Inflammaging in Major Depressive Disorder: Associations between markers of inflammation and brain ageing in depression across two large community-based cohorts

BackgroundMajor Depressive Disorder (MDD) is associated with accelerated ageing trajectories including functional markers of ageing, cellular ageing and markers of poor brain health. The biological mechanisms underlying these associations remain poorly understood. Chronic inflammation is also associated with advanced ageing; however, the degree to which long-term inflammation plays a role in ageing in MDD remains unclear, partly due to difficulties differentiating long-term inflammation from acute cross-sectional measures. MethodsHere, we use a longer-term measure of inflammation: a DNA methylation-based marker of C-reactive protein (DNAm CRP), in a large cohort of individuals deeply phenotyped for MDD (Generation Scotland, GS; N=804). We investigate associations between DNAm CRP and serum CRP using linear modelling with two brain ageing neuroimaging-derived phenotypes: (i) a machine learning based measure of brain-predicted age difference (brain-PAD) and (ii) white matter hyperintensities (WMH). We then examine inflammation by depression interaction effects for these brain ageing phenotypes. We sought to replicate findings in an independent sample of older community-dwelling adults (Lothian Birth Cohort 1936, LBC1936; N=615). ResultsDNAm CRP was significantly associated with increased brain-PAD ({beta}=0.111, p=0.015), which was replicated in the independent sample with a similar significant effect size ({beta}=0.114, p=0.012). There were no associations between the inflammation markers and WMH phenotypes in the GS-imaging sample, however in the LBC1936 sample, DNAm CRP was significantly associated with both Wahlund infratentorial ({beta}=0.15, PFDR= 0.006) and Fazekas deep white matter hyperintensity scores ({beta}= 0.116, PFDR=0.033). There were no interaction effects between inflammation and MDD in either cohort. ConclusionsThis study found robust associations between a longer-term marker of inflammation and brain ageing as measured by brain-PAD, consistent across two large independent samples. However, we found no evidence for interaction effects between inflammation and MDD on any brain ageing phenotype in these community-based cohorts. These findings provide evidence that chronic inflammation is associated with increased brain ageing, which is not specific to MDD. Future work should investigate these relationships in clinical samples including with other inflammatory biomarkers and should furthermore aim to determine causal directionality.