An investigation into the whole transcriptome in Alzheimer's disease - contribution of coding and non-coding RNA

Abstract

Alzheimer’s disease (AD) is a chronic neurogenerative disease affecting millions of people worldwide. While significant strides have been made in understanding the aetiology and pathogenesis of AD, current theories cannot yet fully explain how or why it develops. As such, the vast majority of treatments developed for AD are ineffective, or only capable of slowing disease progression for a short time.

One potential avenue of research that may aid in overcoming these obstacles is non-coding RNA (ncRNA). While protein-coding genes have held the spotlight in research into the transcriptome for many years, work in the past several decades has highlighted important roles for ncRNA, which outnumber protein-coding genes in both number and expression, in a variety of biological processes. In particular, many microRNA (miRNA), long non-coding RNA (lncRNA), and others have been found to contribute to disease processes in AD. Advances in technology are enabling ncRNA to be explored more fully than ever before, and illuminate their multifaceted roles in both health and disease.

To examine whether ncRNA may play roles in AD, this study aimed first to develop a method for identifying and quantifying both protein-coding and non-coding RNA from the same sample simultaneously, using a flexible modified RNA-Seq protocol. It was shown that this method was able to detect the full gamut of RNA species from samples of RNA obtained from both fresh-frozen mouse brain tissue and frozen post-mortem human brain tissue in biologically relevant quantities, whilst also effectively and efficiently removing ribosomal RNA contamination.

Next, the data obtained from this whole transcriptome RNA-Seq method was used to examine changes in the transcriptome of APP/PS1 mice, a commonly used AD model. It was found that 610 genes were differentially expressed between APP/PS1 mice at 15-months-old and wild-type littermates, and that almost half of these were ncRNA (n = 4 each transgenic and wild-type; p < 0.05). Functional enrichment analysis of differentially expressed genes found that changes in gene expression were enriched for processes involved in the glial neuroinflammatory response – notably Complement system activation, TGF-β signalling, astrocyte reactivity, and microglial phagocytosis. Numerous ncRNA genes were identified whose function is yet unknown. Furthermore, eight transcription factors were identified to be responsible for the majority of gene expression changes in these mice – AP-1 (as c-Fos and Junb), Clock, Egr2, Irf8, Lmo2, Nr4a1, Nfe2l2, and Runx1.

Finally, whole-transcriptome gene expression changes were explored in post-mortem human brain tissue from AD patients and controls. These analyses found 1233 genes differentially expressed between AD patients and controls, with around one third corresponding to ncRNA (n = 3 AD, 4 control; p < 0.05). Functional enrichment analysis of these genes revealed three major pathways dysregulated in AD patients – Complement system activation, microglial phagocytosis, and TNF-α signalling via NF-κB. It was also found that six transcription factors underlie most of the changed genes – BCL6, CHD1, CPFP, MYC, EGR1, and ELF1.

Further, these data allowed for a direct comparison of the transcriptome environment between APP/PS1 mice and human AD patients. These analyses revealed 41 genes differentially expressed in both APP/PS1 mice and AD patients, and these genes were strongly enriched for complement cascade activation, implying an aspect of shared pathology. However, this analysis also revealed key differences between the two groups. Notably, while APP/PS1 mice exhibit changes in TGF-β signalling, human AD patients rather show dysregulated TNF-α/NF-κB signalling, highlighting significant differences.

In summary, these results identify a wide range of coding and non-coding RNA dysregulated in both an animal model of AD and in human patients, as well as some of the key biological processes involved in disease pathogenesis, in particular the complement cascade – a process common to both species. These results also identify numerous ncRNA that remain largely unannotated and poorly understood. Further research into these ncRNA species in the context of AD pathology would not only further our understanding of the disease process, but may also lead to novel targets for therapeutic interventions.