The original version of this article unfortunately contained a mistake. Gary S. Laco should not be listed as an author in the author group.

The apolipoprotein (APOE) e4 allele is one of the strongest genetic risk factors for Alzheimer’s disease (AD). While the apoE4 protein is well known for its limited capacity to transport lipids, little examination has been performed on blood lipids in the context of the e4 status and their collective contributions to AD. Lipidomic analyses were performed on serum samples from a longitudinal cohort of cognitively normal subjects, a subset of whom converted to mild cognitive impairment (MCI)/AD within 2 to 3 years. To examine lipids in relation to the APOE genotypes and AD pathology in a mouse model, we performed lipidomic profiling of plasma samples from human APOE targeted replacement mice (APOE-TR) and APOE-TR mice crossed with mice harboring 5 AD mutations (EFAD). To gain a mechanistic understanding of the differences in lipid transport in relation to the e4 allele, we assessed arachidonic acid (AA) and docosahexaenoic acid (DHA) transit in an in vitro blood brain barrier (BBB) model in the presence of different apoE isoforms. Serum from cognitively normal e4 carriers who converted to MCI/AD had increased AA/DHA ratios in major phospholipid (PL) classes compared to non-carriers and cognitively normal e4 carriers. Combining these PL species, e4 status, and the Abeta42/40 ratio gave 91% accuracy for detecting preclinical MCI/AD. Both e4 carriers and non-carriers who consumed fish oil/omega-3 had lower AA/DHA ratios within several PL classes compared to non-users. The AA/DHA ratio within PL was increased in plasma from mice expressing human APOE4 relative to E2 or E3 isoforms and in E4FAD mice compared to EFAD animals with other APOE isoforms. In vitro BBB studies showed the apical-to-basolateral transit of both AA and DHA was diminished using human serum from e4/e4 compared to e3/e3 individuals. The elevated AA/DHA ratios within PL observed in serum of e4 carriers with preclinical MCI/AD could be due to impaired lipid transport into the brain. Finally, these studies indicate that determination of AA and DHA containing PL together with APOE genotype could assist in the detection of preclinical MCI/AD.

The contribution of apolipoprotein E (APOE) E4 to the cerebrovascular dysfunction in Alzheimer’s disease (AD) is characterized by the presence of severe cerebral amyloid angiopathy, increased blood-brain-barrier (BBB) breakdown, reduced cerebral vascularization, and basement membrane thinning in E4 carriers compared to non-carriers. It has also been proposed that the diminished capacity of the apoE4 protein to transport essential polyunsaturated fatty acids (PUFAs) that are required for the structural and functional maintenance and vascular integrity of the brain also contribute to AD pathogenesis. However, it remains to be determined if there are changes in the profiles of phospholipids (PL) and in the expression of lipid transporters within the brain vasculature in relation to the E4 allele and AD diagnosis. We performed liquid chromatography/mass spectrometry based lipidomic analysis of the cerebrovascular and parenchymal fractions from autopsied human brain tissue of pathologically confirmed AD cases and controls stratified by APOE genotype. In order to determine if there were changes in the expression of lipid transporters in relation to the APOE E4 allele, we performed antibody based examination of the major facilitator superfamily domain containing 2A (mfsd2a) protein in the cerebrovasculature from these subjects. Total phosphatidylcholine (PC) was significantly lower in the cerebrovascular fractions of AD patients compared to controls. While docosahexaenoic acid (DHA) containing PL species were lower in heterozygous E4 AD patients compared to E4 controls in both the cerebrovascular and parenchymal fractions, an ether PC species containing arachidonic acid (AA) was elevated within the cerebrovasculature of E4 carriers relative to non-carriers and was highest among E4 AD specimens compared to E4 controls. We also observed an APOE E4 dependent difference in mfsd2a expression. Among AD patients, E4 homozygotes had lower expression of mfsd2a than E4 heterozygotes and non-carriers. These findings demonstrate that deficiencies in DHA within the brains of APOE E4 carriers may, in part, be due to lower expression of mfsd2a. Thus, targeting this transport mechanism may improve the bioavailability of DHA to the brain of APOE E4 individuals providing a novel approach to the treatment of AD.