A nonfucosylated human antibody to CD19 with potent B-cell depletive activity for therapy of B-cell malignancies

Abstract

A human anti-CD19 antibody was expressed in fucosyltransferase-deficient CHO cells to generate nonfucosylated MDX-1342. Binding of MDX-1342 to human CD19-expressing cells was similar to its fucosylated parental antibody. However, MDX-1342 exhibited increased affinity for FcγRIIIa-Phe158 and FcγRIIIa-Val158 receptors as well as enhanced effector cell function, as demonstrated by increased potency and efficacy in antibody-dependent cellular cytotoxicity (ADCC) and phagocytosis assays. MDX-1342 showed dose-dependent improvement in survival using a murine B-cell lymphoma model in which Ramos cells were administered systemically. In addition, low nanomolar binding to cynomolgus monkey CD19 and increased affinity for cynomolgus monkey FcγRIIIa was observed. In vivo administration of MDX-1342 in cynomolgus monkeys revealed potent B-cell depletion, suggesting its potential utility as a B-lymphocyte depletive therapy for malignancies and autoimmune indications.

Introduction

B-lymphocyte-directed immunotherapies, such as anti-CD20 antibody rituximab, have demonstrated activity in both B-cell malignancies and autoimmune diseases [1, 2]. However, a subpopulation of patients does not respond to current therapeutic agents [3, 4], hence there remains a need for an agent with enhanced mechanisms of B-cell depletion. CD19 represents an interesting B-cell target with a number of potential advantages over CD20-directed therapies. CD19 is expressed throughout the developmental stages of B cells yet is not expressed on hematopoietic stem cells, T cells, or other non-lymphoid cells. In addition, CD19 is expressed later than CD20 through the plasmablast stage of B-cell differentiation [1]. Consequently, CD19 expression is relatively high in many pre-B- and immature B-lymphoblastic leukemias and B-cell malignancies in which CD20 is poorly expressed. CD19⁺ plasmablasts may also play a role in the perpetuation of autoimmune diseases [5].

The potential of CD19 as an immunotherapeutic target was recognized many years ago, but initial clinical trials with monoclonal antibodies (mAbs) to CD19 did not result in durable effects despite demonstrating responses in some patients either as a single agent or in combination with other therapeutic agents [6, 7]. A number of approaches have been attempted to improve the activity of CD19-directed therapy, including the use of antibody–drug conjugates [8, 9] and re-direction of immune effector cells with bispecific antibodies [10, 11]. Both of these approaches appear to hold considerable promise for the development of novel therapies, and a recent clinical trial with a bispecific antibody targeting CD19 and CD3 confirmed that such an approach could generate significant antitumor effects at low doses [12].

There has also been recent interest in enhancing the efficacy of antibody therapeutics through increased Fc receptor binding and, consequently, increased activation of effector mechanisms such as antibody-dependent cellular cytotoxicity (ADCC) and phagocytosis. Fc receptor-dependent effector mechanisms have been demonstrated to be particularly important in rituximab-based lymphoma therapy [13, 14]. The Fc receptor on macrophages and NK cells, FcγRIIIa, has two polymorphic isoforms at residue 158, valine (Val) and phenylalanine (Phe). Of these, the FcγRIIIa-Val158 isoform shows increased binding affinity to human IgG1 antibodies resulting in increased effector function. Clinical responses to rituximab have been significantly better in patients carrying the FcγRIIIa-Val158 isoform compared to the lower affinity FcγRIIIa-Phe158 isoform, and this correlates with improved in vitro ADCC activity [14]. Improved binding to both polymorphic isoforms of FcγRIIIa can be obtained by mutating specific amino acid residues within the Fc region of the antibody [15, 16] or by modification of the N-linked carbohydrate attached to the IgG1 CH2 domain [17, 18]. Both approaches have proved successful in increasing ADCC activity at lower antibody concentrations irrespective of FcγRIIIa isoform. Increased activity in animal models has also been demonstrated [19, 20]. One of the most successful modifications to the Fc carbohydrate structure has been the removal of fucose, which results in an improved binding interaction with the FcγRIIIa carbohydrate [21]. A small proportion of native IgG is nonfucosylated, thus the generation of immune reactions to the glycoengineered antibody is unlikely. Moreover, the improved ADCC activity generated with nonfucosylated antibodies may allow enhanced anticancer activity without the immunogenic risk of amino acid substitutions in the human Fc structure.

In this report, we describe the activity of a fully human antibody to CD19 that has been produced in nonfucosylated form to enhance Fc receptor-mediated effector function. This mAb is currently in phase 1 trials for the treatment of chronic lymphocytic leukemia and rheumatoid arthritis.

Materials and methods

Antibody generation

Transgenic mice expressing human immunoglobulin genes [22] were immunized with CD19-expressing cells and purified soluble recombinant CD19 extracellular domain expressed in CHO cells. Human mAbs were generated using standard hybridoma techniques and screened initially for their ability to bind to CD19-transfected CHO cells by flow cytometry, using untransfected cells as controls. After subsequent screening, genes from the lead hybridoma antibody were cloned, sequenced, and re-expressed in CHO cells by standard techniques to produce parental anti-CD19 mAb. To generate MDX-1342, the parental anti-CD19 mAb was expressed in Ms-704PF CHO cells deficient in the FUT8 gene which encodes α1,6-fucosyltransferase, the only enzyme in native CHO cells capable of fucosylating protein [23]. Expression of antibodies in Ms-704PF CHO cells had previously been shown to result in antibodies deficient in fucose [18, 24].

Glycosylation analysis

To characterize MDX-1342 glycans, N-linked oligosaccharides were released from IgG samples (100 μg) by overnight incubation of the samples with 12.5 mU PNGaseF (Prozyme) at 40°C. Following ethanol precipitation to remove protein, supernatant containing glycans was dried by vacuum centrifugation and resuspended in 19 mM APTS (Beckman) in 15% acetic acid and 1 M sodium cyanoborohydride in THF (Sigma-Aldrich). The glycan labeling reaction was allowed to continue overnight at 40°C followed by 25-fold sample dilution in water. APTS-labeled glycans were applied to a Beckman P/ACE MDQ capillary electrophoresis system with laser-induced fluorescence (LIF) and reverse polarity, using a 50 μm internal diameter N-CHO-coated capillary (Beckman) with 50 cm effective length. Separation was carried out at 20°C using Carbohydrate Separation Gel Buffer (Beckman) at 25 kV for 20 min and monitored using a 3 mW argon laser with excitation and emission wavelengths of 488 nm and 520 nm, respectively.

For monosaccharide analysis, IgG samples (200 μg) were subjected to acid hydrolysis using either 2 M TFA (for estimating neutral sugars) or 6 M HCl (for estimating amino sugars) at 100°C for 4 h. Samples were dried by vacuum centrifugation at ambient temperature and were reconstituted in 200 μL water prior to analysis by high-performance anion-exchange chromatography with pulsed amperometric detection (HPAE-PAD; Dionex). Monosaccharides were separated using a CarboPac PA10 4 × 250 mm column with pre-column Amino Trap and Borate Trap (Dionex). Procedures were followed according to Dionex Technical Note 53. Monosaccharide peak identity and relative abundance were determined using monosaccharide standards (Dionex).

FACS analysis of binding to human CD19

To demonstrate CD19 specificity, MDX-1342 binding to CHO cells expressing human CD19 was evaluated by flow cytometry (FACS) using untransfected CHO cells as controls. Aliquots of 5 × 10⁵ cells/well were incubated for 30 min at 4°C with biotinylated MDX-1342 or isotype control antibody. Cells were washed twice in FACS buffer (PBS with 2% FBS + 20% EDTA) before addition of PE-conjugated streptavidin (Jackson ImmunoResearch). Cells were incubated a further 30 min at 4°C, then washed twice in FACS buffer and assayed using a FACSCalibur (BD Biosciences). Binding curves were plotted for geometric mean fluorescence intensity (GMFI) versus MDX-1342 concentration using GraphPad Prism™ 3.0 software with nonlinear regression analysis, sigmoidal dose response. FACS analysis of MDX-1342 binding to human CD19⁺ cells was also measured using a panel of human lymphoma (Raji and Ramos) and myeloma (ARH-77) cell lines as described above.

Binding of MDX-1342 to CHO cells transfected with human FcγRIIIa

CHO cells (1 × 10⁵) stably expressing human FcγRIIIa-Val158 or FcγRIIIa-Phe158 were incubated for 30 min at 4°C with increasing concentrations of biotinylated MDX-1342 or parental anti-CD19 mAb. Cells were washed twice with PBS containing 2% FBS before addition of Streptavidin–PE (BD Biosciences) at 1:50 dilution for 30 min at 4°C. Acquisition was performed by FACS with gating on propidium iodide negative cells. GMFI was plotted against the various concentrations of mAbs used. Streptavidin–PE conjugate alone served as a control. Data were analyzed by GraphPad Prism software.

ADCC activity

Human peripheral blood mononuclear effector cells were purified from heparinized whole blood by standard Ficoll-Paque separation and cultured overnight in the presence of 200 U/mL rhIL-2 (Research Diagnostics). ARH-77 target cells were labeled with 20 μM BATDA (PerkinElmer) and added to effector cells in a final target to effector ratio of 1:50. Serial dilution of MDX-1342 and isotype control antibody was added to the effector/target cell mixtures. After 1 h incubation at 37°C, 20 μL supernatant was harvested from each well and mixed with 180 μL Europium solution (PerkinElmer) in a flat-bottomed 96-well plate. To confirm FcγRIIIa-dependent activity, ADCC was also determined in the absence or presence of 20 μg/mL murine anti-CD16 antibody 3G8 or isotype control antibody. The reaction was read with a PerkinElmer Fusion Alpha TRF reader using a 400 μs delay with 330/80 excitation and 620/10 emission filters. The percent specific lysis was calculated as: (experimental release spontaneous release)/(maximal release − spontaneous release) × 100, where target and effector cells without antibody represented spontaneous release, and target and effector cells in the presence of 3% Lysol control represented maximal release. Percent specific lysis was plotted against antibody concentration and the data were analyzed by nonlinear regression, sigmoidal dose response (variable slope) using GraphPad Prism software.

Phagocytosis assay

Macrophages from human blood samples (Stanford University Blood Bank) were purified by diluting the buffy coat with twice the volume of PBS, then layering on Lymphocyte Separation Media (Mediatech). The cells were centrifuged for 30 min, collected, resuspended in RBC lysis buffer, and incubated for 15 min at room temperature. The cells were spun, resuspended in PBS, and layered over 14 mL hyperosmolar Percoll and centrifuged for 45 min. The top layer, which contained mostly monocytes, was collected and resuspended in RPMI-1640 supplemented with 10% FBS, l-glutamine (Mediatech), 1% HI HS (Gemini Bio-Products), and 10 ng/mL GM-CSF (Research Diagnostics), and then cultured for 5 days. Target cells were labeled with PKH26 (Sigma-Aldrich) following the manufacturer’s suggested protocol. Target cells (1 × 10⁴/100 μL/well), antibody (50 μL/well), and macrophages (2 × 10⁵/50 μL/well) were dispensed into 96-well plates and incubated for 1 h at 37°C. Cells were subsequently stained with CD14-APC Cy7 (BD Biosciences) and read on a FACSArray Bioanalyzer System (BD Biosciences).

Ramos systemic tumor xenograft model in SCID mice

SCID mice were each intravenously injected with 1 × 10⁶ Ramos cells on Day 0. On Day 1, mice were randomized into six groups of eight mice per group and each animal was injected intraperitoneally with 200 μL MDX-1342 (30, 3, 0.3, and 0.03 mg/kg), nonfucosylated human IgG1 (30 mg/kg), or PBS control. Body weights were measured weekly and mice were monitored daily for paralysis symptoms through Day 99. Mice were killed when they reached endpoint (hind limb paralysis) or showed greater than 20% weight loss. Response to MDX-1342 was measured as a function of hind limb paralysis and/or greater than 20% body weight loss. GraphPad Prism software was used to plot the data as a Kaplan–Meier survival graph and calculate the median survival time for each group in the study. A log rank test was used for statistical analysis.

B-cell depletion in the cynomolgus monkey

Cynomolgus CD20⁺ and CD40⁺ B cells were assessed by FACS following single intravenous injections of 0.1 or 1 mg/kg MDX-1342. A 100 μL aliquot of each blood sample was placed into a clean, labeled tube, and appropriate quantities of commercially available fluorochrome-labeled anti-CD20 and anti-CD40 antibodies were added. Samples were incubated for approximately 30 min at room temperature in darkness, and then a commercially available lysing solution was added to remove erythrocytes. The remaining intact B cells (1–2 × 10⁶ cells/mL) were analyzed immediately or stored at approximately 4°C until analysis (within 120 h of blood collection and after cells were slowly warmed to room temperature).

Results

Nonfucosylated anti-CD19 mAb generation

The parental human anti-CD19 mAb was successfully expressed in CHO cells deficient in the FUT8 gene to generate the nonfucosylated mAb termed MDX-1342. Oligosaccharide and monosaccharide analyses of both MDX-1342 and parental anti-CD19 antibody were performed to determine carbohydrate structures. Analysis of mAb oligosaccharides released by PNGase F revealed that, as expected, the parental anti-CD19 mAb contained a mixture of G0, G1, and G2 carbohydrate structures with fucose attached (Fig. 1). In contrast, MDX-1342 contained the same G0, G1, and G2 structures without fucose. The analysis of monosaccharides was carried out following total hydrolysis of the antibodies. No fucose was detected in the MDX-1342 sample, but 2.3 moles of fucose/mole of mAb was detected in the parental mAb.

Fig. 1.

Oligosaccharide analysis of MDX-1342 and parental anti-CD19 antibody by capillary electrophoresis with laser-induced fluorescence detection; assignment of carbohydrate structures by comparison with glycan standards

CD19 binding specificity and affinity

MDX-1342 specificity for human CD19 was verified using human CD19-transfected CHO cells with parental CHO cells as controls. Dose-dependent binding of MDX-1342 was observed in human CD19 transfectants but not in untransfected parental CHO cells (Fig. 2a).

Fig. 2.

FACS titration of MDX-1342 binding on human FcRγIIIa-transfected CHO-S cells (a), human lymphoma cell lines (b), and FcRγIIIa-Val158-transfected and FcRγIIIa-Phe158-transfected CHO cells (c)

Dose-dependent binding of MDX-1342 to CD19 expressed endogenously on a panel of B-cell lines was also demonstrated. The EC₅₀ values for binding were 0.34, 0.47, and 0.74 μg/mL for Ramos, Raji, and ARH-77 cells, respectively (Fig. 2b). No difference in binding was observed when comparing MDX-1342 to its parental mAb on this panel of cell lines (data not shown). Based on FACS data, the rank order of CD19 expression was Ramos followed by Raji and then ARH-77. MDX-1342 reactivity with CD19 was also demonstrated by immunoprecipitation of CD19 from Raji cells (data not shown).

Binding to Fc receptor Val158 and Phe158 alleles

CHO cells were transfected with human FcγRIIIa expressing either Val158 or Phe158 alleles, and stable transfectant clones were isolated with high expression of the Fc receptor. Binding of MDX-1342 and parental anti-CD19 mAb to CHO cells expressing either FcγRIIIa-Val158 or FcγRIIIa-Phe158 was evaluated by flow cytometry (Fig. 2c). As expected, binding of both antibodies was greater to the higher affinity variant, FcγRIIIa-Val158, and binding of the nonfucosylated antibody to both receptor variants was superior to that of the parental mAb. The EC₅₀ values for MDX-1342 binding to FcγRIIIa-Val158 and FcγRIIIa-Phe158 transfected cells were 0.37 and 1.18 μg/mL, respectively, whereas binding of the parental mAb did not reach saturation at the highest concentration tested (Fig. 2c).

ADCC and phagocytosis

MDX-1342 and parental anti-CD19 mAb mediated ADCC activity was measured using ARH-77 target cells and human peripheral blood mononuclear cells (PBMC) as effector cells at a ratio of 1:50. Rituximab, which is known to induce ADCC in vitro, was included in the assay for comparative purposes. MDX-1342 demonstrated robust ADCC with an average EC₅₀ value of 0.01–0.02 μg/mL and maximal lysis averaging 50% (Fig. 3a). ADCC activity was markedly lower with the parental anti-CD19 with an EC₅₀ value of 0.46 μg/mL and maximal lysis of 21%. Rituximab was found to be slightly less potent with maximal lysis at 43% and less efficacious than MDX-1342 with an EC₅₀ of 0.06 μg/mL. The specificity of the ADCC response was demonstrated using a mouse mAb directed against human FcγRIIIa. The resulting blockade inhibited the ADCC activity mediated by rituximab, MDX-1342, and parental anti-CD19 mAb, but had no effect on the isotype control antibody, confirming that cell lysis was FcγRIIIa-dependent (data not shown).

Fig. 3.

Comparison of ADCC (a) and phagocytosis (b) mediated by MDX-1342 and controls

The activity of MDX-1342 was also assessed in a phagocytosis assay in which MDX-1342 mediated dose-dependent phagocytosis of ARH-77 cells by macrophages (Fig. 3b). Similar to observations with ADCC, the potency and efficacy of MDX-1342 was notably higher when compared to parental anti-CD19 mAb.

Ramos systemic tumor xenograft model in SCID mice

The in vivo efficacy of MDX-1342 was tested in a lymphoma systemic model in which human Ramos cells expressing high levels of cell surface CD19 were injected intravenously into SCID mice and antibodies were administered intraperitoneally 24 h later. In vitro studies with mouse B cells had previously demonstrated that MDX-1342 does not cross-react with mouse CD19, therefore any observed efficacy would depend on human CD19 expressed on Ramos cells. Mice treated with a single dose of MDX-1342 at 30, 3, 0.3, or 0.03 mg/kg showed dose-dependent improvement in survival (Fig. 4). The median survival time was 99 days for MDX-1342 at 30 mg/kg, 63 days for MDX-1342 at 3 mg/kg, 71 days for MDX-1342 at 0.3 mg/kg, and 42 days for MDX-1342 at 0.03 mg/kg. Median survival time was 42 days for the PBS control group and 32 days for the isotype control group. Using a log rank test, increases in median survival were statistically significant (P < 0.001) at 30, 3, and 0.3 mg/kg compared to the PBS control group.

Fig. 4.

Kaplan–Meier plot of the Ramos systemic tumor model treated with MDX-1342 at 0.03, 0.3, 3, and 30 mg/kg

B-cell depletion in the cynomolgus monkey

To first establish that MDX-1342 cross-reacted with cynomolgus monkey CD19, cynomolgus PBMC were analyzed by flow cytometry using biotinylated antibody. Dose-dependent binding of MDX-1342 on human B cells (Fig. 5a) and on cynomolgus B cells (Fig. 5b) was similar and reached saturation at 30 μg/mL. This together with ¹²⁵I-MDX-1342 saturation binding K _D values of 12.6 nM for primary cynomolgus B cells and 6 nM for human peripheral blood B cells (data not shown) demonstrated that MDX-1342 does cross-react with cynomolgus monkey CD19.

Fig. 5.

FACS binding of MDX-1342 to CD40⁺ human B cells (a) and cynomolgus monkey CD40⁺ B cells (b), rituximab binding to cynomolgus monkey CD40⁺ B cells (c), and MDX-1342 binding to CHO cells transfected with cynomolgus monkey CD16 (d)

Rituximab with the same biotin substitution ratio (1 biotin molecule/mAb) as MDX-1342 was included as a positive control. Similar to observations with human B cells, the level of CD20 appeared to be higher than CD19 on cynomolgus B cells (Fig. 5b, c). These studies were extended by cloning and expressing cynomolgus CD16 in CHO cells in a similar manner to that used for human CD16. Unlike human CD16 variants, only one CD16 allele has been observed in the cynomolgus monkey. The EC₅₀ of MDX-1342 binding to CHO cells expressing cynomolgus CD16 was 1.1 μg/mL, a value similar to MDX-1342 binding to human FcγRIIIa-Phe158 (Fig. 5c).

Having established that cynomolgus monkey was an appropriate species for testing, the ability of MDX-1342 to deplete circulating B cells was compared to the parental mAb. A single dose of mAb at 1 mg/kg was administered to animals (4 per group) and the circulating levels of B cells were monitored over time. The results (Fig. 6a) demonstrated an initial B-cell depletion of approximately 50% by the parental antibody and 90% by MDX-1342. Furthermore, the duration of B-cell depletion was longer for MDX-1342. B-cell numbers reached a nadir at approximately 3 weeks and slowly recovered.

Fig. 6.

In vivo B-cell depletion in cynomolgus monkeys: FACS analysis of cynomolgus CD20⁺ B cells following a single dose of 1 mg/kg MDX-1342 or parental anti-CD19 mAb (a) and cynomolgus CD40⁺ B cells following a single dose of 0.1 mg/kg MDX-1342, rituximab, or control IgG1 (b)

In a second study, the ability of MDX-1342 to deplete cynomolgus monkey B cells was compared to rituximab and isotype control antibody at a single dose of 0.1 mg/kg (Fig. 6b). The extent and duration of B-cell depletion was similar for the two B-cell-specific antibodies, despite the higher levels of CD20 in circulating cells. The kinetics of depletion appeared to be more rapid with rituximab, with a faster initial drop in circulating cells, although both antibodies depressed total B cell numbers to a similar extent and for a similar duration. This kinetic difference might reflect differences in the mechanism of action of the two antibodies. For example, rituximab is a potent activator of the complement system in addition to having ADCC activity, whereas in vitro complement activation by MDX-1342 has not been detected (data not shown).

Discussion

Fc receptor-mediated functional activity, including ADCC, has been shown to be an important mechanism for the clinical activity of many therapeutic antibodies, and especially important in the therapy of hematologic cancer cells such as lymphomas and leukemias [25]. A promising strategy to further optimize the efficacy of mAbs is to increase Fc receptor-dependent effector functions by removal of the fucose residue [24, 26]. This glycoengineering strategy has been successfully implemented with the generation of nonfucosylated MDX-1342 with potent B-cell depletion in monkeys.

Rituximab in combination with chemotherapy is now the standard of care for B-cell malignancies. The clinical success of rituximab has demonstrated that B-cell depletion can be an effective strategy in the treatment of lymphomas as well as several autoimmune diseases [27, 28]. Rituximab at 375 mg/m² has shown an overall response rate of approximately 40–50% when administered weekly as a single agent in patients with relapsed, refractory low-grade, or follicular NHL. In newly diagnosed NHL patients, overall response rates of 70–80% with a median duration of 18–26 months have been achieved [3]. Meta-analysis of five other trials showed that follicular lymphoma patients with a maintenance dose of rituximab had statistically significant better overall survival. However, the rate of infection-related adverse events was likewise higher [4]. Patients may relapse through a variety of mechanisms, some of which lead to loss of CD20 expression and resistance to further rituximab treatment [29]. CD19-directed therapy has the potential to be active against a wider range of B-cell tumors, including malignancies that are CD20 negative as well as rituximab failures.

The interaction of anti-CD19 with human FcγRIIIa was investigated because it is a major activating FcR present on two types of cytotoxic effector cells, NK cells and macrophages [30], and the absence of fucose improves mAb binding to human FcγRIIIa while having no effect on binding to other human FcRs [26]. Polymorphism in human FcγRIIIa has been predictive of clinical response to IgG1 mAbs such as rituximab [14, 31]. The higher clinical responses of FcγRIIIa-Val158 patients as compared to FcγRIIIa-Phe158 patients have been attributed to enhanced affinity of FcγRIIIa-Val158 receptors for IgG1 antibodies, and may also result from higher FcγRIIIa-Val158 expression levels [32], although this hypothesis is controversial [33].

Demonstration of improved in vivo efficacy of nonfucosylated antibodies over their fucosylated counterparts has proven to be challenging because mouse NK cells have similar binding affinities to fucosylated and nonfucosylated antibodies. Indeed, human PBMC engraftment into SCID mice has been necessary to demonstrate improved potency with nonfucosylated antibodies in vivo [24]. However, murine FcγRIV, which is homologous to human FcγRIIIa yet largely present on murine macrophages, monocytes, and neutrophils [34], has shown improved binding to nonfucosylated antibodies [30]. In the current study, a mouse systemic model was successfully deployed to demonstrate potent, dose-dependent antitumor activity of nonfucosylated MDX-1342 against human lymphoma cells (Fig. 4).

There appears to be only one FcγRIIIa allele found in cynomolgus monkeys that has high homology to the human receptor (>90% identity in amino acid sequence), but contains neither Val nor Phe at the 158 position which is responsible for differential IgG binding affinity in humans. Nevertheless, nonfucosylated MDX-1342 demonstrated greater binding to the monkey receptor than the fucosylated parental mAb. Although the affinity of MDX-1342 for monkey CD19 was slightly lower than that seen for human antigen, MDX-1342 was able to induce potent and sustained depletion of circulating monkey B cells.

There has recently been renewed interest in the development of CD19-directed immunotherapy. The availability of a human CD19 transgenic mouse model has encouraged the development of such therapeutics through improved understanding of the mechanism and effects of CD19-directed therapy [35]. Clinical data with a CD19-directed bispecific antibody have shown promising antitumor activity with a high frequency of partial and complete regressions [12]. This antitumor activity suggests that CD19-directed immunotherapy has the potential to be clinically effective in a wide range of B-cell malignancies. In addition, other strategies to improve effector functions have been successfully used to generate a CD19 antibody with the capacity to deplete B cells [36].

In conclusion, nonfucosylated anti-CD19 mAb MDX-1342 binds with higher affinity to FcγRIIIa receptors, has significantly improved in vitro ADCC and phagocytic activity, and exhibits greater B-cell depletion in primates relative to its fucosylated counterpart. The low doses of antibody required for in vitro ADCC and in vivo activity suggest that MDX-1342 has strong therapeutic potential and could result in improved clinical outcomes in patients with B-cell malignancies.