Membership in high-risk genetic groups predicts Alzheimer's disease and age-at-onset





E.H. Corder, R. Huang, H.M. Cathcart, I.S. Lanham, G.R. Parker, D. Cheng, S.E. Poduslo

Center for Demographic Studies, Duke University, 2117 Campus Drive, Durham, NC 27708-0408, USA



Neurodegeneration is a brickwall when considering extreme longevity. Alzheimer’s brain changes, i.e. the accumulation of neurofibrillary tangles and senile plaque, can begin in early adulthood and are almost universal by age 80. The continued loss of pyramidal neurons in early-affected brain areas, i.e. the hippocampus, and the sequential regional brain spreading of lesions predicts that eventually there will be compromised cognition. The wide variation in the rate of these changes is strongly influenced by inherited factors, APOE polymorphism and other less replicated genetic determinants. But, replication of exactly which variants are relevant is difficult, probably because common combinations vary in frequency by chance from sample to study sample. We identified high and low risk multilocus genotypes for APOE, APOCI, the LDL receptor, cystatin C and cathepsin D genes using a form of fuzzy latent classification called grade-of-membership analysis (8 loci; 180 patients, 120 controls), GoM. Five GoM groups were requested. There were three high-risk combinations. They varied widely in terms of age at onset. One of the two low-risk combinations represented long life without dementia. Membership in the high-risk groups predicted disease status: All 102 subjects having > 80% membership, and none of the 50 subjects having < 20% membership, was affected (p < 0.0001). We conclude that association studies that identify risk sets may be useful in determining the natural history of the dementing process, and for individuals who to treat with preventive interventions, which intervention, and at what age. In addition, this approach should improve replication of the relevant candidate genes: Some variants are risk factors only when found with specific other variants; The combinations of variants identified, based on metabolic fine-tuning, should be stable from sample to sample with chance determining the prevalence of each risk group.




Key words:







Problems or questions regarding this site should be directed to the organiser