We describe three individuals with a unique skeletal phenotype, associated with novel heterozygous missense variants of the PBX1 homeobox domain. PBX1 shows high scores of being loss-of-function intolerant (pLI = 1), and missense Z (3.83) in the constraint metrics. Given the different clinical phenotype of previously reported individuals with PBX1 haploinsufficiency, the molecular pathogenesis of these missense variants would be hypomorphic effects with or without dominant-negative mechanism.
The variant p.Arg290Leu of patient 3 occurs at the same amino acid position p.Arg290Trp reported by Ruscitti et al.[14]. The patient described by Ruscitti et al. had cleidosternal dysostosis, pulmonary hypoplasia, complex cardiovascular anomalies, and intestinal malrotation, suggesting PBX1 missense variants may cause a spectrum of abnormalities in the shoulder girdle. In a previous study by Mathiasen et al., it was observed that the substitution of p.Arg290 into Alanine in the highly conserved homeodomain of PBX1 resulted in a marked reduction in DNA binding activity [15]. Based on the Mathiasen’s results combining their own results, Ruscitti et al. suggested that the structural change of homeobox domain could lead to the loss of function of PBX1 in their patient with p.Arg290Trp [14]. In line with these results, it is assumed that our patients’ phenotypes, with common skeletal manifestations and tightly clustered variants, may be caused by a functional hypomorphs of PBX1.
The skeletal features were unique, and the hallmarks included clavicular dysplasia with coracoclavicular fusion and dysplastic ischia with premature fusion of the ischiopubic synchondrosis. The scapular wing and thorax were mildly hypoplastic. Other common phenotypes include a variable degree of developmental delay, respiratory failure associated with laryngomalacia or tracheomalacia, and minor facial dysmorphism. Unlike previously reported patients with PBX1 pathogenic variants, the three patients did not have CAKUT.
PBX1 has a crucial role in embryonic development. Pbx1-deficient mice (Pbx1-knockout mice) show aplasia of multiple organs and defects of the axial and appendicular skeleton resulting from abnormalities in chondrocyte proliferation and differentiation [3]. The skeletal abnormalities include severely hypoplastic sternum, vertebral and rib malformations, hypoplastic scapulae and clavicles, and a malformed pelvic girdle. Notably, the distal fore- and hindlimbs were intact [3]. The present individuals had malformation of the shoulder and pelvic girdles. However, the skeletal phenotype of our patients was much milder than that of Pbx1-knockout mice. The major abnormalities were confined to the clavicle and ischium. However, the distal clavicular tapering and coracoclavicular ankylosis in the present patients may recapitulate “the attenuated clavicle” [3] and “the coracoid expansion” [2] in Pbx1-knockout mice, respectively. In Pbx1-knockout mice, the ischium was described to be rudimentary [3]. In the present individuals, the ischium appeared thick, and may have shown early ossification as compared with that of Pbx1-knockout mice. The genetic interaction between Pbx1 and Emx2 has a crucial role in shoulder girdle development in interacting with homeodomain proteins [2]. The alteration in the interaction of Pbx1 and Emx2 resulting the expression reduction of Sox9, leading to pelvic girdle malformation [16]. In line with these results, the missense variants of the homeobox domain in PBX1 could alter the function of the PBX1-EMX2 complex, leading to the unique phenotype of our patients.
However, several questions remain. The first is why variants associated with scapuloclavicular and ischial dysplasia are concentrated in the last portion of the homeobox domain of PBX1 (Fig. 2). Skeletal evaluation of patients with homeobox domain variants is limited, and additional patients would need to be added to confirm the relationship between skeletal features and missense variants. Second is why patients with haploinsufficiency of PBX1 do not show overt skeletal abnormalities, including scapulae [4, 7]. There are a few reports that brought attention to alterations of the skeleton including the shoulder girdle in PBX1 haploinsufficiency. However, the skeletal changes of the shoulder girdle are so mild that they are essentially different from what we reported in this report. The correlation between the skeletal abnormalities and variant effects is likely to be incomplete. Finally, pathogenic variants of EMX2 have never been reported to cause skeletal dysplasia in humans [17]. Comprehensive skeletal evaluation in PBX1-related diseases is needed to determine the functional impact of PBX1 and its partners. Further study with Pbx1 knock-in mice with missense variants will be required.
In summary, we described three patients characterized by broad clavicles and ischia, scapuloclavicular fusion and respiratory failure due to laryngomalacia, but without CAKUT phenotype, associated with the missense variants in the homeobox domain of PBX1. Further radiographic analysis of patients with PBX1-related disorder and the generation of animal models with missense variants in the PBX1 homeobox domain will provide insights into the underlying mechanism of cleidoischial dysostosis.