Mapping the landscape of the aged human brain for neurodegenerative disease models

Summary Aging is the main contributor to many human neurodegenerative diseases, such as Alzheimer’s disease (AD), that are common, progressive, and have limited therapeutic options. Our lack of understanding the biology of human brain aging remains a major challenge in the study of age-associated disorders including AD and other

neurodegenerative disorders. While multiple laboratory models such as flies, worms and mice, have uncovered major pathways in the biology of aging, translating these findings to humans is still incomplete, in part due to important cross-species differences. Over the past decade, human induced pluripotent stem cell (hiPSC) models

have emerged as an experimental platform to model many human diseases and have contributed great insights into AD and neurodegeneration broadly. However, one major caveat to hiPSC models is the fetal nature of the cells. Several methods have emerged to try and integrate ‘aging’ factors into hiPSCs or to directly

transdifferentiate ‘aged’ cells. However, these protocols often rely on atypical aging programs or lack the flexibility of the hiPSC system. One advance over the last 5-7-years is the advent of multi-omic data sets from human post-mortem brain. The vast amount of data generated by omics technology has great potential to fill the gap in our understanding

of brain aging and age-dependent, cell-type specific genetic programs. A major current challenge, however, is how to leverage these large, unbiased datasets to identify specific genes that regulate aging pathways. Manipulating candidate genes in a human neural cell experimental system would enable understanding and in

vitro modeling of cellular brain aging in a tractable experimental system. Such experiments may reveal targets that can be modified to improve aging phenotypes in human brain cells. In order to address these challenges, we have assembled a team of experts in explainable artificial intelligence (XAI) technology (S-I. Lee), human brain ‘omic studies (S. Jayadev), and hiPSC disease modeling

for AD (J. Young). We hypothesize that by applying XAI methods to human brain data sets, we can identify a tractable set of molecular drivers of brain aging. We further hypothesize that we can manipulate these drivers in hiPSC models using CRISPR technology to generate aging phenotypes in hiPSC-derived cells. In this two-

pronged proposal, we will first perform proof-of-concept experiments to modulate expression of XAI-identified genes in hiPSC-derived neurons and glia (microglia and astrocytes) and perform phenotypic assays to assess cellular hallmarks of aging (R21 phase). Next, we will increase the complexity of our model by integrating

additional omics layers, further developing and refining the XAI techniques and modulating candidate aging drivers in hiPSC-AD models in a multi-cellular context (R33 phase). These experiments will improve experimental platforms to study human brain aging and further identify pathways that may be developed therapeutically for

neurodegenerative disorders and human longevity.