Comparing common and unique disruptions in transposable elements and gene networks in two humanized mouse models of AD.

Abstract

Alzheimer's Disease (AD) is an widespread neurodegenerative disorder, most common in elderly and Veteran populations, culminating in dementia with additional non-cognitive behavioral and metabolic symptomatology, such as hyperactivity and disrupted sleep/wake patterns, that promote weight loss. Weight loss In demented patients is associated with adverse outcomes like accelerated AD progression and institutionalization. Mouse models of tau deposition, such as Tg4510 mice, and both tau and amyloidosis, such as 3X Tg, develop both cognitive and non-cognitive symptoms. To identify early common and unique factors affecting gene networks in these different molecular models of AD, we analyzed RNA-seq data from young Tg4510 and 3X Tg mice to compare the common and unique transcripts and transposable elements (TE) expression changes that may initiate the behavioral and metabolic derangements in AD. We analyzed RNA-seq from two month old Tg4510 mice (n=5) vs. controls (n=6) and 3X Tg (n=3) and controls (n=3). Using our established bioinformatics TTESA pipeline, reads were aligned to reference genome and measured gene expression or TEs with featureCounts. Pathway enrichment analysis using Visual Annotation Display was used to discover common functional and disease pathways using Gene Ontology, Mammalian Phenotype Ontology, and Protein Ontology. TTESA results detected significantly differentially expressed TEs in both models (p<0.001) albeit 3X Tg expressed four times more TEs. The majority of significantly differentially expressed TEs were retrotransposons (68-85%), mainly of long terminal repeats (LTRs; 50-70%), and some LINE retroelements (10-30%) and a few SINEs (5%) in either transgenic model. VLaD pathway analysis revealed many biological processes upregulated in transgenic mice, including non-cognitive behavioral pathways were downregulated, such as the regulation of circadian and sleep/wake behavior, however with different genes altered within these networks (e.g., Tg3x p<0.032; Ncor1, Rasd1, Rgs16; Tg4510 Ciart, Dbp, Per2, Per3 p<0.0165). There were also unique pathways to each model, e.g., 3X Tg: subcellular organelle structure, cytoskeleton, tubulin and actin binding, whereas Tg4510 dopamine receptor binding and bHLH transcription factor binding. Based on our data, early dysregulation of TEs may be involved altered neural gene expression networks underlying neurogenesis, behavioral changes, and molecular function in two different mouse models of AD.