Transcriptome analysis of zebrafish genetic models to reveal early molecular drivers of Alzheimer’s disease

Alzheimer’s disease (AD) is a complex neurodegenerative disease that still evades effective treatment. Developing treatments to slow or prevent AD will require a detailed understanding of the early stages of AD at the molecular level. Unfortunately, the pathogenesis of AD progresses silently over decades and, in the case of genetically inherited familial AD, may involve subtle changes starting during young adulthood. The difficulties associated with studying human brains at young ages have meant that transcriptome analyses of accurate animal models are essential. This is the approach taken in the work here, which uses a data-driven, bioinformatics-led approach to analyse brain transcriptomes from knock-in zebrafish models of AD developed to resemble the genetic background of human AD. In the Introduction, the importance of transcriptome analysis in understanding AD at the molecular level is explained. Chapters 2, 3 and 4 describe the first brain transcriptome analyses of two different knock-in zebrafish mutation models (psen1K97fs and psen1Q96_K97del) modelling different aspects of human AD. In both zebrafish models, young adult brains revealed notable transcriptome changes, including changes to immune/stress responses and energy metabolism respectively. Gene network and gene set analysis approaches revealed that some of these gene expression changes were preserved in human AD datasets, suggesting the validity and utility of the approach used. The work in Chapter 4 supported an important role for iron dyshomeostasis in AD across animal models and human AD and demonstrated, for the first time, the viability of detecting iron dyshomeostasis changes relevant to AD at the transcriptional level. This was achieved using a unique computational approach to the definition of gene sets based on predicted Iron Responsive Elements. Lastly, Chapter 5 used an advanced dimension reduction method to integrate zebrafish and mouse AD model datasets with human familial and sporadic AD. This resulted in the first preliminary comparison of human familial and sporadic AD transcriptomes as well as confirmation that the brains of a knock-in familial AD mutation-like zebrafish model (psen1Q96_K97del) and the commonly used 5XFAD mouse model show extensive differences at the molecular level. Overall, analysis of young adult zebrafish brains revealed potential molecular mechanisms relevant to early stages of human AD that were significantly preserved in human familial and sporadic AD datasets. This work demonstrates the value to our understanding of AD of a bioinformatics-led approach involving transcriptome analysis of knock-in zebrafish models.Thesis (Ph.D.) -- University of Adelaide, School of Biological Sciences, 202