UCSF

Alzheimer’s Disease (AD) is a complex neurodegenerative disease resulting from both environmental and genetic risk factors that accounts for most dementia cases. AD is one of the top causes of death in the United States, and it is increasing in prevalence along with the growing aging population. AD has no cure, and there are limited therapeutics available to reverse it. Thus, it is of high priority to understand the underlying molecular mechanisms associated with AD progression to gain insight of ways to target it. Among the risk factors of AD are the apolipoprotein E (APOE) ε4 allele and sex. APOE4 has been identified as the largest genetic risk factor for AD, yet its molecular underpinnings are obscure. Additionally, sex differences have also been clinically documented in AD, but the molecular mechanisms explaining these differences remain elusive. To understand how APOE genotype and sex contribute to differing vulnerabilities in AD, we leverage publicly available transcriptomic datasets to explore APOE genotype-specific disease-related changes on a bulk and single-cell level, and sex-specific disease-related changes on a single-cell level.

In Chapter 1, we leverage human AD bulk RNA-sequencing (RNA-Seq) datasets spanning 7 brain regions (temporal cortex, cerebellum, dorsolateral prefrontal cortex, anterior prefrontal cortex, posterior superior temporal gyrus or Wernicke’s area, perirhinal cortex, and inferior frontal gyrus or Broca’s area) containing 494 AD and 262 non-demented controls. We performed a case versus control APOE4-stratified analysis in each brain region separately. We split each dataset into APOE4-negative (“E4NEG”: APOE3/3 (homozygous for allele ε3)) and APOE4-positive (“E4POS”: APOE3/4 (heterozygous ε3/ε4) and APOE4/4 (homozygous for allele ε4)) samples, compared AD to control samples within each subgroup, and examined disease-related gene expression and pathway changes in each subgroup. We identified new and previously studied transcriptomic changes based on the presence of APOE4 with some overlap across brain regions. While we observed neuroinflammatory pathways in all samples, we also observed an emphasis on stress response, hormone and receptor signaling, and epigenetic regulation in E4NEG samples, and an emphasis on metabolic changes, lipid metabolism, clearance and recovery from deleterious events, and ion, iron, and vitamin homeostasis in E4POS samples.

In the last decade, decreased costs and advances in high-throughput sequencing (HTS) technology and analytics tools have transformed the scientific field. With HTS, scientists have been able to generate bigger genomic datasets with unbiased insights in a timely fashion. In chapter 1, we took advantage of this by leveraging publicly available human AD bulk RNA-Seq datasets. However, these were traditional or “bulk” RNA-Seq datasets, which, although they provided enormous insights into the biological changes in AD, these insights were of the average biological changes, and as a result could not reveal more complex changes that explain the heterogeneity of AD. Fortunately, single-cell RNA-Seq emerged to identify and assess subpopulations of cells otherwise considered to be homogeneous. In chapters 2 and 3, we leverage publicly available human AD single-cell RNA-Seq datasets to increase our resolution and determine brain cell type vulnerabilities in AD with regards to APOE genotype and sex.

Following our bulk analysis, to elucidate more complex APOE genotype-specific disease-relevant changes masked by the bulk analysis, we leverage the first two single-nucleus RNA-Seq (snRNA-Seq) AD datasets from human brain samples, including nearly 55,000 cells from the prefrontal and entorhinal cortices (Chapter 2). We performed a case vs control APOE genotype-stratified differential gene expression (DGE) analysis and pathway network enrichment in astrocytes, microglia, neurons, oligodendrocytes, and oligodendrocyte progenitor cells of APOE3/3 and APOE3/4 samples. We observed more global transcriptomic changes in E4POS AD cells and identified differences across APOE genotypes primarily in glial cell types.

Next, to explore sex differences in AD, we also leveraged the same two snRNA-Seq datasets and utilized nearly 74,000 cells from human prefrontal and entorhinal cortex samples (Chapter 3). We performed a case vs control sex-stratified DGE analysis and pathway network enrichment in a cell type-specific manner like Chapter 2. In the prefrontal cortex, we observed sex-specific gene and pathway differences in AD most prominently in glial cells, and in the entorhinal cortex, we observed shared genes and pathways to be perturbed in opposing directions between sexes in AD relative to healthy state.

Ultimately, this dissertation identifies disease-relevant transcriptomic perturbations on a bulk and single-cell level that suggest differing mechanisms of neurodegeneration based on APOE genotype and sex. The findings here highlight the importance of incorporating APOE genotype and sex in future multiomic exploration of AD pathogenesis and progression, and will have implications for precision medicine approaches in the diagnosis and treatment of AD.