Assessing susceptibility to age-related macular degeneration with proteomic and genomic biomarkers

Abstract

Age-related macular degeneration (AMD) is a progressive disease and major cause of severe visual loss. Toward the discovery of tools for early identification of AMD susceptibility, we evaluated the combined predictive capability of proteomic and genomic AMD biomarkers. We quantified plasma carboxyethylpyrrole (CEP) oxidative protein modifications and CEP autoantibodies by ELISA in 916 AMD and 488 control donors. CEP adducts are uniquely generated from oxidation of docosahexaenoate-containing lipids that are abundant in the retina. Mean CEP adduct and autoantibody levels were found to be elevated in AMD plasma by approximately 60 and approximately 30%, respectively. The odds ratio for both CEP markers elevated was 3-fold greater or more in AMD than in control patients. Genotyping was performed for AMD risk polymorphisms associated with age-related maculopathy susceptibility 2 (ARMS2), high temperature requirement factor A1 (HTRA1), complement factor H, and complement C3, and the risk of AMD was predicted based on genotype alone or in combination with the CEP markers. The AMD risk predicted for those exhibiting elevated CEP markers and risk genotypes was 2-3-fold greater than the risk based on genotype alone. AMD donors carrying the ARMS2 and HTRA1 risk alleles were the most likely to exhibit elevated CEP markers. The results compellingly demonstrate higher mean CEP marker levels in AMD plasma over a broad age range. Receiver operating characteristic curves suggest that CEP markers alone can discriminate between AMD and control plasma donors with approximately 76% accuracy and in combination with genomic markers provide up to approximately 80% discrimination accuracy. Plasma CEP marker levels were altered slightly by several demographic and health factors that warrant further study. We conclude that CEP plasma biomarkers, particularly in combination with genomic markers, offer a potential early warning system for assessing susceptibility to this blinding, multifactorial disease.

Fig. 1.

CEP adducts and autoantibodies are elevated in AMD plasma. CEP adduct concentrations (A) and autoantibody titers (B) quantified by ELISA from control (n = 488) and AMD (n = 916) plasma donors are shown with median (▵) results ± first and third quartiles (Q1, Q3) and mean (○) results ± S.D. indicated. p values (two-sided t test) were determined from log-transformed concentrations. These data are presented in Table I by category of AMD progression. Correlation between CEP adduct levels and autoantibody titers is shown for the control (C) and AMD (D) cohorts with horizontal and vertical dashed lines indicating median control values. Significantly more donors with both CEP markers elevated are apparent in AMD patients than in the controls (upper right quadrants in C and D).

Fig. 2.

Western analysis of AMD and control plasma for CEP adducts. Human AMD and control donor plasma samples were subjected to SDS-PAGE (∼15 μg/lane), electroblotted to PVDF, and probed with mouse monoclonal anti-CEP antibody. CEP adducts (immunoreactivity) are shown to be associated with high molecular mass components (>∼40 kDa) after 1-min autoradiography (A) and also with components of ∼25 kDa after 60-min autoradiography (B). Additional components become apparent with longer autoradiography time. Densitometric quantification of CEP immunoreactivity in the indicated bands supports more CEP-adducted proteins in AMD than normal donor plasma (p values are from the two-sided t test). Error bars reflect standard deviation. The Coomassie Blue-stained gel (C) shows that approximately equal amounts of protein were applied per lane for the Western analysis. The age and sex of each donor are listed, and for AMD samples, the asterisk (*) and # symbols denote donors with CNV or geographic atrophy, respectively. F, female; M, male.

Fig. 3.

AMD risk predicted by CEP markers and genotype. A, odds ratios for AMD risk based on elevated CEP markers only, genotype only (specific for the homozygous risk alleles ARMS2, HTRA1, CFH and C3), and the joint effect of both are shown for all AMD (a) and advanced AMD (b) patients. B, odds ratios for both CEP markers to be elevated in AMD risk and non-risk homozygous genotypes are shown for all AMD or advanced AMD patients. Differences in CEP marker concentrations between homozygous risk and non-risk donors were statistically significant (**, p < 0.01; *, p < 0.05; Fisher exact test) for ARMS2 and HTRA1 but not for CFH and C3. Sample size, gene frequencies, OR with 95% CI, and p values are presented in Tables III, IV, and VI. OR, 95% CI, and p values were determined with log-transformed CEP marker concentrations.

Fig. 4.

Plasma CEP adducts and autoantibodies by donor age. Plasma CEP adduct (A) and CEP autoantibody levels (C) in the AMD (▵) and control (•) cohorts are shown plotted by donor age. Pearson's correlation analysis (horizontal color-coded lines and p value from log-transformed data) revealed little change in mean CEP marker concentrations with age except for a gradual increase in CEP autoantibody titer in the control cohort. CEP adduct (B) and CEP autoantibody levels (D) in AMD and control donors are plotted by age group, including controls ≤50 years (y) (n = 98), 51–60 years (n = 138 control, n = 26 AMD), 61–70 years (n = 153 control, n = 123 AMD), 71–80 years (n = 154 control, n = 389 AMD), and >80 years (n = 43 control, n = 378 AMD). -Fold difference in CEP marker concentrations is indicated between the control and AMD groups. Asterisks reflect p values from a two-sided t test (***, p < 0.001; **, p < 0.01; and *, p < 0.05). Error bars reflect standard deviation.

Fig. 5.

Plasma CEP markers stratified by demographic and health factors. Plasma CEP adduct and CEP autoantibody levels in the AMD and control study populations are plotted based on donor status with regard to race, gender, smoking status, hypertension, hyperlipidemia, diabetes, and cardiovascular diseases. Sample size per group is indicated, and asterisks reflect p values from a two-sided t test of log-transformed CEP marker concentrations (***, p < 0.001; **, p < 0.01; and *, p < 0.05). Cauc, Caucasian; Afr Am, African-American; F, female; M, male; S, smoking; NS, non-smoking; w, with; w/o, without. Error bars reflect standard deviation.