A genomic pathway approach to a complex disease: axon guidance and Parkinson disease

Abstract

While major inroads have been made in identifying the genetic causes of rare Mendelian disorders, little progress has been made in the discovery of common gene variations that predispose to complex diseases. The single gene variants that have been shown to associate reproducibly with complex diseases typically have small effect sizes or attributable risks. However, the joint actions of common gene variants within pathways may play a major role in predisposing to complex diseases (the paradigm of complex genetics). The goal of this study was to determine whether polymorphism in a candidate pathway (axon guidance) predisposed to a complex disease (Parkinson disease [PD]). We mined a whole-genome association dataset and identified single nucleotide polymorphisms (SNPs) that were within axon-guidance pathway genes. We then constructed models of axon-guidance pathway SNPs that predicted three outcomes: PD susceptibility (odds ratio = 90.8, p = 4.64 x 10(-38)), survival free of PD (hazards ratio = 19.0, p = 5.43 x 10(-48)), and PD age at onset (R(2) = 0.68, p = 1.68 x 10(-51)). By contrast, models constructed from thousands of random selections of genomic SNPs predicted the three PD outcomes poorly. Mining of a second whole-genome association dataset and mining of an expression profiling dataset also supported a role for many axon-guidance pathway genes in PD. These findings could have important implications regarding the pathogenesis of PD. This genomic pathway approach may also offer insights into other complex diseases such as Alzheimer disease, diabetes mellitus, nicotine and alcohol dependence, and several cancers.

Figure 1. Goodness-of-Fit of Final Model Using Axon Guidance Genes to Predict Susceptibility to PD

Data presented are within our primary whole-genome association dataset [11]. Histogram of predicted probabilities of PD in cases (A) and histogram of predicted probabilities of PD in unaffected siblings (B) are presented. A perfect fit would have all predicted probabilities for cases equal to one and all predicted probabilities for unaffected siblings equal to zero. A model with no explanatory value would have histograms that were indistinguishable from each other. Conditional logistic regression analyses were used to generate the probabilities, and hence the data do not fit a single exponential curve (one end close to zero).

Figure 2. Kaplan-Meier Survival Plot for Age at Onset in PD Patients

Data are grouped by categorized risk score from proportional hazards model within our primary whole-genome association dataset [11]. The model clearly differentiates between early-age-at-onset cases and late-age-at-onset cases. Cases generating the long-dashed line were predicted to be at highest risk for early onset of PD, followed by the medium-dashed, short-dashed, and continuous lines, respectively. Since only cases were included, all of the lines end at 0% free of PD.

Figure 3. Predicted Versus Reported Age at Onset in PD Patients

Data are within our primary whole-genome association dataset [11]. Points indicate reported individual values, and the line represents perfect agreement. In this case, the pattern shows a fairly tight elliptical pattern, reflecting a good fit and high R ² of 0.68. The fit appears equally good from the minimum to maximum reported age at onset.

Figure 4. Convergence of Findings for Whole-Genome Association and Gene-Expression Profiling Datasets

Listed are genes that had one or more SNPs included in at least one of the three genetic models that predicted PD susceptibility or survival free of PD or age at onset of PD (n = 45 genes). Expression data are summarized for each of the three nigrostriatal regions (substantia nigra, putamen, and caudate). Gray, no informative probe sets for the gene and region; white, no difference in expression in cases and controls; red, the gene had increased expression in cases compared to controls; green, the gene had reduced expression in cases compared to controls.