Not all saliva samples are equal: The role of cellular heterogeneity in DNA methylation and epigenetic age analyses with biological and psychosocial factors

Saliva is widely used in biomedical population research, including epigenetic analyses to investigate gene-environment interplay and identify biomarkers. Its minimally invasive collection procedure makes it ideal for studies in pediatric populations. Saliva is a heterogenous tissue composed of immune and buccal epithelial cells (BEC). Amongst the many epigenetic marks, DNA methylation (DNAm) is the most studied in human populations. DNAm profiles are often highly cell type (CT)-specific. CT composition can drive salivary DNAm associations with environments or health as well as epigenetic age acceleration (EAA), which is the discrepancy between chronological and biological age derived from DNAm. To address this, reference-based CT deconvolution and statistical adjustment with estimated CT in DNAm analyses have become a common practice. However, it remains unclear how different CT reference panels-constructed from adult versus pediatric samples-affect DNAm results. Additionally, whether DNAm and EAA associations in saliva primarily originate from immune cells or BECs, or if they persist across saliva samples despite varying CT proportions, still requires more investigations. The current study used salivary DNAm samples obtained from 529 children (mean age=7.26 years, SD=0.26 years) in a community-based cohort, the Family Life Project. Our results demonstrated that the child reference panel outperformed the adult one based on goodness of fit measure and highlighted the impact of estimated CT discrepancies across reference panels on DNAm associations. Upon stratifying the salivary DNAm samples into three subsamples-primarily BECs, primarily immune cells, and an approximately equal mix of both, we found significantly different EAAs across stratified samples when CT proportions were not accounted for. In both the contexts of DNAm and EAA associations, we detected stronger effects of cotinine concentrations, a tobacco smoke-exposure biomarker, in the subsample with primarily immune cells. We discussed the implications of our findings for the interpretation and replication of epigenetic research involving pediatric saliva samples.