Tumour necrosis factor family genes in a phenotype of COPD associated with emphysema

Chronic obstructive pulmonary disease (COPD), a worldwide leading cause of mortality and morbidity, is a complex trait arising from the effects of environmental risk factors, mostly tobacco smoking, on susceptible individuals. In an attempt to further understand the role of genetic factors accounting for individual COPD susceptibility, a number of gene polymorphisms have been investigated in the last few years, including α₁-antitrypsin-deficient variants in an intermediate status, microsomal epoxide hydrolase, glutathione‐S‐transferase, tumour necrosis factor (TNF) gene complex, and haemoxygenase. However, many of the associations found with COPD are controversial, since they have not been replicated in different populations 1–3.

Studies of TNF gene complex, coding for cytokines relevant to the pathophysiology of COPD 2, are the most abundant reports on this issue available in the literature, and they are a peculiar example of such inconsistent results. It was previously reported that the frequency of allele 2 (TNF-308*2), of the biallelic polymorphism located within the promoter region of the TNF‐α gene (TNF-308), associated with higher levels of TNF production 4, was significantly higher in the Taiwanese with chronic bronchitis 5 and in the Japanese with COPD 6 than in appropriate controls. In marked contrast, the association with COPD was rejected by similarly planned case-control studies in Italian 7 and British 8 populations and by a longitudinal study focused on decline of lung function within the Lung Health Study in Northern Americans 9.

A likely explanation for these inconsistent results lies in genetic heterogeneity among different populations, a feature that is well known in other ubiquitous disorders, such as sarcoidosis 10. However, another critical point in planning a genetic investigation on a complex trait is the phenotype choice, and poor definition of the phenotype and/or heterogeneity of the condition are likely to reduce the power to find significant association 11. COPD is a widely heterogeneous condition 12, 13, including pulmonary emphysema, chronic bronchitis, and bronchial hyperreactivity, often in combination 14. Therefore, it cannot be ruled out that an incomplete definition of the phenotype underlies inconsistency in results of genetic studies on COPD 15 and those dealing with the TNF gene in particular.

Working on this assumption and given the lack of association of allele 2 of the TNF-308 (TNF-308*2) with COPD in another Japanese study 16, and the findings of Keatings et al. 17 who identified a COPD phenotype, i.e. poor prognosis possibly associated with the same allele in the British population, the present authors decided to re-evaluate the role of the TNF family genes in the Italian COPD population. In order to do this, a new COPD population was recruited, characterised by narrower functional inclusion criteria: irreversible airflow obstruction and a diffusing capacity <50% predicted, the latter of which was assumed to be a surrogate marker for pulmonary emphysema. The authors evaluated the frequency of the TNF-308 polymorphism and that of the biallelic polymorphism (LtαNcoI) located in the first intron of the lymphotoxin‐α gene (Lt‐α, previously referred to as TNF‐β), in which allele 1 (LtαNcoI*1) is associated with higher levels of TNF production 18, and compared the results with those obtained in appropriate controls. The investigation was also expanded in the two populations by adding an investigation of two biallelic polymorphisms of the TNF receptors 1 and 2 (TNFR1 and TNFR2, also referred to as p55 TNFR and p75 TNFR, respectively), which are type‐I membrane ligands on the cell surface for both TNF‐α and Lt‐α 19. TNFR1 and TNFR2 genes are also thought to be candidate genes involved in mediating the numerous TNF‐α and Lt‐α effector functions 20.

Subjects and methods

All individuals gave their consent prior to entering the study, which was approved by the Local Ethical Committees of the Institutions involved.

Genetic analysis

Genomic deoxyribonucleic acid (DNA) was extracted from whole blood by standard methods. For detection of TNF-308 and LtαNcoI polymorphisms in COPD and healthy control populations, the same primers 18 and conditions adopted in previous investigations 7 were used. For the TNFR1 exon 1 polymorphism, the polymerase chain reaction (PCR) conditions were as follows: PCR amplifications were carried out in a total volume of 25 µl containing 100 ng genomic DNA, 0.5 unit Thermus Aqiaticus (Taq) DNA polymerase (Laboratoires Eurobio, Les Ulis Cèdex, France), 0.5 µM of each PCR primer, 1.5 mM MgCl₂, 200 µM of each dNTP, 67 mM Tris-HCl pH 8.8, 16 mM (NH₄)SO₄, and 0.01% Tween. Amplification was performed using a Gene Amp system 2400 (Perkin Elmer,Norwalk, CT, USA) with the following conditions: initial incubation at 95°C for 5 min, followed by 35 cycles each of 1 min at 95°C, 1 min at 60°C and 1 min at 72°C, and a final incubation at 72°C for 5 min. The primers used were forward 5′ GAGCCCAAATGGGGGAGTGAGAGG 3′ and reverse 5′ ACCAGGCCCGGGCAGGAGAG 3′ 27. The size of the PCR fragment product was 183 base pairs (bp). Restriction fragment length polymorphism (RFLP) analysis identified a single-nucelotide polymorphism (SNP) at position 36 (A→G), which creates a recognition site for the restriction enzyme MspA1 I (New England Biolabs, Beverly, MA, USA) when the G allele, but not the A allele, is present 27. The G allele digestion generates restriction fragments of 108 and 75 bp from the PCR products (fig. 1⇓).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.—

a) Restriction analysis of tumour necrosis factor receptor (TNFR)1 exon 1 polymorphism resolved in 2% agarose gel electrophoresis of polymerase chain reaction (PCR) amplified genomic deoxyribonucleic acid. A 183 base pair (bp) product (A allele) is restricted by the enzyme MspA1 I to fragments of 108 and 75 bp (G allele). Lane 1: molecular weight marker (GeneRuler 100 bp, Fermentas, Vilnius, Lithuana); lane 2: heterozygous AG; lane 3: homozygous AA; lane 4: homozygous GG. b) Restriction analysis of TNFR2 exon 6 polymorphism resolved in 2% agarose gel electrophoresis. Restriction of a 242 bp PCR product (allele R) with the enzyme Nla III results in fragments of 133 and 109 bp (allele M). Lane 1: homozygous MM; lane 2: homozygous RR; lane 3: heterozygous RM; lane 4: molecular weight marker.

For the TNFR2 exon 6 polymorphism, the PCR conditions were the same as above. Primers used were: forward 5′ ACTCTCCTATCCTGCCTGCT 3′ and reverse 5′ TTCTGGAGTTGGCTGCGTTGT 3′ 27. The size of the PCR fragment generated was 242 base pairs. RFLP analysis identified an SNP at position 196 (T→G), which creates a recognition site for the restriction enzyme Nla III (New England Biolabs) when the T allele, but not the G allele, is present. This polymorphism results in an amino acid substitution (Met→Arg). The 242 base pair PCR product is uncleaved in the 196 G allele (also called the R allele), and cleaved into two fragments of 133 and 109 bp in the 196 T allele (also called the M allele) 28 (fig. 1⇑).

Results

Discussion

Phenotype choice and biological reasonability of candidate genes are crucial prerequisites for a genetic investigation on a complex trait such as COPD 21. All previously published papers, including the authors' 5–8, 16, 17, dealing with the TNF gene family in COPD, have used FEV₁ as an objective parameter to define the phenotype. FEV₁ is strongly related to COPD 15 and predictive of all-cause mortality in population-based studies 31. Nevertheless, FEV₁ is relatively nonspecific, reflecting many different aspects of lung pathology in COPD 32, and it does not help provide a better definition of these patients affected by a widely heterogeneous condition 12–14. In an attempt to narrow the phenotype choice to COPD-associated pulmonary emphysema, a severe reduction of D_L,CO was used as an inclusion criterion. FEV₁ % pred is weakly related to the extent of emphysema 24, 33, 34, suggesting that flow obstruction in severe COPD is not totally accounted for by the extent of emphysema. Reduction of D_L,CO is more strongly correlated with the severity of emphysema, as assessed by high-resolution computed tomography analysis 24, reflecting the reduction of the alveolar-capillary surface, although with some limitations 24, 35.

The biological reasonability of the TNF family as candidate genes in COPD has already been emphasised 2, 36. TNF‐α levels are increased in sputum of COPD patients 37, and weight loss in COPD has been associated with increased levels of circulating TNF‐α 38. The potential implications of the two TNF‐α and Lt‐α polymorphisms investigated have already been extensively discussed 1–3, 5–9. In this article, for the first time, the authors have added the genetic study of two other TNF family members, TNFR1 and TNFR2, biologically reasonable candidates involved in mediating a number of TNF‐α and LT‐α effector functions 19. The soluble form of TNFR1 has been shown to correlate with leptin plasma concentration, in turn related to lower fat mass, in emphysema, but not in chronic bronchitis 39. The TNFR2 exon 6 polymorphism investigated in the present paper was previously found to be associated with human narcolepsy in the Japanese 40, and with systemic lupus erythematosus (SLE) 41 and rheumatoid arthritis in British patients 42, but not with SLE in either Spanish or British populations 28 or idiopathic pulmonary fibrosis 43 in British patients.

In the present paper, the authors provide evidence that the two alleles of the TNF‐α and Lt‐α genes (TNF-308*2 and LtαNcoI*1, respectively), related to higher constitutional production of both cytokines, are not linked to moderate-to-severe COPD associated with emphysema in Italians. When the present study was planned, it was debated whether inconsistent results in different populations with respect to association of TNF-308*2 allele with COPD was due to a bias related to heterogeneity of COPD patients. In fact, the paper by Sakao et al. 6 reported a positive association with Japanese COPD patients with chronic obstruction in whom a history of chronic bronchitis had been excluded, thus suggesting, even if not clearly stated in the paper, that those patients were mostly affected by emphysema. By contrast, a positive association with the TNF-308*2 allele was found in Taiwanese subjects meeting the criteria for chronic bronchitis 5. However, it must be underlined that neither study provided any objective verification of the presence (or absence) of emphysema. The negative study reported by Higham et al. 8 in British subjects did not use D_L,CO as an inclusion criterion, but reported a mean±sem of 75±4% pred in their series of 86 COPD patients, a much less severe impairment than that found in the present series (mean±sd 37±12).

TNF receptor polymorphisms were first studied in COPD in the present paper. The data, however, allows rejection of the hypothesis that they are a major genetic risk factor for COPD, at least that defined in terms of pulmonary emphysema.

In this paper, the authors focused attention on biologically reasonable polymorphic sites, such as the TNF-308 polymorphism of the TNF‐α gene and the LtαNcoI polymorphism of the Lt‐α gene, as well as on a polymorphic site whose gene product is of unknown biological effect, such as TNFR2 exon 6 polymorphism, but that has been reported previously to be associated with a number of disorders 40–42. However, a number of other polymorphic sites are described within the TNF family genes 43, 44 and, therefore, their involvement in COPD cannot be excluded.

To conclude, these findings seem to confirm that the polymorphisms of tumour necrosis factor family genes related to higher constitutional production of cytokines are not major genetic determinants for chronic obstructive pulmonary disease in Caucasians, thus confirming the previously reported negative data in this ethnic group. In spite of the negative results from the investigation, the authors would like to stress the importance of narrowing the phenotype spectrum in a heterogeneous condition such as chronic obstructive pulmonary disease. This concept is further strengthened by two recent, independent reports 45, 46, which dealt with linkage analysis of quantitative spirometric phenotypes and demonstrated that forced expiratory volume in one second, forced vital capacity, and forced expiratory volume in one second/forced vital capacity are influenced by different loci. Another strategy would be to introduce a different phenotype of chronic obstructive pulmonary disease, such as qualitative and quantitative assessment of pulmonary emphysema by high-resolution computed tomography scan.