This study identified a number of novel and recurrent mutations in the ABCC6
gene in a large multinational case series of patients with PXE. In all, 82 distinct mutations, of which 39 were novel, previously unreported, were identified in 239 probands. The detection rate of 66% is less than that reported by other groups, and potentially reflects (1) the lower detection rate when using dHPLC for screening,41,42
(2) the inability of the methods used to detect large deletions including heterozygous loss of entire exons or the entire ABCC6
gene, and (3) presence of mutations in regions such as the 5′ regulatory elements, 3′ untranslated region and central intronic sequences of ABCC6
, which were not analysed in this study. Finally, there is the possibility that mutations in genes other than ABCC6
can result in a PXE phenotype, although no concrete evidence for this possibility currently exists. The recurrent mutations R1141X and del23–29 were the most prevalent, consistent with previous reports. In our international case series, mutations in exons 24 and 28 together with del23–29 accounted for the majority of the mutations detected (71.5%) in the ABCC6
gene. Several other recurrent mutations were also identified and comprised around 20% of the total number of mutations.
The NBF domains of the MRP6 protein harboured a large number of missense mutations (22/49), reflecting the strong amino acid sequence conservation in this region, necessary to maintain the function of the protein (fig 2). In addition, intracellular domain 8 harboured several missense mutations (8/49), suggesting the importance of sequence conservation in this region of the protein and perhaps a function that has yet to be identified.
Genotype–phenotype analysis of the 134 patients for whom both ABCC6
mutations and full phenotypic data were available failed to reveal any definitive correlations. The correlation of age of PXE diagnosis with the predicted consequence of the mutation (protein or no protein), as reported by Schulz et al
could not be confirmed. As there is no empirical confirmation of whether there is functional MRP6 protein as a result of each of the mutations, the accuracy of this conclusion rests on the validity of the presumption that the mutation either does or does not lead to functional MRP6 protein. Molecular studies of three different MRP6 missense mutations (V1298F, G1302R and G1321S) performed in vitro found that none of these three mutations resulted in ATP‐dependent substrate transport although ATP binding was normal.11
Although these in vitro data do not address whether the proteins bearing these missense mutations are actually formed in the cell and inserted properly into the membrane, it confirms the hypothesis that a missense mutation in a critical portion of the molecule can completely ablate the function of the protein and result in a phenotype as severe as if no protein was present at all.
Finding no genotype–phenotype correlations in complex mendelian disorders, such as metabolic diseases, is not uncommon.43,44
This lack of association suggests that no simple relationship exists between the type and position of mutations, the mutations themselves and the severity of disease. Thus, it is not possible from these data to predict the type or severity of phenotypic manifestations from mutation studies themselves. Because the type of phenotypic outcome and its severity cannot be predicted from genotype data, all patients with PXE should be screened regularly for potential serious and possibly life‐threatening complications regardless of their genotype.
An interesting observation was the absence of macroscopic skin lesions in four patients (9, 58, 131 and 171), although skin biopsy revealed typical histological characteristics of PXE. Patient 9 had significant ophthalmological complications, suggesting that this patient was not a carrier. In three of these patients, a complete genotype was determined, confirming the clinical diagnosis and emphasising that skin features, although present in the majority of patients with PXE, are not always mandatory for the diagnosis.
This study is limited by possible selection bias: participants volunteered and were contacted through the support group. Thus, it is skewed towards people who seek support, which might select for people with a more severe phenotype. In addition, more women than men responded, although that gender ratio is reflective of the PXE population in general. Furthermore, the phenotypic data were obtained through self‐reports, making their accuracy somewhat questionable, even though matching with medical records in a subgroup yielded a high correlation. Nonetheless, the results of this study, along with results of all previous genotype–phenotype correlation studies, indicate strongly that a straightforward and clinically usable correlation does not exist, as could be expected from the nature of this disorder. However, because of the significant clinical relevance, further work is needed to determine if there are subcellular phenotypes that correlate with genotypes in PXE.42
Although mutation detection in PXE has not been shown to have prognostic value, presymptomatic testing in families with history of PXE can provide early diagnosis and may be of value in surveillance for development of disease. In addition, it will probably prove to be of value in genetic counselling as well as in diagnosis of atypical cases or apparent phenocopies.