The completion of the human genome sequence brought renewed hope that genetic causes and susceptibilities to disease could be identified. But for Alzheimer disease (AD), little has really changed. Scientists still face the fundamental challenge of how to find needles in the haystack. The problem is complex because individually, tiny needles—like Lilliputian swords—have little effect, but en masse, they are strong enough to bring Gulliver down. A flurry of current papers highlight the fronts on which AD genetics advances at this time.

ApoE exemplifies the difficulty in finding genetic variations that associate with AD. Though its epsilon 4 variant is indisputably the biggest genetic risk factor for AD, prior scans of the genome using microsatellite markers have failed to see it. But now an international collaboration led by Dietrich Stephan at the Translational Genomics Research Institute, Phoenix, Arizona, has tried a more sophisticated approach. In the April Journal of Clinical Psychiatry, first author Keith Coon and colleagues reveal that a high-density scan of the genome using more than 500,000 single nucleotide polymorphisms (SNPs) does indeed pull ApoE out of the lineup. The researchers found that on testing DNA samples from 1,086 well-characterized AD cases, a single SNP (rs4420638) lying 14 kb distal to the ApoE locus has a powerful association with late-onset AD (corrected p value was 5.3 x 10 e-34). No other SNP showed as robust an association. The authors estimated that people with two ApoE4 copies have a 25-fold increased risk for developing the disease compared to ApoE3/ApoE3 carriers. This odds ratio is about 10 points higher than that calculated in other recent studies, including Alzgene meta analysis (see Bertram et al., 2007).

Further analysis revealed that the SNP was in linkage disequilibrium (LD) with the two SNPs that account for the genetic variation that distinguishes ApoE4 from ApoE2 and 3. While this finding shows the power of genome-wide SNP analysis, it also reveals a weakness—the association could have easily gone unnoticed. “The ApoE locus would have been missed with the 500K array if not for a single SNP in LD with the functional variant, and speaks to the need for increased SNP density for complete coverage,” write the authors. Indeed, insufficient coverage may explain why the study did not uncover some other genetic associations. For example, a collaboration led by Peter St. George-Hyslop at the University of Toronto; Richard Mayeux at Columbia University, New York; and Steve Younkin at the Mayo Clinic, Jacksonville, Florida, recently found that polymorphisms flanking the gene for SORLA (sortilin-related receptor, low-density lipoprotein receptor class A repeat-containing protein) are strongly associated with late-onset AD (see ARF related news story). The SORLA gene, SORL1, has been touted as perhaps the most important after ApoE. While that remains to be confirmed, in the April Archives of Neurology, Mayeux and colleagues report a follow-up study that supports the original finding.

First author Joseph Lee and colleagues genotyped DNA samples from a community-based cohort of 296 patients and 428 controls representing Caribbean Hispanic, African American, and white non-Hispanic European ethnic groups. The authors focused on SNPs flanking SORL1. Three of the original 29 SNPs (12, 20, and 26) were associated with AD in at least one ethnic group. Oddly enough, while the T allele of SNP 12 was associated with AD in Caribbean Hispanic patients, it was the C allele that was associated with AD in African Americans. Similarly, when the authors analyzed groups of SNPs for association, they found that a CCA haplotype at SNPs 24 to 26 was associated with AD in African Americans, while the CTG haplotype at the same location may be protective. This is the opposite of what was found in a different African American cohort and reported in their original study—in that case, the CTG haplotype was associated with AD and the CCA haplotype protective. “We interpret these opposite allelic associations to mean that one or more risk alleles nearby are in linkage disequilibrium with these haplotypes in this region in African American individuals,” write the authors.

Other haplotypes may have a more consistent association. The TTC haplotype at SNPs 23-25 and the ATA haplotype at SNPs 16-18 were associated with AD in white people in both the original study and in this one. “Our results independently confirm the previous conclusion that multiple genetic variants in SORL1 are associated with AD,” the researchers write. They note that SORL1 and ApoE associations differ markedly. “The discovery of significant association in multiple regions of the gene and the discovery of different AD-associated haplotypes in different data sets support the notion that there may be a high degree of allelic heterogeneity, with disease-associated variants occurring on multiple different haplotype backgrounds,” they write. The next step will almost certainly be identifying the pathogenic variants in SORL1 that are linked to AD.

Diabetes, Genetics, and AD
In looking for genetic susceptibility to AD, researchers are casting a wide net. There is growing evidence, for example, that people with diabetes face an elevated risk for Alzheimer’s. This is supported by a second Archives of Neurology paper from the labs of Mayeux and Jose Luchsinger, also at Columbia University. First author Luchsinger and colleagues report that diabetes is related to a significantly higher risk for all-cause mild cognitive impairment (MCI) and amnestic MCI (aMCI)—the latter is particularly relevant because there are increasing indications that it may be a transition state between normal cognition and AD (see ARF related news story).

The researchers studied 918 elderly volunteers (219 with type 2 diabetes) from northern Manhattan, and during at least 1 year of follow-up, 160 cases of amnestic MCI and 174 cases of non-amnestic MCI emerged. Diabetes increased the risk for all-cause MCI by about 40 percent. The same was true for aMCI, whereas associations between non-amnestic MCI and diabetes failed to reach significance once the data was adjusted for ethnic group, years of education, and ApoE status. Taking cerebrovascular conditions such as heart disease and stroke into account weakened the non-amnestic MCI/diabetes association further. Overall, the researchers found that the risk for MCI attributable to diabetes was 8.4, 11, and 4.6 percent for Hispanic, African American, and non-Hispanic white people, respectively.

The finding speaks to the differences between non-amnestic and amnestic MCI. The latter might be due to AD pathology, such as factors affecting amyloid accumulation. Hyperinsulinemia, for example, might affect levels of insulin-degrading enzyme, which also degrades amyloid-β, and diabetes might also affect levels of advanced glycemic end products, which may contribute to AD pathology (see ARF related news story). Whatever the etiological link, that diabetes was still associated with aMCI, even after taking cardiovascular conditions into account, suggests that vascular changes do not explain diabetic risk for aMCI. “Conversely, the relation of diabetes to non-amnestic MCI was appreciably attenuated and became non-significant after adjustment for stroke and vascular risk factors, suggesting that cerebrovascular disease may mediate the relation between diabetes and non-amnestic MCI,” write the authors.

If diabetes can increase the risk for aMCI and AD, then people who are genetically predisposed to type 2 diabetes may also be at increased risk for AD. A paper in today’s Science highlights one such susceptibility, a variation in the FTO gene. FTO encodes a protein of unknown function and was discovered in a part of mouse chromosome 8 that is deleted in animals with fused toe syndrome. Andrew Hattersley of Peninsula Medical School, Exeter, and Mark McCarthy, University of Oxford, both in the U.K., and an international host of collaborators at The Wellcome Trust and other academic institutions found the FTO link by conducting a genome-wide association study for SNPs that associate with type 2 diabetes. Joint first authors Timothy Frayling, Nicholas Timpson, Michael Weedon, and Eleftheria Zeggini, and their colleagues genotyped 490,000 SNPs in 1,924 patients and 2,938 controls. They found that SNPs in FTO strongly associated with diabetes and also with increased body mass index (BMI). In fact, after adjusting for BMI, the diabetes association disappeared, indicating that the FTO SNPs increased the susceptibility for diabetes through their effect on body mass.

The researchers identified 10 different FTO SNPs in the first intron of the gene that associated with both BMI and type 2 diabetes. Because they all correlated with each other, they chose to examine one of the SNPs (rs9939609) in detail. When the researchers genotyped this SNP in DNA samples from 19,424 European adults and 10,172 European children, they found that the A allele, which occurs at a frequency of 39 percent, was associated with increased BMI. Homozygous carriers of the A allele had higher BMI than heterozygotes. When they broke the data down by body mass, the A allele was associated with increased risk for both being overweight (BMI >25 Kg/sq m) and obese (BMI >30 Kg/sq m). This allele was not associated with body weight at birth, but it did associate with increased BMI by age seven, and still at age 11 and 14 years in different populations. “We conclude therefore that the FTO SNP rs9939609 is not associated with changes in fetal growth but is associated with changes in BMI and obesity in children by the age of 7 which persists into the pre-pubertal period and beyond,” write the authors. The authors found that FTO is most highly expressed in human pancreatic islets and the brain.

Bruce Lamb and colleagues at Case Western Reserve University, Cleveland, Ohio, report how transgenic mice that express human amyloid-β precursor protein (APP) may serve not only as models of pathology, but also of genetic susceptibility. The R1.40 mouse model of AD expresses a full copy of the human APP gene containing the Swedish mutations that cause familial early onset disease. Curiously, in different inbred strains of mice that harbor the R1.40 transgene, amyloid-β levels can be significantly different, even though APP expression levels are the same. In a C57BL/6J background, for example, Aβ levels are 20 percent higher than in DBA/2J animals, and plaques are apparent by 14 months, whereas at 24 months plaques are non-existent in the DBA/2J animals. To Lamb, this raises the possibility of detecting genetic variations responsible for the different Aβ levels (see also Ryman and Lamb, 2006; Lehman et al., 2003).

In the latest paper, first author Davis Ryman tried to answer this question by carrying out whole-genome SNP mapping of 516 C57BL/6J x DBA/2J crosses. As they report in an advance publication in the Neurobiology of Aging, they found nine markers on chromosome 2 that significantly associate with Aβ levels. They also found suggestions of quantitative trait loci (QTL) on chromosomes 1 and 7. Two of the chromosome 2 markers lie in regions of the mouse genome that are homologous to regions in human chromosomes previously linked to AD. These include a region on chromosome 9q31 (see Myers et al., 2002), 9q34 (see Pericak-Vance et al., 2000), and chromosome 2q22-24 (see Lee et al., 2004). Further analysis of these regions may turn up genetic variations that have a significant impact on AD pathology. As a methodology, the approach may be a welcome aid to human genetic studies. “With the complex genetics of human AD proving remarkably difficult to elucidate, identification of QTLs acting on AD phenotypes in mouse models provides a powerful way to isolate and reproducibly study genetic modifiers that are detected,” conclude the authors.—Tom Fagan


Comments on this content

No Available Comments

Make a Comment

To make a comment you must login or register.


News Citations

  1. SORLA Soars—Large Study Links Gene to Late-onset AD
  2. Brains in Transition: Signposts Point the Way from MCI to Alzheimer Disease
  3. AGEing Neurons Waste Away in Fly Model of Neurodegeneration

Paper Citations

  1. . Systematic meta-analyses of Alzheimer disease genetic association studies: the AlzGene database. Nat Genet. 2007 Jan;39(1):17-23. PubMed.
  2. . Genetic and environmental modifiers of Alzheimer's disease phenotypes in the mouse. Curr Alzheimer Res. 2006 Dec;3(5):465-73. PubMed.
  3. . Genetic background regulates beta-amyloid precursor protein processing and beta-amyloid deposition in the mouse. Hum Mol Genet. 2003 Nov 15;12(22):2949-56. Epub 2003 Sep 23 PubMed.
  4. . Full genome screen for Alzheimer disease: stage II analysis. Am J Med Genet. 2002 Mar 8;114(2):235-44. PubMed.
  5. . Identification of novel genes in late-onset Alzheimer's disease. Exp Gerontol. 2000 Dec;35(9-10):1343-52. PubMed.
  6. . Fine mapping of 10q and 18q for familial Alzheimer's disease in Caribbean Hispanics. Mol Psychiatry. 2004 Nov;9(11):1042-51. PubMed.

Further Reading