SNP genotyping technology holds great potential for discovering multigenic contributions to complex neurological disorders. These disorders are usually not inherited in a Mendelian fashion and, in many cases, may be referred to as “sporadic” cases of disease, which are ideally studied using case-control whole genome association studies of outbred populations.
The development of SNP scanning technology now allows the simultaneous testing of more than a million genetic variants, enabling initial studies into the genetics of complex disorders. Specific issues related to the technology of SNP association studies, including sample size, SNP density, and data analysis methods, are reviewed more extensively elsewhere (Craig and Stephan, 2005
; Dunckley et al. 2006
). Here we will discuss how the technology has been applied to neurological disorders and significant findings that have resulted.
In the majority of these diseases, it is striking that there are no single allele variants that account for the bulk of disease risk. The notable exception is that of Alzheimer’s disease, where whole genome SNP association studies confirm that the apolipoprotein E (ApoE) locus confers the largest genetic susceptibility to AD. However, when individuals are grouped and analyzed according to their ApoE genotypes, multiple SNPs within the GRB2 associated protein 2 (GAB2) gene have been reported to be significantly associated with AD risk in carriers of the
4 allele of APOE (Reiman et al. 2007
). These findings await further confirmation. However, this study provides a glimpse into what SNP based genotyping technology offers for the future in terms of uncovering interactions between multiple genetic loci that lead to an aggregate genetic risk for disease.
Association studies in other diseases usually identify few loci with high odds ratios (ORs) for disease. In this regard, Alzheimer’s disease is now clearly an exception among complex human diseases. The typical result appears to be identification of multiple loci with ORs less than 2. For example, a recent study in amyotrophic lateral sclerosis identified a dozen loci reproducibly associated with disease (Dunckley et al. 2007
). However, no locus exceeded a 2-fold OR. Importantly, the multiple loci that were identified highlight distinct functional categories of genes. The most prominent among these groups are genes involved in neurite outgrowth. For example, a SNP within the anaplastic lymphoma kinase gene (ALK) was associated with ALS with an OR of 1.57. Independent studies showed that ALK is critical for pleiotrophin-mediated axonal regeneration in spinal motor neurons and is necessary for neuroprotection in response to glutamate excitotoxicity (Mi et al. 2007
). These observations, coupled with genetic association to sporadic ALS, suggest that variations in the ALK gene could have subtle functional consequences that alter the axonal dynamics of motor neurons, ultimately leading to the development of ALS.
This study in ALS also identified candidate SNPs associated with clinical subclasses of ALS that differ in site of symptom onset, age at onset, and gender-based differences in disease onset. With further validation and confirmation, these SNPs could provide the basis for presymptomatic risk assessment to determine individuals at risk for not only developing ALS, but also for developing specific forms of ALS. This information could then be used to direct differentially treatment options. Importantly, this type of analysis highlights the direction that SNP based diagnostics and risk assessment strategies are headed to impact disease treatment approaches before initial symptoms occur. This may be a particularly important approach in neurodegenerative diseases where substantial neuronal cell loss and pathology has already occurred by the time a clinical diagnosis is made. Treatments are likely to be more effective prior to this advanced cell loss.
The use of SNP association studies for identifying multigenic contributions to complex disease risk is clearly in the early stages. The relatively low risk contributed by individual alleles will necessitate the study and analysis of significantly larger patient populations than have currently been studied. Identification of consistently reproducible genetic variants associated with complex human disease is likely to require tens of thousands of samples. Indeed, large initiatives such as the Well-come Trust Case Control Consortium and the GAIN collaborative research initiative are making the first important steps toward the study of these larger populations. However, these efforts must be expanded upon.
The utility of identifying genetic variants for complex diseases lies in their use as a predictor of disease risk, as a diagnostic for disease, as a predictor of response to therapy, or for the identification of therapeutic targets. For example, currently genetic testing is done one gene at a time using a candidate gene approach. That is, one has a family history of a particular disease for which a common genetic variant is known, such as cystic fibrosis, and can be tested for the presence of that variant within their genome. This information can then be used to guide life decisions, such as reproductive choices or exercise and eating habits in instances of other disorders. However, most human disease is sporadic and multigenic. Risk for these diseases cannot be diagnosed using traditional approaches. SNP analysis can be performed on a genome-wide scale in large case-control association studies of outbred populations to identify all of the genetic variants that contribute simultaneously to a specific disease. These variants can then be packaged into a prognostic test to predict an individual’s overall genetic risk for developing a given disease. These prognostic tests would be most effective when coupled to an effective therapeutic or prevention strategy. However, such a test to assess disease risk could have a significant impact on human health, even in the absence of a specific therapy because environmental influences, which are modifiable, also affect the development and course of disease. For example, high cholesterol diets or smoking have been shown to increase the risk of some forms of cancer and may also be an important factor in the development of Alzheimer’s disease. Knowing that you are at heightened genetic risk for these disorders could motivate significant lifestyle changes. In this way, as the analysis of complex diseases continues to evolve, the ease of use and the versatility of SNP genotyping for identifying relevant genetic variants, such as point mutations, deletions, and insertions, will lead to significant advances in human health.
provides a flow diagram illustrating the interconnected components of the various genomic technologies for the identification of neurological disease biomarkers. These biomarkers can be used as diagnostics or, potentially, as therapeutic targets around which treatments could be directed. Importantly, these are complementary technologies, each providing unique information about the genetic and molecular factors contributing to the disease. However, these technologies though overlapping in their utility also have unique uses. For example, SNP-based screening, and to a lesser extent aCGH, are the likely the preferred options when trying to identify biomarkers that can be used in presymptomatic risk assessment. In contrast, gene expression profiling will only show differences when some component of the disease has already manifest itself and is functionally altering the physiology of the cells or tissues. As such, gene expression profiling can provide unique diagnostic signatures or can be used to identify underlying molecular changes associated with disease. These molecular factors may serve as targets for therapeutic development. A comprehensive genomic biomarker discovery strategy will utilize each of these technologies.
Figure 1 Work flow for genomic biomarker discovery in neurological diseases. Outlined are approaches using both DNA-based and RNA-based technologies. Ultimately these are discovery approaches that then must be translated into clinical application through subsequent (more ...)