Differential expression of HLA class II genes associated with disease susceptibility and progression in rheumatoid arthritis.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 48, No. 10, October 2003, pp 2779–2787 DOI 10.1002/art.11251 © 2003, American College of Rheumatology Differential Expression of HLA Class II Genes Associated With Disease Susceptibility and Progression in Rheumatoid Arthritis Christian Heldt,1 Joachim Listing,1 Osman Sözeri,1 Franca Bläsing,2 Stefan Frischbutter,1 and Brigitte Müller1 ger of the 2 DRB4 promoters, independent of the DRB1 allele. Moreover, DRB1*04 alleles in RA patients showed a reduced association with the DRB4 splice variant, completely preventing DRB4 expression. Conclusion. Our findings represent the first evidence of a correlation between the differential expression of HLA class II genes and both the susceptibility and the progression of RA. Objective. Rheumatoid arthritis (RA)–associated HLA class II genes are assumed to promote susceptibility to and/or progression of the disease. Among the various modes of action proposed so far is the effect of the differential expression of HLA class II genes in different types of antigen-presenting cells on the Th1/ Th2 balance. The aim of this study was to investigate the differential expression of genes encoded within the RA–associated HLA–DR4 superhaplotype and within the neutral DR7 and DR9 superhaplotypes. Methods. The promoters encoded within these 3 haplotypes were first analyzed for sequence polymorphisms. To test for functional consequences, we assumed that the binding of nuclear factors to the promoter elements was correlated with the transcription activity, and we used surface plasmon resonance technology. To that end, oligonucleotides representing the polymorphic regulatory sequences and nuclear extracts from a monocyte cell line and a B cell line were used. Results. While the promoters of the highly polymorphic HLA–DRB1*04, *07, and *09 alleles showed comparable binding of nuclear factors, differential binding was observed for the 2 promoters that drive the relatively nonpolymorphic DRB4 alleles in linkage disequilibrium with DRB1. Interestingly, analysis of RA patients positive for DR4, DR7, and DR9 revealed the segregation of radiographic progression with the stron- The association of HLA class II genes with the development of autoimmune diseases has captivated numerous scientists over the last decades. This association has extensively been studied in rheumatoid arthritis (RA). The DR4, DR1, and DR10 haplotypes have been found to be positively associated with the disease, while others, such as DR2 and DR7, have been reported to be neutral, if not protective (1–3). Several mechanisms have been proposed to explain the association, among them are epitope selection (4) and differential expression (5). The shared epitope hypothesis centers on epitope selection and is intriguingly simple. However, evidence has accumulated from mouse studies supporting an impact of the differential expression of class II major histocompatibility complex (MHC) genes on the T cell cytokine profile, and thus an association with protective/ suppressive effects (5). In this context, differential expression of the various I-A genes has been correlated with a single mismatch in the X2 box of the promoter region (6). Interestingly, the structural requirements for class II MHC promoters (S, X1, and X2, as well as Y, CAAT, and TATA boxes) are well conserved over long evolutionary distances but allow for considerable intraspecies variability. It is assumed that the polymorphism in the regulatory region confers flexibility in the immune response, both for the species and for the heterozygous Supported by a grant from the Fritz Thyssen Stiftung and by the Berlin Senate for Research and Cultural Affairs. 1 Christian Heldt, PhD, Joachim Listing, PhD, Osman Sözeri, PhD, Stefan Frischbutter, Brigitte Müller, PhD: Deutsches RheumaForschungs-Zentrum, Berlin, Germany; 2Franca Bläsing, PhD: Max-Planck-Institut für Molekulare Genetik, Berlin, Germany. Address correspondence and reprint requests to Brigitte Müller, PhD, Deutsches RheumaForschungs-Zentrum, Schumannstrasse 21/22, Berlin 10117, Germany. E-mail: email@example.com. Submitted for publication September 16, 2002; accepted in revised form May 22, 2003. 2779 2780 individual (7). While the variation in the coding sequence is understood to be sustained by balancing selection, concomitant variation in the regulatory regions is thought to result from linkage disequilibrium (8). However, the idea has recently been put forward of a coordinate evolution in the coding and regulatory sequences, which ensures that class II MHC sequences favoring Th2 differentiation associate with class II MHC molecules that are able to present worm epitopes. Conversely, Th1-favoring promoters are predicted to be linked to coding sequences that are able to present endogenous viral epitopes (9). The enormously polymorphic and polygenic organization of the HLA complex and the lack of specific monoclonal antibodies have so far only allowed for a patchy analysis of the differential expression of class II MHC genes in humans. Several hundred HLA–DRB1 alleles showing sequence variation in the coding region (10) as well as in the regulatory region (11) can be grouped into the 10 different superhaplotypes, DR1– DR10, reflecting the 10 major serologically defined HLA–DR specificities. These superhaplotypes each contain multiple haplotypes with unique DRB1 sequences and multiple associations with the appropriate DRB3, DRB4, or DRB5 genes plus additional nonfunctional DRB genes. They are encoded on 5 different allele lineages. One of these lineages encodes the DR4, DR7, and DR9 superhaplotypes containing the DRB4 gene, associated with an HLA–DRB1*04, *07, and *09 allele, respectively (Figure 1). While the DR4 superhaplotype is RA-associated, DR7 and DR9 are considered neutral. So far, more than 40 different DRB1*04 alleles that share the same promoter have been described. Likewise, there is only 1 promoter reading into the single DRB1*09 allele and 1 promoter reading into the 4 DRB1*07 alleles (10). In contrast to the DRB1 gene, the DRB4 gene shows relatively little polymorphism, and of the 8 functional alleles, only 3 show an amino acid substitution within exon 2 that encodes the third hypervariable loop of the class II molecule responsible for epitope selection (11). However, a regulatory polymorphism is associated with the DRB4 gene, consisting of 2 different promoters, DRB4A and DRB4B. Furthermore, the DRB4A promoter has been reported to be in linkage disequilibrium with a splice variant that prevents the formation of the DRB4 gene product altogether (Figure 1) (12). If present, both the DRB1 and the DRB4 gene products combine with the nonpolymorphic ␣-chain to form the heterodimeric HLA–DR molecule. To ensure sufficient immune responses to survive HELDT ET AL Figure 1. Genomic organization of the HLA complex on the human chromosome 6. Top, The HLA–DR region is embedded between the HLA–DQ and the HLA class III region. Middle, Enlargement of the DR region, showing the 5 allele lineages that encode the 10 different DRB1 superhaplotypes (DR1–DR10). Membership in the allele lineages is defined by the DRB1 allele. Bottom, Enlargement of the allele lineage, showing the functional DRB1 alleles, the DRB4 and DRA genes, as well as pseudogenes encoded by this particular allele lineage. Promoters are shown as large arrows, the DRB4 splice variant is indicated by a solid star, and pseudogenes are shown in gray typeface. infections, class II MHC expression on antigenpresenting cells needs to be tightly controlled. The high number of different class II MHC genes and alleles has allowed for a detailed analysis of the structural requirements for this tight regulation, consisting of ubiquitous as well as unique regulatory elements. Among the unique elements are the X1 box, which binds the different RFX factors, which, in turn, bind the class II MHC transcription activator (CIITA) molecule to form the supramolecular transcription complex (13), and the S box, which represents a duplicated X1 box. The ubiquitous regulatory elements consist of the X2 box, which has recently been defined as a CRE box (14), the Y box, which represents a reversed CAAT box, the CAAT box itself, which binds the nuclear factor Y factors, and the TATA box at position –45, which binds the TATAbinding factors (9). It has recently been proposed that the unique elements ensure a basic transcription of the HLA CLASS II GENES IN RA SUSCEPTIBILITY AND PROGRESSION locus, while the ubiquitous elements are responsible for mediating the fine-tuning (15). The differential expression of HLA class molecules has previously been described on various levels. A ranking of transcription activity among the various HLA–DR genes and alleles was established via chloramphenicol acetyltransferase assays performed in EpsteinBarr virus–transformed B cell lines (16). The corresponding steady-state levels of messenger RNA (mRNA) were analyzed by reverse transcription– polymerase chain reaction (RT-PCR) and RNase protection assays using RNA isolated from peripheral B cells (17,18). In the context of RA, an elevated expression of HLA class II genes has been described for peripheral B cells (19). Thus, HLA class II expression has been studied almost exclusively in B cells. However, the recent finding of a differential surface expression of DRB1 and DRB4 on monocytes, but not on B cells, calls for an in-depth analysis of HLA class II expression on all types of antigen-presenting cells (20). In the present study, we set out to compare the RA-associated DR4 superhaplotype and the neutral DR7 and DR9 superhaplotypes, performing binding studies of nuclear extracts from monocytes and B cells to the various promoter elements. Our particular focus was on structural differences in the regulatory regions that lead to functional consequences related to the autoimmune process. PATIENTS AND METHODS Study population. The study patients met the American College of Rheumatology (formerly, the American Rheumatism Association) classification criteria for RA (21) (as described in detail in ref. 2). A total of 73 RA patients and 28 controls expressing the DR4, DR7, and/or DR9 haplotypes were evaluated. The mean age of the RA patients was 51 years (range 19–74 years); 17 of them were men. The mean age of the controls was 37 years (range 23–60 years); half of them were men. Genomic DNA was isolated according to the method described by Sykes (22). Typing of the polymorphism at the HLA–DRB locus was performed using genomic DNA, PCR, and biotinylated sequence-specific oligonucleotide probes (23). Cell lines and fluorescence-activated cell sorting (FACS) analyses. The monocyte cell line THP-1 (ACC 16; DSMZ, Braunschweig, Germany) and the B cell line BJAB (ACC 72; DSMZ) were cultured (at 37°C in an atmosphere of 5% CO2) in complete RPMI 1640 medium containing 10% fetal calf serum, 2 mM L-glutamine, 100 units/ml of penicillin, 100 g/ml of streptomycin, and 50 M 2-mercaptoethanol. Both cell lines were typed at their HLA–DR loci as described above. Cell surface expression of DR molecules on THP-1 and BJAB cells was monitored via staining with the pan–reactive monoclonal antibody L243 (clone HB55; American Type Cul- 2781 ture Collection [ATCC], Manassas, VA), using the CD3specific monoclonal antibody OKT3 (clone CRL-8001; ATCC) as an isotype control. Flow cytometric analyses were performed using a FACSCalibur instrument with CellQuest software (both from Becton Dickinson, Heidelberg, Germany). Nuclear extracts. Cells were left untreated or were stimulated with cAMP (100 M), transforming growth factor ␤1 (TGF␤1; 20 ng/ml), vitamin D3 (250 ng/ml), lipopolysaccharide (LPS; 10 g/ml), interleukin-4 (IL-4; 20 units/ml), TPA (1 nM), IL-1␤ (5 ng/ml), interferon-␥ (IFN␥; 1,000 units/ml), and anti-CD40/anti-IgM (5 g/ml each), respectively. Then, 5 ⫻ 107 THP-1 and BJAB cells were washed twice in ice-cold phosphate buffered saline and resuspended in 500 l of buffer A (10 mM HEPES, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM dithiothreitol, pH 7.9) containing a protease inhibitor cocktail (Roche, Mannheim, Germany). The suspension was kept on ice for 15 minutes, and then Nonidet P40 was added to a final concentration of 0.5%. The lysates were centrifuged at 6,500 revolutions per minute for 20 seconds, and the pellets were resuspended in 150 l of buffer C (20 mM HEPES, 1.5 mM MgCl2, 420 mM NaCl, 0.2 mM EDTA, 25% [weight/volume] glycerol, pH 7.9) containing the protease inhibitor cocktail. The probes were stirred vigorously on ice for 30 minutes and centrifuged at 13,000 rpm for 10 minutes. The nuclear extracts containing supernatants were collected, and the protein concentrations were measured using the bichinolin-4-carbonic acid protein assay reagent kit (Interchim, Montluçon, France). Surface plasmon resonance. The BIAcore System 2000 (Uppsala, Sweden) was used to study the binding affinities of nuclear extracts to promoter fragments bound to streptavidincoated chips. Experiments were performed as described previously (24). Minor variations included using TES buffer (10 mM Tris, 1 mM EDTA, 300 mM NaCl, pH 8) for pretreatment of the chips. Before immobilization on the streptavidin chip, the biotinylated sense strands of the promoter fragments (Table 1) were hybridized to the antisense strands (2-fold excess), and double-stranded oligonucleotides were diluted in TES buffer to a final concentration of 100 nM. Another variation included the immobilization of 400 resonance units (RU) of DNA onto the chips. Measurements were performed at 25°C and at a flow rate of 30 l/minute in modified buffer C (20 mM HEPES, 1.5 mM MgCl2, 420 mM NaCl, 0.2 mM EDTA, 0.005% BIAcore Surfactant P20, pH 7.9). Two hundred fifty microliters of the nuclear extracts (100 g/ml) was injected and, at the end of the association time, was replaced with continuous buffer flow. At the end of each cycle, the sensor chip was regenerated by injecting two 1-minute pulses of 0.05% sodium dodecyl sulfate in TES buffer. The binding response was analyzed by measuring the RU values during each cycle. Data analysis was performed with the Misevaluation software (BIAcore). The binding affinities were determined by subtracting the RU value before injection from the RU value at the end of the association phase. The specific binding affinities were then subtracted from the reference cell containing no DNA and were normalized to the nonspecific DNA. Electrophoretic mobility shift assay (EMSA). The single-stranded oligonucleotides TATA box DRB4A and TATA box DRB4B (Table 1) were end-labeled using ␥-32PATP (Amersham, Freiburg, Germany) and T4 kinase (Promega, Madison, WI). Double-stranded oligonucleotides were 2782 HELDT ET AL Table 1. Oligonucleotides used in the present study* Oligonucleotide To analyze DRB4 splice variants 5⬘ DRB4 (intron 1) 3⬘ DRB4 (exon 2) Sequencing To analyze DRB4 promoters 5⬘ W-Box DRB4 3⬘ ATG-DRB Sequencing To analyze binding affinities 5⬘ Bio S-Box DR4 5⬘ Bio S-Box DR7, DR9 5⬘ Bio S-Box DRB4 5⬘ Bio X-Box DR4 5⬘ Bio X-Box DR7, DR9 5⬘ Bio X-Box DRB4 5⬘ Bio X-Box DRA 5⬘ Bio Y-Box DR4, DR7, DR9 5⬘ Bio Y-Box DRB4A 5⬘ Bio Y-Box DRB4B 5⬘ Bio CCAAT-Box DR4 5⬘ Bio CCAAT-Box DR4, DR7 5⬘ Bio CCAAT-Box DRB4 5⬘ Bio TATA-Box DR4, DR7 5⬘ Bio TATA-Box DR9 5⬘ Bio TATA-Box DRB4A 5⬘ Bio TATA-Box DRB4B 5⬘ Bio random Sequence 5⬘-GAG-GCC-GCC-CGT-GTG-ACC-GGA-3⬘ 5⬘-CAC-AAC-CCC-GTA-GTT-GTA-TCT-GCA-GTA-3⬘ 5⬘-CAC-CTC-GGC-CCG-CCT-CCG-3⬘ 5⬘-GAT-CGC-TAG-CCT-TTA-AAG-AAG-ACA-ATT-GTT-TCA-GAG-3⬘ 5⬘-GAG-CTT-CAG-ACA-CAC-GAT-GCT-GGA-G-3⬘ 5⬘-TTT-CAG-AGG-AGG-ACC-TTC-3⬘ 5⬘-BIO-GTT-TCA-GAA-AAG-GAC-CTT-CAT-ACA-GC-3⬘ 5⬘-BIO-GTT-TCA-GAA-GAG-GAC-CTT-CAT-ACA-GC-3⬘ 5⬘-BIO-GTT-TCA-GAG-GAG-GAC-CTT-CAT-ACA-GC-3⬘ 5⬘-BIO-AGC-ATC-TCT-GAC-CAG-CGA-CTG-ATG-ATG-CTA-TTG-TAC-T-3⬘ 5⬘-BIO-AGC-ATC-TCT-GAC-CAG-CGA-CTG-ATG-ATG-CTA-TTG-TAA-T-3⬘ 5⬘-BIO-AGC-ATC-TCT-GAC-CAG-CAA-CTG-ATG-ATG-CTA-GTG-AAC-T-3⬘ 5⬘-BIO-GAA-CCC-TTC-CCC-TAG-CAA-CAG-ATG-CGT-CAT-CTC-AAA-A-3⬘ 5⬘-BIO-TCA-GAT-GCT-GAT-TCG-TTC-TCC-AAC-AC-3⬘ 5⬘-BIO-TCA-GGT-GCT-GAC-TGG-TTC-TCC-AAC-AC-3⬘ 5⬘-BIO-TCA-GAT-GCT-GAC-TGG-TTC-TCC-AAC-AC-3⬘ 5⬘-BIO-CAC-TAG-ATT-ACC-CAA-TCC-ACG-AGC-AA-3⬘ 5⬘-BIO-CAC-TAG-ATT-ACC-CAA-TCC-ACA-AGC-AA-3⬘ 5⬘-BIO-CAC-GAG-ATT-ACC-CAA-TCA-AGG-AGC-AA-3⬘ 5⬘-BIO-ACT-TCT-TCC-CTA-TAA-TTT-GGA-ATG-TGG-3⬘ 5⬘-BIO-ACT-TCT-TCC-CTA-TAA-TTT-GAA-ATG-TGG-3⬘ 5⬘-BIO-ACT-TCC-TTC-CTA-TAA-TTT-GGA-ATG-TGG-3⬘ 5⬘-BIO-ACT-TCC-TTC-CTA-TAA-TTT-GAA-ATG-TGG-3⬘ 5⬘-BIO-ATC-GGC-TAT-CTA-CGG-ATC-AGT-ACA-CT-3⬘ * Oligonucleotides were synthesized by Tib Molbiol (Berlin, Germany). For the oligonucleotides used to analyze binding affinities, only the biotinylated (Bio) sense strand is shown; the antisense strand did not include biotin. generated by annealing the 32P end-labeled sense and unlabeled antisense oligonucleotides (20-fold excess of antisense). For competition experiments, the 50 nM double-stranded promoter fragment was incubated for 15 minutes with unlabeled oligonucleotide competitors (2–100-fold excess), 40 g of nuclear extract, and 0.5 mg/ml of poly(dI-dC) in binding buffer (40 mM HEPES, 10 mM MgCl2, 100 mM KCl, 30% [w/v] glycerol, 0.02% [w/v] Triton X-100, pH 7.9) containing the protease inhibitor cocktail (Roche) in a final volume of 20 l. Electrophoresis was carried out on a native polyacrylamide gel (4%) using 0.5⫻ Tris–borate–EDTA as the running buffer. After drying, the gel was exposed to the phosphorimager for quantification. DNA amplification and sequencing. Genomic DNA (0.5 g) was amplified in a 50-l PCR reaction containing 1 M each of the 5⬘ and 3⬘ primers (Table 1), 2.5 units of Taq polymerase (Promega), 0.3 units of Pfu polymerase (Promega), and 250 M dNTPs in PCR buffer (670 mM Tris HCl, 1.3 mM MgCl2, 160 mM [NH4]2SO4, 0.1% [w/v] Tween 20, pH 8.8). An initial incubation at 96°C for 10 minutes was followed by 35 cycles at 96°C for 1 minute, 60°C for 1 minute, and 72°C for 1 minute. The reaction was terminated at 72°C for 10 minutes. The PCR products were separated on 1.4% agarose gels and purified using a JetSorb kit (Genomed, Bad Oeynhausen, Germany). The amplified promoter fragments and intron 1/exon 2 splice sites were then sequenced using the dideoxy chain-termination method and the ABI Prism 310 sequencer (Applied Biosystems, Foster City, CA). The sequencing reaction was carried out in a 10-l volume containing 1 l of 0.5 M 5⬘ primer (Table 1), 2 l of PCR product, and 2 l of BigDye (Applied Biosystems). Statistical analysis. Fisher’s exact tests were performed to calculate P values and confidence intervals describing the differences between the fractions shown in Table 2 and Figure 5. RESULTS Mediation of the IFN␥-stimulated expression of HLA–DR via the X box. In this study, we compared the RA-associated superhaplotype DR4 with the neutral DR7 and DR9 superhaplotypes, looking for structural differences that lead to functional consequences. For analysis of the structural differences, we aligned the promoter sequences reading into the HLA–DRB1*04, *07, *09, and the HLA–DRB4 alleles (11), and we identified polymorphic residues in and around the known S, X, Y, CAAT, and TATA boxes. For analysis of the functional consequences, we designed oligonucleotides representing the various polymorphic boxes and tested their binding to nuclear extracts isolated from antigen-presenting cells. We used as antigen-presenting cells the human monocyte cell line THP-1 and the human B cell line BJAB, which carry endogenously the HLA CLASS II GENES IN RA SUSCEPTIBILITY AND PROGRESSION Figure 2. Differential surface expression of HLA–DR molecules on monocytes and B cells. Histograms show the flow cytometric analyses of unstimulated or interferon-␥ (IFN␥)–stimulated (1,000 units/ml for 48 hours) THP-1 and BJAB cells stained with the isotype control or with the HLA–DR–specific L243 monoclonal antibody. DR1/DR2 and DR8/DR5 (or DR6) superhaplotypes, respectively. To confirm that both cell lines were suitable hosts for analyzing DR expression, surface HLA–DR was measured by FACS analysis using the monoclonal antibody L243. Figure 2 shows that 100% of the BJAB cells were positive for HLA–DR, and stimulation with IFN␥ only slightly increased the amount of surface expression, as determined by the mean fluorescence intensity. Likewise, TPA and IL-1␤ only slightly increased HLA class II surface expression, while anti-IgM/anti-CD40, LPS, IL-4, cAMP, vitamin D3, and TGF␤1 did not add to the stimulatory effect induced by IFN␥ (data not shown). In contrast, only very few of the unstimulated THP-1 cells were DR positive. However, stimulation with 1,000 units/ml of IFN␥ for 48 hours (Figure 2), but not with TPA or TGF␤1 (data not shown), induced DR expression on all THP-1 cells. To determine the contribution of the individual promoter boxes to DR expression, we analyzed the binding affinities of THP-1 and BJAB nuclear extracts to the different oligonucleotides representing the polymorphic promoter boxes. The surface plasmon resonance was measured with the BIAcore apparatus, and a random oligonucleotide was used as a control. The binding affinities obtained were then normalized to the affinity of THP-1 or BJAB nuclear extracts to this random sequence and are presented as percentage increase. Figure 3A shows a comparison of the X box sequences of the DRB1*04, *07, *09, DRB4A, and DRB4B promoters. The DRB1*04 promoter differed from the DRB1*07 and *09 promoters by just 1 polymorphic residue 3⬘ of the X2 box. The DRB4A and DRB4B promoters were identical, yet differed from 2783 DRB1*04 by a G3 A substitution within the X box and by T3 G and T3 A substitutions immediately 3⬘ of the X2 box. The X boxes of the DRA promoter were very different from those of the DRB promoters, yet the relative binding of THP-1 and BJAB nuclear proteins to the X box DRA was highest (132% and 191%, respectively) (Figure 3B). For the X box DRB oligonucleotides, binding varied between 106% and 109% for the THP-1 nuclear protein and between 117% and 126% for the BJAB nuclear protein. Activation of THP-1 cells with IFN␥ stimulated the binding of nuclear proteins to the X box and led to a 126% (DRB1*07/*09), 139% (DRB4), 149% (DRB1*04), and 215% (DRA) increase, respectively. For nuclear extracts of IFN␥-stimulated BJAB cells, the increase in transcription factor concentration was less dramatic, but was also strongest for the X box DRA oligonucleotide (Figure 3B). Stimulation of THP-1 or BJAB cells with cAMP, TPA, or TGF␤1 did not lead to a significant increase in transcription factors (data not shown). Of all the oligonucleotides tested, the X box oligo showed the strongest increase in the binding of nuclear extracts isolated from IFN␥-stimulated THP-1 cells, indicating that IFN␥ Figure 3. Mediation of interferon-␥ (IFN␥)–stimulated HLA–DR expression via the X box. A, Sequence alignment of the HLA– DRB1*04, DRB1*07, DRB1*09, DRB4A, DRB4B, and DRA promoters. Shaded areas show the core sequences of the transcription factor target sites. Sequence variations among the alleles/genes are indicated. B, Binding of nuclear extracts from the THP-1 and BJAB cell lines to the X boxes of DRB1*04, DRB1*07/*09, DRB4, and DRA. Surface plasmon resonance was measured and normalized to the random DNA in order to calculate the percentage increase in binding. For isolation of nuclear extracts, cells were either left untreated (solid bars) or were treated with 1,000 units/ml of IFN␥ for 48 hours (hatched bars). Values are the mean and SD. 2784 stimulation of HLA–DR expression is mediated via the X box. Mediation of the differential expression of the relatively nonpolymorphic HLA–DRB4 gene via the TATA box. Alignment of the polymorphic TATA box sequences is shown in Figure 4A. While the box itself was completely conserved, sequence variation was found 5⬘ and 3⬘ of the TATA box. The relative binding affinities of both THP-1 and BJAB nuclear proteins were highest for the DRB4B (203% and 184%, respectively) and lowest for the DRB4A (126% and 106%, respectively) promoter oligonucleotides. The binding of the nuclear proteins to the DRB1-specific oligonucleotides ranged between 174% and 179% for THP-1 cells and between 155% and 168% for BJAB cells, respectively. Stimulation of BJAB cells with IFN␥ increased the concentration of transcription factors that bound the TATA box and led to a 5% (DRB4A), 10% (DRB4B), 29% (DRB1*09), and 37% (DRB1*04/*07) increase in the relative binding. Neither IFN␥, cAMP, nor TPA led to an increase in the transcription factor concentration of THP-1 nuclear extracts (data not shown). The differential binding of nuclear proteins to the DRB4A- and DRB4B-specific oligonucleotides was confirmed by competition assays in which 32P-labeled TATA box DRB4A or DRB4B and cold TATA box DRB4A or DRB4B oligonucleotides competed for binding to the nuclear proteins. The polyacrylamide gel from the EMSA showed a more rapid decline in the signal strength of the upper band when competed out with unlabeled TATA box DRB4B (Figure 4C). Figure 4D summarizes the findings of the competition analysis with 32 P-labeled TATA box DRB4A using a phosphorimager for quantification of the bands. Again, the reduction in signal strength was stronger when competed out with unlabeled TATA box DRB4B, indicating that the DRB4B promoter binds nuclear factors with a higher affinity than does the DRB4A promoter and is predicted to result in stronger transcription activities. Association of an elevated expression of HLA– DRB4B with radiographic progression in early RA. We next addressed whether any of these structural and functional promoter differences correlated with a specific disease course. We therefore evaluated a cohort of RA patients who had previously been evaluated for risk factors of radiographic progression (2). In these patients, a C-reactive protein (CRP) level ⱖ1.5 mg/dl at the start of treatment with disease-modifying antirheumatic drugs was the strongest risk factor for radiographic progression of RA and was independent of genetic predisposition (2). We selected all DRB1*04-, HELDT ET AL Figure 4. Mediation of the differential expression of HLA–DRB4A and HLA–DRB4B via the TATA box. A, Sequence alignment of the HLA–DRB1*04, DRB1*07, DRB1*09, DRB4A, and DRB4B promoters. Shaded areas show the core sequences of the transcription factor target sites. Sequence variations among the alleles/genes are indicated. B, Binding of nuclear extracts from the THP-1 and BJAB cell lines to the TATA boxes of DRB1*04/*07, DRB1*09, DRB4A, and DRB4B. Surface plasmon resonance was measured and normalized to the random DNA in order to calculate the percentage increase in binding. For isolation of nuclear extracts, cells were either left untreated (solid bars) or were treated with 1,000 units/ml of interferon-␥ for 48 hours (hatched bars). Values are the mean and SD. C and D, Competition assays for the binding of nuclear extracts to TATA boxes DRB4B and DRB4A were performed by electrophoretic mobility shift assay. C, Polyacrylamide gel showing that the radiolabeled TATA box DRB4B oligonucleotide was competed out by increasing concentrations of unlabeled TATA box DRB4B and TATA box DRB4A. Arrowhead indicates the band corresponding to nuclear proteins bound to the radiolabeled TATA box DRB4B oligonucleotide. The radiolabeled TATA box DRB4A oligonucleotide was competed out in the same way. D, The intensity of the upper band was calculated in relation to the intensity of the corresponding lane using the phosphorimager. There was a reduction in signal strength resulting from the competition between 32P-labeled DRB4A and unlabeled DRB4 oligonucleotides for binding to nuclear proteins. Binding without competition was set at 1. DRB1*07-, and DRB1*09-positive patients for whom initial CRP values were available. A total of 73 RA patients were evaluated; 35 HLA CLASS II GENES IN RA SUSCEPTIBILITY AND PROGRESSION of them had initial CRP values ⱖ1.5 mg/dl (Table 2). These 35 patients shared a total of 2 DRB1*09, 11 DRB1*07, and 34 DRB1*04 alleles, of which 27 were *0401; 7 alleles were *0403, *0404, and *0408, respectively. The allele distribution in the 38 patients with initial CRP values ⬍1.5 mg/dl was comparable, with 4 DRB1*09, 12 DRB1*07, and 38 DRB1*04 alleles, of which 32 were *0401. All 73 RA patients were evaluated for associations with the DRB4B promoter and the splice variant. Most interestingly, the association with the stronger DRB4B promoter was more pronounced in the group with rapid disease progression (22%) than in the group with moderate disease progression (7%). The 1-sided P value for the difference between these 2 groups was 0.035, with 95% confidence intervals ranging from 0.71 to 0.99 (relative risk) and from 0.064 to 0.098 (odds ratio), indicating statistical significance (Table 2). In contrast, the DRB4 splice variant preventing the formation of a functional DRB4B gene product was not associated with either of the patient groups (data not shown). Five patients carried the combination DRB1*04/*07 and were therefore excluded from the splice variant and DRB4B promoter analyses since an exact assignment of DRB4 to the corresponding DRB1 locus is not possible. It is important to note that all of the DRB4 genes we analyzed share the same exon 2 sequence that forms the epitope-binding groove (data not shown). The association between regulatory polymorphisms and disease progression is therefore not perturbed by additional structural polymorphisms. Association of RA disease susceptibility with the differential expression of HLA–DRB4. We next investigated whether the differential expression of the DRB4 gene also affected the susceptibility to RA. To that end, Table 2. Correlation between an elevated expression of DRB4 and radiographic progression in early RA* RA patient group HLA–DR allele CRP ⱖ1.5 mg/dl at study entry (n ⫽ 35) DR4, 49 (34) DR7, 16 (11) DR9, 3 (2) DR4, 50 (38) DR7, 16 (12) DR9, 5 (4) CRP ⬍1.5 mg/dl at study entry (n ⫽ 38) Association of DRB1*04, *07, and *09 alleles with DRB4B† 22 (10/46) 7 (3/46) * Values are the frequency (no. of positive alleles or no. of positive alleles/total). RA ⫽ rheumatoid arthritis; CRP ⫽ C-reactive protein. † The difference between the 2 groups is 15%. One-sided P ⫽ 0.035, with confidence intervals ranging from 0.71 to 0.99 (relative risk) and from 0.064 to 0.98 (odds ratio), indicating statistical significance. 2785 Figure 5. Differential association of HLA–DRB1*04, DRB1*07, and DRB1*09 alleles with the HLA–DRB4 splice variant, as determined by amplification and sequencing of the respective gene segments in rheumatoid arthritis (RA) patients and controls. ⴱ ⫽ P ⫽ 0.032. 28 healthy DR4-, DR7-, or DR9-positive controls were included in the analysis, and we compared the associations with the DRB4B promoter and splice variant in the RA patients with those in the controls. In the RA patients, a total of 71 DRB1*04, 21 DRB1*07, and 5 DRB1*09 alleles linked to the DRB4A promoter were analyzed for an association with the splice variant. In the controls, we analyzed 9 DRB1*04, 14 DRB1*07, and 1 DRB1*09 alleles. As shown in Figure 5, the frequencies of both the DRB1*07- and DRB1*09-associated splice variants were comparable among the RA patients and the controls. In contrast, the frequencies for the DRB1*04-associated splice variant differed significantly between the RA patients (1.4%) and the controls (22%) (1-sided and 2-sided P ⫽ 0.032). The frequencies of associations between the DRB1*04, *07, and *09 alleles and the DRB4B promoter were also comparable between the RA patients and controls. Again, the DRB1*07 allele showed the highest frequencies (data not shown). Lack of augmentation of the differential expression of HLA–DR genes by the Y, S, and CAAT boxes. The alignment of the Y, S, and CAAT boxes of the DRB1*04, DRB1*07, DRB1*09, and DRB4 promoters revealed several polymorphic residues in and around the various boxes (11), but there was only 1 G3 A substitution 5⬘ of the Y box that distinguished the DRB4A from the DRB4B promoter. However, this substitution did not lead to functional consequences, since the binding of BJAB nuclear factors to the DRB4A and DRB4B Y box oligonucleotides showed comparable increases (14% and 15%, respectively). Likewise, there was a 22% and a 30% increase for the THP-1 nuclear factors compared with the random oligonucleotide (data not shown). Additional structural differences between the Y, S, and 2786 HELDT ET AL CAAT boxes of the DRB1*04, *07, and *09 promoters consisted of 2 single substitutions within the S box and 3⬘ of the CAAT box. These substitutions resulted in only minor differences in the binding of nuclear factors to the S and CAAT box oligonucleotides (14–24% and 1–9%, respectively) (data not shown). DISCUSSION We investigated the differential expression of HLA class II genes as a mechanism involved in the susceptibility and/or progression of RA, concentrating on the molecular analysis of the RA-associated DR4, as well as the neutral DR7 and DR9, superhaplotypes. We compared the various promoters encoded by these haplotypes in an effort to find sequence polymorphisms resulting in functional consequences. Assuming that the specific binding of nuclear factors to the various promoter sequences was related to the transcription activity, we used surface plasmon resonance technology. The promoter sequences we tested correspond to the known S, X, Y, CAAT, and TATA boxes (25), and the nuclear factors were derived from the human monocyte and lymphoma cell lines THP-1 and BJAB (26), respectively. Our data suggest that IFN ␥ -induced upregulation of surface HLA–DR expression in the monocyte cell line is mediated via the X box, since stimulation with IFN␥ led to an increased binding of THP-1 nuclear factors to the various X box oligonucleotides alone (Figure 3). Our results are therefore consistent with previous speculations that the overall expression level, as well as the on/off switch, is regulated via the locusspecific X1 box, which binds the unique RFX proteins and CIITA (15). The identification of binding sites for IFN␥-inducible transcription factors in the CIITA promoter (14) further supports our findings. Moreover, the increased binding of THP-1 nuclear factors to the X boxes was strongest for the DRA-specific sequence, which supports the previously held view that DRA expression is the rate-limiting step in HLA–DR surface expression (16). In BJAB cells, only a minor upregulation of surface HLA–DR expression in response to IFN␥ stimulation was observed. The small shift in the mean fluorescence intensity (Figure 2) correlated with a slightly increased binding of nuclear factors to the X box of DRA (Figure 3) and to the TATA box of the DRB1 alleles (Figure 4). The observed incapacity of BJAB cells to up-regulate surface HLA–DR expression may be characteristic of this particular cell line, but the situation is also reminiscent of that of peripheral blood B lymphocytes, which have recently been described to be refractory to stimulation by LPS and phorbol myristate acetate (20). From our data, it can be concluded that monocytes are very important for the response to environmental stimuli since they are highly activational and capable of affecting the Th1/Th2 balance by differentially expressing HLA class II genes. Likewise, the differential expression of murine class II MHC genes has been shown to be restricted to macrophages and to be associated with protective/suppressive effects via the T cell cytokine profile (5). One approach to addressing the impact of differential HLA–DRB4 expression on the T cell cytokine profile in human RA will involve the analysis of rheumatoid factor isotypes in correlation with the respective DRB4 promoters. Interestingly, our data suggest a differential expression of the relatively nonpolymorphic DRB4 gene, while the highly polymorphic DRB1 alleles seem to be expressed at comparable levels. From an evolutionary point of view, the retention of a second functional, yet coding, nonvariable DRB gene has for a long time remained an enigma. However, the concept that the DRB4 gene regulates the immune response via differential expression is intriguing. Our data thus indicate that selection may also work on the regulatory regions, which were previously thought to be merely flanking sequences in linkage disequilibrium with the coding regions. Independent of stimulation, there was a 1.6–1.8fold increase in the binding activity of nuclear factors to the TATA box DRB4B oligonucleotide compared with the TATA box DRB4A oligonucleotide. Even though this difference seems insignificant at first glance, it was confirmed by EMSA as an alternative assay (Figure 4). A 2-fold difference in promoter activity can well be reconciled with previously published results by Leen and Gorski (18), who reported 3-fold and 5-fold higher levels for the DRB4B-driven pre-mRNA and mRNA, respectively. Measuring the surface plasmon resonance yielded little evidence for an involvement of the S, Y, and CAAT boxes in the differential expression of the DRB1 and DRB4 genes we tested. However, the use of recombinant transcription factors rather than nuclear extracts may result in higher sensitivities. Contrary to the overall level of expression, the fine-tuning of the DRB4A and DRB4B promoters is regulated at a ubiquitous regulatory element, the TATA box. We therefore assume that the differential expression of the DRB4 gene will not be restricted to the classic antigen-presenting cell, including the dendritic cells, but will hold true for any cell expressing HLA class HLA CLASS II GENES IN RA SUSCEPTIBILITY AND PROGRESSION II genes. Interesting candidates include the synoviocytes, which are known to participate in the autoimmune process in RA. In summary, we would like to propose a model whereby both the susceptibility to the development of RA and the progression of RA depend upon the differential expression of the relatively nonpolymorphic HLA–DRB4 gene. Upon grouping the RA patients according to their CRP levels at study entry (which, in this cohort, was the strongest indicator of disease progression), the more potent DRB4B promoter segregated with radiographic progression (Table 2). Interestingly, the frequency of DRB1*0401 alleles was almost identical in RA patients with high and low initial CRP values, ruling out an effect of epitope selection on disease progression. In addition, there was a concomitant reduction in the splice variant frequency associated with DRB1*04 alleles among RA patients, indicating a protective effect of low DRB4 expression. Our speculations about the mechanism include the enhanced signaling between antigen-presenting cells and T cells due to enhanced antigen presentation, favoring the development of inflammatory responses. Comparable shifts toward Th1 responses have previously been reported to result from increases in antigen dose and affinity, respectively (27,28). REFERENCES 1. Lang B, Melchers I, Urlacher A, Tanzi-Fetta RF, Kohlbrenner S, Tongio MM, et al. HLA-DR1 and DRw6 association in DR4negative rheumatoid arthritis patients. Rheumatol Int 1990;10: 171–5. 2. Listing J, Rau R, Muller B, Alten R, Gromnica-Ihle E, Hagemann D, et al, for the Berlin Collaborating Rheumatological Study Group. HLA-DRB1 genes, rheumatoid factor, and elevated C-reactive protein: independent risk factors of radiographic progression in early rheumatoid arthritis. J Rheumatol 2000;27:2100–9. 3. Weyand CM, McCarthy TG, Goronzy JJ. Correlation between disease phenotype and genetic heterogeneity in rheumatoid arthritis. J Clin Invest 1995;95:2120–6. 4. Gregersen PK, Silver J, Winchester RJ. The shared epitope hypothesis: an approach to understanding the molecular genetics of susceptibility to rheumatoid arthritis. Arthritis Rheum 1987;30: 1205–13. 5. Baumgart M, Moos V, Schuhbauer D, Muller B. Differential expression of major histocompatibility complex class II genes on murine macrophages associated with T cell cytokine profile and protective/suppressive effects. Proc Natl Acad Sci U S A 1998;95: 6936–40. 6. Janitz M, Mitchison A, Reiners-Schramm L, Lauster R. Polymorphic MHC class II promoters exhibit distinct expression patterns in various antigen-presenting cell lines. Tissue Antigens 1997;49: 99–106. 7. Cowell LG, Kepler TB, Janitz M, Lauster R, Mitchison NA. The distribution of variation in regulatory gene segments, as present in MHC class II promoters. Genome Res 1998;8:124–34. 8. O’hUigin C, Satta Y, Hausmann A, Dawkins RL, Klein J. The 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 2787 implications of intergenic polymorphism for major histocompatibility complex evolution. Genetics 2000;156:867–77. Mitchison NA, Roes J. Patterned variation in murine MHC promoters. Proc Natl Acad Sci U S A 2002;99:10561–6. Marsh SG, Bodmer JG, Albert ED, Bodmer WF, Bontrop RE, Dupont B, et al. Nomenclature for factors of the HLA system, 2000. Hum Immunol 2001;62:419–68. Louis P, Eliaou JF, Kerlan-Candon S, Pinet V, Vincent R, Clot J. Polymorphism in the regulatory region of HLA-DRB genes correlating with haplotype evolution. Immunogenetics 1993;38:21–6. Sutton VR, Kienzle BK, Knowles RW. An altered splice site is found in the DRB4 gene that is not expressed in HLA-DR7,Dw11 individuals. Immunogenetics 1989;29:317–22. Fontes JD, Jiang B, Peterlin BM. The class II trans-activator CIITA interacts with the TBP-associated factor TAFII32. Nucleic Acids Res 1997;25:2522–8. Moreno CS, Beresford GW, Louis-Plence P, Morris AC, Boss JM. CREB regulates MHC class II expression in a CIITA-dependent manner. Immunity 1999;10:143–51. Mitchison NA. Genetic variation in the MHC class II promoter: lessons for regulation and for comparative genomics. Folia Biol 2001;47:183–6. Louis P, Vincent R, Cavadore P, Clot J, Eliaou JF. Differential transcriptional activities of HLA-DR genes in the various haplotypes. J Immunol 1994;153:5059–67. Vincent R, Louis-Plence P, Gaillard F, Clot J, Eliaou JF. Qualitative and quantitative analysis of HLA-DRB gene expression. J Rheumatol 1997;24:225–6. Leen MP, Gorski J. DRB4 promoter polymorphism in DR7 individuals: correlation with DRB4 pre-mRNA and mRNA levels. Immunogenetics 1997;45:371–8. Kerlan-Candon S, Louis-Plence P, Wiedemann A, Combe B, Clot J, Eliaou JF, et al. Specific overexpression of rheumatoid arthritis–associated HLA–DR alleles and presentation of lowaffinity peptides. Arthritis Rheum 2001;44:1281–92. Czerwony G, Alten R, Gromnica-Ihle E, Hagemann D, Reuter U, Sorensen H, et al. Differential surface expression of HLA-DRB1 and HLA-DRB4 among peripheral blood cells of DR4 positive individuals. Hum Immunol 1999;60:1–9. Arnett FC, Edworthy SM, Block DA, McShane DJ, Fries JF, Cooper NS, et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988;31:315–24. Sykes BC. DNA in heritable disease [letter]. Lancet 1983;2:787–8. Scharf SJ, Griffith RL, Erlich HA. Rapid typing of DNA sequence polymorphism at the HLA-DRB1 locus using the polymerase chain reaction and nonradioactive oligonucleotide probes. Hum Immunol 1991;30:190–201. Blaesing F, Weigel C, Welzeck M, Messer W. Analysis of the DNA-binding domain of Escherichia coli DnaA protein. Mol Microbiol 2000;36:557–69. Benoist C, Mathis D. Regulation of major histocompatibility complex class-II genes: X, Y and other letters of the alphabet. Annu Rev Immunol 1990;8:681–715. Clements GB, Klein G, Povey S. Production by EBV infection of an EBNA-positive subline from an EBNA-negative human lymphoma cell line without detectable EBV DNA. Int J Cancer 1975;16:125–33. Hemmer B, Stefanova I, Vergelli M, Germain RN, Martin R. Relationships among TCR ligand potency, thresholds for effector function elicitation, and the quality of early signaling events in human T cells. J Immunol 1998;160:5807–14. Ise W, Totsuka M, Sogawa Y, Ametani A, Hachimura S, Sato T, et al. Naive CD4⫹ T cells exhibit distinct expression patterns of cytokines and cell surface molecules on their primary responses to varying doses of antigen. J Immunol 2002;168:3242–50.