Identification of Novel Genes in Late-Onset Alzheimer's Disease

Abstract

Four genes affecting Alzheimer's Disease (AD)(AP, PS1, PS2, and APOE) have been identified and a fifth potential gene localized to chromosome 12. Collectively, these genes explain at most half of the genetic effect in AD. Understanding the genetics of AD is critical to developing new treatments. The quest to find the remaining AD genes led us to undertake a large genomic screen using over 466 families (730 affected sibpairs) in late-onset AD. In conjunction with this increase in power, we initiated several novel approaches to identify potential AD-related genes. This included stratification of the data into an autopsy-confirmed subset of 199 AD families. Each of these targeted analyses resulted in the identification of novel regions containing potential AD genetic risk factors. Our most significant finding was on chromosome 9 in the autopsy-confirmed subset where we obtained an MLS of 4.31. These approaches, together with new methodologies such as conditional linkage analysis, generalized family-based association tests (PDT), and a new generation of genetic markers (SNPs), opens the door for additional AD gene discovery. Such strategies are necessary if we are to understand the subtle and complex threads that, woven together, create the intricate tapestry of AD.

Introduction

Alzheimer's Disease (AD) is the leading cause of dementia in the elderly and the most common form of dementia occurring after the age of 40. There are over 2 million affected individuals in the U.S., a number projected to quadruple over the next 50 years as the population ages(Brookmeyer, Gray et al. 1998). AD has a complex etiology with strong genetic and environmental determinants. Pathologically AD is characterized by neurofibrillary tangles found in the neurons of the cerebral cortex and hippocampus and the deposition of amyloid within senile plaques and cerebral blood vessels(Wisniewski, Golabek et al. 1993). Clinically AD is slowly progressive, resulting in memory loss and alterations of higher intellectual function and cognitive abilities(Guttman, Altman et al. 1999).

Although AD was first described in 1907 (Alzheimer 1907), definitive clues to its etiology have emerged only recently. Using the powerful tools of genetic analysis, four AD genes have been identified. Three of these (the amyloid precursor protein [APP](Goate, Chartier-Harlin et al. 1991), and the presenilin 1 and 2 [PS1 and PS2] genes Levy-Lahad, Wasco et al. 1995 were identified using standard positional cloning methods facilitated by simple autosomal dominant inheritance in early-onset AD families. While these three genes account for the majority of early-onset familial AD and their identification represents a tremendous accomplishment, collectively they only account for about 2% of all cases of AD.

The genetic architecture underlying the far more common late-onset AD (Pericak-Vance, Yamaoka et al. 1988) is much more complex. Comparison of the recurrence rates in the siblings of AD patients to the general population prevalence (Risch 1990a; Risch 1990c; Risch 1990b) generates a recurrence risk ratio (λs) that is a rough measure of genetic influence. The estimates of λs for AD are surprisingly constant across studies Breitner, Silverman et al. 1988 and range from 4 to 5. Power studies Hauser, Boehnke et al. 1996 show that the genes responsible for λs as low as 1.5 can be detected with reasonably-sized samples of affected sibpairs.

In 1991, we (Pericak-Vance, Bebout et al. 1991) reported evidence for linkage of late-onset AD to chromosome 19q13. Apolipoprotein E (APOE), a plasma lipoprotein involved in lipid transport and metabolism, also maps to 19q13. The three APOE alleles (-2, -3, -4), when translated, result in different protein isoforms Menzel, Kladetzky et al. 1983. The confluence of biology Namba, Tamonaga et al. 1991 and genetic mapping facilitated our identification of the association between the APOE-4 allele in both familial late-onset and sporadic AD patients Saunders, Strittmatter, et al. 1993. Additional analyses (Corder, Saunders et al. 1993) showed that the APOE-4 allele acts in a dose-dependent manner to increase risk and decrease age of onset in both late-onset familial and sporadic and early-onset sporadic AD Corder, Saunders, et al. 1993. The APOE-2 allele affords protection against late-onset AD Corder, Saunders et al. 1994, but its effect in early-onset sporadic AD is controversial Scott, Saunders et al. 1997. The mechanism by which the specific APOE isoforms uniquely contribute to disease expression is not known. APOE represents the fourth confirmed genetic factor and is the single most significant biological risk factor identified for AD.

Several lines of evidence indicate that APOE does not account for all of the genetic variation seen in AD. While the heritability of AD has been estimated at about 80% (Bergen 1994), more than one-third of AD cases do not have a single APOE-4 allele. The APOE-4 associated risk of AD appears to differ among ethnic groups, suggesting ethnicity may influence genetic risk of AD Farrer, Cupples, et al. 1997. One such example is our studies of the Indiana Amish population (Pericak-Vance, Johnson et al. 1996) which showed familial aggregation of AD despite a low APOE-4 allele frequency. In addition we have estimated the λs for the APOE locus (Roses, Devlin et al. 1995) to be approximately 2. Since the overall λs is estimated to be between 4 and5, then APOE is likely to account for at most 50% of the total genetic effect in AD.

Efforts to identify these additional AD loci have taken two forms: whole-genome scans for linkage in multiplex families, and association tests of candidate genes in case-control samples. Two recent genome scans have implicated several chromosomes (1, 4, 6, 9, 10, 12, 19, and 20) as potential locations of additional AD loci Kehoe, Wavrant-De et al. 1999. Perhaps the most promising of these locations is chromosome 12, where the original linkage report (Pericak-Vance, Bass et al. 1997) has since been supported by results from two independent samples Kehoe, Wavrant-De et al. 1999. Several positional candidate genes in this region have been examined including low-density lipoprotein receptor-related protein (LRP) (Lendon, Talbot et al. 1997)and alpha-2-macroglobulin (A2M)(Blacker, Wilcox et al. 1998). These results have also not been consistently replicated Dow, Lindsey et al. 1999.

In contrast to using this positional candidate gene approach, several investigators have attempted to identify disease loci by focusing on candidate genes selected due to their function. These case-control association studies of candidate genes have been far less successful. Over the past three years, many genes have been reported as being associated with late-onset AD. At best, the evidence for any of these loci is mixed, and some associations have never been replicated. These functional candidate genes include α₁-antichymotrypsin (AACT)(Kamboh, Sanghera et al. 1995), low-density lipoprotein-like receptor (LRP)(Lendon, Talbot et al. 1997), presenilin-1 (PS1)(Wragg, Hutton et al. 1996), ubiquitin(Van Leeuwen, Dekleijn et al. 1998), the HLA complex Curran, Middleton et al. 1997, butyrylcholinesterase K variant (BCHE-K)(Lehmann, Johnston et al. 1997), non-amyloid component of plaques/α-synuclein (NACP/α-synuclein)(Xia, da Silva et al. 1996), and mitochondrial mutations (Hutchin and Cortopassi 1995). Successful examination of these and other candidate genes requires sufficiently large samples to detect modest genetic effects and appropriately matched cases and controls to avoid spurious associations due to biased sampling.

The deconstruction of the complex genetic architecture of Alzheimer's Disease has only begun. AD has become a paradigm for the identification and understanding of susceptibility alleles in diseases of the elderly, and thus is an ideal candidate for the application of new approaches and paradigms. We report here the results of our genomic screen on 466 late-onset AD families.