Associations of VEGF Polymorphisms With Retinopathy of Prematurity

Abstract

Purpose: This study investigated the associations between vascular endothelial growth factor (VEGF) polymorphisms and retinopathy of prematurity (ROP) risk.

Methods: Infants born prematurely at any time from 2009 to 2018 were included. Five single-nucleotide polymorphisms (SNPs) of VEGF were analyzed using real-time PCR in all infants. Multivariate logistic regression was applied to model the associations between VEGF polymorphisms and ROP susceptibility, severity, and premature clinicopathologic characteristics.

Results: A total of 334 patients were included and categorized into three groups: those without ROP, those with mild ROP (i.e., ROP not requiring treatment), and those with severe ROP (i.e., ROP for whom treatment was indicated). Among the female patients with ROP, those with VEGF rs3025035 CT (3.231-fold; 95% confidence interval [CI], 1.238–8.431) and a combination of CT and TT genotypes (2.643-fold; 95% CI, 1.056–6.619) exhibited significantly higher risks of severe ROP compared with those with wild-type genotypes. Female ROP infants with VEGF rs3025010 C (TC + CC) alleles had a lower risk of ROP stage ≥3 (odds ratio [OR] = 0.406; 95% CI, 0.165–0.999) than those with TT homozygotes. ROP patients with the VEGF rs10434 A allele (GA + AA) exhibited higher risks of necrotizing enterocolitis (OR = 2.750; 95% CI, 1.119–6.759) and lower risk of bronchopulmonary dysplasia (OR = 0.390; 95% CI, 0.173–0.877) than those with GG homozygotes did.

Conclusions: VEGF polymorphisms affect ROP risks differently in male and female infants. In female infants, VEGF rs3025035 with T alleles may predict ROP severity, and VEGF rs3025010 with C alleles may protect against severe ROP.

Retinopathy of prematurity (ROP) is a neurovascular complication in preterm infants that leads to severe visual impairment or blindness.¹ Vascular endothelial growth factor (VEGF) is a key regulator of angiogenesis in fetal life and plays a key role in ROP pathogenesis. Among the members of the VEGF family, VEGFA with splicing variants is considered the key angiogenic factor.² In the first phase of ROP, VEGF expression is downregulated by hyperoxia in conjunction with cessation of normal growth and a loss of developing vasculature.³ This causes inadequate perfusion to the developing retina, leading to vasoproliferation; this is the proliferative phase of ROP in which increased VEGF expression plays an essential role.^4,5 Eventually, dysregulated retinal neovascularization leads to fibrovascular tissue formation and even retinal detachment, known as advanced ROP.⁶ 

Genetic polymorphism is a variation in the DNA sequence among individuals, groups, or populations. Sources include single-nucleotide polymorphisms (SNPs), sequence repeats, insertions, deletions, and recombination. The human VEGF gene (OMIM 192240) is located on chromosome 6p12.⁷ Relatively rare SNPs in the 5ʹ- or 3ʹ-untranslated region (UTR) were reported to be associated with ROP.⁸ ROP incidence has been associated with the presence of various VEGF polymorphisms: −460 T/C (rs833061), −634G/C (rs2010963), and +936 C/T (rs3025039).^8,9 However, these studies have mostly focused on Western populations. A study reported that the TNF −308G/A (rs1800629) polymorphism increased the risk of ROP in Chinese infants.¹⁰ 

Approximately 15% of preterm infants with mild ROP eventually experience progression to severe and potentially blinding stages.¹ Identifying genotypes associated with a higher risk of disease progression can aid in the treatment selection for infants who may benefit from close screening or prompt management. In the present study, we investigated the association between VEGF polymorphisms and ROP progression risk, as well as their modification effect on different preterm characteristics in infants with ROP. 

Methods

Study Population

The study was performed in the Linkou and Taipei branches of Chang Gung Memorial Hospital, Taiwan. All infants born prematurely between 2009 and 2018 whose parents consented to their participation in the study were considered for inclusion. All infants with a gestational age (GA) < 30 weeks, birth weight (BW) < 1500 g, or conditions deemed to be at risk of ROP (such as oxygen supplementation) were screened for ROP by at least two experienced ophthalmologists through dilated fundus examination. During the examination, a digital wide-angle retinal imaging device (RetCam III; Clarity Medical Systems, Pleasanton, CA, USA) was used for documentation. We excluded patients whose parents did not provide consent, whose medical records were not complete, who had other eye diseases, and whose follow-up period was <6 months. We divided the included patients into those without and with ROP (no-ROP and ROP groups, respectively). Additionally, ROP was categorized as mild (type 2 and milder than type 2) or severe (type 1) in accordance with the specifications of the Early Treatment for Retinopathy of Prematurity cooperative group trials.¹¹ 

Type 1 ROP, defined as ROP for which treatment was indicated, was subdivided as follows: any stage ROP with plus disease (i.e., a degree of dilation and tortuosity of the posterior retinal blood vessels meeting or exceeding that of a standard photograph) in zone I, stage 3 ROP without plus disease in zone I, or stage 2 or 3 ROP with plus disease in zone II.¹¹ Type 2 ROP was defined as ROP for which treatment was not indicated but required close clinical monitoring, and it was subdivided as follows: stage 1 or 2 ROP without plus disease in zone I and stage 3 ROP without plus disease in zone II.¹¹ Anything milder than type 2 ROP was considered to indicate disease severity that did not reach that of type 2, such as stage 1 ROP in zone III or stage 2 ROP in zone II. The terms zone and stage conform to those described in the International Classification of Retinopathy of Prematurity.¹² 

Selection of VEGF Polymorphisms

More than 60 SNPs in the intron or downstream of the VEGFA region have been documented in the Single Nucleotide Polymorphism Database (dbSNP). Overall, the SNP selection for association analysis in the candidate genes was conducted with adherence to the following principles. (1) Haplotype-tagging SNPs were derived from the analyses of HapMap data and provided a minimal set of markers that guaranteed that at least one marker was in strong linkage disequilibrium with an unmeasured marker according to our estimates of pairwise linkage disequilibrium measures. (2) The most common (P > 0.05) polymorphisms from the coding region (cSNPs) were included in genotyping. On the basis of these principles and with consideration of SNPs reported in studies on VEGF pathophysiology in ocular diseases,^13–15 we included the following SNPs of VEGFA: rs2146323, rs3025010, rs3025035, rs3025040, and rs10434. 

Genomic Data Extraction

Extraction of genomic DNA from the patients was performed using a DNA collection kit (Oragene-DNA; DNA-Genotek, Ottawa, ON, Canada) on 3 mL of blood samples in accordance with the manufacturer's instructions. The DNA samples were dissolved in Tris-EDTA buffer (10-mM Tris at pH 7.8 and 1-mM EDTA) and then quantified by measuring the optical density at 260 nm. The final preparation was stored at −20°C and was used to create templates for PCR. 

Real-Time PCR

Allelic discrimination of the five aforementioned VEGFA polymorphisms was performed using the StepOne Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) and analyzed using SDS 3.0 software (Applied Biosystems) with a TaqMan assay (Applied Biosystems). Each reaction had a final volume of 5 µL, which contained 2.5 µL of TaqMan Genotyping Master Mix, 0.125 µL of TaqMan probe mix, and 10 ng of genomic DNA. An initial denaturation step was performed at 95°C for 10 minutes, followed by 40 cycles of 95°C for 15 seconds and 60°C for 1 minute. 

Statistical Analyses

The Mann–Whitney U test and Fisher’s exact test were used to compare the demographic characteristic distributions between the ROP and non-ROP groups. Adjusted odds ratios (ORs) and 95% confidence intervals (CIs) of the association of genotype frequencies with risks and clinicopathological characteristics were estimated using multiple logistic regression by adjusting for GA and BW. All data were analyzed using SAS 9.4 (SAS Institute, Cary, NC, USA). Between-group differences were considered significant at P < 0.05. 

Results

Clinical Characterization of the Cohort of Premature Infants

The clinicodemographic characteristics of the included premature infants are presented in Table 1. We recruited 334 infants: 156 without ROP (the control group) and 178 with ROP (the case group). Within the ROP group, 85 infants (47.75%) were not indicated for treatment. Significant differences in GA and BW were noted between the non-ROP and ROP groups (both P < 0.001). In subsequent multiple logistic regression models used to determine the risk profiles associated with genetic polymorphisms, GA and BW were listed as confounders for each analysis for adjusted ORs and 95% CIs. 

Table 1.

 

Demographic and Clinical Characteristics of Enrolled ROP and Non-ROP Infants

 

Genotype–Phenotype Correlation Analysis

In our cohort overall, the alleles with the highest distribution frequencies for the VEGF polymorphisms rs10434, rs2146323, rs3025010, rs3025035, and rs3025040 were homozygous for G/G, C/C, T/T, C/C, and C/C, respectively. Supplementary Table S1 shows the association of VEGF SNPs with the GA and BW in patients with ROP. No significant differences in the ORs of ROP incidence were noted in the individuals with VEGF rs10434, rs2146323, rs3025010, rs3025035, or rs3025040 compared with those with wild-type genotypes (Table 2). In addition, no significant difference was observed in the ROP severity risks for individuals with these polymorphisms (Table 3). 

Table 2.

 

Distribution Frequency of VEGF Genotypes of Infants with ROP and Non-ROP

 

Table 3.

 

Distribution Frequency of VEGF Genotypes of Infants With Mild, Severe, or Non-ROP

 

We divided all infants into male and female groups for separate risk analyses. The risk of ROP severity in male infants was not significantly different when comparing VEGF polymorphisms with wild types. For the female infants with ROP, significantly higher risks of developing severe ROP were observed in the individuals with the VEGF rs3025035 CT (3.231-fold; 95% CI, 1.238–8.431) and a combination of CT and TT genotypes (2.643-fold; 95% CI, 1.056–6.619) than in the individuals with wild-type genotypes (Table 4). 

Table 4.

 

Distribution Frequency of VEGF Genotypes of Mild, Severe, and Non-ROP Infant Females

 

Subsequently, we investigated the implications of VEGF polymorphisms for each ROP clinicopathologic characteristic. For rs3025010, the female infants with ROP with the TC/CC genotype had a significantly lower incidence or risk of developing ROP stage ≥3 compared with those with the wild-type genotypes (OR = 0.406; 95% CI, 0.165–0.999) (Table 5). 

Table 5.

 

Characteristics of Infants With ROP, Stratified by Polymorphic Genotypes of VEGF rs3025010

 

We then investigated potential associations between the prevalence of other premature conditions in the patients with ROP and the presence of VEGFA polymorphisms. No wild-type or VEGF polymorphism other than rs10434 was significantly associated with any systemic pathologic conditions in the patients with ROP (data not shown). By contrast, significant differences in the prevalence of pulmonary hypertension, atrial septal defect, patent ductus arteriosus, intraventricular hemorrhage, pneumonia, sepsis, and surfactant use were observed in the infants with ROP with VEGF rs10434. Furthermore, the patients with ROP with VEGF rs10434 A (GA + AA) had a higher prevalence of necrotizing enterocolitis (OR = 2.750; 95% CI, 1.119–6.759) and “lower” prevalence of bronchopulmonary dysplasia (BPD; OR = 0.390; 95% CI, 0.173–0.877) compared with those with wild-type genotypes (Table 6). 

Table 6.

 

Characteristics of Infants with ROP, Stratified by Polymorphic Genotypes of VEGF rs10434

 

Discussion

In this study, we demonstrated potential associations between different VEGF polymorphisms and ROP risks, severity, sex differences, and other premature pathological conditions in an East Asian cohort. Our major finding was that the associations of ROP risks with VEGF polymorphisms differed between male and female infants. Specifically, the presence of VEGF rs3025035 with T alleles may indicate greater ROP severity in female infants. In addition, female infants with VEGF rs3025010 with at least one C allele exhibited lower odds of ROP stage ≥3 than those with TT homozygotes did (OR = 0.406; 95% CI, 0.165–0.999). 

VEGF is a crucial growth factor for vascular development, regulation, and angiogenesis. Its roles include stimulating vascular endothelial cell proliferation and increasing vascular permeability.¹⁶ Although most functions of the SNPs analyzed in our study have not been definitively characterized, the chromosomal interactions between the SNP regions and the VEGF promoters may affect VEGFA activity.¹⁷ Polymorphisms in the promoter, introns, exons, and 3ʹ- and 5ʹ-UTRs in VEGF may affect the synthesis or function of the corresponding protein.¹⁸ In addition, messenger RNA stability may be modified by polymorphisms, resulting in differential expression of VEGF.¹⁷ Understanding the possible mechanisms that cause differential expression and effects of VEGF polymorphisms can help to establish a clearer understanding of the relationship between specific polymorphisms and their corresponding clinical phenotypes or outcomes. 

Other studies have demonstrated associations between certain VEGF polymorphisms and ROP development. Vannay et al.⁹ reported that the VEGF +405 C allele (rs2010963) and VEGF −460 T/T (rs833061), as well as+405 C/C haplotypes, may be associated with a risk of proliferative ROP. Cooke et al.⁸ indicated that the VEGF −634 G/G genotype was associated with an increased risk of threshold ROP. A meta-analysis of 27 case–control studies with 5748 patients with ROP and 6146 controls indicated an association between VEGF −460 polymorphism and increased ROP risks.¹⁹ These associations with ROP risks and severity may be due to increased VEGF production^8,20,21 in VEGF polymorphisms such as −460 or +405, as indicated by in vitro studies.²¹ 

VEGF rs2146323 with A alleles has been reported to be associated with the risk of type 2 diabetic retinopathy^13,22 and serves as a potential effect modifier in anti-VEGF therapy.¹⁵ Located in the intronic region of VEGF, rs3025035 with T alleles was investigated in patients with proliferative diabetic retinopathy²³ and exudative macular degeneration, and the study revealed a better response to anti-VEGF treatment in the latter group.²⁴ Our study determined that VEGF rs3025010 with C alleles lowers the risk of ROP stage ≥3 development. The polymorphisms above have also been proposed as potential biomarkers for anti-VEGF treatment,¹⁵ indicating their potential roles in VEGF-related pathophysiology. Thus, therefore, it is plausible that these polymorphisms may affect ROP risks through the VEGF expression pathway. 

To the best of our knowledge, the current study is the first to report an association between VEGF rs3025035 and an elevated risk of severe ROP in female infants. Although not commonly reported in the ROP literature, multiple studies have observed differential expression of the VEGF genetic polymorphisms between male and female infants.^25,26 This variation has been observed in endothelial progenitor differentiation, and genetic regulatory differences have been identified between the sexes in almost all tissues, with differences particularly noted in brain tissues.^25,26 In bovine retinal vascular endothelial cells, 17β-estradiol increases VEGF expression under normoxic conditions, indicating female sex hormone–mediated modulation of angiogenesis.^27,28 However, under hypoxic conditions, hormones may reduce VEGF gene expression through hypoxia-inducible factor-1α.²⁸ A nationwide population study by Lin et al.²⁹ revealed that the female sex is one of several factors associated with increased ROP risks, along with low BW and oxygen use. The complex interplay between sex and VEGF expression under different oxygen conditions may explain why female premature infants are more susceptible to ROP caused by VEGF polymorphisms. 

Notably, the risk factors for ROP are multifactorial, and polymorphisms might lead to complex interactions among such factors as BW, GA, oxygen usage conditions, or other premature comorbidities. For example, several studies have demonstrated that low BW is a risk factor for vision-threatening ROP when it is correlated with early GA.^29,30 We observed significant differences in the GA and BW between the ROP and non-ROP groups in our study that reflect real-world clinical situations. To improve the robustness of our data, we performed multivariate regression analyses with adjustments for GA and BW in each risk analysis. 

In the literature, rs10434 in 3ʹ-UTR has been reported to be associated with various conditions, such as recurrent pregnancy loss,³¹ renal cell carcinoma,³² and gastric cancer.³³ Altered VEGF expression due to polymorphisms has been associated with spontaneous abortion or preeclampsia.^34–36 In healthy pregnancy, closely coordinated biopathways ensure that the placenta and fetus can receive adequate blood supply. Disruption in the VEGF protein function due to specific variants might result in defective angiogenesis. The presence of VEGF polymorphisms in mothers during pregnancy, such as 936 C/T, was observed to increase the risk of premature birth (GA < 37 weeks).^37,38 This may be due to the inability of normal placental tissues to upregulate VEGF under stress conditions, such as revascularization or placental development, leading to high apoptotic tendency and early birth.³⁷ 

Our study indicated that the rs10434 G>A polymorphism was associated with greater risks of necrotizing enterocolitis but lower risks of BPD in infants with ROP. Consistently, a study including Japanese infants born prematurely demonstrated that G alleles in rs2010963 were correlated with increased risks of BPD, although no association between rs10434 with BPD was reported.³⁹ Low BW is a risk factor for vision-threatening ROP when correlated with early GA.³⁰ In our study, no significant difference was observed in terms of BW among different SNPs in infants with ROP. 

Our study has some limitations. First, the sample size was relatively small for a gene–environment assessment study. This limited the conclusions that could be drawn from our results and the generalizability of our findings because they may not accurately reflect the characteristics of the population of interest. A larger sample size would improve the reproducibility of this study. Nevertheless, a balanced sample size in a genetic study can still yield reliable results.⁴⁰ Second, an association but not a direct effect on gene function was identified between the sequence polymorphisms and ROP risk characteristics of premature infants. This indicates that these variants may also be genetically linked to other loci related to ROP development. A strength of our study is its focus on an East Asian (i.e., Han Chinese) population, which has rarely been investigated with respect to associations between genetic polymorphisms and ROP risks.^9,10 

The determination of the effect of VEGF polymorphisms on ROP is significant because it allows for more accurate outcome predictions. By incorporating personalized risk factors, such as gene polymorphisms with population adjustment, in addition to conventional factors associated with ROP severity and risk, such as GA, BW, and oxygen use, both prediction and management of ROP can be improved. Future studies should include these demographic factors in risk analysis models or clinical algorithms for more effective ROP management.⁴¹ 

Conclusion

Our results indicate that the impact of VEGF polymorphisms on ROP risks varies between male and female infants. In female infants, the presence of VEGF rs3025035 T alleles may predict a more severe ROP. Additionally, VEGF rs3025010 with C alleles may exert a protective effect against the development of stage ≥3 ROP in female infants. In all infants with ROP, VEGF rs10434 was discovered to potentially increase the risk of necrotizing enterocolitis but decrease the risk of BPD. Further studies of these polymorphisms may elucidate the pathways involved in ROP development in premature infants. 

Acknowledgments

The authors thank Wallace Academic Editing for their contribution to the preparation of this manuscript. 

Supported by research grants from Chang Gung Memorial Hospital (CMRPG3M0131-2 and CMRPG3L0151-3) and Ministry of Science and Technology (MOST 109-2314-B-182A-019-MY3). 

Disclosure: X.C. Ling, None; E.Y.-C. Kang, None; K.-J. Chen, None; N.-K. Wang, None; L. Liu, None; Y.-P. Chen, None; Y.-S. Hwang, None; C.-C. Lai, None; S.-F. Yang, None; W.-C. Wu, None