Identification of Non-Traditional Molecular Contributors to Cystic Fibrosis
MetadataShow full item record
Next generation sequencing has enabled the identification of patients with unambiguous recessive Mendelian phenotypes and incomplete genotypes, meaning zero or only one disease-causing mutation are detected by diagnostic sequencing. These patients serve as a rich source of unannotated disease variation, both within the primary disease locus and beyond. We studied two cohorts of cystic fibrosis (CF) patients with incomplete genotypes: one cohort had one CF-causing mutation in the primary locus CFTR after comprehensive molecular diagnostics and one cohort had zero CF-causing CFTR mutations. We developed a novel computational method combining machine learning splice site models with a human-designed algorithm for the selection of high confidence splice variants to identify CFTR deep intronic splice variants in “one mutation” CF patients. Our method is unique in assessing the splice potential of reference sequences and comparing variant sequences as a high quality filter. Candidate deep intronic cryptic splice variants were confirmed by in vitro analysis with expression minigenes containing relevant introns. Additionally, we identified CFTR cryptic splice variants masquerading as missense mutations and experimentally validated two “deep exonic” cryptic splice variants, indicating that exonic cryptic splicing may occur more frequently than commonly appreciated. We performed exome sequencing on CF families with zero CFTR mutations to identify candidate intergenic molecular contributors to CF. Putative deleterious alleles in CA12 encoding carbonic anhydrase (CA) XII segregated with disease in two unrelated families exhibiting an atypical CF phenotype of elevated sweat chloride concentrations and hyponatremic dehydration. Bronchiectasis was identified in an adult Caucasian proband in spite of normal pulmonary function testing, suggesting the early stages of obstructive lung disease. The adult proband carried two mutations which were shown to result in aberrant RNA splicing by analysis of patient nasal epithelia. In the second family, two affected siblings carried a homozygous missense in an essential zinc-coordinating residue of CA XII. This mutation was shown in vitro to cause a complete loss of CA enzymatic activity. CA XII localized to the sweat gland and airway epithelia in normal primary tissue, providing further evidence for loss of CA XII function as a cause of CF.