Alzheimer’s disease (AD), the most common cause of dementia, was the seventh-leading cause of death globally in 2021. As of 2023, dementia affected approximately 55 million people worldwide. AD is a slow-progressing neurodegenerative disorder that primarily impacts memory, cognitive abilities, and daily functioning. The two main pathological hallmarks of AD are extracellular amyloid-beta deposition, forming neuritic plaques, and intracellular hyperphosphorylated tau accumulation, resulting in neurofibrillary tangles (NFTs). These pathological features are believed to drive the degenerative processes in the brain. Emerging evidence highlights the medial temporal lobe as the earliest brain region affected by AD, particularly through the formation of NFTs. As the disease progresses, the associated damages spread throughout the brain in a predictable pattern of distribution. AD is a multifactorial and highly complex disorder, with numerous contributing factors that complicate the understanding of its mechanisms and hinder the development of effective treatments. Bridging these knowledge gaps remains a critical goal in AD research. In this study, I employed systems biology approaches, specifically gene co-expression network analysis, to investigate the complexity of gene interactions underlying AD pathology. Particular focus was placed on long non-coding RNAs (lncRNAs), which remained poorly characterized in the context of AD despite their potential regulatory roles. Additionally, I explored how disease progression is reflected at the level of gene networks between the temporal cortex and frontal cortex. RNA-seq data from two independent studies were used, comprising five datasets with a total of 396 individuals from two human brain regions: the temporal cortex and frontal cortex. Weighted topological overlap was applied to construct gene co-expression networks, and their reliability was enhanced by integrating them into consensus networks. The study aimed to uncover critical genes, pathways, and molecular mechanisms contributing to AD pathology. Differential co-expression analysis revealed significant network rewiring under AD conditions, with shifts in hub genes and the emergence of new key players specific to the disease. Region-specific differences were observed, with the temporal cortex exhibiting more pronounced network alterations compared to the frontal cortex, consistent with the established progression of AD pathology. The temporal cortex showed significant adaptive rewiring of gene interactions, whereas the frontal cortex predominantly displayed a loss of healthy correlations, reflecting different stages of vulnerability and compensation in these brain regions. In the temporal cortex, 46 protein-coding genes were identified as occupying critical positions in the AD network, potentially contributing to the observed disruptions and extensive rewiring. A key objective of this study was to predict lncRNA functions and their associations with AD. The Random Walk with Restart (RWR) algorithm was applied to infer potential functions for lncRNAs by evaluating their proximity to functional protein-coding genes. Using this approach, more than 100 lncRNAs were assigned to biological processes under both AD and non-AD conditions. Of these, 27 lncRNAs were identified as potential key players in AD pathology within the temporal cortex. These findings offered a more comprehensive understanding of the molecular interactions in AD and presented new candidate protein-coding and lncRNA genes for future experimental validation and therapeutic exploration.
