A Polymorphism in the Tumor Necrosis Factor-α Gene Promoter Region May Predispose to a Poor Prognosis in COPD

A Polymorphism in the Tumor Necrosis Factor-␣ Gene Promoter Region May Predispose to a Poor Prognosis in COPD* Vera M. Keatings, MD; Samantha J. Cave, BSc; Michael J. Henry, MD; Kevin Morgan, PhD; Clare M. O’Connor, PhD; Muiris X. FitzGerald, MD, FCCP; and Noor Kalsheker, PhD

Study objectives: To determine whether the adenine (A)-guanine (G) substitution polymorphism at position - 308 on the tumor necrosis factor-␣ gene confers susceptibility to COPD or to the development of a more severe form of disease. Design: A cross-sectional study was undertaken to compare the frequency of the A allele in a group of 106 patients with COPD with that in a control population (n ⴝ 99). Patients were followed up prospectively for a period of 2 years. Participants and setting: Participants included 106 COPD patients recruited from a respiratory outpatient clinic and 99 control subjects recruited from patients admitted for cardiac catheterization. Measurements and results: DNA was extracted from venous blood, and each subject was genotyped for the polymorphism by polymerase chain reaction amplification and restriction digestion using Nco1. There was no increased frequency of the A allele in patients compared to control subjects. AA homozygous patients had less reversible airflow obstruction (p < 0.05) and a significantly greater mortality (both all-cause and respiratory deaths) on follow-up (p < 0.001), despite a shorter cigarette smoking history. Conclusions: This study suggests that homozygosity for this A allele predisposes to more severe airflow obstruction and a worse prognosis in COPD. (CHEST 2000; 118:971–975) Key words: COPD; polymorphism; susceptibility; tumor necrosis factor-␣ Abbreviations: A ⫽ adenine; bp ⫽ base pairs; G ⫽ guanine; IHD ⫽ ischemic heart disease; p ⫽ wild-type; q ⫽ mutant; TNF ⫽ tumor necrosis factor

is the sixth-leading cause of death in the C OPD Western world, accounting for much morbidity and mortality,1 and it is estimated that this will rise to the third-leading cause by the year 2020.2 The major environmental etiologic factor is cigarette smoking. However, as only 15% of cigarette smokers develop clinically significant airway obstruction, host factors

*From the Department of Medicine and Therapeutics (Drs. Keatings, Henry, O’Connor, and FitzGerald), University College Dublin, Ireland; and Division of Clinical Chemistry, School of Clinical Laboratory Sciences (Ms. Cave, Drs. Morgan and Kalsheker), Queens Medical Center, University of Nottingham, Nottingham, UK. Supported by EU Biomed 2 programme, grant No. BMH4CT96 – 0152, and Irish Thoracic Society Travelling Fellowship 1998. Manuscript received September 22, 1999; revision accepted April 5, 2000. Correspondence to: Vera M. Keatings, MD, Department of Medicine, St Vincent’s Hospital, Elm Park, Dublin 4, Ireland; e-mail: [email protected]

appear to play an important part in determining individual susceptibility to the development of disease. Apart from the contribution by the wellrecognized ␣1-proteinase inhibitor deficiency, the genetic factors predisposing to COPD are unknown. It has been shown that smokers develop inflammation dominated by neutrophilic influx in the respiratory tract and that in COPD, the degree of neutrophilic inflammation present is directly proportional to disease severity as assessed by FEV1.3,4 In addition to cellular influx, there are elevated concentrations of the cytokines tumor necrosis factor (TNF)-␣ and interleukin-8 in the airways of patients with COPD. Smokers who do not develop COPD also have elevated numbers of neutrophils and concentrations of interleukin-8 in the airways, suggesting that these are a reflection of smoking per se rather than disease. In healthy smokers, however, the cytokine TNF-␣ is not significantly elevated, suggesting that the presCHEST / 118 / 4 / OCTOBER, 2000

971

ence of increased concentrations of this cytokine may distinguish the inflammation in subjects with COPD from that due to cigarette smoking.4 Several studies indicate that TNF-␣ can induce inflammatory responses that are associated with pathogenicity in COPD. Inhalation of TNF-␣ by humans has been shown to increase the numbers of neutrophils in the airway.5 TNF-␣ also enhances extracellular elastolytic activity by adherent neutrophils.6 Increased proteolytic activity has been considered for many years to be important in the development of emphysema,7 and the enhancement of proteolytic activity by TNF-␣ could contribute to disease progression in COPD. In vitro studies examining the effect of cigarette smoke exposure on alveolar macrophages show a reduction in TNF-␣ production in response to shortterm exposure8 and no increase in production following long-term exposure. In addition, alveolar macrophages from smokers release less TNF-␣ when stimulated with lipopolysaccharide, suggesting that the normal effect of cigarette smoke on alveolar macrophages is to depress the production of TNF␣.9 The finding of normal levels of TNF-␣ in healthy smokers would therefore be in keeping with in vitro studies. In patients with COPD, however, the concentration of TNF-␣ is elevated, leading to our hypothesis that those patients who have an abnormally elevated production of TNF-␣ are more likely to develop COPD following exposure to cigarette smoke. Given that the smokers with no lung disease do not have elevated concentrations of TNF-␣, it is possible that the production of increased quantities of TNF-␣ by alveolar macrophages in the face of cigarette smoking may confer susceptibility to the development of COPD both by increasing neutrophil chemotaxis and increasing the elastolytic capacity of the neutrophils. In recent years, a genomic polymorphism resulting in substitution of the nucleotide adenine (A) for guanine (G) at position - 308 within a regulatory region of the TNF-␣ locus has been discovered.10 Possession of the A substitution has been shown to result in increased binding of nuclear factors and enhanced transcription of the gene.11,12 More recently, it has been demonstrated that, when stimulated with lipopolysaccharide, blood cells from individuals heterozygous for the A allele produce greater quantities of TNF-␣ than from those with GG genotype.13 This study did not include any individuals homozygous for the polymorphism. Our hypothesis was that possession of the A allele, by increasing TNF-␣ production, may confer either an increased susceptibility to the development of COPD in smokers or to the development of more severe disease. 972

Materials and Methods One hundred six volunteers with stable COPD and 99 control subjects were recruited. The study was approved by St Vincent’s Hospital Ethics Committee, and all subjects gave informed consent. Patients with COPD had stable disease and were recruited from the respiratory outpatient clinic at St Vincent’s Hospital, Dublin. Inclusion criteria for patients with COPD were as follows: stable disease, FEV1 ⬍ 70% predicted, FEV1/FVC ⬍ 70%, smoking history ⬎ 20 pack-years, and improvement in FEV1 following inhalation of 200 ␮g albuterol ⬍ 10% of baseline FEV1. Patients were excluded if they had had an exacerbation of COPD within the previous 6 weeks. Nondisease control subjects were subjects with no history of lung disease who were recruited from the cardiology outpatient clinic and from patients admitted to St Vincent’s Hospital for coronary angiography. Inclusion criteria for the control group were as follows: FEV1 ⬎ 85% predicted, FEV1/FVC ⬎ 70%, and smoking history ⬎ 20 packyears. Subjects were excluded if they had any respiratory disease, including asthma. The patients were recruited from among cardiology patients, as this group of patients were likely to be similar in age, sex, and smoking history to the COPD group. Assessment of Disease Severity Comparison of population genotype frequencies in a crosssectional manner alone is an unreliable method of detecting susceptibilities to disease, as the results may be distorted due to survival bias within the control group. We therefore examined indicators of disease severity and prognosis in the patients with COPD. Patients’ smoking history, age, FEV1, FVC, and reversibility of FEV1 following inhalation of a ␤2-agonist were recorded. Two years after the study was started, patient follow-up mortality data were obtained from all the AA (homozygous for the polymorphic allele) patients, the first 44 of the 62 GG (homozygous for the common allele) patients, and the first 24 of the GA (heterozygous) patients. This was done in each case by locating hospital files and by contacting patients’ general practitioners. In many cases, it was not possible to ascertain the cause of death, so all-cause mortality was used as the end point. Polymerase Chain Reaction and Restriction Digestion All patients’ samples were genotyped as using a modification of a previously published method.9 Genomic DNA (100 to 200 ng), isolated from whole blood using the QIAgen QIAamp kit (QIAgen Ltd; Crawley; West Sussex, UK), was amplified using 35 pmol of the primers 5⬘AGGCAATAGG TTTTGAGGGGCAT3⬘and 5⬘AGTTGGGGACACACAAGC ATCA3⬘ in a total volume of 50 ␮L containing 1 U Taq DNA polymerase (Helena BioSciences; Sunderland; Tyne & Wear, UK), deoxynucleoside triphosphates (final concentration, 200 ␮M), and the reaction buffer of the manufacturer, containing 1.5 mM magnesium chloride. The amplification parameters were 94°C for 3 min, 60°C for 1 min, and 72°C for 1 min; followed by 35 cycles of 94°C for 1 min, 60°C for 1 min, 72°C for 1 min; with a final cycle of 94°C for 1 min, 60°C for 1 min, and 72°C for 5 min. The reactions were then held at 4°C. The products were examined on agarose gels (2%) run in 1⫻TAE buffer (40 mM Tris acetate, 1 mM ethylene diamine tetraacetic acid) to ensure successful amplification and that a “no DNA” control was blank. The products were then digested using Nco1 at 37°C for a minimum of 2 h and analyzed by polyacrylamide gel electrophoresis with 10% gels run in 1⫻TBE buffer (89 mM Tris, 89 mM boric acid, 2 mM ethylene diamine tetraacetic acid). The A allele that does Clinical Investigations

Table 1—Characteristics of Patient and Control Groups* Groups

No.

Age, yr

Sex, No.

FEV1, % Predicted

FVC, % Predicted

Pack-yr

COPD Control

106 99

66.71 (6.57) 58.59 (5.91)

59M 61M

45.24 (4.46)† 89.25 (8.91)

65.80 (6.48)† 93.52 (9.35)

43.72 (4.31) 31.67 (3.10)

*Data are presented as mean (SEM) unless otherwise indicated. M ⫽ male. †p ⬍ 0.05; all other data did not differ significantly between patient and control groups.

not contain a Nco1 restriction site gave a band of 151 base pairs (bp), whereas the G allele cuts to yield fragments of 131 bp and 20 bp. The 151/131 bp fragments are readily distinguished on 10% polyacrylamide; the 20 bp fragment is usually lost at the bottom of the gel. Statistical Analysis Genotype frequencies and mortality between groups were compared using Fisher’s Exact Test, and quantitative variables were compared using the analysis of variance. An unpaired t test was used to compare smoking histories between AA genotype and the other patients. The null hypothesis was rejected at p ⬍ 0.05. To assess Hardy-Weinberg equilibrium in COPD and control groups, the expected genotype frequencies were calculated for wild-type (p) and mutant (q) allele frequencies as follows: p2 ⫹ 2pq ⫹ q2. Previous studies have suggested marked differences in genotype frequencies between control and patient groups.14 Based on observed frequencies, power calculations suggest that a 20% difference in genotype frequencies would require 200 samples to have a 80% probability of detecting a difference at p ⫽ 0.05.

Results Patient and control characteristics are shown in Table 1. There was no statistical difference in age or smoking history between groups. There was no difference in the frequency of the GG and GA genotypes in the COPD population compared with control subjects (Table 2). Six patients in the COPD and three patients in the control group were AA homozygous, but this was not statistically significant. The genotype frequencies in patient and control groups did not deviate from Hardy-Weinberg equilibrium. To determine if possession of the A allele predisposed to development of a more severe form of

Table 2—Genotype Frequencies in Patient and Control Groups* TNF-␣-308 Genotype GG GA AA Total *NS ⫽ not significant. †Analysis of variance.

COPD

Controls

p Value†

62 38 6 106

59 37 3 99

NS NS NS

disease, clinical indexes of disease severity were examined according to patient genotype. There was no difference in the degree of reduction of FEV1 according to genotype, but AA-homozygous patients had less reversibility of airflow obstruction following a ␤2-agonist despite a reduced smoking history (Table 3). Follow-up Mortality Data The six homozygous AA subjects and the groups of patients with GA and GG genotypes were followed up for a period of 22 to 23 months (Table 3). Patients homozygous for the polymorphism had a significantly greater all-cause mortality: 4 of 6 patients having died at the time of follow-up, compared with 2 of 24 heterozygous patients and 1 of 44 patients with GG genotype (p ⬍ 0.001). The causes of death were as follows: of the four AA-homozygous patients, two died from acute exacerbations of COPD, one from lung cancer, and one from heart failure secondary to ischemic heart disease (IHD); the two deaths in the heterozygote group were from acute exacerbations of COPD (one secondary to pneumonia), and the patient with GG homozygosity died from an intracerebral hemorrhage. Analyzed separately for deaths due to respiratory failure only, the difference between groups remained significantly higher in the AA group (p ⬍ 0.05). Mortality data for the control group were not available.

Discussion This study demonstrates that there is no difference in frequency of the A and G alleles in patients with COPD compared to control subjects. The patients with AA homozygosity, however, had similar degree of airways obstruction despite a lower total cigarette smoking history. In addition, after almost 2 years of prospective follow-up, the patients with AA homozygosity had a significantly greater mortality than the heterozygote or GG homozygous groups. The major etiologic factor for the development of COPD is cigarette smoking, and apart from the rare ␣1-proteinase deficiency,15 the genetic factors that predispose to the development of this condition are CHEST / 118 / 4 / OCTOBER, 2000

973

Table 3—Indexes of Disease Severity and Mortality in COPD Patients Grouped According to Genotype* Genotypes GG GA AA

Age, yr

FEV1 % Predicted

FEV1 Revers†

Pack-yr

Follow-up Period, mo

Mortality‡

65.4 (1.3) 68.7 (1.3) 67.2 (5.2)

46.98 (2.1) 42.79 (2.4) 43.17 (6.8)

5.67 (0.6) 6.03 (0.8) 0.83 (0.8)§

47.18 (2.7) 40.47 (3.7) 29.17 (4.0)㛳

22 23 23

1 (0)/44 2 (2)/24 4¶ (2§)/6

*Data are presented as mean (SEM) unless otherwise indicated. †FEV1 Revers ⫽ Percent improvement in FEV1 from baseline following inhaled albuterol, 200 ␮g. ‡Total deaths (respiratory deaths)/number followed up. §p ⬍ 0.05 compared to other genotypes. 㛳p ⬍ 0.05 compared to GG and GA genotypes combined. ¶p ⬍ 0.001 compared to other genotypes.

not known. Increased concentrations of the proinflammatory cytokine TNF-␣ are present in the lungs of patients with COPD but not in nondiseased smokers. It was our hypothesis that patients with this TNF-␣ - 308 “A” polymorphism would be more susceptible to the development of COPD through the production of increased concentrations of TNF-␣ in response to a variety of insults (eg, cigarette smoke and infection); basing the study on this hypothesis, the presence of increased concentrations of this cytokine in the airways would accelerate the progression of COPD. This would result in patients with the polymorphism having more severe disease and a worse prognosis. Our results are in keeping with this hypothesis. The highly significant increase in mortality observed in patients homozygous for the polymorphism is strongly suggestive of an association between the AA genotype and poor prognosis. Patients with the AA genotype also displayed a significantly reduced reversibility in FEV1 following inhaled albuterol. In a study examining prognostic factors in ⬎ 900 patients with COPD, it was demonstrated that the FEV1 following a bronchodilator was a better predictor of survival than the initial FEV1,16 suggesting that the degree of “fixed” airway obstruction is what determines prognosis. Thus, reduced reversibility in patients with the AA genotype also indicates an association between this polymorphism and susceptibility to more severe disease in COPD. It was of note that AA patients had a lower smoking history, compared to the other two genotypes combined (p ⬍ 0.05). If the susceptibility to develop COPD was the same between groups, one would expect that this group would have less severe airflow obstruction; however, this group had a similar degree of airflow obstruction despite a history of reduced cigarette exposure. There was no increase in frequency of the polymorphism in patients with COPD compared with control subjects, but since this part of the study was 974

cross-sectional, a true increase in frequency may have been missed due to survival bias in the GA and GG groups. A previous study13 in a Taiwanese population demonstrated an association of the G allele with chronic bronchitis. However, in the Taiwanese study, it is difficult to explain how an allele known to result in reduced TNF-␣ concentrations can be associated with chronic bronchitis. Our study is consistent with other studies in the literature in that the frequency of the less common A allele in white populations ranged from 16 to 27%,17,18 and we have observed a frequency of 24%. This suggests that there may be significant variation in the frequency of this polymorphism between white and Taiwanese populations, as these authors report a frequency of 94.9% in their control population. These authors also suggest that there is an increased concentration associated with the G allele, but this is not consistent with the published literature in which the A allele has been shown to result in increased TNF-␣ concentrations. The TNF-␣ gene is located on the major histocompatibility complex, and the possibility that the association with more severe disease and increased mortality could be explained by linkage disequilibrium between this and another allele should be considered. A search for candidate polymorphisms of the TNF-␣ gene in COPD may reveal other functional mutations, which could lead to disease. This study demonstrates an association between the TNF-␣ - 308 polymorphism and increased mortality due to COPD. This association should further be studied comparing the functional responses of alveolar macrophages in patients with and without the polymorphism. This is beyond the scope of the present study and hampered by the presence of only two survivors with the AA polymorphism, this study is not yet feasible and a prospective search for larger number of patients is underway. Most of our control subjects were recruited from Clinical Investigations

among those admitted to the hospital for coronary angiography and therefore likely to have IHD. The frequency of this polymorphism in patients with IHD compared to a completely disease-free population is not known. From within a hospital setting, it would be extremely difficult to recruit individuals completely free of disease, so it was believed that a suitable age- and cigarette smoking-matched population would be recruited from among those admitted for coronary angiography. The identification of functional polymorphisms such as this one as factors that determine disease susceptibility or severity is an important step in pointing to mechanisms of the inflammatory response in diseases such as COPD, allowing the development of more specific therapeutic interventions. The identification of individuals in whom smoking may lead to more severe disease could also lead to more targeted antismoking intervention leading to prevention of disease. In summary, our data demonstrate that in COPD, those homozygous for the less common allele (A) of the TNF-␣ - 308 polymorphism have significantly greater all-cause mortality (largely due to an increase in respiratory deaths) over a 22- to 23-month period of follow-up, and evidence of more advanced disease, despite a reduced smoking history. This genetic polymorphism may be an important susceptibility marker for patients at risk of more severe disease, and further large independent studies should be undertaken to establish whether this observation can be confirmed. ACKNOWLEDGMENT: The authors also thank Drs. W.T. McNicholas, B. Maurer, and P. Quigley for permission to recruit patients under their care.

References 1 Murray CJL, Lopez AD. Alternative projections of mortality and disability by cause 1990 –2020: global burden of disease study. Lancet 1997; 349:1498 –1504 2 Manfreda J, Mao Y, Litven W. Morbidity and mortality from chronic obstructive pulmonary disease. Am Rev Respir Dis 1989; 140(suppl):S19 –S26 3 Thompson AB, Daughton D, Robbins RA, et al. Intraluminal airways inflammation in chronic bronchitis: characterization

4

5 6

7 8

9 10 11 12

13

14 15 16 17 18

and correlation with clinical parameters. Am Rev Respir Dis 1989; 140:1527–1537 Keatings VM, Collins PD, Scott DM, et al. Differences in interleukin-8 and tumor necrosis factor-␣ in induced sputum from patients with chronic obstructive pulmonary disease and asthma. Am J Respir Crit Care Med 1996; 153:530 –534 Thomas PS, Yates DH, Barnes P. TNF␣ alters human bronchial reactivity and induces inflammatory cell influx. Am Rev Respir Dis 1995; 152:76 – 80 Chamba A, Afford SA, Stockley RA, et al. Extracellular proteolysis of fibronectin by neutrophils: characterization and the effects of recombinant cytokines. Am J Respir Cell Mol Biol 1991; 4:330 –337 Tetley TD. New perspectives on basic mechanisms in lung disease: 6. Proteinase imbalance; its role in lung disease. Thorax 1993; 48:560 –565 Higashimoto Y, Shimada Y, Fukuchi, Y, et al. Inhibition of alveolar macrophage production of TNF␣ by acute in vivo and in vitro exposure to cigarette smoke. Respiration 1992; 59:77– 80 Mc Crea KA, Ensor JE, Nall K, et al. Altered cytokine regulation in the lungs of cigarette smokers. Am J Respir Crit Care Med 1994; 150:696 –703 Wilson AG, di Giovine FS, Blakemore AIF, et al. Single base polymorphism in the human TNF␣ gene detectable by Nco1 restriction of PCR product. Hum Mol Genet 1992; 1:353 Kroeger KM, Carville KS, Abraham LJ. The -308 tumor necrosis factor-␣ promoter polymorphism effects transcription. Mol Immunol 1997; 34:391–399 Wilson AG, Symons JA, McDowell TL, et al. Effects of a polymorphism in the human tumor necrosis factor-␣ promoter on transcriptional activation. Proc Natl Acad Sci USA 1997; 94:3195–3199 Louis E, Franchimont D, Piron A, et al. Tumor necrosis factor (TNF) gene polymorphism influences TNF␣ production in lipopolysaccharide (LPS)-stimulated whole blood cell culture in healthy humans. Clin Exp Immunol 1998; 113:401– 406 Huang SL, Su CH, Chang SC. Tumor necrosis factor-␣ gene polymorphism in chronic bronchitis. Am J Respir Crit Care Med 1997; 156:1436 –1439 Laurell CB, Eriksson S. The electrophoretic alpha 1-globulin pattern of serum in alpha 1-antitrypsin deficiency. Scand J Clin Invest 1963; 15:132–140 Anthonisen NR, Wright EC, Hodgkin JE, et al. Prognosis in chronic obstructive pulmonary disease. Am Rev Respir Dis 1986; 133:14 –20 Messer G, Kick G, Ranki A, et al. Polymorphism of the tumor necrosis factor genes in patients with dermatitis herpetiformis. Dermatology 1994; 189(suppl 1):135–137 Wilson AG, de Vries N, Pociot FS, et al. An allelic polymorphism within the tumour necrosis factor-␣ promoter region is strongly associated with HLA A1, B8 and DR3 alleles. J Exp Med 1993; 177:557–560

CHEST / 118 / 4 / OCTOBER, 2000

975