Immunophenotype and gene signature of long-term survivors with normal allograft function after lung transplantation

Background: The development of chronic lung allograph dysfunction (CLAD) is the leading limitation of long-term survival with good allograft function (LTS) after lung transplantation (LT). The aim of this study was to identify leukocyte subpopulations and to develop a gene model which allows the discrimination of LTS patients. Methods: The cell markers and mRNA expression levels were compared between LTS (n=30) and CLAD patients (n=30) using flow cytometry and microarray techniques. Gene classifiers were built using supervised machine learning methodology and the expression of selected classifier-genes was confirmed by RT-qPCR. Results: In LTS patients, the percentages of CD14^highCD16^- monocytes, CD56⁺CD16^- NK cells, CD4^-CD8^- αβ T cell subset and CD62L⁺granulocytes were elevated whereas the percentage of Vδ1⁺γδ T cell subpopulation was significantly decreased. The computational process led to the identification of the 25 most relevant genes for LT recipients classification. LASSO method with the 25 classifier genes yielded the optimal results with an Area Under the Curve of 0.87 (sensitivity 0.79 and specificity 0.80). RT-qPCR analysis confirmed the differential expression of 17/18 of the genes selected from the microarrays. The combination of these markers in a model could ultimately improve classification performance. Conclusion: This study identifies a set of lymphocyte subsets and genes associated with LTS after LT which should be validated in an independent LT cohort. The combination of these markers may have utility as a medical tool for safe immunosuppression minimization in LT population. Study financed by ISC III (PI13/01076), FEDER, FUCAP, Astellas, Novartis and Chiesi.