Identification of brain aging subtypes in multiple brain regions from cognitively normal individuals

Abstract

AbstractBackgroundAlthough late‐onset Alzheimer’s disease (LOAD) occurs in older individuals (age 65 or older), multiple lines of evidence suggests that the dysregulation of brain functions initiates decades earlier. Through examining multiple large‐scale human brain transcriptomic data, we demonstrated that cognitively normal brain aging from donors in their 40s‐60s can group into several subtypes based on their hippocampal transcriptomes. One subtype represents healthy brain aging and some evidence suggests these donors may remain cognitively normal in much older ages, while some brain aging subtypes showed gene expression changes highly similar to those observed in AD, supporting molecular changes linked to AD have occurred much earlier than the manifestation of cognitive impairment. However, whether the brain aging subtype is a common phenomenon across multiple brain regions or very specific to the hippocampus remains unknown.MethodWe compiled and studied transcriptomic data in another commonly studied human brain region (prefrontal cortex) from donors who were cognitively normal. We performed hierarchical clustering in these samples to infer subtypes, differential gene expression analysis was done between samples from different subtypes. We annotated and compared these DEGs with other AD gene expression signatures.ResultWe discovered that aging subtypes could also be observed in the prefrontal cortical region. Similarly to the hippocampus, the differentially expressed genes among aging subtypes also highly overlapped with AD gene signatures. By comparing aging subtypes defined from hippocampal vs. cortical transcriptomes from the same individuals, we found that an AD similar hippocampal sample does not necessarily correspond to AD similar cortical sample in the same donor, suggesting the brain regional complexity of the disease.ConclusionOur analysis of the transcriptomic data from multiple brain regions from cognitively normal individual supports that gene regulation changes related to AD occurs decades earlier than the manifestation of clinical phenotype. Understanding the molecular changes in these early changes in multiple brain regions could be critical to gain new insights into the mechanisms of AD initiation.