Association between variants in IDE-KIF11-HHEX and plasma amyloid β levels
Article Outline
- Abstract
- 1. Introduction
- 2. Methods
- 3. Results
- 4. Discussion
- Disclosure statement
- Acknowledgements
- Supplementary data
- References
- Copyright
Abstract
Genetic linkage and association studies in late-onset Alzheimer’s disease (LOAD) or its endophenotypes have pointed to several regions on chromosome 10q, among these the ∼ 250 kb linkage disequilibrium (LD) block harboring the genes IDE, KIF1, and HHEX. We explored the association between variants in the genomic region harboring the IDE-KIF11-HHEX complex with plasma Aβ40 and Aβ42 levels in a case-control cohort of Caribbean Hispanics. First, we performed single marker linear regression analysis relating the individual single nucleotide polymorphisms (SNPs) with plasma Aβ40 and Aβ42 levels. Then we performed 3-SNP sliding window haplotype analyses, correcting all analyses for multiple testing. Out of 32 SNPs in this region, 3 SNPs in IDE (rs2421943, rs12264682, rs11187060) were associated with plasma Aβ40 or Aβ42 levels in single marker and haplotype analyses after correction for multiple testing. All these SNPs lie within the same LD block, and are in LD with the previously reported haplotypes. Our findings provide support for an association in the IDE region on chromosome 10q with Aβ40 and 42 levels.
Keywords: Amyloid beta, Alzheimer's disease, Genetics, Insulin-degrading enzyme
1. Introduction
Risk loci for late-onset Alzheimer's disease (LOAD) may be present on chromosome 10q having been identified using case-control status, age-of onset, or plasma amyloid beta (Aβ) levels as the phenotype. One of the functionally plausible candidate genes lying within the genetic region showing evidence for association or linkage reported by these studies is IDE occurring in a ∼ 250 kb haplotype block with KIFF11 and HHEX. Following the initial report on linkage and association of markers around IDE with LOAD (Bertram et al., 2000), some studies that used LOAD as the phenotype did not find an association (Cousin et al., 2009, Reiman et al., 2007) but other independent studies identified haplotypes spanning the IDE-KIFF-HHEX complex that show association with LOAD risk or intermediate LOAD phenotypes (Ertekin-Taner et al., 2004, Prince et al., 2003), including cerebrospinal fluid (CSF) tau levels, Mini Mental State Examination (MMSE) scores, senile plaque and neurofibrillary tangle density, and age-at-onset (Prince et al., 2003). The same haplotypes were associated with plasma Aβ levels in 24 extended Caucasian LOAD families, and with LOAD status in 2 independent case control series (Ertekin-Taner et al., 2004). Three single nucleotide polymorphisms (SNPs) in IDE, that are in linkage disequilibrium (LD) with these haplotypes, have been shown to influence IDE expression in LOAD brains (Zou et al., 2010). The objective in the present paper was to confirm or refute a role of genetic variation in the IDE-KIF11-HHEX complex on chromosome 10q in variation of plasma Aβ40 and Aβ42 levels, the main putative culprits in LOAD.
2. Methods
2.1. Participants
We selected unrelated affected and unaffected individuals from Caribbean Hispanics participating in a population-based study in northern Manhattan (Tang et al., 1998) and single individuals from Caribbean Hispanic families multiply affected by LOAD (Romas et al., 2002). The final analytic sample consisted of 454 Caribbean Hispanic subjects (160 cases, 294 controls) with information on plasma Aβ40 and 42 levels. Dementia diagnosis was based on Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM-IV) criteria and National Institute of Neurological and Communicative Diseases and Stroke/Alzheimer's Disease and Related Disorders Association (NINCDS-ADRDA) criteria.
2.2. Plasma Aβ40 and Aβ42 levels
Plasma was obtained at baseline and follow-up, and was stored within 2 hours after collection at −70 °C. Aβ40 and Aβ42 levels were measured using a combination of monoclonal antibody 6E10 (specific to an epitope present on 1 to 16 amino acid residues of Aβ) and rabbit antibodies specific for Aβ40 (R162) and Aβ42 (R165) in a double-antibody sandwich enzyme-linked immunosorbent assay (ELISA). The detection limit for this assay was 5 pg/mL for Aβ40 and 10 pg/mL for Aβ42.
2.3. Genotyping
Genotyping was performed using the HumanHap650Y BeadChip from Illumina, Inc. Included in the present analyses were 32 SNPs spanning the IDE-KIF11-HHEX region (Supplemental Table 1).
2.4. Statistical methods
Regression analyses were performed individually relating the 32 SNPs with Aβ40 and Aβ42 levels, adjusting for sex, age-at-onset or age-at-examination, apolipoprotein E APOE genotype, education, and population stratification. Finally, we performed 3-SNP sliding window haplotype analyses for Aβ40 and 42 levels. We did not correct for multiple testing as all explored SNPs are in strong LD (Fig. 1) and marker-phenotype associations are therefore not independent (Nyholt, 2001).
Fig. 1.
Linkage disequilibrium (LD) pattern of the 10q region spanning IDE, KIF11, and HHEX in the present sample. The boxes indicate SNPs significant in the present analyses, the black arrows indicate SNPs significant in the study by Ertekin-Taner (2004). The green arrow indicates the SNPs that have been shown to influence IDE expression in LOAD brain samples in the study by Zou et al. (2010)
3. Results
All SNPs were in Hardy-Weinberg equilibrium (HWE). The mean age of the sample was 82.2 ± 6.3 years. The A-allele of IDE SNP rs2421943 was associated with significantly higher Aβ40 levels and the A allele of IDE SNP rs12264682 was associated with significantly lower Aβ40 levels (Table 1, Supplemental Table 2). While the association of rs2421943 with Aβ40 was driven by LOAD cases, the association of rs12264682 was driven by the unaffected persons.
Table 1. Association between SNPs in the IDE-KIF11-HHEX complex and plasma Aβ40 and Aβ42 levels
| All (n = 454) | Cases (n = 160) | Controls (n = 294) | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Gene | SNP | BP | A1l | Aβ (mean) by genotype | BETA | SE | p | Aβ (mean) by genotype | BETA | SE | p | Aβ (mean) by genotype | BETA | SE | p |
| Aβ40 | |||||||||||||||
 IDE | Rs11187060 | 94294112 | T | T/T, T/C, C/C (88.7, 86.8, 94.5) | −4.5 | 3.9 | 0.3 | T/T, T/C, C/C (100.7, 90.1, 108) | −8.1 | 6.3 | 0.2 | T/T, T/C, C/C (79.9, 85.1, 86.7) | −2.4 | 5.0 | 0.6 |
| IDE | Rs2421943 | 94301795 | A | A/A, A/G, G/G (92.62 , 98.1, 83) | 9.4 | 3.8 | 0.01 | A/A, A/G, G/G (112.2, 105.8, 91.43) | 12.3 | 6.5 | 0.05 | A/A, A/G, G/G (83.58, 93.53, 78.66) | 6.8 | 4.6 | 0.14 |
| IDE | Rs12264682 | 94312407 | A | A/A, A/C, C/C (60.55, 71.1, 92) | −21.8 | 9.1 | 0.02 | A/A, A/C, C/C (–, 82.75, 101.5) | −21.9 | 14.9 | 0.15 | A/A, A/C, C/C (60.55, 61.91, 87.02) | −22.7 | 11.7 | 0.05 |
| Aβ42 | |||||||||||||||
| IDE | Rs11187060 | 94294112 | T | T/T, T/C, C/C (36.23, 36.39, 39.83) | −2.3 | 1.8 | 0.2 | T/T, T/C, C/C (37.02, 38.65, 46.19) | −5.6 | 2.9 | 0.05 | T/T, T/C, C/C (79.93, 85.1, 86.66) | −0.5 | 2.3 | 0.82 |
| IDE | Rs2421943 | 94301795 | A | A/A, A/G, G/G (40.4, 37.7, 37.4) | 1.3 | 1.7 | 0.4 | A/A, A/G, G/G (46.7, 43.6, 39.7) | 3.9 | 2.9 | 0.1 | A/A, A/G, G/G (37.4, 34.3, 36.2) | 0.2 | 2.1 | 0.9 |
| IDE | Rs12264682 | 94312407 | A | A/A, A/C, C/C (39.1, 30.9, 41.4) | −7.1 | 4.2 | 0.09 | A/A, A/C, C/C (-, 32.2, 42.9) | −11.5 | 6.8 | 0.09 | A/A, A/C, C/C (39.1, 30.0, 40.1) | −4.6 | 5.3 | 0.3 |
Haplotype analyses confirmed these findings. In 3-SNP sliding window haplotype analyses, haplotypes that included the A alleles of SNP rs2421943 (rs11187062 rs11187064 rs2421943: TTA, β = 7.8, p = 0.04; rs11187064 rs2421943 rs7908111: TAG, β = 7.9, p = 0.03; and rs2421943 rs7908111 rs1999763: AGG, β = 7.8, p = 0.03) were associated with higher Aβ40 levels, and halotypes that included the A allele of SNP rs12264682 (rs7908111 rs1999763 rs12264682: GGA, β = −20.5, p = 0.02; rs1999763 rs12264682 rs7100623: GAC, β = −20.5, p = 0.02; and rs12264682 rs7100623 rs6583826: ACG, β = −22.6, p = 0.01) were associated with lower Aβ40 levels.
The association with Aβ42 levels differed (Table 1). In these analyses, the T allele of IDE SNP rs11187060 was associated with lower Aβ42 levels in LOAD cases. Also this finding was confirmed by haplotype analyses (rs11187025 rs7078413 rs11187060: CCT, β = −9.5, p = 0.02; rs7078413 rs11187060 rs11187062: CTT, p = 0.07; β = −5.6, rs11187060 rs11187062 rs11187064: TTT, β = −8.1, p = 0.01). Of note, the directions of effects for all 3 SNPs were similar for Aβ40 and 42 levels, and all 3 SNPs lie in the same haplotype block as the previously reported SNPs (Fig. 1) (Ertekin-Taner et al., 2004, Prince et al., 2003, Zou et al., 2010). Adjustment for disease duration did not change the associations. When we used Alzheimer's disease AD as the phenotype to explore whether any of the identified SNPs is also associated with the disease phenotype, the C allele of SNP rs11187062 was associated with significantly lower AD risk (odds ratio [OR] = 0.54 ± 0.3, p = 0.03). Of note, this SNP is adjacent and only 2 kb apart from rs11187060 that is associated with Aβ 42 levels.
4. Discussion
Three IDE SNPs (rs2421943, rs12264682, rs11187060) were significantly associated with changes in plasma Aβ40 or Aβ42 levels in single marker and haplotype analyses. We used nominal p-values as all assessed SNPs are in strong LD and marker-phenotype associations are therefore not independent (Nyholt, 2001). Of note, while these SNPs have not been assessed in previous studies, they also lie within the same LD block as the haplotypes reported in previous Caucasian studies (Ertekin-Taner et al., 2004, Prince et al., 2003). In addition, they are in strong LD (D' > 90) with the 3 SNPs that have previously been shown to influence IDE expression in brain samples of 200 LOAD cases (Fig. 1) (Zou et al., 2010).
While SNPs rs2421943 and rs12264682 were associated with changes in Aβ 40 levels, rs11187060 was associated with changes in Aβ42 levels. The directions of associations of all 3 SNPs were consistent for Aβ 40 and Aβ 42 levels. A likely explanation for the differences in the strengths of the associations of the individual SNPs with Aβ 40 and 42 levels is, that Aβ 42 is a stronger surrogate of pathological changes underlying AD than Aβ 40 which is rather a marker of aging. This note is supported by the fact that the association of SNP rs11187060 with Aβ 42 levels is 10-fold stronger in cases than controls (β: −5.6 vs. −0.5), and that this SNP is in close proximity (∼ 2 kb) and strong LD with SNP rs11187062 that is associated with AD. Alternative explanations for differences in the strengths of the associations with Aβ 40 and 42 levels are differences in allele frequencies or power.
IDE binds and degrades Aβ40 and Aβ42 (Perez et al., 2000), and this Aβ degrading activity has been shown to be lower in AD brains than in controls (Perez et al., 2000). In IDE knock-out mice, brain Aβ levels are elevated (Farris et al., 2003), suggesting that IDE activity is 1 of several factors determining the amount of brain Aβ in vivo. Enhanced IDE activity in IDE and APP double transgenic mice decreases their brain Aβ levels, and reduces the formation of AD pathology (Leissring et al., 2003). Finally, polymorphisms in IDE may also contribute to the risk of type 2 diabetes (Rudovich et al., 2009), which itself is associated with LOAD. Taken together the findings reported here support the possibility that the IDE-KIF11-HHEX region on chromosome 10q may contain genetic variants that modify Aβ40 and 42 levels.
Disclosure statement
None of the authors has any actual or potential conflicts of interest.
Appropriate approval and procedures were used concerning human subjects and animals.
Acknowledgements
This work was supported by grants from the National Institute of Health and the National Institute on Aging: R37-AG15473, P01-AG07232 (RM), Paul B. Beeson Career Development Award: K23AG034550 (CR), The Blanchette Hooker Rockefeller Foundation and the Charles S. Robertson Gift from the Banbury Fund (RM). The laboratory under the direction of Dr. St George-Hyslop received additional support from the Alzheimer Society of Canada, Japan-Canada and Canadian Institutes of Health Research Joint Health Research Program (ER), the Canadian Institutes of Health Research, Alzheimer Society of Ontario, Howard Hughes Medical Institute, and The Wellcome Trust.
Supplementary data
Supplemental Table 1.
Supplemental Table 2.
References
- . Evidence for genetic linkage of Alzheimer's disease to chromosome 10q. Science. 2000;290:2302–2303
- . No replication of genetic association between candidate polymorphisms and Alzheimer's disease. Neurobiol. Aging. 2009;Epub ahead of print, Nov. 2
- . Genetic variants in a haplotype block spanning IDE are significantly associated with plasma Abeta42 levels and risk for Alzheimer disease. Hum. Mutat. 2004;23:334–342
- . Insulin-degrading enzyme regulates the levels of insulin, amyloid beta-protein, and the beta-amyloid precursor protein intracellular domain in vivo. Proc. Natl. Acad. Sci. U. S. A. 2003;100:4162–4167
- . Enhanced proteolysis of beta-amyloid in APP transgenic mice prevents plaque formation, secondary pathology, and premature death. Neuron. 2003;40:1087–1093
- . Genetic case-control association studies − correcting for multiple testing. Hum. Genet. 2001;109:564–567
- . Degradation of soluble amyloid beta-peptides 1–40, 1–42, and the Dutch variant 1–40Q by insulin degrading enzyme from Alzheimer disease and control brains. Neurochem. Res. 2000;25:247–255
- . Genetic variation in a haplotype block spanning IDE influences Alzheimer disease. Hum. Mutat. 2003;22:363–371
- . GAB2 alleles modify Alzheimer's risk in APOE epsilon4 carriers. Neuron. 2007;54:713–720
- . Familial Alzheimer disease among Caribbean Hispanics: a reexamination of its association with APOE. Arch. Neurol. 2002;59:87–91
- . Polymorphisms within insulin-degrading enzyme (IDE) gene determine insulin metabolism and risk of type 2 diabetes. J. Mol. Med. 2009;87:1145–1151
- . The APOE-epsilon4 allele and the risk of Alzheimer disease among African Americans, whites, and Hispanics. JAMA. 1998;279:751–755
- . Gene expression levels as endophenotypes in genome-wide association studies of Alzheimer disease. Neurology. 2010;74:480–486
PII: S0197-4580(10)00312-X
doi:10.1016/j.neurobiolaging.2010.07.005
© 2012 Elsevier Inc. All rights reserved.
