Genetic Contributions to the Development of Complications in Preterm Newborns

Aim We aimed to identify specific polymorphisms of genes encoding for vascular endothelial growth factor A (VEGFA), endothelial nitric oxide synthase (eNOS), renin-angiotensin system (angiotensinogen gene [AGT], angiotensinogen type 1 receptor [AGTR1], angiotensin-converting enzyme [ACE]), and heme oxygenase-1 (HMOX-1) in a cohort of preterm infants and correlate their presence with the development of respiratory distress syndrome (RDS) requiring mechanical ventilation (MV), bronchopulmonary dysplasia (BPD), intraventricular hemorrhage (IVH) and retinopathy of prematurity (ROP). Study Design We carried out a retrospective study to evaluate the allele frequency and genotype distribution of polymorphisms of VEGFA, eNOS, AGT, AGTR1, ACE, and HMOX-1 in a population of preterm neonates (n=342) with a gestational age ≤28 weeks according to the presence or absence of RDS requiring MV, BPD, IVH, or ROP. Moreover, we evaluated through the haplotype reconstruction analysis whether combinations of the selected polymorphisms are related to the occurrence of RDS, BPD, IVH and ROP. Results In our population 157 infants developed RDS requiring MV, 71 BPD, 70 IVH, and 43 ROP. We found that TC+CC rs2070744 eNOS (41.7 vs. 25.4%, p=0.01) and GT+TT rs1799983 eNOS (51.8 vs. 35.2%, p=0.01) polymorphisms are independent risk factors for BPD. Haplotype reconstruction showed that haplotypes in VEGF and eNOS are significantly associated with different effects on RDS, BPD, IVH, and ROP in our population. Conclusions We found that TC+CC rs2070744 eNOS and GT+TT rs1799983 eNOS polymorphisms are independent predictors of an increased risk of developing BPD. Haplotypes of VEGFA and eNOS may be independent protective or risk markers for prematurity complications.


Study Design
We carried out a retrospective study to evaluate the allele frequency and genotype distribution of polymorphisms of VEGFA, eNOS, AGT, AGTR1, ACE, and HMOX-1 in a population of preterm neonates (n=342) with a gestational age 28 weeks according to the presence or absence of RDS requiring MV, BPD, IVH, or ROP. Moreover, we evaluated through the haplotype reconstruction analysis whether combinations of the selected polymorphisms are related to the occurrence of RDS, BPD, IVH and ROP.

Results
In our population 157 infants developed RDS requiring MV, 71 BPD, 70 IVH, and 43 ROP. We found that TC+CC rs2070744 eNOS (41.7 vs. 25.4%, p=0.01) and GT+TT rs1799983 eNOS (51.8 vs. 35.2%, p=0.01) polymorphisms are independent risk factors for BPD. Haplotype reconstruction showed that haplotypes in VEGF and eNOS are significantly associated with different effects on RDS, BPD, IVH, and ROP in our population.

Introduction
There is increasing evidence that some genepolymorphismsare implicated in the development of severe complications of preterm birth, including respiratory distress syndrome (RDS), bronchopulmonary dysplasia (BPD), intraventricular hemorrhage (IVH), and retinopathy of prematurity (ROP).
Vascular endothelial growth factor A (VEGFA) is important for the physiological growth of retinal vessels, and experimental and clinical data suggest that VEGF may play a causative role in the development of ROP [1]. Similarly, VEGFA plays a central role in vascular lung repair and its absence results in impaired fetal lung microvascular development [2]. Accordingly, clinical studies suggest a relevant association between VEGFA activity and BPD development in preterm infants [3,4].
Nitric oxide (NO) has several vascular actions which may contribute to the growth and protection of vessels in preterm infants [5]. This seems to be confirmed by the reported association between endothelial NO synthase (eNOS) gene activity and the risk of developing RDS, BPD, and IVH in this population [6].
The renin-angiotensin system (RAS) is important for regulating systemic blood pressure and volume, and its genetic modifications could affect the outcome of preterm infants. Previous studies have reported the possible association between angiotensin-converting enzyme (ACE) and angiotensin type 1 receptor (AGTR1) polymorphisms and the risk of developing BPD [7][8][9][10] and ROP [11][12], but results are still inconclusive.
Heme oxygenase-1 (HMOX-1) is an enzyme degrading heme to iron ions, carbon monoxide, and biliverdin. Products of HO-1 activity perform important physiological functions in the vascular system, which are ultimately linked to the protection of endothelium through cytoprotective, promitogenic, and anti-inflammatory action [13]. The efficacy of this enzyme has been found to be affected by repeat polymorphisms in the promoter of HMOX-1 gene [13]. Although ample clinical data indicate its influence on cardiovascular complications in adult patients, the possible correlation between HMOX-1 gene transcription and the outcome in preterm infants has never been investigated.
On the basis of these considerations, we hypothesized that the occurrence of severe complications in preterm infants-RDS requiring mechanical ventilation (MV), BPD, IVH and ROP -might be affected by polymorphisms in genes coding the VEGFA, eNOS enzyme, RAS system (Angiotensinogen gene [AGT], AGTR1,ACE), and HMOX-1 enzyme. To assess this hypothesis we aimed to genotype specific polymorphisms of these genes in a cohort of preterm infants and correlate their presence to the development of RDS requiring MV, BPD, IVH, and ROP.

Study design
We carried out a retrospective study to evaluate the possible relationship between twelve polymorphisms in genes encoding for VEGFA, eNOS, RAS system, and HMOX-1 and the occurrence of RDS requiring MV, BPD, IVH and ROP in a population of preterm newborns. By using haplotype reconstruction analysis we evaluated whether combinations of the selected polymorphisms are related to the occurrence of the aforementioned prematurity complications.

Patient population
We studied 342 preterm neonates with a gestational age 28 weeks, admitted consecutively to the Neonatal Intensive Care Unit of the Careggi University Hospital of Florence, from January 2004 to December 2012. Exclusion criteria were the diagnosis of major congenital malformations, inherited errors of metabolism, and any other congenital syndrome. All subjects investigated were of Caucasian origin in order to guarantee a homogenous ethnic background.
The DNA analysis was performed on dried blood spots collected from infants at 48 hours of life for the local screening program [14].

Clinical Features and Outcome
Birth weight, gestational age, type of delivery, antenatal steroid treatment, gender, Apgar Score at 5 min, RDS occurrence, need of MV and surfactant, type and duration of respiratory assistance [nasal continuous positive airways pressure (NCPAP), patient-triggered ventilation (PTV),high frequency oscillatory ventilation (HFOV)] were recorded for each newborn infant.
The RDS diagnosis was made as previously reported [15]. Infants were given MV when the pH was <7.20, pO 2 was <50 mmHg with FiO 2 >0.50, and pCO 2 was >65 mmHg.
All clinical data were reported on data sheets designed for this study.

Studied polymorphisms
The criteria for selection of candidate genes in the present study were their potential involvement in the pathogenesis of the aforementioned prematurity complications. The DNA analysis was performed on dried blood spots collected from infants at 48 hours of life for the local screening program using the QIAamp DNA Micro Kit according to the manufacturer's instructions (QiagenGmbh, Hilden, Germany). We studied 12 single nucleotide polymorphisms (SNPs) in the 6 candidate genes (ACE, AGTR1, AGT, eNOS, VEGFA, HMOX-1) according to their demonstrated or putative function based on literature data, localization in the functional regions, as extracted from the dbSNP database (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db = snp&cmd = search&term=) ( Table 1). SNP information was assessed in dbSNP NCBI and ENSEMBL (http://www. ensembl.org/index.html) databases.
Genotyping of ACE gene Insertion/Deletion (I/D) polymorphisms was performed by Taq-Man designed assay, according to Koch et al. [21]. The eNOS 4a4b polymorphism was evaluated by PCR method, as previously described [22].
The (GT)n repeat polymorphism in the promoter of HMOX-1 was detected according to Vazquez-Armenta G et al. [23].

Data analysis
Statistical analysis was performed using the SPSS package v18.0. Continuous variables were expressed as mean and standard deviation, and categorical variables were expressed as frequencies and percentages. Tests for conformity with Hardy-Weinberg equilibrium were performed using a standard χ 2 test. Genotype distributions were compared between groups by χ 2 analysis. We assessed the association between each polymorphism and clinical adverse outcomes (RDS, BPD, IVH, ROP) by using a dominant model of inheritance, which compares individuals with one or two rare alleles (heterozygotes+homozygotes) to a baseline group of homozygous wild type subjects.
The association of the studied polymorphisms with gestational age and birth weight was also evaluated. Comparisons between median values of gestational age and birth weight among genotypes were performed using the non-parametric Mann-Whitney U test. Logistic regression analysis was used to estimate odds ratios (OR) and 95% confidence intervals (CIs). The independent association of polymorphisms with the different clinical endpoints was evaluated by logistic regression analysis adjusted for gestational age and weight at birth. A value of p = 0.05 was chosen as the cut-off level for statistical significance.

Haplotype analysis
In order to perform haplotype reconstruction analysis (to infer the most probable set of polymorphisms of each studied gene that are inherited together from one parent) in our population, data files were processed in R environment (http://www.r-project.org). Haplotype HMOX122q13.1 rs3074372 GT repeat 5' near gene 0 reconstruction and frequency estimation were independently performed using the PHASE v2.1 software [24], and R package haplo.stats by expectation-maximization strategy (EM algorithm). The haplo.stats package was also used to identify statistically significant associations between haplotypes and clinical endpoints by means of generalized linear models: positive or negative coefficient values indicate the risk or the protection conferred by the haplotypes, respectively, and the absolute values of the coefficient indicate the magnitude of the associations. In order to reduce the type I error, we applied the false discovery rate (FDR) multiple testing correction in generalized linear model analysis [25].

Results
We found that 157 infants developed RDS requiring MV, 71 BPD, 70 IVH, and 43 ROP in our study population (S1 Appendix). Table 2 shows clinical and demographic data of the enrolled preterm infants. Genotype distributions and allele frequencies of twelve polymorphisms investigated in our population are shown in Table 3.
We observed a trend toward a different prevalence of AT+TT rs4291 ACE polymorphism of the promoter region, MT+TT rs699 AGT, and CT+TT rs3025039 VEGFA polymorphisms in infants with BPD compared to those without BPD (Table 4), and of GT+TT rs1799983 eNOS in infants with ROP compared to those without ROP (Table 6). However, these associations disappeared after adjusting for gestational age and birth weight (Table 5).   Gestational age and birth weight were not associated with any of the analyzed polymorphisms (Table 7).
According to the multivariate analysis, adjusted for gestational age and birth weight, we found that: 1) VEGFA haplotypes increase (ACCC, ACTC, ACTT, ATCC, TCCT) or decrease (ACCT, TCCC) birth weight (Table 8); 2) the eNOS haplotype (TaG) decreases gestational age (Table 8). We also found that the: 1) VEGFA haplotype (ACTT) decreases the risk of RDS and increases the risk of BPD (Table 9); 2) the VEGFA haplotype (TCCT) decreases the risk of ROP and increases the risk of IVH (Tble 9); 3) the rare (CaT and TaT) and CbT eNOS haplotypes were significantly associated with different effects (risk for RDS and BPD and protection from IVH and ROP) on clinical prematurity complications (Table 9).

Discussion
In this study we evaluated the possible association between VEGFA, eNOS, AGT, AGTR1, ACE, and HMOX-1 gene polymorphisms and the development of the most frequent prematurity complications, namely RDS requiring MV, BPD, IVH and ROP in a large population of preterm infants. Our data demonstrate for the first time that the T allele of the rs1799983 eNOS polymorphism and confirm that C allele of the rs2070744 eNOS polymorphism are independent predictors of an increased risk of BPD in preterm infants. We also discovered that haplotypes of VEGFA and eNOS genes may independently affect birth weight and gestational age. Although further associations emerged from the univariate analysis, these correlations disappeared after data adjustment for gestational age and birth weight, with the exception of TC +CC rs2070744 eNOS and GT+TT rs1799983 eNOS genotypes which were independent risk factors for BPD. We speculated that the lack of correlations of the different polymorphisms "per se" was due to the multifactorial pathogenesis of studied prematurity complications which were mainly governed by gestational age [26][27][28][29][30].
The role of the rs2070744 eNOS polymorphism in the pathogenesis of RDS and BPD has been previously investigated in a cohort (n = 124) of preterm infants by Vannemreddy et al. [6]. They found that the status of rs2070744 allele C carrier(TC+CC) was significantly more frequent in infants with RDS, BPD, and IVH compared to control patients (15.3 vs. 7.3%, p = 0.03) and represents a significant risk factor for the development of these pathologies [6] Our data are in partial agreement with these findings because we did not observe any relationship between the polymorphism and the risk of severe RDS and IVH, but we could confirm that TC+CC rs2070744 eNOS genotypes are a risk factor for BPD. Previous reports have demonstrated that TC heterozygotes and CC homozygotes for the rs2070744 eNOS polymorphism increase the risk of cardiovascular diseases in adults and decreases NO levels [6,31]. Considering the crucial role of NO in promoting lung angiogenesis during the fetal/neonatal period [2,32,33], we can reasonably speculate that the C allele of the polymorphism increases the risk of BPD through the decrease of endothelial NO synthesis that impairs lung vessel growth and in turn alveologenesis.
Similarly, the role of the rs1799983 eNOS polymorphism has been demonstrated in the pathogenesis of cardiovascular diseases in adults [34,35], but has never been investigated in the pathogenesis of prematurity complications. We found that GT+TT rs1799983 eNOS genotypes increase the risk of BPD. It has been demonstrated that the T allele carrier status is associated with reduced function of the enzyme and decreased NO production by endothelial cells compared to GG homozygous subjects [36,37]. Therefore, these polymorphisms might also favor BPD development through a negative effect on lung vessels and alveolar growth.
These considerations seem to be confirmed by the results of Ballard's large randomized controlled study demonstrating that prolonged inhaled NO treatment increases survival without BPD in high-risk preterm infants, although its routine use can not be recommended [38]. Interestingly, the multivariate analysis (adjusted for gestational age and birth weight) of the polymorphism combinations by haplotype analysis of the polymorphisms localized in the same gene showed that haplotypes of VEGFA increase or decrease birth weight and one haplotype of eNOS decreases gestational age. Moreover, the VEGFA haplotype increases the risk of IVH and BPD or decreases the risk of RDS and ROP, while the eNOS haplotype increases the risk of RDS and BPD or decreases the risk of IVH and ROP. Although it is difficult to explain these different effects, if we consider the crucial role of VEGFA [1,2] and NO [5,6] on fetal and neonatal vessel growth and protection, we may speculate that protective haplotypes might enhance their beneficial effects, and promote adequate tissue vascularization and perfusion, while unfavorable combinations might attenuate them and induce inadequate vessel development and blood flow.
Haplotype analysis allow us to identify alleles in which polymorphisms per se associated with the diseases are included, and helps to better identify subjects who are either at risk or not (e.g. VEGFA haplotypes for RDS, BPD, ROP, and IVH). Thus, our findings suggest that different SNPs in VEGFA and eNOS genes of low frequency might change the patients' susceptibility to prematurity complications by altering either their function or their expression.
The limitations of our study include the relatively small size of the study which could affect the demonstration of correlation between studied complications and polymorphisms; the lack of measurement of infants' VEGFA levels and NOS activity which precludes the possibility of relating specific polymorphisms and haplotype combinations with their expression in our patients. Thus, our findings do not allow us to demonstrate definitively whether the aforementioned SNPs and haplotypes are themselves responsible for their beneficial or detrimental effects in our population or whether they reflect linkage disequilibrium with variation in the remainder of the genes or perhaps in a nearby flanking gene. However, given the homogeneity of our population, it is unlikely that the observed associations are a statistical artefact resulting from a population substructure.

Conclusions
We confirm that the C allele of rs2070744 eNOS polymorphism is an independent predictor of an increased risk of BPD in preterm infants and demonstrate for the first time the same thing for the T allele of the rs1799983 eNOS polymorphism. We also find that haplotypes of VEGFA and eNO genes may independently affect birth weight and gestational age, and act as protecting or risk markers for prematurity complications. These results could be useful in identifying patients at high risk of developing prematurity complications in whom prevention strategies would be particularly indicated.
Supporting Information S1 Appendix. Database of clinical characteristics and polymorphisms of studied infants. (XLS)