Multi-Omics of the Iron Homeostasis Pathway in Patient Outcomes after Aneurysmal Subarachnoid Hemorrhage

Patient outcomes after aneurysmal subarachnoid hemorrhage (aSAH) are variable and healthcare providers are often unable to predict those who will do poorly and are in need of more intensive nursing management. This ancillary, longitudinal, observational, candidate gene association study investigated the hypothesis that epigenetic and genetic variability of genes in the iron homeostasis pathway may account for a proportion of variability in patient outcomes post-aSAH. This study was conducted using a two-tier design (targeted discovery and replication) and capitalized on an existing cohort of aSAH participants with genome-wide DNA methylation, genotype, and patient outcome data as well as stored biosamples. This study used group-based trajectory analysis (Aim 1), binary logistic regression (Aims 1 and 2), and Bayesian statistical methods (Aim 2) to assess the relationship between inferred DNA methylation trajectory groups (Aim 1) and tagging SNPs (Aim 2) from iron homeostasis candidate genes (n = 39) and acute (cerebral vasospasm [CV] and delayed cerebral ischemia [DCI]) and long-term (Glasgow Outcome Scale [GOS] and death at 3 and 12-months) patient outcomes post-aSAH. In Aim 1, we identified three DNA methylation sites in three candidate genes (Amyloid Precursor Protein [APP], STEAP3 metalloreductase [STEAP3], and Tumor Necrosis Factor [TNF]) with inferred trajectory groups significantly associated with patient outcomes in our discovery sample. Specifically, associations were identified between cg08866780 (APP) and GOS at 12 months (p = 0.0001) and death at 3 and 12 months (both p=0.0002); cg25713625 (STEAP3) and GOS at 3 and 12 months (p = 0.00005 and p = 0.0005, respectively) and death at 3 and 12 months (p = 0.0013 and p = 0.0015, respectively); and cg08553327 (TNF) with GOS at 3 months (p = 0.00001). Replication data are pending. For Aim 2, we identified two SNPs in two candidate genes (Ceruloplasmin [CP] and Cubilin [CUBN]) that were associated with patient outcomes in both the discovery and replication samples. Specifically, rs10904850 (CUBN) was significantly associated with DCI (mega-analysis combining targeted discovery and replication samples, p = 0.004) and rs17838831 (CP) with GOS at 3 and 12 months (mega-analysis p = 0.0004 and p = 0.057, respectively)