INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Sarcoidosis is a chronic granulomatous disease of unknown etiology, especially involving the lungs. Histopathologically the disease is characterized by the accumulation of activated CD4+ T-lymphocytes and mononuclear phagocytes at disease sites (1), and the formation of granulomas. These granulomas may remit spontaneously or progress to result in scarring and thus interfere with organ function. In patients with lung involvement, roughly 30% progress to significant fibrosis. The variable outcomes of the disease, coupled with the observance of familial clustering, suggest genetic factors may be involved in determining a person's susceptibility to the disease and its subsequent outcome (2).

We previously described an association between this disease and the presence of HLA-DPB1 alleles containing a glutamic acid residue at position 69 (Glu69+) within the UK Caucasoid population (3), a finding in agreement with previous associations described between human leukocyte antigens (HLA) of the major histocompatibility complex (MHC) and sarcoidosis, including HLA-DR17 in Scandinavians (4, 5) and HLA-B8 in British, Italian, and Czech populations (6, 7). More than one third of our patients with sarcoidosis studied were negative for Glu69+ HLA-DPB1 alleles, suggesting the possibility that the association of these DPB1 alleles with sarcoidosis is attributable to linkage disequilibrium with another susceptibility gene within the MHC genetic region. A recent immunogenetic study among African Americans showed associations between HLA-DPB1 position 36 Val and position 55 Asp residues and sarcoidosis, but no association between HLA-DPB1 position 69 residues (8), suggesting that the association we observed previously is not necessarily matched in other ethnic populations but may still be of relevance for some patients with sarcoidosis.

The role of the MHC genes in presenting antigen to T-cells leading to their activation makes them good candidates for involvement in sarcoidosis in which the T-cell activation and accumulation is an essential early event in the disease process. Additional non-MHC genes such as TAP1 and TAP2 encoding the transporter associated with antigen processing (TAP), which participate in the antigen processing pathways prior to its presentation, are also interesting candidates for involvement.

The TAP1 and TAP2 genes located within the Class II MHC region function in the endogenous antigen-processing pathway for subsequent presentation of peptide in association with MHC Class I molecules (9). The two genes encode subunits of a heterodimeric complex that is inserted into the membrane of the endoplasmic reticulum, allowing the transport of antigenic peptide fragments, generated by the proteasome, from the cytosol into the lumen of the endoplasmic reticulum in an ATP-dependent manner. Limited polymorphisms have been described in the coding regions of human TAP1 and TAP2, and these polymorphisms have been investigated in a number of MHC-linked diseases, including ankylosing spondylitis, arthritis, and insulin-dependent diabetes mellitus (10). The aim of this study was to investigate the association between TAP1 and TAP2 gene variations and sarcoidosis in clinically well-defined groups of subjects from two different ethnic populations. By studying two ethnically distinct populations we hoped to determine whether differences in these antigen processing genes contribute to sarcoidosis susceptibility across ethnic boundaries. The apparent failure to identify genetic risk factors for sarcoidosis that are consistent among several ethnic populations (2) suggests that the use of multiple clearly defined ethnic populations may be a useful strategy when performing studies of this disease. Genetic polymorphism within the MHC Class II locus HLA-DPB1 appeared to contribute to sarcoidosis susceptibility in the UK population in our previous study (3). We therefore also examined the relationship between HLA-DPB1 alleles and TAP gene variation in the UK patient and control populations to determine whether any disease-associated variants at the TAP loci were in fact part of an extended haplotype involving the HLA-DPB1 locus.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients and Control Subjects

Two populations of ethnically matched patients and control subjects were studied. The first population comprised of 117 unrelated UK Caucasian patients with sarcoidosis and 290 healthy UK control subjects. The ratio of female to male was 1:1.17 in the UK patients and 1:1.13 in the UK control subjects. Thirty-eight of the UK patients had been included in our previous HLA study (3). The UK control population consisted of sequential unrelated Caucasian cadaveric organ donors from the Oxford Tissue Typing Centre (Churchill Hospital, Oxford). The second population examined comprised 87 unrelated Polish Slavonic patients with sarcoidosis and 158 healthy Polish control subjects. The ratio of female to male was 1:0.85 in the Polish patients. The Polish control group comprised male medical students recruited from the Medical University of Gdansk. In all patients with sarcoidosis clinical features of pulmonary sarcoidosis were observed (13), and the diagnosis of sarcoidosis was confirmed by histiologic findings of noncaseating granulomas from lung biopsy or by Kveim test. Project approval was granted by the ethics committee of the Royal Brompton Hospital.

Genomic DNA Preparation

Genomic DNA extraction performed on blood samples obtained from patients and control subjects used a standard high salt method described previously (14). Briefly, whole blood, collected into 5% Na₂EDTA (BDH Chemicals, Poole, Dorset, UK) in PBS (pH, 7.5) and stored at 20° C, was thawed and washed with TE buffer (10 mM TRIS base; Sigma Ltd., Poole, Dorset, UK), 1 mM Na₂EDTA at pH 8.0. The mixture was centrifuged and the red blood cells were lysed by incubation with a swelling buffer (10 mM TRIS-HCl at pH 7.4, 20 mM NaCl, and 5 mM MgCl₂) at 4° C for 20 min before centrifugation (800 g at room temperature). The resultant pellet was then incubated in red cell and nuclei lysis buffers. Finally, the DNA was precipitated with 5 M NaCl (1/10th volume) and two volumes of absolute ethanol. The DNA was washed with 70% ethanol before resuspension in TE. The concentration of the DNA was determined by measuring the ultraviolet absorption at 260 nm.

ARMS-PCR Analysis of the TAP Genes Polymorphism

An amplification-refractory mutation system polymerase chain reaction (ARMS-PCR) method was used to type two dimorphisms in the TAP1 gene corresponding to amino acid positions 333 and 637, and three dimorphisms in the TAP2 gene corresponding to amino acid positions 379, 565, and 665. The variable amino acids produced by these polymorphisms are isoleucine (Ile) or valine (Val) at TAP1 position 333, aspartic acid (Asp) or threonine (Thr) at TAP1 position 637, isoleucine or valine at TAP2 position 379, and alanine (Ala) or threonine at TAP2 positions 565 and 665 (Table 1). The sequences of the oligonucleotide primers used have been described previously (15). Genomic DNA samples (0.25 µg) were amplified in 25 µl reaction mixtures containing 0.25 µg of each oligonucleotide primer, 200 µM dNTPs, 1× Taq DNA polymerase buffer, and 0.5 units of Taq DNA polymerase (Boehringer Mannheim, Mannheim, Germany) overlaid with mineral oil. Reaction conditions, using a thermal cycler (Techne, Cambridge, UK) were: 95° C for 3 min; 35 cycles of 94° C for 1 min, 58° C for 2 min; 72° C for 2 min; 72° C for 10 min. Reaction products were separated on a 2% agarose gel containing 0.1% wt/vol ethidium bromide (Sigma Ltd.) before photography.

TAP Genotype Interpretation and Analysis

A typical result of genotyping the five biallelic TAP polymorphisms in a single DNA sample is shown in Figure 1. The upper band in each lane is a control product generated by primers flanking the polymorphic site of interest; these provide an "internal" control for each reaction and vary in size (400 to 533 bp) because they are site-specific.

View larger version (110K): [in this window] [in a new window]   Figure 1.   Representative ethidium-bromide-stained agarose gel showing ARMS-PCR products for all five polymorphic sites in one DNA sample. Lanes 1 and 2 represent TAP1 positions 333 and 637; lanes 3-5 represent TAP2 positions 379, 565, and 665, respectively. Lanes 1-5 products consist of one control band and one or two allele-specific differently sized smaller fragments, which define homozygosity (lanes 1-4) or heterozygosity (lane 5) for a particular polymorphic site. Lane 1: 533 bp control band and a 241 bp isoleucine-specific band. Lane 2: 429 bp control band and a 307 bp aspartic acid-specific band. Lane 3: 427 bp control band and a 328 bp valine-specific band. Lane 4: 400 bp control band and a 298 bp alanine-specific band. Lane 5: 408 bp control band, a 326 bp alanine-specific band and a 141 bp threonine-specific band. Lane M: 100 bp spaced molecular weight marker.

The TAP1 and 2 variant amino acids detected by ARMS-PCR are shown in Table 1. In a number of previous studies the results of TAP typing by ARMS-PCR have been reported as allele frequencies (11, 12), using combined data from individual variant sites to give an overall TAP1 and TAP2 allele. A drawback of such an approach is that if the TAP genes are found to be heterozygous at more than one of the variable positions, the exact assignment of alleles is not possible. Because as much as 25% of samples are ambiguous with regard to clearly defining TAP1 or 2 alleles because of such heterozygosity, the results may be interpreted more clearly when expressed as genotype or derived amino acid frequencies (11).

SSP-PCR Analysis of the HLA-DPB1 Locus

To characterize the HLA-DPB1 alleles present in the UK control and patient samples a Sequence Specific Primer (SSP-PCR) method was used as described previously (16).

Statistical Analysis

Statistical analysis was performed on the individual polymorphism results for the two populations studied using chi-square contingency table analysis with the appropriate number of degrees of freedom (df). A stepwise analysis approach was used. Initially phenotype frequencies were compared using a two-by-two chi-square test with 1 df. Where significant differences between case and control phenotype frequencies were observed, determined by a p value of less than 0.05, overall genotype frequencies were compared with the use of a three-by-two chi-square test with 2 df, and where these differed significantly, the frequency of individual genotypes were compared.

In case control studies where the disease is rare in the general population from which cases and control subjects are selected such as sarcoidosis, the odds ratio may be used as an approximation of the relative risk. Odds ratios were calculated for the phenotypes and individual genotypes, which were found to differ significantly.

The UK population data for TAP and HLA-DPB1 were analyzed using the chi-square test by the computer package Knowledge Seeker 4 (ANGOSS Software, Guildford, UK). This allowed associations between different alleles to be analyzeda strong association giving good evidence that these alleles form a haplotype (16).

Our UK patient and control population sizes allowed us to detect with 90% power a difference in phenotype or genotype frequency of 15% (p = 0.05). For our Polish patients and control subjects a 15% difference in phenotype or genotype frequency would be detected with 73.5% power (p = 0.05). If an ethnic independent genetic effect was present, the combined patients and control subjects would, theoretically, detect a 15% increase in exposure within the patient group, with 99% confidence and 95% power.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

A comparison of TAP1 and TAP2 phenotype and genotype frequencies in the UK study population is shown in Table 2, and Table 3 compares the TAP results for the Polish subjects.

UK Population Results

For TAP2 there were significant differences in phenotype frequency at position 565 and position 665. At position 565, the frequency of those positive for the threonine variant was reduced in the patients (odds ratio, 0.40; 95% confidence interval, 0.18 to 0.85). A significant difference was also seen in the overall genotype frequencies at this position (² = 6.86 with 2 df, p = 0.03). Comparison of the three individual genotypes at this locus showed the Ala/Ala genotype to be overrepresented among the patients with sarcoidosis (odds ratio, 2.5; 95% confidence interval, 1.18 to 5.45), whereas the Ala/Thr genotype was underrepresented among the patients (odds ratio, 0.41; 95% confidence interval, 0.19 to 0.87).

At TAP2 position 665, the phenotype frequency of subjects positive for the alanine variant was reduced (odds ratio, 0.51; 95% confidence interval, 0.32 to 0.81). A significant difference was also seen in the overall genotype frequencies at this position (² = 9.01 with 2 df, p = 0.01). Comparison of the three individual genotypes at this locus showed the Thr/Thr genotype to be overrepresented among the patients with sarcoidosis (odds ratio, 1.96; 95% confidence interval, 1.23 to 3.13), whereas the Thr/Ala genotype was underrepresented among the patients (odds ratio, 0.54; 95% confidence interval, 0.34 to 0.87).

When we examined the HLA-DPB1 allele and phenotype frequencies in the UK control and patient groups, no difference was seen in the frequency of participants carrying HLA-DPB1 alleles with a glutamic acid residue at position 69. The HLA-DPB1 allele and phenotype frequencies for the UK population are shown in Table 4. Additionally, we observed no strong associations between any individual TAP variant and DPB1 allele, or between any individual TAP variant and DPB1 Glu69+ alleles. The frequencies of participants carrying HLA-DPB1 alleles with a valine residue at position 36, and aspartic acid residue at position 55, were found to be similiar in patients and control subjects.

Polish Population

For TAP2, there was a significant difference in phenotype frequency at position 379. At position 379, the frequency of participants positive for the valine variant was decreased in the patients (odds ratio, 0.17; 95% confidence interval, 0.02 to 1.0). No significant difference was seen in the overall genotype frequencies at this position.

Comparing the sarcoid groups from both UK and Polish populations showed significant differences between the two sets of patients. The phenotype frequencies at TAP2 position 379 showed a decrease in those positive for valine in the Polish patients (² = 5.5 with 1 df, p = 0.02). No significant difference was seen in the overall genotype frequencies at this position. At TAP2 position 565, the phenotype frequency of those positive for Threonine was increased in the Polish patients (² = 6.2 with 1 df, p = 0.01). A significant difference was also seen in the overall genotype frequencies at this position (² = 6.21 with 2 df, p = 0.01). Comparing the two control groups we observed an increase in the phenotype frequency of participants positive for glycine at TAP1 position 637 (² = 3.7 with 1 df, p = 0.05) in the UK control subjects. The overall genotype frequencies were not found to differ significantly at this position.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study we have investigated the association between TAP1 and TAP2 gene variations and sarcoidosis in clinically well-defined groups of subjects from two different ethnic populations. We detected significant differences in TAP2 between the UK control and patient groups, and in TAP2 between the Polish control and patient groups. When we compared the UK and Polish control groups we observed a significant difference in TAP1. In addition we found a difference in TAP2 between UK and Polish patients. Examination of the relationship between TAP variants and HLA-DPB1 locus variation in the UK Caucasian population showed no strong associations between the two loci. No difference was seen in the frequency of participants carrying HLA-DPB1 alleles with a glutamic acid residue at position 69. These results differ from the association we described previously in our UK patients. This difference may be due to the use of a more precise, sequence-based, HLA-DPB1 typing method, SSP-PCR, or, alternatively, may reflect the use of a larger sample size.

The lack of any strong MHC association in sarcoidosis that is constant across several populations suggests additional genetic factors may be involved in the susceptibility to this disease. In other MHC-linked disease models there are examples of multiple associations that can occur independently of one another. For example, in IDDM where patients carrying either a HLA-DR3 or -DR4 allele have a fivefold increase in relative risk of disease, but heterozygous patients carrying both HLA-DR3 and -DR4 have a fourteenfold increased risk of disease, suggests that multiple genetic factors are involved in the pathogenesis. The location of the polymorphic TAP genes in the HLA Class II region, where numerous sarcoidosis associations have been described, makes them of interestboth from a potential functional relevance because of their role in antigen processing, and from a linkage disequilibrium viewpoint.

The role of TAP proteins in making available peptide for loading onto MHC Class I molecules has been well documented (17), but there is also evidence supporting their role in the loading of certain native peptides onto MHC Class II molecules (18, 19). Although this is likely to be a relatively rare event, if pre-endosomal loading of Class II molecules can occur, circumstances that favor such an event such as polymorphism within the TAP genes might lead to the display of "illegitimate" self-peptides and the subsequent interruption of self-tolerance with pathologic consequences. High affinity binding of the TAP complex as a means of evading immunosurveillance by viruses has been shown to occur (20). In sarcoidosis the nature of the antigen stimulating inflammation and subsequent granuloma formation is unknown. However, it is conceivable that viral peptide binding of the TAP complex influenced to some degree by TAP polymorphism could result in altered trafficking of endogenous peptide toward the MHC Class II restricted antigen presentation pathway. A study by Ishihara and colleagues (21) concluded that TAP1, TAP2, and LMP2 genes are not primarily involved in susceptibility to sarcoidosis in a group of Japanese patients studied. We have shown differences in our UK and Polish patient groups compared with control subjects. This disparity is possibly due to ethnic differences. This view is supported by our findings since differences seen between the sarcoid and control groups in the UK population were not seen in the Polish population studied. The existence of a well-characterized recombination hotspot within intron 2 of the TAP2 locus (22) and the lack of any strong linkage disequilibrium between TAP1 and TAP2 loci, or between the TAP1 and HLA-DR/-DQ loci (23), means that there are unlikely to be strong associations between the HLA-DP and TAP loci. This is supported by our observation in the UK population that there were no strong associations between HLA-DP and the TAP loci. By examining the relationship between TAP variants and HLA-DPB1 locus variation in the UK Caucasian population we have determined that any associations observed at these loci occur independently of one another. Exclusion of linkage was essential considering that the TAP1 and 2 genes are close to the HLA-DPB1 locus.

The finding that HLA-DPB1 Glu69-positive alleles are not associated with sarcoidosis in our enlarged UK population study agrees with a recent study of African Americans (9). In that population Val36- and Asp55-positive alleles, but not Glu69-positive alleles, were found to confer increased risk of the disease. In this study we found no associations with any of these HLA-DPB1 residues in the UK population, suggesting that differing genetic background may account for the increased disease incidence and severity seen in varying ethnic groups. It is possible that within different populations different antigens interact with varying regions of the MHC, including HLA-DP, to produce a common series of pathogenic events.

This is the first study of the TAP genes in the UK and Polish sarcoid populations. We have demonstrated the importance of using multiple defined ethnic populations in helping to define the role genetic factors play in sarcoidosis susceptibility. Definition of the role the MHC genetic region plays in determining susceptibility to sarcoidosis will require additional multiple population studies.