in the present family.