Abstract
Short-term combination therapy with the costimulatory antagonists CTLA4Ig and anti-CD40 ligand induces prolonged suppression of disease onset in New Zealand Black/New Zealand White F1 systemic lupus erythematosus-prone mice. To determine the mechanism for this effect, 20- to 22-wk-old New Zealand Black/New Zealand White F1 mice were treated with six doses each of CTLA4Ig and anti-CD40 ligand Ab over 2 wk. Combination-treated mice, but not mice treated with either agent alone, had prolonged survival and the production of pathogenic IgG anti-dsDNA Ab was suppressed. Twenty weeks after completion of treatment the frequency of activated B cells producing anti-dsDNA Ab was decreased, and the abnormal transition of T cells from the naive to the memory compartment was blocked. Combination treatment partially suppressed class switching and decreased the frequency of somatic mutations in the VHBW-16 gene, which is expressed by pathogenic anti-DNA Abs. Treated mice were still able to respond to the hapten oxazolone when it was given 8 wk after treatment initiation, and they mounted a somatically mutated IgG anti-oxazolone response that was noncross-reactive with dsDNA. Fifty to 60% of previously treated mice, but only 14% of previously untreated mice, responded within 2–3 wk to a second course of therapy given at the onset of fixed proteinuria and remained well for a further 3–4 mo. Although this treatment had no immediate effect on serum anti-dsDNA Abs or on the abnormal T cell activation observed in sick mice, 25% of treated mice lived for >18 mo compared with 5% of untreated controls. These results suggest that the effect of costimulatory blockade in remission induction must be mediated by a different mechanism than is demonstrated in the disease prevention studies.
The anti-dsDNA Abs that are pathogenic in systemic lupus erythematosus (SLE)4 are produced by B lymphocytes that have undergone Ig class switching and somatic mutation under the influence of help from activated T lymphocytes (1, 2). Communication between T and B cells and between T cells and APCs is Ag specific but requires additional activation signals provided by the interaction of costimulatory receptor/ligand pairs, the most important of which are B7/CD28 and CD40/CD40 ligand (CD40L) (3, 4). Both B and T cell function can be modulated by blocking these costimulatory interactions using either CTLA4Ig, which blocks the interaction of B7 with CD28 (5), or anti-CD40L, which blocks the interaction of CD40 with CD40L (6).
We and others found that long-term but not short-term administration of CTLA4Ig to New Zealand Black/New Zealand White (NZB/NZW) F1 mice prevented the onset of SLE, but that disease manifestations appeared 4–8 wk after stopping treatment (7, 8). In contrast, a 2-wk course of anti-CD40L given early in life caused a 2- to 3-mo delay in the onset of disease. This was apparently due to inactivation of autoimmune B cells, with T cell function remaining intact (9). Synergy between these two reagents has been observed in several transplantation models and in induced autoimmunity models where administration of drug can be timed according to the time of Ag exposure (10). In the spontaneous autoimmune disease SLE, Daikh et al. (11) reported that a 2-wk course of combined therapy using CTLA4Ig and anti-CD40L given to prenephritic 5-mo-old mice that already display some evidence of an activated autoimmune phenotype delayed disease onset for up to 6 mo. The mechanism for this effect has not been characterized and the understanding of this issue is crucial to the development of therapeutic regimens for humans with already established autoimmune disease.
The goal of this study was to understand the immunologic effects of combined CTLA4Ig and anti-CD40L therapy on the development, selection, and activation of pathogenic anti-dsDNA Ab-producing B cells and on the activation of T cells in the NZB/NZW F1 murine model of SLE. We found that a 2-wk course of treatment of prenephritic mice with combination CTLA4Ig/anti-CD40L resulted in delayed onset of renal dysfunction, a marked decrease in the frequency of B cells producing IgG anti-DNA Abs, a partial suppression of class switching, and inhibition of T cell activation and switching to a memory phenotype. In contrast, treatment did not appear to affect the abnormal activation or selection of naive autoreactive B cells. Using an autoreactive V region gene as a marker, we were able to show a long-term effect on somatic mutation and on selection specifically of autoreactive B cells without any adverse effect on the response to foreign Ag, indicating that this therapy did not mediate its effects via global immunosuppression. After onset of proteinuria, a repeat course of treatment was able to induce remission in previously treated but not in previously untreated mice and several mice became long-term survivors. These results shed new light on the mechanisms by which short-term costimulatory blockade inhibits autoreactive B cell maturation and prevents SLE onset.
Materials and Methods
Mice
NZB/NZW F1 females were purchased from The Jackson Laboratory (Bar Harbor, ME) and maintained in a conventional animal housing facility throughout the experiment. Fully murine CTLA4Ig was obtained from the serum of SCID mice injected with a CTLA4Ig-expressing adenovirus (8). Mice were treated at the age of 20–22 wk with a combination of 100 μg of CTLA4Ig and 250 μg of anti-CD40L (MR1, a kind gift of Dr. S. Kalled, Biogen, Cambridge, MA), each given i.p. six times over a 2-wk period (24 mice). Control mice received either CTLA4Ig alone (12 mice), anti-CD40L alone (13 mice), or combination mouse IgG2a and hamster IgG (21 mice). Before treatment, mice were tested for anti-dsDNA Abs and were randomized into treatment groups depending on the titer of Abs. Mice were bled every 2–4 wk. Urine was tested for proteinuria by dipstick (Multistick; Fisher, Pittsburgh, PA) every 2 wk. Mice that received combination therapy or that were previously untreated with any therapy or control Ab (14 mice) received a 2-wk course of combination therapy as above when they developed fixed proteinuria of 300 mg/dl. Survival splenectomies were performed at intervals following treatment. One group of 14 combination-treated mice was followed for 2 years. The remaining mice were followed until 18 mo (78 wk) of age. Mice injected with a single dose of adenovirus-CTLA4Ig have been described previously (8).
Anti-DNA Abs
ELISA plates (Falcon) were coated with 100 μl of 100 μg/ml salmon sperm DNA made double stranded by passage through a 45-μm filter (USA Scientific, Ocala, FL). After drying, the plates were blocked and then incubated sequentially for 1 h at 37°C with serial dilutions of serum followed by peroxidase-conjugated F(ab′)2 goat anti-mouse Ig (1/8000) or anti-IgG2a (1/5000) in PBS/1% BSA (Accurate Antibodies, Westbury, NY) and then ABTS substrate (Kirkegaard & Perry, Gaithersburg, MD).
ELISPOT assay
Spleens were harvested 16–20 wk after the first treatment and 4–8 wk after the second treatment. Four to five mice were examined in each group. Threefold serial dilutions of spleen cells were plated in duplicates overnight on DNA-coated plates starting at 1 × 106 cells per well. After washing, the plates were incubated with biotin-conjugated anti-IgM or anti-IgG (Southern Biotechnology Associates, Birmingham, AL) 1/1000 in PBS/1% BSA for 2 h, followed by streptavidin alkaline phosphatase (Southern Biotechnology Associates) 1/1000 in PBS/1% BSA for 45 min. Plates were then developed with 1 mg/ml 5-bromo-4-chloro-3-indolyl phosphate (Sigma-Aldrich, St. Louis, MO) in AMP buffer (0.75 mM MgCl/0.01% Triton-X/9.58% 2-amino-methyl-1-propanol, pH 10.25). Spots were counted using a dissecting microscope. Total numbers of Ig-secreting cells were measured the same way using anti-mouse Igs (Cappel, Westchester, PA) to coat the plates.
Generation of hybridomas
Hybridomas were generated from spleen cells by standard techniques using NSO cells as fusion partner. In some cases, spleen cells were also stimulated with 20 μg/ml LPS (Sigma-Aldrich) for 48 h before fusion. Hybridomas were screened for anti-dsDNA activity by ELISA as above. Positive hybridomas were then isotyped using specific peroxidase-conjugated secondary Abs for IgM and IgG (Cappel).
Analysis of class switching
Evidence of active class switching was sought by semiquantitative RT-PCR analysis of I-C transcripts from spleens of five treated and five age-matched control mice. cDNA was generated using random primers, normalized for cDNA content using actin primers, and subjected to PCR using primers for the Iγ2b and Cγ2b and for the Iγ1 and Cγ1 exons, as previously described (8).
Analysis of the VHBW-16 gene
To understand the molecular basis for the treatment-induced decrease in frequency of B cells producing anti-dsDNA Abs of the IgG isotype, we performed a detailed analysis of the VHBW-16 gene as we have previously described (8). IgG cDNA libraries were constructed by RT-PCR from the spleens of 10 combination-treated mice and eight age-matched controls. Two hundred to 400 colonies from each library were hybridized at 54°C with two VHBW-16-specific oligomers (complementarity-determining region (CDR)1, 5′-CTGCTGCAAGGCTTCTGGTT-3′; CDR2, 5′-GGAATTAATCCTTACTATGGT-3′). The quality of each library was confirmed by stripping and rehybridizing each filter with the IgM or IgG constant region oligomers. Positive colonies were picked 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 basic local alignment search tool search (http://www.ncbi.nlm.nih.gov/blast/). Mutations were analyzed in the V region only and the last four nucleotides of the V region were excluded from the mutation analysis, as there were many differences in these nucleotides due to junctional diversity. Replacement to silent mutation ratios were calculated and the frequency of mutations at the RGYW hot spot on both strands was determined (12). Mutations shared by more than one clonally related sequence were analyzed only once. Comparisons of percentage of hot spot mutations in treated vs untreated mice were performed using the χ2 test. Finally, the number of transitions and transversions was determined and compared with the expected number of randomly generated mutations based on the frequency of each nucleotide in the VHBW-16 sequence.
Flow cytometry
Spleen cells were analyzed for B and T cell markers using Abs to CD4 (Caltag Laboratories, Burlingame, CA), CD8 (Caltag Laboratories), and CD19 (BD PharMingen, San Diego, CA). Presence of activated CD4 cells was determined by double staining with FITC-anti-CD4 and PE-anti-CD69 (BD PharMingen). Presence of naive and activated/memory CD4 cells was determined by triple staining with FITC-anti-CD4, CyChrome-anti-CD44 (BD PharMingen), and PE-anti-CD62L (BD PharMingen).
Oxazolone immunization
To determine whether therapy had long-term effects on the immune response to foreign Ag, groups of five mice were immunized with 750 μg of the hapten oxazolone (Sigma-Aldrich) by skin paint in olive oil/acetone on the shaved abdomen 8 wk after initiation of treatment. Mice were bled at weekly intervals thereafter. Four to 6 wk later mice were boosted with 30 μg of oxazolone by skin paint and spleens were harvested for analysis 3 days later.
Anti-oxazolone Abs
Anti-oxazolone Abs were measured by ELISA as previously described (13). In brief, plates were coated overnight with 100 μg/ml oxazolone-coupled BSA (Ox-BSA), blocked, and then incubated sequentially with serum at either 1/500 or 1/2000 dilution in PBS/1% BSA at 37°C for 1 h, anti-mouse IgG at 1/5000 dilution in PBS/1% BSA at 37°C for 1 h, and ABTS substrate.
Anti-oxazolone ELISPOTs
Plates were coated with Ox-BSA as above and ELISPOTs were performed exactly as described for dsDNA as above.
Analysis of the VH-Ox15 gene
To determine whether the immune response to oxazolone underwent affinity maturation, cDNA libraries were made of the VH-Ox15 gene from spleens of four combination-treated and three control mice. The VH-Ox15 gene dominates the anti-oxazolone response in BALB/c mice (14), and we had previously determined that ∼30% of anti-oxazolone hybridomas from immunized but otherwise unmanipulated NZB/NZW F1 mice use this H chain gene (M. Mihara and A. Davidson, unpublished data). The sequence of the NZB/NZW F1 VH-Ox15 gene differs from the BALB/c sequence by two nucleotides, with a single amino acid difference (Gly-Val) at position 10 (M. Mihara and A. Davidson, unpublished observation). cDNA libraries were generated using a VH-Ox15 leader primer (5′-GCCCCCATCAGAGCATGGC) and an IgG constant region primer (8). Colonies were screened with a leader primer downstream of the primer used for PCR (5′-TGCCTGGTTGCATTTCCAA) and with a framework 3 primer (5′-AGCAAAGACAACTCCAAGAGC). Initial sequence analysis of 30 colonies revealed that only those colonies that hybridized with both primers contained the correct gene so that subsequently only double-positive colonies were picked for sequencing. Positive colonies were then analyzed for mutations as above. For comparison, analysis of the VHBW-16 gene was performed on the same spleens.
Statistical analysis
Proteinuria, survival, and anti-DNA Ab data shown in Figs. 1–3⇓⇓⇓ were analyzed using Kaplan Meier curves and log rank test. Comparisons shown in Figs. 3⇓, 5⇓, 9⇓, and 10⇓ were performed using Wilcoxon rank sum test. Comparisons shown in Figs. 4⇓ and 7⇓ and Table I⇓ were performed using χ2 analysis or Fisher’s exact test. Only significant p values are shown.
Onset of proteinuria. Shown is the cumulative percentage of mice in each group that developed fixed proteinuria (>300 mg/dl) over time (p < 0.0001, combination-treated mice vs hamster/mouse controls).
Survival curves. Shown is the cumulative percentage of mice that had died at each time. All mice were followed to 75 wk. After 75 wk, 14 of 24 mice in the combination treatment group were followed until death or wk 102 (p < 0.0001, combination-treated mice vs controls).
Frequency of spontaneous anti-dsDNA Ab-producing B cells. Shown are results of ELISPOT assays (mean ± SD) for frequency of anti-dsDNA Ab-producing B cells (per 105 cells) and total Ig-producing B cells (per 103 cells) in the spleens of combination-treated and age-matched control mice at 34–38 wk. Six mice were examined in each group and compared with three 22- to 25-wk-old mice. None of the 22- to 25-wk-old mice and all of the 32- to 40-wk-old control mice had high-titer anti-dsDNA Abs. Significant p values were as follows: anti-dsDNA IgG, treated vs control, p = 0.0003; anti-dsDNA IgM, treated vs control, p = 0.021; spleen total IgG, treated vs control, p = 0.0014.
Frequency of autoreactive hybridomas. Hybridomas were generated from eight treated (1811 hybridomas) and six control 34- to 38-wk-old mice (1125 hybridomas) and screened for production of IgG and IgM anti-dsDNA Abs. Hybridomas were also generated from six mice 4–8 wk after the second treatment (1789 hybridomas). There was a marked decrease in spontaneous IgG anti-dsDNA Ab-producing hybridomas in the treated mice after both the first (p < 0.0001) and second (p = 0.0003) treatments. Data are presented as percent of total hybridomas that were positive.
Sequence analysis of VHBW-16 and VHOx15 genes in treated and control mice
Results
Clinical outcomes
Combination therapy with 2 wk of CTLA4Ig and anti-CD40L, but not single therapy with either agent alone, delayed the onset of proteinuria (Fig. 1⇑). Seventeen percent of combination-treated mice had not developed proteinuria by age 74 wk compared with 0% of mouse IgG2a/hamster IgG-treated controls (p < 0.0001). Retreatment with a second course of therapy at the time of onset of proteinuria resulted in complete remission (defined as reversion of proteinuria to 30 mg/dl lasting at least 6 wk) in 11 of 21 combination-treated mice compared with only two of 14 previously untreated controls (p < 0.04). Durable remissions of 29, 38, and >52 wk were observed in three retreated mice. The duration of remission in the other retreated mice was 12.6 ± 3.9 wk. Survival curves are shown in Fig. 2⇑.
Anti-DNA Abs
IgG anti-dsDNA Abs and anti-dsDNA Abs of the IgG2a isotype were measured biweekly. IgG2a Abs are the predominant isotype deposited in the kidneys of NZB/NZW F1 mice (15). Control mice and mice treated with CTLA4Ig or anti-CD40L as single therapy all developed high-titer IgG2a anti-dsDNA Abs by the age of 28.4 ± 6 wk. Combination-treated mice developed IgG2a anti-dsDNA Abs at 42.6 ± 10.3 wk (p < 0.0001, data not shown). After retreatment, anti-dsDNA Abs did not decrease concurrent with the disappearance of proteinuria (data not shown).
ELISPOT assays
We have previously shown that long-term CTLA4Ig treatment does not alter either the polyclonal expansion of IgM Abs or the serum levels of IgG Abs in NZB/NZW F1 mice (8). Similarly, total serum IgM and IgG levels were unaffected by combination CTLA4Ig/anti-CD40L treatment (data not shown). B cells of NZB/NZW F1 mice spontaneously secrete Abs in vitro. To enumerate total numbers of Ig-secreting and anti-dsDNA Ab-producing B cells in the treated mice and follow their kinetics, total Ig and anti-dsDNA ELISPOTs were performed on spleen samples from combination-treated NZB/NZW F1 mice at 16–20 wk after the first treatment and compared with age-matched hamster IgG and IgG2a treated controls. The frequency of IgM anti-DNA-secreting B cells in the spleens of treated mice was similar to that seen in young NZB/NZW mice and was significantly different from that observed in age-matched control mice (p < 0.03). Significant differences were also observed for IgG Ab-producing cells. Sixteen to 20 wk after the first treatment, the frequency of IgG anti-dsDNA Ab-secreting B cells was 60-fold less in treated mice than in age-matched controls (p = 0.0003). This effect was partially accounted for by a 4-fold reduction in overall numbers of IgG-secreting cells in the spleens (p < 0.002), but the majority of the effect appeared to be due to failure of the expansion of IgG anti-dsDNA Ab-secreting cells that occurs with age in unmanipulated mice (Fig. 3⇑). Variable results were observed in mice that had been treated a second time. In these mice, total numbers of IgM- and IgG-secreting B cells were not different from controls, but there was a decrease in IgG anti-DNA-secreting B cells in three of the five mice (0.1–4/105 cells), including two that were studied at the earliest time, 4 wk post-treatment (data not shown).
Analysis of hybridomas
Highly activated B cells from NZB/NZW F1 mice will spontaneously generate hybridomas without the need for immunization. Spleen cell fusions were performed to determine the activation state of anti-DNA Ab-producing B cells and to determine whether they had been anergized. Hybridomas were generated with the standard NSO fusion partner immediately after splenectomy and following a 2-day stimulation with LPS that allows generation of hybridomas from anergic B cells. Far fewer DNA-binding hybridomas of the IgG class were recovered from combination-treated mice than from controls (p = 0.0001, Fig. 4⇑). Furthermore, there was no rescue of autoreactive hybridomas using LPS-stimulated cells (data not shown), indicating that anergy is not a major mechanism for the observed results. These data, together with the ELISPOT data, suggest that the age-related proliferation and activation of IgG anti-dsDNA Ab-producing B cells had been blocked by short-term combination treatment. In the mice studied after the second treatment, IgG anti-DNA-producing hybridomas were recovered, even in mice that had low numbers of anti-DNA-secreting cells by ELISPOT, but the frequency was still less than in untreated controls (p = 0.0003, Fig. 4⇑)
Flow cytometry
Spleen cells from six combination-treated and six age-matched nephritic control mice (34–38 wk) were analyzed by FACS for expression of B and T cell markers. Three 22-wk-old NZB/NZW F1 mice and two BALB/c mice were examined as negative controls. The total number of spleen cells was increased in the control mice compared with either the young or the treated mice (1.33 ± 0.54 × 108 in controls vs 0.7 ± 0.25 × 108 in treated mice; p < 0.02, and 0.9 ± 0.1 × 108 in young mice). Untreated NZB/NZW F1 mice had the same proportions of B and T cells as normal BALB/c controls. Treated mice had decreased numbers of B cells (p < 0.003, Fig. 5⇓). Young NZB/NZW F1 mice had normal numbers of CD8 cells; however, older nephritic mice displayed a loss of CD8 cells with a markedly increased CD4:CD8 ratio. In contrast, the treated mice had normal numbers of CD8+ cells and a normal CD4:CD8 ratio (p < 0.02 vs age-matched controls). Young NZB/NZW F1 mice showed evidence of CD4 T cell activation as demonstrated by increased expression of the early activation marker CD69. Even higher numbers of activated CD4 cells were observed in the older NZB/NZW F1 mice. In treated mice, levels of CD69 expression were comparable with the BALB/c controls (p < 0.0001 vs untreated controls, Fig. 5⇓, left panel). Young NZB/NZW F1 mice had similar numbers of naive (CD44lowCD62Lhigh) and memory (CD44highCD62Llow) CD4 cells as BALB/c mice. However, in older nephritic mice >80% of the CD4 T cells had a memory phenotype and <5% of the cells had a naive phenotype. This phenotypic pattern was also observed in Ad-CTLA4Ig-treated mice (Fig. 5⇓, right panel). However, in combination-treated mice transition of CD4 T cells from the naive to the memory compartment was inhibited (p < 0.003 vs untreated controls).
Flow cytometry analysis of spleen cells. Left panel, Mean percentage of CD4-, CD8-, CD19-, and CD4/CD69-positive cells (± SD) in spleens from six treated and six control NZB/NZW F1 mice aged 34–38 wk, six Ad-CTLA4Ig-treated mice, three young NZB/NZW F1 mice, and two BALB/c mice. CD19 counts were lower in combination-treated mice than in controls (p = 0.0027). Percentage of CD8 cells were higher, CD4:CD8 ratios were lower (p = 0.015), and the percentage of CD4/CD69 cells was lower (p = 0.0007) in treated mice vs age-matched controls and were similar to those found in mice treated with continuous CTLA4Ig. Right panel, Mean percentage of naive (CD44lowCD62Lhigh) and memory (CD44highCD62Llow) CD4 T cells. There was no difference between Ad-CTLA4Ig-treated mice and age-matched controls. In contrast, in combination-treated mice, the proportion of naive and memory cells was no different from young controls (treated vs age-matched control p = 0.0022).
We also examined three mice that were treated a second time after the development of nephritis and entered sustained remission. The B and T cell phenotype of these mice was no different from that of untreated sick controls (data not shown).
Class switching
To determine whether the effects on IgG anti-dsDNA Ab-producing B cells was due to a block in class switching, semiquantitative RT-PCR was performed on spleen mRNA for sterile Iγ2b-Cγ2b class switch transcripts. Sixteen to 20 wk after treatment there was a decrease in the class switch transcripts in the treated mice compared with controls (Fig. 6⇓). A similar decrease in class switch transcripts was observed in three of four spleens harvested 4–8 wk after second treatment of mice that had developed proteinuria (Fig. 6⇓). Similar results were observed for IgG1 class switch transcripts (data not shown).
Effect of treatment on class switching: RT-PCR of spleen mRNA for IgG2b sterile class switch transcripts. A, Combination-treated mice. B, Control age-matched mice. C, Four to 8 wk after second treatment. The upper band in each panel is the IgG2b sterile transcript; the lower band is actin. For each mouse, three 5-fold serial dilutions of cDNA are shown. Lane 1 in each panel is marker.
Somatic mutation
The VHBW-16 gene, a member of the J558 gene family, has been shown to be strongly associated with pathogenic anti-dsDNA Abs both in NZB/NZW F1 and MRL/lpr mice (16). Despite its presence in the germline, VHBW-16 is not expressed in several normal mouse strains, even by naive cells, unless such mice are experimentally induced to mount an anti-dsDNA Ab response by immunization with DNA complexed to a DNA-binding protein. In these cases, the autoantibodies either are all of the IgM isotype or are of low affinity (17). These findings indicate that Abs using this gene are regulated in the peripheral B cell compartment of normal mice. High-affinity anti-dsDNA Ab activity in NZB/NZW F1 mice is further associated with class switching from IgM to IgG and with the presence of basic amino acids, particularly arginine, in CDR2 (acquired by somatic mutation) and CDR3 (acquired during gene rearrangement in the bone marrow) of VHBW-16 (8, 17). We have previously shown that somatic mutations accumulate in this gene with age (8). The VHBW-16 gene is thus 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 combination-treated mice revealed several interesting findings. First, there was no difference in the number of sequences containing arginines in the CDR3 compared with age-matched controls, indicating that regulation of naive bone marrow emigrants did not appear to be affected by therapy (data not shown). Second, the frequency of somatic mutation was decreased in the treated mice compared with untreated controls. The number of sequences containing no mutations or one mutation in treated mice was no different from the number of sequences in young NZB/NZW mice but was significantly different from age-matched controls (p < 0.004, Fig. 7⇓). Together these data point to a long-term effect of treatment on the ability of B cells using this gene to receive or respond to T cell help (Table I⇑). Third, there were striking differences in the nature of the mutations that accumulated in the treated mice. Most mutations in the CDRs were found in CDR2, which has four contiguous RGYW hot spots spanning positions 55–58. All four of these positions were mutated in control mice; however, in treated mice far fewer mutations were observed at positions 55–57. A frequently observed change in positions 55, 56, and 58 in untreated mice was mutation to arginine, a positively charged amino acid that confers increased affinity for DNA. In contrast mutation to arginine was infrequent in the treated mice, the most frequently observed change in combination-treated mice being to threonine, a neutral amino acid (Fig. 8⇓).
Effect of treatment on somatic mutation. Frequency of IgG VHBW-16 sequences containing zero to one, two to five, or more than six mutations per sequence is shown. Twenty-four sequences from young NZB/NZW F1 mice, 44 sequences from 34- to 38-wk-old mice, and 46 sequences from combination-treated mice were analyzed. The percentage of unmutated sequences was significantly different in treated mice than in age-matched 34- to 38-wk-old control mice (p = 0.0006) but was not significantly different from young NZB/NZW F1 mice.
Mutation analysis. Shown is pooled mutation analysis of the CDR2 of VHBW-16 genes from the IgG cDNA libraries compared with the germline sequence. The germline sequence of CDR2 is shown in italics in the middle and pooled data from treated (46 sequences) and control untreated (44 sequences) mice are shown above and below the germline sequence, respectively. Positions 55–58 containing an area of four overlapping RGYW hot spots are underlined. In this area, there are nine mutations to arginine in the sequences from control untreated mice and only one in the sequences from treated mice. Twenty-four sequences from 22-wk-old NZB/NZW F1 mice were also analyzed, but these had few mutations (shown below control group).
Immunization with oxazolone
To determine whether the suppression of the IgG anti-DNA Ab response was due to a global effect on B cell responses or was specific to the autoreactive B cell response, we immunized mice 8–10 wk after treatment with the hapten oxazolone without adjuvant. Control mice mounted a strong IgG anti-oxazolone response. In contrast, mice that had been injected with adenovirus-expressing CTLA4Ig and that were expressing high levels of CTLA4Ig in the serum were unable to respond to oxazolone, indicating an inability to mount an immune response to a T-dependent Ag. Mice treated with either 2 wk of single CTLA4Ig or anti-CD40L therapy or with the short-term combination mounted an anti-oxazolone response equivalent to controls at both serum dilutions tested, indicating that once the CTLA4Ig and anti-CD40L had cleared from the serum the immune response to foreign Ag was not compromised (Fig. 9⇓). To determine the effect of treatment on expansion of Ag-specific B cells, ELISPOT assays were performed on control and combination-treated mice. The frequency of IgG anti-oxazolone-secreting cells was not significantly different between treated mice and untreated controls, but the frequency of IgG anti-DNA-secreting cells was significantly diminished (Fig. 10⇓, p < 0.03). Hybridomas were generated from the spleens and tested for binding to oxazolone and to DNA. The frequency of IgG oxazolone-binding hybridomas was similar in treated and control mice (1.2 vs 1.3%), indicating that there was no effect of treatment on activation of anti-oxazolone-secreting B cells. However, the frequency of IgG anti-DNA hybridomas was diminished 10-fold in the treated mice (0.2 vs 2.0%, p = 0.003). In control mice 36% of the anti-oxazolone hybridomas were cross-reactive with DNA compared with 6% in the treated mice (p = 0.03).
Serum response to immunization with foreign Ag. Mice were immunized with oxazolone as per Materials and Methods. IgG anti-oxazolone Abs were measured by ELISA 1 wk after immunization. Mice that received a high dose of Ad-CTLA4Ig and still had CTLA4Ig in their serum did not mount an Ab response to oxazolone. Mice treated with combination therapy mounted a high-titer anti-oxazolone response (p = 0.0036 vs prebleed serum) that was no different from untreated controls or from mice treated with single agent therapy. Prebleed sera were taken from both treated and untreated controls. Data are shown for the 1/500 serum dilution. Serial dilutions of a hyperimmunized control are shown in the right panel.
Frequency of anti-oxazolone-producing B cells. ELISPOT assays (mean ± SD) for frequency of anti-dsDNA Ab, anti-oxazolone Ab (each per 105 cells), and Ig-producing B cells (per 103 cells) in the spleens of combination-treated and age-matched control mice 3 days after boosting with oxazolone. Four mice were examined in each group. Significant p values: anti-dsDNA IgG, treated vs control, p = 0.029.
To further determine whether affinity maturation proceeded normally in the combination-treated mice, IgG cDNA libraries for the VH-Ox15 gene were generated from the spleens of four treated and four control mice by RT-PCR. We have previously shown that 30–40% of anti-oxazolone hybridomas in untreated NZB/NZW mice use this H chain gene (M. Mihara and A. Davidson, unpublished observation). Very little PCR product was recovered from cDNA of three unimmunized mice and insufficient numbers of clones were recovered for analysis (data not shown). More than 20 sequences from each immunized group were analyzed for mutation frequency. Analysis of the VH-oxazolone sequences revealed evidence for affinity maturation in both treated and control mice with high replacement:silent ratios in the CDRs (Table I⇑). These studies in sum show that treated mice are able to mount a mature IgG response to foreign Ag with concomitant regulation of cross-reactive autoreactive B cells.
Discussion
SLE is characterized by excessive activation of both B and T lymphocytes that results in the production of high-affinity pathogenic autoantibodies (1, 18). NZB/NZW F1 mice have defects at both early (9, 19) and late stages of their immune development. Polyclonal T cell-independent activation of IgM-producing B cells is found early in life, but disease onset does not occur until the age of 4–6 mo, when these naive autoreactive B cell precursors become further activated and their Ig genes undergo class switching and somatic mutation, resulting in the production of high titers of IgG autoantibodies (20, 21, 22). This later process is clearly T cell dependent and is associated with an abnormal accumulation of T cells with a memory phenotype (8).
We have previously shown that long-term therapy with CTLA4Ig has no effect on early polyclonal activation of B cells but that it prevents the later steps of T cell-dependent B cell activation that precedes nephritis onset. However, CTLA4Ig does not prevent accumulation of T cells with a memory phenotype, and disease onset occurs within 4–8 wk of cessation of treatment (8). Furthermore, continuous high-dose CTLA4Ig therapy is immunosuppressive, resulting in complete suppression of humoral immune responses to both alloantigen (23) and oxazolone.
In this study we show that short-term combination therapy resulted in a long delay in onset of disease with suppression of autoreactive B cells that was not due to global long-term immunosuppression. This long-term response to treatment was found to be due to modulation of both T and B cell function. Untreated NZB/NZW F1 mice have excessive T cell activation as evidenced by increased expression of the activation marker CD69 on CD4 cells, a progressive increase in CD4 T cells in the spleen, and accumulation of CD4 T cells bearing the memory phenotype (CD44+CD62L−) (8). The emergence of all of these abnormalities was substantially delayed by 2 wk of treatment with combination costimulatory blockade. However, because T cells can be activated by autoreactive B cells it is not clear whether this was due to a primary effect on T cells or whether the absence of excessive T cell activation was due to absence of autoreactive B cells (24).
There was a clear decrease in the proportion and absolute numbers of B cells in the spleens of treated mice, which was evident even 16–20 wk after treatment. This effect was not observed using long-term administration of either CTLA4Ig (8) or anti-CD40L (X. Wang, W. Huang, M. Mihara, J. Sinha, and A. Davidson, manuscript in preparation) as single agents. CD40L expression on T cells is up-regulated early in the course of murine SLE and may rescue B cells from apoptosis during early B cell activation. B cell expression of B7.2, which occurs early after BCR stimulation, may also rescue B cells from apoptosis by facilitating the T cell release of antiapoptotic cytokines such as IL-4 (25, 26). Therefore, CTLA4Ig and anti-CD40L may act synergistically to block antiapoptotic signals and facilitate apoptosis of B cells during early B cell activation. Despite the decrease in B cell numbers, combination therapy did not appear to have major functional effects on naive autoreactive B cells. However, treatment did clearly prevent proliferation of IgM anti-DNA Ab-producing B cells as well as their activation and class switching to IgG.
Regulation of IgG anti-DNA Abs appeared to occur by several mechanisms. Flow cytometry analysis revealed a decrease in generalized T cell activation as described above. Precursors of autoreactive B cells reside in the marginal zone and B1 B cell compartments (27, 28) and have few somatic mutations. The large number of unmutated VHBW-16 genes found in the cDNA libraries from treated mice could be accounted for by prevention of trafficking of autoreactive precursors to the germinal centers. Therefore, it is of interest that the number of spontaneous germinal centers in the spleen was markedly diminished in the treated mice compared with controls, as observed by immunohistochemistry (data not shown). Alternatively, this effect could be the result of a general decrease in the number of cell divisions secondary to decreased T cell help (29).
Even when somatic mutation occurred, there appeared to be an alteration of selection of B cells producing mutated Abs. It is known that B cells that acquire specificity for both an eliciting foreign Ag and self-Ag through somatic mutation arise frequently in normal mice but are usually eliminated by apoptosis and do not enter the memory compartment (30, 31). In NZB/NZW F1 mice peripheral regulation of autoreactive B cells is altered and they enter both the memory and long-lived plasma cell compartments (31, 32). Examination of the immune response to oxazolone showed that in control-aged NZB/NZW F1 mice many oxazolone/dsDNA double-reactive Abs were elicited by oxazolone immunization and could be rescued as hybridomas without LPS stimulation. In contrast, Ag-selected class-switched Abs to oxazolone could be effectively generated in the combination-treated mice with concomitant regulation of the IgG anti-DNA Ab response. Thus, combination costimulatory blockade appears to allow the mice to maintain this normal aspect of peripheral B cell self-tolerance.
An effect upon B cell selection could also be seen upon analysis of the mutated VHBW-16 genes in the IgG cDNA libraries of the treated mice. This gene is strongly associated with anti-dsDNA Ab activity and with glomerular binding activity in NZB/NZW F1 mice (16), but its expression is regulated in the peripheral B cell compartment of normal mice (17). It has been previously reported that selection against genes with CDR3 bearing arginine, a basic amino acid with affinity for positively charged DNA, is a major mechanism of regulation of the VHBW-16 gene in normal mice that have experimentally been induced to mount an anti-dsDNA Ab response. This selection occurs mainly in the naive B cell repertoire, as these differences could be seen even in unmutated IgM Abs (17). In contrast, the NZB/NZW F1 mice treated with combination costimulatory blockade had no differences in arginine usage in CDR3, suggesting that treatment had no effect on selection of the naive B cell repertoire. The main differences between treated and control groups were observed in the CDR2, where an accumulation of arginine residues was observed in control but not in treated mice. A similar absence of arginine mutations was observed in Ad-CTLA4Ig treated mice (8). Although it has not been formally shown for this gene that arginine mutations in CDR2 are associated with high-affinity DNA binding, these data are suggestive that selection against mutations to arginine occurred in the treated mice. Similar findings have been reported in bcl-2 transgenic mice immunized with the hapten p-azophenylarsonate in which dual anti-arsonate/anti-DNA hybridomas contained a much higher percentage of mutations to arginine than did anti-arsonate hybridomas from wild-type mice (31).
Because onset of proteinuria occurred earlier than expected (11), we administered a second 2-wk course of combination therapy at the time of onset of fixed proteinuria (>300 mg/dl). This treatment effectively induced remission of nephritis in >50% of previously treated mice (defined as proteinuria <30 mg/dl lasting >6 wk after treatment initiation) compared with 17% of previously untreated controls. The mechanism for this remission is not completely clear. Despite the disappearance of proteinuria within 2–3 wk, there was no immediate decrease in circulating anti-dsDNA Abs, nor was there a reversal of the activated T cell phenotype. Although there was a decrease in DNA-binding B cells as evidenced by a decrease in ELISPOTs and hybridomas compared with controls, we could not tell from our experiments whether this preceded or followed the second round of therapy. However, in three mice IgG anti-DNA Ab-producing B cells were virtually absent from the spleen 4–5 wk after therapy. In these mice, serum anti-DNA Abs did not decrease over this period, suggesting that the serum Abs were arising from B cells outside the spleen, most likely from long-lived plasma cells that do not require T cell help to survive (33). Class switching was affected, but examination of VHBW-16 libraries revealed a similar profile of sequences to that observed in control mice (data not shown). It is possible that the rapid response to a second course of treatment may have resulted from an alteration of other inflammatory cells or of effector functions in the target organs. We have recently shown similar results in a disease remission induction protocol in previously untreated mice in which a single dose of cyclophosphamide was added to the 2-wk combination therapy regimen. In these mice, renal remission was induced despite the continued presence of Ig and complementdeposition in the kidneys (34) (J. Sinha, L. Schiffer, X. Wang,W. Huang, A. Akkerman, K. Hiromatsu, and A. Davidson,manuscript in preparation). In sum, these results suggest that costimulatory blockade suppresses autoantibody formation and the initiation of disease when used in prevention studies but can also significantly suppress the effector response to immune complex deposition when used in remission induction studies. Understanding the effects of costimulatory blocking regimens in the various phases of autoimmune disease is essential to optimization of such treatments for use in humans with already established disease.
Acknowledgments
We thank Dr. Charles Hall for assistance with statistical analysis. We thank Drs. B. Diamond and H. Keiser for critical reading of the manuscript.
Footnotes
-
↵1 This work was supported by the Sklarow Trust and by a grant from the National Institutes of Health (AI47291).
-
↵2 Current address: Chugai Pharmaceuticals, Gotemba-shi, Shizuoka, Japan.
-
↵3 Address correspondence and reprint requests to Dr. Anne Davidson, Departments of Medicine and Microbiology and Immunology, Albert Einstein College of Medicine, 1300 Morris Park Avenue, U505, Bronx, NY 10461. E-mail address: davidson{at}aecom.yu.edu
-
↵4 Abbreviations used in this paper: SLE, systemic lupus erythematosus; CD40L, CD40 ligand; CDR, complementarity-determining region.
- Received October 18, 2001.
- Accepted December 13, 2001.
- Copyright © 2002 by The American Association of Immunologists
References
In this issue
