Plasmalogen deficiency and the Alzheimer’s disease risk of apolipoprotein E4

Abstract The ε4 allele of the APOE gene, encoding the E4 isoform of apolipoprotein E, is the leading genetic risk factor for late-onset Alzheimer’s disease. While many potential mechanisms have been proposed to explain this risk, no dominant or unifying process has yet emerged. Here, we explore the primary function of apolipoprotein E in lipid transport and metabolism, by examining its lipid association properties, to establish whether they show isoform dependence and thereby could mediate Alzheimer’s risk. We focus on ethanolamine plasmalogen, a phospholipid subclass known to be depleted in Alzheimer’s disease brain. We purified apolipoprotein E from human cerebrospinal fluid by immunoprecipitation using an anti-pan-apolipoprotein E monoclonal antibody bound to magnetic beads, then conducted lipidomic and proteomic analyses of the precipitates by mass spectrometry. The cerebrospinal fluid samples were obtained from cognitively intact, relatively young individuals with no evidence of amyloid pathology and with known apolipoprotein E isoform status (E3E3, n = 5; E3E4, n = 4; E4E4, n = 5). The molar ratio of ethanolamine plasmalogen to apolipoprotein E was 29.5% lower for E4E4 than for E3E3 (P = 0.007) with a biological gradient: E3E3 > E3E4 > E4E4 (P = 0.03). No similar trends and differences were found for phosphatidyl ethanolamine, a chemically related lipid (P = 0.5). Compared to E3E3, the molar ratio of ethanolamine plasmalogen to phosphatidyl ethanolamine was significantly reduced for E3E4 (P = 0.0016) and E4E4 (P = 0.0001). The latter deficiency was similar in magnitude to that found in Alzheimer’s disease brain relative to control. The finding that ethanolamine plasmalogen is depleted in apolipoprotein E4 relative to E3 strengthens the view that brain deficiency of this same lipid contributes to Alzheimer’s disease causation, rather than being an effect of the neurodegeneration. Simultaneously, these results supply a potential mechanism for the risk of E4 versus E3, the former being less able to counteract the tissue defect. The apolipoprotein E4 lipid depletion cannot itself be a consequence of Alzheimer’s disease, since cerebrospinal fluid samples were taken from individuals with no evidence of the condition. The biological gradient in ethanolamine plasmalogen deficiency mirrors the relationship of Alzheimer’s disease risk (odds ratio) to E4 allelic dose. Ethanolamine plasmalogen deficiency could be linked to, or indeed drive, several metabolic pathways implicated in Alzheimer’s pathogenesis, including amyloid-beta deposition and cholesterol dysregulation. Future studies should extend approaches to therapeutic intervention in Alzheimer’s disease which attempt to reverse this lipid abnormality.