APOE is linked to Alzheimer's disease in a large pedigree

A multigenerational pedigree of late-onset Alzheimer's disease implies new genetic causes

We describe the clinical phenotype and pathology of a new autosomal dominant late-onset familial form of Alzheimer's disease in four extensive kindred originated in a genetically isolated population. Twelve affected and 16 unaffected members of these kindred were examined clinically, and a brain post-mortem study was carried out in one case. The preliminary genetic assessment included complex segregation analysis, evaluation of the power to detect linkage, and exclusion of candidate genes. Dementia has been recorded for six generations in ancestors of examined cases. Review of death certificates allowed linking of all subjects in four extensive pedigrees. Affected individuals examined had progressive memory loss with onset between 57 and 74 years of age, along with seizures, myoclonus and parkinsonism in advanced stages. The brain of the case examined post-mortem showed widespread neocortical neuritic plaques and neurofibrillary tangles (stage VI of Braak), amyloid angiopathy, and Lewy bodies restricted to limbic areas. Sequencing exons 16 and 17 of amyloid precursor protein, and exons 4-12 of presenilin 1 and presenilin 2 genes did not disclose any mutations. Genotyping with markers D21S265, D14S71, D14S77, D1S2850 and D1S479 located 1-3 cM from the previously reported genes further excluded linkage to these genes. Seven out of 12 cases were apolipoprotein E (APOE) epsilon3/3, although the presence of an APOE epsilon4 allele was associated with an increased risk of dementia (odd ratio 6.17; 95% confidence interval: 1.15-33.15), but not to an earlier age of onset. Complex segregation analysis showed that the best model fitting the data was that of a major gene (dominant) with a gene frequency close to 3% in this population. Simulation analysis predicted an average logarithm of odds (LOD) of 2.2 at = 0.05. These four families, which seem to be part of a common extended pedigree originated by a founder arriving in this region in the 18th century, represent an autosomal dominant late-onset familial Alzheimer's disease not linked to previously known genetic loci. The simulation analysis suggests that it will be feasible to locate a novel responsible gene in these kindred.

The complexity of the genotype-phenotype relationship and the limitations of using genetic "markers" at the individual level

Many associations have recently been discovered between phenotypic variation and genetic loci, causing some to advocate what Robert Sinsheimer has called "a new eugenics" that would treat genetic "defects" in individuals prone to a disease. The first premise of this vision is that genetic association studies reveal the biological cause of the phenotypic variation. Once the responsible genes are known, the second premise is that we should focus upon changing "nature" rather than "nurture" by correcting the "defective" genes. The first premise is flawed because associations between genetic markers and phenotypes can be spurious, as shown by an example. Moreover, it is shown that using non-causative but associated genetic markers one at a time (the normal practice) can lead to incorrect predictions of disease risk for many individuals. Going from association to causation is a non-trivial step scientifically that has rarely been done in much of the human genetic research. Even when a particular locus does contribute to the phenotypic variation of interest, the first premise remains flawed because phenotypes in general arise from complex interactions among genes and between genes and environments as shown for genes associations with coronary artery disease (CAD). The ability of current molecular genetic tools to "fix" defective genotypes is extremely limited, but even if the technological problems could be overcome, the studies on CAD reveal no obvious "defective" gene to fix because the genetic effects are so context dependent (upon both other genes and environmental factors). Contrary to the second premise of the new eugenics, the more we learn about how different genotypes show variable responses to environments, the more important the environment becomes for individual treatment. The paradigm of a "defective gene" may work for classical Mendelian genetic diseases that are due to loss-of-function mutations. However, such mutations affect only a small portion of humanity. When the focus is changed to common disease and behavioral phenotypes, the "defective gene" paradigm is biologically meaningless and often harmful when applied to individuals. Thus, even when genes clearly do influence common phenotypic variation, the premises of the "new eugenics" are biologically indefensible.