SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Objective

Autoantibody production in patients with systemic lupus erythematosus (SLE) is associated with abnormalities of B cell function and phenotype. Clinical responses to B cell depletion therapy (BCDT), based on rituximab, are encouraging. Therefore, we undertook this study to investigate the effect of BCDT on antibody profiles.

Methods

Serial sera from 16 patients with active, refractory SLE were assayed for antinucleosome antibodies, anti–double-stranded DNA (anti-dsDNA), anti–extractable nuclear antigen, anti–tetanus toxoid, and antibodies to pneumococcal capsular polysaccharide for at least 1 year following BCDT. Anti-dsDNA antibodies derived from the VH4.34 immunoglobulin germ line gene (9G4+) were also measured.

Results

All patients achieved peripheral B cell depletion and improved clinically for at least 3 months. Antinucleosome and anti-dsDNA antibodies decreased to a mean ± SD of 64 ± 37% and 38 ± 33% of baseline values, respectively, by 6–8 months post-BCDT. Levels of other autoantibodies and antimicrobial antibodies were generally unchanged. In the 9 of 16 patients who were still well at 1 year, anti-dsDNA antibodies fell to 42 ± 36% of baseline values at 6–8 months and to 37 ± 33% at 10–14 months. In patients who had disease flares within 1 year of BCDT, levels of these antibodies decreased to 60 ± 40% and 83 ± 93% of baseline values at 6–8 months and at 10–14 months, respectively. Circulating anti-dsDNA antibodies were positive for 9G4 expression in 4 of 6 patients tested, and flares in 2 of these patients were accompanied by rises in 9G4+ anti-dsDNA antibodies.

Conclusion

These observations suggest that B cell clones committed to producing antinucleosome and anti-dsDNA antibodies, including the VH4.34 subpopulation of anti-dsDNA antibodies, have a relatively rapid turnover compared with B cell clones producing other antibodies. There was also a trend toward a greater and more sustained decrease in anti-dsDNA antibodies in patients with clinical benefit lasting >1 year.

Results of recent clinical studies utilizing the B cell–depleting monoclonal antibody rituximab indicate that B cell–targeted therapy promises to be a major advance in the treatment of autoimmune disease (1). A key feature of the clinical response to B cell depletion therapy (BCDT) is that it can be sustained for up to 4 years following a single cycle of therapy (2). Although it has been suggested that sustained remission following BCDT may indicate that removal of certain autoreactive B cell clones can allow restoration of normal immune tolerance, our understanding is still in its infancy regarding the precise mechanism by which benefit is achieved as well as the impact on protective immune competence (2, 3).

In addition to being the parents of autoantibody-producing plasma cells, B cells may play fundamental roles in both promoting and regulating immune responses (4, 5). The capacity of the B cell receptor (BCR) (i.e., surface Ig) to bind antigen with high specificity and affinity enables the B cell to be an extremely efficient presenter of specific antigen to cognate T cells (6). In turn, soluble antibody is crucial to the regulation of B cell survival through interactions of antigen–antibody complexes with the BCR, complement, immunoglobulin Fc receptors, and Toll-like receptors (7). In addition to direct interactions with T cells, both human and murine studies point to a possible contribution by B cell–derived and other cytokines, notably BAFF, in modulating the immune response (8, 9).

The aberrant immune responses to DNA and DNA-related proteins in systemic lupus erythematosus (SLE) are thought to originate from a complex biologic process involving both T and B cell dysregulation. For example, this is reflected in the ability of CD4+ T cells to respond to subthreshold stimuli, resulting in the triggering of sustained overexpression of CD40 ligand (10). Contributing to this process are genetic and acquired abnormalities of the complement system and of cytokine production, which may fuel defective processing and presentation of apoptotic bodies resulting in autoantibody production (11).

The complex interplay between these factors, combined with the fine specificity and structural properties of the autoantibodies, is then thought to determine the expression and site of disease (12). Evidence from a number of approaches suggests that different subpopulations of autoantibodies play different roles in the development of tissue damage in SLE. Anti–double-stranded DNA (anti-dsDNA) antibodies (13) and, more recently, antinucleosome (14) and anti–α-actinin (15) antibodies have been shown to be linked to the development of tissue damage, especially nephritis. Other specificities such as antihistone and anti-Sm seem to be less closely linked to disease pathogenesis. Studying the changes in levels of different autoantibody populations following BCDT and their relationship to clinical response and disease flare may therefore give clues to the relative importance of specific B cell clones either for disease pathogenesis or for a more subtle role in disease “memory” or progression.

Selective therapeutic depletion of B cells became possible with the availability of the anti-CD20 antibody rituximab. Rituximab is currently given in combination with cyclophosphamide in SLE in most cases, because of inconsistent depletion with rituximab alone. Although cyclophosphamide may have a direct cytolytic effect on plasma cells as well as B cells, this combination provides the best current insight into the effects of depleting B cells on the pathobiology of the disease. We have thus examined the effect of BCDT on different populations of autoantibodies associated with SLE. We have also followed up the levels of anti-dsDNA bearing the idiotope recognized by the 9G4 monoclonal antibody (16). It has been suggested that anti-dsDNA antibodies bearing this idiotope are particularly likely to play a role in SLE (17). In normal individuals, circulating levels of VH4.34-derived immunoglobulins are low due to germinal center (GC)–associated censoring mechanisms (18). Elevated serum levels of VH4.34-derived immunoglobulins have been described in SLE, suggesting that these mechanisms are compromised (19).

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Patients.

Patients were selected on the basis of clinical need in that they had failed to respond adequately to standard immunosuppressive therapy. Sixteen patients with clinically active SLE received 2 doses of intravenous (IV) rituximab (each dose 500 or 1,000 mg) 2 weeks apart with or without 2 doses of IV cyclophosphamide (each dose 750 mg) together with a variable dose of prednisolone (total dose of oral or IV steroid 200–500 mg within 3 weeks; average 250 mg per patient). Therapeutic regimens, times to B cell return, and times to flare are shown in Table 1. These patients form part of a larger cohort treated with BCDT in whom clinical responses have previously been reported (20).

Table 1. Clinical characteristics of the patients*
Patient/age/sexRaceDisease duration, yearsOrgan involvementTreatment, mg/weekTime to B cell return, monthsTime to disease flare, months
  • *

    The treatment protocols for B cell depletion therapy (BCDT) for each patient included 2 doses of intravenous (IV) rituximab (Ritux; 500 or 1,000 mg) with or without 2 concurrent doses of IV cyclophosphamide (CYC; 750 mg) 2 weeks apart. All patients experienced peripheral blood B cell depletion following treatment. The times to peripheral blood B cell return and clinical disease flare are relative to the respective times of BCDT. J = joint; S = skin; CVS = cardiovascular system; R = kidney; CNS = central nervous system; NA = not available; H = hematologic.

2/35/FWhite11J, S, CVSRitux, 2× 500; CYC, 2× 7505>12
3/36/FWhite10J, CVSRitux, 2× 500; CYC, 2× 750>36>12
4/40/FWhite10J, S, CVS, RRitux, 2× 500; CYC, 2× 75067
6/21/FWhite9J, S, RRitux, 2× 500; CYC, 2× 7505>12
8/22/FWhite3J, CNS, RRitux, 2× 1,000; CYC, 2× 7506>12
9/27/MBlack10J, S, CVS, RRitux, 2× 1,000NA8
10/30/FWhite5J, S, RRitux, 2× 1,000; CYC, 2× 75044
11/18/FAsian6J, S, RRitux, 2× 1,000; CYC, 2× 7501111
12/38/FBlack8J, CVS, RRitux, 2× 1,0001010
13/26/FWhite6J, S, CNS, RRitux, 2× 1,000; CYC, 2× 7507>12
14/18/FWhite3J, HRitux, 2× 1,000; CYC, 2× 7508>12
15/18/FWhite4S, J, RRitux, 2× 1,000; CYC, 2× 750>12>12
16/23/MBlack7S, J, RRitux, 2× 1,000; CYC, 2× 7505>12
17/21/FWhite1S, RRitux, 2× 1,000; CYC, 2× 7504>12
19/30/FWhite12J, S, CVS, RRitux, 2× 1,000; CYC, 2× 75054
20/40/FWhite4J, S, CVS, RRitux, 2× 1,000; CYC, 2× 750411

Treatment efficacy was evaluated on the basis of improvement in clinical indices of active disease using the British Isles Lupus Assessment Group (BILAG) index, which distinguishes activity in 8 organs or systems (21). Activity in each system is graded from A (highest) to E (lowest). In this study a flare was defined as development of a BILAG grade A (major flare) or grade B (moderate flare) in any system in which the previous assessment was grade C, D, or E. The categories were combined, since both grades of flare required additional therapeutic intervention. Clinical evaluation was performed and peripheral blood total lymphocyte and B cell (CD19+) counts were measured prior to BCDT, at least twice during the first 6 months after BCDT, and, when possible, every 2–3 months afterward. Serum samples were collected and stored at −80°C until tested for levels of autoantibodies and antimicrobial antibodies. Changes in C3 were used as an objective biochemical measure of disease improvement and flare in serial studies (normal range 0.9–1.8 gm/liter). The study was approved by the local hospital ethics committee, and all patients gave informed consent.

Assessment of B cell depletion and B cell return.

The normal range for CD19+ B cells used by the local pathology laboratory was 0.03–0.40 × 109/liter. Levels less than 0.005 × 109/liter were defined as undetectable. Depletion of B cells in the peripheral blood was deemed to have occurred when CD19+ B cells were undetectable. In all patients, total peripheral blood B cell depletion was achieved for at least 3 months. B cell return was defined as the point at which B cells were again detectable in the peripheral blood (i.e., when the CD19+ cell count was ≥0.005 × 109/liter).

Serum levels of autoantibodies.

Enzyme-linked immunosorbent assays (ELISAs) were used to detect antibodies to dsDNA (phage), histones, SSA, SSB, Sm, and RNP/Sm using the EL-ANA profiles commercial system (TheraTest Laboratories, Lombard, IL) (22). Serial samples from individual patients were tested for each specificity at the same time. Values of greater than 40 IU/ml, 89 units/ml, 83 units/ml, 91 units/ml, and 76 units/ml were cutoff points for a positive result for anti-dsDNA, anti-Sm, anti-RNP/Sm, anti-SSA, and antihistone autoantibodies, respectively. Only 1 patient was positive for antibodies to SSB, and therefore the SSB specificity was not included in the results.

Preparation of nucleosomes from Jurkat cells.

Nucleosomes were prepared as described previously (23). Briefly, the Jurkat human T cell line was cultured and grown to confluence. Nuclei were extracted using a Dounce homogenizer and then digested with a micrococcal nuclease. Integrity of the nucleosome preparation was confirmed on an agarose gel. The nucleosome preparation was diluted 1:1,000 in phosphate buffered saline (equivalent to a concentration of 10 μg/ml dsDNA) and used as the coating antigen for the ELISA. The cutoff point for a positive result was defined as an optical density (OD) of 0.15 nm, calculated as the mean + 3SD of 30 samples from age- and sex-matched normal individuals.

Detection of autoantibodies bearing the 9G4 idiotope.

Expression of the 9G4 idiotope on anti-dsDNA antibodies was measured by capture ELISA using an adaptation of a previously described method (24). Briefly, sera were incubated on dsDNA-coated ELISA plates. Binding of the rat monoclonal antibody 9G4 to the bound autoantibodies was detected with biotinylated goat anti-rat IgG (Serotec UK, Oxford, UK) and a streptavidin–horseradish peroxidase visualization system. A positive result was attributed to sera giving a mean OD greater than the mean + 3SD of a pool of 10 normal human sera in each experiment.

Serum levels of antimicrobial antibodies.

Serum levels of IgG anti–tetanus toxoid (anti-TT) antibodies and antibodies to pneumococcal capsular polysaccharide (a combination of 23 common serotypes) were measured by ELISA (The Binding Site, Birmingham, UK). Levels >0.1 IU/ml for anti-TT antibodies were considered to be protective (25). Levels of antibodies to pneumococcal capsular polysaccharide >35 mg/liter are present in 90% of the general population (26).

Statistical analysis.

Paired t-tests were used for assessing differences in mean percentage binding to antigens. When the data were not normally distributed or the equal variance test was not satisfied, nonparametric testing was performed using the Mann-Whitney rank sum test and the Wilcoxon signed rank test for paired data using GraphPad software for AppleMac (GraphPad Software, San Diego, CA). Correlations were determined by Pearson's product-moment correlation for normally distributed data or by Spearman's rank order correlation for data that did not follow a normal distribution.

RESULTS

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Clinical responses and timing of B cell depletion and return.

Since the clinical characteristics of the majority of this cohort of patients have been previously reported (20), only brief details have been included here (Table 1). The median age of the 16 patients studied was 26.5 years (range 18–40 years), and disease duration ranged from 1 to 12 years (median 6.5 years). All except 3 patients had biopsy-proven renal involvement. Cyclophosphamide was used in combination with rituximab in 14 of 16 patients (patient 9 was allergic to cyclophosphamide and patient 12 refused). All patients achieved peripheral B cell depletion as defined and responded clinically for at least 3 months as confirmed by a decrease in global BILAG scores (20) (Table 1). Repopulation of the peripheral B cell compartment in most patients (13 of 16) had occurred within 1 year following treatment. There was no correlation between length of disease history pre-BCDT and either the time in months to B cell return or the time in months to flare (Pearson's r = 0.22 and r = −0.34, respectively; P > 0.05) (Table 1).

Effect of BCDT on circulating autoantibodies and antimicrobial antibodies.

The majority of patients had antibodies to dsDNA (14 of 16) and to nucleosomes (11 of 16), and all except 1 patient had at least 1 anti–extractable nuclear antigen (anti-ENA) antibody specificity. Changes in the levels of autoantibodies and antimicrobial antibodies from baseline (pre-BCDT) were determined 3 months post-BCDT (when all patients were B cell depleted) and 6–8 months after BCDT (when B cell repopulation was underway in most patients).

As shown in Figures 1A and B, decreases in levels of antinucleosome and anti-dsDNA antibodies approached statistical significance 3 months after BCDT and were statistically significant 6–8 months after BCDT (P = 0.005 for antinucleosome antibodies and P = 0.0005 for anti-dsDNA antibodies, by Wilcoxon signed rank test for paired data). The mean ± SD percentages of the baseline value 3 months and 6–8 months following BCDT were 71 ± 44% and 64 ± 37%, respectively, for antinucleosome antibodies and 53 ± 47% and 38 ± 33%, respectively, for anti-dsDNA antibodies. Mean percentage decreases from baseline values for both of these specificities were significant (P < 0.01 by paired t-test). Median levels of other autoantibodies and antimicrobial antibodies did not change significantly from baseline at either time point (Figures 1C and D and Figures 2A–C). The mean ± SD percentages of baseline values for antibodies to SSA and to histone showed small decreases both at 3 months (to 81 ± 39% and 89 ± 42%, respectively) and at 6–8 months (to 89 ± 44% and 89 ± 50%, respectively). In contrast, small mean ± SD percentage increases were seen for antibodies to Sm, RNP/Sm, TT, and pneumococcal capsular polysaccharide both 3 months post-BCDT (to 142 ± 90%, 103 ± 59%, 124 ± 75%, and 139 ± 79%, respectively) and 6–8 months post-BCDT (to 165 ± 71%, 136 ± 111%, 110 ± 88%, and 104 ± 64%, respectively).

thumbnail image

Figure 1. Effect of B cell depletion therapy (BCDT) on serum levels of anti–double-stranded DNA (anti-dsDNA) and antimicrobial antibodies. Sera from 16 patients with systemic lupus erythematosus were analyzed at baseline (prior to BCDT) and at 3 months and 6–8 months after BCDT. The numbers of patients with significantly elevated levels of autoantibodies at baseline are shown for each specificity. Levels of antinucleosome antibodies (A), anti-dsDNA antibodies (B), antibodies to pneumococcal capsular polysaccharide (anti-PCP) (C), and anti–tetanus toxoid (anti-TT) antibodies (D) are shown. Sera from 15 of 16 patients were used for testing levels of anti-PCP and anti-TT (serum from patient 20 [see Table 1] was not available for testing). The composite results are plotted as box plots. Lines inside boxes indicate the medians; outer borders of boxes indicate the 25th and 75th percentiles; bars extending from boxes indicate the range. P values shown are versus baseline, by Wilcoxon signed rank test for paired data. OD = optical density.

Download figure to PowerPoint

thumbnail image

Figure 2. Effect of B cell depletion therapy (BCDT) on serum levels of antihistone (A), anti-SSA (B), anti-RNP/Sm (anti-RNP) (C), and anti-Sm (D) antibodies. The numbers of patients with significantly elevated levels of these autoantibodies at baseline (prior to BCDT) are shown for each specificity. Levels of these antibodies are shown at baseline and at 3 months and 6–8 months after BCDT. In AC, the composite results are plotted as box plots. Lines inside boxes indicate the medians; outer borders of boxes indicate the 25th and 75th percentiles; bars extending from boxes indicate the range. P values shown are versus baseline, by Wilcoxon signed rank test for paired data. In D, individual values for 5 patients at each time point are shown.

Download figure to PowerPoint

Comparison of autoantibody levels in patients with and those without disease flare within 1 year of followup.

In order to investigate whether there were any particular features of the autoantibody profiles that would possibly be predictive of a more sustained response to BCDT, patients were divided into those who remained well at 1 year (n = 9) and those who experienced a flare within 1 year of BCDT (n = 7) (Table 1). Results for the different autoantibody specificities of individual patients (numbered as in Table 1) in each group of patients are shown in Figures 3 and 4.

thumbnail image

Figure 3. Association of disease flares with serum levels of antinucleosome, anti-dsDNA, and antihistone antibodies. Serial levels of these autoantibodies in individual patients (see Table 1) who experienced a disease flare within 1 year of BCDT are shown at baseline and at 6–8 months and 10–14 months after BCDT (A, C, and E). Levels of these autoantibodies in individual patients (see Table 1) who remained well 1 year after BCDT are also shown (B, D, and F). Significant differences from baseline (P < 0.05 by Wilcoxon signed rank test for paired data) are indicated. See Figure 1 for definitions.

Download figure to PowerPoint

thumbnail image

Figure 4. Association of disease flares with serum levels of anti–extractable nuclear antigen antibodies. Serial levels of anti-Sm, anti-RNP/Sm, and anti-SSA antibodies in individual patients (see Table 1) who experienced a disease flare within 1 year of B cell depletion therapy (BCDT) are shown at baseline and at 6–8 months and 10–14 months after BCDT (A, C, and E). Levels of these autoantibodies in individual patients (see Table 1) who remained well 1 year after BCDT are also shown (B, D, and F). The Wilcoxon signed rank test for paired data did not show any significant changes in levels from baseline (P > 0.05) to either 6–8 months or 10–14 months after BCDT.

Download figure to PowerPoint

There was a tendency for patients experiencing an early flare (at <1 year) to have more than one anti-ENA specificity, especially against SSA and histones. At equivalent time points, comparison between median levels of each autoantibody specificity in patients with an early flare (Figures 3A, C, and E; Figures 4C and E) and those in patients who had a clinical response for >1 year (Figures 3B, D, and F; Figures 4D and F) showed no significant difference (Mann-Whitney rank sum analysis). In the case of antibodies to Sm (Figures 4A and B), the numbers were too small for analysis. Interestingly, there was a significant (P < 0.05 by Wilcoxon signed rank test for paired data) drop from baseline in median levels of anti-dsDNA and antinucleosome antibodies 6–8 months post-BCDT, and also in anti-dsDNA antibodies 10–14 months post-BCDT, in the cohort of patients who remained well for at least 1 year after BCDT (Figures 3B and D). There was also a trend toward a greater decrease in anti-dsDNA antibodies in patients with the longer clinical response compared with those experiencing a disease flare within 1 year (mean ± SD percentage of baseline 55 ± 49% versus 97 ± 92%, respectively, at 3 months; 42 ± 36% versus 60 ± 40%, respectively, at 6–8 months; and 37 ± 33% versus 83 ± 93%, respectively, at 10–14 months). However, comparison of mean percentage baseline values between the 2 groups of patients at matching time points showed no statistically significant differences.

Serial studies of 9G4 expression on circulating anti-dsDNA antibodies.

Pre-BCDT, circulating anti-dsDNA antibodies were positive for 9G4 expression in 4 of 6 patients tested (patients 2, 12, 14, and 16) (Table 1). Results for anti-dsDNA and 9G4-expressing anti-dsDNA in each patient were expressed as a percentage of baseline values and are shown in Figure 5. Improvement in C3 levels followed an antiparallel course in conjunction with declines in anti-dsDNA antibody levels in all 4 patients. 9G4-expressing anti-dsDNA antibodies appeared to decline relatively more rapidly and to a greater extent than the anti-dsDNA population as a whole in 3 of these 4 patients. Two of these 4 patients remained well for at least 1 year post-BCDT despite B cell return 5 months after treatment (Figure 5, top panels). In these 2 patients, both total anti-dsDNA and the 9G4-expressing subpopulation of anti-dsDNA antibodies gradually decreased with time. In the other 2 patients (Figure 5, bottom panels), anti-dsDNA antibodies fell relatively rapidly to nadirs within 2–3 months. These patients experienced disease flares at 5 and 11 months, respectively, coinciding with B cell return in each case. Disease flare in each patient was accompanied or preceded by sharp rises in anti-dsDNA antibody levels and, in particular, the 9G4-expressing subpopulation.

thumbnail image

Figure 5. Serial studies of 4 patients with circulating anti–double stranded DNA (anti-dsDNA) antibodies expressing the 9G4 idiotope. Each panel represents 1 patient. Levels of anti-dsDNA antibodies and of the 9G4-expressing subpopulation of anti-dsDNA antibodies at time points after B cell depletion therapy (BCDT) are expressed as a percentage of baseline values. Also shown are levels of serum C3 (expressed in mg/liter) at baseline and at time points after BCDT. Asterisk indicates disease flare. Horizontal lines indicate period of peripheral B cell depletion (defined as levels of CD19+ B cells <0.005 × 109/liter).

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

In this study of 16 patients with SLE, all of whom achieved effective B cell depletion in the peripheral blood and had a positive clinical response for more than 3 months, there were selective effects on different autoantibody specificities. Antinucleosome and anti-dsDNA antibodies declined 3 months following BCDT, but the decline only reached statistical significance 6–8 months posttreatment. The decrease of anti-dsDNA and antinucleosome antibodies followed similar kinetics in individual patients (Figures 3A–D). Antibodies to ENA did not change significantly from baseline in the patient group as a whole or when patients were divided into those experiencing an early flare (<1 year post-BCDT) and those with sustained responses. In the patient group as a whole, antipneumococcal and anti-TT responses were maintained, as has been reported by others. The combination with cyclophosphamide has been used because rituximab alone in SLE does not reliably achieve profound B cell depletion, and our experience is that without profound depletion, clinical results are poor (20). Thus, the combination is the currently available practical means of BCDT for lupus. Nevertheless, cyclophosphamide alone is both cytotoxic and cytostatic for B cells and may also reduce antibody levels by a mechanism other than purely B cell depletion, by inducing death of plasma cells directly or by effects on other cells of the immune system. The magnitude of such an effect at these rather modest doses is unknown, but it cannot be discounted.

Although there was a tendency for patients with a disease flare within 1 year of treatment to have antibodies to more different ENA specificities (specifically to SSA and histones), the numbers were too small for meaningful statistical analysis. In order to determine whether the levels of any particular autoantibody specificity were predictive of, or associated with, disease flare within 1 year of BCDT, comparisons between the 2 groups were made at baseline and at each time point. There was no difference between the patient groups in the absolute levels for any of the antichromatin or anti-ENA autoantibody specificities. However, there appeared to be a trend toward larger relative decreases in anti-dsDNA antibodies at all 3 time points in the patients who still remained well at the end of the study.

In other studies of rituximab-based BCDT in patients with SLE, significant drops in anti-dsDNA antibodies have been reported, but these were not necessarily associated with a measurable clinical response (27). Looney et al reported an improvement in global disease activity scores following BCDT in the absence of a significant change in anti-dsDNA antibody levels at 3 months posttreatment in 11 of 17 patients in a phase I/II dose-escalation trial (28). However, examination of the data shows that in 4 of 8 patients in whom effective B cell depletion was achieved, significant drops in anti-dsDNA were found. Similarly, we have recently reported on 24 patients with active refractory lupus in whom significant clinical improvement was found in the 23 patients who experienced effective B cell depletion in the peripheral blood (20). Anti-dsDNA titers decreased significantly in the group as a whole at 6 months posttreatment (P = 0.002).

Although the numbers were small in the present study, sustained clinical response appeared to be associated with a trend toward a greater decrease in anti-dsDNA and antichromatin antibodies. The variability in the response of anti-dsDNA antibodies to BCDT most likely reflects the inherent heterogeneity of patients with SLE and, perhaps, the different levels of depletion achieved in solid tissues. We also cannot exclude the possibility that the changes in antibody levels were secondary to the natural history of the disease in individual patients over the period of study. In addition, the complexity of the antibody response in patients with lupus, together with the use of combination therapy such as this, further confounds precise interpretation in terms of cause and effect.

The results of this observational study, however, suggest that different B cell clones show different levels of susceptibility to BCDT. The antinucleosome and anti-dsDNA assays would be expected to detect overlapping populations of antibodies, and it is interesting that the clones producing these antibodies seem particularly sensitive to BCDT. There was also a suggestion that patients whose serum levels of these antibodies did not decrease as greatly were more likely to have early disease flares.

It is probable that the ability of BCDT to remove putative pathogenic B cell clones will depend on both quantitative and qualitative factors. The effectiveness of depletion will depend on the extent of B cell removal from solid lymphoid organs and inflamed tissues as well as from the circulation. In addition, the effect of BCDT on different antibody specificities will reflect the presence of daughter plasma cells with widely differing lifespans (29). Compromised complement function and the possession of the low-affinity Fcγ receptor IIIa (FcγRIIIa) genotype (involved in antibody-dependent cellular cytotoxicity) are both associated with poor levels of rituximab-mediated B cell removal in patients with SLE (30, 31) and contribute to the wide variation between patients in the effectiveness and kinetics of depletion. The apparent sensitivity of different B cell clones to BCDT may also reflect the microanatomic location of cells committed to different auto- and antimicrobial antibody specificities. In mice transgenic for human CD20 expression on B cells, splenic marginal-zone and GC B cells were shown to be more refractory to rituximab lysis due to the presence of intrinsic microenvironmental factors delivered by lymphoid and accessory cells, rather than to lack of tissue penetration, differences in innate resistance, or expression of surface CD20 (32).

Patients with lupus also have elevated levels of circulating plasma cell precursors, plasmablasts (IgD−,CD38++/+++,CD27++,CD19+,CD20−) which are capable of limited cycles of division before maturing into plasma cells (33). Effective treatment with rituximab has been shown to deplete these cells, presumably through the removal of their parent (CD19+,CD20+) B cells. The selective effects of BCDT on anti-dsDNA and antinucleosome antibodies, while leaving the anti-ENA responses intact, which we describe here, suggests that the distribution of anti-dsDNA– and antinucleosome-committed B cells and the longevity of their daughter plasma cells differ from that of B cells recognizing other antigens. This may reflect their generation through T cell–independent extrafollicular responses which usually generate short-lived plasma cells (34).

The relative insensitivity to BCDT of antibodies to nucleic acid–associated antigens and of antimicrobial antibodies seems to indicate that B cell depletion in solid tissues was not complete, either with parent B cells occupying a protected niche and/or with the presence of long-lived plasma cells. For example, in the patients with an earlier disease flare, a significant population of mature anti-dsDNA or nucleosome-committed plasma cells (CD20−) may have become established in niches in the bone marrow; alternatively, similarly committed memory B cells may have been protected by local survival factors. High levels of BAFF and inflammatory cytokines are a well-described feature in patients with SLE and may also influence B cell and plasma cell survival (34). The possible association of an expanded autoantibody repertoire as a consequence of epitope spreading in those patients with a more protracted response to BCDT also supports the possibility that their autoimmune response is more established in secondary lymphoid tissue.

As has previously been reported in patients with rheumatoid arthritis and SLE, the period of clinical benefit can often greatly exceed the period of B cell depletion (20, 35). The aim of using BCDT in SLE is to reset the immune system back to a state in which appropriate regulation of autoreactive B cells is reestablished. Although the identity of the “pathogenic subpopulation of B cells” is not yet known, changes in a potentially important autoreactive B cell population can be monitored. The presence of VH4.34 B cells in GCs and in the memory cell (CD27+) compartment in patients with SLE suggests a breakdown in GC censoring and mechanisms of tolerance (18). Normalization of the VH4.34 autoreactive B cell subset following BCDT has previously been demonstrated in patients with SLE (33). The conventional antigen binding site of VH4.34-derived antibodies can also be specific for a number of autoantigens, including dsDNA (36). Anti-dsDNA antibodies that are 9G4+ have been associated with increased disease activity and inflammatory sites in patients with SLE (19, 37).

We found 9G4+ anti-dsDNA antibodies in 4 of 6 patients tested. Following BCDT, circulating levels dropped in parallel with antibodies to dsDNA as a whole. Steep rises in VH4.34 anti-dsDNA antibodies were seen in the 2 patients who experienced a disease flare at B cell return. The association of VH4.34 anti-dsDNA antibodies with flare at repopulation therefore suggests incomplete depletion of VH4.34 memory B cells specific for dsDNA or the education of naive VH4.34-derived B cells to a dsDNA specificity. It was also of interest that the levels of VH4.34 anti-dsDNA antibodies fell relatively more slowly in patients with longer periods of remission and did not reemerge at B cell return. This may reflect their origin from longer-lived plasma cell populations rather than from overactive GC reactions and a possible return to “normal” of GC censoring mechanisms in patients with a more sustained response.

In conclusion, the efficacy of BCDT in patients with SLE may rely on qualitative features such as the specificity of autoantibodies as well as on the quantity of B cells removed. The success of BCDT in this small group of patients was most associated with reduction in anti-dsDNA and antinucleosome antibody levels, supporting their proposed role in pathogenic effector mechanisms, perhaps most importantly through generation of immune complexes. The results also suggest that these autoantibodies were derived from a rapidly turning over B cell population. The incomplete depletion of memory B cells and the preservation of long-lived plasma cells would explain the apparent insensitivity of anti-ENA specificities to BCDT. B cells with such specificities, and their antibody products, may both have additional roles as the sources of important signals necessary to perpetuate the underlying pathogenic processes.

REFERENCES

  1. Top of page
  2. Abstract
  3. PATIENTS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES