Abstract Rationale Protein and DNA methylation markers are readouts of environmental exposure that are influenced by genetic determinants. COPD and pulmonary fibrosis (PF) proteomics analyses have revealed a subset of inflammatory proteins differentially associated with each disease. Integration of methylation and proteomics may reveal new insights into disease pathways. Methods Lung Tissue Research Consortium (LTRC) cohort plasma Olink proteomics for 451 proteins differentially expressed in COPD versus PF were integrated with DNA methylation from the same blood sample. Methylation data was generated using the Illumina EPIC array. COPD was defined as GOLD spirometry grades 2-4 and PF was defined based on lung pathology and clinical assessment. Epigenome-wide association analysis for each protein was performed using limma, adjusting for age, sex, smoking status (AHRR methylation), smoking pack-years, ancestry, BMI, and cellular heterogeneity. Statistical significance was set at a Benjamini 0.05. Pathway analysis was performed using STRING. CpG sites were ranked based on the number of protein associations. Hierarchical clustering was performed on a binary matrix specifying whether a CpG was associated with a given protein, limited to CpGs with at least 20 protein associations. Results Of the 451 proteins queried, 306 have statistically significant associated CpG sites. 59 proteins were associated with CpG sites within ±1 megabases of the gene transcription start site and 18 proteins had associations within ±10 kilobases. A subset of CpG sites demonstrated multiple protein associations (Figure 1) and these associations formed four clusters. One cluster was enriched with cytokine activity proteins. Five of the CpGs in this cluster mapped to sites near miR-21, a microRNA previously associated with PF, while three sites mapped to sites in RUNX1, a transcription factor implicated in pulmonary disease. The set of proteins in another cluster were enriched in coagulation and fibrin clot formation. A third cluster was enriched with cell adhesion proteins. The last cluster was enriched in extracellular proteins. This suggests that specific sets of CpGs may be involved in control mechanisms for pathways involved in COPD and PF. Conclusions Integration of COPD- and PF-associated protein expression with DNA methylation revealed enrichment for cell signaling, cytokine activity, and wound healing pathways. Integrative omics may reveal new therapeutic targets for smoking-related lung diseases. This abstract is funded by: NIH NHLBI T32 HL007427
Madha et al. (Fri,) studied this question.