PROJECT SUMMARY DISEASE MODEL DEVELOPMENT AND PHENOTYPING (DMP) PROJECT In the initial funding period, the DMP created and evaluated more than forty new models incorporating 33 LOAD- relevant alleles. This included a set of platform strains that incorporated humanized APOE4, Trem2*R47H and humanized Aβ sequence. Platform strains were evaluated from 4 to 24 months through the IU/JAX/PITT MODEL- AD phenotyping pipeline that included PET/CT, multi-omics (transcriptomics, proteomics and metabolomics), neuropathology, and fluid biomarker analysis. The first platform strain, LOAD1 (double homozygous for APOE4 and Trem2*R47H) showed primarily age-dependent affects and few genotype-specific effects. For our second platform strain, LOAD2 (triple homozygous for APOE4, Trem2*R47H and hAβ), we included a high fat/high sugar diet (HFD) group as an environmental risk factor. Compared to control mice, at 12 months, HFD-fed LOAD2 mice showed elevated levels of insoluble Aβ42, loss of neurons in the subiculum, reduced long-term potentiation (LTP), and reduced spine densities in the hippocampus. In addition to the platform models, 24 putative genetic risk variants identified by the Bioinformatics and Data Management core (BDMC) were introduced into LOAD1 or LOAD2 by CRISPR/Cas9 and evaluated by comparing brain transcriptomes to Accelerated Medicines Partnerships in AD (AMP-AD) data. The putative risk variants were located in both previously identified (e.g., Abca7, CR1, Plcg2, Mthfr and Epha1) as well as novel AD risk genes. Through this work, Abca7*A1527G, Plcg2*M28L and Mthfr*677C>T were prioritized for full characterization through our phenotyping pipeline. Collectively, these data support our overarching hypothesis that combinations of LOAD-relevant genetic and/or environmental risk factors, in the absence of familial AD (fAD) mutations, will induce hallmark AD pathologies in mice and therefore create improved mouse models for preclinical testing. In the renewal we will continue to evaluate novel LOAD models, with an increased focus on identifying models for the preclinical testing core (PTC). To achieve this, our phenotyping pipeline has been enhanced and incorporates a combined cross-sectional and longitudinal design, extensive biomarker analysis, MRI, additional PET ligands (e.g., for neuroinflammation and synapse loss), digital spatial profiling, and synaptic and cognitive assessment. We will first complete characterization of three prioritized LOAD strains as well as characterize a novel platform model LOAD3 (triple homozygous for APOE4, hAβ and a humanized MAPT allele) with and without HFD (Aim 1). Hyperphosphorylation and aggregation of TAU is a hallmark pathology for LOAD and absent in our current strains. Therefore, we will next evaluate four MAPT variants (N279K, P301L, R406W and S320F) that are predicted to increase susceptibility to TAU pathology (Aim 2). Finally, no single LOAD model is expected to represent the heterogeneity of human LOAD and we ...