Integrative Network Analysis Unveils Convergent Molecular Pathways in Parkinson's Disease and Diabetes

<div><p>Background</p><p>Shared dysregulated pathways may contribute to Parkinson's disease and type 2 diabetes, chronic diseases that afflict millions of people worldwide. Despite the evidence provided by epidemiological and gene profiling studies, the molecular and functional networks implicated in both diseases, have not been fully explored. In this study, we used an integrated network approach to investigate the extent to which Parkinson's disease and type 2 diabetes are linked at the molecular level.</p><p>Methods and Findings</p><p>Using a random walk algorithm within the human functional linkage network we identified a molecular cluster of 478 neighboring genes closely associated with confirmed Parkinson's disease and type 2 diabetes genes. Biological and functional analysis identified the protein serine-threonine kinase activity, MAPK cascade, activation of the immune response, and insulin receptor and lipid signaling as convergent pathways. Integration of results from microarrays studies identified a blood signature comprising seven genes whose expression is dysregulated in Parkinson's disease and type 2 diabetes. Among this group of genes, is the amyloid precursor protein (<i>APP</i>), previously associated with neurodegeneration and insulin regulation. Quantification of RNA from whole blood of 192 samples from two independent clinical trials, the Harvard Biomarker Study (HBS) and the Prognostic Biomarker Study (PROBE), revealed that expression of <i>APP</i> is significantly upregulated in Parkinson's disease patients compared to healthy controls. Assessment of biomarker performance revealed that expression of <i>APP</i> could distinguish Parkinson's disease from healthy individuals with a diagnostic accuracy of 80% in both cohorts of patients.</p><p>Conclusions</p><p>These results provide the first evidence that Parkinson's disease and diabetes are strongly linked at the molecular level and that shared molecular networks provide an additional source for identifying highly sensitive biomarkers. Further, these results suggest for the first time that increased expression of <i>APP</i> in blood may modulate the neurodegenerative phenotype in type 2 diabetes patients.</p>