Analysis of sputum methylation data in Asthma and COPD

We evaluated the applicability and usability of whole-genome methylomics of sputum samples in molecular profiling of chronic inflammatory lung diseases. Genomic DNA was purified from sputum samples of subjects with Asthma, COPD as well as healthy controls and analyzed on the Illumina Infinium HumanMethylation 450k platform. Induced sputum samples were sourced from the German cohort studies ALLIANCE and COSYCONET. We selected a total of 9 samples from non-smoking subjects with controlled (as indicated by Asthma Control Test score, ACT), eosinophilic/Type-2 Asthma (based on blood/sputum eosinophil counts). A total of 10 samples from subjects with COPD (>12 months after smoking cessation) as well as 10 samples from healthy controls (w/o any pulmonary or systemic-inflammatory disorder) were included in the analysis. Intentionally, we selected sputum samples that were stored/preserved in two different ways in order to explore effects on downstream analysis and simulate challenges in larger-scale biobanking studies (mainly regarding RNA quality in concomitant transcriptome analysis): 1) storage in RLT lysis buffer (Qiagen) at -80 °C following sputum processing. 2) storage in HOPE medium (HEPES-glutamic acid buffer-mediated organic solvent protection effect; DCS innovative diagnostic systems, Hamburg, Germany) after sputum processing, followed by paraffin-embedding and storage at 4°C. Genomic DNA was extracted from the samples, bisulfite converted and hybridized to the Illumina Infinium HumanMethylation450 BeadChip. Parallel transcriptome analysis was performed on extracted RNA. From the initial sample pool, methylation analysis was successfully performed on all Asthma and COPD samples as well as 7/10 control samples. Downstream in silico analysis focused on methylation/transcription differences between Asthma/COPD and healthy controls, respectively. Information availbale about sputum differential cell counts (consisting of alveolar macrophages, neutrophils, eosinophils, lymphocytes, monocytes, ciliated cells and squamous cells) was used to implement and evaluate an in silico model correcting for variation in cell counts across disease conditions.