Brain Imaging of Mice & (Wo)Men: Reverse Translational Development of Mouse Brain Imaging for Alzheimer's Disease

Abstract

AD is a multi-faceted disease, resulting from a combination of genetic and environmental factors. Age and biological sex are important risk factors in this disease, with 90-95% of the clinical AD population suffering from sporadic or late-onset AD (LOAD), aged 65 years and above, and females have twice the lifetime risk of developing AD compared to males. This neurological disorder is associated with brain glucose hypometabolism, loss of white matter integrity (WMI), and neuronal atrophy, which leads to progressive cognitive decline during aging. Currently, approved AD therapeutics are centered around the amyloid cascade hypothesis and have proven to be successful in delaying cognitive decline, but they are not without their shortcomings. This amyloid hypothesis has also influenced the preclinical mouse models used to study this disease and test therapeutics. However, they use dominant mutations in the amyloid precursor protein (APP) and presenilin (PSEN) to induce pathology mirroring familial AD, which accounts for less than 5% of the clinical population. Genome-wide association studies (GWAS) have identified apolipoprotein E4 (APOE4) as the strongest genetic risk factor linked to LOAD. APOE4 also interacts with age and sex to increase the likelihood of developing LOAD, with one copy of the APOE4 allele conferring a 4-fold and two copies conferring a 12-fold risk in APOE4 females relative to APOE3 homozygous males. Furthermore, neuroimaging studies have identified an association between APOE4 carrier status and lower brain glucose uptake, cortical atrophy, and reduction in white matter integrity (WMI). This effect is observed irrespective of diagnosis, ranging from cognitively normal (CN) to mild cognitive impairment (MCI) and AD. However, neuroimaging studies investigating the impact of APOE4 and sex on brain imaging outcomes in preclinical mouse models are lacking. The central hypothesis guiding this doctoral research is that the lipid transport dysfunction of APOE4 can impact brain morphology and metabolic indicators in a genotype and sex-dependent manner at a LOAD-relevant age. I propose that in an aged preclinical mouse model of AD, hAPOE-induced risk modification will result in distinct brain imaging and metabolic profiles influenced by APOE4 allele count and sex. These differences will present in the form of lower brain glucose uptake, volumetric loss, and loss of WMI and will be comparable to the brain imaging outcomes in the clinical population. To test this hypothesis, a forward translational approach was taken that combined preclinical neuroimaging of humanized mouse models with a cross-species framework developed from a large-scale neuroimaging dataset (ADNI) to determine translational correspondence. Impact of APOE and biological sex was assessed in an aged novel hAPOE mouse model of AD, which revealed lower brain glucose uptake and greater total brain volume (TBV) in females. Sex differences in brain volume were observed in the form of larger cortical regions in females relative to males and larger subcortical regions in males relative to females. APOE4 allele dose was linked with greater TBV and lower mean fractional anisotropy(FA) relative to the APOE3/3s. The addition of humanized APP (hAPP) resulted in a reduction in TBV, which was most observable in AD-vulnerable regions of the hypothalamus, hippocampus, and amygdala within the hAPP/APOEs. WMI was linked with greater mean FA in parts of the corpus callosum, right internal capsule, and fimbria in the APP/APOEs, whereas APOEs had greater mean FA in the arbor vitae of the cerebellum and fornix. Analysis of LOAD-relevant ADNI neuroimaging data identified sex and APOE4 genotype effects on brain glucose uptake, with females having greater uptake than males and APOE4 carrier status linked with lower glucose uptake, across diagnoses (CN or MCI). Brain volume was linked with sexually dimorphic outcomes, with females having smaller gray matter, hippocampal, amygdala, and entorhinal cortex volume compared to males. Sex differences were also observed in WMI, with females having greater FA and radial diffusivity(RD) in the fornix, irrespective of the diagnosis status (CN or MCI). Translational correspondence scoring of brain imaging outcomes based on structurally and functionally conserved regions across species revealed the greatest similarities in the volumetric outcome, followed by DTI measures between cognitively normal and hAPOE mice. The addition of APP modified brain imaging outcomes in the preclinical mouse models more towards a clinical AD phenotype. Together, this body of research provides a novel multi-dimensional neuroimaging approach to assess the translational validity of AD risk mouse models. By integrating standardized preclinical imaging with a cross-species correspondence framework, this work offers a rigorous mechanism for evaluating and refining preclinical mouse models. This work lays a foundation for aligning therapeutic testing with clinically relevant outcomes, thereby enhancing the therapeutic pipeline and improving the likelihood of translational success.