Purchase this article with an account.
Eiko de Jong, Yara Lechanteur, Nicolas Schauer, Tina Schick, Sascha Fauser, Carel C B Hoyng, Anneke I Den Hollander; Metabolomics in AMD - identifying metabolic pathways and biomarkers for improved prediction and prevention. Invest. Ophthalmol. Vis. Sci. 2015;56(7 ):395.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
A metabolomics study was performed on serum samples to discover novel biomarkers for (dry) AMD and to uncover clinically relevant metabolic pathways.
132 subjects were categorized into four groups based on their genetic risk profile (CFH and ARMS2 alleles) and their disease status: 1) healthy controls with low genetic risk, 2) AMD patients with low genetic risk, 3) AMD patients with high genetic risk, 4) healthy controls with high genetic risk. Groups were matched for age, sex, smoking, BMI and levels of activated complement. For all subjects, serum was collected and analyzed using a metabolomics approach, employing the parallel application of gas chromatography- mass spectrometry and liquid chromatography- mass spectrometry.<br />
On average, approximately 1200 metabolites per subject were detected. In each of the groups, on average 5 unique metabolites were detected that were not present in any other group. By comparing groups 1 and 2, we identified 17 metabolites significantly associated with AMD, independent of CFH/ARMS2 genotypes. By comparing high risk groups 3 and 4, we identified 32 differentially expressed metabolites. Within the groups, several associated metabolites showed a high degree of correlation with each other. Random forest analysis could predict the groups based on their composition with 58-79% accuracy. Not all metabolites could be annotated directly.
By comparing groups 1 and 2, potentially novel metabolic networks underlying AMD have been identified. By comparing groups 3 and 4, metabolites have been identified that potentially could protect against AMD in the context of high genetic risk. Because there is a high degree of correlation between metabolites in each group, these results are likely to represent meaningful metabolic networks rather than false positives. Further characterization of metabolites that are not yet annotated is required to determine the exact nature of these networks.
This PDF is available to Subscribers Only