Sputum Proteomics in non-Tuberculous Mycobacterial Lung Disease

Abstract Background Non-tuberculous mycobacterial (NTM) infections are difficult to diagnose and treat. Biomarkers to identify patients with active infection or at risk of disease progression would have clinical utility. Sputum is the most frequently used matrix for diagnosis of NTM lung disease. Research Question Can sputum proteomics be used to identify NTM associated inflammatory profiles in sputum? Study Design and Methods Patients with NTM lung disease and a matched cohort of patients with chronic obstructive pulmonary disease (COPD), bronchiectasis (BE) and cystic fibrosis (CF) without NTM lung disease were enrolled from two hospitals in the UK. Liquid chromatography-tandem mass spectrometry was used to identify proteomic biomarkers associated with underlying diagnosis (COPD, BE, CF), the presence of NTM lung disease defined by ATS/IDSA criteria and severity of NTM. A subset of patients receiving guideline concordant NTM treatment were studied to identify protein changes associated with treatment response. Results We analysed 95 sputum samples from 55 subjects (21 BE, 19 COPD, 15 CF). Underlying disease and infection with Pseudomonas aeruginosa were the strongest drivers of sputum protein profiles. Comparing protein abundance in COPD, BE and CF showed 12 proteins were upregulated in CF including MPO, AZU1, CTSG, CAT and RNASE3 with 21 proteins downregulated including SCGB1A1, IGFBP2, SFTPB, GC and CFD. Across CF, BE and COPD, NTM infection (n=15) was not associated with statistically significant differences in sputum protein profiles compared to those without NTM. 2 proteins associated with iron chelation were significantly downregulated in severe NTM disease. NTM treatment was associated with heterogeneous changes in sputum protein profile. NTM patients with a decrease in immune response proteins had a subjective symptomatic improvement. Interpretation Sputum proteomics identified candidate biomarkers of NTM severity and treatment response, however underlying lung disease and typical bacterial pathogens such as P. aeruginosa are also key determinants of sputum proteome profile.