Alzheimer Disease 6



Alzheimer disease (AD) is a neurodegenerative disorder characterized by subtle onset of memory loss followed by a slowly progressive dementia. The great majority of AD cases are of late onset (LOAD) after age 65 years. LOAD shows complex, nonmendelian patterns of inheritance, and most likely results from the combined effects of variation in a number of genes as well as from environmental factors (summary by Grupe et al., 2006).

The Alzheimer disease-6 (AD6) designation refers to a susceptibility locus on chromosome 10q. Although significant associations with several candidate genes on chromosome 10 have been reported, these findings have not been consistently replicated, and they remain controversial (Grupe et al., 2006).

For a discussion of genetic heterogeneity of Alzheimer disease, see 104300.

Clinical Features

Bassett et al. (2005) performed functional magnetic resonance imaging (fMRI) of 9 asymptomatic female offspring from AD families linked to the chromosome 10q region and 6 females from AD families unlinked to this region. These individuals were on average 11.3 years younger than the onset age of their parents. All were APOE4-(107741) negative. During memory encoding tasks, individuals from the 10q-linked families showed more extensive activation in the bilateral temporal and frontal lobes as well as in the thalamus and right parahippocampal gyrus, compared to offspring from the unlinked families who showed only unilateral frontal and temporal lobe activation. These differences were seen despite similar memory performance. Bassett et al. (2005) suggested that there may be different cognitive strategies representing different regional limitations in the brain between AD families linked and unlinked to 10q.


Bertram et al. (2000) performed parametric and nonparametric linkage analyses of 7 genetic markers on chromosome 10q, 6 of which map near the insulin-degrading enzyme gene (IDE; 146680), in 435 multiplex Alzheimer disease families. These analyses revealed significant evidence of linkage for adjacent markers (D10S1671, D10S583, D10S1710, and D10S566), which was most pronounced in late-onset AD (LOAD) families. Furthermore, Bertram et al. (2000) found evidence for allele-specific association between the putative disease locus and marker D10S583, which is located within 195 kb of the IDE gene. Using an autosomal dominant model, the maximum 2-point parametric lod score was 3.3 at theta of 0.22 at marker D10S583. In late-onset families, the highest lod score was 3.4 with theta of 0.16 at marker D10S1671.

Myers et al. (2000) performed a 2-stage genomewide screen in sib pairs with late-onset Alzheimer disease to detect susceptibility loci other than APOE and identified an Alzheimer disease locus on chromosome 10q. Using multiple stages of analysis, they achieved a lod score of 3.83 using 429 sib pairs, close to marker D10S1225. This locus modifies risk for Alzheimer disease independent of APOE genotype. Myers et al. (2000) estimated a relative risk to sibs for this locus to be about equivalent to that for APOE.

Ertekin-Taner et al. (2000) found that there is a quantitative trait locus for high plasma amyloid beta-42 levels that maps to the same region of 10q24 as Alzheimer disease. Plasma amyloid beta-42 is invariably elevated in early-onset familial Alzheimer disease, and it is also increased in first-degree relatives of patients with typical late-onset Alzheimer disease. To detect LOAD loci that increase amyloid beta-42, Ertekin-Taner et al. (2000) used plasma amyloid beta-42 as a surrogate trait and performed linkage analysis on extended Alzheimer disease pedigrees identified through a LOAD patient with extremely high plasma amyloid beta. Ertekin-Taner et al. (2000) reported linkage to chromosome 10 with a maximum lod score of 3.93 at 81 cm, close to D10S1225. They noted that linkage to the same region was obtained in a genomewide screen of LOAD sib pairs by Myers et al. (2000). These results provided strong evidence for a novel LOAD locus on chromosome 10 that acts to increase amyloid beta.

By linkage analysis of 5 Amish families from the midwestern U.S. in which 49 individuals had late-onset dementia or mild cognitive impairment, Hahs et al. (2006) obtained 2-point lod scores greater than 1.5 at marker D10S2327 on 10q22 (lod scores of 2.42 and 1.42). The results were obtained under a model assuming autosomal recessive inheritance.

Grupe et al. (2006) reported evidence suggesting a genetic association between SNP rs498055 on 10q24, upstream of the SORBS1 gene (605264) and located near a putative homolog of ribosomal protein S3a (RPS3A; 180478), and risk of LOAD in 4 of 6 independent case-control samples. However, Bertram et al. (2006) could not corroborate this association in 2 independent family samples. Furthermore, analysis of 3 family-based and 1 case-control datasets by Liang et al. (2008) showed that rs498055 was not associated with an increased risk of late-onset Alzheimer disease.

In a genome screen of individuals from an isolated population from the southwestern area of the Netherlands, ascertained as part of the Genetic Research in Isolated Populations (GRIP) program, Liu et al. (2007) confirmed the AD locus at 10q22-q24 (AD6). Overall, multipoint analysis revealed 4 significant and 1 suggestive linkage peak. Liu et al. (2007) next tested for association between cognitive function as an endophenotype of AD and 4,173 single-nucleotide polymorphisms in the linked regions in an independent sample consisting of 197 individuals from the GRIP region. After adjusting for multiple testing, they detected significant association with cognitive function at 10q22-q24. A study of potential disease-causing genes in the region pointed to the HTR7 (182137), MPHOSPH1 (605498), and CYP2C (124020) cluster.

By genomewide analysis of 3 large cohorts totaling 723 affected relative pairs with late-onset AD, Hamshere et al. (2007) found linkage to a locus on chromosome 10q21.2 (lod score of 3.3 at D10S464). There was no evidence to suggest that more than 1 locus was responsible for the linkage to 10q21.2, although this region may harbor more than 1 susceptibility gene. Evidence for an interaction was observed between loci on chromosomes 10 and 19.

By genotyping SNPs spanning 80.2 Mb on chromosome 10q in a family-based data set containing 1,337 discordant sib pairs from 567 multiplex families and an independent case-control data set containing 483 cases and 879 controls, Liang et al. (2007) found that 22 SNPs in 5 candidate genes yielded significant association results in at least 1 data set, including SNPs in CTNNA2 (p = 0.03) in both data sets. However, the results did not converge with linkage analysis, suggesting that there is more extensive heterogeneity on chromosome 10 than had been appreciated.

Liang et al. (2009) examined 28 genes on chromosome 10 for association with LOAD in a Caucasian case-control cohort of 506 cases and 558 controls. Two SNPs in 2 genes remained significant after correction for multiple testing: rs10508533 in PTPLA (610467) on 10p14-p13 (p = 0.0022) and rs17277986 in SORCS1 (606283) on 10q23 (p = 0.0025). The SORC1 association was not replicated in the validation and family-trio data sets. Multifactor dimensionality reduction yielded a significant association with the SORCS1 SNP rs17277986 in females (p = 0.00002; OR, 1.7).

Reitz et al. (2011) analyzed genotypic data for 16 SNPs in the SORCS1 gene from 6 independent data sets totaling 2,809 patients and 3,482 controls. Inherited variants in SORCS1 were associated with AD in 5 of the 6 datasets, but the specific alleles and direction of association differed. Metaanalysis of single-SNP association data yielded significant associations in all datasets (p = 0.001 to 0.049). Using microarray gene expression and RT-PCR, Reitz et al. (2011) found decreased expression of SORCS1 in the amygdala of 19 AD brains compared to 10 controls. Expression levels appeared to correlate with certain disease-associated SNPs. In HEK293 cells with an AD-associated APP mutation (104760.0008), overexpression of SORCS1 resulted in a significant decrease in amyloid-beta-40 and -beta-42 secretion, whereas suppression of SORCS1 in HEK293 cells increased beta-amyloid-40 levels. The findings indicated that SORCS1 can influence APP processing, and Reitz et al. (2011) suggested that variation in the SORCS1 gene may be associated with risk of LOAD.

Molecular Genetics

IDE Gene

Abraham et al. (2001) identified 8 single-nucleotide polymorphisms (SNPs) in the IDE gene. They found no association between late-onset Alzheimer disease and any of the individual SNPs or with any haplotypes.

Prince et al. (2003) used a SNP genetic association strategy to investigate Alzheimer disease in relation to a 480-kb region encompassing the IDE gene. They interpreted the results as providing 'substantial' evidence that genetic variation within or very close to IDE impacts both disease risk and traits related to the severity of Alzheimer disease.

Grupe et al. (2006) reported no association between AD and the IDE gene.

In 176 patients with LOAD, Zou et al. (2010) measured cerebellar expression levels of 12 LOAD candidate genes. Genomewide association analysis using 564 cis-SNPs identified a significant association between rs7910977, located 4.2 kb beyond the 3-prime end of the IDE gene, and 2-fold higher cerebellar expression levels of IDE (p = 2.7 x 10(-8)). In the combined group of 565 patients with LOAD, the minor allele of rs7910977 was associated with 1.45-fold increased cerebellar expression of IDE (p = 2.6 x 10(-5)). When results from 7 independent case-control series involving 2,280 LOAD cases and 2,396 controls were combined, the minor allele of rs7910977 yielded a protective effect (OR, 0.81; p = 0.0046).


Finckh et al. (2003) noted that the urokinase gene (PLAU; 191840) maps to the AD6 critical region on chromosome 10. They genotyped a frequent C/T SNP of the PLAU gene, which results in a pro141-to-leu (P141L) mutation (rs2227564; 191840.0001), in 347 patients with LOAD and 291 control subjects. LOAD was associated with a CC genotype in the whole sample as well as in all subsamples stratified by gender or APOE4 (107741) carrier status. The odds ratio for LOAD due to a CC genotype was 1.89. Finckh et al. (2003) suggested that PLAU is a susceptibility gene for LOAD, with allele C (P141) being a recessive risk allele and allele T (L141) conferring protection.

Ertekin-Taner et al. (2005) found significant association between PLAU 5-SNP haplotypes and LOAD in 3 independent series of 204, 148, and 113 AD patients with matched controls. Plasma A-beta-42 levels among individuals over 50 years old in 10 extended LOAD families were also significantly associated with PLAU haplotypes (p = 0.005). The CT and TT genotypes of rs2227564 were associated with LOAD (p = 0.05) and with age-dependent elevation of plasma A-beta-42 in 24 extended LOAD families (p = 0.0006). In Plau-knockout mice, plasma A-beta-42 and A-beta-40 levels, but not levels in brain, were significantly elevated in an age-dependent manner.

Grupe et al. (2006) reported no association between AD and the PLAU gene.


Alpha-T-catenin (CTNNA3; 607667) is a binding partner of beta-catenin (CTNNB1; 116806). In turn, CTNNB1 interacts with PSEN1 (104311), which has many mutations that elevate A-beta-42 and cause early-onset familial AD. The CTNNA3 gene maps to the AD6 region for which Ertekin-Taner et al. (2000) previously reported significant linkage to LOAD. Ertekin-Taner et al. (2003) identified 2 intronic CTNNA3 SNPs in strong linkage disequilibrium that showed highly significant association with plasma A-beta-42 levels in 10 extended LOAD families.

Grupe et al. (2006) reported no association between AD and the CTNNA3 gene.

In a study of 363 Japanese LOAD patients, with validation in a second set of 336 Japanese LOAD patients, Miyashita et al. (2007) found that 7 SNPs, spanning about 38 kb in intron 9 of the CTNNA3 gene, were associated with late-onset AD in females (p values of 5.95 x 10(-6) to 7.66 x 10(-4) after correction). Multiple logistic regression analysis of a total of 2,762 subjects (1,313 LOAD patients and 1,449 controls) demonstrated significant interaction between rs713250 and female patients but not with male patients. The data suggested that variation in the CTNNA3 gene may affect LOAD through a female-specific mechanism independent of the APOE4 allele.


Dreses-Werringloer et al. (2008) identified a 394C-T SNP (rs2986017), resulting in a pro86-to-leu (P86L) substitution, in the CALHM1 gene (612234) on chromosome 10q24.33. In 5 independent case-control populations including a total of 2,043 patients with late-onset AD and 1,361 controls, the authors observed an association between the T allele of this SNP and disease (odds ratio of 1.44, p = 2 x 10(-10)). In vitro functional expression studies showed that the polymorphic P86L channel exhibited decreased CALHM1-induced calcium permeability, resulting in decreased cytosolic calcium levels, as well as impaired ability to inhibit generation of insoluble beta-amyloid. The findings indicated a role for calcium signaling in APP (104760) processing and suggested that variations in the CALHM1 gene may influence susceptibility to late-onset AD.

Beecham et al. (2009) were unable to replicate the findings of Dreses-Werringloer et al. (2008) in a case-control study of 510 patients with late-onset AD and 524 controls. Beecham et al. (2009) found no association with rs2986017 despite adequate power.