Epigenetic training of human bronchial epithelium cells by repeated rhinovirus infections

Background Humans are subjected to various environmental stressors (bacteria, viruses, pollution) throughout life. As such, an inherent relationship exists between the effect of these exposures with age. The impact of these environmental stressors can manifest through DNA methylation (DNAm). However, whether these epigenetic effects selectively target genes, pathways, and biological regulatory mechanisms remains unclear. Due to the frequency of human rhinovirus (HRV) infections throughout life (particularly in early development), we propose the use of HRV under controlled conditions can model the effect of multiple exposures to environmental stressors. Methods We generated a prediction model by combining transcriptome and DNAm datasets from human epithelial cells after repeated HRV infections. We applied a novel experimental statistical design and method to systematically explore the multifaceted experimental space (number of infections, multiplicity of infections and duration). Our model included 35 samples, each characterized by the three parameters defining their infection status. Results Trainable genes were defined by a consistent linear directionality in DNAm and gene expression changes with successive infections. We identified 77 trainable genes which could be further explored in future studies. The identified methylation sites were tracked within a pediatric cohort to determine the relative changes in candidate‐trained sites with disease status and age. Conclusions Repeated viral infections induce an immune training response in bronchial epithelial cells. Training‐sensitive DNAm sites indicate alternate divergent associations in asthma compared to healthy individuals. Our novel model presents a robust tool for identifying trainable genes, providing a foundation for future studies. This study aimed to identify trainable genes responsive to repeated HRV infections by examining changes in DNA methylation and gene expression in human bronchial epithelial cells (BEAS‐2B). By using an innovative experimental design (DoE) and integrating transcriptome and methylation data, the developed method successfully served as a prediction model. Key findings included the identification of 77 trainable genes (TGs). The tracking of TGs methylation sites within an asthmatic pediatric cohort revealed differential associations with disease status and age. Additionally, distinct methylation sites associated with asthma and age highlight potential biomarkers for early dia