The genetics of late-onset Alzheimer's disease (LOAD) has taken impressive steps forwards in the last few years. To date, more than six-hundred genes have been linked to the disorder. However, only a minority of them are supported by a sufficient level of evidence. This review focused on such genes and analyzed shared biological pathways. Genetic markers were selected from a web-based collection (Alzgene). For each SNP in the database, it was possible to perform a meta-analysis. The quality of studies was assessed using criteria such as size of research samples, heterogeneity across studies, and protection from publication bias. This produced a list of 15 top-rated genes:
Alzheimer’s disease (AD) is the leading cause of dementia in developed countries. It afflicts 5.3 million individuals in the US. Total direct and indirect cost is US$ 172 billion per year [
Clinically, AD is characterized by progressive impairments in memory and other cognitive domains. Behavioral and psychiatric symptoms (BPSDs), clustered into agitation/aggression, mood disorders, and psychosis, may occur with disease progression [
We used the AlzGene database to identify those genes that had the strongest association with LOAD, for which there was a qualitatively high level of evidence. AlzGene is a web-based synopsis of published association studies on AD [
Genes with strong (A) and moderate (B) associations were included in a query set of the gene ontology database AmiGO to discover shared biological functions. The Gene Ontology [GO (
The Alzgene database (updated 13 September 2010) includes 1,380 studies and 666 genes. The number of meta-analyses is 380.
Forty-two genes have at least one positive meta-analysis (see Table
Top-rated genes associated with LOAD.
Gene | Ch | N° minor | Quality | Caucasian | Asian | All ethnic groups |
---|---|---|---|---|---|---|
APOE | 19 | 4,167 | AAA | 3.77 (3.29–4.32) | 3.99 (2.86–5.57) | 3.61 (3.20–4.08) |
CLU | 8 | 53,712 | AAA | 0.87 (0.85–0.90) | n.a | 0.88 (0.86–0.91) |
PICALM | 11 | 44,358 | AAA | 0.89 (0.86–0.92) | n.a | 0.90 (0.86–0.93) |
EXOC3L2 | 19 | 13,519 | AAA | 1.17 (1.12–1.23) | n.a | 1.17 (1.12–1.23) |
BIN1 | 2 | 24,713 | AAA | 1.14 (1.08–1.21) | n.a | 1.14 (1.08–1.21) |
CR1 | 1 | 18,779 | AAA | 1.14 (1.08–1.20) | n.a | 1.16 (1.09–1.22) |
SORL1 | 11 | 1,734 | AAA | 1.07 (1.00–1.15) | 1.30 (1.13–1.50) | 1.10 (1.02–1.17) |
TNK1 | 17 | 3,538 | AAA | 0.84 (0.76–0.93) | n.a | 0.84 (0.76–0.93) |
IL8 | 4 | 1,157 | AAA | 1.26 (1.01–1.58) | n.a | 1.26 (1.01–1.58) |
LDLR | 19 | 1,228 | AAA | 0.85 (0.72–0.89) | n.a | 0.85 (0.72–0.89) |
CST3 | 20 | 1,203 | AAA | 1.28 (1.04–1.56) | n.a | 1.23 (1.03–1.48) |
CHRNB2 | 1 | 227 | BAA | 0.69 (0.51–0.95) | n.a | 0.67 (0.50–0.90) |
SORCS1 | 10 | 567 | BAA | 1.34 (1.09–1.65) | n.a | 1.34 (1.09–1.65) |
TNF | 6 | 301 | BAA | n.a | 1.37 (1.05–1.79) | 1.35 (1.39–1.77) |
CCR2 | 3 | 308 | BAA | 0.73 (0.56–0.97) | n.a | 0.73 (0.56–0.97) |
OR values are referred to the best SNP for each gene.
n.a: one study or none; meta-analysis could not be performed.
HuGENet classification was used to assess the quality of studies (see text).
The gene encoding apolipoprotein E (chromosome 19q 13.2) was associated with AD in thirty-eight case-control samples (Caucasian = 28; Asian = 4; African descent = 2; Hispanic descent = 1; mixed ethnic groups = 3) and four family-based studies. Overall OR was 3.77 (95% CI 3.29–4.32; I2 = 13) in Caucasian samples and 3.99 (95% CI: 2.86–5.57 I2 = 20) in Asian samples.
Clusterin (apolipoprotein J) is a chaperone molecule that appears to be involved in membrane recycling and apoptosis. Clusterin, like apolipoprotein E, is found in amyloid plaques [
Phosphatidylinositol-binding clathrin assembly proteins is a key component of clathrin-mediated endocytosis. It recruits clathrin and adaptor protein 2 (AP-2) to the plasma membrane and, along with AP-2, recognizes target protein. The attached clathrin triskelions cause membrane deformation around the target proteins enclosing them within clathrin-coated vesicles [
Exocyst complex component 3-like 2 is also involved in vesicle targeting during exocytosis of proteins and lipids that is essential to neuron outgrowth and integrity [
Bridging integrator 1 is a member of the BAR adapter family which has been implicated in endocytosis and intracellular endosome trafficking [
Complement component receptor 1 regulates complement cascade via the inhibition of both classical and alternative pathway C3 and C5 convertases [
Sortilin-related receptor (SorLA) is a sorting receptor that regulates trafficking and processing of APP. SorLA acts as a retention factor for APP in trans-Golgi compartments/trans-Golgi network, preventing the release of the precursor into regular processing pathways [
Nonreceptor tyrosine kinase 1 is involved in intracellular transduction pathways, and it was shown to enable TNF-alpha-induced apoptosis [
Interleukin 8 is a proinflammatory cytokines. Cerebrospinal fluid levels of IL-8 were found to be increased in AD and mild cognitive impairment [
Low density lipoprotein receptor is implicated in cholesterol metabolism via endocytosis. Recently, it has been discovered that overexpression of brain LDLR is associated with decrease in APOE levels and beta amyloid due to either inhibited deposition or enhanced clearance [
Cystatin C, a potent inhibitor of lysosomal proteinases, was shown to bind beta amyloid and to prevent beta-amyloid aggregation and deposition in mouse models [
Each nAChR protein is made up of a combination of five subunits, usually two alpha (
SorCS proteins (like SorLA) are members of the Vps10p family of sorting receptors. SorCS1 binds to nerve growth factor (NGF) propeptide. Pro-NGF is increased in AD brains, and its binding to neurotrophin receptor p75 induces apoptotic cell death in neurons [
Tumor necrosis factor alpha induces the production of beta amyloid [
Chemokine receptor 2 is IL-8 receptor. It is coupled with MAP-kinase pathway to modulate signaling transduction.
Gene ontology analysis identified 146 GO terms more represented in test dataset (LOAD genes) than in UniProtKb collection (
The genetics of late-onset AD is a complex one. More than six-hundred genes have been investigated as susceptibility factors. They represent 2.9% of all genes with known function (
Five genes of our compilation (
A second-pathway was endocytosis. This is supported by five genes (
Seven genes (
These pathways are actually interconnected. One such network is lipoprotein-inflammation apoptosis. Central links in this chain are
Several characteristics of AD patients, not merely diagnostic identification, are affected by AD genes [
Gene-gene interactions may account for a substantial genetic variability in LOAD. Gene-gene interactions were reported between
Response to antidementia drugs is also affected by genetic factors. Pharmacogenomics in AD is still in its infancy, with genes associated with AD pathogenesis and genes responsible for drug metabolism (cytochrome P450) [
This review was based on the most comprehensive collection of published studies about the genetics of AD. The best genes were classified according to qualitative criteria such as size of research samples, heterogeneity across studies, and control for various sources of bias including small effect size (OR) and publication bias. However, there were also important limitations, mainly due to Alzgene design. First, meta-analyses were restricted to allele contrast, which is less powerful than genotype-based test and allows no inference of the true underlying mode of inheritance, and there was no genetic information at haplotype level. Moreover, only the main effect was investigated, that is diagnostic association with AD, while other clinical phenotypes and endophenotypes could not be considered alongside gene-gene and gene-environment interactions. On the contrary, a nonnegligible effect of LOAD genes may be directed to these secondary targets as suggested elsewhere. Gene ontologies were developed to provide a shared representation of genes and gene products across species. GO terms contain broad definitions of biological processes in the living cell. Hence, these terms are suitable to identify areas for genomic exploration (e.g., all genes implicated in cholesterol metabolism) but not to elucidate pathogenic mechanisms in depth.
Notwithstanding these caveats including all published studies in a single open-access database (Alzgene) highlights the most important pathophysiological mechanisms, which show the convergence of many genes, and it more easily prompts new biological hypotheses.