Introduction

Age-related macular degeneration (AMD) is an irreversible, progressive, and sight-threatening macular disease in the elderly population, affecting over 50 million individuals worldwide.1 Clinically, the late stage of AMD can be divided into two subgroups: geographic atrophy and exudative/neovascular forms.2 Exudative AMD involves choroidal neovascularization (CNV), which is an ingrowth of new vascular vessels from choriocapillaries beneath retinal pigment epithelium (RPE) arising toward subretinal space and the neuroretina. These new vessels have greater tendency of leakage and bleeding.3 CNV invades into the retina via the Bruch’s membrane and RPE, leading to serous or hemorrhagic detachments of either the RPE or neuroretina. Consequently, this leads to subretinal fibrosis4 and ultimately irreversible blindness. The exudative form accounts for >80% of significant rapid visual loss from AMD.5, 6

Inflammation has an influential role in exudative AMD.7 Drusen, the hallmark of AMD, contain inflammatory components, such as complement activators and inhibitors, membrane attack complex as well as inflammasome.8 Local RPE cell damage by aging activates and recruits choroidal dendritic cells, contributing to the formation of drusen and CNV.9 The choroidal infiltrate is also composed of macrophages and lymphocytes.10, 11 The activated inflammatory cells secrete collagenase and elastase, which erode the thinned Bruch’s membrane and facilitate migration of choroidal capillaries.12 Moreover, interaction between inflammatory cells and vascular endothelial growth factor (VEGF) also indicates the role of inflammation in CNV development.13 In addition, significant association of inflammatory markers with AMD suggests that inflammation could participate in AMD pathogenesis.14, 15

As inflammation is involved in AMD pathogenesis, inflammation-related genes should be studied in AMD patients. The formyl peptide receptor 1 (FPR1) gene, located on chromosome 19q13.4,16 encodes a seven transmembrane domain G protein-coupled receptor. It is expressed mainly by mammalian phagocytic leukocytes and involved in inflammatory responses by activation of chemotatic peptides.17 This receptor allows phagocytic cells to recognize the presence of exogenous organisms, such as bacteria, and mediates the traffic of phagocytes to the sites of tissue damage.18 Therefore, FPR1 is a crucial molecule in innate immunity. Moreover, its ligand, the N-formyl peptides, is produced only by bacteria and mitochondrial proteins in nature.19 Mitochondrial dysfunction is also associated with AMD pathogenesis.20 Hence, we hypothesized that FPR1 could have a positive role in AMD pathogenesis.

Polypoidal choroidal vasculopathy (PCV) shares many similarities in clinical features with exudative AMD. But they differ in clinical course, imaging feature, response to treatment, and prognosis.21 In order to identify the role of FPR1 in exudative AMD, the whole-coding region of FPR1 gene was sequenced in this study. In addition, the FPR1 sequence was screened in PCV and compared with that of exudative AMD. Furthermore, the interactions and combined effects of FPR1 with complement factor H (CFH) and high temperature requirement factor A1 (HTRA1) were also determined as CFH and HTRA1 are associated with both exudative AMD and PCV.22, 23

Materials and methods

Study subjects

A total of 554 participants were recruited at the Prince of Wales Hospital Eye Centre, including 155 exudative AMD patients, 179 PCV patients and 220 age-matched control subjects (Supplementary Table 1). All study subjects were given complete ophthalmic examinations. AMD was graded according to an international classification and grading system.24 Patients with exudative AMD had non-drusenoid RPE detachment, choroidal neovascularization, serous or hemorrhagic retinal detachments, subretinal or sub-RPE hemorrhage, or fibrosis. The diagnosis of PCV was distinguished from exudative AMD by fluorescein angiography and indocyanine green angiography (ICGA).25 PCV patients had subretinal red or orange nodules and hemorrhagic pigment epithelial detachment and characteristic sacculated vascular abnormalities in the inner choroid as visualized on ICGA. The control subjects were recruited from elderly people >60 years of age. According to the complete ophthalmic examinations, the control subjects did not have any identifiable signs of AMD, PCV, or other retinal or optic nerve diseases except for mild senile cataract and refractive errors. Smoking habits were also recorded. A smoker was defined as a person who had smoked at least five cigarettes daily for >1 year. The study subjects were divided into two groups: those who had never smoked, those who were ex-smokers, and current smokers. The study protocol, approved by the Ethics Committee for Human Research at the Chinese University of Hong Kong, is in accordance with the tenets of the Declaration of Helsinki. Informed consent was obtained from each study subject.

Sequence analysis

Genomic DNA was extracted (Qiagen QIAamp DNA Blood Mini kit, Qiagen, Hiden, Germany) from peripheral blood samples following the supplier’s instructions. All samples were screened for sequence alterations in the entire coding region (exon 2) and intron–exon junctions of FPR1 (ENSG00000171051) by PCR with specific primer sets (Supplementary Table 2), followed by direct DNA sequencing (BigDye Terminator Cycle Sequencing Reaction Kit, v3.1, Applied Biosystems, Foster City, CA, USA) on a DNA sequencer (ABI 3130XL, Applied Biosystems). In addition, all samples were genotyped for CFH rs800292 and HTRA1 rs11200639 using direct DNA sequencing with specific primers.22, 23

Amino-acid substitution analysis

For the non-synonymous variants identified in FPR1, functional effects of the amino-acid substitutions were evaluated and predicted by two web-based analyzing programs: PolyPhen (polymorphism phenotyping, http://genetics.bwh.harvard.edu/pph/, provided by the Bork Group and the Sunyaev Lab, Brighamand Women’s Hospital, Harvard Medical School, Boston, MA, USA) and SIFT (Sorting Intolerant from Tolerant, http://sift.jcvi.org/, provided in the public domain by the J Craig Venter Institute, Rockville, MD, USA). The PolyPhen score>1.0 would be considered as ‘probably damaging’ and otherwise would be considered as ‘benign’. The SIFT score ≤0.05 is predicted to be damaging, and >0.05 is predicted to be tolerated.

Statistical analysis

All the identified polymorphisms were assessed for Hardy–Weinberg equilibrium in controls using χ2 analysis. Allelic and genotypic distributions were compared using the χ2-test or Fisher’s exact test (SPSS, version 16.0; SPSS Science, Chicago, IL, USA). Bonferroni’s correction was used for multiple testing corrections. Linkage disequilibrium (LD) plots and haplotype-based association analyses were generated by Haploview (v4.2, http://www.broadinstitute.org/).26 All associated polymorphisms in FPR1 were assessed by logistic regression (SPSS). Logistic regression analysis included only gender as indicator because the recruited control subjects were older, excluding the potential patients with major late-onset ocular diseases. SNPs that remained significant after adjusting for other individual SNPs were selected for interaction and combined effect analyses. The pair-wise association of FPR1 rs78488639 with CFH rs800292, HTRA1 rs11200638 or smoking was evaluated using the χ2-test or Fisher’s exact test (SPSS).

Results

The gender distribution of PCV patients in our study was ∼2 : 1 (male:female, 123 : 51), whereas that of exudative AMD patients (84 : 70) and control subjects (101 : 119) was ∼1 : 1 (Supplementary Table 1). This was compatible to a previous study.27

A total of 28 single nucleotide polymorphisms (SNPs) were identified in FPR1 (Table 1). Variants rs7253355 (c.1-342C>T) and c.513G>A (T171T) were excluded for further analysis as they violated Hardy–Weinberg equilibrium in the control subjects. Among the remaining 26 polymorphisms, 18 were rare SNPs (minor allele frequency <3%) and did not show significant association. They were also excluded for further analysis. The triallelic variant rs1042229 (c.576T>G/C, N192K/N) was analyzed independently. The homozygous of risk allele G was associated with exudative AMD (P=0.0394, odds ratio (OR)=2.27, 95% confident interval (CI): 1.08–4.74), but not with PCV (P=0.241) or in comparison between exudative AMD and PCV (P=0.137). Besides, the codon change caused by risk allele C was synonymous and did not show any association under either autosomal recessive or dominant model (data not shown). Thus, totally seven polymorphisms (c.117C>T, rs78488639, rs2070745, rs2070746, rs5030880, rs867228, and rs867229) were further analyzed (Table 1). SNP rs78488639 (c.289C>A, L97M) was associated with both exudative AMD (P=0.043) and PCV (P=0.029), whereas rs867229 (c.1053+196C>T) was associated with only exudative AMD (P=0.0026). Genotypic difference between exudative AMD and PCV was detected in rs867229 (P=0.014). Haplotype analysis revealed a LD block including two SNPs, rs2070746 and rs5030880, in both exudative AMD and PCV (Figure 1). However, haplotype-based association analysis could not identify any associated haplotype in exudative AMD or PCV (data not shown).

Table 1 Sequence variants detected in FPR1
Full size table
Figure 1
figure 1

Haplotype block structure for the seven common polymorphisms. The haplotype analysis revealed a LD block lying across rs2070746 and rs5030880 in (a) AMD and (b) PCV.

Full size image

The heterozygous genotype of rs78488639 contributed a 2.05- and 2.27-fold of increased risk, respectively, to exudative AMD (P=0.043) and PCV (P=0.016; Table 2). Homozygous genotype of rs2070745 was associated with PCV (P=0.034, OR=1.80, 95% CI: 1.04–3.09). Heterozygous genotypes of rs2070746 and rs867229 were associated with a decreased risk in exudative AMD (P=0.019, OR=0.57, 95% CI: 0.35–0.91; P=0.0082, OR=0.54, 95% CI: 0.34–0.86, respectively). Heterozygous genotype of rs2070746 was also different between exudative AMD and PCV (P=0.0086, OR=0.51, 95% CI: 0.31–0.85). Moreover, five novel rare variants were identified. Amino-acid substitution prediction suggested that the protein function of c.368G>A (R123H) could be affected with a SIFT score of 0.01 (Table 3). This was further supported by the PolyPhen prediction with a score of 2.75. However, neither individual rare variant nor pooling of the variants were associated with exudative AMD or PCV (data not shown). The association became not significant after Bonferroni’s correction (P=0.05/28=0.0018).

Table 2 Odds ratios of FPR1-associated polymorphisms
Full size table
Table 3 Amino-acid substitution prediction of novel rare FPR1 variants
Full size table

The G allele of CFH rs800292 increased the risk for both exudative AMD (homozygous: OR=2.60, 95% CI: 1.27–5.31, P=0.0074; heterozygous: OR=1.41, 95% CI: 0.68–2.92, P=0.36) and PCV (homozygous: OR=3.12, 95% CI: 1.42–6.86, P=0.0035; heterozygous: OR=2.50, 95% CI: 1.14–5.51, P=0.020). The risk was also increased by the A allele of HTRA1 rs11200638 in exudative AMD (homozygous: OR=12.48, 95% CI: 6.53–23.83, P=1.07 × 10−16; heterozygous: OR=2.46, 95% CI: 1.31–4.64, P=0.0046) and PCV (homozygous: OR=5.24, 95% CI: 2.89–9.50, P=1.79 × 10−8; heterozygous: OR=2.51, 95% CI: 1.47–4.29, P=5.94 × 10−4). Logistic regression analysis revealed that only FPR1 rs78488639 remained significant after adjusting for gender and other individual-associated SNPs in both exudative AMD (P=0.032) and PCV (P=0.022). Further interaction analysis identified significantly positive interactions among FPR1 rs78488639, CFH rs800292, and HTRA1 rs11200638 in exudative AMD (P=0.022) and PCV (P=0.023). In the combined effect analysis, ORs of combined FPR1 rs78488639 CA and CFH rs800292 GG genotypes were approximately twice greater than the individual ORs of FPR1 rs78488639 CA or CFH rs800292 GG in both exudative AMD (P=0.0062, OR=4.83, 95% CI: 1.51–15.51) and PCV (P=0.019, OR=4.03, 95% CI: 1.22–13.28; Table 4). A combined risk OR of 10.47 (P=2.22 × 10−4, 95% CI: 2.72–40.29) was identified in the heterozygous risk alleles of these two variants in the PCV patients, but not in exudative AMD (P=0.133). In addition, ORs of combined FPR1 rs78488639 CA and HTRA1 rs11200638 AA genotypes were at least 1.5-fold higher than the individual ORs of HTRA1 rs11200638 AA and >6-fold higher than FPR1 rs78488639 CA in both exudative AMD (P=1.02 × 10−4, OR=19.47, 95% CI: 3.75–100.97) and PCV (P=7.45 × 10−4, OR=14.19, 95% CI: 2.72–74.20). Furthermore, combined ORs of FPR1 rs78488639 CA and smoking were >5-fold higher than FPR1 rs78488639 CA in both exudative AMD (P=0.010, OR=10.93, 95% CI: 1.30–92.10) and PCV (P=9.96 × 10−4, OR=16.94, 95% CI: 2.06–139.38; Table 5).

Table 4 Combined effects of FPR1 rs78488639 with CFH rs800292 and HTRA1 rs11200638 in exudative AMD and PCV
Full size table
Table 5 Combined effects of FPR1 rs78488639 with smoking in exudative AMD and PCV
Full size table

Discussion

Inflammation is an important component in AMD etiology. Chronic localized inflammation initiated and stimulated by drusen, together with the accumulation of lipofuscin, compromise RPE function.28, 29 The injured RPE actively recruits choroidal dendritic cells,29 microglia, macrophages,30 and leukocytes,31 which express VEGF and stimulate CNV growth,32 the hallmark of exudative AMD. The initiation of AMD by inflammation and infections has been supported by clinical observations of Chlamydia pneumonia and cytomegalovirus infections in AMD patients.33, 34 As FPR1 is a receptor for phagocytic cells recognizing the presence of exogenous organisms and mediating the traffic of phagocytes to the sites of tissue damage,18 FPR1 could have an important role in AMD initiation by infections.

The involvement of FPR1 in innate immunity is supported by multiple evidences. First, the release of formylated peptides from mitochondria secondary to cell death might attract phagocytic leukocytes through the receptor.35 Second, the Fpr1−/− mice model had accelerated mortality and increased bacterial burden in the liver and spleen early after infection,36 showing a role of Fpr1 in the host defense. Moreover, rs5030879 (c.348C>T, I116I) in FPR1 is associated with aggressive periodontitis,37 which is an infectious–inflammatory disease. In addition, c.329T>C (F110S), c.378C>G (C126W), and rs5030878 (c.32C>T, I11T) of FPR1 are significantly correlated with defective polymorphonuclear neutrophil (PMN) function and C-reactive protein.38, 39 However, limited data reported the association of FPR1 with AMD, and also PCV. In this study, the results of FPR1 sequencing in both exudative AMD and PCV revealed five associated polymorphisms (Table 1), indicating a role of FPR1 in the inflammation mechanisms of exudative AMD and PCV.

Formyl peptide receptors (FPRs) are expressed mainly in PMNs and monocytes, but also found in neurons, microglial, and glial cells.40 In the inflammatory response, phagocytes are induced to move directionally through chemotaxis.41 When the ligands (N-formyl peptides) bind to FPR1, FPR1 will direct the activated phagocyte migration through the PI3k/PTEN pathways. This implies an important role of FPR1 in inflammatory cell migration. FPR1 also induces superoxide production, and this oxidative stress could further damage the RPE cells.42, 43 Furthermore, FPR1 expression can be enhanced by inflammation in Alzheimer’s disease,44 which is associated with age as AMD and involves accumulation of amyloid β (Aβ) in senile plaques. Aβ has been found in AMD drusen.45 Therefore, FPR1 likely has an etiological role in AMD.

CFH and HTRA1 are well recognized as susceptibility genes for AMD.22, 23 CFH is an inhibitor in alternative complement pathway, controlling complement activation.46 Abnormal regulation of complement activation by CFH might contribute to AMD pathogenesis.47 HTRA1 is also involved in inflammatory responses.48, 49 In this study, FPR1 significantly interacted with CFH and HTRA1, with respective combined ORs of 10.47 and 19.47 for increasing disease risk (Table 4). Moreover, the risks of FPR1 to exudative AMD and PCV were also increased by smoking, with respective combined ORs of 10.93 and 16.94 (Table 5). The combined effect with smoking has been demonstrated in our previous studies on CFH and HTRA1.22, 23 Therefore, the role of FPR1 as well as the contribution of inflammation genes and smoking to AMD and PCV are further implicated.

SNP rs1042229 is a triallelic (T, G or C) variant, whereas the majority of SNPs are biallelic. Its G allele specifies lysine and T/C specifies asparagines at codon 192. The N192 locates in the center of the second extracellular loop, which the lysine substitution may alter the ligand-binding specificity and binding affinity through the binding pocket for N-formyl peptides.50 In this study, the G allele is the risk allele of rs1042229, and homozygous G genotype of rs1042229 was associated with 2.27-fold of increasing risk in exudative AMD, but not in PCV. Another variant (L97M; rs78488639), located in the first extracellular loop, markedly decreases in the affinity of ligand binding.51 This variant showed significant association with both exudative AMD and PCV, contributing an approximately twofold increasing disease risk (Table 2). Differential associations of various FPR1 SNPs with AMD and PCV are also noted, suggesting that exudative AMD and PCV might have different gene variant association patterns.

In summary, significant interactions and combined effects of FPR1 with CFH, HTRA1, and smoking in exudative AMD and PCV were revealed in this study. FPR1 could have a role in AMD and PCV pathogenesis, and its biological functions in the disease mechanisms need to be elucidated.