Genetic and functional dissection of HTRA1 and LOC387715 in age-related macular degeneration

Abstract

A common haplotype on 10q26 influences the risk of age-related macular degeneration (AMD) and encompasses two genes, LOC387715 and HTRA1. Recent data have suggested that loss of LOC387715, mediated by an insertion/deletion (in/del) that destabilizes its message, is causally related with the disorder. Here we show that loss of LOC387715 is insufficient to explain AMD susceptibility, since a nonsense mutation (R38X) in this gene that leads to loss of its message resides in a protective haplotype. At the same time, the common disease haplotype tagged by the in/del and rs11200638 has an effect on the transcriptional upregulation of the adjacent gene, HTRA1. These data implicate increased HTRA1 expression in the pathogenesis of AMD and highlight the importance of exploring multiple functional consequences of alleles in haplotypes that confer susceptibility to complex traits.

Figure 1. Genes and the main SNPs in the AMD 10q26 region.

The schematic diagram showing a 100 kb region subject to re-sequencing, the AMD-associated 21.5 kb region containing 22 SNPs in the same LD block with a disease haplotype. SNPs in bold were chosen for further studies.

Figure 2. The haplotype block in D' of rs2736911, rs10490924, in/del/Wt, and rs11200638 in the Utah case-control cohort.

The risk haplotype C-T-in/del-A is strongly associated with advanced AMD (P = 4.05×10⁻²⁸); the non-risk haplotype T-G-Wt-G is significantly associated with protection of advanced AMD (P = 4.00×10⁻⁴). In/del and Wt refers to presence or absence of 54 base pair insertion and 443 base pair deletion, while the letters G, T, A refer to nucleotides for the respective SNPs at rs2736911, rs10490924, and rs11200638.

Figure 3. Endogenous expression studies comparing effects of genotype on HTRA1 and LOC387715.

mRNA Expression levels of LOC387715 and HTRA1 in human placenta tissues according to haplotypes defined by SNPs rs2736911, rs10490924, in/del/Wt, and rs11200638. Human placenta tissues from seven individuals with a homozygous disease haplotype (C-T-in/del-A), five individuals with a homozygous protective haplotype (C-G-Wt-G), and five individuals with heterozygous C/T alleles at rs2736911 and homozygous protective haplotype at rs10490924, in/del/Wt, and rs11200638 (C/T-G-Wt-G) were harvested and analyzed. (A) In comparison to the haplotype C-G-Wt-G, mRNA levels of LOC387715 with a homozygous disease haplotype C-T-in/del-A and a haplotype C/T-G-Wt-G were 4.7-fold and 2.3 fold lower, respectively. The mean ± SD is given for each genotype. (B) mRNA levels of HTRA1 with C-T-in/del-A was 2.7-fold higher compared to that of C-G-Wt-G, and there was comparable expression levels between C-G-Wt-G and C/T-G-Wt-G. The mean ± SD is given for each genotype.

Figure 4. Heterologous Luciferase assays in vitro and in vivo.

Effects of the in/del variants on HTRA1 expression in cultured human RPE cells and mouse RPE in vivo. (A) Schematic diagram of constructs for luciferase reporter assays. The L and the S denote long and short promoter constructs, A/G represents the allele at rs11200638, while in/del (black circle) refers to long promoter construct with an in/del. Different HTRA1 promoter sequences corresponding to risk and wild-type alleles of in/del and rs11200638 (WT(L-A)), WT(L-G), (MT(L+in/del)), (WT(L+in/del), WT(S) and MT(S)) were placed upstream of a pGL3 reporter. (B) Luciferase activities in cultured human RPE cells transfected with different HTRA1 promoter reporter constructs. The pGL3-Basic vector without insert (negative) was transfected into human RPE cells as a negative control. Renilla luciferase plasmid pTK-RL was cotransfected with each construct as an internal control for normalization. Normalized luciferase activity was measured in five independent experiments. The mean ± SD is given for each construct. (C) Luciferase activities in mouse RPE in vivo corresponding to different HTRA1 promoter reporter constructs. Reporter constructs were injected into the subretinal space and electroporated into mouse RPE cells. Eighteen, nineteen and sixteen eyes were injected with WT(L), MT(L+in/del) and WT(L+in/del) constructs respectively. The firefly luciferase value was divided by the Renilla luciferase value to give the normalized luciferase activity. This was divided by the normalized luciferase activity of empty-PGL3 and pRL-CMV to give the relative luciferase activity ratio. Each bar represents the mean (± SEM).