Purpose:
To establish molecular biomarkers for age-related macular degeneration (AMD) susceptibility and progression. We have expanded AMD biomarker studies of three oxidative protein modifications [carboxymethyllysine (CML), pentosidine and carboxyethylpyrrole (CEP)] and four AMD risk genotypes [Y402H in CFH, R80G in C3, A69S in ARMS2, and rs11200638 in HTRA1].
Methods:
CML and furosine were quantified in AMD and control plasma by LC MS/MS, pentosidine by LC fluorimetry, and CEP/CEP autoantibody by ELISA. DNA was genotyped using TaqMan SNP genotyping assays. Logistic regression modeling for c-statistics, odds ratios and p values was performed with SAS 9.2. The risk for AMD was predicted based on genotype alone or in combination with proteomic markers. Methods and results from 90 plasma specimens have been reported for CML, pentosidine and furosine (2009 Mol Cell Proteomics 8, 1921) and from 1404 plasma for CEP/CEP autoantibody (2009 Mol Cell Proteomics 8, 1338).
Results:
CML, pentosidine and furosine measurements have been obtained from ~700 plasma samples and compared with available CEP data. CML and pentosidine levels were elevated significantly in all categories of AMD but highest in advanced atrophic AMD, up 23% and 65% respectively, relative to controls. No difference in the concentration of furosine (fructosyl-lysine), a marker of early glycation, was detected in control and non-diabetic AMD plasma, supporting a direct association of CML and pentosidine with AMD. Combining CML and pentosidine plasma concentrations with CEP marker levels yielded discriminatory accuracy (c-statistics) of ~89% for all AMD and >90% for geographic atrophy. Similar to that previously reported for CEP, the AMD risk (odds ratio) for all disease categories combined predicted for those with elevated CML or pentosidine in combination with genotype was 2-5x higher than risk based on genotype alone (p < 0.001, Fisher exact test), with joint effect odds ratios in the range of 7.6-25.5.
Conclusions:
Molecular prognostics for identifying those susceptible to progression to advanced AMD could help slow or prevent vision loss. Genomic markers alone are insufficient for prognosis, as many carrying AMD risk genotypes never develop AMD. In combination, CML, pentosidine, CEP and genomic markers significantly improve AMD risk predictions.
Keywords: age-related macular degeneration • proteomics • genetics