Discussion

The increased allele frequency of potentially pathogenic CFTR variation we report may be partially attributed to the smaller sample size of the deeply characterized SARP cohort, but is consistent with prior reports of a higher frequency of potentially pathogenic CFTR variants, especially F508del, in individuals with asthma from the general population.(2, 3) In NHW individuals from SARP, the allele frequency for F508del was >5-fold higher than that reported in European general populations (Table 2 ).(5) Although we hypothesized biologic plausibility, we found no evidence that CFTR variation influences asthma severity. The lack of an association could result from a lack of pathogenicity data. For example, we identified 91 coding variants in the exons of CFTR of which we were only able to find 62 catalogued in CFTR2 or by ACMG. Another potential reason for a lack of a severity association is that determinants of asthma are heterogeneous and complex stemming from numerous genes potentially masking anyCFTR -specific effects in SARP.
Twelve African American individuals carried two African descentcis potentially pathogenic variants in perfect linkage disequilibrium that were both CFTR variant targets for CFTR modulators. This phenomenon of potentially pathogenic cis-occurring variants has been described previously in the literature and results in a complex allele with additive effects for these and other European descent variants associated with CF or congenital bilateral absence of the vas deferens(CBAVD).(6) Interestingly, when associated with CBAVD or mild CF there is usually an additional in cis variant (c.601G>A; p.Val201Met; legacy: V201M) that we did not find in these individuals.(6) Even though we found a smaller number of potentially pathogenic CFTR variants in AA individuals, we found a substantial number of African ancestry-specific variants in AA that we were unable to categorize because current databases primarily consist of data from European descent white individuals. Of the variants we identified with confirmed or potential pathogenicity, eight were found in AA while 21 were found in NHW individuals. A similar lack of identifiable pathogenic variation was described in both individuals with clinical CF from Puerto Rico and the Dominican Republic.
Four non-Hispanic white individuals with asthma of varying severity were either homozygous or compound heterozygous for potentially pathogenic CFTR variation of which three would have qualified for CFTR modulator therapy in the appropriate clinical setting. None of the potentially pathogenic genotypes we identified consisted of two CF-causing variants, but always had at least one VVCC. In the absence of concomitant demonstration of measured CFTR dysfunction, these individuals would not meet criteria for a confirmed diagnosis of CF. However, the clinical significance of these CFTR genotypes might relate to CF masquerading as asthma in individuals with CF-asthma overlap or as an asthma disease-modifying genotype in the absence of CF, and further evaluation for CF would be warranted in such individuals. The analysis of next-generationCFTR sequencing data in larger asthma and general populations will be required and are currently underway ( https://topmed.nhlbi.nih.gov/) to further characterize and confirm our findings.