Introduction
Systemic lupus erythematosus (SLE) has a heterogeneous clinical and immunologic presentation, with the commonality of B cell hyperactivity across the patient population. B cell-activating factor (BAFF, also known as B lymphocyte stimulator [BLyS]) is crucial to the survival and differentiation of B cells [
1‐
5]. In animal models of SLE, BAFF overexpression promotes features of SLE [
6‐
10] and BAFF antagonists attenuate disease progression [
8,
11‐
13]. In humans with SLE, circulating BAFF levels are increased in as many as 50 % of patients [
14‐
16] and numerous studies have found a positive correlation between BAFF expression and disease activity [
17‐
21].
Given its likely contributory role to the maintenance of SLE disease activity, BAFF is an appealing target in the treatment of SLE. Two phase 3 clinical trials of the antisoluble BAFF monoclonal antibody belimumab added to standard-of-care treatment demonstrated significant effects on disease activity [
22,
23]. The “peptibody” blisibimod (A-623, AMG 623) is also a BAFF antagonist, but unlike belimumab, it binds to both membrane-bound and soluble BAFF. Blisibimod inhibits the interaction of BAFF with any of its three receptors (half-maximal inhibitory concentration [IC
50] = 0.2 nM at 1 nM BAFF), inhibits the in vitro BAFF-driven increases in B cell proliferation (IC
50 = 0.1 nM at 0.2 nM BAFF), reduces in vivo peripheral B cell numbers in non-autoimmune-prone mice, and delays the onset of proteinuria in SLE-prone (NZBxNZW)F1 mice [
24]. Neutralization of both soluble and membrane-expressed BAFF by blisibimod may produce an immunologic change different from that upon neutralization of soluble BAFF alone.
We present the safety, tolerability, and pharmacokinetic and pharmacodynamic profiles of blisibimod in subjects with SLE in two multicenter, double-blind, placebo-controlled trials: a single-dose phase 1a first-in-human study and a multiple-dose phase 1b study.
Methods
General
The studies were conducted at 12 sites in the United States. The protocols were approved by each center’s institutional review board (IRB) (please see the Acknowledgements), and all subjects provided written informed consent. The studies were developed and sponsored by Amgen, Inc., and blisibimod is now licensed to Anthera Pharmaceuticals, Inc. (Hayward, CA, USA).
Subjects
Men and women 18–65 years of age with mild disease that was either stable or inactive (as assessed by the investigator) were eligible for enrollment if they had a diagnosis of SLE for at least 1 year, and met the updated American College of Rheumatology classification criteria [
25,
26]. Subjects were excluded if they did not have a current or previous positive antinuclear antibody test; had active vasculitis, central nervous system lupus, acute renal disease, a glomerular filtration rate < 60 mL/min, uncontrolled hypertension or diabetes, or active infection within the previous 30 days; or had antibodies to human immunodeficiency virus, hepatitis B surface antigen, or hepatitis C. Those who were pregnant or nursing were also excluded, as were those who had either received monoclonal antibody therapy as an investigational drug within the previous year or participated in other investigational drug or device trials within 30 days. Subjects who had received cyclosporine, tacrolimus, sirolimus, cyclophosphamide, other alkylating agents, or the equivalent of > 100 mg/day of prednisone (pulse therapy) within 60 days or rituximab within 12 months were also not eligible. However, subjects were allowed to be treated with other SLE medications, including ≤ 10 mg/day of prednisone or its equivalent. Other than the proscribed medications above, investigators were permitted to prescribe concomitant medications or treatments as clinically warranted. Subjects had to be willing to practice effective methods of contraception.
Study design
In the first-in-human phase 1a ascending single-dose study, subjects were sequentially enrolled into one of seven dose cohorts: 0.1, 0.3, 1.0, or 3.0 mg/kg subcutaneous (SC) or 1.0, 3.0, or 6.0 mg/kg intravenous (IV) blisibimod. The 0.1 mg/kg dose was chosen as the starting dose on the basis of multiple-dosing preclinical toxicology findings in cynomolgus monkeys; the remaining doses were based on preclinical pharmacokinetic modeling and toxicology studies that showed no adverse effects (AEs) at weekly doses up to 150 mg/kg IV or SC [data on file]. Within each dose cohort, subjects were randomized (3:1 ratio) to receive blisibimod or placebo, and were followed for 28 (0.1 and 0.3 mg/kg) or 42 days (1.0, 3.0, and 6.0 mg/kg) post-dose, and those in the highest dose cohort (6.0 mg/kg) completed a 4-week extension follow-up visit after their end-of-study visit. In the phase 1b ascending multiple-dose study, subjects were sequentially enrolled into one of four dose cohorts: 0.3, 1.0, or 3.0 mg/kg SC or 6.0 mg/kg IV. Within each dose cohort, subjects were randomized (4:1 ratio) to receive four weekly doses of blisibimod or placebo, and the pharmacokinetics of blisibimod were followed to day 63 (0.3 and 1.0 mg/kg) or day 112 (3.0 and 6.0 mg/kg). For both studies, blisibimod was given as a 30 mg/mL formulation in vehicle. Those assigned to the placebo group received an equivalent volume of the corresponding vehicle.
End points
The primary end point of both studies was the characterization of the safety and tolerability of blisibimod. Safety and tolerability assessments included the incidence of treatment-emergent AEs, and clinically significant changes in vital signs, electrocardiograms (ECGs), and clinical laboratory safety tests. AEs were classified using Medical Dictionary for Regulatory Activities (MedDRA) terminology, and clinical laboratory results were assessed using the grading criteria of the National Institute of Health Common Toxicity Criteria for Adverse Events (NIH CTCAE). Secondary end points were the pharmacokinetic and pharmacodynamic profiles of blisibimod following single or multiple SC or IV doses. Exploratory end points included changes in peripheral blood B cell counts and expression of B cell surface markers.
Pharmacokinetic analysis
Serum concentrations of blisibimod were assessed at each visit beginning on day 1 (the day before dosing) and continuing through the end of the studies, with extensive serum collections after the first and fourth doses (phase 1b study only). Pharmacokinetic parameters were derived for each subject from the blisibimod serum concentrations: these parameters included area under the plasma-concentration curve from time of treatment to the last measurable concentration (AUC0-t), maximum observed serum concentration (Cmax), estimated initial concentration after IV bolus dosing (C0), time to Cmax (tmax), half-life (t1/2,z), clearance (CL), apparent clearance after SC dosing (CL/F), and accumulation ratio (AR). Blisibimod concentrations over time, as well as pharmacokinetic parameter values were summarized by route and dose.
Immunogenicity analysis
The presence of neutralizing serologic reactivity to blisibimod was analyzed using two validated assays at screening in both studies, at the end of the phase 1a study, and on day 29 in the phase 1b study. The first was a Biacore immunoassay to establish the presence of binding antibodies, and the second was an in vitro cell-based bioassay to detect neutralizing or inhibitory effects toward blisibimod in pre- and post-dose samples. All samples that tested positive in the immunoassay were tested in the bioassay along with their corresponding pre-dose samples, even if the pre-dose samples did not test positive in the immunoassay. If a sample was positive in both the Biacore assay and the cell-based assay, the subject was defined as positive for neutralizing antibodies.
Exploratory B cell analysis
In the phase 1a study, whole blood specimens were collected at screening visit day 1, pre-dose on the day of dosing, post-dose on days 14 and 28 for all dose groups, day 42 for the 3.0 and 6.0 mg/kg dose groups, and at the extension/follow-up for all dose groups. In the phase 1b study, whole blood specimens were collected twice before study inclusion (at screening day 28, re-screen); pre-dose on day 1; post-dose on days 15, 29, 50, 64, 78, 110, 134, 164, 190, 218; and early termination (when applicable). Whole blood specimens were shipped at ambient temperature to ICON Central Laboratories within the previously established 48-hour stability time frame for standard hematology assessment and flow cytometry determinations [data on file]. A standard clinical hematology analyzer was used to enumerate total lymphocyte counts. Detailed methods are provided elsewhere [
27].
Antibody combinations were chosen to permit enumeration of total B cells (CD19
+CD20
+), naïve B cells (CD19
+sIgD
+CD27
-), IgD
+ memory B cells (CD19
+sIgD
+CD27
+), IgD
- memory B cells (CD19
+sIgD
-CD27
+), total T cells (CD3
+CD16/CD56
-), and natural killer cells (CD3
-CD16/CD56
+). Plasmablasts were previously assessed and determined not to be sufficiently stable for assessment in the phase 1 studies [
27]; although they were analyzed in the phase 1a study, the data were not used because of cell instability. The antibody configurations and clone numbers were as follows: Tube 1: CD20 FITC (L27), CD45 APC (2D1), CD19 PerCP (SJ25C1); Tube 2: CD3 FITC (SK7), CD16 PE (B73.1), CD56 PE (MY31), CD45 APC (2D1); Tube 3: CD27 FITC (L128), CD38 PE (HB7), CD19 PerCP (SJ25C1), CD45 APC (2D1); Tube 4: CD27 FITC (L128), IgD PE (IA6-2), CD19 PerCP (SJ25C1), CD45 APC (2D1), Tube 5: CD138 FITC (MI15), CD38 PE (HB7), CD19 PerCP (SJ25C1), CD45 APC (2D1). The entire panel was performed in the phase 1a study; in the phase 1b study, the routine panel (tubes 1–3) was performed at screening visit day 28, re-screen, and days 64, 110, 134, 164, and 190. The extended panel (all tubes) was performed on phase 1b specimens drawn on days 1, 15, 29, 50, 78, 218, and early termination. Monitoring was not required if the change in B cell level was ≤ 20 %, or levels had returned to either within 20 % of baseline or a level from a reference range of healthy volunteers, whichever was lower; subjects could be released from monitoring as early as day 190. Those that did not meet these criteria were asked to return monthly until the criteria were met or they were released by medical board review.
Statistical analyses
Demographic, safety, pharmacokinetic, and B cell data for each route and dose were analyzed with descriptive statistics. After review, data were combined for the placebo groups by any route. The primary safety analysis was performed for all subjects who received an investigational product (blisibimod or placebo).
The relationship between blisibimod dose and changes in lymphocyte phenotypes was explored graphically using a cubic spline [
28] to visualize potential trends over time by dose and route. After reviewing these graphs, exploratory modeling of the relationship between dose/route and select lymphocyte phenotypes was analyzed using a mixed-effects model, with baseline value and actual treatment as fixed effects, and subject as a random effect. Based on graphical review, post-dose times were grouped as dosing (days 15–42), post-dosing (days 43–100), and extended follow-up (> day 100) phases, with exploratory modeling performed separately for these three post-dose time ranges. Because they were similar, blisibimod 1.0 mg/kg and 3.0 mg/kg dose groups were pooled and compared with the placebo group for each post-dose phase using an F test (
P values reported for these comparisons should be considered descriptive). Reference ranges of the mean, standard error, and coefficient of reliability of healthy volunteer and SLE subsets were established in a previously published study [
27].
Discussion
In the present studies, treatment with blisibimod produced significant changes in B cell subpopulations and B cell surface molecule density were observed, with no notable changes in T cells or natural killer cells. There was an initial transient increase of the B cell pool in subjects treated with single or multiple doses of blisibimod, but this was not statistically significant and was similar to previous findings in SLE patients treated with the BAFF antagonists atacicept [
29,
30] and belimumab [
31,
32]. Blisibimod showed consistency in half-life and linearity across dose ranges for both IV and SC routes of administration, with half-life ranging from 7 to 12 days for the 0.3 to 6.0 mg/kg IV and SC dose groups, and clearance estimates constant over the 1.0 to 6.0 mg/kg IV and SC doses ranging from 34 to 79 mL/hr. The safety and tolerability of single or multiple SC or IV doses of blisibimod were comparable with those associated with placebo administration, and AE and treatment-related events did not increase with increasing doses of blisibimod.
The main finding of these studies was the statistically significant decrease in B cell counts (Table
5, > day 100,
P = 0.10), although the reduction was not as large as those seen during treatment with the lytic agent rituximab [
33]. The constituency of the B cell pool also changed following treatment with blisibimod: there was a decrease in the number of naïve B cells (24–76 % reduction, day 43–100) and a relative increase in the memory B cell compartment, with the greatest effect on IgD
-CD27
+ (53–154 % increase, day 15–42). One possible mechanism for these results may be related to differences in the expression profiles of BAFF receptors on various B cell subsets, as BAFF is known to differentially support the stages of B cell development [
1‐
5]. The relative increase in the memory B cell compartment was not associated with safety signals or elevations of immunoglobulin or anti-blisibimod antibody levels. In fact, when comparing these data with those of healthy volunteers from our previous study [
27], blisibimod seemed to correct the B cell compartment to be more like that of healthy subjects, suggesting that if normalization of these cell types can be associated with clinical improvement, these may be a treatment goal in future studies.
Markers of the metabolic activity of targeted cells were also examined, as BAFF deprivation might suppress excess metabolic activity of those cells that were not depleted. The fluorescence intensity of CD19 was measured because CD19 is part of the signaling complex that modulates the threshold of activation of the B cell receptor, and it is known to exhibit genetic variations that correlate to expression levels and autoimmunity [
34]. The changes in CD19 density were variable, with increases of 119–154 % for all doses except 6.0 mg/kg IV, which produced a 43 % reduction > day 100 (Table
5, CD19 MFI day 43–100 and day > 100,
P = 0.04 and
P < 0.01, respectively). However, interpretation of these results is difficult because the CD19 MFI values at baseline were statistically significantly different between the blisibimod- and placebo-treated groups (Table
5; CD19 on B cells MFI, pre-dose,
P < 0.01). The variability in this dataset may be due to inconsistencies in the laboratory methods that were used; however, as the coefficient of reliability from our validation exercises all exceeded 0.8, it is more likely attributable to biologic variability in our subject pool. Nonetheless, CD19 MFI was increased during administration of blisibimod, raising the possibility that blisibimod promoted a reduction in B cell activation that reversed the down-modulation of CD19 from the B cell surface in SLE patients [
35]. Reduction of CD38 was also seen in blisibimod-treated subjects, occurring as early as day 15–42 of treatment (Table
5; CD38 on B cells MFI, day 15–42,
P < 0.01). This brought the density of this marker to within the range of healthy volunteers (362–384 vs. 295 CD38 on B cells MFI).
Other agents that reduce the number of circulating B cells have been tested in SLE: the anti-CD20 monoclonal antibody rituximab, a targeted B cell depletor, has been used in the clinical setting in patients with SLE despite not being approved for use in this population [
36,
37]. The anti-BAFF monoclonal antibody belimumab in two phase 3 trials in SLE demonstrated improvements in clinical measurements and biomarkers of disease activity [
22,
23], and is approved for the treatment of SLE in the United States. The BAFF- and a proliferation-inducing ligand (APRIL)-neutralizing atacicept also reduces B cell counts in patients with SLE, underscoring the versatility of options in modulating this pathway [
29,
38]. The success of belimumab and the absence of efficacy above standard of care from the rituximab clinical trials may be multifactorial and may dependent on trial design, but these results could suggest that efficacy in treating SLE may not solely be a function of reductions in B cell numbers. Yet, the positive results seen with belimumab confirm the possibility of therapeutic potential of BAFF antagonists in SLE.
A key question is whether agents that antagonize both cell-bound and circulating BAFF (e.g., blisibimod or tabulumab) will have a different spectrum of efficacy than those that bind only to soluble BAFF (e.g., belimumab) [
39]. In a 24-week, phase 2b, placebo-controlled study of blisibimod in 547 patients with SLE, treatment with blisibimod decreased B cell counts, reduced disease activity, and prolonged time to first severe disease flare [
40]. The reductions in B cell counts in this study were comparable with those reported with belimumab [
23]. Likewise, a preliminary report confirmed some B cell reductions in phase 3 trials of the soluble and bound BAFF antagonist tabulumab [
41].
Little is known about the nature and purpose of membrane-bound BAFF. The phenotype of genetically engineered mice that express membrane-bound BAFF and almost no soluble BAFF is similar to that of BAFF-deficient mice, with marked reductions in B cells, serum immunoglobulin (Ig) levels, and antigen-specific Ig responses [
42]. Based on this, patients with SLE in whom both membrane-expressed and soluble BAFF have been neutralized may have a clinical response no different from that of patients in whom only soluble BAFF has been neutralized (e.g., belimumab). However, since expression of membrane-bound BAFF in the genetically engineered mice was low [
42], there may not have been sufficient membrane-expressed BAFF to clarify its function.
In addition to effects on B cells, BAFF can co-stimulate in vitro proliferation of and cytokine production by T cells [
43,
44], and in mice it can skew in vivo inflammatory responses toward a T helper (Th) 1 cell profile and away from a Th2 cell profile in a B cell-independent manner [
45]. Moreover, BAFF promotes generation of Th17 cells at least in part through direct effects on T cells [
46]. Whether membrane-expressed BAFF might contribute to these BAFF-driven effects on T cells is unknown.
These were a priori first-in-human studies, and the reason for the large portion of subjects testing positive for neutralizing antibodies is unknown. Ten and 4 % of subjects in the phase 1a and 1b studies, respectively, tested positive for antibodies prior to dosing. As previous use of an investigational product was an exclusion criteria (belimumab was an investigational product at the time of the studies), prior exposure to belimumab would not explain these findings. Of the three subjects who had neutralizing antibodies during the studies, the nature and frequency of adverse events did not differ from the group as a whole, and all three subjects were negative for neutralizing antibodies at the end of the study. Given that a large portion of the subjects in these studies were positive for antibodies at any time, we believe this may be a characteristic of this population or perhaps a limitation of the assay, which may require further development [
40].
A limitation of these studies is that this population had mild stable or inactive disease (mean SLE Disease Activity Index score of approximately 3) and baseline immune-phenotypic profiles that were more similar to those of the healthy population than patients with SLE [
27]. This may have made it difficult to appreciate significant immunologic changes seen in these studies. Moreover, many patients were being treated with immunosuppressives and steroids, which could have impacted baseline B cell levels. There were also some baseline differences between the dosing groups in some B cell measurements. Although this most likely reflects the heterogeneity of the population, this may have muted or amplified treatment responses in these small treatment groups.
Competing interests
WS has been a member of scientific advisory boards for Eli Lilly, a consultant for Novartis and Celgene, received preclinical grant support from Xencor and clinical trial support from Pfizer and GlaxoSmithKline; he has no other conflicts of interest to declare. JTM has received consulting fees and grant support from Human Genome Sciences and GlaxoSmithKline. RJL has served as a consultant to Genentech, Biogen Idec, and Roche; she has no other conflicts of interest to declare. JB has received consulting fees from GlaxoSmithKline and Human Genome Sciences, Inc; she has no other conflicts of interest to declare. MHW has received consulting fees and research support from Centocor; he has no other conflicts of interest to declare. EMG was a consultant to EMD Serono, Inc, GlaxoSmithKline, and Human Genome Sciences; she has no other conflicts of interest to declare. BC, AK, TCC, ER, JF, and WT are and DH, KRK, SSB, and DJZ were employees of Amgen, Inc; they have no other conflicts of interest to declare. This study was funded by Amgen, Inc., and blisibimod is now licensed to Anthera Pharmaceuticals, Inc.
Authors’ contributions
WS, BC, DH, AK, KRK, TCC, ER, JF, SSB, and DJZ designed the studies, collected and interpreted the data, drafted and revised the manuscript critically for important intellectual content. JTM, RJL, JB, DJW, MHW, EMG, and WT designed the studies, collected and interpreted the data, and revised the manuscript critically for important intellectual content. All authors approved the final version to be published.