Relative Leukocyte Telomere Length and Genetic Variants in Telomere-Related Genes and Serum Levels Role in Age-Related Macular Degeneration

Telomere shortening is well known to be associated with ageing. Age is the most decisive risk factor for age-related macular degeneration (AMD) development. The older the individual, the higher the AMD risk. For this reason, we aimed to find any associations between telomere length, distribution of genetic variants in telomere-related genes (TERT, TERT-CLPTM1, TRF1, TRF2, and TNKS2), and serum TERF-1 and TERF2 levels on AMD development. Methods: Our study enrolled 342 patients with AMD and 177 healthy controls. Samples of DNA from peripheral blood leukocytes were extracted by DNA salting-out method. The genotyping of TERT rs2736098, rs401681 in TERT-CLPTM1 locus, TRF1 rs1545827, rs10107605, TNKS2 rs10509637, rs10509639, and TRF2 rs251796 and relative leukocyte telomere length (T/S) measurement were carried out using the real-time polymerase chain reaction method. Serum TERF-1 and TERF2 levels were measured by enzymatic immunoassay (ELISA). Results: We found longer telomeres in early AMD patients compared to the control group. Additionally, we revealed that minor allele C at TRF1 rs10107605 was associated with decreases the odds of both early and exudative AMD. Each minor allele G at TRF2 rs251796 and TRF1 rs1545827 C/T genotype and C/T+T/T genotypes, compared to the C/C genotype, increases the odds of having shorter telomeres. Furthermore, we found elevated TERF1 serum levels in the early AMD group compared to the control group. Conclusions: In conclusion, these results suggest that relative leukocyte telomere length and genetic variants of TRF1 and TRF2 play a role in AMD development. Additionally, TERF1 is likely to be associated with early AMD.


Relative Leukocyte Telomere Length
Relative leukocyte telomere length (T/S) was determined for all study subjects and compared between study groups. We found longer telomeres in early AMD patients compared to the control group I (T/S (median (IQR  Relative leukocyte telomere length in early AMD patients and the control group I. Relative leukocyte telomere length (T/S) in early AMD patients versus healthy controls are presented as boxand-whisker plots with the median and IQR. Mann-Whitney U test was used to assess T/S differences between patients with early AMD and control groups; p < 0.001. Relative leukocyte telomere length in early AMD patients and the control group I. Relative leukocyte telomere length (T/S) in early AMD patients versus healthy controls are presented as box-and-whisker plots with the median and IQR. Mann-Whitney U test was used to assess T/S differences between patients with early AMD and control groups; p < 0.001. Relative leukocyte telomere length (T/S) in exudative AMD patients versus healthy controls are presented as box-and-whisker plots with the median and IQR. Mann-Whitney U test assessed T/S differences between patients with exudative AMD and control groups; p = 0.842.

Figure 2.
Relative leukocyte telomere length in exudative AMD patients and the control group II. Relative leukocyte telomere length (T/S) in exudative AMD patients versus healthy controls are presented as box-and-whisker plots with the median and IQR. Mann-Whitney U test assessed T/S differences between patients with exudative AMD and control groups; p = 0.842.

Figure 2.
Relative leukocyte telomere length in exudative AMD patients and the control group II. Relative leukocyte telomere length (T/S) in exudative AMD patients versus healthy controls are presented as box-and-whisker plots with the median and IQR. Mann-Whitney U test assessed T/S differences between patients with exudative AMD and control groups; p = 0.842.

Figure 3.
Relative leukocyte telomere length in early and exudative AMD patients. Relative leukocyte telomere length (T/S) in early AMD patients versus exudative AMD patients are presented as box-and-whisker plots with the median and IQR. Mann-Whitney U test was used to assess the differences in T/S between patients with early and exudative AMD; p = 0.064. 3.3. The Genotyping of TERT rs2736098, rs401681, TRF1 rs1545827, rs10107605, TNKS2 rs10509637, rs10509639 and TRF2 rs251796 Hardy-Weinberg equilibrium (HWE) was evaluated for all SNPs in both control group I and control group II. Only one SNP TRF1 rs10107605 did not follow the HWE, which might be caused by low sample size in study groups.
The genotype and allele frequencies of TERT rs2736098, rs401681, TRF1 rs1545827, rs10107605, TNKS2 rs1050963, rs10509639, and TRF2 rs251796 in the early, exudative AMD and control groups are shown in Table 2. Statistical analysis was performed to compare the genotype and allele frequencies between the early AMD group and the control group I as well as between the exudative AMD group and the control group II. We found that TRF1 rs10107605 minor allele C was statistically significantly less frequent in early AMD patients than in control group I subjects: 8.5% vs. 14.1%, respectively, p = 0.018 (Table 2). Additionally, we revealed that TRF1 rs10107605 genotypes (A/A, A/C, and C/C) distribution differed statistically significantly between exudative AMD and control group II subjects: 87.1%, 12.9%, and 0% vs. 78.8%, 14.4% and 6.8%, respectively, p = 0.004, and the minor allele C was less frequent in exudative AMD patients than in control group II subjects: 6.5% vs. 14.0%, respectively, p = 0.003 (Table 3).    Binomial logistic regression was performed to evaluate the impact of TERT rs2736098, rs401681, TRF1 rs1545827, rs10107605, TNKS2 rs1050963, rs10509639 and TRF2 rs251796 on early and exudative AMD development. Our results revealed that genotype CC at rs10107605 was associated with about 75% decreased odds of early AMD development under the codominant and recessive models (OR = 0.251; 95% CI: 1.333-3.870; p = 0.037 and OR = 0.260; 95% CI: 1.483-3.970; p = 0.041, respectively). Additionally, the analysis showed that minor allele C was associated with decreased odds of early and exudative AMD as well (OR = 0.632; 95% CI: 1.333-3.870; p = 0.038 and OR = 0.490; 95% CI: 1.483-3.970; p = 0.010, respectively) ( Table 4). The leukocyte telomeres were divided into short and long telomeres by all study subjects' median telomere length. Statistical analysis was performed to compare the genotype and allele frequencies between the two groups. Statistically significant differences were found only by comparing genotype and allele frequencies of TRF2 rs251796 between the long and short telomere groups (p = 0.043 and p = 0.011, respectively) ( Table 5). The binomial logistic regression analysis revealed that G/G genotype compared to A/A genotype was associated with 2-fold increased odds of having short telomeres (OR = 2.039; 95% CI: 1.050-3.961; p = 0.035); A/G+G/G genotypes carriers compared to A/A genotype had 1.5-fold increased odds of having short telomeres (OR = 1.499; 95% CI: 1.056-2.128; p = 0.023), and overall each allele G at TRF2 rs251796 was associated with 1.4-fold increased odds of having short telomeres (OR = 1.418; 95% CI: 1.078-1.866; p = 0.013) ( Table 6). Additionally, we determined that TRF1 rs1545827 C/T genotype and C/T+T/T genotypes compared to C/C genotype carriers were associated with 1.6 and 1.5-fold increased odds of having short telomeres under the codominant and dominant models (OR = 1.555; 95% CI: 1.055-2.293; p = 0.026 and OR = 1.518; 95% CI: 1.054-2.186; p = 0.025, respectively) ( Table 6).    TERF1 and TERF2 serum levels were measured in duplicates for 20 early AMD patients, 20 exudative AMD patients, and 20 control subjects. Analysis showed elevated TERF1 serum levels in the early AMD group compared to control subjects (median (IQR): 0.850 (1.025) ng/mL vs. 0.546 (0.526) ng/mL, p = 0.004) (Figure 4). However, there were no statistically significant differences between exudative AMD and control groups (median (IQR): 0.493 (0.459) ng/mL vs. 0.546 (0.526) ng/mL, p = 0.607) ( Figure 5). Additionally, no statistically significantly results were found analysing TERF2 serum levels between early AMD vs. control group, and exudative AMD vs. control group (median (IQR): 4.476 (2.200) ng/mL vs. 3.743 (4.235) ng/mL, p = 0.160 ( Figure 6); 3.911 (2.462) ng/mL vs. 3.743 (4.235) ng/mL, p = 0.829 (Figure 7), respectively). 0.850 (1.025) ng/mL vs. 0.546 (0.526) ng/mL, p = 0.004) (Figure 4). However, there were no statistically significant differences between exudative AMD and control groups (median (IQR): 0.493 (0.459) ng/mL vs. 0.546 (0.526) ng/mL, p = 0.607) ( Figure 5). Additionally, no statistically significantly results were found analysing TERF2 serum levels between early AMD vs. control group, and exudative AMD vs. control group (median (IQR): 4.476 (2.200) ng/mL vs. 3.743 (4.235) ng/mL, p = 0.160 ( Figure 6); 3.911 (2.462) ng/mL vs. 3.743 (4.235) ng/mL, p = 0.829 (Figure 7), respectively). Figure 4. Serum levels of TERF1 in patients with early AMD and control group. AMD: age-related macular degeneration; Serum protein values are presented as median and IQR. Mann-Whitney U test was used to assess serum TERF1 levels differences between patients with early AMD and control groups; p = 0.004.  Mann-Whitney U test was used to assess serum TERF1 levels differences between patients with exudative AMD and control groups; p = 0.607. Figure 5. Serum levels of TERF1 in patients with exudative AMD and control group. AMD: agerelated macular degeneration; Serum protein values are presented as median and IQR. Mann-Whitney U test was used to assess serum TERF1 levels differences between patients with exudative AMD and control groups; p = 0.607. Figure 6. Serum levels of TERF2 in patients with early AMD and control group. AMD: age-related macular degeneration; Serum protein values are presented as median and IQR. Mann-Whitney U test was used to assess serum TERF2 levels differences between patients with early AMD and control groups; p = 0.160. Figure 6. Serum levels of TERF2 in patients with early AMD and control group. AMD: age-related macular degeneration; Serum protein values are presented as median and IQR. Mann-Whitney U test was used to assess serum TERF2 levels differences between patients with early AMD and control groups; p = 0.160.

Discussion
The discovery and explanation of telomere complex role in biology warranted the 2009 Nobel Prize in medicine [38]. Understanding basic biological mechanisms and the emerging impact of telomerase and telomere biology in medicine provides a unique opportunity to study age-related diseases such as age-related macular degeneration.
Many studies are analyzing TL association with ageing, and results are bewildering and wondrous. Some studies have proved that LTL declines by 0.003 ln(T/S) per year on Figure 7. Serum levels of TERF2 in patients with exudative AMD and control group. AMD: agerelated macular degeneration; Serum protein values are presented as median and IQR. Mann-Whitney U test was used to assess serum TERF2 levels differences between patients with exudative AMD and control groups; p = 0.004.

Discussion
The discovery and explanation of telomere complex role in biology warranted the 2009 Nobel Prize in medicine [38]. Understanding basic biological mechanisms and the emerging impact of telomerase and telomere biology in medicine provides a unique opportunity to study age-related diseases such as age-related macular degeneration.
Many studies are analyzing TL association with ageing, and results are bewildering and wondrous. Some studies have proved that LTL declines by 0.003 ln(T/S) per year on average from the age of 50 to 90 [33,39], while other studies have stated that there is no significant association between the shortened LTL and ageing [40,41]. Muezzinler and colleagues have determined that a reduction in LTL is approximately 0.5/(10 years) in cohorts of various age groups [42]. For example, in a Japanese study, Arai and colleagues have concluded that inflammation is a crucial malleable driver of ageing up to extreme old age in humans. Still, telomere length is not a predictor of successful ageing in centenarians and semi-supercentenarians [43]. Another study has stated that individuals with shorter telomeres are characterized by a higher mortality rate, nearly twice as high as those with longer telomeres [44]. A study conducted by Mons et al. [45] analyzed more than 12,000 subjects of two population-based studies (ESTHER and Nurses' Health Study) and identified that subjects with shorter telomeres (1st quintile) have a higher hazard ratio for all-cause mortality (1.66, 95% CI 1.09-2.53, p = 0.018) compared to those with longer telomeres (5th quintile), in agreement with the study conducted by Goglin et al. [44].
The first study (2015) analyzing the exudative AMD association with LTL was conducted on the Han Chinese population [33]. The researchers analyzed 197 AMD cases (both exudative and atrophic: 76 GA cases (38.57%), 52 CNV cases (26.40%), and 69 advanced AMD cases lacking further subtype information (35.03%)) and 259 healthy controls. Scientists revealed a strong association between AMD and LTL (OR = 2.24; 95% CI = 1.68-3.07; p = 0.0001) after adjustment for age and sex. Furthermore, their results showed a significant association between the GA subtype and the LTL (OR = 4.81; 95% CI = 3.15-7.82; p = 0.0001), also after adjustment for age and sex. They proved that LTL plays a role in AMD's pathological mechanisms, mainly in the GA subtype but not in the CNV [33]. Meanwhile, another study by Immonen et al. [39] did not reveal any statistically significant results comparing the mean (SD) telomere length in AMD patients (0.68) and the control group (0.69) (p = 0.485). The corresponding proportions of telomeres <5 • kb were 10.60 (2.76) and 10.05 (2.64) (p = 0.197). In this study, a hundred (82.6%) patients had neovascular AMD in the worse eye, seventeen (14%) large drusen, and four (3.3%) central geographic atrophy. There were no differences in the telomere length between patients with drusen or exudative AMD [39]. Two studies (Immonen et al. [39] and Weng, X. et al. [33]) have presented different results. Still, the difference between the studies could be explained by the different proportions of AMD subtypes among the studied AMD patients. In Immonen et al.'s study, AMD patients had mostly neovascular AMD, and only 3% had geographic atrophy AMD. In the other study, Weng, X. et al. evaluated 76 GA cases (38.57%) and 52 CNV cases (26.40%), comparing them with 259 healthy individuals separately. Differences in telomere length could be due to the phenotypic difference between GA and CNV.
It has been thought that the leukocyte telomere length may reflect the systemic telomerase capacity of an individual. Alternatively, shortened leukocyte telomeres may be associated with the chronic activation of the immune system beyond the reparative capacity of telomerase [53]. Such chronic systemic inflammation has been reported in the pathogenesis of cardiovascular disease and AMD [54].
Drigeard Desgarnier et al. found telomere length differences in different human eye structures [55]. Moreover, Bell et al. observed a unique telomere DNA expansion phenotype in the rod cells but not in other retinal cells [56].
Analysing the AMD pathogenesis, scientists suggested that the senescence of the RPE cells might play a role in AMD development [57] through several pathways, including oxidative stress response [58]. Oxidative stress damages telomeres due to their guanine-rich DNA structure. While it can be more challenging to repair, in some cases, the telomerase may extend oxidative stress-shortened telomeres, preventing further RPE cell degeneration following AMD progression [59].
In the experimental model, telomere shortening inhibited neovascularization [60]. It is possible that telomere shortening might have a role in the pathogenesis of geographic atrophy. Unfortunately, the number of patients with geographic atrophy was too small (n = 4 and n = 76 [33,40]) that this hypothesis could not be approved; besides, it was drawn in different populations (Chinese and European (Finland)). It is also stated that telomere length does not correlate with mortality and morbidity in the very old [61].
However, normal age-related shortening of telomeres may still be one of the ageing changes that make conditions favourable for the action of specific pathogenic factors of AMD development.
Based on Mendelian randomization approaches, a score built from SNPs associated with LTL was suggested as a critical risk marker ("teloscore" which explains 2.2% of the telomere variability) [62]. Moreover, genome-wide association and candidate gene studies have shown that SNPs in the TERT gene and telosome complex genes were associated mostly with cancer risk. While it has been hypothesized that polymorphisms in the TERT gene might be related to cancer via their effects on the expression of TERT, others have revealed the associations between TERT polymorphisms and increased TERT transcription activity [63,64]. Controversial results have been published, showing differences among the types of cancer [65,66]. Furthermore, a meta-analysis has revealed an association between the allele A at TERT rs2736098 G > A (located on 5p15.33) and cancer development. Besides, the ethnicity-specific effect has also been found while analyzing different subgroups [67]. Another meta-analysis has shown that allele C at TERT-CLPTM1L rs401681 (located on 5p15.33) was a low-penetrance risk allele for the development of lung, bladder, prostate cancers, and basal cell carcinoma but also a potential protective allele for melanoma and pancreatic cancer [68]. Additionally, the TRF2 rs251796 was significantly associated with lung cancer as well [69]. No other experiments were performed in eye disease studies with SNPs, which could be compared to our results. Our study revealed that the TRF1 rs10107605 was associated with decreased odds of early and exudative AMD development, while the TRF2 rs251796 and TRF1 rs1545827 variants were linked to shorter telomeres.
No other study has investigated the link between AMD and TRF1/TRF2 expression. However, only one study has been conducted on the association between TRF1/TRF2 and TL-shortening disease [52]. Wu et al. reported that the expression of TRF1 was elevated in Alzheimer's patients than in the control group, while the TRF2 expression was significantly lower than in control subjects. This investigation showed that TRF1 and TRF2 could be associated with TL-shortening and may be closely connected with Alzheimer's disease progression [70]. Several authors examined TRF1/TRF2 expression in many types of cancers. In the study by Shi et al., the link between the expression of TRF1 protein and human leukaemia was analyzed. The study showed that TRF1 protein concentration was lower in leukaemia patients than in control subjects [71]. However, Lin and other authors reported that the expression of TRF1 was significantly lower in lung cancer tissues than in normal tissues; no significant differences were found between TRF2 [72]. Our study revealed that the expression of TRF1 in serum was elevated in the early AMD group compared to control subjects. While our study revealed significant results, it is still important to highlight that further studies with larger sample size are necessary to confirm such associations. Moreover, other AMD risk factors should be included into future studies.

Conclusions
Our study revealed longer telomeres in early AMD patients compared to the control group (T/S (median (IQR): 1.207 (1.319) vs. 0.778 (1.057), respectively, p < 0.001). Addition-ally, TRF2 rs251796 and TRF1 rs1545827 variants were linked to shorter telomeres, while TRF1 rs10107605 was associated with decreased odds of early and exudative AMD development. We also found elevated TERF1 serum levels in the early AMD group compared to the control group (p = 0.004).