ClinicalTrials.gov identifiers: NCT00424476 and NCT00410384.
To assess the effects of the B lymphocyte stimulator (BLyS)–specific inhibitor belimumab on immunologic biomarkers, including B cell and T cell populations, and maintenance of antibody titers to prior vaccines in autoantibody-positive systemic lupus erythematosus (SLE) patients.
Pooled data from 2 phase III trials, the Study of Belimumab in Subjects with SLE 52-week (BLISS-52) and 76-week (BLISS-76) trials, comparing belimumab 1 mg/kg or 10 mg/kg versus placebo (plus standard SLE therapy for each group) were analyzed for changes in autoantibody, immunoglobulin, and complement levels. BLISS-76 patients were also analyzed for changes in B cell and T cell populations and effects on prior vaccine-induced antibody levels.
Belimumab-treated patients experienced significant sustained reductions in IgG and autoantibodies and improvement in C3/C4 levels, resulting in greater positive-to-negative conversion rates for IgG anti–double-stranded DNA (anti-dsDNA), anti-Sm, anticardiolipin, and anti–ribosomal P autoantibodies and normalization of hypergammaglobulinemia and low C3/C4 levels. Belimumab-treated patients experienced significant decreases in the numbers of naive and activated B cells, as well as plasma cells, whereas memory B cells and T cell populations did not decrease. Belimumab did not substantially affect preexisting antipneumococcal or anti–tetanus toxoid antibody levels. Post hoc analysis showed greater reductions in SLE disease activity and the risk of severe flares in patients treated with belimumab 10 mg/kg (P ≤ 0.01) who were anti-dsDNA positive and had low C3/C4 levels at baseline. Normalization of the C3 or anti-dsDNA level by 8 weeks, irrespective of therapy, was predictive of a reduced risk of severe flare over 52 weeks.
Belimumab appears to promote normalization of serologic activity and reduce BLyS-dependent B cell subsets in serologically and clinically active SLE. Greater serologic activity may predict a better treatment response to belimumab.
Systemic lupus erythematosus (SLE) is a heterogeneous autoimmune disease associated with considerable morbidity, increased mortality, and poor health-related quality of life (1, 2). Immunologic features of SLE include abnormal activation of B and T lymphocytes and elevated titers of autoantibodies (3, 4).
B lymphocyte stimulator (BLyS) is a 285–amino acid type II protein that is a member of the tumor necrosis factor ligand superfamily (5, 6). In vitro and in vivo studies have demonstrated BLyS to be a vital B cell survival factor (7–9), with important roles in the differentiation of immature to mature B cells (10, 11) and in immunoglobulin class switching and production (12, 13).
There is a strong connection between BLyS and SLE. In mice that otherwise are not prone to autoimmunity, constitutive overexpression of BLyS leads to SLE-like disease (14–16). In human SLE, circulating levels of BLyS are elevated in as many as 50% of patients (17, 18), and a large 2-year longitudinal study documented a significant correlation between BLyS levels and clinical disease activity (19).
Belimumab is a human IgG1λ monoclonal antibody that inhibits B cell survival and differentiation by neutralizing soluble BLyS, without directly causing B cell death (20, 21). Two international phase III trials, BLISS-52 and BLISS-76 (ClinicalTrials.gov identifiers NCT00424476 and NCT00410384, respectively), evaluated the safety and efficacy of belimumab in patients with autoantibody-positive (seropositive) SLE (defined as a serum antinuclear antibody [ANA] titer of ≥1:80 and/or positive results on a test for serum anti–double-stranded DNA [anti-dsDNA]) (22, 23). Both trials showed that belimumab 10 mg/kg plus standard therapy for SLE (including corticosteroids, immunosuppressive agents, and/or antimalarial agents administered alone or in combination) was generally well tolerated and met the primary end point of a significantly improved SLE Responder Index (SRI) at week 52 compared with standard therapy alone.
Several SLE studies have suggested that assessment of immunologic biomarkers, such as anti-dsDNA and complement, may be useful in predicting disease activity, the occurrence of flares, and the response to treatment (24–27). The present study focused on the effects of belimumab treatment in the BLISS-52 and BLISS-76 trials on biomarkers, including levels of Ig in serum, autoantibodies, complement components, and select B cell and T cell populations, biomarkers as predictors of treatment response, and the ability of patients treated with belimumab to maintain titers of antipneumococcal and anti–tetanus toxoid antibodies from vaccinations received prior to study entry.
PATIENTS AND METHODS
BLISS-52 (n = 865) and BLISS-76 (n = 819) were multicenter randomized, double-blind, placebo-controlled trials comparing belimumab 1 mg/kg or belimumab 10 mg/kg with placebo in patients with active SLE. In addition to the study drug, all patients received standard therapy for SLE (including corticosteroids, immunosuppressive agents, and/or antimalarial agents administered alone or in combination). The two trials had similar designs, which have been described in detail previously (22, 23). Briefly, all patients had a score of ≥6 on the Safety of Estrogens in Lupus Erythematosus National Assessment (SELENA) version of the Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) (28) at screening, were autoantibody-positive (ANA ≥1:80 or anti-dsDNA ≥30 IU/ml), and had received stable dosages of standard therapy for ≥30 days prior to the specific study. Patients received belimumab or placebo via intravenous infusion on days 0, 14, and 28, and then every 28 days through week 48 (BLISS-52) or through week 72 (BLISS-76), plus standard therapy throughout the entire studies. Patients had progressive restrictions on concurrent immunosuppressive and antimalarial medications as well as on prednisone during the trials. The primary end point in both trials was 52-week response rate as assessed by the SRI, which was defined as a ≥4-point improvement in the SELENA–SLEDAI score, with no new British Isles Lupus Assessment Group A organ domain score and no more than 1 new B organ domain score (29, 30), and no worsening (<0.3-point increase) in the physician's global assessment score (31).
Since the changes in biomarkers in the BLISS-52 and BLISS-76 trials through week 52 (conclusion of BLISS-52) were similar (data not shown) and each trial had the same basic clinical design and met its primary end point, the results were pooled to increase the power to detect the belimumab treatment effect and to present a more representative overall effect of biomarker changes observed in a subgroup of patients. Biomarkers were determined at baseline to evaluate the relationship between their presence and the overall treatment effect, as measured by the SRI and by SLE flares.
Details on the methods by which some of the biomarkers (i.e., IgG, IgM, IgA, C3, C4, anti-dsDNA, and ANA) were measured in BLISS-52 and BLISS-76 have been reported previously (22, 23). Other biomarkers analyzed in both studies included anti-Sm antibodies, anticardiolipin antibodies (aCL), and anti–ribosomal P autoantibodies, which were measured by enzyme-linked immunosorbent assays. All autoantibodies evaluated were for IgG subtypes, except for aCL, for which IgG, IgM, and IgA subtypes were measured. Changes in B cell and T cell subsets were analyzed only in BLISS-76 and are presented through week 76. B cell subsets consisted of CD20+ and activated (CD20+CD69+), memory (CD20+CD27+), and naive (CD20+CD27–) subsets, and plasma cell subsets consisted of plasmacytoid (CD20+CD138+), plasma (CD20–CD138+), short-lived plasma (CD20–CD27bright), and SLE (CD19+CD27brightCD38bright) subsets. (Bright indicates high levels of expression of the cell surface marker [32, 33].) T cell subsets consisted of CD3+CD4+ and CD3+CD8+ subsets.
Complement and anti-dsDNA levels were measured at baseline and every 4 weeks through week 52. All other autoantibodies were measured at baseline and week 52. In BLISS-76, immunoglobulin levels were measured at baseline, every 8 weeks through week 40, and week 52. In BLISS-52, IgG levels were measured at baseline and at weeks 8, 24, 40, and 52, and IgM and IgA levels were measured at baseline and at week 52. B cell and T cell subsets were assessed only in BLISS-76, being measured at baseline and at weeks 8, 24, 52, and 76.
A substudy of BLISS-76 assessed the effects of belimumab on vaccine-associated antibody levels in patients receiving pneumococcal or tetanus vaccine within 5 years of the study. Antibody levels were measured at baseline and week 52 using multianalyte immune detection (lower limit of quantitation 0.3 mg/ml) for the 7 type-specific pneumococcal polysaccharide antigens included in the 12-valent conjugate vaccines, i.e., 1, 3, 4, 8, 9(9N), 12(12F), 14, 19(19A), 23(23F), 26(26B), 51(7F), and 56(18C) (34). An enzyme-linked immunosorbent assay was used to measure anti–tetanus toxoid; based on the reported protective levels of 0.10 IU/ml (35), IgG anti–tetanus toxoid levels ≥0.50 IU/ml were conservatively considered to be protective on this platform.
The percentage change in biomarker levels were compared between the placebo group and each of the belimumab groups using Wilcoxon's rank sum test. The proportions of patients whose biomarker levels normalized or whose antibody levels were maintained were compared between the placebo group and each belimumab group using the likelihood ratio chi-square test. When >20% of the expected contingency table cell counts were <5, the proportion of patients whose antibody levels were maintained was analyzed in the placebo group versus each belimumab group using Fisher's exact test. Among patients with elevated BLyS values at baseline and those with low C3 or C4 levels and positive anti-dsDNA levels at baseline, the SRI response rate at week 52 was compared between the placebo group and each belimumab group using a logistic-regression model. The time to first severe SLE flare was compared between the placebo group and each belimumab group using a Cox proportional hazards model. Both analyses were adjusted for baseline randomization factors. The association of biomarker normalization with the SRI response and with SLE flare was assessed using the likelihood ratio test and the log rank test, respectively. The association analysis was performed in all study patients combined and in the belimumab-treated patients separately.
Effects on serologic biomarkers.
Baseline demographic features and biomarker values in the BLISS-52 and BLISS-76 cohorts are presented in Table 1.
Table 1. Baseline demographic features, disease characteristics, and biomarker levels in the BLISS-52 and BLISS-76 cohorts and baseline median B cell and T cell values in the BLISS-76 cohort, by treatment group*
BLISS-52 cohort (n = 865)
BLISS-76 cohort (n = 819)
Placebo (n = 287)
Belimumab 1 mg/kg (n = 288)
Belimumab 10 mg/kg (n = 290)
Placebo (n = 275)
Belimumab 1 mg/kg (n = 271)
Belimumab 10 mg/kg (n = 273)
All patients in the Study of Belimumab in Subjects with SLE 52-week (BLISS-52) and 76-week (BLISS-76) trials received standard therapy plus intravenous infusions of belimumab or placebo on days 0, 14, and 28, and then every 28 days through week 48 (BLISS-52) or through week 72 (BLISS-76), respectively. SLE = systemic lupus erythematosus; SELENA–SLEDAI = Safety of Estrogens in Lupus Erythematosus National Assessment version of the SLE Disease Activity Index; BILAG = British Isles Lupus Assessment Group; pred. = prednisone; ANA = antinuclear antibody; anti-dsDNA = anti–double-stranded DNA; aCL = anticardiolipin antibody; EU = enzyme-linked immunosorbent assay units; BLyS = B lymphocyte stimulator.
Sex, % female
Age, mean ± SD years
36.2 ± 11.8
35.0 ± 10.6
35.4 ± 10.8
40.0 ± 11.9
40.0 ± 11.4
40.5 ± 11.1
Duration of SLE, mean ± SD years
5.9 ± 6.2
5.0 ± 4.6
5.0 ± 5.1
7.4 ± 6.7
7.9 ± 7.1
7.2 ± 7.5
SELENA–SLEDAI score, mean ± SD
9.7 ± 3.6
9.6 ± 3.8
10.0 ± 3.9
9.8 ± 4.0
9.7 ± 3.7
9.5 ± 3.6
BILAG organ domain involvement, %
≥1 A or ≥2 B domain scores
≥1 A domain score
Glucocorticoid use, %
Pred. or equivalent >7.5 mg/day, %
Dosage, mean ± SD mg/day
11.9 ± 7.9
12.9 ± 8.6
13.2 ± 9.5
9.4 ± 8.9
8.7 ± 7.6
8.4 ± 7.9
Other immunosuppressant use, %
% with ≥30 IU/ml
Mean ± SD IU/ml
144.4 ± 63.7
146.2 ± 61.8
144.4 ± 62.0
150.8 ± 61.6
138.5 ± 63.4
143.2 ± 61.9
Anti-Sm ≥15 units/ml
Anti–ribosomal P >25 EU/ml
Low C3 (<900 mg/liter)
Low C4 (<160 mg/liter)
High IgG (>16.2 mg/ml)
Low IgA (<0.81 mg/ml)
Low IgM (<0.48 mg/ml)
BLyS, % >0.5 ng/ml
B cells and plasma cells, median counts
No. of patients evaluated
CD20+CD27+ (memory), ×103/ml
CD20+CD27− (naive), ×103/ml
CD20−CD27bright (short-lived plasma),/ml
T cells, median counts
No. of patients evaluated
By week 8, improvements in IgG levels were significantly greater with belimumab at either dose than with placebo and remained significantly greater through week 52 (median reduction 13.8% and 15.3% for 1 mg/kg and 10 mg/kg belimumab, respectively, versus 2.5% for placebo; P < 0.001) (Figure 1A and Table 2). In addition, belimumab reduced autoantibody levels, and significantly more patients treated with belimumab converted from seropositive to seronegative for anti-dsDNA (both doses), anti-Sm (10 mg/kg), anti–ribosomal P (10 mg/kg), and IgG aCL (both doses) autoantibodies by week 52 (Figure 1B and Table 2). Of note, both doses of belimumab resulted in significantly lower anti-dsDNA levels compared with placebo treatment as early as week 8 and at every measurement thereafter (median reduction at week 52 36.6% for 1 mg/kg and 40.8% for 10 mg/kg versus 10.2% for placebo; P < 0.001), and the median percentage reduction in IgG anti-dsDNA levels among anti-dsDNA–positive patients receiving belimumab was more than 2 times higher than the reduction in overall IgG among all patients receiving belimumab (Figure 1A and Table 2).
Table 2. Changes from baseline in levels of immunoglobulins, autoantibodies, and biomarker predictors of SRI and severe flare reduction (pooled data)*
Belimumab 1 mg/kg
Belimumab 10 mg/kg
Data from the Study of Belimumab in Subjects with SLE 52-week (BLISS-52) and 76-week (BLISS-76) trials were pooled and analyzed together. ANA = antinuclear antibody; aCL = anticardiolipin antibody.
P < 0.001 versus placebo.
Only patients positive at baseline are represented.
Although the values presented in the table are the median, the P values are for the entire distribution (mean ± SEM value of 39.02 ± 10.15, 19.21 ± 8.46, and 15.91 ± 9.21 for the placebo, belimumab 1 mg/kg, and belimumab 10 mg/kg groups, respectively).
P < 0.05 versus placebo.
P < 0.01 versus placebo.
Both anti–double-stranded DNA (anti-dsDNA) and low complement were removed from the Safety of Estrogens in Lupus Erythematosus National Assessment version of the Systemic Lupus Erythematosus Disease Activity Index score.
Baseline B lymphocyte stimulator (BLyS) levels were divided into quartiles from lowest to highest, and the Systemic Lupus Erythematosus Responder Index (SRI) response rate at week 52 for pooled data for each treatment group was calculated for each quartile. The number of patients in each quartile for each treatment group ranged from 125 to 155.
SRI response rate at week 52, by BLyS quartile at baseline, %††
<1.00 ng/ml of BLyS
1.00 to <1.35 ng/ml of BLyS
1.35 to <1.92 ng/ml of BLyS
≥1.92 ng/ml of BLyS
C3 and C4 levels increased in patients treated with belimumab. Among patients with low C3 or C4 levels at baseline, significant increases were observed with belimumab as early as week 4 and were maintained through week 52 (Figure 1C). Normalization of low C3 or C4 levels and hypergammaglobulinemia (IgG ≥16.2 gm/liter) at baseline occurred in significantly more patients with belimumab 1 mg/kg and 10 mg/kg versus placebo (Figure 1D). Of note, patients with low C3 or C4 levels at baseline had greater improvements in these levels with the higher belimumab dose. Grade 3 hypogammaglobulinemia (<400 ng/dl) was rare (no grade 4 reported) during therapy, with 1 patient each in the placebo and belimumab 10 mg/kg groups.
Overall, belimumab-treated patients experienced lower rates of worsening of biomarker values than did placebo-treated patients (standard therapy alone) (Table 2). Seroconversion to anti-dsDNA positivity was infrequent and occurred significantly more often in patients receiving placebo than in those receiving 10 mg/kg of belimumab (P = 0.02). Conversion from normal C3/C4 levels to low C3/C4 levels also occurred more often in patients treated with placebo than in those treated with belimumab 1 mg/kg (P not significant/P < 0.001) or 10 mg/kg (P < 0.001/P = 0.002). Among patients with normal levels of IgG at baseline, more patients treated with placebo developed IgG hypergammaglobulinemia than did those treated with either belimumab dose (P < 0.001).
Effects on lymphocytes.
In the BLISS-76 cohort, belimumab treatment significantly reduced median levels of CD19+ and CD20+ B cells through week 76 (55–58%; P < 0.001), while preserving memory B cell and T cell populations (Figures 2A and 3). Median reductions in naive (CD20+CD27–) B cells were significantly greater with either belimumab dose by week 8 and through week 76 than with placebo, while significant reductions in activated (CD20+CD69+) cells were observed with belimumab 10 mg/kg at week 52 and with both belimumab doses at week 76 (Figure 2A). Both belimumab doses significantly reduced the numbers of plasmacytoid cells (CD20+CD138+) by week 24, and these reductions were sustained through week 76. The effects of belimumab on CD20–CD138+, CD20–CD27bright, and CD19+CD27brightCD38bright plasma cell subsets appeared to be dose related, with significant reductions observed with the 10 mg/kg dose by week 8 and maintained through week 76 (Figure 2B). There was a significant increase in memory B cells (CD20+CD27+) with belimumab, which was not dose dependent, and peaked at week 8 and gradually returned toward baseline thereafter.
We observed no decreases from baseline in the median CD3+, CD3+CD4+, or CD3+CD8+ T cell counts between belimumab and placebo at week 52 or week 76 (Figure 3). There was a modest (10–15%), although not statistically significant, expansion of CD3+CD4+ and CD3+CD8+ cells between baseline and week 52 in patients receiving belimumab, but no differences were apparent in the ratio of CD4+ cells to CD8+ cells at week 52 across all treatment groups.
Findings of the vaccine substudy.
Effects on preexisting vaccine antigen–specific antibody levels were assessed in patients who had received pneumococcal or tetanus vaccine within 5 years of the start of treatment in BLISS-76. In all, 26 patients treated with placebo, 28 treated with belimumab 1 mg/kg, and 22 treated with belimumab 10 mg/kg were evaluated for IgG antipneumococcal antibody levels and 33, 33, and 25 patients, respectively, were evaluated for IgG anti–tetanus toxoid. Of the 6 serotypes reported by the CDC as causing the most frequent drug-resistant pneumococcal infection in the US (34), 5 correspond to serotypes tested in this study. At week 52, there were no significant differences across treatment groups in the percentages of patients maintaining IgG antipneumococcal antibody titers to these 5 serotypes (Table 3); of the 7 additional pneumococcal vaccine serotypes tested, significantly lower titers were noted only for serotype 12F (data not shown). Titers of IgG anti–tetanus toxoid antibodies were not significantly decreased (Table 3). These changes from baseline in the titers of IgG anti–tetanus toxoid antibody and IgG antipneumococcal antibody did not significantly alter the proportions of patients in whom detectable titers of antibodies for these vaccines were maintained (data not shown).
Table 3. Change in preexisting titers of antipneumococcal and anti–tetanus toxoid antibodies from baseline to week 52*
Belimumab 1 mg/kg
Belimumab 10 mg/kg
The percentage change in titers of IgG antivaccine antibodies was calculated among patients with available data who had been vaccinated prior to day 0 and were not revaccinated during the study. Values are the mean ± SEM (median) % change.
Pneumococcal serotypes 9N and 26B overlap with serotypes 9V and 6B, respectively, as described elsewhere (34).
Clinical correlations with biomarker status at baseline and biomarker changes during the study.
A post hoc analysis was performed of pooled data from the 2 trials to investigate whether the ultimate clinical response correlated with the BLyS level, the autoantibody profile, or the presence of low complement levels at baseline. We also assessed whether a reduction of serum autoantibodies to normal levels or a normalization of complement levels during the course of the trials was accompanied by clinical benefit.
BLyS was detected in 98% of patients at baseline (Table 1). There was no correlation between baseline BLyS levels (quartiles) and the SRI response at week 52, irrespective of therapy (Table 2). Among the 287, 284, and 305 patients receiving placebo, belimumab 1 mg/kg, and 10 mg/kg, respectively, who had low levels of C3 or C4 and were anti-dsDNA positive at baseline, SRI rates at week 52 were significantly higher with belimumab 1 mg/kg (41.5%; P = 0.002) and 10 mg/kg (51.5%; P < 0.001) than with placebo (31.7%). Furthermore, the risk of a severe flare (as assessed by the SELENA–SLEDAI Flare Index) was reduced with belimumab 1 mg/kg (20.4%; hazard ratio 0.67 [P = 0.02]) and belimumab 10 mg/kg (19.0%; hazard ratio 0.61 [P = 0.004]) versus placebo (29.6%). A modified SRI analysis that did not count normalization of anti-dsDNA or low complement in the calculation of a ≥4-point improvement in the SELENA–SLEDAI score demonstrated that the differential SRI response with belimumab versus placebo was maintained in patients who were anti-dsDNA positive and had low complement levels at baseline (Table 2), as well as in all patients (data not shown) irrespective of normalization of these biomarkers.
Biomarker changes while receiving study therapy were explored as potential predictors of improvement in the SRI or reduction in the severe flares. Based on the modified SRI, evaluating all patients (irrespective of therapy), those with normalization of C4 levels had significantly greater response rates than did those without normalization at week 4 (52% versus 41%; P = 0.02), week 28 (54% versus 42%; P < 0.01), and thereafter (except for weeks 36 and 44). Among only the belimumab-treated patients, the modified SRI response rates were significantly greater in those with normalization of C4 levels than in those without normalization at week 4 (56% versus 44%; P = 0.02), week 16 (54% versus 45%; P < 0.05), and week 28 (59% versus 43%; P < 0.001), and thereafter (except for week 44). Patients with normalization of IgG levels, irrespective of treatment, had significantly greater SRI response rates than did patients without normalization at week 24 (65% versus 50%; P < 0.001), week 40 (69% versus 54%; P < 0.001), and week 52 (66% versus 57%; P = 0.049). Belimumab-treated patients with normalization of IgG levels had significantly greater SRI response rates than did patients without normalization at week 24 (69% versus 53%; P = 0.002) and week 40 (71% versus 56%; P = 0.002). Risk of severe flare was significantly lower in all patients (irrespective of therapy) who experienced early normalization of C3 levels at week 4 (19% versus 27% in the non-normalized group; P = 0.04) and week 8 (17% versus 27%; P = 0.02) and early normalization of anti-dsDNA levels at week 8 (11% versus 22%; P = 0.04). In patients treated with belimumab, the risk of severe flare was numerically lower with early normalization of C3 levels at week 8 (16% versus 24%; P = 0.08) and of anti-dsDNA levels at week 8 (10% versus 19%; P = 0.12).
Changes over 52 weeks in B cell and plasma cell subsets (measured only in BLISS-76) were evaluated by quartile or absolute minimum percentage change from baseline to determine if reductions in B cells or plasma cells during therapy correlated with the SRI response or with a reduction in the rate of severe SLE flares. Only the naive B cell subset, irrespective of therapy, showed a consistent association of a greater percentage reduction from baseline with a greater likelihood of an SRI response at week 52 (in all patients, 60.2% versus 46.0% SRI response rate with a >70% reduction [n = 171] versus a ≤70% reduction [n = 361]; P = 0.002) and lower risk of severe flare over 52 weeks (in all patients, 7.6% versus 14.7% risk with a >70% reduction versus a ≤70% reduction; P = 0.02).
Belimumab treatment combined with standard therapy (including corticosteroids, immunosuppressive agents, and/or antimalarial agents administered alone or in combination) led to significant sustained reductions in autoantibody levels and increases in C3 and C4 levels over 52 weeks as compared with placebo plus standard therapy in patients with autoantibody-positive SLE. Significantly more patients treated with belimumab had normalization of their high IgG levels, conversion from low to normal/high C3 and C4 levels, and conversion from positive to negative serologic status for anti-dsDNA antibodies. In addition, treatment with belimumab 10 mg/kg resulted in significantly greater rates of conversion to seronegative status for anti-Sm, IgG aCL, and anti–ribosomal P autoantibodies.
Treatment with belimumab 1 mg/kg or 10 mg/kg was associated with significant reductions in the numbers of CD20+ B cells and in multiple B cell and plasma cell subsets, including naive and activated B cells, as well as in CD20+CD138+ plasma cell precursors (plasmablasts), while preserving the memory B cell subset and T cell populations. Increases in plasma cell subsets have been associated with increased SLE disease activity (32). It is therefore noteworthy that in the present study, plasma cell subsets decreased in a dose-dependent manner with belimumab treatment and were significantly different in those receiving 10 mg/kg as compared with standard therapy alone. These collective results are consistent with the results of the phase II clinical trial of belimumab (33, 36) and are consistent with the mechanism of action of belimumab as an antibody that specifically targets and inhibits the activity of soluble BLyS.
Consistent with the preservation of memory B cells and T cells, belimumab treatment did not result in a significant reduction in the preexisting titers of antibodies to pneumococcal and tetanus toxoid vaccines that had been administered prior to the start of the study. Preservation of memory B cells appears to be due to their lack of dependence on BLyS for survival (37). The initial expansion of memory B cells in the blood (peaking at week 8 and then gradually returning to baseline) may be secondary to their release from disrupted germinal centers where these B cells reside, may be due to inhibition of their return to these lymphoid tissues, or may be consequent to promotion of differentiation of naive B cells to memory B cells (21, 36, 38). In any case, similar percentages of patients taking belimumab and placebo maintained specific antivaccine antibody titers at week 52, including antibodies against 5 of the 6 serotypes (i.e., 6B, 9V, 14, 19F, and 23F; the 6B and 9V serotypes overlap with 26B and 9N reported here) that most frequently cause invasive drug-resistant pneumococcal infection in the US (34). In the BLISS studies, there were similar rates of serious and/or severe infections in the belimumab-treated groups as compared with the placebo-treated group, who received standard therapy alone (22, 23). Taken together, these results suggest that belimumab treatment does not compromise the immune response to infection.
The post hoc analyses pointed to likely meaningful correlations between biomarker changes and the SRI response and/or the incidence of severe SLE flares. Normalization of C4 levels as early as treatment week 4 and at other times was correlated with an increased SRI response in all patients, as well as in those treated with belimumab, regardless of whether normalization of low complement levels or anti-dsDNA antibody levels was counted in the calculation of the improvement in SELENA–SLEDAI scores. In addition, analysis of biomarker changes in all patients, as well as in those treated with belimumab, demonstrated that early normalization of C3 or anti-dsDNA values was a predictor of a reduced risk of severe flares. These results are consistent with those of previous studies that suggested a correlation of anti-dsDNA positivity and low complement levels on the one hand with increased disease activity and risk for flares on the other (24, 27). In addition, SLE patients with >70% reductions in naive B cell populations, irrespective of therapy, had a greater likelihood of achieving an SRI response or having a lower incidence of severe flares. The naive B cell subset had the greatest percentage reductions with belimumab therapy of all B cell and plasma cell subsets evaluated in BLISS-76.
In a phase II trial of belimumab in patients with SLE, belimumab was not associated with a significant benefit in improving SELENA–SLEDAI scores as compared with placebo (36). Exploratory analyses in that study showed that belimumab was associated with a significant benefit in improving SELENA–SLEDAI scores in patients with serologic activity (ANA ≥1:80 or anti-dsDNA >30 IU/ml), providing the rationale for phase III testing in, and for enrichment of the study population for, patients with serologically active SLE. A post hoc analysis in the BLISS studies showed significantly higher SRI rates and reduced risk of severe SLE flares in belimumab-treated patients who had low C3 or C4 levels and anti-dsDNA–positive status at baseline. This is consistent with the correlation of higher serologic activity with higher BLyS levels (>98% of patients had detectable BLyS in the two BLISS studies), which may permit autoantibody-producing cells to survive and to continue producing autoantibodies. Furthermore, the median percentage reduction in anti-dsDNA IgG among anti-dsDNA–positive patients was >2-fold greater than the percentage reduction in overall IgG among all patients treated with belimumab, which suggests that the effects of belimumab on autoantibody-producing cells may have been disproportionately greater than the effects on non–autoantibody-producing cells.
Higher BLyS levels at baseline did not correlate with a greater SRI response. Although higher serologic activity was associated with a better response to belimumab, the SRI response was not driven by normalization of anti-dsDNA or low complement, as the magnitude of the treatment effect was similar in patients with and without normalization of these serologic parameters. These findings suggest that the efficacy of belimumab is greatest in SLE patients with high levels of disease activity and serologic activity.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Stohl had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Pineda, Migone, Zhong, Freimuth.
Analysis and interpretation of data. Stohl, Hiepe, Latinis, Scheinberg, Clarke, Aranow, Abud-Mendoza, Hough, Pineda, Migone, Zhong, Freimuth, Chatham.
Medical monitor. Freimuth.
ROLE OF THE STUDY SPONSOR
Human Genome Sciences, Inc. was involved in the conception, design, implementation, and supervision of the BLISS-52 and BLISS-76 studies, the data analysis and interpretation, the statistical analysis, and the manuscript drafting, revision, and approval. GlaxoSmithKline was involved in the design of the BLISS-52 and BLISS-76 studies and the manuscript drafting, revision, and approval. Human Genome Sciences, Inc. and GlaxoSmithKline agreed to submit the manuscript for publication and approved the content of the manuscript. Editorial support was provided by Matt Stenger, Eleanore Gross, and Geoff Marx (BioScience Communications, New York, NY). Publication of this article was contingent upon the approval of all of the authors as well as the approval of Human Genome Sciences, Inc. and GlaxoSmithKline.