© 2002 American Thoracic Society DOI: 10.1165/rcmb.4782 HLA-DQB1*0201A Marker for Good Prognosis in British and Dutch Patients with SarcoidosisClinical Genomics Group, Department of Occupational and Environmental Medicine, Imperial College of Science, Technology and Medicine, National Heart and Lung Institute, London, United Kingdom; Heart Lung Center Utrecht, Department of Pulmonology, Sint Antonius Hospital, Nieuwegein, The Netherlands; Department of Gastroenterology, Radcliffe Hospital, Oxford, United Kingdom; Department of Medical Microbiology & Immunology, Sint Antonius Hospital, Nieuwegein; and Heart Lung Center Utrecht, Department of Pulmonology, University Medical Center, Utrecht, The Netherlands Address correspondence to: Professor R.M. du Bois, M.D., Clinical Genomics Group, National Heart and Lung Institute, 1B Manresa Road, London SW3 6LR, UK. Email: r.dubois{at}rbh.nthames.nhs.uk
Human leukocyte antigen (HLA)-DQB1 is one of the intriguing candidate genes in sarcoidosis. We performed high resolution molecular HLA-DQB1 typing on two groups of white patients (British [UK] and Dutch [NL]) in order to investigate the relationship between 19 DQB1 alleles and disease severity and progression. A total of 803 individuals were studied (133 UK and 102 NL patients, and 354 UK and 214 NL control subjects). Disease severity data included extrapulmonary disease, chest X-ray stage, lung diffusing capacity for carbon monoxide, and FVC. Progression was evaluated on follow-up chest radiographs (2 and 4 yr). The results showed DQB1*0201 to be strongly protective against severe sarcoidosis in both populations; in other words, it localized to Stage I at all time points (P < 0.0001, Pcorrected (Pc) = 0.002), whereas another DQB1 allele, *0602, tended to have opposite effects. Further, a clear association was found between the *0201 allele and Löfgren's syndrome (P < 0.0001, Pc = 0.001). More importantly, carriage of this allele reduced the risk of disease progression. In contrast, the other common split of DQB1*02, *0202, did not affect disease severity but was mildly protective against sarcoidosis in the UK population (P = 0.02, pc not significant). In conclusion, this study shows that DQB1*0201 is a strong marker for mild sarcoidosis. Additional mapping across the DQB1*0201-DRB1*0301 haplotype, including specific alleles at genes such as DRB3, tumor necrosis factor, lymphotoxin-  , I-kappa-B-like protein, and B-associated transcript 1, is necessary for a final localization of the protective effect on this haplotype.
Abbreviations: bilateral hilar lymphadenopathy, BHL • degrees of freedom, df • diffusing capacity of the lung for carbon monoxide, DLCO • erythema nodosum, EN • human leukocyte antigen, HLA • major histocompatibility complex, MHC • P values corrected for the number of DQ alleles tested, Pc • polymerase chain reaction, PCR • sequence-specific primers, SSPs
Sarcoidosis is a multiorgan inflammatory disorder of unknown etiology, characterized by the accumulation of activated CD4+ T lymphocytes and macrophages at disease sites and by the formation of noncaseating epithelioid cell granulomas (1). The majority of T lymphocytes accumulating in the affected organs express the /ß T cell receptor that identifies antigenic peptides in the context of human leukocyte antigen (HLA) Class II molecules on antigen presenting cells (2). The genes encoding human leukocyte antigen (HLA) molecules fall into two major classes, Class I (HLA-A, -B, and -C) and Class II (HLA-DP, -DQ, and -DR), which are clustered within the major histocompatibility complex (MHC) region on the short arm of chromosome 6 (3). HLA molecules are pivotal to the adaptive immune response, and various associations with diseases, including sarcoidosis, have been described (4). The most frequently reported immunogenetic associations have been found with HLA Class II alleles, but these associations have varied in reported studies of patients from different races and populations. This has made the interpretation of the possible role(s) of these genes in the initiation of sarcoidosis difficult. Nonetheless, as evidence continues to accumulate for genetic susceptibility to sarcoidosis, the HLA Class II genes remain major candidate genes for further research (5, 6, 7, 8). Previous linkage studies in familial sarcoidosis have confirmed the importance of these loci (9, 10). In addition, a recent genome-wide linkage study by Schürmann and colleagues has pointed to the importance of the MHC for candidate genes in sarcoidosis, based on the observation that the greatest linkage score was found in this region (11). Most of the HLA Class II associations in sarcoidosis are found with the HLA-DRB1 and HLA-DQB1 loci. In Japanese patients with sarcoidosis, a strong association with the HLA-DR5, -DR6 and -DR8 alleles has been reported (12, 13). In German patients, HLA-DR5 was found to be associated with chronic disease, whereas in Polish patients, HLA-DR3 was associated with acute onset of disease and short disease duration (14, 15). In Scandinavians, a strong association between HLA-DRB1*03 (DR3) and acute onset and short duration of disease was found, while the alleles HLA-DR14 and DR15 were found to be associated with chronic disease (16). In a study of Polish patients, an association between the HLA-DRB1*03 alleles and Löfgren's syndrome was reported (17). In addition, a recent study by our group demonstrated a protection against sarcoidosis in cases of HLA-DRB1*01 and *04 carriage, which was probably related to the presence of hydrophobic residues at position 11, located within a pocket of the HLA-DR complex antigen binding groove (18). With regard to HLA-DQB1, the DQB1*0301 allele was found to be significantly increased in Japanese patients (19), whereas in Scandinavians DQB1*0201/2 was more common (16). A second Japanese study showed a significant increase in the HLA-DQB1*0601 allele in cardiac sarcoidosis (20). In a familial sarcoidosis study of Germans, DQB1*0603 and *0604 alleles were found to be over-represented in sarcoidosis (9). Finally, we previously reported an increased risk of sarcoidosis with the HLA-DQB1*0602 allele at the HLA-DQ locus; however, HLA-DQB1 typing was performed at low resolution for many of the other alleles, which interfered with the clarification of the role of this locus in sarcoidosis (18). With this background, the aim of the present study was to explore the association of alleles at the HLA-DQB1 locus with disease severity and progression, using high resolution HLA-DQB1 locus-molecular typing in two clinically well-defined groups of white patients with sarcoidosis.
Subjects The UK patients with sarcoidosis (n = 133) were recruited consecutively from the Royal Brompton Hospital, London (tertiary referral center taking patients from mainly the southeast of the UK), and the Dutch patients with sarcoidosis (n = 102) were from the Sint Antonius Hospital, Nieuwegein (secondary referral hospital in the Utrecht region). All patients were native British or Dutch whites and unrelated. The diagnosis of sarcoidosis was established when clinicoradiologic findings were supported by histologic evidence of noncaseating epithelioid cell granulomas, and after exclusion of other known causes of granulomatosis. Verbal and written patient consent for genetic analysis was obtained prior to phlebotomy and authorization was given by the Ethics Committees of both hospitals. The UK control population comprised 354 unrelated and unaffected native UK whites mainly from the southeast of the UK. The Dutch control group comprised 214 native Dutch whites from the Blood Transfusion Service of Utrecht, which takes donors mainly from the Utrecht region.
Evaluation of Pulmonary Disease Severity Radiography. Chest radiographs at presentation and at 2 yr and 4 yr were collected for each patient and assessed blind for disease severity by a pulmonary physician using standard radiographic staging for sarcoidosis. In brief, this staging is comprised of five stages: Stage 0, normal; Stage I, bilateral hilar lymphadenopathy (BHL); Stage II, BHL and parenchymal infiltration; Stage III, parenchymal infiltration without BHL; and Stage IV, irreversible fibrosis with loss of lung volume. For radiologic follow-up, one or more chest radiographic stage changes was regarded as significant. Radiologic improvement was defined as change from a higher into a lower radiographic stage and progression was defined as change from a lower to a higher stage. Radiologic appearances were otherwise classified as unchanged. Presentation chest radiographic data were available on 215 patients (115 UK and 100 NL). The distribution of chest radiographic stages at presentation is presented in Table 1. Chest radiographs, taken at 2 yr after initial presentation, were available on 163 patients (81 UK and 82 NL). Of this group, a total of 137 (64 UK and 73 NL) patients had radiographic follow-up at 4 yr. For five patients (4 UK and 1 NL), no chest radiograph was available for evaluation at 2 yr but was available at 4 yr. The difference in patient numbers at presentation compared with 2 and 4 yr is mainly due to the inclusion of some patients in whom a diagnosis had only recently been made. These patients, therefore, had not had enough follow-up to be included in the 2 and 4 yr analysis.
Pulmonary function testing. Both groups of patients had pulmonary function tests performed at the referral center. FVC and lung diffusing capacity for carbon monoxide (DLCO) were used to assess the presence of lung function impairment at presentation of disease. Lung function impairment was defined as FVC or DLCO < 80% of the predicted value. Lung function data were available for 121 UK patients (91.0%) and 88 Dutch patients (86.3%; Table 1).
Evaluation of Extrapulmonary Disease Severity
Analysis of Genetic Polymorphisms in the HLA-DQB1 Gene
PCR Conditions
Gel Electrophoresis
Statistical Analysis
HLA-DQB1 in Patients with Sarcoidosis and Control Subjects Table 2 summarizes the phenotype frequencies of HLA-DQB1 alleles in UK patients with sarcoidosis and unaffected UK control subjects, and Table 3 summarizes the results for the NL subjects. We found a decreased carrier frequency of the DQB1*0202 allele in the UK sarcoidosis population (15.0% compared with 25.7% in controls, P = 0.02 [Pc = not significant (ns)]), but this was not confirmed in the NL population. In the other DQB1 alleles, no relevant differences were found.
A comparison between the phenotype frequencies in UK and NL patients with sarcoidosis showed no significant differences although NL patients tended to have a lower DQB1*0602 allele carriage frequency and showed a higher DQB1*0604 frequency (P = 0.007 [Pc = ns]).
HLA-DQB1 in Relation to Sarcoidosis Subgroups
Another DQB1 allele, *0602, was found to be associated with unfavorable clinical characteristics. A higher incidence of uveitis and a higher stage on the initial chest radiograph was found in patients carrying this allele (P = 0.03 and 0.04, respectively; Pc = ns for both; see Table 4). Twelve of 72 *0602 carriers (16.7%) had uveitis compared with 11 of 163 non-*0602 carriers (6.7%). Forty-six of 64 *0602 carriers (71.9%) presented with Stage II or higher compared with 84 of 151 patients (55.6%) who did not carry this allele. In addition, the strongest association between *0602 carriage and chest radiographic stage was found with Stages II and III (  2 = 5.55 with 1 df, P = 0.02).
HLA-DQB1 in Relation to Pulmonary Disease Progression
Evaluation of radiographic evolution (from presentation to 4 yr) was performed in 135 patients, excluding the patients with stable Stage IV disease. Complete resolution of pulmonary disease, improvement toward Stage I, or stable Stage I disease was observed in 60 patients (Group A), whereas 75 patients showed progressive disease or persistent Stage II/III disease (Group B). In Group A, 41.7% of patients carried the HLA-DQB1*0201 allele compared with 10.7% in Group B ( 2 = 15.71 with 1 df, P < 0.0001 [Pc = 0.002]; see Figure 1). Therefore, the risk of disease progression or persistent Stage II/III disease was 24.2% in HLA-DQB1*0201 carriers and 65.7% in noncarriers. In other words, possession of an HLA-DQB1*0201 allele reduced the risk of disease progression or persistent Stage II/III disease in this patient cohort by  65%. Excluding Löfgren's patients caused a *0201 carrier frequency of 35.3% in Group A and 10.7% in Group B (2 = 9.79 with 1 df, P = 0.002 [Pc = 0.04]). Analysis of the protective effect of carriage of the *0201 in relation to corticosteroid treatment indicated no significant differences between treated and untreated patients.
Although the HLA-DQB1*0602 allele carrier frequency was higher in Group B compared with Group A (30.7% versus 20.0%), this difference did not reach statistical significance (Figure 1).
In this study we have investigated the relationship between HLA-DQB1 alleles and sarcoidosis severity at presentation and progression of disease. The results showed HLA-DQB1*0202 to be mildly protective against sarcoidosis in the white UK population, but this finding was not confirmed in the NL. In contrast, the other common split of DQB1*02, namely DQB1*0201, did not affect susceptibility to the disease but was associated with mild disease and favorable radiologic outcome, the strongest association being found between patients presenting with Löfgren's syndrome and *0201. However, even after excluding patients presenting with Löfgren's syndrome, there remained a significant association between *0201 carriage, Stage 0/I disease, and good prognosis. In patients showing complete resolution or Stage I disease after a 4 y follow-up period from presentation, 40% carried the HLA-DQB1*0201 allele compared with 10% in patients showing progressive pulmonary disease or persistent Stage II/III disease. This comparison suggests that the possession of the HLA-DQB1*0201 allele significantly reduces the risk of disease progression or persistent diffuse lung abnormalities during the first follow-up years. A risk reduction percentage in *0201 carriers of 65% was estimated.
The HLA-DQB1*0201 allele is in tight linkage disequilibrium with the HLA-DRB1*0301 allele. This allele has been found in approximately two-thirds of Scandinavian patients with nonchronic sarcoidosis (i.e., patients who showed complete recovery within 2 y) (16). The strongest association was found between the HLA-DRB1*0301 allele and Löfgren's syndrome, a relationship that has been confirmed in Polish patients with sarcoidosis (17). In a study by Berlin and colleagues, patients were also typed for HLA-DQB1 alleles; alleles *0201 and *0202 were not, however, separated (16). The authors reported a higher frequency of the combined HLA-DQB1*0201/2 allele in the sarcoidosis group compared with controls, and the highest frequency in patients with nonchronic sarcoidosis. However, this association was less strong than with the HLA-DRB1*0301 allele. Remarkably, the present study showed that the HLA-DQB1*0201 allele is a strong marker for mild sarcoidosis, and the *0202 allele might have a protective effect in particular populations (i.e., the allele carrier frequency of this allele was lower in UK patients with sarcoidosis compared with unaffected control subjects). This characteristic could only be revealed by separating these two alleles. Furthermore, not separating the *0201 and *0202 allele could explain, at least in part, the finding of the weaker association between the combined HLA-DQB1*0201/2 allele frequency and milder sarcoidosis phenotypes in the study by Berlin and coworkers. Therefore, at present it remains to be elucidated which of the two tightly linked alleles—HLA-DRB1*0301 or HLA-DQB1*0201—is really responsible for the association with a milder sarcoidosis phenotype. In addition, since the haplotype HLA-DRB1*0301, DQB1*0201 includes specific alleles at the HLA-DRB3, and the genes encoding tumor necrosis factor- Aside from the mild disease effect of the HLA-DQB1*0201 allele, we found the DQB1*0602 allele to be associated with less favorable clinical characteristics of sarcoidosis (e.g., uveitis and higher chest radiographic stage at presentation). Although there was no significant association between *0602 and an unfavorable radiographic evolution during follow-up, this allele tended to be more common in the severe disease evolution group compared with the mild disease evolution group. The HLA-DQB1*0201 and *0602 alleles act in an intriguing opposite direction in the severity and progression of sarcoidosis. Our HLA-DQB1*0602 result is in agreement with that reported by Berlin and colleagues, who reported an association between this allele and chronic sarcoidosis (16). In addition, Tang and colleagues reported an association between HLA-DR15 and uveitis (22). In four patients, they performed molecular subtyping of the HLA-DR15 specificity and detected the HLA-DRB1*1501 allele in all of them. HLA DQB1*0602 allele is in tight linkage disequilibrium with HLA-DRB1*1501 in white populations. A Japanese study showed an association between an HLA-DQB1*06 allele (*0601) and a severe sarcoidosis phenotype, cardiac sarcoidosis, further suggesting a role for this cluster of alleles in determining severe sarcoidosis phenotypes (20). UK and NL patients with sarcoidosis showed a strong similarity in HLA-DQB1 allele frequencies, supporting genetic homogeneity between these populations. The only significant difference was the HLA-DQB1*0604 allele which was increased in Dutch patients with sarcoidosis. Further, the *0602 allele tended to be lower in Dutch patients, but this could have been related to the less severe disease characteristics in the Dutch sarcoidosis group as a whole. The differences in disease severity phenotype between the two sarcoidosis populations is likely caused by a different patient referral practice in the Sint Antonius Hospital (secondary referral center) compared with the Royal Brompton Hospital (tertiary referral center). The observed genetic homogeneity and the fact that the same trends in HLA-DQB1*0201 and *0602 associations were found in each of the patient populations separately, allowed us to combine both populations for genotype–disease phenotype analysis. HLA Class II molecules are synthesized in the endoplasmic reticulum and delivered by way of the Golgi apparatus into primary lysosomes. Foreign proteins are taken up by endocytosis or phagocytosis and sequestered into endosomes. After fusion of lysosome and endosome, engulfed proteins are degraded into peptides, which are loaded onto the Class II molecules with help from HLA-DM molecules. The peptide-laden Class II molecules are then exported to the surface of the antigen presenting cell (B cell, macrophage, dendritic cell) (3). Subsequently, the HLA Class II-peptide complex interacts with the T cell receptor on the surface of CD4+ T lymphocytes, initiating an adaptive immune response. The loading of a particular peptide onto a Class II molecule depends on the characteristics of the peptide-binding groove of this molecule, which consists of three parts: a floor and two walls (23). Therefore, as with previous HLA associations, those reported in the present study might indicate that specific peptides are presented in the initial phase of the pathogenesis of sarcoidosis. Furthermore, some peptides might result in milder forms of sarcoidosis than others, or give rise to particular disease phenotypes. Whether the HLA-DQB1*0201 allele-derived ß chain is directly involved in the binding of a particular sarcoidosis-triggering peptide in a more favorable way, or merely indirectly, by operating as a marker for other immunoregulating molecules, needs further elucidation. This study provides strong evidence for the genetic basis of sarcoidosis and, further, it provides data that help tease out the complexity of the initiation of sarcoidosis. The characteristics of the peptide-binding groove of the HLA molecules are at least in part determined by the amino acid sequences of its chains, and these are encoded by genes such as HLA-DQB1. Knowledge of the involvement of particular allelic variants in sarcoidosis might help in finding a particular peptide motif which could be the key to identifying the antigenic trigger(s) of this disease (24). These findings are clinically relevant in that they may inform the future development of genetic screening to identify likely disease severity and prognosis when patients first present. These tests might therefore ultimately help pulmonary physicians in making disease management decisions on a more individualized basis. In conclusion, in this study we identified two HLA-DQB1 associations with opposite effects on disease phenotype. The HLA-DQB1*0201 allele was associated with mild disease and showed a strong relationship with favorable disease outcome, such as complete resolution of chest radiograph or Stage I disease. The HLA-DQB1*0602 allele, on the other hand, tended to be associated with less favorable disease characteristics, such as uveitis and Stage II, III, or IV disease. Furthermore, another split of the HLA-DQB1*02 allele, *0202, might have a protective effect against sarcoidosis in UK whites. Additional fine mapping across the MHC region is necessary to clarify linkage disequilibrium problems with HLA-DRB1*0301 and other alleles, such as the TNF locus, and to determine unequivocally the identity of the protective allele.
The authors thank E.M. Bik and S. Chocron for their help with the isolation of the Dutch DNA, and Dynal Biotech (Bromborough, Wirrall, UK) for providing the primers for HLA-DQB1 typing. Supported in part by grants from the European Respiratory Society, the Mr. Willem Bakhuys Roozeboom Fonds, and the Prof. Dr. Jaap Swierenga Stichting (J.C.G.).
* These authors contributed equally to this work.  Received in original form November 21, 2001 Received in final form July 8, 2002
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
II, and impairment of DLCO (i.e. DLCO < 80% of predicted value) (P = 0.03, P = 0.01, and P = 0.03, respectively; Pc = ns for all; see 
-B-like protein, and HLA-B-associated transcript 1 loci, and less so HLA-B*0801 and Cw*0701, additional fine mapping across the MHC region is required for a final localization of the protective effect of this common haplotype. 
