To identify and characterize a fully human antibody directed against B lymphocyte stimulator (BLyS), a tumor necrosis factor–related cytokine that plays a critical role in the regulation of B cell maturation and development. Elevated levels of BLyS have been implicated in the pathogenesis of autoimmune diseases.
A human phage display library was screened for antibodies against human BLyS. A human monoclonal antibody, LymphoStat-B, specific for human BLyS was obtained from the library screening and subsequent affinity optimization mutagenesis. The antibody was tested for inhibition of human BLyS in vitro and in an in vivo murine model. Additionally, the consequences of BLyS inhibition were tested in vivo by administration of LymphoStat-B to cynomolgus monkeys.
LymphoStat-B bound with high affinity to human BLyS and inhibited the binding of BLyS to its 3 receptors, TACI, BCMA, and BLyS receptor 3/BAFF-R. LymphoStat-B potently inhibited BLyS-induced proliferation of B cells in vitro, and administration of LymphoStat-B to mice prevented human BLyS-induced increases in splenic B cell numbers and IgA titers. In cynomolgus monkeys, administration of LymphoStat-B resulted in decreased B cell representation in both spleen and mesenteric lymph nodes.
A fully human monoclonal antibody has been isolated that binds to BLyS with high affinity and neutralizes human BLyS bioactivity in vitro and in vivo. Administration of this antibody to cynomolgus monkeys resulted in B cell depletion in spleen and lymph node. This antibody may prove therapeutically useful in the treatment of autoimmune diseases in humans.
B lymphocyte stimulator (BLyS; trademark of Human Genome Sciences, Rockville, MD) protein (also known as BAFF, THANK, TALL-1, TNFSF13B, and zTNF4), a member of the tumor necrosis factor (TNF) ligand superfamily, is synthesized as a 285–amino acid type II membrane protein and exists in both membrane and cleaved 152–amino acid soluble forms (1–6). BLyS is expressed on monocytes, macrophages, and monocyte-derived dendritic cells, and is up-regulated in response to interferon-γ and interleukin-10 (IL-10) (7). In vitro, recombinant human BLyS enhances B cell proliferation and Ig secretion through interaction with receptors expressed predominantly on B cells (1). In vivo, recombinant human BLyS causes splenic hyperplasia in mice, primarily due to increases in the number of mature B cells. BLyS administration to mice also causes increases in the total serum Ig concentration and enhanced humoral responses to both T cell–independent and T cell–dependent antigens (1, 2, 8).
BLyS has been shown to bind with high affinity to 3 receptors, all of which are members of the TNF receptor family (6, 9, 10). Two of the receptors, BCMA and TACI, also bind APRIL, another TNF family member that is the most homologous to BLyS (11–13). The precise function of these two receptors is not well understood, but studies in knockout mice suggest that BCMA is functionally redundant, since BCMA-deficient mice have a normal B cell phenotype (14). TACI has been shown to have an inhibitory role in B cell development, since TACI knockout animals exhibit increased peripheral B cells, reduced responses to T cell–independent antigens (15, 16), and lymphoproliferative disorders and autoimmune disease (17). In contrast, BLyS-deficient mice show a phenotype of severe loss of mature B cells in the spleen, peripheral blood, and lymph nodes (18, 19). A third receptor, BLyS receptor 3 (BR-3; BAFF-R), is specific for BLyS (9, 10). The A/WySnJ mouse strain, which harbors a naturally occurring truncation of BR-3, exhibits a phenotype similar to that of BLyS-deficient mice (9, 10, 20), suggesting that BR-3 is the primary mediator of the effects of BLyS on B cell survival and maturation.
Several lines of evidence suggest that elevated levels of BLyS may be involved in the pathogenesis of B cell–mediated autoimmune diseases. First, constitutive overexpression of BLyS in transgenic animals results in manifestations of autoimmune-like symptoms, including anti-DNA antibodies, rheumatoid factor, circulating immune complexes, and deposition of immune complexes in the kidney leading to glomerulonephritis. These symptoms resemble those of systemic lupus erythematosus (SLE) and some aspects of rheumatoid arthritis (RA) (21, 22). Second, elevated levels of BLyS have been found in other murine models of SLE, including MRL-lpr/lpr and (NZB × NZW)F1 strains (6). Finally, elevated levels of BLyS have been observed in the serum of patients with SLE and RA (23, 24) as well as Sjögren's syndrome (25, 26). Significant correlations between BLyS levels and autoantibody production were shown in these studies.
The association of BLyS overproduction with manifestations of several autoimmune diseases suggests that modulation of BLyS levels could be a novel therapeutic approach to the treatment of such diseases. Indeed, studies with soluble BLyS receptors as antagonists have shown efficacy in reducing disease manifestations in murine models of both SLE and RA (6, 27, 28). With the aim of developing a therapeutic agent for autoimmune disease, LymphoStat-B (Human Genome Sciences) antibody, a human, neutralizing monoclonal antibody against human BLyS, was generated. We describe herein the generation of human antibodies against BLyS using phage display, as well as the in vitro and in vivo characterization of LymphoStat-B.
MATERIALS AND METHODS
Phage display libraries and selections.
Nonimmunized human single-chain variable fragment (scFv) phage display libraries, recently expanded from 1010–1011 clones (29), were used for lead-scFv isolation. BLyS, at 10 μg/ml in phosphate buffered saline (PBS), was immobilized on immunotubes (Nunc, Naperville, IL), and BLyS-binding phage was isolated by 3 sequential rounds of panning (29). The optimized variants were isolated by selection from randomized libraries in solution using biotinylated BLyS captured on streptavidin-coated paramagnetic beads (Dynal, New Hyde Park, NY) (30). The hBLySsc-1 and hBLySsc-2 VH third complementarity-determining region (CDR3) randomized repertoires were constructed by polymerase chain reaction (PCR) using mutagenic oligonucleotides to replace the last 6 VH CDR3 amino acids with randomized codons. The mutated DNA for both lineages was ligated into the phagemid vector pCANTAB6 (31) and electroporated into Escherichia coli TG1 (32). Libraries of 2 × 108 and 6 × 108 individual clones were generated for hBLySsc-1 and hBLySsc-2, respectively. The randomized libraries were subjected to 1 round of panning on 10 μg/ml of immobilized BLyS, followed by 12 rounds of soluble selection using decreasing concentrations of biotinylated BLyS, from 50 nM down to 100 pM.
Preparation and screening of scFv in a receptor binding inhibition assay.
Expression of scFv was induced in 2TY medium (1.6% weight/volume tryptone, 1% w/v yeast extract, 0.5% w/v NaCl) containing 1 mM IPTG for 3 hours at 30°C. Periplasmic extracts (33) were prepared, and His-tagged scFv were purified on nickel agarose (Qiagen, Chatsworth, CA) and buffer exchanged into PBS. BLyS was biotinylated using N-hydroxysuccinimide (NHS)–biotin (Pierce, Rockford, IL) at a molar ratio of 20:1 of biotin to BLyS. IM9 cells were immobilized onto poly-L-lysine–coated 96-well plates at a density of 1 × 105 cells/well, and scFv was added in the presence of 25 ng/ml (0.5 nM) of biotinylated BLyS. Bound biotinylated BLyS was detected via streptavidin–Delfia (Perkin Elmer, Boston, MA). The 50% inhibition concentration (IC50) values for the competition of each scFv for BLyS binding to its receptor were then determined.
Cloning and expression of antibodies.
Variable heavy and light chain fragments from scFv clones were subcloned as separate heavy and light chains into expression vectors obtained from Cambridge Antibody Technology (34). Expression constructs were used for transfection of mammalian cell lines (HEK293T or NSO) to produce whole IgG. Conditioned media were collected 72 hours posttransfection, and antibodies were purified by protein A chromatography.
Binding to BLyS in solid phase (direct enzyme-linked immunosorbent assay [ELISA]).
ELISA plates (Immulon-II; Dynatech, Chantilly, VA) were coated with 1 μg/ml of streptavidin (Sigma, St. Louis, MO) in PBS. After washing in PBS and blocking with 3% bovine serum albumin (BSA), 100 μl of biotinylated BLyS solution in PBST (1 μg/ml in PBS, 0.1% Tween 20, 0.1% BSA) was added to each well, and the wells were incubated for 1 hour at room temperature. After washing, 100 μl of serial dilutions of antibodies (66–6.6 × 10–6 nM) in diluent buffer (PBS, 0.1% Tween 20, 0.1% BSA) was dispensed into individual wells of BLyS-coated plates. The plates were incubated for 2 hours at room temperature, washed 4 times with PBST, and then 100 μl of peroxidase-conjugated goat anti-human IgG (1 μg/ml in diluent buffer; Vector, Burlingame, CA) was added to each well. Plates were incubated for 1 hour at room temperature, washed 4 times with PBST, and developed with tetramethylbenzidine substrate (Sigma). Absorption at 450 nm was measured with a SpectraMax 3000 instrument (Molecular Devices, Sunnyvale, CA).
Binding to BLyS in solution (competition ELISA).
Serial dilutions of BLyS (103–10–4 nM) were prepared in diluent buffer, and 50-μl aliquots of each dilution were distributed to individual wells of BLyS-coated plates (prepared as described above). Antibodies were diluted to the 50% maximum response concentration (EC50), and 50 μl of the resultant dilution was added to the wells already containing BLyS. The plates were further processed as described above.
Murine splenocyte in vitro proliferation assay.
Serial 3-fold antibody dilutions were prepared in a 96-well plate. The final antibody concentrations ranged from 100 nM to 0.01 nM; a preparation containing no antibody served as a medium control. BLyS (3 ng/ml final concentration) that had been diluted in complete medium (RPMI 1640 with 10% fetal bovine serum containing 100 units/ml of penicillin, 100 μg/ml of streptomycin, 4 mML-glutamine, and 5 × 10–5M β-mercaptoethanol) was added to each well. Medium alone (without BLyS) was used as the negative control for the assay background. Human IgG1 isotype controls were used as negative controls for the antibodies.
Plates were incubated for 30 minutes at 37°C in an atmosphere containing 5% CO2. Splenocytes (final concentration 1 × 106/ml) containing Staphylococcus aureus Cowan strain (Calbiochem, La Jolla, CA) were added to the wells. The plates were further incubated for 72 hours at 37°C in an atmosphere containing 5% CO2. Each well was supplemented with complete medium containing 0.5 mCi of 3H-thymidine (6.7 Ci/mmole; Amersham, Arlington Heights, IL), and the cells were incubated for an additional 20–24 hours at 37°C. The proliferation of each sample was measured by 3H-thymidine incorporation and graphed as a function of antibody concentration. A 4-parameter regression analysis was performed for each data series, and IC50 values were derived.
Neutralizing activity of LymphoStat-B antibody in mice.
All animal experimentation was done in accordance with the Guide for the Care and Use of Laboratory Animals and under the supervision of the Institutional Animal Care and Use Committee. On day 1, female BALB/c mice (ages 8–9 weeks) were injected intravenously with diluent (PBS) or with LymphoStat-B or control human monoclonal IgG1 antibody at doses of 0.05, 0.15, 0.5, 1.5, or 5 mg/kg. Recombinant human BLyS at 0.3 mg/kg or BLyS vehicle (12.5 mM citrate, 125 mM NaCl, pH 6) was injected subcutaneously 1 hour after antibody injection. On day 3, mice received a second dose of antibody (1 hour before BLyS injection). BLyS was injected subcutaneously once a day for 4 consecutive days. On day 5, all mice were killed 24 hours after the last BLyS injection. Sera were analyzed for IgA content by ELISA, wet spleen weights were recorded, and splenocytes were analyzed by flow cytometry using phycoerythrin (PE)–conjugated anti–B220 (anti-CD45R) and fluorescein isothiocyanate (FITC)–conjugated anti-ThB (anti-Ly6D) as markers, as previously described (1, 35).
Effects of LymphoStat-B in cynomolgus monkeys.
Studies with cynomolgus monkeys (Macaca fascicularis) were performed at Charles River Laboratories, Sierra Biomedical Division (Sparks, NV). Monkeys (n = 10 [5 males and 5 females] per dosage group) were given an intravenous injection of vehicle or LymphoStat-B at 5, 15, or 50 mg/kg once a week for 4 weeks. On day 29 (4 weeks of treatment), 3 monkeys of each sex from each dosage group were necropsied. The remaining monkeys were necropsied on day 57, after an additional 4-week treatment-free period. Periodically during the study, the mononuclear cell populations in peripheral blood were evaluated by flow cytometry using a Coulter Epics XL-MCL instrument (Beckman-Coulter, Miami, FL). In addition, relative mononuclear cell populations in the spleen and mesenteric lymph nodes were determined for each monkey at the time of necropsy. For both peripheral blood and tissues, the following markers were used to distinguish immunophenotype: CD2 for total lymphocytes, CD20 for B cells, CD20 and CD21 for mature B cells, CD3 for total T cells, CD3 and CD4 for helper T cells, CD3 and CD8 for suppressor/cytotoxic T cells, and CD14 (CD3–) for monocytes. The PE-labeled anti-CD2 antibody was obtained from Beckman-Coulter. FITC-labeled anti-CD3 and anti-CD20 antibodies, as well as PE-labeled anti-CD4, anti-CD8, anti-CD14, and anti-CD21 antibodies, were obtained from BD PharMingen (San Diego, CA).
Inhibition of BLyS binding to BLyS receptor proteins.
BLyS was labeled with the electrochemiluminescent reporter ORI-TAG–NHS (IGEN International, Gaithersburg, MD), and binding to each of the 3 known BLyS receptors was measured in the presence of LymphoStat-B, human IgG1 isotype control antibody, free receptor fusion protein, or control fusion protein herpesvirus entry mediator ([HVEM]–Fc) essentially as previously described (36). Briefly, fusion proteins of the extracellular domain of each of the 3 BLyS receptors (TACI, BCMA, and BR-3) and HVEM were biotinylated with NHS–LC–biotin (Pierce) at molar ratios ranging from 9:1 to 6.6:1. Twelve 2-fold serial dilutions of LymphoStat-B, receptor, or controls at final concentrations ranging from 53 nM to 0.006 nM were prepared in assay diluent, and 50-μl aliquots of each dilution were added to individual wells of a 96-well plate. Fifty microliters of ORI-TAG–labeled BLyS at a final concentration of 40 ng/ml in assay diluent was added to each well, and the plate was incubated for 60–90 minutes at room temperature, with gentle shaking.
The biotinylated receptor proteins were each separately diluted to 50 ng/ml in assay diluent, and Dynabeads M280–streptavidin (Dynal) were used at a final concentration of 250 μg/ml. One hundred microliters of this mixture was added to each well of the assay plate, and the plate was incubated for 45 minutes at room temperature, with vigorous shaking. The plate was read on an Origen M8 series electrochemiluminescence (ECL) analyzer (IGEN International), and the resulting ECL signal was plotted against the log nanomolar antibody or receptor concentration.
Inhibition of receptor binding in a cellular system (flow cytometry).
Flow cytometry was used to assess the capacity of LymphoStat-B to inhibit binding of BLyS to HEK293T cells transfected with complementary DNA encoding full-length BLyS receptors. Cells were transfected with TACI-, BCMA-, or BR-3–encoding plasmids and used for binding studies 24 hours posttransfection, as previously described (12). Cells were stained with 1 μg/ml of biotinylated BLyS either alone or in the presence of a 5-fold molar excess of LymphoStat-B or an isotype control antibody. BLyS binding was assessed with streptavidin–PE conjugate and analyzed on a FACScan instrument (Becton Dickinson, Mountain View, CA).
Flow cytometry analysis of LymphoStat-B binding.
Peripheral blood mononuclear cells were isolated from whole human blood using lymphocyte separation medium (ICN Biochemicals, Irvine, CA). Cells (106 in 100 μl) in fluorescence-activated cell sorting (FACS) buffer (PBS with 0.1% sodium azide and 0.1% BSA) were dispensed to FACS tubes. Human IgG1 (20 μg in 10 μl) was added to each tube, and the tubes were incubated for 20 minutes at 4°C to block nonspecific Fc receptor–mediated binding. Biotinylated antibody or isotype control (1–2 μg) was added to each tube along with FITC-conjugated lineage-specific marker antibody. Antibody binding was assessed with streptavidin–PE (Dako, Carpinteria, CA) measured in a FACScan instrument. The cell line K562 was used as a positive control for the presence of membrane-bound BLyS.
Isolation of BLyS binding antibodies.
Large nonimmunized repertoires of human scFv fragments displayed on phage (29) were used for selection of single-chain antibodies that bound to human BLyS protein. More than 1,200 distinct BLyS-binding scFv were identified and were ranked by their ability to inhibit BLyS binding to receptors on IM9 cells, a myeloma cell line expressing significant levels of BLyS receptors. A panel of 50 candidate scFv demonstrating the greatest inhibitory activity was selected for further analysis. Inhibition of binding activity for 6 of these scFv is shown in Figure 1A. These scFv were converted to full IgG molecules and assessed for their ability to neutralize the activity of human BLyS protein in a murine splenocyte in vitro proliferation assay. Examples of antibodies demonstrating the greatest neutralizing activity in this assay as full IgG are shown in Figure 1B. Antibodies demonstrating the best inhibitory profile as full IgG molecules were hBLySmAb-1 and hBLySmAb-2. Both antibodies were demonstrated to specifically recognize BLyS and did not bind to other TNF ligand family members, including APRIL, LIGHT, TNFα, Fas ligand, TRAIL, and TNFβ, or to IL-4 or IL-18 (data not shown).
Optimization of lead scFv.
Optimization of the scFv corresponding to the lead candidates hBLySmAb-1 and hBLySmAb-2 was performed to identify antibodies with enhanced inhibitory activity. Randomized libraries were generated from the hBLySsc-1 and hBLySsc-2 sequences in which the terminal 6 residues of the VH CDR3 were randomly replaced in order to increase diversity within the VH domain. Optimized scFv were again selected for binding to human BLyS.
The ability of the selected scFv to inhibit binding of biotinylated BLyS to its receptors on the surface of IM9 cells was assessed, and >30 scFv were identified with improved inhibitory profiles. Figure 2 illustrates the improvements that were observed for the 2 most potent scFv from each lineage, hBLySsc-1.1 and hBLySsc-2.1, compared with their respective parents. The hBLySsc-1.1 isolate (IC50 0.7 nM) showed ∼10-fold improvement over hBLySsc-1 (Figure 2A), and hBLySsc-2.1 (IC50 0.2 nM) was improved ∼20-fold over hBLySsc-2 (Figure 2B).
The top 30 scFv candidates were converted to full IgG molecules and assayed for neutralization activity in the murine splenocyte proliferation assay. Most full IgG showed increased activity over the scFv form. The most potent antibody in the neutralization assay, hBLySmAb-1.1, was designated LymphoStat-B, and selected for further characterization (Figure 3). The hBLySmAb-2.1 antibody, which showed the greatest inhibitory activity as an scFv, did not demonstrate the greatest activity as a full IgG.
In vitro characterization of LymphoStat-B.
Binding characteristics of LymphoStat-B were determined using a BLyS-specific ELISA in which biotinylated BLyS was immobilized in wells coated with streptavidin. As shown in Figure 4, LymphoStat-B bound to BLyS with an EC50 value of 0.024 nM (Figure 4A). The binding of LymphoStat-B was readily inhibited with soluble BLyS, with an IC50 value of 8.5 nM (Figure 4B). Thus, LymphoStat-B binds BLyS in both a solid-phase capture assay and in solution.
Specificity of LymphoStat-B for soluble BLyS.
In studies using a series of monoclonal antibodies against BLyS, differential recognition of membrane-bound BLyS has previously been reported (7), although the biologic basis for this remains unclear. The ability of LymphoStat-B to recognize membrane-bound BLyS on the surface of cells was examined by flow cytometry using primary human cells as well as the K562 myelogenous leukemia cell line that was previously characterized as expressing membrane-bound BLyS (Table 1). The presence of membrane-bound BLyS was demonstrated by staining with a murine monoclonal antibody, 12D6, which was previously characterized as recognizing membrane-bound BLyS. LymphoStat-B failed to stain cells that were positive for monoclonal antibody 12D6. Furthermore, interferon-γ, which has previously been demonstrated to enhance BLyS expression on monocytes, led to increased specific binding of the 12D6 antibody but not LymphoStat-B. These results suggest that LymphoStat-B does not recognize membrane-bound BLyS.
Table 1. Analysis of LymphoStat-B binding to membrane-bound BLyS*
Two-color fluorescence-activated cell sorter analysis was performed on human peripheral blood mononuclear cells (PBMCs) or the cell line K562 (positive control for the presence of membrane-bound B lymphocyte) using biotinylated LymphoStat-B and fluorescein isothiocyanate–labeled lineage-specific antibodies. PBMCs from 3 different donors were studied. A representative binding profile for 1 donor is presented. In all 3 donors, the 4 subpopulations of PBMCs were negative for LymphoStat-B binding. In 2 of the 3 donors, the murine monoclonal antibody detected BLyS on CD14+ monocytes and possibly on a minor subpopulation of CD20+ cells.
Negative control for LymphoStat-B binding.
Negative control for 12D6 binding.
K562 cell line
T lymphocytes (CD3+)
B lymphocytes (CD20+)
Natural killer cells (CD56+)
Monocytes (CD14+) plus interferon-γ
LymphoStat-B inhibition of BLyS binding to TACI, BCMA, and BR-3.
Recent studies have determined that BLyS binds to 3 cellular receptors, TACI, BCMA, and BR-3. Because the role of these receptors in B cell development and function is not fully understood, it was important to assess the ability of LymphoStat-B to inhibit BLyS binding to each receptor. Accordingly, the extracellular domains of TACI, BCMA, and BR-3 were expressed as fusion proteins, and the ability of LymphoStat-B to inhibit labeled BLyS binding to each was evaluated in an electrochemiluminescence detection assay. As shown in Figures 5A–C, LymphoStat-B inhibited BLyS binding to all 3 receptors with equivalent potency (IC50 0.10–0.11 nM).
The ability of LymphoStat-B to inhibit ligand–receptor interactions was also evident in a cell-based assay in which full-length TACI, BCMA, and BR-3 were transiently expressed on the surface of HEK293T cells (Figure 6). BLyS-specific binding was readily detected on all receptor-expressing cells but not on cells transfected with the corresponding vector control. Addition of a 5-fold molar excess of LymphoStat-B, but not control IgG, blocked BLyS binding. Taken together, these findings demonstrate that LymphoStat-B blocks BLyS binding to cells via TACI, BCMA, and BR-3.
LymphoStat-B inhibition of the pharmacologic effects of human BLyS in mice.
Although human and murine BLyS share significant amino acid homology (63% identity), LymphoStat-B does not neutralize the effects of murine BLyS (Human Genome Sciences: unpublished data) and is therefore unable to antagonize endogenous BLyS activity in either normal mice or murine models of autoimmune disease. In order to assess the in vivo efficacy of human-specific BLyS antagonists, a mouse model was developed in which the ability of LymphoStat-B to inhibit the pharmacologic effects of exogenously administered human BLyS was tested. This model allows the biologic actions of BLyS to be measured as increases in spleen weight, increases in serum IgA levels, and increases in the population of mature B cells in the spleen (B220+/ThB+ splenocytes; the human equivalent of the markers are CD45R and Ly6D, respectively) (1, 35).
As shown in Figure 7, subcutaneous administration of 0.3 mg/kg of human BLyS for 4 consecutive days resulted in increases in spleen weights, in the representation of CD45R+ (B220+)/ThB+ splenocytes, and in the total serum IgA concentrations. Coadministration of LymphoStat-B intravenously resulted in a dose-dependent inhibition of these human BLyS-induced effects, with complete inhibition observed between 1.5 and 5.0 mg/kg of monoclonal antibody. No inhibition of BLyS-induced effects was achieved by administration of the Ig isotype control monoclonal antibody (human IgG1). These results demonstrate that LymphoStat-B is an effective in vivo antagonist of human BLyS bioactivity.
Effects of administration of LymphoStat-B to cynomolgus monkeys.
The in vivo consequences of BLyS inhibition was evaluated in cynomolgus monkeys. Cynomolgus monkeys were considered a suitable animal model since cynomolgus BLyS is 96% identical to human BLyS, and LymphoStat-B inhibits the in vitro activity of cynomolgus monkey BLyS with similar efficiency as human BLyS (Human Genome Sciences: unpublished data). Animals were injected with 5, 15, or 50 mg/kg of LymphoStat-B or vehicle once every 7 days for 4 treatment cycles. Animals were necropsied on study day 29 or, following a 28-day treatment-free recovery period, on day 57.
Consistent with the phenotype of BLyS-deficient mice (18, 19), a decrease in B cell representation was observed in spleen and mesenteric lymph nodes from LymphoStat-B–treated monkeys on both day 29 and day 57, as monitored by flow cytometry (Table 2). As might be expected, there was a reciprocal increase in the percentage of tissue T cells, but there was no corresponding increase in circulating T cells, nor was there a consistent T cell immunophenotype that was increased (i.e., CD8+ versus CD4+). Therefore, it is likely that the observed increase in the relative percentage of T cells simply reflects the B cell depletion in these tissues. Consistent with this, histologic analysis of these tissues demonstrated a mild lymphoid depletion primarily in B cell–dependent areas (data not shown). No significant changes were observed in peripheral blood mononuclear cells detected by flow cytometry or in serum Ig concentrations (data not shown). These results demonstrate that LymphoStat-B is able to inhibit BLyS in vivo and that this inhibition results in depletion of B cell populations after a relatively short course of treatment.
Table 2. Effect of LymphoStat-B on tissue immunophenotype in cynomolgus monkeys*
MNCs analyzed, study day, LymphoStat-B dose
LymphoStat-B was administered at the indicated doses once every 7 days for 4 treatment cycles. Lymph node and spleen mononuclear cells (MNCs) were analyzed by flow cytometry using the indicated markers. Values are the mean ± SD percentage of spleen or lymph node MNCs.
P < 0.05 versus control group, by Dunnett's test (analysis of variance).
BLyS plays a critical role in the normal regulation of B cell development and immune response. This was demonstrated by the severely depleted B cell phenotype observed in BLyS-deficient animals (18, 19). As is the case for other cytokines, BLyS levels must be carefully regulated in order to maintain “immune homeostasis” (37). Consistent with this notion are findings that overproduction of BLyS is associated with increased immunoglobulin production and autoimmunity. Transgenic mice overexpressing either murine or human BLyS show increased numbers of mature B cells in the spleen and lymph nodes, elevated levels of total immunoglobulins, rheumatoid factor, anti–double-stranded DNA antibodies, and circulating immune complexes, all of which are manifestations associated with autoimmune disease. The mice eventually develop proteinuria and glomerulonephritis, symptoms that are specifically associated with SLE (21, 22). Elevated levels of BLyS have also been detected in the (NZB × NZW)F1 and MRL-lpr/lpr mouse models of lupus. BLyS levels in these animals increase with age, in parallel with increased autoantibody production, proteinuria, and other disease manifestations, again suggesting an important role for BLyS in the development of autoimmune disease in animals (6).
Evidence that BLyS is associated with human autoimmune disease comes from several studies demonstrating elevated levels of circulating BLyS in serum samples from patients with SLE and RA (23, 24). In RA patients, elevated levels of BLyS were also found in synovial fluid and were consistently greater than the serum levels in the same patient, possibly suggesting that there is local production of BLyS at sites of inflammation (38). Increased BLyS levels have also been found in sera from patients with Sjögren's syndrome, another autoimmune disease characterized by B cell hyperreactivity and autoantibody production (25, 26). Mice transgenic for BLyS also develop symptoms of Sjögren's syndrome (25), consistent with clinical observations of the association of Sjögren's syndrome in subsets of patients with SLE (39). Taken together, these observations suggest that elevated levels of BLyS may contribute to B cell hyperactivity and autoantibody production in multiple autoimmune diseases by enhancing the survival of B cells. Specific targeting of BLyS may therefore provide a novel treatment for these diseases. Toward that end, we have generated a human antibody specific for human BLyS that can neutralize the biologic functions of BLyS in vitro and in vivo.
LymphoStat-B was generated from a naive human Ig library using phage display technology. More than 1,200 antibodies with distinct amino acid sequences were isolated that specifically bound soluble human BLyS. Optimization of selected main candidates allowed further improvements in activity, leading to the selection of the clinical development candidate, LymphoStat-B.
Interactions between BLyS and its receptors are complex. At least 3 distinct receptors exist for BLyS. The major receptor for mediating BLyS-dependent B cell development appears to be the BLyS-specific receptor BR-3. This is based on data showing that the A/WySnJ mouse strain, which has a defect in BR-3, mirrors the severe deficits in peripheral mature B cell populations seen in BLyS-deficient mice (9, 10, 18–20). Although BLyS binds to 2 additional receptors, BCMA and TACI, analyses of mice deficient in these receptors show no B cell deficits in the BCMA knockout mice (14) and increased levels of B cells in TACI-deficient mice (15), suggesting a redundant role for BCMA and an inhibitory role for TACI in the regulation of B cell maturation (17). Importantly, we show by 2 independent methods that LymphoStat-B is able to neutralize BLyS interactions with all 3 BLyS receptors.
BLyS, like many TNF family members, exists in both soluble and membrane-bound forms. Most available data support the action of BLyS as a soluble cytokine (1), and elevated levels of soluble BLyS are correlated with autoimmune disease (23, 24). Any specific activity of membrane-bound BLyS remains to be determined. In fact, the exact nature of membrane-bound BLyS is not fully understood, but it has been shown that a panel of antibodies can differentially recognize membrane-bound BLyS (7). The BLyS antibodies described in the present study were selected specifically for their ability to recognize and inhibit the effects of soluble human BLyS. LymphoStat-B does not recognize membrane-bound BLyS and does not bind to the surface of human cells that bear membrane-bound BLyS. Thus, the specific activity of soluble BLyS is targeted, and any nonspecific effects that might be associated with antibody binding to BLyS-producing cells should be prevented.
BLyS has also been shown to exist in a complex with APRIL to form heterotrimers. Although the biologic significance of these heterotrimers is not clearly defined, they have in vitro biologic activity and interact with BLyS receptors (40). While LymphoStat-B was raised against soluble “homotrimeric” BLyS, interactions with heterotrimers have been shown in in vitro assays (Smith R, Roschke V: unpublished observations), although the affinity of these interactions was 7–8-fold lower than that of the heterotrimer recognizing antibodies reported previously. The significance of this inhibition awaits a clearer delineation of the in vivo role of BLyS/APRIL heterotrimers in B cell regulation.
In an in vitro assay, LymphoStat-B was found to inhibit human BLyS-induced stimulation of B cells from both murine splenocytes and human tonsillar B cells (data not shown). In vivo, LymphoStat-B neutralizes the observable effects of exogenously administered human BLyS in mice. These effects include stimulation of B cell proliferation, as evidenced by increases in spleen weight, representation of mature marginal zone B cells, as demonstrated by increases in B220+/ThB+ B cells, and immunoglobulin secretion, as evidenced by increases in IgA levels. LymphoStat-B at dosages of 1–5 mg/kg completely prevented these BLyS-induced activities. In this model system, LymphoStat-B has similar efficiencies when administered via the intravenous or the subcutaneous route (Hilbert D: unpublished observations). It should also be noted that the dosing schedule of 4 daily injections of BLyS used in the present study has only modest effects on increasing IgM levels and little, if any, effects on IgG levels (35). Nevertheless, LymphoStat-B can reduce BLyS-induced changes in IgM (data not shown).
Administration of LymphoStat-B to normal cynomolgus monkeys resulted in tissue B cell depletion that was sustained for up to 4 weeks after the treatment period. Although no significant depletion of circulating B cells was observed at these time points, it has been suggested that the spleen and lymphoid tissue act as reservoirs for the supply of B cells into the bloodstream (41), and effects on circulating B cells may not be observed until reserves of B cells in lymphoid tissues are adequately suppressed. No significant changes were observed in serum immunoglobulin levels following LymphoStat-B administration. However, since the B cell–depleting drug rituximab results in severe (>90%) and prolonged B cell depletion, with only minor concomitant effects on serum immunoglobulin levels (42), it is perhaps not surprising that changes in serum immunoglobulins were not observed in the study reported here. Although the effects of LymphoStat-B administration over a short period of time might not be expected to fully reflect the phenotypes of BLyS knockout mice, one key similarity with the BLyS knockout mice, namely a reduction of tissue B cell populations, was observed following drug administration.
LymphoStat-B possesses many properties that make it ideally suited for use as a therapeutic agent. First, it binds with high affinity to soluble human BLyS and inhibits its activities both in vitro and in vivo. Second, it does not recognize other TNF family members, including its closest homolog human APRIL, and is therefore specific for the factor implicated in the pathogenesis of several autoimmune diseases. As an antibody, an additional advantage of LymphoStat-B is its long terminal half-life. In mice and monkeys, the half-life of LymphoStat-B is 2.5 days and 11–14 days, respectively (Riccobene T: unpublished observations).
Overproduction of BLyS has been associated with SLE, RA, and Sjögren's syndrome. The etiologies of these diseases are unknown, but they share the common features of B cell hyperactivity and autoantibody production, which are strongly implicated in the pathogenesis of the disease process. Support for the ability of an antagonist of BLyS to affect these diseases comes from several animal studies with soluble BLyS receptors. Administration of TACI-Fc or BCMA-Fc has been shown to inhibit antigen-specific antibody responses and decrease B cell numbers and germinal centers (43, 44). TACI-Fc and BR-3-Fc have also been shown to inhibit the development of SLE disease manifestations in the (NZB × NZW)F1 mouse (6, 28). TACI-Fc has also prevented disease onset in the collagen-induced arthritis model in mice (27).
The critical role BLyS plays in the regulation of B cell homeostasis suggests that deregulation of BLyS may underlie these as well as additional B cell–mediated diseases. LymphoStat-B is being developed as a novel treatment for autoimmune diseases and is currently being evaluated in a phase I clinical trial in SLE patients. Studies are also under way to identify additional diseases in which BLyS might play a role and LymphoStat-B may have therapeutic potential.
The authors wish to thank members of the DNA sequencing facility at Human Genome Sciences for antibody sequencing, Steve Barash for bioinformatics analysis, Steve Ullrich for antibody purification, and Jeff Carrell, Ernest Boyd, and Dana Bresette for flow cytometry.