The ABCA3 (ATP-binding cassette transporter A3) protein transports phospholipids from the cytoplasm into lamellar bodies in alveolar type II cells (AEC2s) and is critical for pulmonary surfactant synthesis and function. Biallelic pathogenic ABCA3 variants cause diverse pulmonary phenotypes, including lethal neonatal respiratory distress syndrome, childhood interstitial lung disease, and idiopathic pulmonary fibrosis (1–5). Current therapies, including steroids, hydroxychloroquine, and azithromycin, are nonspecific and have limited efficacy (3). Lung transplantation remains the definitive treatment option but is associated with 50% 5-year mortality and significant morbidities (6). The majority of more than 300 diseaseassociated ABCA3 variants identified to date are private. Among the fewer than 10% of disease-associated variants that have been functionally characterized in vitro, two mechanistic classes, disruption of intracellular trafficking and disruption of phospholipid transport into the lamellar bodies, have been described (7, 8). Although a genotype–phenotype correlation exists between biallelic frameshift or nonsense variants and neonatal respiratory distress syndrome (RDS) and death before 1 year of age without lung transplant, the disease presentation, severity, and course are difficult to predict for missense variants (2, 3). The similarities between ABCA3 and CFTR (encoded by ABCC7) and the clinical success of personalized therapies for patients with cystic fibrosis (9) support the potential for development of variant-specific therapies for the treatment of ABCA3 deficiency. A major barrier for developing such therapies for ABCA3 deficiency is the lack of a suitable in vitro system that mimics the biology and function of ABCA3 in AEC2s with a suitable readout for normal ABCA3 function amenable to highthroughput screening.

In this issue of the Journal, Forstner and colleagues (pp. 382–390) used a human pulmonary epithelial cell line (A549) and machine-learning algorithms to develop a phenotypic assay to detect morphologic differences between stably transfected cells expressing wild-type ABCA3 or missense variants (10). They next used this phenotypic assay to screen 1,280 US Food and Drug Administration (FDA)-approved small molecules and identified cyclosporine as a corrector of several ABCA3 mutants that disrupt intracellular trafficking. The authors then validated the results of the phenotypic cell-based assay using established functional small format assays (7, 8). Importantly, this phenotypic screen was able to distinguish between mutants that disrupted intracellular trafficking and those that impaired phospholipid transport. Notably, cyclosporine treatment did not correct all trafficking mutants, demonstrating that, like CFTR, mutant-specific therapies may be necessary despite similar mechanisms of variant-encoded disruption of ABCA3.