Abstract
Translational validity of mouse models of human aging and late-onset Alzheimer's disease (LOAD) risk are essential for both fundamental mechanistic science and therapeutic development. Given that the strongest risk factors for LOAD are age, female sex, and APOE-ε4 carriership, models must reflect these biological variables and disease phenotypes. The use of mouse models with humanized APOE (hAPOE) is a key strategy to advance translational validity. To initially address translational validity of the hAPOE mouse model, we conducted ex-vivo magnetic resonance imaging analysis of brain volumes in male and female mice across APOE genotypes (ε3/ε3, ε3/ε4, ε4/ε4) and ages corresponding to a human lifespan of approximately 30-70 years (6-25 months in mice). The primary outcomes indicated that total MRI brain volume increased with age (an average of 2.12mm3 per month), irrespective of sex or APOE genotype. Additionally, APOE-ε4 carriers had greater total brain volumes than non-carriers. No sex differences were observed in total brain volume. Voxelwise analyses revealed a pattern of localized morphometric changes independent of differences in total brain volume. Age-related volumetric increases were distributed across subcortical regions (e.g., thalami, hippocampi), while age-related decreases were evident across cortical regions, notably the bilateral parieto-temporal and frontal lobes. Additionally, sex differences were evident after controlling for total brain volume, with females showing greater cortex-dominant volumes while males showed a pattern of greater volumes in regions including cerebellar cortices, olfactory bulbs, and striata. No genotypic effects were observed in the voxelwise analysis after correcting for multiple comparisons, suggesting that APOE genotype does not drive localized volume differences independent of total brain volume. These findings indicate that, at the MRI level of analysis, this humanized APOE mouse model does not recapitulate the volumetric atrophy typically seen in human brain aging and Alzheimer's disease. The results suggest that humanized APOE alone is insufficient to induce the LOAD atrophy phenotype. This model may better serve as a platform for studying vulnerable aging rather than a primary model for progressive neurodegenerative atrophy.