Integrated DNA methylome and transcriptome signatures of lung CD4 + T-cells in sarcoidosis

Rationale Activated pulmonary CD4 + T lymphocytes are a critical part of sarcoidosis pathogenesis. To identify molecular changes associated with sarcoidosis risk primarily and progression secondarily, we profiled and integrated DNA methylome and transcriptome data in lung CD4 + T-cells. Methods Bronchoalveolar lavage (BAL) CD4 + T-cells were isolated from 29 sarcoidosis (16 progressive sarcoidosis (SarcP) and 13 non-progressive sarcoidosis (SarcNP)) cases and 19 healthy controls. mRNA was sequenced and DNA methylation profiled on Illumina HumanMethylationEPIC BeadChips. Linear models were fitted to test for effect of diagnosis, adjusting for age and sex. Methylation models were adjusted for unknown batch effects, inflation and bias. We used Data Integration Analysis for Biomarker discovery using Latent cOmponents (DIABLO) to integrate methylome and transcriptome datasets and validated targets using single-cell RNA sequencing (RNA-seq). Measurements and main results Using transcriptome-wide significance, differentially expressed (DE) transcripts were identified in participants with sarcoidosis (39), SarcNP (23) and SarcP (7), compared with controls. Using genome-wide significance, 37 CpGs and 47 regions were identified as differentially methylated (DM) in sarcoidosis participants, compared with controls. Focusing on DE genes, we identified DM CpGs in 11 genes associated with sarcoidosis, 3 with SarcNP and 7 with SarcP. In DIABLO, the first latent component distinguished cases and controls based on transcripts and CpGs associated with T-cell signalling, T-cell differentiation, apoptosis and epigenetic regulation. Among the DE/DM transcripts identified by integrative omic analysis and validated by single-cell RNA-seq is interleukin 18 receptor accessory protein ( IL18RAP ). Conclusions Integrative methylome–transcriptome profiling of lung CD4 + T-cells identified coordinated immune and epigenetic alterations, including IL18RAP , as candidate biomarkers.