Abstract
Purpose::
Toll-like receptors (TLRs) are pattern recognition receptors that recognize pathogen associated molecular patterns. TLRs play an important role in innate immunity as primary sensors of invading pathogens. TLR4 binds lipopolysaccharide (LPS), and coding polymorphisms within its extracellular domain were associated with AMD in a previous study. We sought to replicate this observation and determine if variation in the 10 human toll-like receptor (TLR) genes was associated with AMD.
Methods::
Sixty-eight single nucleotide polymorphisms (SNPs) were genotyped in a Caucasian cohort of 396 subjects with AMD and 181 controls. The SNPs were selected based on the criteria that included a minor allele frequency of 0.10 or greater, predicted non-synonymous changes in protein sequence, and linkage disequilibrium between SNPs within each locus. We also evaluated 2 coding SNPs in TLR4 used in the previous report. Statistical associations were reported for allele and genotype (GT) association.
Results::
A Leu412Phe polymorphism (rs3775291) in TLR3 (4q35) was significantly associated with AMD (Allele, P <0.01; GT, P < 0.02). This significance level was comparable to that of the most significant SNP (rs547154) at the BF/C2 locus in this cohort (allele, P < 0.01; GT, P < 0.02). Six other SNPs were not (P > 0.25). Haplotype analyses revealed a highly significant risk haplotype (likelihood ratio test, P < 0.001) and suggested that more than one SNPs are required to explain the association with AMD. We were unable to replicate association with the previously reported coding SNPs in the TLR4 locus (9q33.1). However, an intronic SNP (rs1927911) in TLR4 showed a trend for association with AMD (Allele, P < 0.10; GT, P < 0.04) while the other 4 SNPs did not (P > 0.1400). The TLR10, TLR1, and TLR6 genes are adjacent to each other on 4p14. The coding SNP Gly381Asp (rs11466655) and the 2 non-coding SNPs (rs4274855 and rs11466617) within TLR10 trended toward an association with AMD (Allele, P = 0.07, 0.08 and 0.08 respectively). The other 23 SNPs across these three genes were not associated with the risk of AMD (P > 0.12). Ten SNPs in TLR5 (1q41-42), TLR9 (3p21.3)and TLR2 locus (4q32) were not associated with AMD. Some SNPs in the X-linked TLR genes (TLR7 and TLR8) showed association with AMD, but were not in Hardy Weinberg equilibrium.
Conclusions::
The coding variation Leu412Phe within TLR3 was significantly associated with increased risk for AMD. Other TLR genes may be associated with AMD, but require further study to replicate our observations and better understand the molecular changes that may be associated with AMD.
Keywords: age-related macular degeneration • receptors • retina