PSEN1 and PSEN2
Many rare missense and other gain of function mutations increasing levels ofAP42
APOE e4 ' risk allele
Figure 2.6 Factors influencing the accumulation of Ap42, thought to be important in the pathogenesis of Alzheimer disease based on the amyloid hypothesis. Adapted by permission from Macmillan Publishers Ltd: Nature Genetics (Hardy 2006), copyright 2001.
with early onset disease, the APOE e4 allele is thought to play a key role in the accumulation of Ap protein. A small number of cases, estimated at less than 1%, are caused by trisomy 21 as part of Down syndrome (Box 3.3), the characteristic neuropathological changes of Alzheimer's disease being seen among almost all individuals with Down syndrome after 40 years.
This review falls into two sections. Firstly, a discussion of early onset familial Alzheimer's disease, which illustrates many of the issues and approaches inherent to investigating a mendelian trait. Secondly, a consideration of late onset Alzheimer's disease and the role of the APOE e4 allele, which illustrates the success of a candidate gene approach in defining genetic determinants of common multifactorial disease.
2.5.1 Early onset familial Alzheimer's disease: rare variants underlying a mendelian trait
The prevalence of early onset familial Alzheimer's disease in the French city of Rouen was estimated to be
41.2 per 100 000 people at risk aged between 41 and 60 years; of these 13% showed clear cut autosomal inheritance (Campion et al. 1999). Study of such families showing segregation consistent with autosomal dominant i nheritance was highly informative for both linkage analysis and identification of specific mutations. The link with Down syndrome was highly suggestive that a risk locus may be present on chromosome 21. Aß was thought to play a key role in disease pathogenesis and the determination of the peptide sequence from patients with Alzheimer's disease and with Down syndrome (Glenner and Wong 1984a, 1984b; Masters et al. 1985), including material purified from senile plaques, allowed cloning of the APP gene (Goldgaber et al. 1987; Kang et al. 1987; Robakis et al. 1987; Tanzi et al. 1987). The gene was mapped to chromosome 21q21 and, together with initial evidence of linkage among families with early onset familial Alzheimer's disease, further heightened interest in a potential disease susceptibility locus on chromosome 21 dubbed 'familial Alzheimer disease 1', AD1 (OMIM 104300).
The first pathogenic sequence variation of APP was reported among Dutch families with a rare autosomal dominant disorder involving amyloid protein, called hereditary cerebral haemorrhage with amyloidosis (OMIM 609065). Levy and colleagues sequenced the exons encoding the Ap domain of the protein in two patients and defined a G to C transversion which resulted in the substitution of glutamic acid to glutam-ine (p.E693Q) (Levy et al. 1990).
Among patients with early onset familial Alzheimer's disease showing autosomal inheritance, Goate and colleagues used additional markers on the long arm of chromosome 21 to further define linkage to chromosome 21 which included APP (Goate et al. 1991). Sequencing of exon 17, involved in encoding the Ap domain, showed a G to A transition cosegregating with disease in the family kindred analysed; two individuals with the same variant were defined in a second affected family. The nucleotide substitution was a further missense mutation, changing the amino acid sequence in the encoded protein from valine to isoleucine at position 717 (c.2149G>A, p.V717I) (Goate et al. 1991).
A total of 29 pathogenic mutations have now been identified at the APP gene, the majority being missense mutations, and are listed at the Alzheimer Disease and Frontotemporal Dementia Mutation Database (www.molgen.ua.ac.be/ADMutations) (Cruts and Van Broeckhoven 1998a, 1998b). Cumulatively, however, APP sequence variants account for a minority of early onset familial disease cases (estimated at 10-15%) and overall a very small number of Alzheimer's disease cases. They appear to be gain of function mutations, resulting in increased levels of the Ap42 protein thought to be central to the pathogenesis of Alzheimer's disease (Scheuner et al. 1996).
A gene dosage effect is suggested by the neuro-pathology seen in Down syndrome, with trisomy of chromosome 21 thought to result in increased lifetime accumulation of Ap. It is striking that gene duplication events involving APP have now been identified among families with early onset familial Alzheimer's disease showing that overexpression of the normal protein can lead to disease (Rovelet-Lecrux et al. 2006). Among 65 families, individuals from five families were defined with duplications ranging in size from 0.58 to 6.37 Mb, encompassing five to 12 known genes including APP;
the affected families included individuals with evidence of cerebral amyloid angiopathy suggesting a potentially more complex phenotype.
It was clear that not all families with early onset Alzheimer's disease showed linkage to chromosome 21q21 (St George-Hyslop et al. 1990). A number of groups found evidence instead of linkage to chromosome 14q24.3 (Schellenberg et al. 1992; Van Broeckhoven et al. 1992; Campion et al. 1995), leading to definition of the early onset familial Alzheimer disease-3 (AD3) subtype (OMIM 607822) due to rare mutations affecting the PSEN1 gene. Analysis of short tandem repeat markers using the lod score method resolved strong linkage for 14q24.3. A positional cloning approach was then used by Sherrington and colleagues to define a novel gene containing specific disease-associated sequence variants (Sherrington et al. 1995). Additional markers helped resolve the region of linkage with haplotypic analysis defining recombination boundaries. The authors hybridized cDNA generated from human brain mRNA to a series of genomic DNA fragments to perform transcription mapping, then looked for sequence differences. Analysis of affected family members from tightly linked pedigrees versus normal brain samples resolved a novel gene encoding the 'S182 protein' with five nonsynonymous sequence variants specific to affected individuals (p.M146L, p.A246E, p.286V, p.C410Y) (Sherrington et al. 1995).
The novel gene now called PSEN1, at chromosome 14q24.3, encodes presenilin 1 which is part of the gamma secretase complex involved in cleavage of the amyloid precursor protein to form Aß protein. At least 168 different pathogenic sequence variants have been reported at PSEN1, the vast majority being missense mutations (Alzheimer Disease and Frontotemporal Dementia Mutation Database). Overall, mutations at AD3 involving PSEN1 account for 20-70% of cases of early onset familial Alzheimer's disease. The PSEN1 variants are thought to be gain of function mutations, increasing production of the Aß42 peptide (Scheuner et al. 1996; De Strooper et al. 1998). Reports suggest the p.M146L variant has been associated with calcium dys-regulation affecting Aß processing (Cheung et al. 2008) while p.V97L was associated with increased Aß protein through an effect on insulin degrading enzyme (Qin and Jia 2008).
A further gene locus associated with early onset famil- et al. 1995). Volga Germans are thought to have settled ial Alzheimer's disease was reported in 1995 by two inde- in the Volga region of Russia in the 1760s but to have pendent groups taking different approaches (Levy-Lahad remained genetically isolated, emigrating to the United et al. 1995; Rogaev et al. 1995). Rogaev and colleagues States between 1870 and 1920. A founder effect had used their finding of the novel 'S182' gene (PSEN1) to been proposed within this population group who show identify homologous genes. They screened cDNA librar- a number of families with early onset familial Alzheimer's ies and identified a 2.3 kb transcript which they mapped disease inherited in an autosomal dominant manner (Fig.
to chromosome 1. The new gene (now called PSEN2) 2.7), the disease manifesting on average at 55 years of encoded a protein with 63% amino acid identity to age (in contrast to AD3 at 45 years of age) (Bird et al.
presenili n-1 and the authors sought to identify under- 1988, 1996). Analysis of 162 markers showed linkage to lying variants associated with disease among 23 family chromosome 1q31-q34 (Levy-Lahad et al. 1995), the pedigrees where variation at APP or PSEN1 had been early onset familial Alzheimer locus 'AD4' (OMIM 606889)
excluded. Two missense mutations were identified: an A in which PSEN2 was identified by Rogaev and colleagues.
to G substitution leading to a methionine to valine sub- Mutations of PSEN2 are rare and account for a very small stitution at codon 239 (p.M239V) was found in all four proportion of cases of early onset familial Alzheimer's affected members of an Italian origin extended pedigree; disease; to date ten missense mutations of PSEN2 have and a further mutation in three of four pedigrees of Volga been identified and associated with Alzheimer's disease
German origin (p.N141I) (Rogaev et al. 1995). among 18 families, some with incomplete penetrance
By contrast, Levy-Lahad and colleagues adopted a (Alzheimer Disease and Frontotemporal Dementia linkage-based approach to identify the same gene locus Mutation Database). Like PSEN1 and APP mutations, the by analysing seven related families of Volga German consequences are thought to relate to increased AP42
ancestry, which included 94 affected people (Levy-Lahad accumulation.
67| 45 I 85 I 4^ 501 491 40 I#501 56 1.43 I 45 L 70 I 45 L 541 64 « 54 42 55 47 81
534 5i 46i 4i
Figure 2.7 Family pedigree of Volga German ancestry showing autosomal dominant segregation for early onset familial Alzheimer disease (AD). Redrawn from Levy-Lahad et al. (1995), reprinted with permission from AAAS.
2.5.2 APOE e4 and late onset Alzheimer's disease
The association of a particular isoform of the ApoE protein with age of onset in late onset Alzheimer's disease is a remarkable story. In the context of common multifactorial disease traits, it was an early example of the potential power of a candidate gene approach yet has proved subsequently in many respects to be unusual: this is a common variant associated with a large effect size whose association with disease susceptibility has been robustly replicated. Biological plausibility as a candidate gene combined with some evidence of linkage to the region of chromosome 19q where the APOE gene is found (19q13.2) led to reports in 1993 of association between the e4 allele and late onset Alzheimer's disease. Pericak-Vance and colleagues had reported in 1991 evidence that familial late onset Alzheimer's disease was linked to the proximal region of the long arm of chromosome 19, in contrast to early onset familial disease for which they found evidence of linkage to chromosome 21. Linkage studies in late onset disease were more problematic due to the advanced age of cases and lack of clear mendelian segregation, but they were able to derive some data using affected pedigree member linkage analysis looking for differences in independent segregation between markers and disease (Pericak-Vance et al. 1991).
The linkage data were suggestive but the findings that ApoE protein was a significant component of amyloid plaques, neurofibrillary tangles, and vascular amyloid highlighted the biological plausibility of diversity involving the APOE gene in disease pathogenesis. APOE had been found to be strongly expressed in the brain by astrocytes, with increased levels following injury or in chronic neurodegenerative disease including in Alzheimer's disease. Strittmatter and colleagues found that when one patient from each of 30 predominantly late onset families with Alzheimer's disease was analysed, the frequency of a particular APOE allele was significantly increased among cases compared to 91 age-matched controls (Strittmatter et al. 1993). Amino acid sequencing had previously determined three major isoforms of ApoE, e3, e4 and e2, which differed by particular amino acid substitutions (Fig. 2.8) (Mahley 1988). These could be determined by genotyping using different methodologies including restriction enzyme digestion with the Hhal restriction enzyme (Hixson and Vernier 1990). It was the e4 allele that was significantly increased in frequency among cases (0.5 ± 0.06) versus controls (0.16 ± 0.03) (P = 0.01) (Strittmatter et al. 1993).
As well as familial late onset Alzheimer's disease, the APOE e4 allele was found to be associated with sporadic cases of the disease, including an early report detailing 176 cases of sporadic disease where diagnosis was confirmed at post mortem and the e4 allele frequency was found to be 0.40 ± 0.03 (P = 0.00001) (Saunders et al. 1993). No association was found with early onset familial Alzheimer's disease. The disease association was robustly replicated for different populations by numerous independent researchers, showing similar results in men and women. These studies, together with very many other published disease association studies in Alzheimer's disease, are described in the 'AlzGene' online database maintained by the Alzheimer Research Forum (www.alzforum.org) (Bertram et al. 2007a). The strongest association is found when a family history of dementia is present and with two copies of the e4 allele (e4/e4). There is also some evidence of a protective effect associated with possession of the APOE e2 allele (Corder et al. 1994). Analyses of polymorphic variants in the region of APOE showed no evidence that the observed associations were due to linkage disequilibrium with other genetic variation in the locus (Strittmatter and Roses 1996).
A meta analysis published in 1997 included data for 5930 patients with Alzheimer's disease and 8607 controls, predominantly for populations of European ancestry (Fig. 2.9) (Farrer et al. 1997). Among Caucasians from clinics or with post mortem proven disease, possession of a copy of the e4 allele (e4/e3) versus the common haplotype e3/e3 carried an odds ratio of 3.2 (2.8-3.8) for Alzheimer's disease, and possession of two copies (e4/e4) versus e3/e3 an OR of 14.9 (10.8-20.6); there was also evidence that possession of the e2 allele was associated with some protection (e2/e3) versus e3/ e3 with an OR of 0.6 (0.5-0.8) (Farrer et al. 1997). A more recent study among African Americans demonstrated similar age-related disease risk with ORs versus e3/e3 of 2.6 (1.8-3.7) for e4/e3 and 10.5 (5.1-21.8) for e4/e4 (Graff-Radford et al. 2002); among Hispanics the association also appears robust with a strong association with
Was this article helpful?