INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Hypersensitivity pneumonitis (HP) represents a group of granulomatous lung disorders that results from repeated inhalation of finely dispersed organic particles including mammalian and avian proteins, fungi, thermophilic bacteria, and certain small-molecular-weight chemical compounds (1). The disease is characterized by a diffuse and predominantly mononuclear cell inflammation of the small airways and pulmonary parenchyma; however, the pathogenic mechanisms involved in the inflammatory lung response have not been completely elucidated. Strong evidence supports a pathogenic role of exaggerated local T cell activation, albeit in the acute clinical presentation the deposition of immune complexes may also participate (1).

Pigeon breeders disease (PBD) is a common form of HP provoked by the inhalation of avian antigens. Patients exposed to proteins from bird droppings and feathers, usually in the domestic environment, develop a subacute or chronic illness and some of them evolve to diffuse pulmonary fibrosis (5).

Interestingly, only a small proportion of individuals exposed to potential HP-causing antigens develops the disease, suggesting that additional host/environmental-promoting factors may play a role (1).

The major histocompatibility complex (MHC) is essential for an appropriate immune response. Particularly, the expression of MHC class II molecules is required for the activation of T lymphocytes by antigen-presenting cells during the immune response triggering. In this context, associations between MHC alleles and a large variety of diseases, mainly those with a presumably immunopathological etiology, have been extensively studied (6). Thus, alleles from loci HLA-DRB1 and HLA-DQB1 have been associated with a number of diseases, primarily those with damage mechanisms related to either autoantibody production or abnormal T cell response (7).

Regarding HP, a few studies using serological techniques suggested that some HLA class II alleles might provide susceptibility to the development of the disease, although results were not conclusive (11). However, polymerase chain reaction (PCR)-based MHC class II typing has not been carried out in this disease.

Genetic polymorphisms of the cytokine tumor necrosis factor- (TNF-), which is encoded within the MHC class III region of chromosome 6, have also been associated with predisposition to disease development (14, 15).

Two polymorphisms (guanine [G] versus adenine [A]) have been identified in the 5' regulatory region of the TNF- gene at positions 308 and 238 (16, 17). More recently, it has been shown that the allele TNF-308.2 (A at 308) is associated with higher constitutive and inducible levels of this cytokine (17).

Few studies on the putative role of TNF- in HP have been performed. It has been reported that serum TNF- was higher in peat moss-processing plant workers who develop HP (18), and that macrophages obtained from bronchoalveolar lavage of patients with farmers lung produced significantly higher levels of this cytokine (19). However, polymorphisms of the 5' promotor region of the TNF- gene have not been previously explored in this disease.

The aim of this study was to determine whether any MHC class II alleles or TNF- promoter alleles were linked with susceptibility to or protection against the development of HP induced by avian antigens.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

Forty-four consecutive patients with biopsy-proved HP were included in this study (34 females and 10 males, 42.2 ± 12 yr [mean ± SD]). The protocol was approved by the Ethical Committee of the National Institute of Respiratory Diseases, and informed written consent was obtained from each subject. Diagnosis of PBD was established according to the following criteria (1, 4, 5): (1) exposure to pigeon preceding the beginning of the disease, and positive serum antibodies to avian antigens, as detected by ELISA (20); (2) progressive dyspnea with partial improvement after avoidance of the avian antigen exposure; (3) clinical, radiological, and functional features of an interstitial lung disease; (4) more than 40% of lymphocytes in the bronchoalveolar lavage; and (5) lung histology compatible with HP (1, 21). All patients enrolled in this study fulfilled the five described criteria.

Tissue samples were taken by open lung biopsy or by videoassisted thoracoscopy, usually 1 wk after hospital admission. None of the patients had been treated with corticosteroids or immunosuppressive drugs at the time of biopsy. Histological evaluation of the tissue samples showed diffuse interstitial inflammation with mononuclear predominance, mainly lymphocytes, and frequent multinucleated giant cells in terminal and respiratory bronchioles, as well as in the alveolar walls. Small and loosely arranged granulomas were observed in the interstitium. Biopsy cultures were negative for bacteria and fungi, and no changes suggestive of other interstitial lung disease (ILD) were found.

Two control groups were studied, both integrated by individuals with the same ethnic origin as the PBD group and with at least two generations born in Mexico, as was the case for PBD patients. The first control group was composed of 99 (54 females, 45 males; 30.5 ± 9 yr old) healthy unrelated Mexican mestizo subjects (HC group) with no family history of HP. These patients were randomly selected from the Mexico City population, and their healthy status was determined by a detailed questionnaire, clinical history, physical examination, and routine blood counts and chemistry. TNF- promoter polymorphisms were determined in 55 of these subjects.

The second control group included 50 exposed but asymptomatic subjects (EAS group). Direct exposure was determined by a detailed questionnaire. They were sequential, unrelated, healthy blood donors from the National Institute of Respiratory Diseases (23 females, 27 males; 29.8 ± 5.1 yr old). The Transfusion Department of this institute serves the geographic area from which HP patients were recruited. This random sample of adult population went through a complete clinical history, physical examination, routine blood counts and chemistries, HIV test, spirometry, and chest X-rays. All individuals in this group were analyzed for HLA-class II and TNF- promoter polymorphisms, and plasma levels of TNF- were also determined. Twelve of them, according to their willingness to participate, were submitted to a bronchoalveolar lavage for cell profile and TNF- analyses. Finally, to confirm the presence of inferred haplotypes in linkage disequilibrium, a family-based study was additionally made in the HP group. For this purpose, four or five direct relatives of patients carrying out these putative haplotypes were examined.

DNA Typing for HLA Class II Alleles

Genomic DNA was extracted from 20 ml of peripheral blood by standard salting out methods (22).

Amplification of Genomic DNA

Generic HLA-DRB1 typing was performed by PCR-SSO reverse dot-blot using the Amplicor Kit (Hoffmann-La Roche, Basel Switzerland). DRB3, DRB4, and DRB5 amplification was done by PCR using Taq polymerase (Promega, Madison, WI) as previously described (23). The primers used for the amplification were DRBAMP-B for the region 3' of the exon 2 in all cases, and DRBAMP-1, DRBAMP-2, DRBAMP-3, DRBAMP-4, DRBAMPB-5, and DRBAMP-52 for the region 5' of the exon 2 for each group of the specific amplification. Primers DQBAMP-A and DQBAMP-B were used to achieve DQB1 amplification. They were synthesized in a DNA-SM automated synthesizer (Beckman, Palo Alto, CA) from information given at the Twelfth International Histocompatibility Workshop (24).

Dot-blot Hybridization

Five percent of the amplified DNA was denatured in 0.4 mol/L NaOH, 10 min, neutralized in 1 mol/L ammonium acetate, and transferred to a Hybond-N membrane (Amersham, Amersham, Bucks, UK). The filters were prehybridized at 42° C for 30 min in a solution containing 6× SSPE (30× SSPE: 4.5 mol/L NaCl, 0.3 mol/L NaH₂PO₄, 30 mmol/L ethylenediaminetetraacetic acid [EDTA], pH 7.4), 5× Denhard solution (2% bovine serum albumin, 2% polyvinylpyrrolidone 40, 2% Ficoll 400), 0.1% lauryl-sarcosine and 0.02% sodium dodecyl sulfate (SDS). Then, oligonucleotide probes labeled with digoxigenin dideoxyuridine triphosphate (Dig-11-ddUTP) were added and hybridized at 42° C for 3 h. Filters were washed twice in 2× SSPE, 0.1% SDS at room temperature for 10 min, once in TMAC solution (50 mmol/L Tris-HCl, pH = 8.0, 3 mol/L tetramethylammonium chloride, 2 mmol/L EDTA, 0.1% SDS) at room temperature for 10 min, and twice at 60° C for 10 min. The dots were revealed using the Dig Nucleic Acid Detection Kit (Boehringer Mannheim Biochemical, Mannheim, Germany).

Oligonucleotide Probes

Information on the sequences and specificities of the DRB1, DRB3, DRB5, and DQB1 oligonucleotides was obtained from the Twelfth International Histocompatibility Workshop (24). The oligonucleotide synthesis was made using the cyanoethyl phosphoramidite technique in a Beckman DNA-SM automated DNA synthesizer, following the manufacturer's protocol.

TNF-238 Polymorphism

TNF- promoter polymorphisms were typed using an ARMS (amplification refractory mutation system)-PCR in a single PCR reaction.

TNF-G primers were 5'-AGACCCCCCTCGGAATCG-3' and 5'-CCGGATCATGCTTTCAGTGC-3'; TNF-A primers were 5'-GCCCCTCCCAGTTCTAG-TTCTATC-3' and 5'-CACACTCCCCATCCTCCCTGGTCT-3'. PCR products obtained with primers had 447 and 209 base pairs (bp), respectively. As a positive control, a DNA segment (608 bp) spanning the mutation site was amplified using primers 5'-GCCCCTCCCAGTTCTAGTTCTATC-3' and 5'CCGGATCATGCTTTCAGTGC-3'. PCR conditions were 5 cycles of 94°, 67°, and 72° C for 60 s each, followed by 25 cycles of 94°, 62°, and 72° C for 60 s each.

TNF-308 Polymorphism

TNF1 (G³⁰⁸) primers were 5'-GGCAATAGGTTTTGAGGGGCGTGG-3' and 5'-AAGCGGTAGTGGGCCCTGCACCTT-3' (216 bp amplified), and for the variant TNF2 (A³⁰⁸) 5'-GCCCCTCCCAGTTCTAGTTCTATC-3' and 5'-ACCCTGGAGGCTGAACCCCGGCCT-3' (139 bp amplified). For the positive control the primers were 5'GCCCCTCCCAGTTCTAGTTCTAT-3' and 5'-AAGCGGTAGTGGGCCCTGCACCTT-3' (354 bp amplified). PCR conditions were 94°, 62°, and 72° C (60 s each) during 30 cycles. The PCR products were visualized by ultraviolet exposure after 2% agarose gel electrophoresis, stained with ethidium bromide (25).

Bronchoalveolar Lavage (BAL)

BAL samples were obtained by a slight modification of the method already described (26). One subsegmental bronchi from the middle lobe was washed with six 50-ml aliquots of sterile saline solution. The recovered fluid was quantified, strained through a surgical gauze to remove mucus and debris, and centrifuged at 400 × g for 10 min at 4° C. Cell pellets were resuspended in 5 ml phosphate-buffered saline (PBS), and total cells were counted in a hemocytometer. Aliquots were fixed in carbowax, and three slides per sample were stained with hematoxylin- eosin, Giemsa, and toluidine blue, and used for differential cell count. Supernatants were stored at 70° C until use.

ELISA

Plasma and BAL TNF-. BAL fluid and plasma TNF- concentrations were evaluated in all PBD patients, and in 9 and 12 subjects from the HC and EAS groups, respectively. Measurements were done using high sensitivity immunoassay kits from R&D Systems Inc. (Minneapolis MN), following the instructions of the manufacturer.

Antibodies to avian antigen. Microtiter plates were coated with 100 ng/well of pooled pigeon sera, and patient sera diluted 1:1,500 in phosphate-buffered saline/bovine serum albumin were added to the plates (100 µl/well). Afterward, a peroxidase-conjugated goat polyclonal antibody against human immunoglobulin G (IgG) diluted 1:10,000 (gamma chain specific; Organon Teknika, Cappel Hill, Durham NC) was added and incubated for 1 h. Finally, for color development orthophenylenediamine plus H₂O₂ was added and the reaction was stopped with 1 N H₂SO₄. Optical density was determined at 490 nm in an ELISA reader (20).

Statistical Analysis

Differences between groups were evaluated through the Mantel-Haenszel chi-square test with Yates correction (27), or the Fisher exact test when appropriate, by using the Epi Info v5.0 statistical program (Stone Mountain, GA). Correction for multiple comparisons was done through the Bonferroni's method. Odds ratios (OR) and 95% confidence intervals (95% CI) were also calculated. Relationship between TNF- and lymphocytes observed in BAL fluids was assessed by using the Pearson's correlation coefficient. Data in the text and tables correspond to mean ± SD.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Demographic comparison of studied populations showed differences in sex and age. The PBD group was predominantly composed of women (p < 0.01), and was older than the control groups (42.2 ± 12 yr versus 29.8 ± 5.1 [EAS] and 30.5 ± 9 [HC] yr; p < 0.01). No differences were noticed in the duration of exposure to birds (63 ± 41 mo in PBD versus 52 ± 37 mo in EAS). The BAL cell profile displayed significant differences between PBD patients and either control group. PBD patients had a noteworthy increment of lymphocytes (72.5 ± 17.8% versus 8.0 ± 5.4% [EAS] and 17.5 ± 4.9% [HC]; p < 0.0001), with a concomitant decrease of macrophages (25.2 ± 17% versus 91.8 ± 5.5% [EAS] and 82 ± 4.9% [HC]; p < 0.0001).

Major Histocompatibility Complex Class II Alleles

Table 1 shows the gene frequency of HLA-DRB1 alleles in PBD patients. An increase in the gene frequency of HLA-DRB1*1305 was observed when compared with both control groups (p < 0.001, OR = 15.4, 95% CI = 3.18-102.6 [HC], and p < 0.001, OR = 17.11, 95% CI = 2.25-357.8 [EAS]). Additionally, a significant decrement in the HLA-DRB1*0802 frequency was noticed in PBD patients compared with HC (p < 0.05, OR = 0.34, 95% CI = 0.11-0.96) and EAS (p < 0.05, OR = 0.26, 95% CI = 0.08-0.78).

Regarding the HLA-DQ locus (Table 2), an increase in the HLA-DQB1*0501 allele frequency was observed in PDB patients compared with HC (p < 0.05, OR = 2.93, 95% CI = 1.21-7.15), and EAS (p < 0.05, OR = 2.96, 95% CI = 1.0- 9.14). In addition, a decreased HLA-DQB1*0402 allele frequency was detected in PBD patients when compared with the EAS group (p < 0.05, OR = 0.38, 95% CI = 0.14-0.96), but not when compared with HC subjects.

Haplotype frequency at the population level is presented in Table 3. HLA-DRB1*1305-DQB1*0301 was significantly increased in PBD patients when compared with HC and EAS (p < 0.005, OR = 8.47, 95%CI = 1.57-60.36, and p < 0.05, OR = 9.88, 95% CI = 1.21-214.85, respectively). There were also three patients with HLA-DRB1*1305 associated to DQB1*0302, a combination not seen in any of the control individuals. This difference reached significance (p < 0.05) when compared with the HC group, but not with the EAS one, probably due to the smaller sample size. HLA-DRB1*0701-DQB1*0302 was also significantly increased in PBD patients when compared with HC patients (p < 0.005, OR = 14.31, 95% CI = 1.67-310.4).

The familial study corroborated the presence of haplotypes HLA-DRB1*1305-DQB1*0301 and HLA-DRB1*0701- DQB1*0302 in patients, as well as in some relatives. In addition, PBD patients displayed a significant decrease of the haplotype DRB1*0802-DQB1*0402 compared with either control group (p < 0.05, OR = 0.34, 95% CI = 0.11-0.96 [HC], and p < 0.05 OR = 0.26, 95% CI = 0.08-0.78 [EAS]).

TNF- Promoter Polymorphisms

The TNF- promoter polymorphism analysis revealed a probable relevance of the TNF-2³⁰⁸ allele in the susceptibility to HP, as it was found in 7 of the 44 PBD patients but only in 1 of the 55 HC subjects and in 1 of the 50 EAS individuals (Table 4, p < 0.05, OR = 10.9, 95% CI = 1.34-236.9, and OR = 9.9, 95% CI = 1.22-215.4, respectively). The genotype analysis showed that six patients were 1/2 heterozygous (p < 0.05 when compared with both control groups). On the other hand, a decrease of the 1/1 genotype (wild-type homozygote) and TNF-1³⁰⁸ allele was observed in the patients group (p < 0.05).

Clinical features were compared among patients with and without the TNF-2³⁰⁸ allele (Table 5). Patients with this polymorphism were significantly younger (33.9 ± 14.6 versus 44.2 ± 10.4 yr; p < 0.05), and were exposed to avian antigens for less time before the beginning of symptoms (15.3 ± 11.1 versus 80.0 ± 77.6 mo, p < 0.05). Likewise, PBD patients having TNF-2³⁰⁸ allele displayed a significantly higher percentage of BAL lymphocytes (88.0 ± 12.1% versus 68.9 ± 17.2%; p < 0.05), which was accompanied by a decrease in alveolar macrophages. Other parameters, including severity of dyspnea, pulmonary function tests, hypoxemia, degree of fibrosis in the lung biopsy, and outcome, were similar (data not shown).

Plasma levels of TNF- were significantly higher in PBD patients compared with either control group (5.89 ± 3.06 pg/ ml versus 1.12 ± 0.28 pg/ml [HC] and 3.23 ± 0.81 pg/ml [EAS], p < 0.05).

TNF- in BAL was undetectable in almost all HC and EAS control groups, whereas it was present in HP patients with an average of 1.04 ± 0.87 pg/ml.

Comparison of TNF- concentrations, either in BAL or plasma, showed no differences between patients exhibiting TNF-2³⁰⁸ and those with other TNF- polymorphisms (0.9 ± 0.6 versus 1.0 ± 0.6 pg/ml in BAL, and 6.8 ± 3.7 versus 5.7 ± 3.0 pg/ml in plasma, Table 5). However, TNF- levels in BAL fluid from PBD patients correlated with the percentage of BAL lymphocytes (r = 0.5, p < 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In spite of the wide spectrum of organic particles capable of provoking HP, the incidence of this disease in the general population is low, suggesting that host susceptibility factors and other environmental conditions are necessary to develop the disease (1). Accordingly, a possible association of pigeon breeder's disease and MHC class II and TNF- polymorphisms was assessed.

To avoid overestimation of an allele or haplotype in PBD patients, two ethnic matched control groups were studied, one of them including subjects exposed to avian antigens but without the disease. In addition, because multiple comparisons may result in spurious associations, we corrected p-values by multiplying for the number of comparisons at each locus.

Our results revealed a significant increase of one HLA-DRB1 allele and one HLA-DQB1 allele, as well as of one haplotype (HLA-DRB1*1305-DQB1*0301) in the PBD patients group, as compared with both control groups. Interestingly, some patients also showed the haplotype HLA-DRB1*1305- DQB1*0302, which differs from the former by the replacement of aspartic acid by alanine at the pocket 9 in the position 57 of the DQB chain (8).

The haplotype HLA-DRB1*0701-DQB1*0302 was also in linkage disequilibrium in the PBD group in relation to the general nonexposed population. It is important to emphasize that in Mexico DRB1*0701 is usually linked to DQB1*0201, whereas DQB1*0302 is commonly linked to DRB1*0407. Therefore, the haplotype DRB1*0701 and DQB1*0302 is virtually nonexistent in the Mexican population (28). Additionally, PBD patients exhibited an important decrease of the allele DRB1*0802, compared with both control groups, and of the allele DQB1*0402, when compared with the EAS group. In agreement with this, similar differences were observed for the haplotype DRB1*0802-DQB1*0402. These findings suggest that these alleles might play a protective role against lung inflammation induced by exposure to organic particles.

In general, our findings suggest that certain MHC class II genes modulate, either increasing or decreasing, the susceptibility to develop HP. Although we do not have any direct evidence that the associated alleles/haplotypes are themselves implicated in the pathogenesis of PBD, the known functions of MHC class II molecules suggest this possibility.

The MHC comprises closely linked genes controlling highly polymorphic proteins involved in the presentation of peptides to the T cell receptor. Specifically, the MHC class II molecules are implicated, among other functions, in the activation of T lymphocytes and transmembrane signaling in response to superantigens (29). Hypersensitivity pneumonitis represents a group of lung disorders mainly characterized by a T cell-mediated delayed-type hypersensitivity reaction to organic dusts. Therefore, an uncontrolled activation of T lymphocytes could be associated with the expression of some peculiar MHC class II alleles/haplotypes. Supporting this point of view, it has been demonstrated that MHC class II genes seem to control susceptibility to experimental HP in mice (30). However, it is plausible that genetic factors other than MHC class II might play a role in determining the host susceptibility for the expression of the disease.

Another highly polymorphic region of the genome is the TNF cluster. Expression levels of TNF- gene have been associated with certain MHC class II alleles (31), and also with the polymorphisms in the TNF gene complex itself. Among them, a single-nucleotide polymorphism located into the gene promoter, as G to A transition located at position 308 define the TNF-1 and TNF-2 alleles, and it has been suggested that the latter is related to increased mRNA transcription in reporter gene constructs (17). Thus, it has been found that TNF-2 homozygous individuals have higher circulating TNF levels than TNF1 homozygous subjects, which in turn has been associated with the development of autoimmune and infectious diseases (32).

In our study, an increase in the promoter allele variant TNF-2³⁰⁸ was observed in the PBD group; almost all of these patients were heterozygous and did not show higher levels of soluble TNF- either in plasma or in the bronchoalveolar lavage, compared with the TNF1 patients. These findings agree with a recent study in which three new single-nucleotide polymorphisms in the TNF- gene promoter were identified. That study demonstrated that neither the novel polymorphisms nor the previously described polymorphisms at positions 238 and 308 had an effect on the gene expression in activated lymphocytes (33). These results suggest that TNF- promoter polymorphisms might serve as markers for neighboring genes encoding for other MHC molecules that may influence disease predisposition and/or severity.

TNF2³⁰⁸ was found in younger patients presenting the disease after less time of pigeon exposure, who also exhibited a higher lymphocytic response in BAL. However, no differences were found with regard to clinical and functional characteristics, morphological findings, and outcome. On the other hand, in this study the increase of the TNF-2³⁰⁸ allele was independent of the alleles/haplotypes in linkage disequilibrium.

TNF- plays a pivotal role in the inflammatory response, and high levels of this cytokine have been found in patients with HP (1, 19, 34, 35). Likewise, TNF- appears to play an essential role in the development of a farmer's lung model in mice (36). Thus, animals receiving the actinomycete Faenia rectivirgula developed HP-like inflammation that was accompanied by an increase in TNF-. Additionally, the use of an antibody against TNF- completely abrogated the development of the disease (36). In the same way, spontaneous regression of HP lesions in mice has been related to changes in natural killer cell activity and lung cell cytokine profile, including the secretion of TNF- (37).

In summary, our findings suggest that an association between some alleles/haplotypes of the MHC class II genes and PBD is present in the Mexican population. Our results also show that the TNF-2³⁰⁸ promoter polymorphism may be a component of the genetic predisposition to HP. The involvement of different genes in various MHC subregions suggests that the lung inflammatory response to avian antigens is modulated by a complex gene interplay rather than by single alleles.