S18 The sputum proteome and its relationship to cystic fibrosis lung disease: using global proteomics to develop clinically useful biomarkers

Introduction There is potential for protein biomarkers to assist with clinical decision making in cystic fibrosis (CF) patients, such as when to commence and cease antibiotic therapy. Several small proteomic studies in CF populations have identified large numbers of proteins within respiratory samples, with some correlating with measures of lung function. Here initial data is presented from a larger cohort, with a comprehensive evaluation of protein abundance relative to baseline and longitudinal lung function. Methods Spontaneous sputum samples were collected from CF subjects (n=37) and induced from healthy volunteers (HV) (n=33). Bottom-up shotgun proteomic analysis of samples was undertaken using liquid chromatography-mass spectrometry. Spirometry was noted at baseline, and for the previous 12 months. For comparison of relative protein abundance between cohorts, principal component analysis (PCA) and correlative statistics were undertaken. Results Principal component analysis (PCA) highlighted significant differences between the CF and HV sputum proteome, with large increases in inflammatory proteins, predominantly neutrophil granulocyte proteins. There were significant differences in the sputum proteome between CF patients with normal/mild (FEV1% predicted ≥ 70%, n=4) and severe lung disease (FEV1% predicted ≥ 70%, n=5). The top ten most influential proteins from this PCA were further examined. Within the entire CF cohort (n=37), seven of these ten proteins significantly correlated (p Discussion These data confirm findings from previous smaller studies that differences in the sputum proteome relate to baseline severity of lung disease. However, it does not appear to relate to longitudinal changes in lung function over 12 months. A biomarker might be only able to inform over shorter time periods, potentially because the proteome is in a state of flux. Further work is required to evaluate if longitudinal assessment of the proteome allow prediction of FEV1% decline, or if proteome changes are predictive of a pulmonary exacerbation.