Convergence of linkage, gene expression and association data demonstrates the influence of the RAR-related orphan receptor alpha (RORA) gene on neovascular AMD: a systems biology based approach

Abstract

To identify novel genes and pathways associated with AMD, we performed microarray gene expression and linkage analysis which implicated the candidate gene, retinoic acid receptor-related orphan receptor alpha (RORA, 15q). Subsequent genotyping of 159 RORA single nucleotide polymorphisms (SNPs) in a family-based cohort, followed by replication in an unrelated case-control cohort, demonstrated that SNPs and haplotypes located in intron 1 were significantly associated with neovascular AMD risk in both cohorts. This is the first report demonstrating a possible role for RORA, a receptor for cholesterol, in the pathophysiology of AMD. Moreover, we found a significant interaction between RORA and the ARMS2/HTRA1 locus suggesting a novel pathway underlying AMD pathophysiology.

Fig. 1

IPA-generated network containing microarray genes that were identified as significant and that showed a greater than two fold difference when comparing neovascular AMD patients and their unaffected siblings. The molecules which are colored are those that were identified in our study. Green indicates down-regulation while red indicates up-regulation when comparing affected to unaffected siblings. Solid lines indicate direct relationships and dotted lines indicate indirect relationships as identified in previously published literature (http://www.ingenuity.com/index.html). The individual shapes represent the family of molecule, for example, the shape of RORA indicates a ligand-dependent nuclear receptor. A full listing of these proteins, their official name, and the family of molecules to which they belong can be found in Supplementary Table 3.

Fig 2

Quantitative real time-PCR analysis reveals reduction of RORA expression in AMD patients of the extremely discordant sibling pairs. The graph represents data from 5 families depicting reduction of RORA expression in AMD patients relative to their respective unaffected sibling pair; p < 0.04. Each bar is labeled with the family identification number and error bars indicating standard error of the mean.

Fig. 3

Linkage disequilibrium plots for the extremely discordant sibling cohort (n = 150 sibling pairs). Haplotype blocks were constructed in Haploview using the method proposed by Gabriel et al. (Gabriel et al. 2002) Boxes were shaded increasingly darker to represent higher percentage of LD and the numbers listed in each square represent the D′ (A) and r² values (B) unless the box is completely shaded in representing complete LD. Two haplotype blocks were generated for this cohort.

Fig. 4

Linkage disequilibrium plots for the AREDS 2 discordant sibling cohort (n = 46 sibling pairs). Haplotype blocks were constructed in Haploview using the method proposed by Gabriel et al. (Gabriel et al. 2002) Boxes were shaded increasingly darker to represent higher percentage of LD and the numbers listed in each square represent the D′ (A) and r² values (B) unless the box is completely shaded in representing complete LD. A single haplotype block was generated for this cohort.

Fig. 5

Linkage disequilibrium plots for unrelated case-control cohort from Central Greece. Haplotype blocks were constructed in Haploview using the method proposed by Gabriel et al. (Gabriel et al. 2002) Boxes were shaded increasingly darker to represent higher percentage of LD and the numbers listed in each square represent the D′ (A) and r² values (B) unless the box is completely shaded in representing complete LD. A single haplotype block was generated for this cohort.

Fig. 6

IPA-generated network containing microarray genes that were identified as significant and that showed a greater than two fold difference when comparing neovascular AMD patients and their unaffected siblings with the addition of HTRA1 and HTRA1 interacting molecules. The molecules which are shaded are those that were identified in our study. Green indicates down-regulation while red indicates up-regulation when comparing affected to unaffected siblings. HTRA1 is indicated in blue. Solid lines indicate direct relationships and dotted lines indicate indirect relationships as identified in previously published literature (http://www.ingenuity.com/index.html). The individual shapes represent the family of molecule, for example, the shape of RORA indicates a ligand-dependent nuclear receptor. A full listing of these proteins, their official name, and the family of molecules to which they belong can be found in Supplementary Table 5.