Clinical and molecular analysis of Williams-Beuren syndrome : a search for factors determining clinical variability in a contiguous deletion syndrome

Williams-Beuren syndrome (WBS) is clinically characterized by a distinctive "elfin" facial appearance, short stature, mental retardation, a unique cognitive profile, friendly outgoing personality, and various congenital heart malformations with supravalvular aortic stenosis (SVAS) being the most common. Clinical features seen with less frequency include infantile hypercalcaemia, hoarse voice, hyperacousis, renal hypoplasia, clinodactyly of the 5th finger, and teeth hypoplasia. The syndrome arises as a consequence of a 2 cM deletion of 7qll.23 in 90 - 95% of all clinically typical cases reported to date. This region includes a number of genes that include elastin (ELN) and LEVI kinase 1 (LTMK1). ELN mutations have been found in isolated cases of SVAS while LIMK1 has been implicated in the unique WBS cognitive profile. However, not all patients present with congenital cardiovascular malformations and the severity to which cognition is impaired in WBS individuals varies with each patient. This variability of clinical phenotype presentation in clinically typical cases has not been explained. In order to identify the underlying cause of this variability, clinical and molecular data were analyzed from 108 clinically diagnosed WBS cases. A deletion of 7qll.23 was identified in 85 cases and extended from D7S489U to D7S1870 (including these markers) in the majority of cases. Only two cases were found to have a more distal deletion that also encompassed D7S489L. In these 85 cases, no influence on phenotypic variability was found for deletion extent, parental origin of deletion, ELN and LLMK1 polymorphisms, or gender using a Bonferroni corrected level of significance (α'=4.6x10⁻⁴). Nonetheless, a number of interesting 'suggestive' associations with p<0.05 were identified. Correlations between clinical features of WBS were examined to identify those features which may have a common etiology. Identifying correlates may also identify specific phenotypic features that can be used as a prognostic tool. This has revealed significant correlations between: 1) low birth weight (<10th centile) and low weight at the time of physical examination (