Abstract Background Restrictions on mouse models have significantly impacted research towards understanding the most common genotype contributing to dementia in the human population – APOE ε3/ε4 . To address this, as part of MODEL-AD, we created new versions of humanized APOE ε4 and APOE ε3 mice on a C57BL/6J background that allow for unrestricted distribution and breeding. Methods To determine similarities and differences between APOE ε3/ε4 and APOE ε4/ε4 risk genotypes, we analyzed peripheral lipid concentrations as well as performed unbiased transcriptional profiling of the cortex at two and four months of age, comparing APOE ε3/ε4 and APOE ε4/ε4 to the reference APOE ε3/ε3 . To further compare APOE genotypes, cohorts of APOE ε3/ε3 , APOE ε3/ε4 , and APOE ε4/ε4 mice were exercised by voluntary running from 1 month to 4 months of age. Results Cholesterol composition was significantly influenced by APOE genotype as early as 2 months, while triglycerides were affected by APOE genotype at 4 months. Importantly, RNA-sequencing of the cortex followed by linear modeling or weighted gene co-expression network analysis (WGCNA) revealed that the APOE ε3/ε4 genotype showed unique transcriptomic signatures to that of APOE ε4/ε4 . Functional enrichment of the APOE ε3/ε4 , but not APOE ε3/ε4 genotype, revealed sulfur and heparin binding as significant terms at 2 months, and extracellular matrix and blood coagulation at 4 months. Further, cell specific contributions of significant genes identified endothelial cells as overrepresented in the APOE ε3/ε4 but not APOE ε4/ε4 genotype. WGCNA analysis confirmed findings from linear modeling but also predicted that running at a young age affects myelination and gliogenesis across APOE genotypes. Conclusions In summary, APOE ε3/ε4 genotype-specific effects were observed in cortical transcriptional profiles, suggesting therapies aimed at modifying APOE biology to treat dementias may need to be targeted to specific APOE genotypes.