The Journal of Immunology

IL-6 Modulates CD5 Expression in B Cells from Patients with Lupus by Regulating DNA Methylation1 Soizic Garaud,2* Christelle Le Dantec,2* Sandrine Jousse-Joulin,† Catherine Hanrotel-Saliou,† Alain Saraux,*† Rizgar A. Mageed,‡ Pierre Youinou,3*† and Yves Renaudineau*† B lymphocytes from patients with systemic lupus erythematosus (SLE) are characterized by reduced expression levels of membrane CD5. Recent studies from our laboratory have revealed that the level of membrane CD5 is determined by the relative level of two alternative CD5 isoforms; CD5-E1A, which is expressed on the membrane, and CD5-E1B, which is retained in the cytoplasm. Using bisulfite sequencing and methylation-sensitive endonuclease assays we show that the promoter for the alternative CD5-E1B isoform is demethylated in B cells from patients with SLE but not in healthy controls. We go on to show that differential methylation is more pronounced following BCR engagement. As a result of this demethylation, CD5-E1B mRNA is transcribed at the expense of CD5-E1A mRNA transcription. We provide further evidence that production of high IL-6 levels by SLE B cells abrogates the ability of SLE B cells to induce DNA methyl transferase (DNMT1) and then to methylate DNA, an effect that is reversed in the presence of a blocking Ab to the IL-6 receptor. The pattern of demethylation of CpG islands in the CD5-E1B promoter in SLE B cells is similar to those in B cells from healthy controls stimulated in the presence of IL-6, or treated with the methylation inhibitor PD98059. The study reveals that engagement of the BCR with constitutive IL-6 down-regulates the level of membrane CD5, which negatively regulates BCR signaling, in SLE B cells. This altered signaling could, in turn, promote the activation and expansion of autoreactive B cells in SLE patients. The Journal of Immunology, 2009, 182: 5623–5632.

S

ystemic lupus erythematosus (SLE)4 is associated with diverse clinical manifestations (1). The main features of autoimmunity in SLE are B cell hyperactivity, spontaneous lymphocyte proliferation, and the production of pathogenic Abs to self-Ags. B cell abnormalities in SLE also include excess cytokine production, autoantigen presentation to T cells and modulation of the function of other immune cells (2). Thus, SLE is generally considered a B cell disease, a theme strengthened by the efficacy of therapies targeting B cells. For example, therapeutic approaches using depleting anti-CD20 has proved to be highly beneficial in treatment of SLE (3). Further, neutralization of cytokines that promote B cell responses such as IL-6, or interruption of cognate T cell and B cell interactions has been successful in early clinical trials (4, 5). All this provides the rationale for further investigation

*Research Unit EA2216 Immunology and Pathology, IFR148 ScInBioS, Universite´ de Brest and Universite´ Europe´enne de Bretagne; †CHU Brest, Hoˆpital Morvan and Hoˆpital de la Cavale Blanche, Brest, France; ‡William Harvey Research Institute, Barts and the London School of Medicine and Dentistry, Queen Mary University of London, London, United Kingdom Received for publication July 30, 2008. Accepted for publication February 18, 2009. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1

This work was supported by grants from the Conseil Regional de Bretagne, the Conseil Ge´ne´ral du Finiste`re, and the French Ministry for Education and Research.

2

S.G. and C.L.D. contributed equally to the work.

3

Address correspondence and reprint requests to Prof. Pierre Youinou, Laboratory of Immunology, Brest University Medical School Hospital, BP824, F-29609, Brest, France. E-mail address: [email protected]

4

Abbreviations used in this paper: SLE, systemic lupus erythematosus; ChIP, chromatin immunoprecipitation; DNMT, DNA methyl transferase; HERV, human endogenous retrovirus; HC, healthy control; MBD, methyl-CpG-binding domain protein; LTR, long terminal repeat; SLEDAI, SLE disease activity index. Copyright © 2009 by The American Association of Immunologists, Inc. 0022-1767/09/$2.00

www.jimmunol.org/cgi/doi/10.4049/jimmunol.0802412

of mechanisms of B cell involvement in driving autoimmunity and to develop more selective therapeutic targets. The central role played by B cells in immunity necessitates that its responses are tightly regulated. B cell responses are initiated by signaling through the BCR. Signaling initiated following BCR engagement is regulated by coreceptors and by a network of protein tyrosine kinases and phosphatases. Recent findings suggest that defects in BCR-mediated signaling can result in lupus autoimmunity. For example, there is an association between SLE autoimmunity and mutations in a number of genes that encode B cellspecific signaling molecules including protein tyrosine kinases, non-receptor phosphatase type 22, B cell scaffold protein with ankyrin repeats 1 and the inhibitory IgG Fc␥RIIb (6). In addition to direct evidence for the effect of mutations in signaling molecules on BCR-mediated signaling, the contribution of epigenetic factors has also been proposed (7). The most commonly observed epigenetic abnormality implicated in SLE pathology is altered DNA methylation at the 5-carbon position of cytosines of CpG dinucleotides. DNA methylation is regulated by three DNA methyl transferases (DNMTs), DNMT1, DNMT3a, and DNMT3b. DNMT1 preferentially targets hemi-methylated DNA over non-methylated DNA. DNMT3a and DNMT3b, in contrast, exhibit de novo activities. The methyl-CpG-binding domain (MBD) protein-4 is a DNA glycosylase that acts preferentially on hemi-methylated CpGs and initiates demethylation by replacing a 5-methylcytosine with an unmethylated cytosine (8). The action of MBD2 on DNA de-methylation has been suggested, but remains controversial (9). DNA de-methylation activates the expression of many genes, such as CD21 in B cells in patients with rheumatoid arthritis (10), CD40 ligand in T cells from patients with SLE (11), and the expression of silenced retroviral genes (12). Increased expression of genes resulting from de-methylation was confirmed using DNA methylation inhibitors in T cells from patients with SLE. Among those inhibitors the ras signal blocker PD98059 appears to be the more

5624

IL-6 CONTROLS DNA METHYLATION AND CD5 IN LUPUS B CELLS

Table I. Demographic and clinical characteristics of SLE patients and HCs included in the study

Age (years) Female (%) Ethnicity, Caucasian SLEDAI score SLE manifestations Diseases duration Skin disease Arthritis CNS disease Lupus nephritis Heart ANA Anti-dsDNA Medication usage (%) Steroids Antimalarials Cytotoxic agents a

Patients

Controls

n ⫽ 25

n ⫽ 25

47.6 ⫾ 15.1 关22–74兴a 19 (76) 25 3.7 ⫾ 5.3 关0–19兴

43.3 ⫾ 10.7 关28–68兴 17 (68) 25 –

11.1 ⫾ 8.1 关0–19兴 52% 72% 12% 48% 32% 100% 64% 60% 72% 20%

– – – – – – – – – – –

Mean ⫾ SD 关min ⫺ max兴.

relevant inhibitor because it induces cellular defects similar to those observed in SLE (13). B cells that express the CD5 protein, also known as B1 cells, primarily express BCRs that are autoreactive, have a reduced capacity to enter the cell cycle, and have a longer lifespan. One model for the role of CD5 in intracellular signaling suggests that surface CD5 modulates signaling from the BCR and thereby controls autoreactivity (14). According to this model, it is necessary that the level of membrane CD5 is maintained to control the threshold of BCR-mediated signaling. In humans, three mechanisms have been shown to be involved in regulating the level of membrane CD5. These are shedding (15), internalization of the protein (16), and transcriptional regulation through altered level of expression of two alternative CD5 isoforms (17). The first isoform, designated CD5-E1A, corresponds to the full-length CD5 protein which is synthesized and translocated to the cell membrane. The second isoform derived from an alternative promoter, designated CD5-E1B, encodes a truncated form of the protein that lacks a leader peptide and is retained intracellularly. Indeed, CD5-E1B is a fusion transcript from a new exon (4) that incorporates the 5⬘ long terminal repeat (LTR) of a human endogenous retrovirus (HERV) sequence integrated at the time of old and new world monkeys divergence estimated at ⬃25–30 million years ago (18). We recently observed that transcription of mRNA for this isoform inversely correlates with the level of DNMT1 mRNA (19). Here, we provide evidence that relative to B cells from healthy controls (HCs), the level of DNA methylation in BCR-mediated B cell activation in patients with SLE is reduced. A consequence of this hypomethylation is increased expression of CD5-E1B. Excess production of IL-6 augments CD5-E1B transcription. Based on this observation, we propose that modulation of B cell autoreactivity in SLE could be achieved by targeting IL-6.

Materials and Methods B lymphocyte isolation PBMCs were isolated from the blood of 25 patients with SLE and 25 HCs by centrifugation on Ficoll-Hypaque. All patients fulfilled the 1982 American College of Rheumatology criteria for SLE (20). SLE activity was assessed by the SLE disease activity index (SLEDAI), and those with SLEDAI of ⱖ5 were considered active. The characteristics of SLE patients and HCs are shown in Table I. The cells were stained with FITC-anti-CD19 and PE-anti-CD5 Abs, and CD5⫺CD19⫹ B cells sorted on an Epics Elite FACS (Beckman-Coulter). All sorted cells were ⬎98% CD19⫹. Informed consent was obtained from the patients before taking blood, and the study

protocol approved by the Institutional Review Board at Brest University. The Daudi human B cell line was purchased from American Type Culture Collection.

FACS analysis FITC-anti-CD19 (clone J4-119) and PE-anti-CD5 (clone BL1a) were obtained from Beckman-Coulter, whereas anti-DNMT1 and anti-p27kip1 were obtained from Abcam. Intracellular staining was performed after permeabilization of the cells with 70% methanol. Binding of primary unconjugated Abs was revealed with FITC-conjugated anti-mouse Abs (Jackson ImmunoResearch). In pilot experiments, the number of CD5 molecules per cell was quantified by determining the amount of Ab binding to the cells (ABC) at saturating concentrations, using the Quantum Simply Cellular kit (Flow Cytometry Standards). Arbitrary ABC value was then determined from a standard ABC curve generated from the mean fluorescence intensity obtained from the FACS analysis of 50 ␮l calibrated microspheres stained with 20 ␮l of the same anti-CD5 Ab.

Cell culture FACS-sorted B cells were suspended in RPMI 1640 supplemented with 10% heat-inactivated FCS, 2 mM L-glutamine, 200 U/ml penicillin and 100 ␮g/ml streptomycin. B lymphocytes were seeded at 2 ⫻ 105 cells per well, and incubated with 1 ␮g/ml anti-IgM Ab-coated Sepharose beads (BioRad) and 10 U/ml IL-2, in the presence or absence of 10 – 40 ng/ml anti-IL6RAb (R&D Systems), or 100 ng/ml rhIL-6 (Immuno Tools). Repression of DNMTs was achieved by incubating the cells with 50 ␮M of the ras signal blocker PD98059. IL-6 and IFN-␥ were detected in sera and IL-6 in the supernatant of cultured cells using ELISA kits according to the manufacturer’s instructions (Beckman Coulter).

mRNA extraction and quantitative RT-PCR Total mRNA was extracted using the RNAble method (Eurobio), and cDNA synthesized by reverse transcription in 20 ␮l volume with Superscript II RNase H-RT (Invitrogen Corporation). Quantitative RT-PCR was conducted in 20 ␮l mixtures containing 50 ng template cDNA, 1X Sybr Green PCR Master mix (Applied Biosystems), and 500 nM of each primer (Table II). As described in detail elsewhere (17), CD5 promoter usage was evaluated using two sets of primers. All assays included a negative control which consisted of the reaction mixture with no template, and a positive control which consisted of the mixture with 18S rRNA. Comparison of cycle thresholds was completed with the 2⫺⌬⌬ct method using 18S as an internal control.

RACE The 5⬘ transcript ends were amplified from mRNA using SMART-RACE kit (Clontech). As described previously (17), the first strand of cDNA was synthesized using the sense UPM primer and the gene-specific antisense primer CD5 E5 (Table II). The PCR protocol included an initial denaturation step at 94°C for 5 min, starting 5 touchdown-PCR cycles of denaturation at 94°C for 30 s and annealing at 72°C for 3 min. These cycles were followed by another 5 cycles of 94°C for 30 s, 70°C for 30 s, and 72°C for 3 min, then with a decreasing temperature for 35 cycles of 94°C for 30 s, 68°C for 30 s, and 72°C for 3 min. A nested PCR was performed using the sense NUP primer and the gene-specific antisense primer CD5 E3. The second PCR round was for 40 cycles of 30 s at 94°C, 1 min at 56°C, and 1 min at 72°C with a final extension at 72°C for 10 min.

Methylation-specific PCR This assay is based on the inability of some restriction enzymes to digest a methylated 5⬘-CmG-3⬘ site. Genomic DNA was purified using QIAmp 96 DNA blood kit (Qiagen) and digested with 20 U of the methylation-sensitive restriction enzymes HpaII, HaeII, FauI, HgaI, or the methylationinsensitive restriction enzyme MspI for 3 h at 37°C. Undigested genomic DNA was used as positive control. The PCR primers were positioned upstream and downstream of the recognition site in the promoters of E1A and E1B of the cd5, cd19, cd70, Pax5, Syk, and HRES-1 genes (11, 19, 21, 22). The PCR protocol included an initial denaturation at 94°C for 5 min, followed by 35 cycles of denaturation at 94°C for 30 s, annealing at 56°C for 1 min, and primer extension at 72°C for 1 min; PCR cycles were followed by final extension at 72°C for 10 min. The PCR products were separated on agarose gel and visualized with 0.5 ␮g/ml ethidium bromide.

The Journal of Immunology

5625

Table II. Oligonucleotide primers used in the studya Oligonucleotide

Sequence

Method

CD5-E1A sense CD5-E1B sense CD5-E3 antisense DNMT1 sense DNMT1 antisense DNMT3a sense DNMT3a antisense DNMT3b sense DNMT3b antisense MBD2 sense MBD2 antisense MBD4 sense MBD4 antisense 18S sense 18S antisense CD5-E5 antisense CD5-E3 antisense CD5 promoter-1B sense CD5-E1B antisense CD5 promoter-1A sense CD5-E1A antisense CD5 promoter-1B sense CD5-E1B antisense CD5 promoter-1B sense CD5-E1B antisense CD5 promoter-1A sense CD5-E1A antisense CD5 promoter-1A sense CD5-E1A antisense CD5-E1B sense CD5-E1B antisense GAPDH sense GAPDH antisense CD19 promoter sense CD19 promoter antisense CD70 promoter sense CD70 promoter antisense Pax5 promoter sense Pax5 promoter antisense Syk promoter sense Syk promoter antisense HRES-1 promoter sense HRES-1 promoter antisense

5⬘-ATGCCCATGGGGTCTCTGCAAC-3⬘ 5⬘-GCTGAACACAGGGAAGGAAC-3⬘ 5⬘-GCCTGGAAATCTGGGTCATA-3⬘ 5⬘-CCTGTACCGAGTTGGTGATGGT-3⬘ 5⬘-CCTTCCGTGGGCGTTTC-3⬘ 5⬘-CTCCTGTGGGAGCCTCAATGTTACC-3⬘ 5⬘-CAGTTCTTGCAGTTTTGGCACATTCC-3⬘ 5⬘-CTCGAAGACGCACAGCTGACGAC-3⬘ 5⬘-CCTATAACAACGGCAAAGACCGAGC-3⬘ 5⬘-CCATGGAACTACCCAAAGGTCTT-3⬘ 5⬘-CAGCAGATAAAAGGGTCTCATCATT-3⬘ 5⬘-TCTAGTGAGCGCCTAGTCCCAG-3⬘ 5⬘-TTCCAATTCCATAGCAACATCTTCT-3⬘ 5⬘-GGCTACCACATCCAAGGAAGGCAG-3⬘ 5⬘-CCAATTACAGGGCCTCGAAAGAGTC-3⬘ 5⬘-AGTCTCTGGCAGCTGCTTCAAGAA-3⬘ 5⬘-TGCCATCCGTCCTTGAGGTAGAC-3⬘ 5⬘-GTGAAGGGCTGCTTACTGCACACAGC-3⬘ 5⬘-CAGCCACTGCGTTGATCCTCCACAG-3⬘ 5⬘-CTGGAAGGGTAAAGCAGGGTTCTC-3⬘ 5⬘-GGAGTCTGCAACAAGAACTGGCATC-3⬘ 5⬘-TTTTATTTGTGAAATGGAAAGTTGT-3⬘ 5⬘-AAATTCCCAAAACCAATCCTATC-3⬘ 5⬘-TGTGAAATGGAAAGTTGTGTTTATT-3⬘ 5⬘-AACATACCATAAATAATTAAACCAC-3⬘ 5⬘-GAATTGGTATTATGTTGTTTATTTTT-3⬘ 5⬘-CTCCCTACCAACCTAAAAACTACTC-3⬘ 5⬘-ATTGTTTTAGTTTTGGGTATTTTGG-3⬘ 5⬘-CCCCCTACAATCTCTCTTACACTAA-3⬘ 5⬘-TCTGCCAAAGAGGTTCAAGC-3⬘ 5⬘-TGCATGCACCGGTAATTAGA-3⬘ 5⬘-TACTAGCGGTTTTACGGGCG-3⬘ 5⬘-TCGAACAGGAGGAGCAGAGAGCGA-3⬘ 5⬘-AGCGTGGCAGGGAGGAGGCAAGTGTT-3⬘ 5⬘-GCGAGGAGGTGGCATGGTGGTCAGA-3⬘ 5⬘-TCACCCAAGGTCAGGAGTTC-3⬘ 5⬘-CCATCTCAACTCACCCCAAG-3⬘ 5⬘-GCAATAGTCAGGACCCCAAC-3⬘ 5⬘-TTCTCGCCAACATCACAAGA-3⬘ 5⬘-GGCAGCCCCACCTTCTCT-3⬘ 5⬘-CGCGGCTCTTCCTCATTT-3⬘ 5⬘-GCATATGCACTGGGAAAGGT-3⬘ 5⬘-CCGCCTTTTCAAGTTTCCTC-3⬘

Quantitative PCR Quantitative PCR Quantitative PCR Quantitative PCR Quantitative PCR Quantitative PCR Quantitative PCR Quantitative PCR Quantitative PCR Quantitative PCR Quantitative PCR Quantitative PCR Quantitative PCR Quantitative PCR Quantitative PCR 5⬘RACE 5⬘RACE Methylation-PCR Methylation-PCR Methylation-PCR Methylation-PCR Methylation-PCR Bisulfite sequencing Bisulfite sequencing Bisulfite sequencing Bisulfite sequencing Bisulfite sequencing Bisulfite sequencing Bisulfite sequencing ChIP ChIP ChIP ChIP Methylation PCR Methylation PCR Methylation PCR Methylation PCR Methylation PCR Methylation PCR Methylation PCR Methylation PCR Methylation PCR Methylation PCR

a

E1A, exon 1A; E1B, exon 1B; DNMT, DNA methyltransferase; MBD, methyl-CpG binding domain protein.

Bisulfite sequencing To determine the methylation status of DNA, non-methylated cytosines were converted to uridines by bisulfite treatment using the EZ-DNA methylation-Gold kit according to the manufacturer’s instructions (Zymo Research). Unmodified DNA (100 ng) was amplified 40 times at 56°C using specific primers. The bisulfite-converted DNA was amplified by nested PCR using two rounds of 40 cycles each at 56°C with primers specific for methylated cytosines (Table II). PCR products were purified using the high pure PCR product purification kit (Roche), and directly sequenced with internal primers by means of the BigDye Terminator Cycle Sequencing kit using an automated ABI-310 genetic analyzer (Applied Biosystems). The electrophoregram T and C peaks were quantified and methylation status determined as [peak (C)/peak (T) ⫹ peak (C)] ⫻ 100. At the same time, the unmodified DNA was amplified and sequenced using specific primers.

Chromatin immunoprecipitation Chromatin immunoprecipitation (ChIP) was conducted using the EpiQuik kit (Epigentek Group) according to the manufacturer’s instructions to evaluate activation of the CD5-E1B promoter. In brief, sonicated DNA (200 – 1000 bp) was transferred into strip wells precoated either with mouse antiRNA polymerase II, or with a nonspecific mouse IgG, used as a negative control. After a 90-min incubation at room temperature and extensive washes, precipitated DNA-protein complexes were treated with 250 ␮g/ml proteinase K in the DNA release buffer for 15 min, and left in the same buffer for 90 min at 65°C. The DNA samples were collected by the P-spin

columns, washed with ethanol, and eluted. Using purified DNA as a template, PCR was performed using GAPDH and CD5-E1B-specific primers (Table II) and 40 cycles at 56°C. PCR products were separated on agarose gel, and visualized with 0.5 ␮g/ml ethidium bromide.

Computational promoter analysis Putative transcription factor binding sites were identified by using Alibaba, version 2.1, the transcription element search system (http://www.cbil. open.edu/tess/index.html) and Genomatix (http://www.genomatix.de).

Statistics The results were expressed as arithmetic means with SD. Comparisons were made using the Mann-Whitney U test for unpaired data and the Wilcoxon test for paired data. Correlations were established using Spearman’s rank correlation.

Results CD5 expression in B cells The percentage of CD5-expressing B cells was similar in the 25 SLE patients and 25 HCs (Fig. 1A). However, membrane density of CD5 on B1 cells was lower in the patients (Fig. 1B) compared with the controls (49,971 ⫾ 15,124 vs 80,703 ⫾ 22,462; p ⬍ 0.001). A representative example is depicted in Fig. 1C.

5626

IL-6 CONTROLS DNA METHYLATION AND CD5 IN LUPUS B CELLS

FIGURE 1. CD5 expression in B cells. A, A scattergram depicting percentage of CD5-expressing B cells in 25 SLE patients and 25 HCs. B, The number of CD5 molecules per cell is given in the scattergram. The numbers are expressed as the amount of anti-CD5 Ab bound to the cell membrane. C, Representative FACS profile for cell surface expression of CD5 in blood B cells from one SLE patient (bold line) and one HC (thin line). D, Quantitative RT-PCR results presenting as histograms revealing that a 24-h stimulation of B cells with antiIgM increased CD5-E1B transcription in B cells from SLE patients but not HCs. Incubation of antiIgM-activated B cells from HCs with PD98059 resulted in up-regulation of CD5-E1B. Mean and SD of data from 10 SLE patients and 15 HCs are shown. E, CD5 5⬘-RACE analysis of cDNA revealing that CD5-E1B (639 bp) and CD5-E1A (259 bp) are induced when B cells from the HCs were stimulated with anti-IgM in the presence of PD98059, whereas only CD5-E1A (259 bp) is induced in the presence of anti-IgM. These data indicate that MAPK/Erk has a role in regulating CD5E1B up-regulation.

Expression of CD5-E1B, which is retained in the cytoplasm, reduces membrane expression of CD5-E1A (17, 19). This suggests that CD5-expressing B cells may, under certain physiological conditions (such as activation), preferentially up-regulate CD5-E1B relative to CD5-E1A. To test this proposition, levels of CD5-E1A and CD5-E1B mRNA were determined using a 5⬘ specific real time PCR following 24h anti-IgM stimulation of FACS-sorted

CD5-negative B cells from the blood of 10 SLE patients and 15 HCs (Fig. 1D). BCR engagement of the cells raised the level of CD5-E1B mRNA in B cells from the patients (54.4 ⫾ 68.1-fold) but not in those from the controls (0.5 ⫾ 0.2-fold). In contrast, after BCR engagement expression of CD5-E1A is increased both in SLE patients and in HCs (7.7 ⫾ 4.9-fold and 5.5 ⫾ 5.6-fold, NS). The 5⬘-RACE analysis of CD5 cDNA using B cells from HCs

FIGURE 2. Amplification of methylation-sensitive, endonuclease-digested genomic DNA reveals methylation status of the alternative promoters of CD5 in resting B cells. A, Schematic representation of the affected promoters. The CD5-E1B promoter arises from a LTR element subdivided into U3, R, and U5 regions. The two splice donors (SD) are indicated, and positions of DNA-specific primers shown. The 1177-bp CD5-E1B amplicon contains six HpaII/MspI motifs (5⬘-CCGG-3⬘) at positions ⫺8736, ⫺8726, ⫺8522, ⫺8268, ⫺8114, and ⫺7795 bp, and one HaeII motif (5⬘(A/G)GCGC(T/C)3⬘) at position ⫺8393bp according to numbering from the established ATG initiation site in exon 1A (22). The 783-bp CD5-E1A amplicon contains one HgaI motif (5⬘-GACGC(N)5-3⬘) at position ⫺151, one FauI motif (5⬘-CCCGC(N)4-3⬘) at position ⫺96, and two HpaII/MspI motifs at positions ⫺65 and ⫹87. B, Analysis of CD5-E1B promoter methylation by amplification of genomic DNA digested with methylation-sensitive HaeII, HpaII, or methylationinsensitive MspI enzymes. C, Analysis of the CD5-E1A promoter region using HgaI, FauI, HpaII, and MspI enzymes. Undigested genomic DNA is amplified by PCR, separated on agarose gel, and visualized with ethidium bromide.

The Journal of Immunology

5627

FIGURE 3. The U3-LTR HERV-CD5 region is de-methylated in B cells from SLE patients. A, Regulatory motifs for transcription factor binding and TATA box location within the U3 region. Circles and boxes identify the CpG and U3-R-U5 regions. The HpaII and HaeII sites are also indicated. B, The level of CpG methylation was determined by bisulfite sequencing using genomic DNA obtained from six SLE patients (white) and seven HCs (black). CpG sites are numbered as in A. C, Correlation between CD5 cell surface expression and methylation status for CpG sites 3 to 6. White circle, SLE patients; black circle, HC subjects. ⴱ, p ⬍ 0.05; p ⬍ 0.001.

confirmed that BCR engagement induce CD5-E1A but did not induce CD5-E1B transcripts (Fig. 1E). To test the hypothesis that epigenetic changes in the cd5 gene contribute to the generation of alternative transcripts in activated B cells from patients with SLE, BCR engagement was re-evaluated in the presence of PD98059 which decreases DNA methylation. In these conditions, an up-regulation of CD5-E1B upon anti-IgM/ PD98059 stimulation was observed in B cells from HCs using real-time PCR (61.5 ⫾ 11.9-fold) and 5⬘-RACE analysis (Fig. 1, D and E). The CD5-E1B promoter is hypomethylated in resting SLE B cells To determine whether epigenetic changes result in increased CD5E1B in SLE B cells, genomic DNA from six randomly selected patients and seven HCs were analyzed for the level of DNA methylation restriction enzyme treatment of the DNA followed by PCR (Fig. 2A). The protocol is based on the inability of methylationsensitive restriction enzymes to cut methylated DNA. The results of treatment with HaeII enzyme revealed that the CD5-E1B promoter was de-methylated in B cells from all 6 SLE patients but in none of the 7 HCs (Fig. 2B). Treatment with HpaII confirmed demethylation in two of six SLE patients but in none of the HCs. The positive control was the DNA methylation of Daudi cell line

cells (23). The negative control was the HpaII isoschizomer, MpsI, that cut the CpG sequences so that there were no PCR products. In addition to this data, the methylation status of the CD5-E1A promoter was sought using HgaII, FauI and HpaII (Fig. 2C). The results showed that the CpG motifs were de-methylated, or hemimethylated in B cells from both the SLE patients and HCs. The U3 region in CD5-E1B is hypomethylated in B cells from SLE patients The presence of HaeII site (site 4) in the U3 region of the 5⬘-LTR element of HERV-CD5, and three HpaII sites upstream U3-LTR (site 1) or in the R region of the 5⬘-LTR (sites 7 and 13) raises the possibility that the HERV U3-R-U5 regulatory elements may be de-methylated in B cells from patients with SLE (Fig. 3A). To address this issue, bisulfite-treated genomic DNA were amplified and sequenced. Bisulfite C3 T transition in SLE CD5-E1B amplicons were analyzed and quantified as shown in Fig. 3B and supplemental Fig. 1.5 Among 26 CpG sites studied in the 1439-bp amplicon of CD5EIB, five CpG sites are hemi- or de-methylated. These CpGs (sites 2 to 6) are located in the U3-LTR at positions ⫺8466 from the 5

The online version of this article contains supplemental material.

5628

IL-6 CONTROLS DNA METHYLATION AND CD5 IN LUPUS B CELLS

FIGURE 4. Effect of BCR engagement on cd5 gene methylation. A, Methylation of CD5-E1B promoter analyzed by restriction enzymes and bisulfite sequencing. Enzymes and symbols used are the same as in the legend to Fig. 3. B, Effect of restriction enzymes on CD5E1A. C, The level of CpG methylation measured by bisulfite sequencing with (white) or without (black) BCR engagement in six SLE patients. D, CpG methylation in B cells from six HCs stimulated with anti-IgM in the presence (gray) or absence (white) of PD98059. ⴱ, p ⬍ 0.05.

known ATG1 initiation site (24), ⫺8396, ⫺8393 (HaeII), ⫺8347, and ⫺8363, respectively. De-methylated CpG motifs contain, or are located near, the binding sites of E2/Rb family proteins (consensus TTT(C/G)(C/G)CGC, position ⫺8394 to ⫺8391), E-box family proteins (CA(N)3TG, ⫺8391 to ⫺8388), and STAT family proteins (inverted consensus TT(N)4AA, ⫺8353 to ⫺8344). When comparing SLE patients to HCs, the cytosine residues near the E2/Rb binding sites 3 and 4, plus the CpG at site 5, were significantly less affected by the bisulfite treatment in the SLE patients (Fig. 3B, p ⬍ 0.05). Furthermore, the level of methylation of CpG sites 3 to 5, are inversely correlated with CD5 cell surface expression in CD5⫹ B cells (Fig. 3C, p ⬍ 0.01). However, demethylation of the CD5-E1B promoter identified in B cells from the SLE patients could be due to gene polymorphisms. Such a proposition was discounted because only one association between a methylated CpG motif at position ⫺8001 and a SNP site was observed (supplemental Fig. 2). In contrast to CD5-E1B promoter, the 791bp amplicon for CD5E1A which contains 12 CpG was equally de-methylated in B cells from the SLE patients and HCs (Fig. 2C and supplemental Fig. 1). De-methylated CpGs were located at positions: ⫺148, ⫺123, ⫺102, ⫺73 and ⫺64. These regions contain, or are located nearby to consensus binding sites for AP-1 (⫺151), Sp-1 (⫺120, ⫺96), E-box family proteins (⫺50) and the INR transcriptional start site (⫺61).

cells from the HCs were stimulated with anti-IgM in the presence of PD98059, CpG sites 3, 4, and 5 were prone to demethylation similar to what was seen in the SLE B cells. All these results are consistent with previous findings that the U3-LTR regions in HERV elements are prone to demethylation (25).

Activation modifies methylation of the CD5-E1B promoter

Table III. Level of DNMTs and MBDs mRNA in resting B cells from 10 SLE patients and 15 HCsa

To assess the influence of BCR engagement on methylation of the CD5 locus, FACS-sorted CD5-negative B cells from six SLE patients and six HCs were stimulated with anti-IgM for 24 h. Methylation patterns were then compared in these activated cells with the pattern of methylation in the cells before activation (Fig. 4). In the SLE patients, engagement of the BCR did not modify CD5 promoter methylation status, whereas in the HCs, it resulted in methylation of the CD5-E1B promoter at positions 3 and 4 (Fig. 4D), but not that of CD5-E1A (not shown). Interestingly, when B

Analysis of DNA methylation enzymes To gain insights into the cause(s) of de-methylation of the CD5E1B promoter in B cells from SLE patients, levels of mRNA for DNMTs and MBDs were determined in B cells from ten SLE patients and 15 HCs. Levels of DNMT1, DNMT3a, DNMT3b, MBD2, and MBD4 were not different in B cells from the patients and the controls (Table III). With regard to SLE activity, we failed to find correlations between patients with SLEDAI ⱖ 5 (n ⫽ 3) and changes in DNMTs or MBDs activity. On the basis of our observation that B cell activation through the BCR influences U3LTR methylation, we measured the level of mRNA for DNMT1 and MBDs by real-time PCR in B cells stimulated for 24 h with anti-IgM from ten SLE patients and 15 HCs (Fig. 5A). The expression of DNMT1 was increased by 2.3 ⫾ 0.2-fold in B cells from the patients and by 16.6 ⫾ 13.4-fold in the controls ( p ⬍ 0.005). Importantly, the expression of MBDs was not affected by

DNMT1 DNMT3a DNMT3b MBD2 MBD4

Patients

Controls

14.8 ⫾ 8.4 1.03 ⫾ 0.52 undetectable 44.2 ⫾ 22.9 34.2 ⫾ 21.5

15.4 ⫾ 10.4 0.61 ⫾ 0.76 undetectable 37.1 ⫾ 19.4 23.5 ⫾ 12.5

a cDNA were subjected to quantitative RT-PCR (primers indicated in Table I). Their relative expression was adjusted to 18S levels (⫻10⫺6).

The Journal of Immunology

5629

FIGURE 5. Involvement of DNMTs on CD5-E1B expression. A, Quantitative PCR measurement of DNMT1, MBD2, and MBD4 mRNA levels in ten SLE patients and 15 HCs following BCR engagement. Role of activation of MAPK/Erk pathway in DNMT1 induction was examined in HCs using the PD98059 inhibitor. B, Cytoplasmic staining of DNMT1 in methanol-permeabilized non-activated, or B cells activated with anti-IgM.

stimulation of the BCR or by the addition of the DNMTs inhibitor PD98059. FACS analyses verified that BCR engagement modulated DNMT1 expression. Results from four of ten independent experiments are depicted in Fig. 5B. Interestingly, DNMT1 staining showed two peaks, DNMT1dim and DNMT1bright. According to this dichotomy, resting B cells could be divided into 76.1 ⫾ 7.9% DNMT1dim and 23.8 ⫾ 7.9% DNMT1bright cells in the patients and into 74.7 ⫾ 12.3% DNMT1dim and 25.3 ⫾ 12.3% DNMT1bright cells in the controls ( p ⫽ NS). A 24-h culture of B cells with anti-IgM increased the proportion of the DNMT1bright population to 86.1 ⫾ 12.2% in the HCs compared with 46.1 ⫾ 7.8% in the SLE patients ( p ⬍ 0.05). Thus, these results indicate that induction of DNMT1 following BCR engagement is reduced in patients with SLE.

The role of IL-6 in de-methylation and CD5-E1B expression Based on our previous observations that IL-6 over-expression controls the cell cycle in BCR-activated in B cells from SLE patients, we predicted that IL-6 by arresting the cell cycle at late G1 phase may control the expression of DNMT1 and its capacity to methylate DNA and subsequently CD5 cell surface expression (26 –28). To test whether IL-6 acts on its own, or requires engagement of the BCR, FACS-sorted B cells from six HCs were stimulated for 48 h with rhIL-6 in the presence or absence of anti-IgM. Expression of CD5-E1B increased by 4.1 ⫾ 3.13-fold in B cells cultured with rhIL-6, and by 54.8 ⫾ 11.3-fold in B cells cultured with rhIL-6 and anti-IgM (Fig. 6A, p ⬍ 0.05). The induction of CD5E1B upon IgM/rhIL-6 stimulation was also demonstrated by ChIP

FIGURE 6. The effect of IL-6 on CD5-E1B expression and promoter methylation. A, FACS-sorted B cells from HCs were incubated with IL-6 in the presence, or absence of anti-IgM, or anti-IL-6R Abs. Quantitative PCR measurement of CD5-E1B (white boxes) and DNMT1 (black boxes) in six HCs. B, CD5-E1B promoter ChIP analysis using a nonspecific mouse IgG as negative control (C⫺), or a mouse anti-RNA polymerase as positive control (C⫹). C, Cytoplasmic staining of DNMT1 and p27kip1 as a marker of cell cycle arrest in the G1 phase in methanol-permeabilized B cells from the HCs incubated with IL-6 in the presence or absence of anti-IgM.

5630

IL-6 CONTROLS DNA METHYLATION AND CD5 IN LUPUS B CELLS The induction of DNMT1 mRNA by anti-IgM was negated in the presence of rhIL-6 (Fig. 6A). The experiments were repeated in the presence of anti-IL-6R Ab to inhibit the activity of IL-6. DNMT1 induction was restored (10.0 ⫾ 3.6 vs 0.63 ⫾ 0.25-fold without anti-IL-6R Ab, p ⬍ 0.05) and CD5-E1B was reduced (54.8 ⫾ 11.3 vs 35.6 ⫾ 5.2-fold without anti-IL-6R Ab, p ⬍ 0.05). Moreover, FACS analyses showed that the number of DNMT1bright cells was reduced after anti-IgM stimulation in the presence of rhIL-6 (86.1 ⫾ 12.2% vs 33.5 ⫾ 3.7%, p ⬍ 0.05). A representative example of two experiments is shown in Fig. 6C. We also studied whether these differences could be attributed to a cell cycle blockade. As suspected, the cyclindependent kinases inhibitor p27kip1 was over-expressed in antiIgM/rhIL-6-stimulated B cells compared with rhIL-6 or antiIgM stimulation alone (Fig. 6C). Overall, IL-6 appears to control CpG methylation in SLE B cells resulting, probably, from its effect on arresting cells at the late G1 phase of the cell cycle. Effects of IgM/IL-6 stimulation on promoters of other genes

FIGURE 7. Anti-IgM-induced methylation of promoters for CD19, CD70, Pax5, Syk, and HRES-1. A, B cells from HCs were incubated with anti-IgM in the presence or absence of anti-IL-6R Ab. The methylation status of all five promoters was determined by PCR using predigested genomic DNA with methylation-sensitive HpaII, or methylation-insensitive MspI. B, The methylation of the HRES-1 promoter was quantified by calculating the ratio of HpaII-digested to undigested bands in six HCs. ⴱ, p ⬍ 0.05.

To determine the extent of IL-6 effect on methylation, we determined the methylation status of other promoters (CD19, CD70, Pax 5, Syk, and HRES-1) that are known to be regulated by methylation (11, 21, 22). In the six HCs studied (Fig. 7), the promoters were hypomethylated in resting cells. Stimulation of B cells with anti-IgM only increased the methylation of the HRES-1. This effect by anti-IgM was reversed in the presence of IL-6. Effect of IL-6 on SLE B cells

analysis. This experiment showed that RNA polymerase II was recruited to the CD5-E1B promoter upon anti-IgM/rhIL-6 stimulation (Fig. 6B). Thus, rhIL-6 induces CD5-E1B expression and this effect is more pronounced when B cells are activated through the BCR.

FIGURE 8. IL-6-dependent modulation of methylation in SLE B cells. CD5 cell surface expression in relation to the detection level of IL-6 (A) or IFN-␥ (B) in the serum of 18 SLE patients. The detection limit used corresponds to the sensitivity limit of the ELISA. C, FACS-sorted B cells from six SLE patients were cultured with anti-IgM in the presence or absence of 40 ng/ml anti-IL-6R Ab. All SLE patients, except one, were inactive (SLEDAI ⬍ 5). Dose effect response at 10 and 20 ng/ml anti-IL-6R were performed on three SLE patients. Quantitative PCR measurement of CD5-E1B (white boxes) and DNMT1 (black boxes). D, Effect of anti-IgM/anti-IL-6R (gray) on CD5-E1B promoter methylation status.

To validate the role of IL-6 in the altered methylation status of CD5-E1B in B cells in SLE patients, membrane CD5 expression levels were compared between patients whose sera were IL-6 positive and negative. The presence of IL-6 in serum was associated with decreased membrane CD5 expression (Fig. 8A, p ⫽ 0.01). In

The Journal of Immunology contrast, changes in membrane CD5 expression were not associated with serum IFN-␥ (Fig. 8B). CD5 cell surface expression was reduced in patients who had SLEDAI ⱖ 5 (39,789 ⫾ 11,450 vs 51,294 ⫾ 15,215), but this was not significant (data not shown). Finally, the effect of a blocking anti-IL-6R Ab was studied because we previously observed that anti-IgM stimulation of B cells from SLE patients produced higher levels of IL-6 than matched HCs (514.1 ⫾ 159.7 pg/ml vs 99.0 ⫾ 116.0 pg/ml, data not shown) (26). Blocking the autocrine loop of IL-6 increased DNMT1 expression (58.9 ⫾ 13.5-fold with 40 ng/ml anti-IL-6R) and contributed to the methylation of the U3-LTR sites 3 to 5 (Fig. 8, C and D). In addition, expression of CD5-E1B was reduced (17.2 ⫾ 3.5fold with 40 ng/ml anti-IL-6R vs 54.4 ⫾ 68.1-fold without, p ⬍ 0.05). Further, this effect was dose-dependent.

Discussion This study provides evidence for the notion that reduction in the level of membrane CD5 on circulating CD5⫹ B cells SLE is due to increased expression of the cytoplasmic isoform of CD5, CD5E1B. As a consequence of reduced membrane CD5, increased CD5-E1B limits the negative regulatory effects of CD5 on BCRmediated signaling. The available evidence indicates that membrane CD5 increases the threshold of BCR-mediated responses (14). The implication of such data is that regulation of CD5-E1B, which in turn regulates expression of CD5, is involved in maintaining the anergy status of autoreactive B cells. Results generated from our previous studies indicated that the modulatory effects of CD5 protein on BCR signaling are attributed to the CD5 isoform which contains E1A, the previously documented exon 1. Introduction of cDNA for the recently identified CD5-E1B isoform into CD5⫹ T and B cells reduced membrane expression of CD5 protein (17, 19). Coexpression experiments have also shown that in the presence of CD5-E1B, CD5-E1A was not translocated to the membrane but associated with CD5-E1B in cytoplasmic aggregates (19). The exact dynamics of the interaction between CD5-E1B and CD5-E1A remain unknown. However, we have established that the stretch of amino acids from positions 286 – 400 in CD5-E1B is crucial for reducing membrane CD5-E1A translocation (19). Functionally, these data suggest that up-regulated CD5-E1B expression in B cells could reduce the BCR activation threshold by Ags and thus, promote autoimmunity. Because CD5-E1B arises from the integration of an HERV element known to be silenced by DNA methylation (19), we predicted that epigenetic modifications could be involved in regulating CD5 isoform expression in B cells. Normally, methylation of DNA is evident in CpG-poor regions and in regions of repetitive sequences including LTR in HERV sequences. In contrast, CpG islands present within gene promoters resist to methylation. In normal B cells, this paradigm holds true for the cd5 locus, where the U3-LTR element is methylated while the CD5-E1A promoter is not. However, in SLE B cells, the rate of de-methylation is higher in the U3 region of the HERV-CD5 element. Consistent with this observation is the finding that another HERV element, HRES-1, was also integrated at the stage of old world primate divergence (22). Indeed, expression of HRES-1 retroviral proteins, HRES-1/ p28 and HRES-1/Rab4 (29), and our observations on HRES-1 promoter methylation status (Fig. 7) may hint to a global defect in controlling repetitive elements in SLE. Our preliminary results are indeed indicative that HRES-1 methylation pattern was not increased in SLE patients upon BCR engagement. In this respect, global DNA methylation has been reported to be altered in almost all forms of SLE including drug-induced lupus (30). As a consequence, it was postulated that epigenetic transformation modifies B cell physiology resulting in polyclonal activation, IgG1 class-

5631 switching (31), V(D)J rearrangement by RAG1/RAG2 enzymes (32) and a shift in cytokine profiles (33). Altered methylation also influences T cells in SLE causing up-regulation of T cell costimulating molecules (CD70, CD40 ligand), and a shift in cytokine profile from Th1 to Th2 (for review, see Ref. 34). We propose that these alterations, together with our findings on CD5 down-regulation, contribute to promoting autoreactivity. De-methylation of DNA can be result from the inhibition or lack of DNMTs or be due to specific enzymatic reactions including MBDs or a combination of both. Thus, DNA methylation patterns can be altered owing to a shift in the balance between MBD demethylation and DNMT methylation enzymes. Quantitative PCR assays were performed to evaluate the relative levels of MBD and DNMT transcripts in resting and in anti-IgM stimulated B cells. At basal level, B cells from SLE patients and HCs had similar levels of DNMT1, DNMT3a, MBD2, and MBD4. The amount of DNMT3b was very low or undetectable in both SLE patients and HCs as previously described for T cells (35). A recent study reported reductions in DNMT1 in a subgroup of SLE patients (36). However, this study was conducted using a relatively small cohort and was not confirmed either in CD4⫹ T cells (35) or in B cells by the present study. In addition, the recently reported increase in MBD2 and MBD4 in patients with SLE (37) was also not confirmed in our study. These differences may be because we have examined B cells instead of CD4⫹ T cells, or that the patient population included in the other study was different. Interestingly, analysis of stimulated B cells revealed that expression of DNMT1 was reduced following engagement of the BCR in SLE in agreement with findings in stimulated CD4⫹ T cells (13). Our results, suggesting that DNA hypomethylation in activated SLE B cells, could be attributed to cell cycle arrest in G1 is surprising. The surprising aspect of the observation is based on the fact that proliferation of lymphocytes did not differ between the patients and HCs implying that decreased DNMT1 activity in SLE was not due to a cell cycle arrest (13). However, inhibition of the MAPK/Erk2 pathway was clearly associated with DNMT1 reduction in those patients (13, 38). Further, inhibition of this pathway is likely to elicit cell cycle arrest through p27kip1 induction (39, 40). However, inhibition of MEK/Erk2 pathway using PD98059 in pro-B induces cell growth arrest with p27kip1 accumulation (39). Therefore, in such a setting, it was not surprising that PD98059 decreased methylation in HCs as in stimulated SLE B cells. Hence, these data indicate that decreased activation of the Erk pathway could be important for the development of autoimmunity (7). This was previously confirmed in animal models showing that treating normal mice with inhibitors of DNA methylation, or with Erk inhibitors causes a lupus-like disease (41). Further, Tg mice expressing dominant-negative MEK leads to overexpression of methylation-sensitive genes and the production of anti-dsDNA autoantibodies by B cells (38). The proposed involvement of IL-6 in promoting autoreactivity is supported by in vitro and in vivo studies. For example, high levels of serum IL-6 is correlated with lupus disease activity and targeting IL-6 with blocking Abs is an effective treatment for SLE (4). However, the precise mechanism by which IL-6-mediated B cell proliferation, differentiation, and autoantibody production in SLE remains to be elucidated. Repression of IL-6 is associated with hypermethylation of its promoter in HCs whereas its overexpression in SLE is associated with promoter hypomethylation (42). Interestingly, by treating B cells from HCs with IL-6 in the presence of anti-IgM, we have shown that CD5-E1B expression can be significantly increased and that this correlates with a cell cycle arrest in late G1 phase. In addition, this effect is associated with an Ig gene rearrangement following RAG re-expression (26). The loss

5632

IL-6 CONTROLS DNA METHYLATION AND CD5 IN LUPUS B CELLS

of, or reduction in, the inhibitory effects of CD5 is associated with RAG re-expression in B cells in patients with SLE and these may synergize to induce autoantibody production. In conclusion, the finding that IL-6 activates CD5-E1B transcription is relevant for understanding mechanism of action and therapeutic benefits of anti-IL-6. Thus, treatment of patients with SLE with anti-IL-6R mAb could inhibit autoreactive B cell expansion by restoring DNA methylation and cell cycle progression. In addition, the data provide further evidence for the pivotal role of CD5 in maintaining anergy in autoreactive B cells.

Acknowledgments We are grateful to the Conseil Re´gional de Bretagne, the Conseil Ge´ne´ral du Finiste`re, the College Doctoral International, and the Universite´ Europe´enne de Bretagne for support. Thanks are also due to Cindy Se´ne´ and Simone Forest for excellent secretarial assistance.

Disclosures The authors have no financial conflict of interest.

References 1. Rahman, A., and D. A. Isenberg. 2008. Systemic lupus erythematosus. N. Engl. J. Med. 358: 929 –939. 2. Renaudineau, Y., P. O. Pers, B. Bendaoud, C. Jamin, and P. Youinou. 2004. Dysfunctional B cells in SLE. Autoimmun. Rev. 3: 516 –523. 3. Leandro, M. J., J. C. Edwards, G. Cambridge, M. R. Ehrenstein, and D. A. Isenberg. 2002. An open study of B lymphocyte depletion in systemic lupus erythematosus. Arthritis Rheum. 46: 2673–2677. 4. Illei, G., C. Yarboro, Y. Shirota, E. Tackey, L. Lapteva, and T. Fleisher. 2006. Tocilizumab (humanized anti IL-6 receptor monoclonal antibody) in patients with systemic lupus erythematosus (SLE): safety, tolerability and preliminary efficacy. Arthritis Rheum. 54 (Suppl): 4043 (Abstract). 5. Finck, B. K., P. S. Linsley, and D. Wofsy. 1994. Treatment of murine lupus with CTLA4Ig. Science 265: 1225–1227. 6. Hom, G., R. R. Graham, B. Modrek, K. E. Taylor, W. Ortmann, S. Garnier, A. T. Lee, S. A. Chung, R. C. Ferreira, P. V. Pant, et al. 2008. Association of systemic lupus erythematosus with C8orf13-BLK and ITGAM-ITGAX. N. Engl. J. Med. 358: 900 –909. 7. Ballestar, E., M. Esteller, and B. C. Richardson. 2006. The epigenetic face of systemic lupus erythematosus. J. Immunol. 176: 7143–7147. 8. Zhu, B., Y. Zheng, H. Angliker, S. Schwarz, S. Thiry, M. Siegmann, and J. P. Jost. 2000. 5-Methylcytosine DNA glycosylase activity is also present in the human MBD4 (G/T mismatch glycosylase) and in a related avian sequence. Nucleic Acids Res. 28: 4157– 4165. 9. Hendrich, B., and A. Bird. 1998. Identification and characterization of a family of mammalian methyl-CpG binding proteins. Mol. Cell Biol. 18: 6538 – 6547. 10. Schwab, J., and H. Illges. 2001. Regulation of CD21 expression by DNA methylation and histone acetylation. Int. Immunol. 13: 705–710. 11. Lu, Q., A. Wu, L. Tesmer, D. Ray, N. Yousif, and B. Richardson. 2007. Demethylation of CD40LG on the inactive X in T cells from women with lupus. J. Immunol. 179: 6352– 6358. 12. Piotrowski, P. C., S. Duriagin, and P. P. Jagodzinski. 2005. Expression of HERV clone 4-1 may correlate with blood plasma concentration of anti-U1 RNP and anti-Sm nuclear antibodies. Clin. Rheumatol. 24: 620 – 624. 13. Deng, C., M. J. Kaplan, J. Yang, D. Ray, Z. Zhang, W. J. McCune, S. M. Hanash, and B. C. Richardson. 2001. Decreased ras-mitogen-activated protein kinase signaling may cause DNA hypomethylation in T lymphocytes from lupus patients. Arthritis Rheum. 44: 397– 407. 14. Hippen, K. L., L. E. Tze, and T. W. Behrens. 2000. CD5 maintains tolerance in anergic B cells. J. Exp. Med. 191: 883– 890. 15. Jamin, C., G. Magadur, A. Lamour, L. MacKenzie, P. M. Lydyard, P. Katsikis, and P. Youinou. 1992. Cell-free CD5 in patients with rheumatic diseases. Immunol. Lett. 31: 79 – 84. 16. Lu, X., R. C. Axtell, J. F. Collawn, A. Gilson, L. B. Justement, and C. Raman. 2002. AP2 adaptor complex-dependent internalization of CD5: differential regulation in T and B cells. J. Immunol. 168: 5612–5620. 17. Renaudineau, Y., S. Hillion, A. Saraux, R. A. Mageed, and P. Youinou. 2005. An alternative exon 1 of the CD5 gene regulates CD5 expression in human B lymphocyte. Blood 106: 2781–2789.

18. Renaudineau, Y., S. Vallet, C. Le Dantec, S. Hillion, A. Saraux, and P. Youinou. 2005. Characterization of the human CD5 endogenous retrovirus-E in B lymphocytes. Gene Immun. 6: 663– 671. 19. Garaud, S., C. Le Dantec, C. Berthou, P. M. Lydyard, P. Youinou, and Y. Renaudineau. 2008. Selection of the alternative exon 1 from the cd5 gene down-regulates membrane level of the protein in B lymphocytes. J. Immunol. 181: 2010 –2018. 20. Tan, E. M., A. S. Cohen, J. F. Fries, A. T. Masi, D. J. McShane, N. F. Rothfield, J. G. Schaller, N. Talal, and R. J. Winchester. 1982. The 1982 revised criteria for the classification of SLE. Arthritis Rheum. 25: 1271–1277. 21. Ushmorov, A., F. Leitha¨user, O. Sakk, A. Weinhau¨sel, S. W. Popov, P. Mo¨ller, and T Wirth. 2006. Epigenetic processes play a major role in B-cell-specific gene silencing in classical Hodgkin lymphoma. Blood 107: 2493–2500. 22. Perl, A., J. D. Rosenblatt, I. S. Chen, J. P. DiVincenzo, R. Bever, B. J. Poiesz, and G. N. Abraham. 1989. Detection and cloning of new HTLV-related endogenous sequences in man. Nucleic Acids Res. 17: 6841– 6854. 23. Ruchusatsawat, K., J. Wongpiyabovorn, S. Shuangshoti, N. Hirankarn, and A. Mutirangura. 2006. SHP-1 promoter 2 methylation in normal epithelial tissues and demethylation in psoriasis. J. Mol. Med. 84: 175–182. 24. Jones, N. H., M. L. Clabby, D. P. Dialynas, H. J. Huang, L. A. Herzenberg, and J. L. Strominger. 1986. Isolation of complementary DNA clones encoding the human lymphocyte glycoprotein T1/Leu-1. Nature 323: 346 –349. 25. Reiss, D., Y. Zhang, and D. L. Mager. 2007. Widely variable endogenous retroviral methylation levels in human placenta. Nucleic Acids Res. 35: 4743– 4754. 26. Hillion, S., S. Garaud, V. Devauchelle, A. Bordron, C. Berthou, P. Youinou, and C. Jamin. 2007. IL-6 is responsible for aberrant BCR-mediated regulation of RAG expression in SLE. Immunology 122: 371–380. 27. Robertson, K. D., K. Keyomarsi, F. A. Gonzales, M. Velicescu, and P. A. Jones. 2000. Differential mRNA expression of the human DNA methyltransferases (DNMTs) 1, 3a and 3b during the G(0)/G(1) to S phase transition in normal and tumor cells. Nucleic Acids Res. 28: 2108 –2113. 28. Brown, S. E., M. F. Fraga, I. C. Weaver, M. Berdasco, and M. Szyf. 2007. Variations in DNA methylation patterns during the cell cycle of HeLa cells. Epigenetics 2: 54 – 65. 29. Pullmann, R., Jr., E. Bonilla, P. E. Phillips, F. A. Middleton, and A. Perl. 2008. Haplotypes of the HRES-1 endogenous retrovirus are associated with development and disease manifestations of systemic lupus erythematosus. Arthritis Rheum. 58: 532–540. 30. Zhou, Y., and Q. Lu. 2008. DNA methylation in T cells from idiopathic lupus and drug-induced lupus patients. Autoimmun. Rev. 7: 376 –383. 31. Vigorito, E., K. L. Perks, C. Abreu-Goodger, S. Bunting, Z. Xiang, S. Kohlhaas, P. P. Das, E. A. Miska, A. Rodriguez, A. Bradley, et al. 2007. microRNA-155 regulates the generation of Ig class-switched plasma cells. Immunity 27: 847– 859. 32. Wang, H., J. Feng, C. F. Qi, Z. Li, H. C. Morse 3rd, and S. H. Clarke. 2007. Transitional B cells lose their ability to receptor edit but retain their potential for positive and negative selection. J. Immunol. 179: 7544 –7552. 33. Pang, Y., Y. Norihisa, D. Benjamin, R. R. Kantor, and H. A. Young. 1992. IFN-␥ gene expression in human B-cell lines: induction by IL-2, PKC, and possible effect of hypomethylation on gene regulation. Blood 80: 724 –732. 34. Huber, L. C., J. Stanczyk, A. Ju¨ngel, and S. Gay. 2007. Epigenetics in inflammatory rheumatic diseases. Arthritis Rheum. 56: 3523–3531. 35. Balada, E., J. Ordi-Ros, S. Serrano-Acedo, L. Martinez-Lostao, M. Rosa-Leyva, and M. Vilardell-Tarre´s. 2008. Transcript levels of DNA methyltransferases DNMT1, DNMT3A and DNMT3B in CD4⫹ T cells from patients with systemic lupus erythematosus. Immunology 124: 339 –347. 36. Ogasawara, H., M. Okada, H. Kaneko, T. Hishikawa, I. Sekigawa, and H. Hashimoto. 2003. Possible role of DNA hypomethylation in the induction of SLE: relationship to the transcription of HERV. Clin. Exp. Rheumatol. 21: 733–738. 37. Balada, E., J. Ordi-Ros, S. Serrano-Acedo, L. Martinez-Lostao, and M. Vilardell-Tarre´s. 2007. Transcript overexpression of the MBD2 and MBD4 genes in CD4⫹ T cells from SLE patients. J. Leukocyte Biol. 81: 1609 –1616. 38. Sawalha, A. H., M. Jeffries, R. Webb, Q. Lu, G. Gorelik, D. Ray, J. Osban, N. Knowlton, K. Johnson, and B. Richardson. 2008. Defective T-cell ERK signaling induces IFN-regulated gene expression and overexpression of methylation-sensitive genes similar to lupus patients. Genes Immun. 9: 368 –378. 39. Qiang, Y. W., M. Kitagawa, M. Higashi, G. Ishii, C. Morimoto, and K. Harigaya. 2000. Activation of MAPK through alpha5/␤1 integrin is required for cell cycle progression of B progenitor cell line, Reh, on human marrow stromal cells. Exp. Hematol. 28: 1147–1157. 40. Gysin, S., S. H. Lee, N. M. Dean, and M. McMahon. 2005. Pharmacologic inhibition of RAF–⬎MEK–⬎ERK signaling elicits pancreatic cancer cell cycle arrest through induced expression of p27Kip1. Cancer Res. 65: 4870 – 4880. 41. Richardson, B., D. Ray, and R. Yung. 2004. Murine models of lupus induced by hypomethylated T cells. Methods Mol. Med. 102: 285–294. 42. Mi, X. B., and F. Q. Zeng. 2008. Hypomethylation of IL-4 and -6 promoters in T cells from SLE patients. Acta Pharmacol. Sin. 29: 105–112.