Introduction

Age-related macular degeneration (AMD) is a major cause of severe, irreversible visual impairment among elderly individuals in developed countries [1–3]. Despite recent advances in anti-VEGF therapy [4, 5], AMD is often difficult to control, owing to deterioration of macular function [6]. The origins of AMD are both environmental, for example smoking [7], and genetic [8–11]. Thus, current smoking increases the odds of AMD development 3.5-fold, and having the TT allele at age-related maculopathy susceptibility 2 (ARMS2) A69S (rs10490924) or the CC allele at complement factor H (CFH) Y402H (rs1061170) increases it 6.2-fold or 4.1-fold, respectively [12]. Recently, the AMD Gene Consortium performed a genomewide association study (GWAS) and reported 19 AMD-susceptibility single-nucleotide polymorphisms (SNP) [13]. The association of many of these loci with development of AMD among Asian populations has been confirmed, and it is also recognized that their effect can differ between Asian and Caucasian individuals [13, 14].

Although GWAS achieved great success in identifying common (usually defined as minor allele frequency (MAF) of ≥1 %) AMD-susceptibility SNP, these SNP do not sufficiently explain the heritability of AMD (at best, they explain 65 % of the heritability) [13, 14]. To explain the missing heritability, researchers changed their focus to rare variants. In contrast with common SNP, rare susceptibility variants tend to have larger effects, because they sometimes result in loss of function. Studies using next-generation sequencing have revealed that rare variants located within CFH, C3, CFI, and C9 confer a high risk of AMD [15–19]; in particular, CFH R1210C has a large effect (odds ratio 18.8), and both relatively early-onset and drusen-rich AMD have been reported for individuals with this mutation [19]. This mutation is also known to be responsible for atypical hemolytic uremic syndrome (aHUS) [20] and primary glomerulonephritis [21]. However, its association with AMD has been reported for Caucasian cohorts only, not for other ethnicities. Because rare variants are sometimes over-represented in particular ethnic groups, information about whether this mutation occurs among Japanese populations and causes AMD is of interest, especially considering its large effect. Therefore, in this study, we evaluated 1364 Japanese patients with AMD for the distribution of CFH R1210C.

Patients and methods

All procedures adhered to the tenets of the Declaration of Helsinki. The institutional review boards and ethics committees of Kyoto University Graduate School of Medicine, Fukushima Medical University, and Kobe City General Hospital approved the protocols. All patients were fully informed of the purpose and procedures of the study, and written consent was obtained from each patient.

All of the patients were recruited from Kyoto University Hospital, Fukushima Medical University, and Kobe City General Hospital. Comprehensive ophthalmic examinations were conducted, including dilated fundus examination, optical coherence tomography (OCT), fluorescein angiography (FA), and indocyanine angiography (ICGA). Exudative AMD was diagnosed by retina specialists in accordance with the International Classification System for age-related maculopathy [22], as described elsewhere [23]. Patients without a clear diagnosis of AMD subtypes (owing to an older lesion) were excluded from the analysis.

One thousand three hundred and sixty-four patients were included in the study. Genomic DNA samples were prepared from peripheral blood by use of a DNA extraction kit (QuickGene-610L; Fujifilm, Tokyo, Japan). The genotype of CFH R1210C (rs121913059) was determined for all samples by use of a custom-made Taqman genotyping assay (Taqman SNP assay with the ABI Prism 7700 system; Applied Biosystems, Foster City, CA, USA). The primer and probe sequences used for this assay were: 5′-AGGTGGACAGCCAAACAGAAG-3′ (forward primer); 5′-AGTTTCCCATCCCAACATGTTGT-3′ (reverse primer); TGTGTGAGAACGTGATGAA (Taqman probe—VIC labeled); and TGTGTGAGAACATGATGAA (Taqman probe—FAM labeled). This mutation was validated by conventional Sanger sequencing, which was performed with an Applied Biosystems 3130xl Genetic Analyzer (Life Technologies, Grand Island, NY, USA). We used the same primer set as in our previous report: Fwd-5′-CCCTAATTCTCATACATTAAACATCG-3′ and Rev-5′-GACACAACCGTTAGTTTTCCAG-3′ [19]. To determine the frequency of the mutation among healthy individuals, we referred to the human genetic variation database (HGVD; http://www.genome.med.kyoto-u.ac.jp/SnpDB/; in the public domain). The HGVD contains genetic variations determined by whole-exome sequencing of 1208 Japanese controls.

We reviewed the medical records of individuals with the CFH R1210C variant to obtain details of both their clinical course and their disease characteristics. Glomerular filtration rate (GFR) was estimated from serum creatinine levels by using the CKD-EPI equation for Japanese females: 194 × SCr−1.094 × Age−0.287 × 0.739.

Samples from 1300 of the AMD patients were also genotyped by use of the HumanOmniExpress or HumanExome BeadChip array (Illumina, San Diego, CA, USA) then imputation by use of a 1000-Genomes Project cosmopolitan reference (details are described in the Supplementary Note). Thus, we looked up the genotypes of other AMD-susceptibility SNP whose associations with AMD were confirmed for both Caucasian [13] and Asian [14] populations. These SNP include ARMS2 (rs10490924), CFH (rs800292), C2/CFB (rs429608), C3 (rs2241394), APOE (rs4420638), CETP (rs3764261), VEGFA (rs943080), TNFRSF10A (rs13278062), CFI (rs4698775), TGFBR1 (rs334353), and ADAMTS9 (rs6795735). We constructed a multilocus genetic risk score (GRS) by summing up the number of risk alleles for each SNP, weighted by its reported effect size (log odds ratio). Because rs13278062 in TNFRSF10A could not be well imputed, we excluded rs13278062 when constructing the score. The effect sizes used in this analysis are summarized in Supplementary Tables 1 and 2.

Results

Details of the patients’ ages, sex, and AMD subtypes are summarized in Table 1

The Taqman assay revealed that only 1 of the 1364 patients with neovascular AMD had the CFH R1210C mutation (MAF = 0.037 %), in contrast with 0.83 % reported for Caucasian patients with AMD [19]. Direct sequencing around this mutation (Fig. 1), performed for this patient, confirmed a heterozygous allele change of CGT to TGT (arginine to cysteine). This mutation was not detected in the 1208 Japanese controls (MAF = 0 %).

Table 1 Patients’ (N = 1364) characteristics
Full size table
Fig. 1
figure 1

Sanger sequencing confirmation of CFH R1210C. A heterozygous allele change of CGT to TGT (arginine to cysteine) was observed in only 1 of the 1364 Japanese patients with AMD

Full size image

The distribution of genetic risk scores for 1300 of the AMD patients is shown in Fig. 2. The patient with the CFH R1210C variant had a GRS of 3.54 located at the 47.2 percentile, meaning that the patient had almost median genetic susceptibility to AMD, except for the presence of the CFH R1210C mutation.

Fig. 2
figure 2

Distribution of the genetic risk score for Japanese patients with AMD. The genetic risk score were normally distributed. The patient with CFH R1210C had a genetic risk score of 3.54, which was in the lower 47.2 percentile of the Japanese patients with neovascular AMD

Full size image

Case study: patient with AMD and the CFH R1210C mutation

The patient was a 70-year-old Japanese woman who visited Fukushima Medical University Hospital with progressively blurred vision associated with metamorphopsia in the right eye. She had neither a significant medical history nor a history of previous ocular disease. Her family medical history was unavailable. On examination, the best-corrected visual acuity was 18/20 OD and 20/20 OS. No particular abnormality was found in the anterior segment, but moderate cataracts were found in both eyes. Dilated funduscopic examination of the left eye was unremarkable. In the right eye, a hemorrhagic pigment epithelial detachment (PED) with extravagant exudation was present, but drusen were absent (Fig. 3a). FA revealed hyperfluorescence pooling in the PED, and the ICGA clearly revealed a polypoidal lesion at the edge of the PED. Except for high blood sugar (205 mg/dL), the results of the hematologic tests were within normal limits, including an estimated glomerular filtration rate of 74.1 mL/min.

Fig. 3
figure 3

Case presentation of the patient with AMD and the CFH R1210C variant. a A 70-year-old Japanese woman visited Fukushima Medical University Hospital with progressively blurred vision associated with metamorphopsia in the right eye. Dilated funduscopic examination of the left eye was unremarkable. In the right eye, there was a hemorrhagic pigment epithelial detachment with extravagant exudation, but no drusen. Fluorescein angiography revealed the classic pattern of choroidal neovascularization, and indocyanine green angiography clearly revealed a polypoidal lesion (arrow). b Although 1 treatment using intravitreal bevacizumab and 2 treatments using focal photocoagulation stabilized the disease activity, her visual acuity remained 1/50 OD. At her final visit (6 years and 9 months after the first visit), dilated funduscopic examination revealed several drusen in the left eye and diffuse RPE damage and a massive fibrotic scar in the right eye

Full size image

One month after her first visit, the patient’s visual acuity had decreased to 8/20 OD because of subretinal hemorrhage; we therefore conducted photodynamic therapy (PDT). After 4 rounds of PDT at intervals of 3–6 months, disease activity was controlled. However, subretinal hemorrhage occurred again 16 months after the final PDT, which caused further visual acuity loss to 1/20 OD. The patient was treated once with intravitreal bevacizumab and twice by focal photocoagulation, which stabilized disease activity but resulted in no visual acuity improvement (visual acuity = 1/50 OD). At her final visit (6 years and 9 months after her initial visit), dilated funduscopic examination revealed several drusen in the left eye. She also had diffuse RPE damage and a massive fibrotic scar in the right eye (Fig. 3b).

Discussion

Age-related macular degeneration is a complex disease the origins of which are both genetic and environmental. Environmental risk factors have been identified in several epidemiologic studies, and genetic aspects of AMD have also been intensively investigated. In particular, GWAS have contributed substantially to understanding of the effect of common variants [8, 13, 14]. The effect of rare variants has not been elucidated until recently, however. In this study we evaluated one such variant, CFH R1210C, and found that it rarely occurs in Japanese patients with neovascular AMD.

CFH R1210C is a missense mutation whose association with AMD was reported in 2011 [19]. Because of an arginine-to-cysteine amino acid change, the C-terminal function of CFH is disrupted, leading to defective binding to C3d, C3b, heparin, and endothelial cells [24, 25]. Although this mutation is also responsible for aHUS and primary glomerulonephritis [20, 21], the only patient who had CFH R1210C in our study did not have a medical history of aHUS or renal dysfunction. This is not surprising, given that the penetration of this mutation is not high. A previous study of patients with this mutation found that they, also, did not have these conditions [19]. However, the fact that AMD shares genetic determinants with other diseases and that this mutation was rarely found among a Japanese population, compared with findings for Caucasian populations, is of interest. For example, CETP D442G (rs2303790), which increases the risk of AMD and reduces the risk of myocardial infarction, is exclusive to East Asian populations [14, 26]. These observations suggest two things to us. First, rare variants are sometimes ethnicity-specific. Thus, there might be additional rare AMD-susceptible variants within CFH that can be found only in Asian individuals. Identifying such variants would facilitate our understanding of the effect of CFH in AMD development. Second, rare variants can be associated with several diseases because many of these variants directly affect function. In that respect, rare aHUS susceptibility variants within CFH [27] might be good candidates for examination of novel variants for susceptibility to AMD.

Individuals with the CFH R1210C variant develop signs of AMD earlier than those without it, with a median age of onset of 65 years (range 35–75 years). In addition, the retinal phenotype for these individuals typically includes numerous small, medium, and large drusen [19]. The patient with CFH R1210C identified in this study did not have drusen at the time of diagnosis, however; drusen were not observed until her final visit. Furthermore, her AMD subtype was PCV, which has been predominantly observed among Asian populations. This finding indicates that dysfunction of CFH does not solely regulate the retinal phenotype. Retinal phenotypes, for example drusen and polypoidal lesion formation, might arise as a result of an interactive effect between CFH and other genes. Although target resequencing, whole-exome sequencing, and whole-genome sequencing using next-generation sequencers have been performed for more than 2000 Caucasian individuals with AMD, and have identified several rare AMD-susceptibility variants within known AMD-susceptibility genes [15–19], such sequencing has not been performed for Asian populations. Considering the ethnicity-specific nature of rare variants, high-throughput sequencing should also be applied to Asian patients with AMD; this, with the previous research on Caucasian patients, will help to clarify the mechanisms of AMD and its variable phenotypes (Table 2).

Table 2 Genotype distribution of CFH R1210C
Full size table

As far as we are aware, this is the first study to evaluate the CFH R1210C variant among a large cohort of Japanese patients with AMD. We showed that the CFH R1210C variant was present in 1 of the 1364 Japanese patients with AMD. This patient had PCV without initial drusen, which is not a typical phenotype for this variant. The ethnicity-specific nature of rare variants encourages us to perform more analysis with next-generation sequencing among Japanese cohorts with AMD; this may help further elucidate genetic differences between Japanese and Caucasian populations with AMD.