Effects of anti-CD154 treatment on B cells in murine systemic lupus erythematosus.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 48, No. 2, February 2003, pp 495–506 DOI 10.1002/art.10929 © 2003, American College of Rheumatology Effects of Anti-CD154 Treatment on B Cells in Murine Systemic Lupus Erythematosus Xiaobo Wang,1 Weiqing Huang,1 Lena E. Schiffer,1 Masahiko Mihara,2 Alla Akkerman,1 Kenji Hiromatsu,1 and Anne Davidson1 attenuated IgM response to the hapten oxazolone and produced no IgG antioxazolone antibodies. Conclusion. Anti-CD154 is a B cell–depleting therapy that affects multiple B cell subsets. Activation of both B and T cells is prevented during therapy. After treatment cessation, autoreactive B cells progress through a series of activation steps before they become fully competent antibody-producing cells. The general immunosuppression induced during treatment will need to be taken into account when using B cell–depleting regimens in humans. Objective. To determine the immunologic effects of anti-CD154 (CD40L) therapy in the (NZB ⴛ NZW)F1 mouse model of systemic lupus erythematosus. Methods. Twenty-week-old and 26-week-old (NZB ⴛ NZW)F1 mice were treated with continuous anti-CD154 therapy. Mice were followed up clinically, and their spleens were studied at intervals for B and T cell numbers and subsets and frequency of anti–doublestranded DNA (anti-dsDNA)–producing B cells. T cell– dependent immunity was assessed by studying the humoral response to the hapten oxazolone. Results. IgG anti-dsDNA antibodies decreased during therapy and disease onset was delayed, but immune tolerance did not occur. During treatment, there was marked depletion of CD19ⴙ cells in the spleen; however, autoreactive IgM-producing B cells could still be detected by enzyme-linked immunospot assay. In contrast, few IgG- and IgG anti-dsDNA– secreting B cells were detected. Eight weeks after treatment cessation, the frequency of B cells producing IgG anti-dsDNA antibodies was still decreased in 50% of the mice, and activation and transition of T cells from the naive to the memory compartment were blocked. AntiCD154 treatment blocked both class switching and somatic mutation and induced a variable period of relative unresponsiveness of IgG anti-dsDNA–producing B cells, as shown by decreased expression of the CD69 marker and failure to generate spontaneous IgG antidsDNA–producing hybridomas. Treated mice mounted an The interaction of CD40 on B cells with its ligand CD154 (CD40L) on activated T cells provides a B cell costimulatory signal that induces B cell proliferation and formation of germinal centers (1–3). Further cell–cell interactions within the germinal center, also involving CD40/CD154, lead to immunoglobulin isotype switching, somatic mutation, clonal expansion of high-affinity B cells, and terminal differentiation to memory cells or antibody-forming cells (1,4,5). In addition, activation of antigen-presenting cells by CD40 ligation is required for optimal priming and clonal expansion of CD4⫹ T cells (2). CD40–CD154 interactions are also important in the effector arm of the inflammatory response (6,7). In particular, CD40 engagement plays a role in activation of endothelial cells, including up-regulation of adhesion molecules and chemokines that help direct cell traffic to sites of inflammation (7). The multiple functions of CD40/CD154 in the immune and inflammatory response have made it an attractive target for therapeutic intervention in autoimmune disease. Disease expression in 3 independent murine models of systemic lupus erythematosus (SLE) has been shown to depend on T cell help mediated by CD40– CD154 interactions. In the MRL/lpr mouse, the absence of CD40 in T cell receptor–intact mice resulted in a decrease in IgG levels but not IgM levels, markedly Supported by the Sklarow Trust and by NIH grant AI-47291. 1 Xiaobo Wang, MD, Weiqing Huang, MD, Lena E. Schiffer, MD, Alla Akkerman, BSc, Kenji Hiromatsu, PhD, Anne Davidson, MBBS: Albert Einstein College of Medicine, Bronx, New York; 2 Masahiko Mihara, PhD: Chugai Pharmaceuticals, Shizuoka, Japan, and Albert Einstein College of Medicine, Bronx, New York. Address correspondence and reprint requests to Anne Davidson, MBBS, Albert Einstein College of Medicine, 1300 Morris Park Avenue, U505, Bronx, NY 10461. E-mail: email@example.com. Submitted for publication August 16, 2002; accepted in revised form November 13, 2002. 495 496 WANG ET AL lower titers of IgG autoantibodies, and decreased expression of clinical renal disease (8). In the (SWR ⫻ NZB)F1 model, 3 doses of anti-CD154 antibody given early in life delayed disease onset by an average of 3 months (9). In these mice, a high-dose continuous treatment regimen was able to stabilize or reverse early nephritis, although it had much less effect on mice that had already developed proteinuria of ⬎300 mg/dl (10,11). In (NZB ⫻ NZW)F1 mice, long-term administration of anti-CD154 antibody prevented development of high-titer autoantibodies and disease onset (12). Finally, a 2-week combination protocol of anti-CD154 and CTLA4-Ig was highly synergistic in the (NZB ⫻ NZW)F1 model, resulting in prolonged survival of the mice (13,14). The goal of this study was to determine the effects, and to further evaluate the mechanism of action, of continuous anti-CD154 treatment in the (NZB ⫻ NZW)F1 model with respect to its effect on autoreactive B cells. We confirmed that disease onset is prevented in mice treated 20 weeks before the development of anti– double-stranded DNA (anti-dsDNA) antibodies. In 26week-old prenephritic mice, prevention of disease onset necessitated higher doses of anti-CD154. There was a substantial decrease in the frequency of B cells in the spleens of mice during treatment, showing that anti-CD154 is a B cell–depleting agent. Secretion of autoantibodies by IgM anti-dsDNA antibody–producing B cells was relatively unaffected by treatment, but class switching and somatic mutation were suppressed during therapy, as was the IgG humoral response to a foreign antigen. Activation of T cells was also suppressed, and there was a delay in the abnormal accumulation of T cells bearing the activated/memory phenotype. After cessation of therapy, there was a variable period during which autoreactive B cells were not fully activated, and there was a consequent delay in disease onset for an average of 4 months. This maintenance of an immunosuppressive effect suggests that CD154 blockade might be useful for intermittent therapy. MATERIALS AND METHODS Mice. Female (NZB ⫻ NZW)F1 mice were purchased from The Jackson Laboratory (Bar Harbor, ME) and maintained in a conventional animal housing facility. Mice were treated at the age of 20 weeks or 26 weeks with 250 g of hamster anti-CD154 (MR1; a kind gift of Dr. Susan Kalled, Biogen, Cambridge, MA) given intraperitoneally biweekly for 6 months (groups A20 and A26, respectively; n ⫽ 10 mice per group) until age 46 weeks. Because this protocol was ineffec- tive in 26-week-old mice, a group of 26-week-old mice (group B; n ⫽ 10 mice) was treated for 10 weeks with a higher dose protocol of 500 g/week for 3 weeks followed by biweekly treatment for a further 4 doses (total of 7 doses given from ages 26–36 weeks). This protocol was designed to replicate the dosing schedule of a simultaneous human clinical trial of an anti-CD154 antibody based on the half-lives of the two antibodies (15). Control mice received either hamster IgG (group C; n ⫽ 14 mice) or no treatment (group D; n ⫽ 17 mice). Prior to treatment, mice were randomized into treatment groups depending on the titer of anti-dsDNA antibodies. Mice were bled every 2–4 weeks. Urine was tested for proteinuria by dipstick (Multistick; Fisher, Pittsburgh, PA) every 2 weeks. Survival splenectomies were performed at intervals following treatment. Mice were followed up until death. Anti-dsDNA antibodies. Enzyme-linked immunosorbent assay (ELISA) plates (Falcon, Springfield, NJ) were coated with 100 l of 100 g/ml salmon sperm DNA made double stranded by passage through a 45 filter (USA Scientific, Ocala, FL). After drying, the plates were blocked with 5% fetal calf serum (FCS)/3% bovine serum albumin (BSA) in phosphate buffered saline (PBS) and then incubated sequentially for 1 hour at 37°C with serial dilutions of serum followed by peroxidase-conjugated F(ab⬘)2 goat anti-mouse immunoglobulin (1:8,000) or anti-IgG2a (1:5,000) in PBS/1% BSA (Accurate, Westbury, NY) and then by ABTS substrate (Kirkegaard & Perry, Gaithersburg, MD). Anti-hamster antibodies. Anti-CD154 antibody (BD PharMingen, San Diego, CA) was plated at 2 g/well onto ELISA plates (Falcon) overnight at 4°C. Plates were blocked with 5% FCS/3% BSA in PBS for 90 minutes at 37°C and then incubated with a 1:500 dilution of serum taken 6 weeks after the first injection of anti-CD154 antibody or hamster Ig. This was followed by peroxidase-conjugated anti-mouse IgG2a (1: 5,000; Accurate) or anti-mouse total Ig (1:5,000; Accurate) for 1 hour at 37°C and then by ABTS peroxidase substrate. A high-titer serum was plated in serial dilutions on each plate for quantification. Enzyme-linked immunospot (ELISpot) assay. Spleens were harvested by survival splenectomy either 2 or 8 weeks after cessation of treatment. Splenectomy has been shown not to affect the course of disease in (NZB ⫻ NZW)F1 mice (16), and it allowed us to correlate the immunologic phenotype with the eventual clinical outcome. Threefold serial dilutions of spleen cells in Dulbecco’s modified Eagle’s medium (DMEM)/ 10% FCS were plated in quadruplicate on DNA-coated plates starting at 1 ⫻ 106/well. The plates were spun and incubated for 2 hours at 37°C, after which 50 l of biotin-conjugated anti-IgM or anti-IgG (Southern Biotechnology, Birmingham, AL) diluted 1:700 in DMEM/10% FCS was gently added to each of two wells, and the plates were allowed to incubate overnight. The cells were then washed off, and the plates were incubated with streptavidin–alkaline phosphatase (Southern Biotechnology) diluted 1:1,000 in PBS/1% BSA for 45 minutes. Plates were then developed with 1 mg/ml of BCIP (Sigma, St. Louis, MO) in AMP buffer (0.75 mM MgCl2, 0.01% Triton X-405, 9.58% 2-amino-2-methyl-1-propanol, pH 10.25). Spots were counted using a dissecting microscope. Total numbers of immunoglobulin-secreting cells were measured the same way using anti-mouse immunoglobulins (Cappel, West Chester, PA) to coat the plates. ANTI-CD154 IN MURINE SLE Generation of hybridomas. Highly activated B cells from (NZB ⫻ NZW)F1 mice will spontaneously generate hybridomas without the need for immunization (14). Hybridomas were therefore generated from spleen cells by standard techniques using NSO cells as fusion partners. In some cases, spleen cells were also stimulated with 20 g/ml lipopolysaccharide (LPS; Sigma) for 48 hours prior to fusion. Hybridomas were screened for anti-dsDNA activity by ELISA as above. Positive hybridomas were then isotyped using specific peroxidase-conjugated secondary antibodies for IgM and IgG (Cappel). Analysis of class switching. Evidence of active class switching was sought by semiquantitative reverse transcription–polymerase chain reaction (RT-PCR) analysis of I-C transcripts from the spleens of 4–6 treated mice from groups A20 and B, and from the spleens of 5 nephritic mice from group C. Complementary DNA (cDNA) was generated using random primers, normalized for cDNA content using actin primers, and subjected to PCR using primers for the I␥2b and C␥2b exons as previously described (17). Analysis of the VHBW-16 gene. To determine the effect of treatment on somatic mutation, we performed a detailed analysis of the VHBW-16 gene as we have previously described (17). Briefly, IgG cDNA libraries were constructed by RTPCR from the 8 week–posttreatment spleens of 6 group A20 mice, 6 group B mice, and 8 age-matched controls from groups C and D. Between 200 and 400 colonies from each library were hybridized at 54°C with two VHBW-16–specific oligomers (for the first complementarity-determining region [CDR1], 5⬘-CTGCTGCAAGGCTTCTGGTT-3⬘ and for CDR2, 5⬘GGAATTAATCCTTACTATGGT-3⬘). Positive colonies were selected, and inserts from purified plasmids were sequenced in the Albert Einstein College of Medicine sequencing facility. Sequences were compared with the germline VHBW-16 sequence using a BLAST search (online at http:// www.ncbi.nlm.nih.gov/blast/). Mutations were analyzed in the V region only. Replacement mutation to silent mutation ratios were calculated, and the frequency of mutations at any position in the RGYW hot spot on either strand was determined (18). Mutations shared by more than one clonally related sequence were analyzed only once. Flow cytometry. Spleen cells were analyzed for B and T cell markers using antibodies to CD4 (Caltag, Burlingame, CA), CD8 (Caltag), and CD19 (BD PharMingen). The presence of activated CD4 cells was determined by double staining with fluorescein isothiocyanate (FITC)–anti-CD4 and phycoerythrin (PE)–anti-CD69 (PharMingen). The presence of naive and activated/memory CD4 cells was determined by triple staining with FITC–anti-CD4, Cy-Chrome–anti-CD44 (PharMingen), and PE–anti-CD62L (PharMingen). The presence of activated B cells was determined by double staining with FITC–anti-CD19 and PE–anti-CD69. Transitional, marginal zone, and follicular B cells were stained as previously described (19) using anti-CD19 allophycocyanin, PE–anti– heat-stable antigen (PE–anti-HSA), FITC–anti-CD21, and biotin–anti-CD23, followed by streptavidin–peridin chlorophyll protein. CD5⫹ cells were identified as B220intermediate and CD5⫹ using PE–anti-B220 and Cy-Chrome–anti-CD5. All antibodies to B cell markers were obtained from PharMingen. 497 Oxazolone immunization. To determine whether therapy had affected the T cell–dependent immune response to foreign antigen, 5 mice from group B and 5 age-matched controls from group C were immunized with 750 g of the hapten oxazolone (Sigma) by skin paint in olive oil/acetone on the shaved abdomen 3 days after the fourth dose of treatment (age 30 weeks). Mice were bled at weekly intervals thereafter for 3 weeks. Antioxazolone antibodies were measured by ELISA as previously described (20). Briefly, plates were coated overnight with 100 g/ml of oxazolone coupled to BSA in PBS, blocked, and then incubated sequentially with serum at 1:500 and 1:2,000 dilutions in PBS/1% BSA at 37°C for 1 hour, peroxidase-conjugated anti-mouse IgM, IgG1, IgG2a, or IgG3 (Accurate) at a 1:5,000 dilution in PBS/1% BSA at 37°C for 1 hour, and ABTS substrate. Statistical analysis. Proteinuria and survival data shown in Figure 1 were analyzed using Kaplan-Meier curves and the log rank test. Comparisons shown in Figures 3 and 8 were performed using the Wilcoxon rank sum test. Comparisons shown in Figures 4 and 6 and in Table 2 were performed using chi-square analysis or Fisher’s exact test. Only significant P values are shown. RESULTS Clinical outcomes. Mice treated with 250 g of anti-CD154 biweekly from age 20 weeks until age 46 weeks (group A20) and mice treated with 7 doses of 500 g of anti-CD154 from age 26 weeks until age 36 weeks (group B) had a markedly delayed onset of proteinuria (Figure 1). Proteinuria onset was delayed for an average of 19 weeks after cessation of therapy in both groups. Similarly, death was delayed by an average of 27–28 weeks after treatment cessation in both groups (Figure 1). Once proteinuria became manifest, the course of disease was similar to that in controls. There was no difference in disease course or survival between mice in the two control groups. The 26-week-old mice treated with 250 g of anti-CD154 biweekly (group A26) did not respond to therapy and were not studied further. This response failure was associated with the prompt development of anti-hamster antibodies in this group, but not in the other groups (data not shown), indicating that higher doses of therapy are needed to suppress the immune response as the degree of B cell activation increases with aging (12). Anti-dsDNA antibody findings. IgG anti-dsDNA antibodies and anti-dsDNA antibodies of the IgG2a isotype, the predominant isotype deposited in the kidneys of (NZB ⫻ NZW)F1 mice (21), were measured biweekly. Development of anti-dsDNA antibodies was delayed by an average of 20 weeks in group A20 (data not shown), a time equivalent to the treatment window. Six of 10 group B mice had high titers of anti-dsDNA 498 WANG ET AL Figure 1. Percentages of mice with fixed proteinuria ⬎300 mg/dl (left) and of mice surviving (right) after each treatment regimen (see Materials and Methods for a description of the treatment groups). Comparisons for proteinuria yielded the following P values: P ⫽ 0.008 for hamster IgG–treated control group (n ⫽ 14) versus group A20 (treatment commenced at 20 weeks); P ⫽ 0.024 for hamster IgG–treated control group versus group B (treatment commenced at 26 weeks); P ⫽ 0.0016 for untreated controls (n ⫽ 17) versus group A20; P ⫽ 0.001 for untreated controls versus group B. Comparisons for survival yielded the following P values: P ⫽ 0.0037 for hamster IgG–treated control group versus group A20; P ⫽ 0.043 for hamster IgG–treated control group versus group B; P ⫽ 0.0042 for untreated controls versus group A20; P ⬍ 0.02 for untreated controls versus group B. Group A26 mice did not respond to therapy and were not studied further. antibodies at the beginning or within the first 2–3 weeks of treatment initiation. In all these mice, anti-dsDNA antibody titers stabilized or diminished during treatment and rose again after the treatment course was completed (Figure 2). ELISpot assays. We have previously shown that neither long-term CTLA4-Ig nor short-term combination CTLA4-Ig/anti-CD154 treatment alters either the polyclonal expansion of IgM antibodies or the serum Figure 2. Measurement of IgG2a anti–double-stranded DNA (antidsDNA) antibodies (see Materials and Methods for description of the treatment groups). A, Data for 6 individual mice from group B that had serum IgG anti-DNA antibodies prior to treatment. Anti-dsDNA titers stabilized and, in most instances, decreased between 8 and 16 weeks after the start of therapy, reaching a low point at 16 weeks. Levels of autoantibodies rose again, peaked, and then fell concomitant with the onset of proteinuria. Mice were followed up until death. B, Representative data from 4 control mice. Anti-DNA antibody titers rose between 28 and 32 weeks and fell at the time of onset of proteinuria. All 4 mice were dead by age 40 weeks. C, Twofold serial dilutions of a high-titer serum. OD ⫽ optical density. levels of IgG antibodies in (NZB ⫻ NZW)F1 mice (14,17). Similarly, total serum IgM and IgG levels were unaffected by anti-CD154 treatment throughout the treatment course (data not shown). B cells of (NZB ⫻ NZW)F1 mice spontaneously secrete anti-dsDNA antibodies in vitro. To enumerate the frequency of these cells, ELISpot assays were performed on spleen samples from (NZB ⫻ NZW)F1 mice after cessation of anti-CD154 treatment. Frequencies of antibody-producing cells were compared with those in nephritic controls from groups C and D. Two mice in group B chosen at random from those whose antidsDNA antibody titers had declined after treatment were examined 2 weeks after cessation of treatment. These two mice displayed a marked suppression of IgG-secreting B cells, indicating a substantial suppression of splenic B cell maturation and class switching during anti-CD154 treatment. No effect was seen in these mice on the frequency of IgM anti-dsDNA antibody–secreting B cells, suggesting that CD40 ligation is not required for the selection of autoreactive B cells into the naive B cell compartment (Figure 3). The remaining mice in groups A20 and B were studied 8 weeks after cessation of treatment, at which time the anti-CD154 antibody had cleared from the serum. Eight mice from each group were examined. There was no difference in the frequency of IgMsecreting or IgM anti-dsDNA–secreting B cells between any of the treated mice and age-matched controls. In contrast, the effect of treatment on the frequency of IgG ANTI-CD154 IN MURINE SLE Figure 3. A, Frequencies of total IgM- and IgG-producing cells per 103 spleen cells and of IgM and IgG anti–double-stranded DNA (anti-dsDNA)–producing B cells per 105 spleen cells measured by enzyme-linked immunospot assay 2 weeks (n ⫽ 2 mice) and 8 weeks (n ⫽ 16 mice) after the final treatment (see Materials and Methods). Bars show the mean and SD. 22–25w ⫽ 22–25-week-old predisease control mice; 34–38w ⫽ 34–38-week-old nephritic control mice from groups C and D (see Materials and Methods for a description of the treatment groups). B, Dot-plot analysis of 16 anti-CD154–treated mice. Eight weeks after cessation of treatment, 8 mice from the treated groups had frequencies of IgG anti-dsDNA–producing B cells comparable with those in untreated controls, and 8 mice had fewer than 5 detectable IgG anti-dsDNA–secreting B cells per 105 spleen cells. anti-dsDNA antibody–secreting B cells was variable. (NZB ⫻ NZW)F1 mice age 22–25 weeks had low frequencies of IgG anti-dsDNA antibody–producing cells by ELISpot, and the frequencies increased ⬃20fold by the time of nephritis onset (Figure 3A). Eight of the 16 treated mice examined by ELISpot had a frequency of IgG anti-dsDNA–secreting cells that was not different from that in group C and D controls. However, these 8 mice had 4–10-fold lower titers of IgG antidsDNA antibodies in their serum than did age-matched controls (not shown) and had not yet developed renal disease. The other 8 treated mice had fewer than 5 IgG anti-dsDNA–secreting B cells per 105 spleen cells, a frequency comparable to that in pretreatment controls (Figure 3B). The degree of suppression of the anti-dsDNA response 8 weeks after cessation of treatment did not correlate with the treatment dose or the age at which treatment was instituted. The suppression of IgG antidsDNA–producing B cells was transient, however, and nearly all mice eventually developed high-titer IgG anti-dsDNA antibodies in the serum. Nevertheless, all 3 long-term survivors (⬎18 months) were in the group of 8 mice that demonstrated suppression of IgG antidsDNA antibody–secreting cells 8 weeks after treatment cessation, with a trend toward longer survival in this group (P ⫽ 0.09). 499 Analysis of hybridomas. Spleen cell fusions were performed to determine the activation state of antidsDNA antibody–producing B cells. Hybridomas were generated with the standard NSO fusion partner immediately after splenectomy and following a 2-day stimulation with LPS. In 8 control mice from group C, an average of 3.7% of the hybridomas bound dsDNA, and the ratio of IgG:IgM DNA-binding hybridomas was 3.5:1. No increase in hybridoma recovery or change in this ratio was observed after 2 days of LPS stimulation. In contrast, 10-fold fewer dsDNA-binding hybridomas were recovered from 9 anti-CD154–treated mice from either group A20 or group B 8 weeks after cessation of treatment (P ⬍ 0.0001), and the ratio of IgG to IgM DNA-binding hybridomas was 2:3 (P ⬍ 0.03), indicating that the IgG-producing B cells detected by ELISpot were not activated enough to produce hybridomas. After LPS stimulation, an increased number of hybridomas of the IgM isotype were recovered from 4 mice that had anti-dsDNA–producing cells detected by ELISpot (Figure 4). No increase in the frequency of either IgM or IgG anti-dsDNA–producing hybridomas was found after LPS stimulation in 5 treated mice that had a low frequency of IgG anti-dsDNA–producing B cells by ELISpot. Flow cytometry findings. Spleen cells from 6 nephritic controls from groups C and D (age 34–38 weeks), from 6 anti-CD154–treated mice from group Figure 4. Analysis of hybridomas. Hybridomas were generated from 9 mice 8 weeks after cessation of therapy (anti-CD154 treated) and from 8 control mice ages 34–38-week. Spleen cells from each mouse were divided in two, and fusions were performed immediately after harvest and after 2 days of stimulation with lipopolysaccharide (LPS). Data are shown as the mean and SD percentage of hybridomas of each isotype that were anti–double-stranded DNA (anti-dsDNA) positive. A total of 1,273 hybridomas from controls and 2,375 from treated mice were screened. The frequency of IgM or IgG anti-dsDNA–producing hybridomas from 2 mice that were examined 2 weeks after cessation of treatment was 2 of 437, with no increase after LPS stimulation (data not shown). See Materials and Methods for a description of the treatment groups. 500 WANG ET AL Table 1. B and T cell subsets in the spleens of control mice and mice examined 8 weeks after cessation of anti-CD154 treatment* Mice† No. of spleen cells (⫻ 107) CD19⫹‡ CD19⫹,CD69⫹§ CD19⫹,HSA⫹§ CD4⫹‡ CD4:CD8 ratio CD4⫹,CD69⫹‡ CD4⫹,CD44⫺,CD62L⫹¶ CD4⫹,CD44⫹,CD62L⫺¶ CD8⫹‡ Age 14–16 weeks (n ⫽ 4) Age 22–25 weeks (n ⫽ 3) Age 34–38 weeks (n ⫽ 6) Group A20 (n ⫽ 6) Group B (n ⫽ 4) ND ND 2.2 ⫾ 0.8 44.2 ⫾ 7.3 ND ND ND ND ND ND 9 ⫾ 1.0 42 ⫾ 13.6 ND ND 29.5 ⫾ 8.4 1.9 ⫾ 0.4 7.4 ⫾ 3.3 44 ⫾ 11.5 35.9 ⫾ 6.9 15.7 ⫾ 4.5 13.3 ⫾ 5.4 46 ⫾ 6.2 13 ⫾ 4.1 19.1 ⫾ 3.6 34.4 ⫾ 8.9 6.6 ⫾ 2.4 15 ⫾ 6.2 9.2 ⫾ 4.0 74.4 ⫾ 6.6 6.1 ⫾ 3.1 7.5 ⫾ 2.5 38.6 ⫾ 3.8 5.5 ⫾ 1.0 24.2 ⫾ 6.9 25.5 ⫾ 4.0 2.1 ⫾ 0.8 4.1 ⫾ 1.3 36.6 ⫾ 5.8 43.9 ⫾ 5.4 12.8 ⫾ 2.9 5.3 ⫾ 1.3 39.4 ⫾ 15.8 3.2 ⫾ 2.0 29 ⫾ 5.7 30.6 ⫾ 6.0 1.7 ⫾ 0.1 2.5 ⫾ 1.1 44.7 ⫾ 8.4 45.2 ⫾ 8.2 17.5 ⫾ 2.7 * Except where indicated otherwise, values are the mean ⫾ SD percentage of the given cell type. ND ⫽ not determined; HSA ⫽ heat-stable antigen (see Materials and Methods for detailed description of treatment groups). † Negative controls were ages 14–16 weeks and ages 22–25 weeks. Nephritic controls (from groups C and D) were ages 34–38 weeks. Group A20 mice were treated starting at age 20 weeks. Group B mice were treated starting at age 26 weeks with a higher dose protocol. ‡ Percentage of live cells. § Percentage of CD19⫹ gated cells. ¶ Percentage of CD4⫹ gated cells. 34–38-week-old controls) and 50 ⫾ 4.4% in 4 mice in which these cells were detected, compared with 42.9 ⫾ 4.2% in 34–38-week-old controls (P not significant). Flow cytometry was also performed to determine whether particular B and T cell subsets were preferentially depleted by anti-CD154 therapy. Treated mice had fewer CD19⫹,CD69⫹ B cells than 34–38-week-old controls (Tables 1 and 2). There was no change in the frequency of B220intermediate,CD5⫹ cells between 22–25week-old and 34–38-week-old control mice or between treated and control groups (data not shown). Mice ages 14–16 weeks had more immature (HSA⫹) cells than did controls ages 34–38 weeks, and the treated mice had a phenotype intermediate between these 2 groups (Table 1). There was a trend toward an increase in marginal zone B cells and follicular B cells with age, but there was no statistically significant difference between treated mice and either 14–16-week-old or 34–38-week-old controls (data not shown). In both treatment groups, anti- A20, and from 4 anti-CD154–treated mice from group B were analyzed by flow cytometry for expression of B and T cell markers 8 weeks after cessation of treatment. Four 14–16-week-old and three 22–25-week-old mice were examined as negative controls. The total number of spleen cells was decreased in the treated mice compared with either the 22–25-week-old or the 34–38-week-old control mice (Table 1). There was also a decrease in B cell frequency during treatment compared with 34–38week-old controls; this decrease persisted in some mice after cessation of treatment and correlated with a decrease in the frequency of Ig-secreting cells and with the presence or absence of anti-dsDNA antibody–producing cells by ELISpot. The mean ⫾ SD frequency of CD19⫹ B cells was 21.8 ⫾ 6.6% in the 2 mice whose spleens were harvested immediately after high-dose treatment ended. The frequency of B cells was 30.6 ⫾ 9% in 4 mice from the group with no detectable anti-dsDNA antibody–producing cells by ELISpot (P ⬍ 0.03 versus Table 2. Significant P values for the flow cytometry data shown in Table 1* Comparison No. of spleen cells (⫻ 107) CD19⫹, CD69⫹† CD19⫹, HSA⫹† CD4:CD8 ratio CD4⫹, CD69⫹‡ CD4⫹,CD44⫺, CD62L⫹§ CD4⫹,CD44⫹, CD62L⫺§ 34–38-week-old group vs. group A20 34–38-week-old group vs. group B 34–38-week-old group vs. 22–25-week-old group 22–25-week-old group vs. group B 0.03 0.004 NS 0.04 0.003 0.024 0.001 NS NS NS 0.03 NS 0.002 0.001 0.024 NS 0.003 0.01 0.03 NS 0.0001 0.0001 0.0001 NS 0.0001 0.0001 0.0001 NS * HSA ⫽ heat-stable antigen; NS ⫽ not significant (see Table 1 and Materials and Methods for detailed description of treatment groups). † Percentage of CD19⫹ gated cells. ‡ Percentage of live cells. § Percentage of CD4⫹ gated cells. ANTI-CD154 IN MURINE SLE Figure 5. Effect of treatment on class switching. Shown are the results of reverse transcription–polymerase chain reaction of spleen mRNA for IgG2b sterile class-switch transcripts. The upper band in each panel is the IgG2b sterile transcript. The lower band is actin. Three 5-fold serial dilutions of cDNA are shown for each mouse. Panels 1–6, Mice treated at age 26 weeks with high-dose therapy (group B). Panels 1 and 2, Treated mice whose spleens were harvested 2 weeks after cessation of treatment (IgG2b undetectable). Panels 3–6, Mice whose spleens were harvested 8 weeks after cessation of treatment (mean ⫾ SD IgG2b:actin ratio 0.08 ⫾ 0.10, n ⫽ 5 mice). Panel 7, A representative untreated control (IgG2b:actin ratio 0.98 ⫾ 0.43, n ⫽ 5 mice). Panels 8–12, Mice treated at age 20 weeks with low-dose therapy (group A20) whose spleens were harvested 8 weeks after cessation of treatment (IgG2b:actin ratio 0.52 ⫾ 0.32). Single lanes at left show the 2.1-kb marker. P not significant for group A20 versus control mice; P ⫽ 0.008 for group B versus control mice. See Materials and Methods for a description of the treatment groups. CD154 treatment prevented the loss of CD8⫹ T cells that occurs in these mice with age and inhibited the accumulation of both activated and memory CD4⫹ T cells as assessed by the markers CD69, CD44, and CD62L (Tables 1 and 2). Class switching. To determine the effect of treatment on class switching, semiquantitative RT-PCR was performed on spleen messenger RNA for sterile I␥2b– C␥2b class-switch transcripts. No transcripts were detected using a single round of PCR in the 2 mice whose spleens were harvested 2 weeks after cessation of treatment, indicating that anti-CD154 suppresses class switching. In the mice examined 8 weeks after treatment, class-switch transcripts were detected in all those in group A20, and results were variable in group B. This points to a dose-dependent effect of treatment on the length of time needed to recover the ability to class switch to IgG (Figure 5). Class-switch transcripts were absent only in mice that had a low frequency of IgG anti-dsDNA antibody–producing B cells detectable by ELISpot. Somatic mutation. The VHBW-16 gene, a member of the J558 gene family, has been shown to be strongly associated with pathogenic anti-dsDNA antibodies in (NZB ⫻ NZW)F1 mice (22). VHBW-16 is not expressed in normal mice, even by naive cells, unless tolerance to DNA is broken by immunization with DNA complexed to a DNA-binding protein. In these cases, the 501 autoantibodies are either all of the IgM isotype or of low affinity (23). These findings indicate that antibodies using this gene are regulated in the peripheral B cell compartment of normal mice. High-affinity anti-dsDNA antibody activity in (NZB ⫻ NZW)F1 mice is associated with class switching from IgM to IgG and with the presence of basic amino acids in the CDRs (17,23). We have previously shown that somatic mutations accumulate in the CDR2 of this gene with age (17). Thus, the VHBW-16 gene is an excellent marker to examine the effect of costimulatory blockade, both on the selection of naive autoreactive B cells and on activated B cells undergoing somatic mutation. Analysis of the VHBW-16 sequences from antiCD154–treated mice revealed findings similar to those we have previously reported with other forms of costimulatory blockade (14,17). In 1 of the 2 mice from group B harvested 2 weeks after cessation of treatment, we could not produce any IgG VHBW-16 product by PCR. In the other mouse, we recovered a few sequences that had 0–1 mutations, consistent with the ELISpot data. VHBW-16 IgG was detected in all 4 group A20 mice examined and in 6 of the 8 group B mice. There was no difference in the number of sequences containing arginines in the CDR3 compared with group C or D controls, indicating that regulation of naive bone marrow emigrants did not appear to be affected by therapy (data not shown). However, the frequency of somatic mutation was decreased in both the low- and highdose–treated mice compared with controls. Sixty percent of sequences had 0–1 mutations in mice from group A20, but ⬍30% had 0–1 mutations in mice from group B, suggesting that the effect of treatment on somatic mutation decreased with increasing disease activity. Nevertheless, ⬎90% of the sequences had 0–4 mutations in group B mice compared with ⬍60% in the controls, suggesting that accumulation of mutations could be blocked even in older mice (Figure 6). As we have previously observed with costimulatory blockade, there were also differences in the nature of the mutations that accumulated in the CDR2 of the treated mice. In nephritic mice, a mutation to arginine was a frequently observed change in positions 55, 56, and 58, where 4 RGYW hot spots overlap. In contrast, mutation to arginine was infrequent in the treated mice, being present in 9 of 44 control sequences, but in only 1 of 54 sequences from the treatment group (P ⬍ 0.005) (Figure 7). Analysis of RGYW hot-spot mutations was performed in the early treatment group in which mutations were occurring during therapy. The analysis was performed only for silent mutations that are not subject 502 WANG ET AL Figure 6. Effect of treatment on the frequency of somatic mutation of the VHBW-16 gene. In the early treatment group (group A20), 60% of the sequences had 0–1 mutations, compared with only 20% of the sequences in 34–38-week-old controls (P ⫽ 0.0023). In the late treatment group (group B), the frequency of somatic mutation was still significantly lower than that in 34–38-week-old controls (P ⫽ 0.023). See Materials and Methods for a description of the treatment groups. to selection bias. Despite what appeared to be a normal targeting of replacement mutations to the hot spots in the CDR2, there was a significant decrease in the frequency of silent hot-spot mutations, but not in that of silent non–hot-spot mutations, in the treated mice versus the hamster IgG–treated or untreated controls (P ⬍ 0.05) (Table 3). Immunization with oxazolone. In order to determine whether anti-CD154 treatment resulted in immunosuppression of humoral immune responses, we immunized mice in group B with the hapten oxazolone by skin painting without adjuvant 3 days after the fourth anti-CD154 treatment. Control age-matched mice from group C mounted a strong IgG antioxazolone response. In contrast, mice treated with anti-CD154 were unable to mount a response of either the IgG2a or the IgG1 isotype to oxazolone. Anti-CD154–treated mice developed an IgM antioxazolone response at week 1 postimmunization, but this had stabilized or diminished by week 3 postimmunization, whereas it continued to increase in controls (Figure 8). These studies show that treated mice are immunosuppressed during treatment, with failure to maintain or expand a mature classswitched humoral response after initial priming. DISCUSSION The SLE-like disease that develops in (NZB ⫻ NZW)F1 mice is dependent on the spontaneous activation of both T and B cells. The CD40/CD154 receptor/ ligand pair has been an attractive target for therapeutic intervention in SLE because it is involved both in costimulation of antigen-presenting cells and B cells by T cells (2) and in the effector arm of the immune response (7,24). Up-regulation of CD154 expression on T and B cells has been found very early in the life of lupus-prone mice (9), and similar findings have been reported in humans with SLE (25,26). CD40/CD154 blockade has been successful in preventing or stabilizing SLE nephritis in spontaneous murine models, even at relatively late stages of disease (10). In both murine and human SLE, there is spontaneous overproduction of polyclonal IgM antibodies, including dsDNA-binding antibodies. It is possible that aberrant expression of CD154 on T cells early in the course of murine SLE (9) might rescue autoreactive B cells from apoptosis during early B cell activation. If this were the case, treatment with anti-CD154 might induce deletion of naive autoreactive B cells. In contrast to observations of mice made deficient in CD40 or CD40L from birth (27,28), we found that spleens were smaller in treated mice than in young controls, and the percentage of spleen B cells was decreased by 50% during CD154 blockade. This effect was not specific for autoreactive B Figure 7. Pooled mutation analysis of the second complementaritydetermining region (CDR2) of VHBW-16 genes from the IgG cDNA libraries compared with the germline sequence. The middle of the diagram shows the germline sequence of the CDR2 (boldface); pooled data from treated (54 sequences) and control untreated (44 sequences) mice are shown below and above the germline sequence, respectively. Positions 57–60 (underlined) of the germline sequence contain an area of 4 overlapping RGYW hot spots. In this area, there are 9 mutations to arginine in the sequences from control untreated mice and only 1 such mutation in the sequences from treated mice. ANTI-CD154 IN MURINE SLE 503 Table 3. Mutation frequencies and patterns of the VHBW-16 gene* Group No. of sequences analyzed No. of mutations analyzed No. of mutations per sequence† Ages 34–38 weeks Anti-CD154 treated Ages 22–25 weeks 44 54 24 211 132 30 4.8 2.4 1.3 Replacement mutation: silent mutation ratio Framework CDR No. of silent mutations in hot spots‡ 1.8 2.7 2.4 3.3 3.6 3.3 19 7 – No. of silent mutations in non–hot spots 18 12 – * CDR ⫽ complementarity-determining region (see Materials and Methods for a description of the treatment groups). † P ⬍ 0.0001 for 34–38-week-old control mice versus pooled anti-CD154–treated mice; P ⫽ 0.001 for 22–25-week-old control mice versus anti-CD154–treated mice. ‡ Thirty control sequences and 28 early treatment sequences were included in this analysis. There are 86 RGYW hot spots and 202 non–hot spots in the VHBW-16 gene. Hot spot targeting was seen in the control sequences (ratio 2.47:1), but not in the treatment sequences (ratio 0.58:1) (P ⬍ 0.05). cells, since there was no change in the relative frequency of naive IgM anti-dsDNA–producing B cells enumerated by ELISpot. In addition, there was no apparent effect of anti-CD154 on selection of B cells into the B1 or marginal zone subsets that are a significant source of IgM anti-dsDNA antibodies (29,30). Finally, it appears that CD40 ligation is not required for priming of an initial IgM response to the hapten oxazolone in this autoimmune strain. Anti-CD154 treatment resulted in a block in the expansion and activation of autoreactive B cells, and class switching from IgM to IgG did not occur during the treatment period. Furthermore, as indicated by the studies using the marker gene VHBW-16, the frequency of somatic mutation was decreased and its pattern was altered. We observed a decrease in the frequency of RGYW hot-spot targeting in unselected mutations in our early treatment group, supporting the contention Figure 8. IgM and IgG isotype–specific antioxazolone antibodies. Antibodies were measured by enzyme-linked immunosorbent assay at weeks 1 and 3 after immunization with oxazolone in treated and age-matched control (NZB ⫻ NZW)F1 mice. Data for the 1:500 dilution of serum are shown for IgM and for the 1:2,000 dilution for IgG. Bars show the mean and SD. There was an increase in IgM antioxazolone titer between weeks 1 and 3 in all the control mice (P ⫽ 0.0039), but not in the treated mice. Treated mice did not mount an IgG antioxazolone response. Naive mice had optical density (OD) values ⬍ 0.05. that CD40 signaling has an effect on hot-spot targeting (31), but these data need to be confirmed with a larger number of sequences. Anti-CD154–treated mice had fewer mutations to arginine in the CDR2 than did untreated (NZB ⫻ NZW)F1 controls, a finding that we have previously observed following treatment with CTLA4-Ig (17) and short-term combination CTLA4-Ig/ anti-CD154 (14). We speculate that costimulatory blockade, through its effects on T cells or dendritic cells (32), alters B cell signaling thresholds, leading to regulation of high-affinity autoreactive B cells by deletion. The absence of high-affinity autoantibody-producing B cells will in turn result in the formation of less pathogenic immune complexes. Since activated autoreactive B cells have additional functions, including antigen presentation to T cells and secretion of cytokines (33,34), antiCD154 treatment should result in a global decrease in immune activation and consequent beneficial effects on disease. It is of potential clinical relevance that treatment with high-dose anti-CD154 therapy was successful even in mice that had already developed IgG anti-dsDNA antibodies. The fall in anti-dsDNA titers in these mice was associated with a decrease in total IgG-secreting B cells and with the disappearance of anti-dsDNA– secreting B cells from the spleen. These findings indicate that most of the anti-dsDNA response in early disease derives from B cells that are CD40 dependent. It appears from the results of ELISpot assays and hybridoma analysis that B cell function resumed in a stepwise manner after cessation of anti-CD154 treatment. During treatment, B cell numbers were substantially diminished. In the early stage of recovery, total B cell frequency increased, but the frequency of B cells producing IgG or IgG anti-dsDNA antibodies remained 504 low. In the next stage of recovery, high frequencies of IgG anti-dsDNA–producing B cells were readily detectable by ELISpot assay and were similar to those found in older (NZB ⫻ NZW)F1 mice with renal disease. However, serum titers of IgG anti-dsDNA antibodies were still 4–10-fold lower than in untreated age-matched controls, and no spontaneous IgG anti-dsDNA– producing hybridomas could be recovered from these mice even after LPS stimulation. The lack of response to LPS did not appear to be due to anergy, since spontaneous secretion of anti-dsDNA antibodies could be detected by ELISpot. Nevertheless, at this stage, the B cells of treated mice remained functionally different from those in the diseased controls. Flow cytometry analysis revealed that the frequency of CD69⫹ B cells in all mice examined 8 weeks after cessation of therapy was similar to that in young controls, even when anti-dsDNA antibody–producing cells were detected by ELISpot, showing that excessive B cell activation is a late step in the process of development of a high-titer serum antidsDNA response. A possible explanation for the inability of the IgG anti-dsDNA–producing B cells detected by ELISpot to secrete high titers of antibodies in vivo could be a lack of sufficient T cell help in the treated mice, since T cell function was also affected by treatment with antiCD154. We observed a striking difference in the effect of long-term administration of anti-CD154 on T cell activation compared with long-term administration of CTLA4-Ig. In mice treated with CTLA4-Ig, the abnormal accumulation of memory T cells that is associated with disease in older mice was not blocked despite the absence of the early activation marker CD69 on these T cells. This may be because CTLA4-Ig, in addition to blocking B7–CD28 interactions, blocks negative regulatory signals mediated via CTLA4 (17,35,36), or because activation of naive T cells by resting or anergic B cells results in only partial T cell activation in the presence of CTLA4-Ig (37). In contrast, anti-CD154 blocked memory T cell accumulation even after anti-dsDNA titers were beginning to rise, perhaps reflecting the need for CD40 ligation on antigen-presenting cells and/or memory B cells to generate or maintain an activated T cell response (38). Permanent tolerance was not achieved in the majority of treated mice. After a variable period of time following cessation of treatment, the T cell phenotype of the treated mice became indistinguishable from that of aged controls (data not shown), and the B cells appeared to attain a fully activated phenotype, in that pathogenic IgG anti-dsDNA antibodies began to appear in high titer WANG ET AL in the serum, and the mice developed proteinuria and progressive renal disease. Recently, two clinical trials of anti-CD154 antibody treatment in humans have been performed using humanized mouse anti-CD154 antibodies different from the antibody used in the present study. In one placebocontrolled trial, no efficacy of treatment was observed (39). It is possible that the dosage of drug used in that study was too low to achieve efficacy in most of the patients who already had active disease, analogous to the low-dose, 26-week-old group in our study. A second phase II study that used a dose of 10 mg/kg in a regimen analogous to the high-dose, 26-week-old group we describe here was stopped prematurely because of thrombotic events in 2 patients. Nevertheless, some efficacy was shown in the 18 patients who received at least 4 doses of the drug (15). In that study, only a 40–50% decline in anti-dsDNA titers was observed following anti-CD154 treatment despite the rapid disappearance of plasmablasts from the peripheral blood (40,41). As discussed above, most autoreactive B cells in the (NZB ⫻ NZW)F1 SLE model are CD40 dependent early in the disease course. However, after renal disease has developed, anti-CD154 treatment does not alter anti-dsDNA titers in (NZB ⫻ NZW)F1 mice (42). These antibodies are probably derived from long-lived plasma cells that are thought to be responsible for maintaining long-term antibody synthesis (43–45). These cells do not appear to require either T cells or the continuing presence of antigen to survive or secrete antibody (44), and they are therefore resistant to costimulatory blockade. Thus, patients with early disease (who have not yet developed a large autoreactive plasma cell population) may be more likely than patients with late disease that has proven refractory to other treatments to benefit from anti-CD154 treatment and other B cell–depleting treatments that spare plasma cells. Although the monoclonal antibody approach to CD40/CD154 blockade has not yet been successful in human clinical trials in SLE, the sequence of events that we have described here may be relevant to the application of other B cell–depleting therapies. Treatment with such agents may be useful in early disease or when given intermittently in combination with T cell blockade. With the appropriate combination of drugs and timing of treatment, it may be possible to achieve long-lasting effects on autoreactive cells with less global immunosuppression (14), pointing the way to safer and more effective therapy of human autoimmune disease. ANTI-CD154 IN MURINE SLE 505 ACKNOWLEDGMENTS We thank Drs. B. Diamond and H. Keiser for critical reading of the manuscript. 18. 19. REFERENCES 1. Foy TM, Laman JD, Ledbetter JA, Aruffo A, Claassen E, Noelle RJ. gp39-CD40 interactions are essential for germinal center formation and the development of B cell memory. J Exp Med 1994;180:157–63. 2. Grewal IS, Flavell RA. CD40 and CD154 in cell-mediated immunity. Annu Rev Immunol 1998;16:111–35. 3. Lederman S, Cleary AM, Yellin MJ, Frank DM, Karpusas M, Thomas DW, et al. The central role of the CD40-ligand and CD40 pathway in T-lymphocyte-mediated differentiation of B lymphocytes. Curr Opin Hematol 1996;3:77–86. 4. Han S, Hathcock K, Zheng B, Kepler TB, Hodes R, Kelsoe G. Cellular interaction in germinal centers: roles of CD40 ligand and B7-2 in established germinal centers. J Immunol 1995;155:556–67. 5. Morimoto S, Kanno Y, Tanaka Y, Tokano Y, Hashimoto H, Jacquot S, et al. CD134L engagement enhances human B cell Ig production: CD154/CD40, CD70/CD27, and CD134/CD134L interactions coordinately regulate T cell-dependent B cell responses. J Immunol 2000;164:4097–104. 6. Van der Voort R, Keehnen RM, Beuling EA, Spaargaren M, Pals ST. Regulation of cytokine signaling by B cell antigen receptor and CD40-controlled expression of heparan sulfate proteoglycans. J Exp Med 2000;192:1115–24. 7. Thienel U, Loike J, Yellin MJ. CD154 (CD40L) induces human endothelial cell chemokine production and migration of leukocyte subsets. Cell Immunol 1999;198:87–95. 8. Ma J, Xu J, Madaio MP, Peng Q, Zhang J, Grewal IS, et al. Autoimmune lpr/lpr mice deficient in CD40 ligand: spontaneous Ig class switching with dichotomy of autoantibody responses. J Immunol 1996;157:417–26. 9. Mohan C, Shi Y, Laman JD, Datta SK. Interaction between CD40 and its ligand gp39 in the development of murine lupus nephritis. J Immunol 1995;154:1470–80. 10. Kalled SL, Cutler AH, Datta SK, Thomas DW. Anti-CD40 ligand antibody treatment of SNF1 mice with established nephritis: preservation of kidney function. J Immunol 1998;160:2158–65. 11. Kalled SL, Cutler AH, Ferrant JL. Long-term anti-CD154 dosing in nephritic mice is required to maintain survival and inhibit mediators of renal fibrosis. Lupus 2001;10:9–22. 12. Early GS, Zhao W, Burns CM. Anti-CD40 ligand antibody treatment prevents the development of lupus-like nephritis in a subset of New Zealand black ⫻ New Zealand white mice: response correlates with the absence of an anti-antibody response. J Immunol 1996;157:3159–64. 13. Daikh DI, Finck BK, Linsley PS, Hollenbaugh D, Wofsy D. Long-term inhibition of murine lupus by brief simultaneous blockade of the B7/CD28 and CD40/gp39 costimulation pathways. J Immunol 1997;159:3104–8. 14. Wang X, Huang W, Mihara M, Sinha J, Davidson A. Mechanism of action of combined short term CTLA4Ig and anti-CD40L in murine SLE. J Immunol 2002;168:2046–53. 15. Boumpas DT, Furie RA, Manzi S, Illei GG, Balow JE, Vaishnaw A. A short-course of BG9588 (Anti-CD40L antibody) improves serologic activity and decreases hematuria in patients with proliferative lupus glomerulonephritis [abstract]. Arthritis Rheum 2001;44 Suppl 9:S387. 16. Theofilopoulos AN, Dixon FJ. Murine models of systemic lupus erythematosus. Adv Immunol 1985;37:269–390. 17. Mihara M, Tan I, Chuzhin Y, Reddy B, Budhai L, Holzer A, et al. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. CTLA4Ig inhibits T cell-dependent B-cell maturation in murine systemic lupus erythematosus. J Clin Invest 2000;106:91–101. Neuberger MS, Ehrenstein MR, Klix N, Jolly CJ, Yelamos J, Rada C, et al. Monitoring and interpreting the intrinsic features of somatic hypermutation. Immunol Rev 1998;162:107–16. Grimaldi CM, Michael DJ, Diamond B. Cutting edge: expansion and activation of a population of autoreactive marginal zone B cells in a model of estrogen-induced lupus. J Immunol 2001;167: 1886–90. Steinbrink K, Sorg C, Macher E. Low zone tolerance to contact allergens in mice: a functional role for CD8⫹ T helper type 2 cells. J Exp Med 1996;183:759–68. Dang H, Harbeck RJ. A comparison of anti-DNA antibodies from serum and kidney eluates of NZB ⫻ NZW F1 mice. J Clin Lab Immunol 1982;9:139–45. Katz MS, Foster MH, Madaio MP. Independently derived murine glomerular immune deposit-forming anti-DNA antibodies are encoded by near-identical VH gene sequences. J Clin Invest 1993;91:402–8. Ash-Lerner A, Ginsberg-Strauss M, Pewzner-Jung Y, Desai DD, Marion TN, Eilat D. Expression of an anti-DNA-associated VH gene in immunized and autoimmune mice. J Immunol 1997;159: 1508–19. Kuroiwa T, Lee EG, Danning CL, Illei GG, McInnes IB, Boumpas DT. CD40 ligand-activated human monocytes amplify glomerular inflammatory responses through soluble and cell-to-cell contactdependent mechanisms. J Immunol 1999;163:2168–75. Desai-Mehta A, Lu L, Ramsey-Goldman R, Datta SK. Hyperexpression of CD40 ligand by B and T cells in human lupus and its role in pathogenic autoantibody production. J Clin Invest 1996;97: 2063–73. Koshy M, Berger D, Crow MK. Increased expression of CD40 ligand on systemic lupus erythematosus lymphocytes. J Clin Invest 1996;98:826–37. Renshaw BR, Fanslow WC 3rd, Armitage RJ, Campbell KA, Liggitt D, Wright B, et al. Humoral immune responses in CD40 ligand-deficient mice. J Exp Med 1994;180:1889–900. Castigli E, Alt FW, Davidson L, Bottaro A, Mizoguchi E, Bhan AK, et al. CD40-deficient mice generated by recombinationactivating gene-2-deficient blastocyst complementation. Proc Natl Acad Sci U S A 1994;91:12135–9. Hirose S, Yan K, Abe M, Jiang Y, Hamano Y, Tsurui H, et al. Precursor B cells for autoantibody production in genomically Fas-intact autoimmune disease are not subject to Fas-mediated immune elimination. Proc Natl Acad Sci U S A 1997;94:9291–5. Zeng D, Lee MK, Tung J, Brendolan A, Strober S. Cutting edge: a role for CD1 in the pathogenesis of lupus in NZB/NZW mice. J Immunol 2000;164:5000–4. Monson NL, Foster SJ, Brezinschek HP, Brezinschek RI, Dorner T, Lipsky PE. The role of CD40-CD40 ligand (CD154) interactions in immunoglobulin light chain repertoire generation and somatic mutation. Clin Immunol 2001;100:71–81. Kalled SL, Cutler AH, Burkly LC. Apoptosis and altered dendritic cell homeostasis in lupus nephritis are limited by anti-CD154 treatment. J Immunol 2001;167:1740–7. Chan OT, Madaio MP, Shlomchik MJ. The central and multiple roles of B cells in lupus pathogenesis. Immunol Rev 1999;169: 107–21. Lipsky PE. Systemic lupus erythematosus: an autoimmune disease of B cell hyperactivity. Nat Immunol 2001;2:764–6. Riley JL, Blair PJ, Musser JT, Abe R, Tezuka K, Tsu T, et al. ICOS costimulation requires IL-2 and can be prevented by CTLA-4 engagement. J Immunol 2001;166:4943–8. Salomon B, Bluestone JA. Complexities of CD28/B7: CTLA-4 costimulatory pathways in autoimmunity and transplantation. Annu Rev Immunol 2001;19:225–52. 506 37. Ho WY, Cooke MP, Goodnow CC, Davis MM. Resting and anergic B cells are defective in CD28-dependent costimulation of naive CD4⫹ T cells. J Exp Med 1994;179:1539–49. 38. Howland KC, Ausubel LJ, London CA, Abbas AK. The roles of CD28 and CD40 ligand in T cell activation and tolerance. J Immunol 2000;164:4465–70. 39. Kalunian KC, Davis J, Merrill JT, Petri M, Buyon J, Ginzler E, et al. Treatment of systemic lupus erythematosus by inhibition of T cell costimulation [abstract]. Arthritis Rheum 2000;43 Suppl 9:S271. 40. Huang W, Sinha J, Newman J, Reddy B, Budhai L, Furie R, et al. The effect of anti-CD40 ligand antibody on B cells in human systemic lupus erythematosus. Arthritis Rheum 2002;46:1554–62. 41. Grammer AC, Shinohara S, Vazquez E, Gur H, Illei G, Lipsky PE. Normalization of peripheral B cells following treatment of active WANG ET AL 42. 43. 44. 45. SLE patients with humanized anti-CD154 mAb (5C8, BG9588) [abstract]. Arthritis Rheum 2001;44 Suppl 9:S282. Burns CM, Quesada S, Noelle RJ, Schned A. CD40-CD154 interactions in the pathogenesis of murine lupus: the beneficial effects of early and late anti-CD154 antibody treatment appear to be mediated through different mechanisms [abstract]. Arthritis Rheum 2001;44 Suppl 9:S397. Manz RA, Radbruch A. Plasma cells for a lifetime? Eur J Immunol 2002;32:923–7. Manz RA, Lohning M, Cassese G, Thiel A, Radbruch A. Survival of long-lived plasma cells is independent of antigen. Int Immunol 1998;10:1703–11. Slifka MK, Ahmed R. Long-lived plasma cells: a mechanism for maintaining persistent antibody production. Curr Opin Immunol 1998;10:252–8.