CA2226962A1

CA2226962A1 - Use of binding agents to cd47 and its ligands in the treatment or the prophylaxis of various inflammatory, autoimmune and allergic diseases and in the treatment of graft rejection

Info

Publication number: CA2226962A1
Application number: CA002226962A
Authority: CA
Inventors: Marie Sarfati
Original assignee: Individual
Current assignee: Individual
Priority date: 1998-02-16
Filing date: 1998-02-16
Publication date: 1999-08-16
Also published as: WO1999040940A1; AU2508499A

Description

TITLE
Use of binding agents to CD47 and its ligands in the treatment or the prophylaxis of various inflanvnatory, autoimmune and allergic diseases and in the treatment of graft rejection.
FIELD OF THE INVENTION
The present invention relates to new uses of binding agents to CD47 antigen Zo or its ligands, and more particularly to monoclonal antibodies specific to the CD47 or thrombospondin, in the treatment or prophylaxis of various inflammatory, autoimmune and allergic diseases as well as in treatment of tumor metastasis, cachexia and graft rejection.

Integrin superfamily of adhesive receptors are transmembrane heterodimeric molecules which function in cell-matrix and cell-cell adhesion (4, 5). The CD47 Ag, a surface glycoprotein of ~50 KD lvlW is physically and ftmctionally associated

2 o wide X33 integrin mainly av(33 (the vitronectin receptor) on a variety of cell types (6, 7). lntegrin av(33 is a highly promiscuous receptor recognizing Arg-Gly-Asp (RGD) in a wide variety of proteins as well as being expressed by many cell types including endothelial cells, osteoclasts, monocytes, activated lymphoid cells, platelet, fibroblasts and malignant cells (8, 9). The av(33 integrin plays a role in diverse 25 biologic processes such as cell migration and differentiation, himor cell invasion, angiogenesis, bone resorption and immune response (9-l2). Taken together, the role of av~33 in cell directed mobility underlies the importance of this molecule in development, wound repair, cancer and inflammation.
The cDNA sequence of CD47 Ag predicts a multispanning membrane protein with 5 transmembrane domains; the large extracellular N-terminal domain is homologous to members of LgV superfamily and harbours several potential N-glycosylation sites ( 13, I4). It appears that CD47 Ag affects both ligand affinity and signal transduction of (33 integl-ins (6, 15). Indeed, mAb directed against CD47 Ag inhibited ligand binding to av(33 receptor and blocked activation of PMN
to phagocytes, respiratory burst, chemotaxis as well as stimulation of Ca+2 entry in endothelial cells in response to RGD containing proteins. Recent studies have shdwn that CD47 Ag is also involved in transendothelial and transepithelial migration of neutrophils ( I G, 17).
Although CD47 does not bind to vitronectin, its natural ligand was recently identified as being thrombospondin (TSI) ( 18), one of the several ligands of av(33 to which it binds via RGD-containing sequence. TSI interacts with CD47 through its cell binding domain (non RGD sequence). Both anti-TSI and anti-CD47 mAbs partially inhibited TSI-stimulated Ca' 2 entrance in fibroblasts providing a possible 2o mechanism for TSI directed cell mobility via CD47. Lt is also speculated that TSI-CD47 interactions would modulate the function of av(33 during angiogenesis ( 18).
Interestingly, CD47 Ag is an ubiquitous molecule, present on a variety of cell types including lymphocytes and erythrocytes which express low levels or no av(33 integrins, respectively ( 19).
It is also known that the infiltrate at the extravascular site of inflammation during acute (ex: following invasion by microorganisms) or chronic (ex:

3 autoimmune) diseases consists of diverse accumulation of leukocytes (i.e., T, B, g~ranulocytes and macrophages). Despite the specific immune reaction that triggers the disease, the majority of cells found in the inflammatory infiltrate are non-specifically activated leukocytes. Recovery from the disease is intimately related to the regression of this infiltrate and this can be achieved by the elimination of the few Ag specific T cells strongly suggesting that rare cells may retnllate the recruitment and most likely the function of the vast majority of non-specific cells ( 1 ). Recently an "in vitro model" of non-specific T cell activation has been developed whereby cocultures of human resting T cells with autologous monocytes 1 o and LL-2 or IL-12 lead to large production of IFN-~y in the absence of Ag (2).
Results obtained with this method indicates that CD40-CD40L interactions as well as 1 W I 2 are key relmlators of this bystander T cell activation and that sCI)23 further amplifies it by triggering monokine release by monocytes (ex: TNF-a, IL- l ) (3).
SUMMARY OF THE INVENTION
The present invention is directed to a plurality of new uses of binding agents to CD47 antigen (Ag), and more particularly to new uses of monoclonal antibodies 2 0 (mAbs) specific to the CD47. The invention i s based on the discovery that mAbs directed against CD47 Ag strongly abrogate both IFN-y production and monokine release (i.e. LL-1, IL-12 and TNF-a), and dowrlregulate cytokine production by anti-CD3 or allogenic cell-stimulated T cells. These results were obtained with both an antibody named IOG2 produced by the applicant, and with different commercial 2 5 anti-CD47 antibodies, including an antibody named B6H 12 (ATCC HB-9771 ).
Thus, it is believed that anti-CD47 antibodies in general, and CD47 ligands such as thrombospondin could be usefiil in the treatment or prophylaxis of various

4 inflammatory, autoimmune and allergic diseases as well as in treatment of tumor metastasis, cachexia and graft rejection. These human diseases include rheumatoid arthritis, lupus erythematosus, multiple sclerosis, diabetes, uveitis, ulcerative colitis, Crohn's disease, inflammatory bowel disease, thyroiditis, glomemlonephritis, Sjogren disease, graft versus host disease (GVH), allergies, asthma, rhinitis and eczema.
Preferred binding agents include ligands, antibodies, fragments thereof or artificial constmcts comprising antibodies or fragments thereof or artificial i o constructs designed to mimic the binding of antibodies or fragments thereof. Such binding agents are discussed in Dougall et al in Tibtech ( 1994) 12:372-379.
They include complete antibodies, F(ab')2 fragments, Fab fragments, Ev fragments, ScFv fragments, other fragments, CDR peptides and mimetics. These can be obtained/prepared by those skilled in the a.rt. For example, enzyme digestion can be used to obtain F(ab')2 and Fab fragments (by subjecting an LgG to molecule to pepsin or papain cleavage respectively). References to "antibodies" in the following description should be taken to include all of the possibilities mentioned above.
Recombinant antibodies may be used. The antibodies may be humanized or chiinerised. The CDRs may be derived from a rat or mouse monoclonal antibody.
2 o The framework of the variable domains, and the constant domains, of the altered antibody may be derived from a human antibody. Such a humanized antibody elicits a negligible immune response when administered to a human compared to the immune response mounted by a human against a rat or mouse antibody.
Alternatively, the antibody may be a chimeric antibody. A chimeric antibody comprises an antigen binding region and a non-immunoglobuiin region. The antigen binding region is an antibody light chain variable domain or heavy chain variable domain. Typically, the chimeric antibody comprises both light and heavy chain variable domains. The non-immunoglobulin region is fused as its C-terminus to the antigen binding region. The non-immunoglobulin region is typically a non-immlmoglobulin protein and may be an enzyme region, a region derived ftom a

5 protein having known binding specificity, ftom a protein toxin or indeed ftom any protein expressed by a gene. The two regions of the chimeric antibody may be connected via a cleavable linker sequence.
The antibody may be a human IgG such as IgG 1, IgG2, IgG3, IgG4, IgM, io IgA, LgE or 1gD carrying rat or mouse variable regions (chimeric) or CDRs (humanized). Primatizing techniques may also be used.
The binding agents of the present invention may be used alone or in combination with immunosuppressive agents such as steroids, cyclosporin, or antibodies such as an anti-lymphocyte antibody or more preferably with a tolerance-inducing, anti-autoimmune or anti-inflammatory agent such as a CD4 ' T cell inhibiting agent, e.g. an anti-CD4 antibody (preferably a blocking or non-depleting antibody), an anti-CD8 antibody, a TNF antagonist e.g. an anti-TNF antibody or TNF inhibitor e.g. soluble TNF receptor, or agents such as NSAIDs.
The binding agent will usually be supplied as part of a sterile, pharmaceutically acceptable composition. This pharmaceutical composition may be in any suitable form, depending upon the desired method of administering it to a patient. It may be provide in unit dosage form and may be provided as part of a kit. Such a kit would normally (although not necessarily) include instnlction for use.

6 Binding agent administrations are generally given parenterally, for example intravenously, intramuscularly or sub-cutaneously. The binding agents are generally given by injection or by infitsion. For this purpose a binding agent is formulated in a pharmaceutical composition containing a pharmaceutically acceptable carrier or diluent. Any appropriate carrier or diluent may be used, for example isotonic saline solution. Stabilizers may be added such as a metal chelator to avoid copper-induced cleavage. A suitable chelator would be ED'1'A or sodium citrate. 'hhey may be given orally or nasally by means of a spray, especially for treatment of respiratory disorders. They may be formulated as creams or ointments, especially for use in io treating skin disorders. They may be formulated as drops, or the like, for administration to the eye, for use in treating disorders such as vernal conjunctivitis.
For injectable solutions, excipients which may be used include, for example, water, alcohols, polyols, glycerine, and vegetable oils.
15 The pharmaceutical compositions may contain preserving agents, solubilizing agents, stabilizing agents, wetting agents, emulsifiers, sweeteners, colorants, odorants, salts (substances of the present invention may themselves be provided in the form of a phannaceutically acceptable salt), buffers, coating agents or antioxidants. They may also contain other therapeutically active agents.
Suitable dosages of the substance of the present invention will vary, depending upon factors such as the disease or disorder to be treated, the route of administration and the age and weight of the individual to be treated. Without being bound by any particular dosages, it is believed that for instance for parenteral administration, a daily dosage of from 0.01 to 50 mg/kg of a binding agent of the present invention (usually present as part of a pharmaceutical composition as indicated above) may be suitable for treating a typical adult. More suitably the dose

7 might be 0.05 to 10 mg/kg, such as 0.1 to 2 mg/kg. This dosage may be repeated as often as appropriate. Typically administration may be 1 to 7 times a week.
If side effects develop the amount and/or frequency of the dosage can be reduced.
A
typical unit dose for incorporation into a pharmaceutical composition would thus be at least 1 mg of binding agent, suitably 1 to 1000 mg.
BRIEF DESCRIPTION Oh' TIDE DRAW1NGS
io The invention will be better understood from the following detailed description given with reference to the accompanying drawings.
Fig. 1 and Fig. 2 are graphical representations which show 1 OG2 and B61-i 12 mAbs recognizing CD47 antigen. Untransfected COS and COS cell lines transiently i5 expressing CD47 Ag (Fig. 1) or stable CHO-transfected cell line with CD47 cDNA
(Fig. 2) were stained with either 1 OG2 (Fig. l and 2) or B6H 12 (Fig. 2) mAbs or isotype-matched cont mAbs as described in materials and methods. One representative experiment out of 3.
2 o Figures 3 to 5 are graphical representations which show the measurements of IFN-y production by autologous monocytes. T cells ( 1 X l 0''/ml) were cultured in the presence of autologous monocytes (2.Sx105/ml) and stimulated by LL-2 (50 U/ml) (Fig. 3), LL-2 (2 U/ml) plus sCD23 (25 ng/ml) (Fig. 4) or IL-:15 (100 ng/ml) plus sCD23 (25 ng/ml) (Fig. 5) in the presence or absence of anti-CD47 mAbs used 2 5 at 5 ~.g/ml final concentration. IFN-y production was measured in the CSN
by RIA
after 3 days culture. Data represent mean ~ SE1VI of 5 experiments (p<0.001 ).

8 Figures 6 is a graphical representation which shows that anti-CD47 mAbs suppressed in a dose-dependent manner IL-2 plus sCD23 induced IFN-y production.
Cultures of T cells and monocytes were stimulated with IL-2 ( l 0 U/ml) and sCD23 (25 ng/ml) in the presence of various concentration of anti-CD47 tnAb (clone 1 OG2) (Panel A), F(ab')2 fragments of clone B6H 12 (Panel B) or different anti-CD47 mAbs (Panel C). IFN-'y production was measured after 3 days cultures. Data is one representative experiment out of 3.
Figures 7 to 9 are graphical representations which show Anti-CD47 mAbs i o suppressed IL- I 2 producti on. T cells ( 1 X 10~/ml) were cocultured with autologous monocytes (2. S X 105/ml) as described in Fi gures 3-5 . After 3 days cultures, I L- l 2 p40 release was measured in the culture supernatant by ELISA. Data represent mean ~ SEM of 5 (Fig. 7), 4 (Fig. 8) and 1 out of 3 experiments (Fig. 9) (p<0.01 ).
Figures IO is a graphical representation which shows that Anti-CD47 mAbs suppressed IL-l 2 plus sCD23-induced IFN-y production. T cells (l X 10''/ml) were cultured with autologous monocytes (2.SX105/ml) and stimulated by IL-l 2 (40 pM) and sCD23 (25 ng/ml) in the presence or absence of anti-CD47 mAbs (2.5 ~.g/ml).
IFN-~y production was measured in the CSN after 3 days culture. Data represent a o mean ~ SD of 4 experiments (p<0.0 I ).
Figures I 1 is a graphical representation which shows that F(ab')2 and Fab fragments of anti-CD47 mAb suppressed IFN-y production. T cells ( 1X 106/ml) were cocttltt~red with monocytes (2.SX 105/ml) and stimulated by IL-2 (20 U/ml) with or without F(ab')2 or Fab fragments of anti-CD47 mAb (clone B6H 12). IFN-y was measured after 3 days of culhtre. Shown is one representative experiment out of 2.

9 Figures 12 is a graphical representation which shows that clones 1 OG2 and B6H.12 recognize different CD47 epitopes. .Iurkat T cell line (Panel A) and Tl-monocyte cell line (Panel B) were stained with various concentrations of clone l OG2, B6H l2 tnAbs and isotype-control matched mAbs as described in materials and methods.
Figures 13 is a graphical representation which shows the cellular distribution of l OG2 and B6H12 antigens on dendritic cells and erythrocytes. Erythrocytes (Fig.
13a) and dendritic cells (Fig. 13b) were stained with either 1 OG2 or B6H 12 mAbs i o as described in materials and methods. One representative experiment out of 3 .
Figures 14 is a graphical representation which shows that anti-CD47 mAbs suppressed TNF-a production by purified monocytes. Enriched monocytes (2X 105/ml) were stimulated by sCD23 (25 ng/ml) or LPS ( 10 ~.g/ml) in the presence i5 or absence of 2.5 pg/mI anti-CD47 mAbs (clone 1 OG2 or B6HI2). After overnight culture, TNF-a, IL-1 Vii, IL-8 and PGE2 were measured in the CSN by ELISA.
Data represent mean, ~ SD of 8 (Panel A) and 3 experiments (Panel B) (p<0.001 ).
Figures 15 is a graphical representation which shows that sCD23 ao costimulates IL-2 or IL,-15-induced IL-12 p40 release. T cells (106/ml) were cultured with autologous monocytes (2XIOShnl) and stimulated with IL-2 (50 U/ml) or IL-(100 ng/ml) in the presence or absence of sCD23 (25 ng/ml). Anti-CD47 or isotype-control matched mAb (5 ~ g/ml) were added to the cultures. After 3 days culhtre, IL-12 p40 release was measured in the CSN. Data represent mean ~ SD
of 25 4 experiments.

Figures 16 is a graphical representation which shows that anti-CD47 mAb suppresses IL-l2 p75 production induced by T-cell dependent or independent costimulatory signals. Monocytes (106/ml) were stimulated with SAC alone, SAC
plus IFN~y or sCD40L plus IFN~y and GM-CSF in the presence of anti-CD47 or 5 control mAb. After overnight culhire, IL-12 p75 release was measured in the CSN.
Data represent mean ~ SD of 7 experiments.
Figm-es 17 is a graphical representation which shows the effect of anti-CD47 mAb on SAC or SAC plus IFNy-induced monokine release. Monocytes (10~/ml) s o were stimulated with SAC plus 1 FN~y in the presence of anti-CD47 or isotype control-matched mAb. Monokine release (i.e. 1L-1 (3, 1L-6, '1'NFa and 1L-10) were measured in the CSN after overnight culture. Data represent mean ~ SD of 5 experiments.
Figures 18 is a g~rapllical representation which slows that anti-CD47 mAb suppresses Ag-dependent T cell LFN-y response. Purified T cells ( l X 1 U~'/ml) were stimulated by soluble anti-CD3 mAb (clone 64.1 ) plus I L; 2 (25 U/ml) or I L-12 (60 PM) with or without anti-CD47 mAb (5 tlg/ml). 3H thymidine uptake was measured during the last 16 hrs of 5 days culhare and CSN were collected for the measurement of IFNy production. Data represent mean ~ SD of 3 experiments (p<0.05).
Figures 19 is a graphical representation which shows that anti-CD47 mAbs suppresses allogeneic mixed lymphocyte reaction. T cells (0. S X 106/ml) were coculhired with allogeneic mitomycin C-treated dendritic cells (0.3XI05/ml) in a 5 well U-bottom plate in the presence or absence of anti-CD47 mAb (5 ~
g/ml). 'H-thymidine uptake was measured after 5 days culture. Data are one representative experiment out of 3.

Fig~ires 20 is a graphical representation which shows that anti-CD47 mAbs specifically suppresses IgE synthesis with no effect on B cell proliferation.
Tonsillar B cells ( 1 X 106/ml) were stimulated by soluble CD40-ligand (sCD40L) ( 1 tlg/ml) and IL-4 (10 ng/ml) in the presence or absence of anti-CD47 mAbs (5 ~g/ml) (clone I OG2 or B6H12). 3H-thymidine uptake (B-cell proliferation) was measured after days culture and IgE production after 14 days. Data represent mean ~ SD of 8 experiments (p<0.001 ).
1o DETAILED DESCRIPTION OF THE INVENTION
Production of mAbs Clone l OG2 secreting anti-CD47 antibodies (IgM class) has been produced according to conventional procedures such as described by Kohler and Milstein (Nature ( 1975) 256:495-497). Accordingly, a non-IgG secreting mouse myeloma cell line (NSI) rendered azag~.ianine resistant are ftised to spleen cells from immunized mice with Jurkat T cell line to obtain hybrid cells that produce large amounts of monoclonal antibody. This method employed polyethylene-glycol (PEG) as the fusing agent followed by selection in HAT medium (hypoxanthine, 2 o aminopterin and thymidine). Screening of mAbs was performed according to their "anti-inflammatory biological activity: i.e. inhibition of IFN-y response" in T
cells/monocytes coculture system in the absence of TCR engagement.
Cell separation and culture conditions Morrocyte.s: PBMC were isolated by density gradient centrifugation of heparinized blood from normal healthy volunteers using Lymphoprep (Nycomed, Oslo, Norway). Monocytes were prepared as described (2). Briefly, 1'BMC were resuspended at 50 X 106 cells/ml in RPMI 1 G40 containing 10% FCS
(BioWittaker, Inc., Walkersville, MD) and incubated 40 min at 4°C under rotation (to allow aggregation of monocytes) followed by 10 min incubation on ice. Pellets of aggregated enriched monocytes were further separated from non-aggregated PBMC
by a gradient of FCS and another 10 min incubation on ice. Enriched monocyte preparations were further depleted in T and/or NK cells by rosetting with S-(2-aminoethyl) isothiouronium bromide (Aldrich Chemical Co., Milwaukee, W I) treated sheep red blood cells (AET-SRBC). Monocyte purity was shown to be >95% by flow cytometry (FACScan, Becton Dickinson) using phycoerythrin-1 o conjugated anti-CD 14 mAb (Becton Dickinson). For some experiments, monocytes were positively selected according to CD14 expression by means of a FACSort (Becton Dickinson), and monocyte purity was >99% CD 14' cells. Cellular viability was >90% using trypan blue exclusion.
15 T cells: Enriched T cell populations were obtained ftom the monocyte-depleted PBMC by rosetting with AET-SRBC and treatment with ammonium chloride. To obtain highly pm-ified T cells, rosette forming cells were washed and incubated for 20 min at 37°C in Lympho-Kwik T (One Lambda, Los Angeles, CA). Cell purity was assessed by flow cytometry (FACScan, Becton Dickinson) using phycoerythrin-2o conjugated anti-CD3 mAb (Becton Dickinson) and shown to be >98% in all the cases.
All cultures were performed in complete seem-free HB 1 O1 medium (lrvine Scientific, Santa Ana, CA) supplemented with 2 mM glutamine, 1 W M sodium 25 pymvate, 10 mM I~epes, l00 LU penicillin and I00 yg/ml streptomycin. When cultured alone, monocytes were incubated in Sml sterile Falcon tubes (Becton Dickinson, Lincoln Park, NJ) at 2 X 105 cells/ml for cytokine measurement in the presence of polytnyxin B ( 10 Itg/ml) (Sigma Chem., St. Louis, MO). For coculture experiments, T cells ( 106 cells/ml) were incubated with monocytes or B cells (2 X
105 cells/ml) in 24-well Falcon plates.
Reagents Human recombinant IL-2 was kindly provided by Dr. D. Bron (Instihtt Bordet, Brussels, Belgium). IL-4 and soluble CD40L, was a gift from Immunex (Seattle, WA), IL-10 was received from Dr. K. Moore (DNAX, Palo Alto, CA), 1L-12 was a generous gift from Dr. M. Gately (HofFmann-La Roche, Nutley, NJ) and 1 o used at 40 pM. Endotoxin-free (< 15 pg/ml as determined by the chromogenic Limulus amebocyte lysate, QCL-1000, BioWhittaker Inc., Walkersville, MD) affinity-purified sCD23 was prepared in our laboratory from CSN of CHO cell line transfected with human cDNA encoding for as 148 to 321 of the CD23 molecule.
The concentration of 25 ng/ml sCD23 used throughout this shtdy was selected on the basis of previously reported dose-response curves. Recombinant TNFa was kindly provided by Dr. W. Fiers (State University, Ghent, Belgimn). B6H I 2 mAb was purchased, at the ATCC (clone HB-9771: U.S. Pa.t. 5,057,604).
Cytotluorimetric analysis 2 o Immlunofluorescence was performed on various cells and cell lines according to standard techniques using both anti-CD47 mAbs in the presence of normal human Igs (150 gg/ml). PE-conjugated streptavidin were obtained from Ancell and biotinylated goat anti-human IgG + IgM was purchased from Tago. After staining, cells were analyzed with a FACScan (Becton Dickinson & Co.).

Lymphokine determinations IFN-'y and IL-10 were measured by a sandwich solid-phase RIA. Anti-LFN-y mAb (clone 42.25) was used to coat the solid phase and'25I-labeled anti-IFN-y mAb (clone KM48, purchased from Dimension Labs. lllC., Mississauga, Ont., Canada) as detecting probe; anti-I I~ 10 clone 9D7 for coating and anti-I L-10 clone 1268 for labelling. TNF-a was assessed using a sandwich ELISA employing mouse mAb to human TNF-a (clone T 144. B, kindly provided by Dr. T. Nakajima, St. Marianna University School of Medicine, Kawasaki, Japan) and a polyclonal rabbit anti-TNF-a received from Dr. J. Tavernier (Roche Research Institute, Ghent, 1o Belgium). IFN-'y, 1L-10 and TNF-a assays were calibrated against international standards obtained from the National Institute of Biological Standards and Control (Hertfordshire, England). The detection limit for the IFN-y and IL-10 RLA is pg/ml and is 45 pg/ml for the TNF-aELLSA. LL-1 (3 was measured by ELISA kits purchased from R & D Systems. IL-12 p40 and IL-12 p75 were measured by a 15 two-site sandwich ELISA employing clone 2.4 A 1 or clone 20C2 as capture mAbs and clone 4D6 as second mAb. Samples were analyzed in serial 5 fold dilutions in duplicate; the sensitivity of the assay is 10 pg/ml.
Immunoglobulin determinations a o IgE was measured by sandwich radioimmlu~oassays (RIA). Clone 89 (mAb anti-IgE) was used as coating mAb and ' 25I clone 4.15 (anti-IgE) mAb as detecting probe; sensitivity of the assay was < 150 pg/ml.
Expression cloning of molecule recognized by lOG2 mAb 25 A cDNA library was prepared from Jurkat T cell line, that expressed high level of lOG2 epitope, according to the method described by Seed et al (Proc.Natl.Acad.Sci. USA ( 1987) 84:3365-3369). Briefly, the cDNA library was used to transfect COS cells by DEAF-dextran method. After 3 days transfection, COS cells were harvested and incubated with 1 OG2 mAb at 4°C for 1 h.
Unbound mAb was removed by washing and the COS cells were incubated in petri dishes coated with goat anti-mouse IgM antibodies. After 2 hrs, the unbound cells were 5 extensively washed with PBS; the cells adhering to the plates were lysed and the episomal DNA was prepared. The cDNA was used to transform bacteria. The antibiotics resistant colonies were amplified a.nd pooled. Plasmids were prepared from pools of 50 colonies and used to transfect COS cells. Single colonies from the positive pools were amplified and their plasmids were tested for their ability to 1o transfer lOG2 epitope into COS cells. The positive clone was subsequently cloned and cDNA was sequenced.
Statistical analysis Paired Shident's t test have been used to assess level of significance (*<0.05;
15 **p<0.01, ***p<0.001 ) Explanation and significance of the foregoing results are as follows:
ao Clone lOG2 mAb is directed against CD47 antigen.
1t has been previously reported that soluble CD23 activated monocytes to contribute to the antigen-independent stimulation of T cells (2). During the screening for mAbs that might regaate this bystander T cell response, the applicant found clone lOG2, this clone secreting antibodies having anti-inflammatory properties. Using mammalian vector expression cloning method and 1 OG2 mAb, the cDNA encoding CD47 Ag was cloned from Jurkat T cell line cDNA library. The CD47 cDNA was transiently expressed in COS 7 cell line. As shown in Fig. l , lOG2 mAb strongly reacted with CD47-transfectants with no staining of untransfected cell lines. Next, the applicant prepared stable CD47 transfectants in CHO cell line and found similar pattern of staining with clone l OG2 and B6H

mAb (a commercially available CD47 mAb binding to CD47 Ag) mAbs establishing that lOG2 recognized CD47 Ag (Fig. 2).
Anti-CD47 mAbs suppressed I L.-2 and I L-1 S-induced 1 F'N-y production in the presence or absence of sCD23.
As shown in Fig. 3-S, anti-CD47 mAbs (clone l OG2 and B6H 12), inhibited 1o IL-2 and IL-15 stimulated 1FN-y response not only in the presence but also in the absence of sCD23. The inhibitory effect of anti-CD47 mAbs is dose-dependent (Fig.
6). Interestingly, thrombospondin, the naW ral ligand of CD47, similarly suppressed IL-2-induced IFN~y production in T/monocytes cocultures (Table Il). The applicant has previously demonstrated that the IL-2 or IL-15 induced LFN-y production was strictly dependent on CD40-CD40L interactions and on endogenous IL- I 2 production as shown by the inhibitory effects of both anti-IL-12 and anti-Abs on IFN-'y, production and of anti-CD40L mAb on IL-12 release (3). The applicant therefore examined the efFect of these anti-CD47 mAbs on 1L-12 secretion. The data in Fig. 7-9 indicated that, like anti-CD40L mAb, both anti-2o mAbs strongly suppressed IL-2 or IL-15-induced IL-12 production; however, by contrast to anti-CD40L mAb, addition of exogenous II~ 12 failed to restore IFN-y production (Fig. 10). Taken together, these data indicated that anti-CD47 mAbs inhibited IFN-y production not only by reducing 1L-12 release belt also by diminishing T cell response to IL-12.
To further analyze the mechanisms of inhibitory activity of anti-CD47 mAbs in T cells/monocytes coculW res, the applicant prepared F(ab')2 and monovalent Fab fragments of CD47 mAb. As shown in Fig. 11, divalent or monovalent ftagments of CD47 mAb significantly suppressed IFNy response suggesting that the anti-CD47 mAb mediated its activity by either delivering a negative signal to the cell through the cross-linking of CD47 Ag via its divalent Fab (not via its Fc fragment bound to FcyR) or by inhibiting the interactions between CD47 and its natural ligand, the thrombospondin-derived macrophages.
Cellular distribution of Ag binding to AIM mAbs.
The applicant next examined a panel of cell lines for their reactivity to clone 1 OG2. As shown in Table I, clone IOG2 reacted to most of the cell lines (T, B, monocytic and erythroleukemia cell lines) with tle exception of THP 1 monocytic cell line exclusively stained by BH612 mAb and not by 1 OG2 mAb (Fig. 12).
Both anti-CD47 mAbs stained all leukocytes (T, B and macrophages) (Table I);
erythrocytes (Fig. 13a) and dendritic cells (Fig. 13b) are reacting with BI-i612 but s 5 not with 1 OGZ mAb. All together, these results strongly suggested that both mAbs reacted with different epitopes or different molecular forms of a common antigen.
Anti-CD47 mAbs inhibit inflammatory mediators release by monocytes.
IFN-'y and TNF-a are directly implicated in the pathogenesis of chronic a o inflammatory disease as shown by in vivo studies using neutralizing mAbs.
Monocyte-dependent T cell IFN-'y production involved TNF-a and IL-12 production as well as interactions between costimulatory surface molecules (CD40-CD40L; LFA3-CD2, B7-B7 ligands). The applicant therefore examined the reg~~latory activity of anti-CD47 mAbs on monokine release by pm-ified monocytes.
25 Bacterial stimuli (i.e. lipolysaccharides (LPS) or staphylococcus aureus Cowan 1 (SAC) or sCD23 were used to trigger TN F-a, IL-1 ~3 or IL- I 2 production. The data in Fig. I4 and table IV indicated that anti-CD47 mAbs strongly inhibited sCD23 induced TNF-a release, without affecting LPS induced TNF-a production.
Similarly, sCD23-induced 1L-I ~3 and PGE2 production were suppressed by anti-CD47 mAbs. Although sCD23 did not trigger IL- I 2 release by purified monocytes, it costimulated IL-12 production in T cells/monocytes cocultures system. As shown in Fig. 15, CD47 mAb strongly decreased sCD23 costimulatory activity on IL-12 secretion. Most sfikingly, CD47 mAb also suppressed IL-l 2 production by purified monocytes stimulated by T-cell independent (i.e. SAC) or dependent signals (i.e.
sCD40L) (Fig. 16). Of interest, thrombospondin significantly reduced SAC and IFN~y-activated IL-12 p70 release by monocytes (Table III). As reported for LPS-1o induced TNFa, CD47 mAb failed to inhibit SAC plus LFNy induced TNFa release (Fig. 17). The SAC-induced secretion of other monocyte products (i.e. IL1~3, IL-6, 1L-10) remained largely unaffected.
Taken together, these results indicated that anti-CD47 mAbs displayed potent suppressing activity on inflammatory mediators release underlying their inhibitory effect on IFN-y production.
Anti-CD47 mAbs suppress IL-12 and anti-CD3-induced IFN-y production and allogeneic mixed lymphocyte reaction.
a o The applicant next examined the biological activity of anti-CD47 mAbs on Ag-dependent T cell stimulation. As depicted in Fig. 18, anti-CD47 mAbs inhibited anti-CD3-induced T cell proliferation and IFN-y production by purified T
cells.
Similar selective suppression of anti-CD3-induced T cell response to LL-12 were obtained using purified CD4 or CD8 subpopulations (not shown). Of interest, pokweed mitogen-induced IFN-y is also suppressed by anti-CD47 mAbs (data not shown). The inhibitory activity by anti-CD47 mAbs of Ag-dependent T cell activation was also observed in mixed lymphocyte reaction (MLR). Irradiated allogeneic non-T cells enriched populations or purified dendritic cells were used as allostiinulators of adult peripheral purified CD4+ T cells. As shown in Fig.
19, anti-CD47 mAbs inhibited primary mixed lymphocyte reaction as measured by 3H
thyrnidine uptake; the applicant also tested the effect of mAbs on secondary stimulation, and found that triggering of CD47 Ag in primary cultures lead to a state of hyporesponsiveness of T cells in secondary cultures (not detailed).
Anti-CD47 mAbs specifically suppress IgE synthesis without significantly affecting B cell proliferation.
1o Since resting B cells strongly express CD47 Ag, the applicant examined the effect of CD47 ligation on B cell proliferation and differentiation. Purified tonsillar B cells were stimulated by trimeric soluble CD40L (sCD40L) in the presence of IL-4. As shown in Fig. 20 and table V, anti-CD47 mAbs did not interfere with B
cell proliferation as measured by 3H thymidine uptake at day 5, while they strongly 1 s inhibited in a dose-dependent manner (not shown) II~ 4-induced IgE
synthesis.
Moreover, anti-CD47 mAbs were likely to block IgE-class switching since they also suppressed 1gE synthesis by 1L-4-stimulated naive ( sIgM' slgD' ) B cel Is and inhibited the expression of germ line E transcripts (not shown). The inhibitory effect was not reversed by addition of neutralizing mAbs to TGF-(3, a potent inhibitor of a o Ig synthesis nor by IL-6 (not shown). Taken together, anti-CD47 mAbs appeared as strong anti-inflammatory agents since they also interfered with the humoral response of the allergic reaction.
IL-I2 is a potent proinflammatory and immunoregulatory cytokine which 25 plays a crucial role in innate and adaptive Th 1 response. IL-12 is released during the early stage of infection caused by a large variety of bacteria, intracellular pathogens, fungi and certain viruses. IL-12 is also produced in the absence of infection, following interaction of CD4+ T and dendritic cells/monocytes. It has been reported that Thl response associated with chronic or autoimmune disease remains IL-12-dependent. Indeed, IL-12 neutralization (by anti-IL-12 Ab) reveals to be an effective treatment of experimental bowel disease, auto-immure 5 encephalitis and insulin-dependent diabetes.
Therefore, the inhibition by CD47 ligation of proinflammatory cytokine (including IL-12) production by monocytes and IL-12 responsiveness by T cells could permit to downregulate inflammatory response on which new therapeutic io strategies to chronic disorders could be based.

REFERENCES
1. Brocke S., K. Gijbels, M. Allegretta, I. Ferber, C. Piercy, T.
Blankenstein, R. Martin, U. Utz, N. Karin, D. Mitchell, T. Veromaa, A. Waisman, A. Gaur, P. Conlon, N.
Ling, P.J. Fairchild, D.C. Wraith, A. O'Garra, C.G. Fathman, and L. Steinman.
1996.
Treatment of experimental encephalomyelitis with a peptide analogue of myelin basic protein. Nature. 379:343-346.
2. Armant M. , M. Rubio, G. Delespesse, and M. Sarfati. 1995. Soluble CD23 directly activates monocytes to contribute to the antigen-independent stimulation of resting T cells.
J.Immunol. 155:4868-4875.
3. Armant M. , R. Armitage, N. Boiani, G. Delespesse, and M. Sarfati. 1996.
Functional CD40 ligand expression on T lymphocytes in the absence of T cell receptor engagement:
involvement in interleukin-2-induced interleukin-12 and interferon-y production.
Eur.J.Immunol. 26:1430-1434.
4. Stewart M. , M. Thief, and N. Hogg. 1995. Leukocyte integrins. Curr.Opin.
in Cell Biol.
7:690-696.
5. Diamond M. S. , and T. A. Springer. 1994. The dynamic regulation of integrin adhesiveness. Curr.Biol. 4:506-517.

' CA 02226962 1998-02-16 ~a 6) Brown E. , L. Hooper, T. Ho, and H. Grresham. 1990. Integrin-associated protein: A 50-kD plasma membrane antigen physically and functionally associated with integrins. J. Cell Biol. 111:2785-2794.
7. Blystone S.D., F.P. Lindberg, S.E. LaFlamme, and E.J. Brown. 1995. Integrin cytoplasmic tail is necessary and sufficient for regulation of asf31 phagocytosis by a~133 and integrin-associated protein. J.Cell Biol. 130:745-754.
8. Felding-Habermann B. , and D. A. Cheresh. 1993. Vitronectin and its receptors.
Curr.Opin. in Cell Biol. 5:864-868.
9. Gladson C. L. , and D. A. Cheresh. 1994. The av integrins. In: Integrins:
The biological problems. CRC Press Page 83-99.

10. Halvorson M.J., J.E. Coligan, and K. Sturmhofel. 1996. The vitronectin receptor (aVB3) as an example for the role of integrins in T lymphocyte stimulation. Immunol.
Res. 15:16-29.

11. Friedlander M. , P. C. Brooks, R. W. Shaffer, C. M. Kincaid, J. A. Varner, and D. A.
Cheresh. 1995. Definition of two angiogenic pathways by distinct a~ integrins.
Science.
270:1500-1502.

~3

12. Brooks P. C. , A. M. P. Montgomery ) M. Rosenfeld, R. A. Reisfeld, T. Hu, G. Klier, and D. A. Cheresh. 1994. Integrin a~1~3 antagonists promote tumor regression by inducing apoptosis of angiogenic blood vessels. Cell. 79:1157-1164.

13. Lindberg F.P., H.D. Gresham, E. Schwarz, and E.J. Brown. 1993. Molecular cloning of integrin-associated protein: An immunoglobulin family member with multiple membrane-spanning domains implicated in a~f33-dependent ligand binding. J.
Cell Biol.
123:485-496.

14. Mawby W.J., C.H. Holmes, D.J. Anstee, F.A. Spring, and M.J.A. Tanner.
1994.
Isolation and characterization of CD47 glycoprotein: a multispanning membrane protein which is the same as integrin-associated protein (IAP) and the ovarian tumour marker OA3. Biochem J. 304:525-530.

15. Zhou M. , and E.J. Brown. 1993. Leukocyte response integrin and integrin-associated protein act as a signal transduction unit in generation of a phagocyte respiratory burst.
J.Exp.Med. 178:1165-1174.

16. Cooper D. , F. P. Lindberg, J. R. Gamble, E.J. Brown, and M. A. Vadas.
1995.
Transendothelial migration of neutrophils involves integrin-associated protein (CD47).
Proc. Natl. Acad. Sci. USA. 92:3978-3982.

L/

17. Parkos C.A., S.P. Colgan) T.W. Liang, A. Nusrat, A.E. Bacarra, D.K.
Carnes, and J. L. Madara. 1996. CD47 mediates post-adhesive events required for neutrophil migration across polarized intestinal epithelia. J. Cell Biol. 132:437-450.

18. Gao A.-G., F.P. Lindberg, M.B. Finn, S.D. Blystone, E.J. Brown, and W.A.
Frazier.
1996. Integrin-associated protein is a receptor for the C-terminal domain of thrombospondin. J.Biol.Chem. 271:21-24.

19. Lindberg F.P., D.M. Lublin, M.J. Telen, R.A. Veile, Y.E. Miller, H. Donis-Keller, and E.J. Brown. 1994. Rh-related antigen CD47 is the signal-transducer integrin-associated protein. J.Biol.Chem. 269:1567-1570.

Table 1 Cellular distribution : of lOG2 antigen on human cell lines Freshly Human cell lines isolated human leucocytes sCD23 lOG2 sCD23 lOG2 T cells +++ ++++ T cell lines ++ ++++

(Jurkat, CEM, HUT 78) B cells ++ ++ B cell lines ++ ++

(RPMI 822f, Raji, WIL-2, Daudi) Monocytes + + Monocyte cell lines ++ ++

( 11937) THP-I - _ Erythrocytes- - Erythroleukemia cell lines + +

(K5621 KS62-CR2) a~

U

O

U
O ~ M

M V1 0~0 U

O

z U w E~

N

1~ I~ O v0 b G ~ M ~ c o E~

H

U

N
z + d' N O ~ M
p 00 ~ N O M

+

H

O

O

o z U
A'' ~ N M wt V'1 W v~

. CA 02226962 1998-02-16 w O

O O O
...

4 .

F, O

G

_ b o E

~ o g N ~ 0~0 .-r aw ~ v p', 0 0 w ~ N
H

O ~ ~3 z N ~ O O

cd 4, _ w ~ pp O O
o ..-) 3 '-' CV 1b4 O

~o .-i O

G
~! .., 'G

~ G

N M

TABLE IV (Appendix to I~g. 14a) Effect of lOG2 and B6H12 mAbs on sCD23-induced TNFa production by purified monocytes.
TNFa (pg/ml) Cont mAb lOG2 Exp. 1 993 347 Exp. 2 1392 483 Exp. 3 396 110 Cont mAb B6H12 Exp. 4 845 347 Exp. 5 1177 166 Exp. 6 1152 252 Monocytes were stimulated overnight with sCD23 (25 ng/ml) in the presence of anti-CD47 mAbs (clone l OG2 or B6H 12) or isotype-matched cont mAbs. TNFa was measured in the CSN by specific ELISA.

TABLE V (Appendix to Fig. 20) Effect of lOG2 and B6H12 mAb on IIr4-induced IgE synthesis by CD40-activated B
cells.
3H-Thymidine Uptake (X103 CP1V1) IgE (ng/ml) A Cont mAb lOG2 Cont mAb 1062' Exp. 1 2.8 3.9 51 26 Exp. 2 33 42 31 15 Exp. 3 47 48 79 28 B Cont mAb B6H12 Cont mAb B6H12' Exp. 4 4.8 4.5 23 4.5 Exp. 5 87.1 60.4 97 18 Exp. 6 77.1 63.1 80 27.5 lOG2 mAb and B6H12 mAbs are used at 20 ~g/ml and 10 ~g/ml respectively.