IL-8 Gene Variants are Associated with Lung Function Decline and Multidimensional BODE Index in COPD Patients But Not with Disease Susceptibility: A Validation Study

Abstract

Background and objective: COPD is a leading cause of dead worldwide and tobacco smoking is its major risk factor. IL8 is a proinflammatory chemokine mainly involved in the acute inflammatory reaction. The aim of this study was to test the association of IL-8, CXCR1 and CXCR2 gene variants and COPD susceptibility as part of a replication study and explore the effect of these variations in disease progression. Methods: 9 tagSNPs were genotyped in 728 Caucasian individuals (196 COPD patients, 80 smokers and 452 non-smoking controls). Pulmonary compromise was evaluated using spirometry and clinical parameters at baseline and annually over a 2 years period. We also determined plasma levels of TNF-α, IL-6, IL-8 and IL-16 in COPD patients. Results: There was a lack of association between gene variants or haplotypes with predisposition to COPD. No correlation was observed between the polymorphisms and cytokines levels. Interestingly, significant associations were found between carriers of the rs4073A (OR = 3.53, CI 1.34-9.35, p = 0.01), rs2227306C (OR = 5.65, CI 1.75–18.88, p = 0.004) and rs2227307T (OR = 4.52, CI = 1.49–12.82, p = 0.007) alleles in the IL-8 gene and patients who scored higher in the BODE index and showed an important decrease in their FEV₁ and FVC during the 2 years follow-up period (p < 0.05). Conclusions: Despite no association was found between the studied genes and COPD susceptibility, three polymorphisms in the IL-8 gene appear to be involved in a worse progression of the disease, with an affectation beyond the pulmonary function and importantly, a reduction in lung function along the follow-up years.

Keywords:

• COPD
• cytokine
• genetics
• IL-8
• FEV₁
• FVC

Background

Chronic Obstructive Pulmonary Disease (COPD), a major cause of morbidity and mortality worldwide, is predicted to be the fourth cause of death and disability in the world by the year 2030 (1). COPD is thought to result from an abnormal pulmonary inflammatory response to noxious particles or gases (1) in which CD8+ Tc1 lymphocytes together with macrophages and neutrophils are supposed to play an important role (2–4). Many different inflammatory markers appear to be increased in the serum of stable COPD patients, or were associated with COPD risk, and worse outcomes of the disease (5–9).

Among the proinflammatory cytokines and chemokines, IL-8, on chromosome 4q12–q21, is mainly involved in the initiation and amplification of acute inflammatory reaction and in the chronic inflammatory process. Therefore, it plays an important role in diseases like COPD, in which inflammation is a substantial pathophysiological feature (10, 11). IL-8 mediates chemoattraction via two different receptors: CXCR1 (­previously called IL8RA) and CXCR2 (or IL8RB), which are encoded by 2 single-copy genes and located on chromosome 2q34–q35. CXCR1 and CXCR2 have 77% amino acid identity and bind to IL-8 with high affinity, and transduce the signal through a G-protein-activated second messenger system. CXCR1 is more specific for IL-8, while CXCR2 is less specific for IL-8 and will bind other related chemokines with high ­affinity (12, 13).

Resequencing of the promoter region of IL-8 identified a common single nucleotide polymorphism (SNP) at position –251. The IL-8 -251A allele tends to result in greater expression of IL-8 in leukocytes when stimulated with lipopolysaccharide (LPS) (14). Moreover, a genome-wide screen study revealed the chromosome region in which CXCR1 and CXCR2 lie as the one with the highest linkage to spirometric phenotypes for early onset COPD (15). In this context, a study performed in a German cohort significantly associated a SNP in CXCR1 with chronic obstructive pulmonary disease (COPD) and asthma (16).

In this study, we evaluated common genetic variations across the entire IL-8, CXCR1 and CXCR2 genes using a haplotype tagging SNP approach in an attempted to replicate previous findings that associated variants in these genes with COPD susceptibility. Then, we aimed to investigate the effect of variations in the IL-8, CXCR1 and CXCR2 genes in the disease progression and its relationship with circulating levels of IL-6, IL-8, IL-16 and TNF-α in COPD patients.

Methods

Study subjects

Caucasian individuals included in this study were divided into three groups: COPD patients, smokers without COPD and non-smoking controls. COPD outpatients were recruited from a pulmonary clinic at the Hospital Universitario N. S. de Candelaria (Canary Islands, Spain) from August 2002 to September 2009. Patients with a wide range of disease severity were included using the following criteria: age > 40 years, smoking history > 10 pack-years and a post-bronchodilator FEV₁/FVC ratio < 0.70 measured 20 min after the administration of 400 mg of inhaled albuterol. Pulmonary function test, spirometry and lung volumes were measured according to ATS-ERS guidelines (17). The bias carried by using the FEV₁/FVC< 0.7 was corrected by using the FEV₁/FVC<LLN (lower limits of normal) (18) and the predicted values calculated using the software PredValues v3.1.0.118 (www.spirxpert.com).

The arterial oxygen tension (PaO₂) was measured at rest. Exercise capacity was also tested using the best of two 6-min walking distance (6MWD) tests separated by at least 30 min following the ATS recommendations (19). The body mass index (BMI) was calculated as the weight in kilograms divided by the square of the height in meters. Dyspnea was evaluated by the modified Medical Research Council (mMRC) scale (20). The BODE (Body Mass Index, Airflow Obstruction, Dyspnea, Exercise Performance) Index was calculated as previously described (21). To quantify the degree of co-morbidity we used the Charlson index (22). All patients were clinically stable (no exacerbation for at least 6 weeks) at the time of evaluation. Patients were excluded if they had history of other respiratory diseases like asthma or bronchiectasis.

As a control group we included 452 individuals ­without respiratory disease and no smoking history from an on-going general adult population cohort named CDC project in the Canary Islands (23). Finally, the third group included 96 current smokers with a cumulative smoking history of > 10 pack-years and normal lung function (FEV₁ > 80%; FEV₁/FVC ≥ 0.70). In order to minimize any possible effect of population structure on our estimations, cases and controls were recruited if they have at least two generations of Canarian ancestry (24). The Ethics Committee of Clinical Investigation of the Hospital approved this study (approval No. 36) and written informed consents were obtained from all ­subjects.

SNPs selection and genotyping

The haplotype-tagging SNPs in the IL-8, CXCR1 and CXCR2 genes selection was based on the resequencing data for European individuals of the SeattleSNPs database (http://pga.gs.washington.edu). A multimarker tagging algorithm criteria of r2 > 0.8 and minor allele frequency (MAF) ≥ 0.5 was used with the Haploview program (version 1.0) (25). Genomic DNA was extracted from whole blood sample using the commercially available GFX Kit (GE Healthcare). Ten SNPs in these genes were genotyped in the patients and controls samples using Sequenom (Hamburg, germany) iPLEX SNP genotyping protocol, available at the CEGEN-ISCIII (Genotyping National Centre, www.cegen.org). Genotyping was blind to case or control status and samples yielded the 95–100% genotypes. To ensure the quality of genotyping data, 10% of randomly selected samples were interspersed in the plates as replicates and also sequenced to verify the genotyping results.

Serum measurements

COPD patients’ serum was prepared by centrifugation of coagulated blood tubes at 3200 rpm for 10 minutes at room temperature, and stored at −80°C until laboratory analysis. Samples were tested for IL-6, IL-8, IL-16 and TNF-α by quantitative sandwich enzyme-linked immunosorbent assay (ELISA) according to the manufacturer's instructions (R&D Systems Europe, Minneapolis). The detection limits were 0.7, 3.5, 6.2 and 5.0 pg/ml, respectively. The within-assay coefficient of variation for all assays was less than 10%.

Statistical analysis

Differences in demographic and clinical variables among groups were identified using Student t-test, Mann-Whitney U-rank test, or Pearson χ² test as appropriate. All the statistical analysis were performed by using the package SPSS v19 (SPSS, Inc). Contingency tables and the χ² test for independence were used for comparisons in the allele and genotype distribution between patients and control groups. Disequilibrium values (LD) and departures from Hardy-Weinberg equilibrium were calculated using the GENEPOP v3.4 and Haploview v3.2 (4.1) software programmes, as appropriate (25, 26). Haplotype analyses and reconstruction from biallelic polymorphisms was conducted using PHASE v2.1 (27). To test haplotype associations and for checking for consistency of the inferred data, haplotypes were reconstructed from the best average goodness-of-fit output of six PHASE 2.1 runs with 1,000 permutations (27).

Multivariate logistic regression to estimate the genotype-specific odds ratio (OR) and 95% confidence interval (CI) for COPD susceptibility was used with SNPStats software (28). Biallelic tests were conducted for SNPs and an additive and/or co-dominant genetic model was first used. Potential risk and confounding factors (age, sex, pack-years of smoking) were adjusted by a multivariate logistic regression analysis. We also analysed the gene variants association with the degree of airways obstruction as well as with other prognostic parameters obtained in the clinical practice using as thresholds the ones that have shown to predict mortality in patients with COPD: the BODE index (> 3) and its components (FEV₁ < 50%, BMI < 21, dyspnea mMRC > 2, 6MWD < 350 meters), PaO₂ (< 60 mm Hg), IC/TLC (< 0.25) and K_CO (< 80%) (Table 1).

Table 1.   Demographic characteristics of COPD patients, smokers and non-smoking individuals included in the association study

Variable	COPD cases (n = 196)	Smokers (n = 96)	Non-smoking controls (n = 452)	p-value
Gender (male%)	74	48	75	<0.0001
Age (years)*	62 ± 10	49 ± 10	53 ± 8	<0.0001
Smoking history (pack-yrs)^†*	64 ± 26	38 ± 20	1 ± 3	<0.0001
FEV1 (L)*	1.48 ± 0.7	3 ± 0.7	__	NA
FEV1 (% pred)*	55 ± 20	101 ± 14	__	NA
FVC (% pred)*	87 ± 22	107 ± 15	__	NA
FEV1 / FVC (%)*	51 ± 12	78 ± 7	__	NA
PaO2 (mmHg)*	69.6 ± 12.9	__	__	NA
KCO (%)*	74 ± 26	__	__	NA
IC/TLC (%)*	33 ± 10	__	__	NA
6MWD (m)*	494 ± 94	__	__	NA
BMI (Kg/m^2)*	27 ± 5	__	__	NA
mMRC dyspnea^**	1 (0–2)	__	__	NA
BODE index^**	2 (1–3)	__	__	NA
BODE index ≤ 3^‡	130	__	__	NA
BODE index > 3^‡	24	__	__	NA
GOLD I-II^‡	99	__	__	NA
GOLD III-IV^‡	87	__	__	NA
Charlson index^**	3 (2–4)	__	__	NA

*Data are presented as mean ±SD

^**Data are presented as median (25th–75th)

^†Number of packs of cigarettes smoked per day × number of years smoking

FEV₁: forced expiratory volume in one second; FVC: forced vital capacity;% pred.: per cent predicted; 6MWD: six-min walk distance test; BMI: body mass index; NA: Non Applicable.

^‡Number of subjects in the two groups of GOLD and BODE index considered to analysis.

A multivariate logistic regression analysis adjusted by age, sex and pack-yrs of smoking was performed for the analysis of severity. We conducted a longitudinal analysis to evaluate the progression of the disease of clinical and pulmonary function variables using the SNPs as comparative factor by using a general linear modelling for repeated measures (GLIM). ANOVA was performed to test the effect of the polymorphisms and serum proteins levels. In every case, a two-sided p value of < 0.05 was considered statistically significant.

Results

The demographic data and baseline characteristics of the study groups are summarized in Table 1. COPD patients (147 men and 49 women) include a wide range of the airflow obstruction and were distributed in GOLD levels as follows: I (14.3%), II (42.5%), III (34.6%) and IV (8.6%).

All SNPs conformed to HWE with the exception of rs4674259 (p < 0.05) in the CXCR2 gene, and was then excluded from further analyses. Genotype frequencies for control and case subjects are presented in additional Table S1.

Table 2 shows the odds ratios and allele frequencies for the IL-8, CXCR1 and CXCR2 polymorphisms. When comparing COPD patients with non-smoking control subjects and smokers without COPD, we did not observed a significant effect of any IL-8, CXCR1 or CXCR2 ­polymorphism in the predisposition to COPD.

Table 2.   Association of SNPs in the IL8, CXCR1 and CXCR2 genes with Chronic Obstructive Pulmonary Disease (COPD)

Gene and SNPs	Risk allele	COPD cases MAF%	Non-smoking controls MAF%	Control Smokers MAF%	ORad (95% CI)^a	ORad (95% CI)^b
IL-8
rs4073	A	0.44	0.46	0.48	0.93 (0.70–1.24)	0.83 (0.50–1.36)
rs2227307	G	0.4	0.41	0.43	0.96 (0.72–1.28)	0.88 (0.52–1.50)
rs2227306	T	0.36	0.39	0.42	0.91 (0.66–1.21)	0.78 (0.47–1.31)
CXCR1
rs1008563	T	0.52	0.49	0.52	1.19 (0.88–1.61)	0.82 (0.49–1.38)
rs1008562	G	0.39	0.4	0.43	0.90 (0.66–1.21)	0.81 (0.49–1.35)
CXCR2
rs1126579	T	0.38	0.4	0.41	0.87 (0.64–1.17)	0.81 (0.48–135)
rs4674261	C	0.52	0.48	0.51	1.25 (0.92–1.69)	0.84 (0.50–1.42)
rs7607437	A	0.39	0.4	0.41	0.88 (0.65–1.20)	0.87 (0.52–1.45)

Data are presented as MAF: minor allele frequency (%). OR_adj: odds ratio adjusted: Age and gender were included in a multivariate logistic regression analyses as potential confounding factors in an additive model; CI: confidence interval. ^aCOPD cases vs. non-smoking control individuals. ^bCOPD cases vs. smokers. p-values were >0.05.

We examined all possible haplotypes in each chromosome for differences between cases and the control groups. Table 3 shows the 2 common haplotypes inferred from the polymorphisms analysed in each chromosome. Similar with the results obtained from genotype analyses, no significant associations (p > 0.01) were observed between any of the analysed groups and the inferred haplotypes.

Table 3.   Association of IL8, CXCR1 and CXCR2 haplotypes with Chronic Obstructive Pulmonary Disease (COPD)

	COPD cases vs. Non-smoking controls			COPD cases vs. Smokers
Gene and haplotypes	Freq (%)	ORad (95% CI)	p-value*	Freq (%)	ORad (95% CI)	p-value*
IL-8 (chr4)^a
TTC	0.54	—		0.54	—
AGT	0.38	0.91 (0.68–1.23)	ns	0.38	0.78 (0.47–1.32)	ns
ATC	0.04	0.88 (0.44–1.73)	ns	0.04	1.14 (0.31–4.16)	ns
Rare	0.03	1.25 (0.47–3.33)	ns	0.03	0.78 (0.16–3.84)	ns
CXCR2- CXCR1 (Chr2)^b
CCGTC	0.49	—		0.5	—
TGACG	0.39	0.83 (0.60–1.14)	ns	0.38	0.76 (0.44–1.33)	ns
CGGCC	0.09	0.78 (0.47–1.31)	ns	0.07	0.88 (0.33–2.33)	ns
Rare	0.03	1.13 (0.43–2.94)	ns	0.03	0.87 (0.23–3.30)	ns

Data are presented as MAF: minor allele frequency (%). OR_adj: odds ratio adjusted; CI: confidence interval. Age, gender and pack/year were included in a multivariate logistic regression analyses as potential confounding factors in an additive model. *p-values >0.05. ns: non-significant

^ars4073/rs2227307/rs2227306.

^brs1126579/rs4674261/rs7607437/rs1008563/rs1008562.

When exploring the effect of the polymorphisms studied on the respiratory and clinical variables registered in the COPD patients, we found that the rs4073A, rs2227306C and rs2227307T alleles in the IL-8 gene were significantly associated with patients who scored higher in the BODE index (BODE ≤ 3 vs. BODE > 3) in a multiple logistic regression analyses, adjusting for age, sex and pack-years (Table 4). When analysed independently, the variables that define the BODE index (BMI, mMRC dyspnea, FEV₁% and 6MWD) did not statistically associate (p > 0.05) with a worse progression of the disease. The other clinical and pulmonary function parameters (IC/TLC, K_CO, PaO₂) as well as the comorbidities (­Charlson index) registered in our COPD patients did not associate with any of the genetic ­variants studied (p > 0.05, data not shown).

Table 4.   Association test of IL8 rs4073, rs2227306 and rs2227307 polymorphisms with COPD severity

	BODE≤3 vs BODE>3
Included variable^a	ORadjusted	95% CI	p-value
Dominant model
rs4073 (TA+ AA)	3.53	1.34–9.35	0.011*
rs2227306 (CT+CC)	5.65	1.75–18.88	0.004*
rs2227307 (GT+TT)	4.52	1.49–12.82	0.007*

^aAge, gender and smoking history (pack/years) were included in the regression analyses as confounding variables. OR: odds ratio; CI: confidence interval, *p values < 0.05.

Interestingly, patients that presented the same IL-8 gene variants rs4073A, rs2227306C and rs2227307T showed an important decrease in their lung function, measure by FEV₁ and FVC, during the two follow-up years of study (Table 5). The association found between these SNPs and FVC was not statistically significant at baseline, but we could appreciate FVC differences between carriers of the rs4073A, rs2227306C and rs2227307T alleles and non-carriers, in each one of the two following moments analysed. In the same way, we observed an important decrease in FEV₁ over time (Table 5) even though we could not appreciate differences between the allele carriers and non-carriers, individually at each moment (p > 0.05).

Table 5.   Association of IL8 rs4073, rs2227306 and rs2227307 gene polymorphisms with pulmonary function (FVC and FEV₁) in COPD patients over time

A
SNPs	Genotype	FVC			p-value^†
		Baseline	1° year	2° year
rs4073	TT	2.77 (2.29–3.54)	2.69 (2.24–3.55)	2.66 (2.13–3.68)	0.008
	AT+AA	2.56 (2.14–3.26)	2.33 (1.96–2.84)	2.18 (1.75–2.69)
	P-value	ns	0.01	0.005
rs2227306	TT	2.76 (2.21–3.54)	2.58 (2.16–3.45)	2.63 (2.08–3.60)	0.051
	CT+CC	2.55 (2.25–3.29)	2.36 (1.98–2.89)	2.20 (1.81–2.67)
	P-value	ns	0.059	0.037
rs2227307	GG	2.78 (2.28–3.56)	2.69 (2.22–3.50)	2.71 (2.18–3.72)	0.015
	TG+ TT	2.55 (2.20–3.36)	2.36 (2.06–2.87)	2.18 (1.72–2.67)
	P-value	ns	0.032	0.004
B
		FEV1
SNPs	Genotype	Baseline	1º year	2º year	p-value^†
rs4073	TT	1.43 (1.05–1.88)	1.27 (0.91–1.75)	1.32 (0.81–1.85)	0.047
	AT+AA	1.33 (0.93–1.71)	1.09 (0.92–1.54)	0.99 (0.79–1.31)
	P-value	ns	ns	ns
rs2227306	TT	1.46 (1.06–1.94)	1.24 (0.96–1.79)	1.27 (0.83–1.89)	0.041
	CT+CC	1.27 (0.92–1.69)	1.09 (0.83–1.51)	0.96 (0.72–1.39)
	P-value	ns	ns	ns
rs2227307	GG	1.46 (1.08–1.87)	1.27 (0.96–1.83)	1.35 (0.84–1.94)	0.011
	TG+ TT	1.25 (0.90–1.77)	1.09 (0.91–1.53)	0.96 (0.67–1.32)
	P-value	ns	ns	ns

Data are presented as median (25th–75th pc) of the forced volume capacity (A) forced expiratory volume in the first second (B) and compared by ANOVA test.

^†Global p-value for GLIM test for repeated measures inter-groups. ns: non-significant

When analysing the inflammatory cytokines no correlation was found between any of the IL-8, CXCR1, CXCR2 polymorphisms and the circulating levels of IL-6, IL-8, IL-16 and TNF-α (p > 0.05, data not shown).

Discussion

The main finding of this study arises in the association found between IL-8 polymorphisms (rs4073, rs2227307, rs2227306) and an affectation beyond pulmonary function and a reduced lung function in the COPD patients during the follow-up period. In this study, we also sought to validate previous findings on genetic variants in the IL-8 and its receptors genes, CXCR1 and CXCR2, with the risk for COPD and related outcomes.

Several studies have investigated the role of IL-8 polymorphisms in chronic inflammatory diseases in different ethnical populations with controversial results (14, 29, 30). Few studies were performed studying the effect of IL-8 polymorphisms and COPD. Four of them performed on Caucasian populations, when studying the distribution of the IL-8 promoter SNP (-251 A/T) failed to detect an association with disease susceptibility, respiratory symptoms or lung function (16, 31–34). In opposite, Shen et al. (2008) found an increased risk of COPD in patients presenting the rs4073 AA genotype in the Chinese population (31). In concordance with these previous findings we did not find an association with any IL-8 variant and COPD susceptibility at least in the studied population.

CXCR1 and CXCR2 genes are both located within the region of linkage to FEV₁/FVC on chromosome 2q (35). Stemmler et al. (2005) identified two polymorphisms in the coding region of the CXCR1 gene that were significantly increased in patients with COPD (16). Shen et al. (2008) found a borderline but not significant association of the +860G/C SNP in the CXCR1 gene in Chinese COPD patients (31). In our study no association was found between either of the two CXCR1 and CXCR2 genes variants and COPD risk susceptibility, nor with any clinical or pulmonary function parameter.

High concentrations of IL-8 were observed in induced sputum from patients with COPD. Tumour necrosis factor-alpha (TNF-α) and IL-8 may be involved in the inflammation in COPD. Increased concentrations of these cytokines were found in patients with COPD compared with smoking and non-smoking control subjects (36). However, in our study, none of the polymorphisms analysed was associated with any of the inflammatory cytokines included. With respect to this result we ­cannot discard that cytokine levels could be altered due to the pharmacological treatment in use in the COPD patients at time of the study.

Interestingly, three polymorphisms in the IL-8 gene significantly associated with the multidimensional (pulmonary and extrapulmonary domains) impairment of the disease expressed by the BODE index. The rs4073A, rs2227306C and rs2227307T alleles were found associated with patients who scored higher in the BODE index. This multidimensional index incorporates four factors: the body mass index (BMI), the degree of airflow obstruction assessed by the FEV₁, the functional dyspnea assessed by the modified Medical Research Council (mMRC) questionnaire, and the exercise capacity assessed by the 6 minutes walking distance (6MWD) test. The BODE index can better represents a heterogeneous disease like COPD and is proved to be a good predictor of poor outcomes, compared to FEV₁ alone (19, 37, 38).

An important finding of this research that deserves to be highlighted is that COPD patients presenting the rs4073A, rs2227306C and rs2227307T alleles in the IL-8 gene showed an important decrease in their lung function, measure by the forced expiratory volume in the first second (FEV₁) and the forced vital capacity (FVC) during the two follow-up years of study. Silverman et al (2002) in a GWAS found significant linkage of FEV₁/FVC to chromosome 2q indicating genetic influencing to the development of airflow obstruction or dysanapsis. They also postulated that regions of linkage to both FEV₁ and FVC may represent genetic determinants of lung size whereas regions of linkage to FEV₁ and FEV₁/FVC, but not to FVC, may represent genetic determinants of airway obstruction (35). The association found between IL-8 variants and FVC in our study was statistically stronger that the one found between the studied SNPs and FEV₁. Anyway, as some genetic regions showed evidence for linkage only to FEV₁ and others to FVC, the chromosome 4q locus may represent mechanical locus, and FVC phenotypes may be more powerful at detection of such loci.

There are some limitations in our study. First, we did not find significant differences in the distribution of the genetic markers studied in COPD cases vs. the control groups, and besides we replicate previous findings, we cannot rule out this could be due to our sample size, thereby generating a β type II error. However, the sample size could be relatively small to evaluate COPD susceptibility but it could be sufficient enough to appreciate differences in the progression of disease. In fact, it has recently been shown that COPD is not always progressive and presents a high variability among patients. Second, since few women were included, the results here reported cannot be extended to both genders. Finally, our findings should be replicated in other cohort. Replication of genotype–phenotype associations is the gold standard to identify genes conferring susceptibility (39). However, the number of supportive studies is not necessarily an indication that a gene is consistently replicated. In fact, excluding SERPINA1 (encoding the alpha-1 antitrypsin protein), none of the other genes are well-proven susceptibility genes for COPD (40).

In summary, in a well-characterized cohort of stable COPD patients, we observed that rs4073A, rs2227306C and rs2227307T alleles in the IL-8 gene appear to be involved in a COPD affectation beyond the lung and a consistent longitudinal worsening of the lung function. Our findings support the limitations in the use of the FEV₁ alone in the assessment of patients with COPD and the importance of their longitudinal follow-up analysis. We did not find association between IL-8, CXCR1 and CXCR2 gene variants and COPD susceptibility or the proposed cytokine serum levels. Further studies should confirm our findings.

Acknowledgments

The authors thank Angela Montejo de Garcini and Candelaria González Díaz for their excellent technical assistance in the pulmonary function tests.

Authors Córdoba-Lanús and Baz-Dávila contributed equally to the study.

Declaration of Interest Statement

The authors report no conflicts of interest. This research was supported by grants of the Fundación Canaria de Investigación y Salud (FUNCIS 55/04), Fondo de Investigación Sanitaria (FIS 06/1144) and Sociedad Española de Patología Respiratoria (SEPAR 2005).

The authors alone are responsible for the content and writing of the paper.