A novel human CD32 mAb blocks experimental immune haemolytic anaemia in FcgammaRIIA transgenic mice

research paper

A novel human CD32 mAb blocks experimental immune haemolytic anaemia in FccRIIA transgenic mice

Annet van Royen-Kerkhof,1,2 Elisabeth A. M. Sanders,2 Vanessa Walraven,2 Marleen Voorhorst-Ogink,3 Eirikur Saeland,1 Jessica L. Teeling,3 Arnout Gerritsen,3 Marc A. van Dijk3, Wietse Kuis,2 Ger T. Rijkers,2 Laura Vitale,4 Tibor Keler,4 Steven E. McKenzie,5 Jeanette H. W. Leusen1,* and Jan G. J. van de Winkel1,3,* 1

Immunotherapy Laboratory, Department of

Immunology, 2Department of Pediatric Immunology, University Medical Centre Utrecht, 3

Genmab, Utrecht, the Netherlands, 4Medarex, Annandale, NJ, and 5Cardeza Foundation for Hematologic Research, Thomas Jefferson University, Philadelphia, PA, USA

Summary A fully human IgG1 kappa antibody (MDE-8) was generated, which recognised Fc-gamma receptor IIa (FccRIIa) molecules on CD32 transfectants, peripheral blood monocytes, polymorphonuclear cells and platelets. This antibody blocked FccRIIa ligand-binding via its F(ab¢)2 fragment. Overnight incubation of monocytes with F(ab¢)2 fragments of MDE-8 leads to a c. 60% decrease in cell surface expression of FccRIIa. MDE8 whole antibody induced a concomitant c. 30% decrease of FccRI on THP-1 cells and monocytes. In humans FccRIIa plays an important role in the clearance of antibody-coated red blood cells in vivo. As an equivalent of FccRIIa does not exist in mice, the in vivo effect of MDE-8 was studied in an FccRIIa transgenic mouse model. In these mice, antibody-induced anaemia could readily be blocked by MDE-8. These data document a new human antibody that effectively blocks FccRIIa, induces modulation of both FccRIIa and FccRI from phagocytic cells, and ameliorates antibody-induced anaemia in vivo.

Received 17 February 2005, accepted for publication 25 April 2005 Correspondence: Dr Jeanette H. W. Leusen,

Keywords: Fc receptors, antibodies, autoimmunity, transgenic mice.

Immunotherapy Laboratory, Department of Immunology, KC 02.085.2, Lundlaan 6, 3584 EA Utrecht, the Netherlands. E-mail: [email protected] *These two authors contributed equally to this study.

Receptors for IgG (Fc-gamma receptors, FccR) are important immune-response modulating molecules. In man, three classes of leucocyte FccRs are known. FccRI (CD64) is primarily expressed on monocytes and macrophages, and can be induced on granulocytes. FccRIII (CD16) molecules are expressed on neutrophils, macrophages and natural killer cells. FccRI and IIIa are activatory receptors and their function depends on a common FcRc-chain. FccRIIa (CD32) represents a low-affinity receptor, interacting only with immune-complexed IgG. It is the only IgG receptor with an immunoreceptor tyrosine-based activation motif signalling motif within its ligand-binding chain. FccRIIa represents the most widely distributed FccR subclass and is present on a variety of blood cells, including neutrophils, monocytes and platelets (King et al, 1990; Daeron, 1997; Deo et al, 1997). Autoimmune haemolytic anaemia (AIHA) is an acquired disorder resulting from auto antibodies directed against doi:10.1111/j.1365-2141.2005.05571.x

erythrocytes. AIHA includes warm AIHA, cold agglutinin syndrome, paroxysmal cold haemoglobinuria, mixed-type AIHA and drug-induced AIHA (Gehrs & Friedberg, 2002). The clinical entities depends on the characteristics of the autoantibodies, whether they react most strongly at 37
C (warm AIHA, predominantly IgG) or at lower temperatures (cold AIHA, mainly IgM and, to a lesser extent, IgG). Corticosteroids are effective in inducing a remission, but relapses are frequent. Immune haemolytic anaemia can also result from alloantibodies, as can be frequently observed in neonates or after allogeneic bone marrow transplantation because of alloantibodies to ABO or non-ABO antigens (Drobyski et al, 1996; Hashimoto, 1998). In the latter cases, the response to conventional treatment is generally unsatisfactory, thus requiring prolonged courses of immunosuppressive therapy, which, in turn, might influence engraftment and increase the risk for viral infections. Especially in these cases a

ª 2005 Blackwell Publishing Ltd, British Journal of Haematology, 130, 130–137

CD32 mAb Blocks Anaemia in huCD32 Tg Mice high mortality rate is observed (O’Brien et al, 2004), therefore warranting novel therapeutic strategies. We investigated the immunotherapy of AIHA in a murine model. In mice the prominent role of FccR in extra vascular clearance of IgG-coated red blood cells (RBC) was demonstrated in complement deficient mice. Upregulation of FccRI induced an accelerated clearance of IgG-opsonised RBC (Shibata et al, 1990; Berney et al, 1992; Sylvestre et al, 1996). Both FccRI and FccRIIa have been implicated in AIHA in man (Kumpel & Hadley, 1990; Clynes & Ravetch, 1995; Dijstelbloem et al, 2000; Miescher et al, 2004). We generated a panel of seven fully human monoclonal antibodies (mAbs) specific for FccRII (CD32) in human Ig transgenic (Tg) mice (Fishwild et al, 1996). One of them, MDE-8, was selected for further characterisation and was shown to effectively block and modulate FccRIIa in vitro. We furthermore investigated its capacity to affect immune haemolytic anaemia in vivo in an FccRIIa Tg mouse model.

Materials and methods Cells The IIA1.6 cells transfected with FccRIIa-R131(van den HerikOudijk et al, 1994), IIa-H131 (van den Herik-Oudijk et al, 1994), Fca receptor I (Morton et al, 1995), Jurkat cells, naturally expressing FccRIIIa, IIA1.6 cells expressing FccRIIb1* (van den Herik-Oudijk et al, 1994), as well as Raji cells, expressing CD20, were cultured in Roswell Park Memorial Institute (RPMI) 1640 medium (GibcoBRL, Grand Island, NY, USA) supplemented with 10% heat-inactivated fetal calf serum (FCS; Hyclone, Logan, UT, USA) and penicillin/streptomycin. The human monocytic cell-line THP-1 (American Type Culture Collection, Rockville, MD, USA) was cultured in RPMI 1640 medium with 10% FCS and penicillin/streptomycin. Mononuclear cells from healthy donors, allotyped for FccRIIa by polymerase chain reaction (Carlsson et al, 1998), were isolated from heparinised venous blood using Ficoll-Histopaque (Sigma, St Louis, MO, USA) density gradient centrifugation. Isolated cells were washed twice and resuspended either in RPMI 1640 medium with 10% FCS, or phosphate-buffered saline (PBS) containing 1% bovine serum albumin and 0Æ02% sodium azide (immunofluorescence buffer, IFB).

197 was used as unlabelled IgG to block FccRI in select experiments (Pfefferkorn & Fanger, 1989). The CD20 mAb 1F5 (mIgG2a) (Coulter, Miami, FL, USA) and Fc-receptor antibodies Gran-1 against CD16 (Sanquin, Amsterdam, the Netherlands), and A77 against CD89 (Medarex) were used as positive controls in binding studies. CD89 mAb 14.1 (human IgG1; Medarex) was used as isotype control (van Spriel et al, 2002). Murine IgG1 anti-glycophorin A was purchased from Dako (Cambridge, UK) and mouse IgG1 anti-murine erythrocyte antibody 105-2H was kindly provided by Dr Izui (Department of Pathology, University of Geneva, Geneva, Switzerland) (de Sa Oliveira et al, 1996).

Generation of human mAbs Human Ig-Tg (Co7) mice (Fishwild et al, 1996) were immunised intraperitoneally (IP) with 50 lg FccRIIa-H131 conjugated to human serum albumin (HSA; a generous gift of Dr P.M. Hogarth, Austin Research Institute, Heidelberg, Vic, Australia) in complete Freund’s adjuvans, and several times with IIA1.6 cells transfected with either FccRIIa-H131 or FccRIIa-R131 in PBS (van den Herik-Oudijk et al, 1994). Mice with positive antibody titres were sacrificed, and splenocytes were fused with SP2/0 myeloma cells according to standard laboratory protocols. Resulting hybridomas were screened by enzyme-linked immunosorbent assay (ELISA) for human IgG, j antibodies, and in flow-cytometric assays with IIA1.6 FccRIIa-R131 and IIA1.6 FccRIIa-H131 cells. Hybridomas producing human jCD32 antibodies were subcloned by at least two rounds of limiting dilution. Select human antibodies were purified by affinity chromatography, using Sepharosecoupled protein A (Pharmacia, Uppsala, Sweden). Fractions were analysed by electrophoresis on 4–15% sodium dodecyl sulphate (SDS) gradient gels, and stained with Coomassie brilliant blue. Protein concentrations were determined by optical densitometry at 280 nm, and a Pierce assay (Rockford, IL, USA). F(ab¢)2 fragments were generated by standard methods of digestion of whole antibody, using pepsin and citric acid (Coligan et al, 2002). The preparations were depleted of residual Fc portions by protein A adsorption chromatography. F(ab¢)2 fragments were purified by gel filtration, and checked by 10% SDS-polyacrylamide gel electrophoresis.

Measurement of CD32 antibody affinity Monoclonal antibodies Monoclonal antibodies directed to FccRI (CD64), mAb 22 (mIgG1), mAb 32 (mIgG1) and mAb 197 (mIgG2a), as well as anti-FccRIIa mAb IV.3 (mIgG2b) and anti-FccRIIa-R131 mAb were from Medarex (Annandale, NJ, USA). Anti-FccRIIa (CD32) mAb FLI8.26 (mIgG1) was obtained from Research Diagnostics (Flanders, NJ, USA). Antibody FLI8.26 and mAb32 were both used as fluorescein isothiocyanate (FITC)labelled IgG in direct immunofluorescent staining, and mAb

biacore avidity measurements were performed using the biacore 3000 to obtain the affinity of MDE-8 for HSAFccRIIa-H131. Briefly, HSA-FccRIIa-H131was coupled to a CM5 sensor chip according to the manufacturer’s protocol (biacore 3000 control software, version 3.1.1; Biacore International SA, Neuchatel, Switzerland). In each kinetic analysis the binding of the analyte MDE-8 was compared with the binding to a reference control CM5 sensor chip. Serial dilutions of MDE-8 were tested (maximum dose 40 lg/ml).

ª 2005 Blackwell Publishing Ltd, British Journal of Haematology, 130, 130–137

131

A. van Royen-Kerkhof et al Association and dissociation curves were fitted for monomeric interaction in the model ‘Langmuir 1:1’, to determine ka and kd and to calculate KA and KD. All data were analysed using BIAEvaluation, version 3.1 (Biacore International SA).

Binding assays IIA1.6 transfectants (FccRIIa-H131, FccRIIa-R131or FccRIIb*, CD64 and CD89), Jurkat cells (FccRIIIa) and Raji cells (CD20) were incubated with MDE-8 IgG (30 min, 4
C). Cells were washed twice, and incubated with FITC-labelled goat F(ab¢)2 anti-human-kappa serum (Jackson, West Grove, PA, USA). Separate panels of FccRIIa-R131 and FccRIIa-H13III transfectants were preincubated with 41H16 and IV.3, respectively, at optimal concentrations. Subsequently, MDE-8 (c. 80% of optimal binding concentration) was added and suspensions were further incubated for 30 min at 4
C. Cells were washed and treated as mentioned above. Binding of monoclonal antibody MDE-8 to FccRIIa was tested on monocytic THP-1 cells, expressing both FccRI and FccRIIa (Fleit & Kobasiuk, 1991). IIA1.6-CD89 transfectants were used as negative control. Assays were performed on ice in 96-well, roundbottomed polypropylene micro titre plates (Nunc, Roskilde, Denmark). THP-1 cells (105) were seeded in 100 ll IFB. Cells were incubated 30 min at 4
C with MDE-8, with or without preincubation with FccRI- blocking mAb 197 (Medarex) (Pfefferkorn & Fanger, 1989), or with MDE-8 F(ab¢)2 fragments. Cells were washed twice, and incubated with FITC-labelled goat F(ab¢)2 anti-human-kappa serum (Jackson). Flow cytometric analyses were performed using a FACSCalibur flow cytometer (Becton Dickinson, San Jose, CA, USA).

Antibody sensitised erythrocyte rosetting Monocytes of healthy FccRIIa-R/R131 donors, or IIA 1.6 FccRIIa-R131transfectants were incubated with 10 lg/ll MDE-8 IgG or 10 lg/ml MDE-8 F(ab¢)2 fragments (1 h, 4
C). CD32-blocking mAb IV.3 and hIgG1 CD89 mAb 14.1 (Medarex) (van Spriel et al, 2002) were used as controls. Human erythrocytes were opsonised (30 min at 37
C) with mouse IgG1 anti-glycophorin A mAb (Dako), binding selectively to FccRIIa-R131 (van de Winkel et al, 1989; Braakman et al, 1992). Monocytes and transfectants were incubated with opsonised erythrocytes at a ratio of 1:50. Cells and erythrocytes were pelleted by centrifugation (10 min at 249 g) and incubated at 4
C for 1 h. Cells were then gently resuspended in RPMI 1640 medium. Cells with at least three bound erythrocytes were microscopically scored as antibody sensitised erythrocyte (EA)-rosettes.

FccR modulation THP-1 cells were grown overnight at 37
C in RPMI 1640 medium with 10% FCS and recombinant interferon-c 132

(300 U/ml) (Amgen, Thousand Oaks, CA, USA) to enhance FccRI expression (Guyre et al, 1983). Cells were washed twice in RPMI 1640 medium and divided into two tubes. The first suspension was kept at 4
C in IFB, the second was resuspended in RPMI 1640 medium at 37
C. MDE-8 IgG was added to both sets of cells in different concentrations and incubated overnight. Cells were washed twice with IFB, and kept at 4
C. FITC-labelled mAb FLI8.26 (van de Winkel & Anderson, 1995) and mAb 32.2 (Shen et al, 1986) binding to FccIIa, and FccRI outside their respective ligand-binding regions, were added at 10 lg/ml to detect FccRII, and FccRI expression. Flow cytometric analyses were performed on a FACS scan. Decrease in receptor cell surface expression was related to mean fluorescence intensity of cells incubated in IFB at 4
C, which was set at 100% (Wallace et al, 1997).

Blocking of FccR phagocytosis Mononuclear cells were isolated from whole blood by Ficoll (Amersham) separation, subsequently erythrocytes were lysed to obtain polymorphonuclear cells (PMN). Cells were washed and 4 · 107 cells were green fluorescently labelled with PKH2 (Sigma) according to the manufacturer’s instructions. Cells were washed three times with 10% FBS (Sigma) and adjusted to 4 · 106/ml. Autologous RBC were generated by washing 1 ml whole blood three times with PBS (Invitrogen, Carlsbad, CA, USA). Cell suspensions were diluted with 10 ml of PBS, then 1 ml was pelleted and stained red with PKH26 (Sigma). Cells were washed three times with 10% FBS and adjusted to 4 · 105/ml. A total of 50 ml of mAb MDE-8 was incubated at various concentrations with a maximum concentration of 10 lg/ml, in 96-well, round-bottomed polypropylene micro titre plates (Nunc). In parallel, CD32 mAb IV.3, or intravenous immunoglobulins (IVIg) was added at the same concentrations. A mouse IgG1 mAb to glycophorin A (Dako) was added at a final concentration of 1 lg/ml. RBC and PMN were added and incubated for 1 h at 37
C. Cells were pelleted and resuspended in lysis buffer for 10 min at 37
C. Phagocytosis was analysed by flow cytometry.

Granulocyte activation Human PMN were incubated with different concentrations of MDE-8 (1 h, 37
C). Cells were centrifuged and supernatants were taken for quantification of lactoferrin release in a sandwich ELISA. A 96-well maxisorp plate (Nunc) was coated with rabbit anti-lactoferrin IgG (Sigma) (1 h, 37
). Alkaline phosphate-conjugated rabbit anti-lactoferrin IgG (1:5.000) (Sigma) was used to detect bound lactoferrin. ELISA plates were analysed in a spectrophotometer (Ultrospec 2000; Pharmacia, Cambridge, UK). Unstimulated granulocytes were used as a negative control, whereas formyl-methionyl-leucylphenylalanine (FMLP, 10 lg/ml; Sigma) plus cytochalasine B (10)6 mol/l; Sigma) were used as positive control.

ª 2005 Blackwell Publishing Ltd, British Journal of Haematology, 130, 130–137

CD32 mAb Blocks Anaemia in huCD32 Tg Mice Heparinised whole blood samples were incubated with MDE-8 for 30 min at 37
C, or with FMLP (Sigma) as a positive control. Reactions were stopped by adding ice-cold IFB and cells were kept at 4
C. Cell surface expression of CD63-PE (Pharmingen, Becton Dickinson, San Diego, CA, USA) and CD66-b (Immunotech, Vaudreuil-Dorion, Quebec, Canada) was assessed by flow cytometry.

Platelet activation Platelet-rich plasma (PRP) was prepared from fresh citrated blood by centrifugation at 200 g for 15 min at 22
C. PRP was adjusted to a platelet concentration of 20 000 platelets/ml. Autologous platelet poor plasma was obtained from PRP by centrifugation at 3000 g for 10 min. Subsequently, 25 ll PRP were incubated with 50 ll MDE-8 (10 lg/ml) for 30 min at 37
C. To assess the effect of FccR cross-linking, Fcc-specific goat anti-human IgG (Jackson) was added at a fivefold excess. Reactions were stopped at different time points by adding 1% paraformaldehyde in HEPES tyrode buffer. Platelets were pelleted and incubated for 30 min at 4
C with FITC-labelled CD62-P mAb (Pharmingen) as a marker for platelet degranulation (Joutsi-Korhonen et al, 2003). Platelets incubated with FITC-labeklled anti-glycoprotein IIIa IgG (Rubinstein et al, 1991) served as a positive control. Platelet activation was further studied in platelet aggregation assays. PRP was incubated with MDE-8 at 0, 0Æ1, 1 and 10 lg/ml with or without cross-linking by Fcc-specific goat anti-human IgG (Jackson). Platelet aggregation was determined from light transmission in PRP measured at 37
C in a PAP-4 4-channel aggregometer (Bio/Data, Horsham, PA, USA) over 7 min, using adenosine diphosphate (1Æ25–20 lmol/l; Sigma) as a positive control.

Auto-immune haemolytic anaemia mouse model The FccRIIa-Tg mice in an FcR c-chain knock-out (KO) background were used, as previously described (McKenzie et al, 1999). Haemolytic anaemia was induced by a single IP injection of 400 lg mIgG1 anti-mouse RBC mAb 105.2H (Fossati-Jimack et al, 2000) in 8–12-week-old female FccRIIaTg mice, and in non-Tg (NTg) mice in an FcRc-chain KO background (controls). For FccRIIa blockade 100 lg MDE-8 was injected intravenously 60 min prior to IP administration of mAb 105.2H. Control mice were injected with equal volumes of physiological saline (0Æ9% NaCl). Blood samples were obtained daily (days 0–8) from the retro orbital plexus, or tail veins and were collected in heparinised tubes. Erythrocyte counts were measured in whole blood using a Cell-Dyn1700 counter (Abbott, Abbott Park, IL, USA).

Statistical analyses Mean values and standard deviations were calculated and differences between various groups were assessed by a one-way

analysis of variance (anova) test, with significance accepted at the P < 0Æ05 level.

Results Generation of human antibodies specific for CD32 We generated a panel of seven mAb directed against FccRII (CD32) in human immunoglobulin Tg mice (Fishwild et al, 1996), which all bound CD32 on cell lines, peripheral blood monocytes and PMN. All were IgG1, j-isotype. One antibody (MDE-9) selectively recognised the FccRIIa-H131 allotype and was characterised in more detail (van Royen-Kerkhof et al, 2004) Monoclonal antibody MDE-8 was selected for further development based on its superior performance in ligandbinding inhibition over the other CD32 mAb.

MDE-8 binding characteristics The specificity of antibody MDE-8 was tested in studies using IIA1.6 transfected cells. MDE-8 bound to IIA1.6 cells transfected with FccRIIa-H131, FccRIIa-R131 or FccRIIb*. No binding was observed to Jurkat cells naturally expressing FccRIIIa (CD16), or to IIA1.6 cells expressing FcaRI (CD89) or FccRI (CD64) (Fig. 1, panel A). Binding to FccRIIa-H131 and FccRIIa-R131 could be blocked by preincubation of murine CD32 mAb. MDE-8 bound, furthermore, to human peripheral blood monocytes, PMN, platelets and B cells. We next assessed whether MDE-8 interacted with FccRII via its F(ab) fragment, by studying the binding to THP-1 cells and monocytes expressing both FccRI and FccRIIa (Fleit & Kobasiuk, 1991), of whole Ab MDE-8, as well as purified F(ab¢)2 fragments. Binding curves of MDE-8 F(ab¢)2 fragments were similar to those obtained with whole IgG. Binding curves to THP-1 cells in the presence of the FccRI-blocking mAb 197 were also nearly identical. MDE-8 showed no specific-binding to IIA1.6-CD89 transfectants (Fig. 1, panel B). Similar results were obtained using human peripheral blood monocytes (data not shown, n ¼ 7). Subsequently, binding of MDE-8 was studied by Biosensor analyses. The avidity of MDE-8 was tested by immobilising HSA-FccRIIa-H131 to a CM-5 sensor chip. Curves were fitted for monomeric interaction in the model ‘Langmuir 1:1’ to determine kd. MAb MDE-8 IgG bound to FccRIIa with a intermediate KD of 9 · 10)9 mol/l. Avidity measurements were performed in three individual experiments, and yielded similar results. We next assessed the effect of MDE-8 on FccRIIa ligandbinding on THP-1 cells and peripheral blood monocytes. Erythrocytes sensitised with a mouse IgG1 anti-glycophorin mAb were used to assay the in vitro functional binding to FccRIIa (van de Winkel et al, 1988, 1989). Formation of EArosettes could be blocked by both IgG, and F(ab¢)2 fragments of MDE-8 at similar levels as mAb IV.3. The human IgG1 CD89 mAb (negative control) did not block EA-rosette

ª 2005 Blackwell Publishing Ltd, British Journal of Haematology, 130, 130–137

133

A. van Royen-Kerkhof et al

Fig 1. Binding characteristics of MDE-8. (A) Binding of MDE-8 to a panel of (transfected) cells determined by flow cytometry. Open bars represent binding of positive control monoclonal antibody (mAb) (mAb IV.3 for CD32-IIa, mAb AT10 for CD32-IIb1, mAb H22 for CD64, mAb Gran-1 for CD16, mAb A77 for CD89, mAb 1F5 for CD20). Select samples of FccRIIa-R131 and -H131 transfectants were preincubated with murine CD32 mAb 41H16 and IV.3, respectively. MDE-8 binds both allotypes of FccRIIa, but not to other FccR. Binding to FccRIIa could be blocked by CD32 mAb 41H16 and IV.3. Data are representative of three individual experiments, yielding identical results. (B) Binding of MDE-8 to FccRI and FccRIIaexpressing monocytic cell line THP-1. Cells were incubated with different concentrations of MDE-8 IgG ( ), MDE-8 IgG in the presence of FccRI-blocking mAb 197 ( ), or MDE-8 F(ab¢)2 fragments (h). Binding of MDE-8 IgG to CD89-transfecants was used as negative control ( ). Binding was detected with fluorescein isothiocyanate-labelled goat anti-human IgG-kappa and analysed by flow cytometry. Data are representative of three experiments, yielding similar results.

formation. MDE-8 effectively blocked phagocytosis of mouse IgG1 antibody opsonised RBC [selectively targeting to FccRIIa (van de Winkel et al, 1988)] in a dose-dependent manner, similar to the CD32 mAb IV.3 (at a concentration of 10 lg/ml). IVIg did not block FccRIIa-mediated phagocytosis.

FcR modulation In addition to direct blocking of CD32, MDE-8 may reduce FccR-mediated phagocytosis of antibody-coated cells by down-modulating additional FccR. We, therefore, studied the effect of MDE-8 on FccR expression on THP-1 cells and monocytes. After overnight incubation of THP-1 cells or monocytes with MDE-8, membrane expression of FccRIIa and 134

Fig 2. Effect of MDE-8 on FccR modulation. THP-1 cells were incubated overnight with MDE-8 IgG ( ) or MDE-8 F(ab¢)2 fragments (h) at different concentrations. As a control, cells were incubated with the humanised CD64 mAb H22 ( ). Expression of FccRIIa was assessed the next day with fluorescein isothiocyanate (FITC)-labelled mAb FLI8.26 (panels A and C). FccRI expression was assessed with FITClabelled CD64 mAb 32.2 (panels B and D). Levels of FccRIIa and FccRI on non-treated cells were set at 100%. Data are expressed as percentage of receptor surface expression of non-treated cells. Data represent mean values ± SD of five individual experiments, performed on separate days.

FccRI (CD64) was assayed with directly labelled mAbs, binding outside the ligand-binding regions of FccRIIa and FccRI (FLI8.26 and 32.2 respectively) (Pfefferkorn & Fanger, 1989; Ierino et al, 1993; van de Winkel & Anderson, 1995). MDE-8 IgG and MDE-8 F(ab¢)2 fragments induced a 60–70% decrease of FccRIIa membrane expression, both on THP-1 cells and human monocytes (Fig. 2, panel A). Notably, MDE-8 IgG induced a c. 30% decrease in membrane expression of FccRI as well, whereas MDE-8 F(ab¢)2 fragments did not (Fig. 2, panel B). This indicates the Fc-part of MDE-8 IgG to be responsible for the reduction in FccRI surface expression. As a control, we used mAb H22, a humanised antibody directed against FccRI, which selectively modulates CD64 (Wallace et al, 1997). Incubation with CD64 mAb H22 had no effect on FccRIIa expression (Fig. 2, panel C), but induced a reduction of FccRI membrane expression of c. 60% (Fig. 2, panel D). These data documented MDE-8 IgG’s ability to down-modulate both FccRIIa and FccRI on phagocytic cells. Fig. 2 shows the results of five representative experiments performed on THP-1 cells.

Granulocyte and platelet activation As MDE-8 is a human IgG1 mAb, cross-linking of FccRI and FccRIIa may potentially initiate granulocyte activation. Specific-granule mobilisation was analysed by studying lactoferrin release in human PMN. Incubation with MDE-8 did not

ª 2005 Blackwell Publishing Ltd, British Journal of Haematology, 130, 130–137

CD32 mAb Blocks Anaemia in huCD32 Tg Mice

Fig 4. MDE-8 blocks development of anaemia in FccRIIa-Tg mice. Eight-week-old female FccRIIa transgenic mice or NTg mice (as a control) received an intravenous dose of MDE-8 (5 lg/g) or saline. After 60 min, mIgG1 anti-mouse erythrocyte antibody 105.2H was given intraperitoneally. Hb levels (g/dl) were determined at various time-points using a Cell-Dyne 1700 multiparameter haematology analyser. Data represent mean values ± SD from four mice. **P < 0Æ01, *P < 0Æ05. The experiment was performed twice yielding essentially identical results.

Fig 3. Effect of MDE-8 on granulocyte and platelet activation. (A) Isolated human polymorphonuclear cells (PMN) were incubated with MDE-8 at different concentrations for 60 min (37
C). Formylmethionyl-leucyl-phenylalanine plus cytochalasine B was used as positive control, and unstimulated PMN as negative controls. Supernatants were tested for the presence of lactoferrin by ELISA. Experiments were repeated three times, yielding essentially identical results. (B) Isolated human platelets were incubated with MDE-8 (10 lg/ml) in the absence (grey bars) or presence (white bars) of antihuman IgG. Anti-glycoprotein IIIa mAb was used as positive control. Reactions were stopped at different time points. Increases in CD62-p surface expression were measured by flow cytometry. Data are representative for three experiments with almost identical results.

induce a significant increase in lactoferrin release (Fig. 3A). Activation of granulocytes leads to an increase of CD63 and CD66 surface molecules as can be detected by flow cytometry. However, incubation of whole blood samples with MDE-8 did not induce increases in CD63 or CD66b expression. Only when a secondary antibody was added, leading to MDE-8 crosslinking, an increased CD63 and CD66b surface expression was observed (n ¼ 3, data not shown). The FccRIIa cross-linking on platelets might induce platelet activation. Platelet aggregate formation and increases in CD62p surface expression were studied to assess this. Incubation of isolated human platelets with MDE-8 did not induce an aggregatory response. No increased CD62-p surface expression could be detected by flow cytometric analyses (n ¼ 3). Only cross-linking of MDE-8 by an Fcc-specific goat anti-human IgG led to increased CD62-p expression (Fig. 3B). Although activation was only observed under these experimental conditions, we cannot fully rule out the presence of anti-idiotypic antibodies in vivo, which might induce cross-linking and activation.

Because MDE-8 exhibited a weak-binding to FccRIIb (Fig. 1), we studied the activation, as well as apoptosis induction, of human B-cell lines in vitro. MDE-8 did not modulate these B-cell functions in four independent experiments (data not shown).

Prevention of auto-immune haemolytic anaemia by MDE-8 In order to study the potential application of MDE-8 in vivo, we used a recently documented human FccRIIa Tg mouse model. A human FccRIIa Tg mouse was bred into an FcR c-chain KO strain, lacking all murine activatory Fcc-receptors (van Vugt et al, 1996; Park et al, 1998). The sole activating FccR in these mice is a human FccRIIa (McKenzie et al, 1999). A single intravenous injection of 80 lg MDE-8 1 h prior to the infusion of mAb105.2H effectively blocked the induction of haemolytic anaemia, comparable with levels observed in control (NTg) mice (Fig. 4). Untreated mice and protected mice showed a slight decrease in erythrocyte counts, which was most likely because of frequent blood sampling.

Discussion In this study, we generated a novel human monoclonal antibody directed to FccRII, MDE-8. Monoclonal antibody MDE-8 bound to FccRIIa on cell lines, peripheral blood monocytes, PMN, platelets and blocked ligand binding via this receptor. FccRIIa membrane expression was decreased upon overnight incubation of monocytic cells with MDE-8. Notably, also the high-affinity FccRI (CD64) receptor expression was decreased under these conditions because of interaction with the Fc-tail of MDE-8. This implicates that mAb MDE-8 may block both classes of activatory Fcc receptors on mononuclear and phagocytic cells. Although MDE-8 IgG bound to both FccRIIa allotypes, as well as to FccRIIb, its in vitro effects were restricted to FccRIIa.

ª 2005 Blackwell Publishing Ltd, British Journal of Haematology, 130, 130–137

135

A. van Royen-Kerkhof et al A number of earlier studies addressed the role of FccR in autoimmune cytopenic diseases, in which antibody-opsonised blood cells were prematurely destroyed by phagocytic cells. The presence of murine FccRIII was found crucial for the induction of anaemia in a murine AIHA model (Meyer et al, 1998). It has also been shown that FccRIIa plays an important role in the clearance of human IgG-coated RBC in man (Dijstelbloem et al, 2000; Miescher et al, 2004). As mice lack an FccRIIa ‘homologue’, the role of FccRIIa could only be evaluated in human FccRIIa-Tg mice, bred into an FcR c-chain KO background (leading to absence of murine FccRI and FccRIII) (Takai et al, 1994). In a murine model for immune thrombocytopenia, McKenzie et al (1999) studied the role of FccRIIa in the immune clearance of platelets. However, as the Tg thrombocytes expressed FccRIIa (McKenzie et al, 1999), some antiplatelet antibodies may cross-link FccRIIa, and direct activation of platelets may thus contribute to the observed decrease in thrombocytes. To further assess the role of human FccRIIa in autoimmune cytopenic disease, we set-up a murine model for autoimmune haemolytic anaemia, using a pathogenic murine IgG1 anti-mouse RBC mAb 105.2H, directed against the erythrocyte anion channel band 3 on murine RBCs (de Sa Oliveira et al, 1996). In our model, huFccRIIa Tg mice (in an FcR c-chain KO background) developed anaemia, although not as profound as that observed in wild-type mice (Meyer et al, 1998). These data implicate that hFccRIIa functions as an equivalent of murine FccRIII in this model (Hazenbos et al, 1998; Meyer et al, 1998). In FccRIIa Tg mice, anaemia was strongly decreased upon injection of the FccRIIa-blocking mAb MDE-8 (prior to injection of anti-RBC mAb 105.2H). NTg mice in an FcRcchain KO background, did not develop anaemia, as was to be expected because of the absence of activating Fcc receptors. In control experiments, we did not observe activation of platelets or granulocytes upon binding of MDE-8 IgG, unless the mAb was cross-linked by a secondary antibody. Although binding of MDE-8 to FccRIIb in vitro did not have functional effects on B cells, the mAb may well affect B cells in vivo, which remains to be addressed. In summary, we characterised a novel fully human CD32 mAb that blocks the function of FccRIIa and possibly of FccRI. In vivo experiments showed a reduction of anaemia when this mAb was applied in an antibody-induced anaemia model in an FccRIIa Tg mouse. These findings justify further investigations on FccRIIa as a therapeutic target in immune cytopenic diseases.

Acknowledgements The authors wish to thank Ingrid van den Brink for excellent technical assistance in animal handling. This study was supported in part by research funding from Medarex to MV-O, AG, MD, TK, JL, JW and in part by the US Public Health Service, NHLBI R01 HL61865 and P01 HL40387 to SMcK. 136

References Berney, T., Shibata, T., Merino, R., Chicheportiche, Y., Kindler, V., Vassalli, P. & Izui, S. (1992) Murine autoimmune hemolytic anemia resulting from Fcc receptor-mediated erythrophagocytosis: protection by erythropoietin but not by interleukin-3, and aggravation by granulocyte-macrophage colony-stimulating factor. Blood, 79, 2960– 2964. Braakman, E., van de Winkel, J.G.J., van Krimpen, B.A., Jansze, M. & Bolhuis, R.L. (1992) CD16 on human cd T lymphocytes: expression, function, and specificity for mouse IgG isotypes. Cellular Immunology, 143, 97–107. Carlsson, L.E., Santoso, S., Baurichter, G., Kroll, H., Papenberg, S., Eichler, P., Westerdaal, N.A., Kiefel, V., van de Winkel, J.G.J. & Greinacher, A. (1998) Heparin-induced thrombocytopenia: new insights into the impact of the FccRIIa-R-H131 polymorphism. Blood, 92, 1526–1531. Clynes, R. & Ravetch, J.V. (1995) Cytotoxic antibodies trigger inflammation through Fc receptors. Immunity, 3, 21–26. Coligan, J.E., Kruisbeek, A.M., Margulies, D.H., Shevach, E.M. & Strober, W. (2002) Current Protocols in Immunology. Wiley & Sons, New York, NY. Daeron, M. (1997) Fc receptor biology. Annual Review of Immunolgy, 15, 203–234. Deo, Y.M., Graziano, R.F., Repp, R. & van de Winkel, J.G.J. (1997) Clinical significance of IgG Fc receptors and Fcc R-directed immunotherapies. Immunology Today, 18, 127–135. Dijstelbloem, H.M., Bijl, M., Fijnheer, R., Scheepers, R.H., Oost, W.W., Jansen, M.D., Sluiter, W.J., Limburg, P.C., Derksen, R.H., van de Winkel, J.G.J. & Kallenberg, C.G. (2000) Fcc receptor polymorphisms in systemic lupus erythematosus: association with disease and in vivo clearance of immune complexes. Arthritis and Rheumatism, 43, 2793–2800. Drobyski, W.R., Potluri, J., Sauer, D. & Gottschall, J.L. (1996) Autoimmune hemolytic anemia following T cell-depleted allogeneic bone marrow transplantation. Bone Marrow Transplantation, 17, 1093– 1099. Fishwild, D.M., O’Donnell, S.L., Bengoechea, T., Hudson, D.V., Harding, F., Bernhard, S.L., Jones, D., Kay, R.M., Higgins, K.M., Schramm, S.R. & Lonberg, N. (1996) High-avidity human IgG kappa monoclonal antibodies from a novel strain of minilocus transgenic mice. Nature Biotechnology, 14, 845–851. Fleit, H.B. & Kobasiuk, C.D. (1991) The human monocyte-like cell line THP-1 expresses FccRI and FccRII. Journal of Leukocyte Biology, 49, 556–565. Fossati-Jimack, L., Ioan-Facsinay, A., Reininger, L., Chicheportiche, Y., Watanabe, N., Saito, T., Hofhuis, F.M., Gessner, J.E., Schiller, C., Schmidt, R.E., Honjo, T., Verbeek, J.S. & Izui, S. (2000) Markedly different pathogenicity of four immunoglobulin G isotype-switch variants of an anti-erythrocyte autoantibody is based on their capacity to interact in vivo with the low-affinity Fcc receptor III. Journal of Experimental Medicine, 191, 1293–1302. Gehrs, B.C. & Friedberg, R.C. (2002) Autoimmune hemolytic anemia. American Journal of Hematology, 69, 258–271. Guyre, P.M., Morganelli, P.M. & Miller, R. (1983) Recombinant immune interferon increases immunoglobulin G Fc receptors on cultured human mononuclear phagocytes. Journal of Clinical Investigations, 72, 393–397. Hashimoto, C. (1998) Autoimmune hemolytic anemia. Clinical Reviews in Allergy and Immunology, 16, 285–295.

ª 2005 Blackwell Publishing Ltd, British Journal of Haematology, 130, 130–137

CD32 mAb Blocks Anaemia in huCD32 Tg Mice Hazenbos, W.L., Heijnen, I.A., Meyer, D., Hofhuis, F.M., Renardel de Lavalette, C.R., Schmidt, R.E., Capel, P.J., van de Winkel, J.G.J., Gessner, J.E., van den Berg, T.K. & Verbeek, J.S. (1998) Murine IgG1 complexes trigger immune effector functions predominantly via FccRIII (CD16). Journal of Immunology, 161, 3026–3032. van den Herik-Oudijk, I.E., Westerdaal, N.A., Henriquez, N.V., Capel, P.J. & van de Winkel, J.G.J. (1994) Functional analysis of human FccRII (CD32) isoforms expressed in B lymphocytes. Journal of Immunology, 152, 574–585. Ierino, F.L., Hulett, M.D., McKenzie, I.F. & Hogarth, P.M. (1993) Mapping epitopes of human FccRII (CDw32) with monoclonal antibodies and recombinant receptors. Journal of Immunology, 150, 1794–1803. Joutsi-Korhonen, L., Smethurst, P.A., Rankin, A., Gray, E., IJsseldijk, M., Onley, C.M., Watkins, N.A., Williamson, L.M., Goodall, A.H., de Groot, P.G., Farndale, R.W. & Ouwehand, W.H. (2003) The lowfrequency allele of the platelet collagen signaling receptor glycoprotein VI is associated with reduced functional responses and expression. Blood, 101, 4372–4379. King, M., McDermott, P. & Schreiber, A.D. (1990) Characterization of the Fcc receptor on human platelets. Cellular Immunology, 128, 462– 479. Kumpel, B.M. & Hadley, A.G. (1990) Functional interactions of red cells sensitized by IgG1 and IgG3 human monoclonal anti-D with enzyme-modified human monocytes and FcR-bearing cell lines. Molecular Immunology, 27, 247–256. McKenzie, S.E., Taylor, S.M., Malladi, P., Yuhan, H., Cassel, D.L., Chien, P., Schwartz, E., Schreiber, A.D., Surrey, S. & Reilly, M.P. (1999) The role of the human Fc receptor FccRIIA in the immune clearance of platelets: a transgenic mouse model. Journal of Immunology, 162, 4311–4318. Meyer, D., Schiller, C., Westermann, J., Izui, S., Hazenbos, W.L., Verbeek, J.S., Schmidt, R.E. & Gessner, J.E. (1998) FccRIII (CD16)deficient mice show IgG isotype-dependent protection to experimental autoimmune hemolytic anemia. Blood, 92, 3997–4002. Miescher, S., Spycher, M.O., Amstutz, H., De, H.M., Kleijer, M., Kalus, U.J., Radtke, H., Hubsch, A., Andresen, I., Martin, R.M. & Bichler, J. (2004) A single recombinant anti-RhD IgG prevents rhesus D immunization: association of RhD positive red blood cell clearance rate with polymorphisms in the FccRIIA and IIIA genes. Blood, 103, 4028–4035. Morton, H.C., van den Herik-Oudijk, I.E., Vossebeld, P., Snijders, A., Verhoeven, A.J., Capel, P.J. & van de Winkel, J.G.J. (1995) Functional association between the human myeloid immunoglobulin A Fc receptor (CD89) and FcR c-chain. Molecular basis for CD89/FcR c-chain association. Journal of Biological Chemistry, 270, 29781– 29787. O’Brien, T.A., Eastlund, T., Peters, C., Neglia, J.P., Defor, T., Ramsay, N.K. & Scott Baker, K. (2004) Autoimmune haemolytic anaemia complicating haematopoietic cell transplantation in paediatric patients: high incidence and significant mortality in unrelated donor transplants for non-malignant diseases. British Journal of Haematology, 127, 67–75. Park, S.Y., Ueda, S., Ohno, H., Hamano, Y., Tanaka, M., Shiratori, T., Yamazaki, T., Arase, H., Arase, N., Karasawa, A., Sato, S., Ledermann, B., Kondo, Y., Okumura, K., Ra, C. & Saito, T. (1998) Resistance of Fc receptor- deficient mice to fatal glomerulonephritis. Journal of Clinical Investigations, 102, 1229–1238. Pfefferkorn, L.C. & Fanger, M.W. (1989) Cross-linking of the high affinity Fc receptor for human immunoglobulin G1 triggers tran-

sient activation of NADPH oxidase activity. Continuous oxidase activation requires continuous de novo receptor cross-linking. Journal of Biological Chemistry, 264, 14112–14120. van Royen-Kerkhof, A., Sanders, E.A., Wijngaarden, S., van Roon, J.A., Voorhorst-Ogink, M., Walraven, V., Gerritsen, A., van Dijk, M.A., Kuis, W., Rijkers, G.T., Keler, T., Leusen, J.H.W. & van de Winkel, J.G.J. (2004) Flow cytometric determination of FccRIIa (CD32) polymorphism. Journal of Immunological Methods, 294, 135–144. Rubinstein, E., Kouns, W.C., Jennings, L.K., Boucheix, C. & Carroll, R.C. (1991) Interaction of two GPIIb/IIIa monoclonal antibodies with platelet Fc receptor (FccRII). British Journal of Haematology, 78, 80–86. de Sa Oliveira, G.G., Izui, S., Ravirajan, C.T., Mageed, R.A., Lydyard, P.M., Elson, C.J. & Barker, R.N. (1996) Diverse antigen specificity of erythrocyte-reactive monoclonal autoantibodies from NZB mice. Clinical Experimental Immunology, 105, 313–320. Shen, L., Guyre, P.M., Anderson, C.L. & Fanger, M.W. (1986) Heteroantibody-mediated cytotoxicity: antibody to the high affinity Fc receptor for IgG mediates cytotoxicity by human monocytes that is enhanced by interferon-c and is not blocked by human IgG. Journal of Immunology, 137, 3378–3382. Shibata, T., Berney, T., Reininger, L., Chicheportiche, Y., Ozaki, S., Shirai, T. & Izui, S. (1990) Monoclonal anti-erythrocyte autoantibodies derived from NZB mice cause autoimmune hemolytic anemia by two distinct pathogenic mechanisms. International Immunology, 2, 1133–1141. van Spriel, A.B., Leusen, J.H.W., Vile, H. & van de Winkel, J.G.J. (2002) Mac-1 (CD11b/CD18) as accessory molecule for FcaR (CD89) binding of IgA. Journal of Immunology, 169, 3831–3836. Sylvestre, D., Clynes, R., Ma, M., Warren, H., Carroll, M.C. & Ravetch, J.V. (1996) Immunoglobulin G-mediated inflammatory responses develop normally in complement-deficient mice. Journal of Experimental Medicine, 184, 2385–2392. Takai, T., Li, M., Sylvestre, D., Clynes, R. & Ravetch, J.V. (1994) FcR c-chain deletion results in pleiotrophic effector cell defects. Cell, 76, 519–529. van Vugt, M.J., Heijnen, A.F., Capel, P.J., Park, S.Y., Ra, C., Saito, T., Verbeek, J.S. & van de Winkel, J.G.J. (1996) FcR c-chain is essential for both surface expression and function of human FccRI (CD64) in vivo. Blood, 87, 3593–3599. Wallace, P.K., Keler, T., Coleman, K., Fisher, J., Vitale, L., Graziano, R.F., Guyre, P.M. & Fanger, M.W. (1997) Humanized mAb H22 binds the human high affinity Fc receptor for IgG (FccRI), blocks phagocytosis, and modulates receptor expression. Journal of Leukocyte Biology, 62, 469–479. van de Winkel, J.G.J. & Anderson, C.L. (1995) CD32 cluster workshop report V. In: Leucocyte Typing (ed. by S.F. Scholossman, L. Boumsell, W. Gilks, J.M. Harlan, T Kishimoto, C. Morimoto, J. Ritz, S. Shaw, R. Silverstein, T.A. Springer, T.F. Tedder, R.F. Todd), pp. 823–826. Oxford University Press, Oxford. van de Winkel, J.G.J., van Duijnhoven, H.L., van Ommen, R., Capel, P.J. & Tax, W.J. (1988) Selective modulation of two human monocyte Fc receptors for IgG by immobilized immune complexes. Journal of Immunology, 140, 3515–3521. van de Winkel, J.G.J., Boonen, G.J., Janssen, P.L., Vlug, A., Hogg, N. & Tax, W.J. (1989) Activity of two types of Fc receptors, FccRI and FccRII, in human monocyte cytotoxicity to sensitized erythrocytes. Scandinavian Journal of Immunology, 29, 23–31.

ª 2005 Blackwell Publishing Ltd, British Journal of Haematology, 130, 130–137

137