Alzheimer’s disease (AD) is a progressive, neurodegenerative disorder known to occur in the elderly, with an onset age of around 60 years. AD represents the most common form of elderly dementia, characterized by memory loss (92% of all patients), confusion (71%), short attention span (63%), a declining sense of direction (53%), and personality changes (31%) [1]. Autopsy studies on brains of affected patients reveal the neuropathological hallmarks of AD, which include senile plaques of amyloid β (Aβ) protein and intraneuronal tangles of hyperphosphorylated tau protein, as well as gross cortical atrophy and ventricular dilation [2]. Approximately 1 in 8 people over the age of 65 years suffer from this debilitating disease, 65% of which are women [1]. Alzheimer’s disease poses a tremendous public health challenge to the world, as the average human life expectancy continues to improve. Todate, a total of ~ 46.8 million cases of dementia are recorded worldwide, with an associated health-care cost in the billions. By the year 2030, the number of cases will almost double [3].

Extensive research on the genetics of AD using traditional and modern genetic tools has, to a large extent, helped elucidate the etiology of AD, although many questions remain. This article presents a review of the genetics of AD attributed to the hallmark neuropathology of AD, including accumulation of Aβ and tau protein.

There are two main strategies used to identify the genetic risk factors of AD. The first strategy employs a phenotype to genotype approach. Researchers examine polymorphic genomic markers such as short tandem repeats found commonly in families with a high burden of AD (two or more cases of AD among first–degree relatives) across multiple generations to identify broad genomic regions co-transmitted with the disease. This method, used to determine linkage or tendency of genes to be inherited together because of their spatial proximity, is known as linkage analysis. The technique is typically followed by positional cloning of candidate genes to test their contribution to AD. Linkage analyses were instrumental in the identification of genes associated with familial or early-onset AD, which is described in more detail below [4].

The second strategy for identifying the genetic risks of AD employs a genotype to phenotype approach. Researchers compare allele frequencies of available polymorphic markers such as single nucleotide polymorphisms (SNPs) between healthy (control) and sick (case) people to identify genes associated with AD. These studies are known as genome-wide association studies and represent modern genetic tools for studying complex or sporadic forms of AD [5] or other parameters such as the concentration of soluble TREM2 in cerebrospinal fluid [6].

Table 1 lists each method in detail, including its advantages and disadvantages and its contribution to the genetics of AD.

[enlarge]

Figure 1. Categories of Alzheimer’s Disease.

AD is commonly categorized based on the age of onset of symptoms. Early-onset AD (EOAD) typically begins before 65 years of age, while late-onset AD (LOAD) most commonly starts after 60 (see Figure 1). EOAD, also known as familial AD, is caused by inheritance of autosomal dominant mutations in three genes: APP, PSEN1, and PNSEN2, and accounts only for ~ 5% of all AD cases. A rare mutation of ADAM17 was also found to cause late-onset familial Alzheimer disease with an autosomal-dominant pattern [7]. The majority of AD cases fall under the LOAD category, also known as sporadic AD, which represents a much more complex disorder, and is associated with a multitude of genetic risk factors. While EOAD is 100% inheritable, LOAD is only 60-80% inheritable; the remaining cases of LOAD can be attributed to environmental influences (i.e., diet, brain injury, lifestyle, certain medications) [1].

Linkage analyses of the affected families led to the discovery of APP, PSEN1, and PSEN2 (Table 2), with approximately 250 mutations identified in these genes that are responsible for disrupting their normal biological function, leading to deposition of Aβ plaques in the brain. APOE E4 may explain around 10.1% of the variance of EOAD and rare coding variants in APOB may also contribute to EOAD [8].

The APP gene is on chromosome 21. Researchers have identified over 50 AD-associated APP mutations with new missense/deletion APP mutations being reported [14, 15]. The gene encodes the APP protein that undergoes proteolytic cleavage by three distinct proteases called α-, β- and γ-secretase. Proteolysis of APP by β and γ-secretase produces two amyloid-β species called sAPPβ and β-CTF. Aβ, the protein responsible for the senile amyloid plaques seen in AD-affected brains, is produced from further processing of βCTF. While most AD-causing APP mutations are dominant, some recessive mutations have also been described [1]. Down syndrome patients older than 65 years, with the trisomy of chromosome 21 (hence the overexpression of APP gene), has a prevalence of nearly 80% for Alzheimer disease [16] ; for those older than 55 years, in another study, the prevalence is 32.7% [17]. A short, 17-amino acid peptide in sAPP binds to the sushi 1 domain of GABA_BR subunit 1a and is a physiological ligand for metabotropic GABA receptors [18]. APP gene, more specifically, the beta amyloid peptide derived from it, is the most common therapeutic target for Alzheimer's disease [19].

The presenilin-1 (PSEN1) gene is located on Chromosome 14. The encoded protein is a component of the γ-secretase complex that is needed in APP processing. Over 200 AD-associated PSEN1 mutations have been identified so far. PSEN1 are responsible for 50% of all EOAD, with new ones being identified. For example, Jia L et al examined 404 pedigrees and identified 10 new missense PSEN1 mutations [27]. The PSEN2 gene is located on Chromosome 1. New mutations are identified constantly. Van Giau V et al locate a new mutation at Trp 165 of PSEN1 associated with EOAD [28]. Itzcovich T et al identified a novel T119I dominant mutation in PSEN1 in an Argentine family with early and late-onset Alzheimer’s disease [29]. Kim YE et al reported both pathogenic variants of PSEN1 gene and variants of unknown significance in early onset Korean Alzheimer's patients [30]. PSEN2 mutations are much rarer; a total of only 13 pathogenic mutations have been identified in 29 families. Similar to PSEN1, PSEN2 is also a component of the γ-secretase complex and is needed for APP processing [1].

LOAD represents a complex form of Alzheimer’s disease. However, one gene, Apolipoprotein E (APOE) has the strongest causative effect on LOAD. APOE was first identified in 1991 using linkage analysis on chromosome 19 [31]. The encoded protein is produced by astrocytes in the nervous system and is responsible for transporting cholesterol to neurons, as well as neuronal growth, repair response to tissue injury, nerve regeneration, immunoregulation, and activation of lipolytic enzymes [32]. APOE is a polymorphic gene, which is expressed in three different isoforms (Table 3). These include E2, E3, and E4. Isoform E4 represents a genetic risk factor for LOAD that can increase a person’s risk for LOAD by 3 to 8 times. Isoform E4 is associated with approximately 36.7% of LOAD cases [33] and likely accelerates amyloid deposition, possibly beginning as early as the 30-40 years of age [34]. ApoE E4/4 homozygotes also tend to have olfactory dysfunction [35]. Interestingly, isoform E2 may decrease a person’s risk of developing LOAD [33], while E3 appears to have a neutral biological effect. A mutation at ApoE3 (R136S) potentially delays or prevents the onset of Alzheimer’s disease in an autosomal dominant PSEN1 E280A mutation carrier [36]. A gain-of-function mutation of RELN, which, like ApoE, binds to VLDLr and APOEr2 receptors, delays the onset of dementia [37].

Until 2009, APOE was the only genetic risk factor described for LOAD. However, only about 30% of all LOAD cases can be attributed to APOE [32]. Genetic risk factors that can account for the remaining cases of LOAD include polymorphisms or mutations of multiple genes involved in several biological pathways. Lately, emerging techniques such as Genome-wide Association Studies (GWAS), whole exome/genome sequencing, and candidate gene analysis studies are spearheading the identification of the genetic risk factors contributing to onset and/or progression of LOAD (Table 4).

Among the genetic risk factors described in the table above, some have been identified repeatedly in multiple genetic studies. Thus, a more prominent association with LOAD is presumed. These genes include CLU, CR1, BIN1, PICALM, CD33, EPHA1, TREM2, ABCA7, SORL1, and ADAM10 [5]. The contribution of common and rare variants of SORL1 in Alzheimer's disease is summarized in a recent review [71]. The encoded proteins of these genes are involved in a multitude of biological pathways such as immune response, protein trafficking, endocytosis and sorting, lipid metabolism, APP processing, tau pathology, and gene regulation. Mutations in the TREM2 gene, in particular, are associated with 3 to 5 times higher risk of developing AD [54]. Variants in TREM2 and also in another gene, NOTCH3, may result in not only Alzheimer's disease but also other dementias (Nasu-Hakola disease and cerebral autosomal-dominant arteriopathy with subcortical infarcts and leukoencephalopathy) [72]. Increased soluble TREM2 in CSF correlates with reduced cognitive and clinical decline in Alzheimer’s disease [73], which has been explored to identify variants in the MS4A gene region as modulators of Alzheimer disease risk [6]. TREM2 and APOE forms an axis that might convert microglia from a homeostatic phenotype to a neurodegenerative one after phagocytosis of apoptotic neurons [74, 75]. TREM2, present in tumor-infiltrating macrophages, is also explored in immuno-oncotherapy [76]. INPP5D transmits the signalling of CD22, a regulator of microglial phagocytosis [77].

Table 5 includes details on each genetic risk factor. It is important to note that some genetic risk factors may weigh heavier in some populations than in others. For example, a rare variant in TM2D3 is disproportionately-enriched in Icelanders (~0.5% compared to <0.05% in other European populations) and is associated with a 7.5 times increased risk of LOAD, as shown in an exome-wide association analysis-based study [78], while a rare variant of ADAM17 causing late-onset familial AD is identified only in Europeans.

The major biological pathways associated with AD pathogenesis include protein sorting, an immune/inflammatory response such as inflammasome and ASC specks [80], cholesterol and lipid metabolism, and APP processing. Table 6 lists the key biological pathways associated with LOAD that have been discovered with GWAS.

The Tau protein, encoded by the MAPT gene, is associated with microtubules in the neurons and functions to stabilize them. However, hyperphosphorylated variants of tau proteins, especially those phosphorylated at Ser²⁶² [81], fail to interact with microtubules and clump together to form intraneuronal tangles. The result is a collapsed neuronal cytoskeletal system, rendering the neurons unable to function properly. Increased tau proteins in cerebrospinal fluid (CSF) correlate with risk of AD onset. GWAS reveal the role of APOE and TREM2 in the variability of tau levels in the CSF. Other genes implicated in tau pathology include BIN1, PICALM, and FERMT2 [32].

Common and rare TBK1 variants are not significantly represented in early-onset Alzheimer disease in a European cohort of 1253 patients; they are known to cause frontotemporal dementia and amyotrophic lateral sclerosis [82].

Alzheimer’s disease is a major public health concern as people’s life expectancy improves worldwide. While mutations on key genes can explain familial AD or early-onset AD, i.e., APP, PSEN1, and PSEN2, identified using linkage analyses, much of sporadic AD or late-onset AD stems from a combination of several etiological risk factors. These include both environmental (i.e., diet, brain injury, lifestyle, certain medications) and genetic factors. genome-wide association studies have been instrumental in identifying several genetic risk factors that can predispose people to the disease. These risk factors include polymorphisms or mutations in several genes that play important roles in a multitude of biological pathways protein sorting, immune/inflammatory response, cholesterol, and lipid metabolism, and APP processing. These genes are CLU, CR1, BIN1, PICALM, CD33, EPHA1, TREM2, ABCA7, SORL1, and ADAM10. Scientists continue to make progress in gaining a deeper understanding on the neuropathology of the disease by relying on traditional research models such as transgenic and non-transgenic animals and modern research models including induced pluripotent stem cells, cultured brain tissue, and molecular simulation models. Research on the genes involved in Alzheimer's disease is essential to develop effective preventive and/or therapeutical intervention approaches. Table 7 lists the antibodies against the genes discussed in this article, based on Labome's Validated Antibody Database.