Single Nucleotide Polymorphisms In The Nos2 Biology Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Background: Polymorphisms in nitric oxide synthase (NOS) genes (NOS1, NOS2 and NOS3) have been suggested to have a major impact on fraction of exhaled nitric oxide (FENO), a biomarker of airway inflammation. Recently, NOS2 was reported to be an important genetic determinant of FENO in children. The aim of the study was to thoroughly investigate the association between single nucleotide polymorphisms (SNPs) in all three NOS genes and FENO in an adult population.

Method: In 1737 adults from a Swedish general population sample, FENO was measured and genetic variation in the NOS genes was assessed using 49 SNPs. The genetic effect of NOS polymorphisms on FENO, asthma and atopy was estimated using multiple regression models.

Results: In a multi-SNP model, two SNPs in NOS2 and one in NOS3 showed independent associations with levels of FENO based on stepwise regression analysis. For NOS2 SNP rs9901734, subjects had 5.3% higher levels of FENO per G allele (p = 0.016) and for rs3729508, subjects with CC or CT genotypes had 9.4% higher levels compared with TT (p = 0.004). For NOS3 SNP rs7830, subjects with GT or TT had 5.6% higher levels than GG (p = 0.034). These associations did not differ significantly according to asthma or atopy status.

Conclusion: We identified three polymorphisms in the NOS2 and NOS3 genes that were independently associated with elevated levels of FENO.

Key words: FENO, NOS, polymorphisms, Association study


ADONIX: Adult-onset asthma and exhaled nitric oxide

ATS: American thoracic society

CR: Call rate

ERS: European respiratory society

FENO: Fraction of exhaled nitric oxide

FEV1: Forced expiratory volume in 1 second

FVC: Forced vital capacity

HWE: Hardy-Weinberg Equilibrium

IgE: Immunoglobulin E

NO: Nitric oxide

NOS: Nitric oxide synthase

eNOS; NOS3: Endothelial NOS

iNOS; NOS2;NOS2A: Inducible NOS

nNOS; NOS1: Neuronal NOS

PCR: Polymerase chain reaction

ppb: Parts per billion

SNP: Single nucleotide polymorphism


The fraction of exhaled nitric oxide (FENO) is a non-invasive biomarker of airway inflammation and a useful clinical tool for diagnosing and monitoring asthma [1, 2]. It is easy to measure and results can be obtained in real time (online) making it an attractive method in clinical practice. However, one obstacle is that the inter-individual variation is relatively large, and there is limited knowledge about factors explaining this variation.

Genetic factors have been suggested to have a major impact on FENO. In a Norwegian twin study, genetic factors explain 60% of the variability in FENO [3]. There was also a small but significant association between FENO and atopy, and between FENO and airway hyper-responsiveness especially among those with atopy, which appeared to be explained by genetic factors. No specific genotypes explaining this variation were identified.  

Nitric oxide (NO) is synthesized by specific NO synthase (NOS) enzymes, with three distinct isoforms: neuronal NOS (nNOS; NOS1), endothelial NOS (eNOS; NOS3) and a more inducible form (iNOS; NOS2) [4, 5]. The NOS1 and NOS3 genes are constitutively expressed in the lung resulting in a low basal synthesis of NO [6, 7], whereas expression of NOS2 seems to be regulated by gene transcription factors such as nuclear factor kappa B (NFKβ), resulting in a profoundly greater NO production [8]. Although all three isoforms are present in the lung tissue [9], mainly iNOS (NOS2) is measurable, and contributes to variation in FENO levels [10].

Genetic polymorphisms are known to exist in all NOS genes, but only a few studies have examined the role of various NOS polymorphisms for FENO levels and results have been inconsistent [11-15]. An early finding of an association between AAT repeats in intron 20 in the NOS1 gene with higher FENO levels in asthmatic adults [11], was not replicated in children [12, 16]. In another study, no significant difference in mean concentration of FENO was seen for many (≥ 14) CCTTT repeats in the NOS2 promoter region vs. less repeats [14]. In a recent population-based study, genetic variants in NOS2 were significantly associated with higher levels of FENO among children [17]. For NOS3, an association of the G894T variant (rs1799983) to elevated FENO levels was reported in adult subjects with asthma but not in Chinese children [15, 16].

The aim of this study was to examine the association of common genetic variation of the three NOS genes with levels of FENO, by studying selected, previously reported, associated single nucleotide polymorphisms (SNPs) complemented with a set of tag SNPs for each gene selected to capture the genetic variation. In addition, we wanted to examine whether any risk variants might be associated with asthma, atopy or lung function and if the genetic effect on FENO is modified by these respiratory phenotypes.

Material and Methods

Study population and design

The investigated study forms part of the ADONIX (adult-onset asthma and exhaled nitric oxide) study including 2198 subjects, which is a subproject based on responded cohort of 3610 of 8625 eligible subjects, from INTERGENE research program that assessed the INTERplay between GENEtic susceptibility and environmental factors for risk of chronic diseases in western Sweden [18-20]. The ADONIX study population consisted of randomly selected men and women aged 25-75 years from the city of Gothenburg and surrounding municipalities in Sweden, recruited between June 2001 and December 2003. All the participants received a postal questionnaire and an invitation to a clinical examination, which included FENO measurements, lung function measurements (forced expiratory volume in 1 second [FEV1] and forced vital capacity [FVC]) as well as blood samples. Participants were classified as never, former and current smokers, based on information from the questionnaire. Asthma was defined as a positive response to at least one of the following questionnaire items: 1) self-reported asthma; 2) doctor-diagnosed asthma; 3) asthma attack during the last 12 months; 4) asthma attack during the month preceding the study. Atopy was defined as presence of specific IgE antibodies to inhalation allergens (house dust, mite, cat, dog, timothy grass and local allergen), determined by the Phadiatop test (Pharmacia Diagnostics; Uppsala, Sweden). A participant was considered atopic if the specific IgE was ≥ 0.35 kU/1 [21, 22]. Spirometry was performed with a dry wedge spirometer (Vitalograph; Buckingham, UK), and FEV1/FVC ratio and percentages of predicted FEV1 and FVC were calculated based on age, sex and height [23].

Exhaled Nitric Oxide measurements

FENO was measured before spirometry at an expiratory flow rate of 50 mL/s using a chemiluminescence method (NIOX-system; Aerocrine AB; Stockholm; Sweden) according to ATS/ERS recommendations [24, 25]. Exhalations were registered three times during the study period June 2001 to January 2003 and twice during February 2003 to December 2003 according to revised ATS/ERS recommendations, and the mean concentration of these was used.

SNP selection and Genotyping

Fifty-four SNPs of the NOS1, NOS2 and NOS3 genes were selected, a few based on literature, and these were then complemented with tagging SNPs to capture genetic variation across each gene (Supplementary Table 1). Tag SNP selection was done with a pairwise approach at an r2 cutoff value of 0.8 and minor allele frequency cutoff of 0.05, including 10 kb upstream and 10 kb downstream of the genes in the European ancestry genotype information from the HapMap phase III database ( SNPs were genotyped using a Sequenom MassARRAY platform (Sequenom San Diego, CA, USA) or a competitive allele specific PCR system, KASPar (KBioscience, Hoddesdon Herts, Great Britain). Potential genotyping problems were identified by call rate (threshold 95%) and Hardy-Weinberg Equilibrium (HWE) testing.

Participants of non-European origin were excluded from the study. In total, 1737 subjects had complete information on NOS polymorphisms and FENO values and were included in the analysis.

Statistical Analysis

Association between NOS polymorphisms and FENO was analyzed under five different genetic models: additive, recessive, dominant, co-dominant (genotype-specific risk) and over-dominant (heterozygote risk) [26].

Initially, we performed a linear regression analysis for each of the 49 SNPs, adjusting for age, sex, height and smoking, where the associated SNPs were identified and ranked according to their p-value of association.

Next we performed a forward stepwise multiple linear regression analysis in two stages to identify a reduced set of the strongest independently associated SNPs and the most adequate genetic model for each final SNP.

In the first stage, SNPs with p ≤ 0.2 from the additive genetic model in the single-SNP association were entered into the stepwise regression analysis, which aimed to determine the most strongly associated SNP or SNPs in a model allowing for independent effects of several SNPs.

In the second stage, significant SNPs (p ≤ 0.05) from the first stage stepwise regression model, as well as any SNPs from the initial single-SNP association analyses with p ≤ 0.05 from any non-additive (dominant, recessive, over-dominant and co-dominant) genetic model were included. This selection aimed to ensure that genetic effects not following an additive model would not be missed in this final stage. All of the SNPs thus selected were again codified into all five genetic models, and using this set of SNPs coded in the various genetic models as input, we then performed a forward stepwise regression analysis in order to identify the most predictive SNPs, each with the best-fitting genetic model, allowing for independent effects of several SNPs.

In both stages, the p-value for a SNP covariate to enter and remain in the model were set at p = 0.10 and p = 0.20, respectively. The covariates age, sex, height, atopy and smoking were forced into the model.

For the estimation of effect size, risk genotype was modelled as the genotype that was associated with higher levels of FENO.

In addition to the overall effects, stratified analyses were also performed for the final SNP model by asthma or atopy status, respectively, in a common interaction model where the genetic effect was estimated separately for each stratum. Likelihood ratio test was used to test differences among the strata with appropriate interaction terms. Models were adjusted for age, sex, height and smoking.

We also performed logistic regression models to examine association between identified risk genotypes and asthma and atopy, respectively, under a dominant genetic model to optimize statistical power. We also analyzed the relationship between the risk genotypes and lung function, as assessed by FEV1/FVC and percent predicted FVC and FEV1.

Single-SNP association analyses were performed with the R statistical package 'SNPassoc'. Further regression analyses were done with SAS version 9.2 (SAS Institute; Cary, NC, USA). Since FENO50 had a skewed distribution, the values were log transformed before model fitting. A P-value ≤ 0.05 was considered significant. All reported effect estimates are recalculated and expressed as percentage change. Odds ratios (ORs) and 95 % CI were used to estimate effects of risk genotypes on asthma and atopy.


Baseline characteristics for the study population are presented in Table 1. Four SNPs had call rates below 95% and one SNP showed departure from HWE (p = 0.05), leaving 49 SNPs (20 NOS1, 17 NOS2 and 12 NOS3) for use in the analyses. Minor allele frequencies, HWE p-values and call rates are shown in Supplementary Table 1. Pairwise LD among SNPs of all three NOS genes were shown separately in Supplementary Figures 1, 2 and 3.

Thirteen SNPs with p ≤ 0.2 from the additive model in the single-SNP analysis (Figure 1, Supplementary Table 2) were entered into a first stage stepwise regression analysis. Only NOS2 rs9901734 (p = 0.01) was significantly associated with FENO.

The SNP rs9901734 and seven SNPs with p â‰¤ 0.05 from non-additive genetic models in the single-SNP analysis (Figure 1) were entered into the second stage stepwise regression analysis. Two SNPs in the NOS2 gene (rs9901734 and rs3729508) and one SNP in the NOS3 gene (rs7830) were all significantly associated with FENO, each with different genetic models (Table 2). For rs9901734, subjects had 5.3% higher levels of FENO per each copy of the G allele (additive model). For rs3729508, the optimal fit was provided by a negative recessive genetic effect model for the minor T allele implying that subjects with CC or CT genotype had 9.4% higher FENO levels compared TT genotype (i.e. dominant effect for the major C allele); and for rs7830, subjects with GT or TT had 5.6% higher levels of FENO than GG genotype (dominant model).

When we stratified by asthma or atopy, the effect estimate for NOS2 rs9901734 in "healthy" subjects (i.e. no asthma, wheeze and atopy; n = 958) was marginally strengthened and remained significant (Table 3). The NOS2 rs3729508 estimate was unchanged, while for NOS3 rs7830 little effect on FENO was seen. In subjects with asthma or atopy, effect estimates for both NOS2 and NOS3 variants were in the same direction as the overall results, although due to the statistical power limitations no differences between groups were significant (p-value for interaction > 0.05 for all models; Table 3).

We could not identify any significant association between the risk genotypes of NOS SNPs and asthma, atopy or lung function parameters in the cohort (Supplementary Tables 3 & 4).


In this large Swedish study of an adult general population, we have investigated genetic variation in the NOS genes, using SNPs previously reported in literature as well as tag SNPs. Of 49 NOS SNPs examined, 2 SNPs in NOS2 and one SNP in NOS3 were independently associated with variation in FENO levels. The NOS2 gene, also known as inducible NOS, may be considered a priori and the most likely candidate for a rise in FENO associated with airway inflammation. A population-based study among children in California recently provided strong evidence supporting that polymorphisms of NOS2 are an important determinant of exhaled NO [17]. A gene expression study in children also demonstrated that the expression of iNOS in human airway epithelium cells is an important determinant of NO in exhaled breath [10].

The third SNP, rs7830, that was significantly associated with higher levels of FENO was located in the NOS3 (endothelial NOS) gene. A missense polymorphism (G849T) in the NOS3 gene has previously been reported to be associated with higher levels of FENO in adult subjects [15], but this polymorphism is not in linkage disequilibrium (LD) with our rs7830 (r2 = 0.03).

Previous studies have also reported an association between polymorphisms in the NOS1 gene and FENO [11]. However, no clear association was identified between any of the NOS1 SNPs and FENO in our data. This is well in line with gene expression data, where NOS2 is highly expressed in the human lung, while NOS3 is expressed to a lesser extent and NOS1 in principle is undetectable [10] .

SNP rs3729508 is located in the intron region of NOS2, rs7830 is located in the 3-prime untranslated region of NOS3 and rs9901734 is located in the intergenic region, close to NOS2. These SNPs may affect regulation of gene expression, directly affecting FENO. Alternatively, they may be in LD with another functional SNP or SNPs that actually affect variation in FENO. Further replication of these NOS SNPs can provide more information to better understand their role in airway inflammation and respiratory health.

Even though FENO has been used for more a decade, it has been difficult to establish measurement of FENO as a tool for diagnosis and monitoring in asthma and it has not yet fully found its place in the clinical setting. This has mainly been due to the challenge in monitoring disease activity for clinical studies to show a beneficial clinical effect when basing clinical action on a measured FENO value. In the diagnosis of asthma, a great advantage of FENO is to aid in ruling out the disease, at least to rule out eosinophilic airway inflammation responsive to steroid treatment. When trying to use a single FENO measurement for guiding diagnosis one obstacle is the large variation in a "normal FENO value". We have previously endeavored to establish normal values for FENO [20] based on FENO measurement from the same population as in this study. Upper normal values (95th percentiles) ranged from 22 ppb to 57 ppb depending on age, height and presence of atopy. FENO is most probably also associated with sex, but this association did not reach statistical significance in this data set. So far, these normal values have not been applied in any diagnostic study, so whether they will improve the utility of FENO in the diagnosis of asthma is still unclear. At present it is not possible to evaluate the FENO value in the scope of the presence/absence a NOS polymorphisms in a clinical setting, and the practical implications of the results are hence unclear. In a not too-far-away future this may however be a reality, and would enhance the interpretation of an individual FENO value. Based on our results, an individual with all of the studied risk genotypes (21% of population) will have on average 35% higher FENO than an individual none of the risk genotypes (4% of the population).

This study has several strengths. First, it was conducted on a relatively large homogenous Swedish population sample and restricted to subjects of Caucasian origin, to limit population stratification. Second, we comprehensively evaluated common genetic variation within all of the important human NOS genes (NOS1, NOS2 and NOS3), by first selecting SNPs with previous evidence of relevance to airway inflammation, asthma and other respiratory diseases, and then using tag SNPs to complement and achieve good coverage. Thus, our analysis focused on characterizing in detail NOS regulation of FENO, based on a relatively strong prior hypothesis.

Some potential weaknesses of the present study should be noted. Asthma was defined based on questionnaire responses with at least one positive answer to one similar question, which produce a potential risk of misclassification. On the other hand, this broader asthma definition increased the number of asthmatic subjects and potentially the statistical power to detect effect differences between healthy and asthmatic subjects. Despite this, we did not observe any significant difference. According to the manufacturer (Aerocrine AB; Stockholm; Sweden) the variability in FENO measurements is around 3 ppb. This variability would induce some level of independent misclassification unrelated to genetic variation, and will on average tend to slightly bias our results towards the null.

To best investigate the potential true association, we performed a forward stepwise multiple linear regression analysis in two stages. In the context of genetic association studies, stepwise regression analysis has been proposed as a useful method to select the most relevant typed variants and thereby potentially also capture effects of untyped causal functional polymorphisms within the gene or where sufficient LD structure is present also in adjacent genes [27]. In the first stage, we aimed to determine the most strongly associated SNP or SNPs in a model allowing for independent effects of several SNPs. In the second stage, we aimed to ensure that genetic effects not following an additive model would not be missed. Thus we identified the most predictive SNPs, each with the best-fitting genetic model, allowing for independent effects of several SNPs.

In conclusion, two polymorphisms in the NOS2 ("inducible NOS", iNOS) and one polymorphism in the NOS3 ("endothelial NOS", eNOS) gene but no polymorphisms in the NOS1 ("neuronal NOS", nNOS) gene were independently associated with elevated levels of FENO. Our results are consistent with recent results reported in children, suggesting that NOS2 and as well as NOS3 are the major NOS genes contributing to some of the variations in FENO.

Tables and Figures

Table 1 Basic characteristics of the study population and FENO levels overall and in subgroups (Limited to those with genotype and FENO data at baseline, n =1737)


Mean (SD)

Age, years

49.2 (13.6)

Height, cm

172.3 (9.1)

FEV1/FVC ratiob

0.8 (0.1)

FVC (%predicted)b

97.1 (12.6)

FEV1 (% Predicted)b

93.8 (13.6)

FENO levels, ppb

All (n =1737)

15.9 (1.8)

Asthma (n = 298)c

17.3 (1.9)

Atopy (n = 432)

18.4 (1.9)

Never-smokers (n = 878)

17.3 (1.7)

Former-smokers (n = 534)

16.9 (1.7)

Current-smokers (n = 325)

11.2 (1.8)

FENO: Fraction of exhaled nitric oxide; ppb: parts per billion; SD: Standard Deviation;

FEV1: Forced Expiratory Volume in 1 sec; FVC: Forced Vital Capacity; bn = 425 missing; cn = missing 4.

Table 2 Association of SNPs in NOS genes with FENO among adults


Genetic model


n = 1737

* Difference in



FENO (%), 95% CI






10.7 (1.9 - 19.5)



5.3 (1.0 - 9.7)


CC (ref)






9.4 (3.1 - 15.2)










5.6 (0.4 - 11.1)





GG (ref)




*FENO values were expressed as percentage change, adjusted for age, sex, height, atopy and smoking habits.

Table 3 Percent change in levels of FENO in healthy, asthmatic and atopic subjects according to presence of risk genotypes of NOS SNPs.







n = 1737


(CG/GG vs. CC)


(CC/CT vs. TT)


(GT/TT vs. GG)


n = 958

5.8 (0.6 - 11.1 )

8.9 (0.6 - 17.8)

1.5 (−4.4 - 7.7)


n = 298

2.4 (−9.1 - 15.4)

11.3 (−9.1 - 36.4)

21.9 (4.6 - 42.0)



n = 432

3.8 (−5.5 - 14.0)

13.6(−2.0 - 31.8)

4.9 (−6.5 - 17.7)


* % difference in mean values of FENO (risk vs. no-risk genotype), adjusted for age, sex, height and smoking habits; † healthy=no asthma, wheeze and atopy. a By likelihood ratio test.

Figure 1

Figure 1 Single-SNP association bweteen FENO and 49 NOS SNPs. Results for 5 genetic models are shown. Each circle represents the minus log 10 of the p-value for one single NOS SNP for each genetic model. The horizontal dotted pink line shows statistcally signifcant levelat p value 0.05. ↑SNPs with p-value ≤ 0.2 from additive model. SNP with p-value ≤ 0.05 from non-additive models.