RAD51B (rs8017304 and rs2588809), TRIB1 (rs6987702, rs4351379, and rs4351376), COL8A1 (rs13095226), and COL10A1 (rs1064583) Gene Variants with Predisposition to Age-Related Macular Degeneration

Background Age-related macular degeneration (AMD) is a progressive neurodegenerative disease of a central part of the neural retina (macula) and a leading cause of blindness in elderly people. While it is known that the AMD is a multifactorial disease, genetic factors involved in lipid metabolism, inflammation, and neovascularization are currently being widely studied in genome-wide association studies (GWAS). The aim of our study was to evaluate the impact of new single nucleotide polymorphisms (SNPs) in RAD51B, TRIB1, COL8A1, and COL10A1 genes on AMD development. Methods Case-control study involved 254 patients diagnosed with early AMD, 244 patients with exudative AMD, and 942 control subjects. The genotyping of RAD51B (rs8017304 and rs2588809), TRIB1 (rs6987702, rs4351379, and rs4351376), COL8A1 (rs13095226), and COL10A1 (rs1064583) was carried out using TaqMan assays by a real-time polymerase chain reaction (RT-PCR) method. Results Statistically significant difference was found in genotype (TT, TC, and CC) distribution of COL8A1 rs13095226 between exudative AMD and control groups (60.2%, 33.6%, and 6.1% vs. 64.9%, 32.3%, and 2.9%, respectively, p = 0.036). Also, comparing with TT+TC, rs13095226 CC genotype was associated with 3.5-fold increased odds of exudative AMD development (OR = 3.540; 95% CI: 1.415-8.856; p = 0.007). Conclusion Our study revealed a strong association between a variant in COL8A1 (rs13095226) and exudative AMD development.


Introduction
Age-related macular degeneration (AMD) is a progressive neurodegenerative disease of a central part of neural retina (macula) [1]. AMD is a leading cause of central vision loss, and while it is diagnosed in elderly people (in those aged 60 and over) [1], the first signs of the disease can occur in people aged 40 [2]. One of the main processes involved in AMD pathogenesis is called drusogenesis [3]. Drusen are described as small particles of lipid, protein, and collagen detachments accumulated between the retinal pigment epithelium (RPE) and the Bruch's membrane (BrM) in the retina [3]. RPE controls fluid transportation between the choriocapillaris and the retina including lipid transportation and metabolism as well as oxygen transportation [4]. The BrM is a five-layer extracellular matrix located between the RPE and the choroid. It regulates metabolic exchange between RPE cells and blood flow from the choroid through a semipermeable filtration barrier [5]. Any alteration in the structure of BrM might impact dysfunction of the RPE and the outer retina [6]. One of the most complicated biological processes which may affect BrM is ageing: age-related processes urge the accumulation of incompletely digested phospholipids [7], combining with oxidative stress results of lipid peroxidation [8] and lysosomal defects in photoreceptor outer segments [9]. Oxidative stress-induced lipid and protein accumulation leads to RPE injury or early RPE cell death with changes in BrM [10,11]. Damaged RPE cells release huge amounts of different cytokines and chemokines which could result in chronic inflammation over time [12][13][14]. One of the cytokines is a vascular endothelial growth factor A (VEGF-A) which can also induce angiogenesis [15][16][17]. This process leads to a growth of new fragile and leaky vessels resulting in exudative hemorrhage and acute vision loss [18]. According to the Age-Related Eye Disease Study (AREDS), AMD can be classified into early, intermediate, and late stages [19]. Early AMD is described as the appearance of drusen and retinal pigment abnormalities; when at least one large druse, numerous medium-sized drusen, or geographic atrophy (GA) without extension to the center of the macula occurs, the intermediate stage of AMD is diagnosed; late AMD is divided into dry AMD with the GA of the RPE and neovascular AMD which is characterized by choroidal neovascularization with detachments of the RPE, hemorrhages, and/or scars [20]. While it is known that the AMD is a multifactorial disease, genetic factors involved in lipid metabolism, inflammation, and neovascularization are currently being widely studied in genome-wide association studies (GWAS). Products of a collagen gene family are composed of types I, II, III, and V of fibril-forming interstitial and type IV basement membrane collagens. Collagens keep the architecture and function of normal tissues as well as the structure and function of the extracellular matrix [21,22]. Recent studies detected an intronic variant in COL8A1 gene (rs13095226 T/C), suggesting an association with advanced AMD [23][24][25][26][27]. The COL8A1 gene is located on human chromosome 3 and encodes one of the two alpha chains of type VIII collagen which is a central component of multiple basement membranes of corneal endothelial cells, including BrM, choroidal stroma, and endothelia of blood vessels, playing a role in the maintenance of vascular integrity and structure [28,29]. BrM and new vessel formation play a key role in pathogenesis of AMD [30]. Another collagen gene family member is COL10A1 gene, located on human chromosome 6, which encodes the alpha chain of type X collagen. This shortchain collagen is expressed by hypertrophic chondrocytes during endochondral ossification, and it was shown that the expression of COL10A1 was significantly downregulated in patients with osteoarthritis [31], but its expression was higher in diverse solid tumors and correlated with tumor vasculature [32]. The new variant near FRK/COL10A1 (rs1999930) was found to be associated with advanced AMD development [24] and suggested the COL10A1 as a potential locus for future association studies. While the oxidative stress has been linked to the various types of DNA damage playing a significant role in ageing and age-related disorders [33,34], few studies [35,36] revealed a significant association between oxidative stress-linked DNA damage and AMD development. Recently, GWAS pinpointed new genetic variations in the RAD51B gene associated with AMD [25,26,37,38]. RAD51B is involved in homologous recombinational repair of DNA double-strand breaks by promoting the activity of the central recombinase [39]. Absence of RAD51B protein is thought to disrupt the formation of the RAD51 nucleoprotein filament, which is the initial stage of homologous recombinational repair [40]. G-protein-coupled receptorinduced proteins, playing a role in the mitogen-activated protein kinases-(MAPK-) related signaling cascade [41,42] which mediates cell proliferation, differentiation, and apoptosis [43] and can regulate lipid metabolism through this pathway [44], are encoded by tribbles pseudokinase 1 (TRIB1) gene located in human chromosome 8. It has also been suggested that the TRIB1 expression is regulated by inflammatory stimulation [45]. There was found one SNP in TRIB1 gene (rs6987702) to be associated with AMD in African American and Mexican American population [46], suggesting a significant role of TRIB1 in AMD development.
Association studies of these SNPs, however, have not been conducted in both (early and exudative) AMD forms. Genetic differences between developments of early and exudative AMD still need to be researched and understood better which might lead to future studies with focus on certain molecular pathways involved in the pathogenesis of AMD. Also, genetic variations combined with other risk factors or molecular changes can benefit physicians in creating personalized genetic therapies and/or lifestyle programs for increased risk individuals.
Considering previously studied links between genetic variations and AMD, we chose three previously described SNPs in RAD51B (rs8017304), TRIB1 (rs6987702), and COL8A1 (rs13095226) and four new genetic loci in RAD51B (rs2588809), TRIB1 (rs4351376 and rs4351379), and COL10A1 (rs1064583) genes as potential biomarkers for early and exudative AMD. The aim of our study was to determine and evaluate their impact on the development of early and exudative AMD in the Lithuanian population.

Materials and Methods
The study was conducted in the Department of Ophthalmology, Hospital of Lithuanian University of Health Sciences and Laboratory of Ophthalmology, Neuroscience Institute, Lithuanian University of Health Sciences. Ethical approval was obtained from the Ethics Committee for Biomedical Research (number: BE-2-/13).
Ophthalmological evaluation, study group formation, DNA extraction and genotyping methods, and statistical analysis were described in detail in our previous studies [47,48].

Demographic Characteristics of the Study Groups.
Our study involved 498 patients (254 patients with a diagnosis of early AMD and 244 patients with exudative AMD) and 942 healthy controls. The control group was formed of 942 subjects classified into different genders, which matched gender classification in the early and exudative AMD group structure; however, subjects of the control group were younger than AMD patients (p < 0 001) ( Table 1).

Hardy-Weinberg Equilibrium.
When genotyping results showed that G allele at TRIB1 rs4351376 was observed in all study subjects (100%), it was not included in further analysis. Hardy-Weinberg equilibrium (HWE) analysis showed that the distribution of genotypes of rs6987702 and rs1064583 deviated from the HWE in the control group and those two SNPs were excluded from further statistical analysis as well (Table 2).

Haplotype Associations with AMD.
In this section, we performed haplotype association analysis and evaluated the impact of haplotype constructed of two SNPs: rs8017304 and rs2588809 on early and exudative AMD development. Linkage disequilibrium (LD) analysis was assessed by both D′ and r 2 measures. The r 2 values were 0.055 and 0.063 in early AMD and exudative AMD analysis, respectively; for the haplotype block, D′ values were 0.425 and 0.457 in early AMD and exudative AMD analysis, respectively.
Unfortunately, haplotype analysis did not reveal any significant associations with early or exudative AMD development after adjustment for age.

Discussion
Our study involved 498 AMD patients (254 patients with the diagnosis of early AMD and 244 patients with exudative AMD) and 942 healthy control subjects. Since we chose 7 SNPs, 5 of those were newly selected as risk factors for AMD development, RAD51B (rs2588809), TRIB1 (rs6987702, rs4351376, and rs4351379), and COL10A1 (rs1064583), and 2 were already associated with AMD development, RAD51B (rs8017304) and COL8A1 (rs13095226). Unfortunately, our study revealed only one SNP (rs13095226) which was associated with 3.5-fold increased odds for exudative AMD development in the population of Lithuania (p = 0 007).
One of the SNPs near the TRIB1 locus called rs17321515 was found affecting lipoprotein metabolism. It was strongly associated with triglycerides (TG), low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C) and even with increased risk of coronary heart disease (CHD), but results remain controversial among the studies [49][50][51]. Another SNP (rs2954021) located in TRIB1 gene was associated with increasing plasma TG levels and coronary artery disease (CAD) risk [52,53] as well as the other three variants: rs2001945, rs2001845, and rs2001844 [54]. Furthermore, rs2001844 was associated with increased plasma TG and reduced HDL-C [54]. Several association studies reported a significant association between HDL-C and AMD development as well [55][56][57][58][59]; unfortunately, others did not reveal any significant associations between HDL-C and AMD development [60][61][62][63][64][65][66]. On the other hand, while the association of HDL-C and AMD is controversial, variant in lipid trait-associated gene TRIB1 rs6987702 showed a significant association with AMD development, but those results did not survive strict corrections for multiple testing [46]. In our study, none of the SNPs in TRIB1 reached a statistically significant association with early or exudative AMD development.  Studies of variants in RAD51B revealed statistically significant associations of SNPs in RAD51B with breast cancer occurrence [67][68][69][70]. Also, RAD51B rs911263 was associated with rheumatoid arthritis (RA) [71,72]; two SNPs (rs11158728 and rs927220) were associated with nasopharyngeal carcinoma development [73], and one (rs34094401) was associated with Parkinson's disease [74]. Also, literature review explains that risk alleles at RAD51B rs8017304, rs13081855 near COL8A1/FILIP1L locus, and rs3812111 in COL10A1 are associated with a greater risk for advanced AMD development [25]. Other three GWAS proved RAD51B gene association with AMD as well [26,37,38]. Seddon et al. conducted a 12-year follow-up study of 2765 individuals and revealed that RAD51B (HR: 0.8; 95% CI: 0.60-0.97, p = 0 03) was significantly related to AMD progression [26]. In another study, Seddon and coauthors [38] analyzed the progression of AMD and found 834 from 2951 subjects, who progressed from no AMD, early AMD, or intermediate AMD to an advanced AMD form, and also confirmed that RAD51B was associated with AMD progression. Fritsche et al. [25] also proved a strong RAD51B association (p < 5 × 10 −8 ) with advanced AMD in Europeans and Asians. Chu et al. analyzed RAD51B influence on AMD using two cohorts from Caucasian and Han Chinese populations, as well. Scientists identified two new SNPs in RAD51B (rs17105278 and rs4902566) and confirmed the association between rs8017304 and AMD development in Caucasians [37], suggesting a significant role of RAD51B associated with DNA damage/DNA repair mechanism in AMD pathogenesis.
Recent studies indicated that variants in COL8A1 gene were associated with AMD development [24][25][26]38]. One of the GWAS studies was performed on 979 patients with advanced AMD and 1709 control subjects and revealed a significant association between rs13095226 and AMD (p = 2 5e − 06) in the European population [23]. Also, in this study, authors found that the frequencies of minor allele C at rs13095226 in patients with polypoidal choroidal vasculopathy (PCV), neovascular AMD (nAMD), and controls were similar to the frequencies in Americans, while differences were found between Americans and Asians [23]. Metaanalysis of advanced AMD confirmed previously found associations to advanced AMD [24], while in contrast, Yang et al.'s study including 300 cases with nAMD, 300 cases with PCV, and 300 control subjects did not show the association of COL8A1 rs13095226 with nAMD or PCV, suggesting that  this variant may not be a risk factor for nAMD or PCV in Chinese subjects [75]. Also, another variant of the same gene, polymorphism of COL8A1 rs13081855, was linked to exudative AMD development [27]. As in non-AMD patients, one of the common variants in COL8A1 gene (rs669676) was found to be associated with myopic choroidal neovascularization (OR: 1.88; 95% CI: 1.18-2.98, p = 0 0076) but did not survive multiple testing [74]. Our study of the Lithuanian population discovered a strong association between a variant in COL8A1 (rs13095226) and exudative AMD development; genotype rs13095226 CC was determined to be associated with 3.5-fold increased odds of exudative AMD development.
We also found a study, which showed the contribution of the extracellular collagen matrix (FRK/COL10A1) pathways to the development of advanced AMD. A coding variant rs3812111 in COL10A1, a collagen protein, could be involved in maintaining the structure and function of the extracellular matrix [24]. Ferrington et al. analyzed that the COL10A1 rs3812111 A/T (nonrisk allele/risk allele) ratio was 0.69/0.65 in controls/age-related macular degeneration [76]. To our knowledge, there are no studies analyzing COL10A1 (rs1064583) polymorphism in patients with AMD.
The following limitations of this study have to be declared: (i) Patients with atrophic AMD have not been analyzed (ii) Sample size of AMD and healthy controls was relatively small (iii) Possibility of other risk factors was not included in the study A thorough medical examination of the study objects can be acknowledged as one of the main strengths of our study. All patients were examined for chronic infectious and noninfectious diseases by a general practitioner. It is also important to highlight that RAD51B (rs8017304 and rs2588809), TRIB1 (rs6987702, rs4351379, and rs4351376), COL8A1 (rs13095226), and COL10A1 (rs1064583) gene variants have never been studied in the Baltic states, namely, in the Lithuanian population, and our study was the first of its type.

Conclusion
Our study revealed a strong association between variant in COL8A1 (rs13095226) and exudative AMD development.

Data Availability
The genotyping data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
None of the authors have any proprietary interests or conflicts of interest related to this submission.