Blood mRNA biomarkers for detection of treatment response in acute pulmonary exacerbations of cystic fibrosis

Abstract

Background Acute pulmonary exacerbations accelerate pulmonary decline in cystic fibrosis (CF). There is a critical need for better predictors of treatment response. Objective To test whether expression of a panel of leucocyte genes directly measured from whole blood predicts reductions in sputum bacterial density. Methods A previously validated 10-gene peripheral blood mononuclear cell (PBMC) signature was prospectively tested in PBMC and whole blood leucocyte RNA isolated from adult subjects with CF at the beginning and end of treatment for an acute pulmonary exacerbation. Gene expression was simultaneously quantified from PBMCs and whole blood RNA using real-time PCR amplification. Test characteristics including sensitivity, specificity, positive and negative predictive values were calculated and receiver operating characteristic curves determined the best cut-off to diagnose a microbiological response. The findings were then validated in a smaller independent sample. Results Whole blood transcript measurements are more accurate than forced expiratory volume in 1 s (FEV1) or C reactive protein (CRP) alone in identifying reduction of airway infection. When added to FEV1, the whole blood gene panel improved diagnostic accuracy from 64% to 82%. The specificity of the test to detect reduced infection was 88% and the positive predictive value for the presence of persistent infection was 86%. The area under the curve for detecting treatment response was 0.81. Six genes were the most significant predictors for identifying reduction in airway bacterial load beyond FEV1 or CRP alone. The high specificity of the test was replicated in the validation cohort. Conclusions The addition of blood leucocyte gene expression to FEV1 and CRP enhances specificity in predicting reduced pulmonary infection and may bolster the assessment of CF treatment outcomes.