Gene therapy to treat hereditary disorders conventionally delivers the normal allele to compensate for loss-of-function mutations. More effective gene therapy may be achieved using a gain-of-function variant. We tested the hypothesis that AAVrh.10-mediated CNS delivery of the human APOE2 allele with the Christchurch mutation (R136S) (E2Ch) will provide superior protection against APOE4-associated Alzheimer's disease (AD) in mice compared to the unmodified APOE2 allele (E2). The vectors were assessed in two mouse strains with humanized APOE4: APP.PSEN1/TRE4 "amyloid mice" and P301S/TRE4, "tau mice." Both the E2Ch and E2 vectors prevented Aβ42 and Aβ40 accumulation and decreased β-amyloid aggregates in amyloid mice, but only the E2Ch vector suppressed tau tangles in tau mice. Microglial activation and reactive astrocytes were significantly suppressed by both vectors in amyloid mice but only the E2Ch vector mediated significant suppression of Iba1 and glial fibrillary acidic protein (GFAP) in tau mice. In four behavioral assays, the E2 and E2Ch vectors had similar benefits in amyloid mice, but E2Ch outperformed E2 in tau mice. In summary, while E2 is effective in suppressing amyloid pathology, the novel E2 variant E2Ch more effectively treats both the amyloid and tau pathology of murine AD in APOE4 background, supporting the development of AAVrh.10APOE2Ch as a therapy for APOE4-associated AD.