MXPA00010681A - Binding molecules derived from immunoglobulins which do not trigger complement mediated lysis - Google Patents

Abstract

Disclosed are binding molecules which are recombinant polypeptides comprising:(i) a binding domain capable of binding a target molecule, and (ii) an effector domain having an amino acid sequence substantially homologous to all or part of a constant domain of a human immunoglobulin heavy chain;characterised in that the binding molecule is capable of binding the target molecule without triggering significant complement dependent lysis, or cell mediated destruction of the target, and more preferably wherein the effector domain is capable of specifically binding FcRn and/or Fc&ggr;RIIb. These are generally based on chimeric domains which are derived from two or more human immunoglobulin heavy chain CH2 domains. In preferred embodiments the regions 233-236, and 327-331, are modified, as are further residues to render the molecule null allotypic. The binding domain may derive from any source appropriate to the (usually clinical) application for the molecule and may be from e.g. an antibody;an enzyme;a hormone;a receptor;a cytokine or an antigen;a ligand and an adhesion molecule. Also disclosed are nucleic acids, host cells, production processes and materials, and uses e.g. to inhibit B cell activation;mast cell degranulation;phagocytosis, or to inhibit the binding of a second binding molecule to the target molecule. Pharmaceutical preparations are also disclosed.

Description

LINKED MOLECULES DERIVED FROM IMMUNOGLOBULINS WHICH DO NOT BEGIN THE MEDIATED LYSIS BY THE COMPLEMENTS TECHNICAL FIELD The present invention relates to binding polypeptides having amino acid sequences derived from a constant, modified region of heavy chain immunoglobulin G (IgG). The invention further relates to methods and materials for producing these polypeptides, and to methods and materials that employ them.

PREVIOUS TECHNIQUE Immunoglobulins Immunoglobulins are glycoproteins which help defend the host against infection. These generally consist of heavy and light chains, the N-terminal domains of which form a variable or a V domain capable of binding the antigen. Domain V is associated with a constant or a C-terminal domain which defines the class (and some REF: 124409 times the subclass [isotype] and allotype [isoalotype]) of the immunoglobulin. Thus in mammalian species, immunoglobulins exist as IgD, IgG, IgA, IgM and IgE. The IgG class in turn exists as 4 subclasses in humans (IgGl, IgG2, IgG3, IgG4). The C domain in the IgGs comprises three C? L, C? 2, and C? 3 domains, which are very similar among these subclasses (over 90% homology). The domains C 1 and C 2 are joined by a joint. The role of subclasses seems to vary between species. It is known that the C domain is responsible for various effector functions of the immunoglobulin (see Clark (1997) "IgG Effector Mechanisms" in "Antibody Engineering" Ed. Capra, Pub. Chem Immunol, Basel, Kurger, Vol 65 pp 88-110 , for a detailed review). In summary, the functions of IgGs are generally achieved by means of the interaction between the Fe region of the Ig and an Fc receptor. (Fc? R) or another binding molecule, sometimes in an effector cell. This can activate the effector cells to destroy the target cells to which the antibodies bind via their variable regions (V). Also, antibodies directed against soluble antigens could form immune complexes which are targeted for FcγRs which results in the uptake (opsonification) of immune complexes or the activation of effector cells and the release of cytokines. In humans, three classes of Fc? R have been characterized, although the situation is further complicated by the occurrence of multiple forms of receptors. The three classes are: (i) Fc? RI (CD64) binds to a monomeric IgG with high affinity and is expressed in macrophages, monocytes, and sometimes neutrophils and eosinophilic cells. (ii) Fc? RII (CD32) binds to a complex IgG with a medium for low affinity and is widely expressed. These receptors can be divided into two important types, the Fc? RIIa and the Fc? RIIb.

The a-form of the receptor is found in many cells involved in destruction (eg, macrophages, monocytes, neutrophils) and appears to be capable of activating the destruction process and occurs as two alternative alleles.

The b 'form seems to play a role in the inhibitory processes and is found in B cells, in macrophages and in mast cells or Mastzelle and eosinophilic cells. In B cells it seems to function to further suppress immunoglobulin production and switching or isotype connection by saying, for example, the IgE class. In macrophages, form b acts to inhibit phagocytosis when mediated through FcγRIIa. In eosinophilic cells and mast cells or Mastzelle, form b can help suppress the activation of these cells through the binding of IgE to its separate receptor. (iii) Fc? RIII (CD16) binds to an IgG with a medium for low affinity and exists as two types. The Fc? RIIIa is found in NK cells, macrophages, eosinophil cells and some monocytes and T cells and mediates ADCC. Fc? RIIIb is highly expressed in neutrophils. Both types have different allotypic forms. As well as the binding to FcγRs, IgG antibodies can activate complement and this can also result in cell lysis, opsonification or cytokine release and inflammation. The Fe region also mediates these properties as the transport of the IgGs to the neonate (by means of the commonly called "FcRn"); the increased life span (also believed to be affected by means of a FcRn type receptor - see Ghetie et ard (1997) Immunology Today 18, 592-598) and self-aggregation. The Fe region is also responsible for the interaction with protein A and protein G (the interaction that appears to be analogous to the FcRn link).

Design or Modification of Immunoglobulins Many of the properties mediated by Fe, discussed above may be desirable in naturally occurring or artificially constructed antibodies. However, there are circumstances where, in particular, the destruction of cells, or the release of cytokines and the resulting inflammation, is inappropriate and undesirable. However, in the same way it may be desirable to retain certain functions mediated by the Faith, for example the long life span of the plasma. It is known that human IgG4, for example, does not activate the complement and human IgG2 does not bind to the high affinity Fc? RI receptor and so that these have previously been used in some situations (the fusion protein of the receptor of TNF was made with the Fe of IgG4). However, no human subclass lacks all the functions of Fe effector activation, relevant or complement activation in all circumstances, possibly due to the existence of various forms of Fc? Rs. Thus, for example, IgG4 can activate antibody-dependent cellular cytotoxicity (ADCC) in some people and IgG2 binds to an allelic form of the FcγRIIz receptor and also activates complement. An alternative approach has been to mutate the sequence of Fe to substitute the crucial residues for a function. Certain target residues have been identified and published (see the magazine by Clark 1997, supra). These include the N-linked carbohydrate bound to the conserved site in the CH2 domain, certain residues in the lower articulation region (eg the ELLGGP sequence) and a proline residue at position 331 and an ExKxK sequence at positions 318-322 . A recent example is described by Cole et al. (1997) Journal of immunology 159, 3613-3621. In that description, residues 234, 235 and 237 were mutated to Alaninas (or in the case of 235, sometimes to Glu). However, these are all unusual residues in these positions in human IgG, in this way the presence of such inappropriate amino acids can make the Fe more immunogenic or antigenic and can also lead to the loss of certain desirable functions of Fe. Again, this strategy has been used for the construction of a therapeutic aglycosylated CD3 antibody (see Routledge et al., 1993 Eur J Immunol 23: 403-411; see also UK PA 9206422.9) and for an inhibitory CD18 antibody. However, a disadvantage here is that the new recombinant constructs have unusual sequences and can be recognized and rejected by the immune system as foreign. Aglycosylated antibodies also lack linkage to the inhibitory FcγRIIb receptor, while maintaining this link may be advantageous for some applications. Other approaches for modifying the immunoglobulins are described in WO 92/16562 (Lynxvale Ltd) which deals with the modification of the allotype of the humanized IgGl antibody CAMPATH1H which has binding affinity for the antigen CD52. The CD52 antigen is found in human lymphocytes and monocytes and has been used as a therapeutic target for the treatment of T and B cell lymphomas and leukemias, immunosuppression of organ and bone marrow transplant recipients and also the treatment of some autoimmune disorders. and related such as rheumatoid arthritis and systemic vasculitis. WO 95/05468 (Lynxvale Ltd) also describes the modification of allotypic determinants in Igs (or derivatives) having the desired binding function or other effector functions. It can be observed from the above that the provision of methods or materials which facilitate the design or modification of the regions of Faith such as reducing the undesired effects, while retaining or increasing the desirable properties, will provide a contribution to the technique.

DESCRIPTION OF THE INVENTION The present inventors have used new combinations of human IgG subclass sequences to generate chimeric polypeptides comprising non-natural, human mimetic Fe sequences, which nevertheless do not activate the complement or activate the cytotoxic activities through 'l Fc? R. At the same time, certain desirable properties of IgG have been retained. For example, the polypeptides do not contain non-human amino acids, and therefore are likely to have reduced immunogenicity. In addition, these are still bound to protein A, which is consistent with being able to cross the human placenta through interaction with the FcRn (neonatal Fe receptor). The manner in which the sequences were developed, and certain demonstrated properties, will be discussed in more detail later. However, in summary, the inventors formulated numerous constructs based on three different IgG sequences (1, 2 and 4). Although the relevant regions of these antibodies share homology, they do not correspond precisely in terms of length, due to this the process of generating derivative sequences which retain activities of the natural sequences is complicated. Constructed antibodies were compared with control antibodies, parenteral in the context of model antigen systems RhD (Fogl) and CD52 (CAMPATH-1H). Surprisingly, a number of sequences were developed with the required combination of activities not found in the molecules of origin. Generally speaking, they contained 1 or more regions or blocks which contained a modification (in general, 2, 3 or 4 amino acids) which was in accordance with the corresponding region of a different subclass. Two particular regions or blocks of interest were 233-236 and 327, 330, 331. Thus, in a first aspect of the present invention there is disclosed a polypeptide binding molecule comprising (i) a binding domain capable of binding a target molecule, and (ii) an effector domain having an amino acid sequence substantially homologous to all or part of a constant domain of a human immunoglobulin heavy chain; characterized in that the binding molecule is capable of binding the target molecule without initiating complement-dependent lysis, significant, or destruction mediated by the target cells, and preferentially, whereby the effector domain is capable of binding specifically FcRn or Fc? RIIb, more preferably both FcRn and Fc? RIIb. The specific binding of FcRn can be evidenced by the ability to specifically bind protein A. In this way, the binding molecules according to the present invention have improved the clinical properties (for example in the context of blocking antibodies). ') This is achieved by the provision of an effector domain derived from Fe which has a reduced affinity for Fc? RI, Fc? RIIa and Fc? RIII, but which retains the ability to bind protein A (and therefore the FcRn, therefore, allows neonatal transport and a long lifespan) and / or the Fc? RIIb In this way, the residues responsible for the linkage of the FcRn in the IgGs do not need to be modified with respect to a natural Fe region. The molecules of the present invention In general, the reduction in affinity that the effector region has for the FcγRI receptor (compared to a Fe region from which it is derived) can, in the preferred embodiments, be of the order of 100. go For certain lower affinity receptors discussed above, the reduction in affinity may be lower for example about 2-10 times, although in the most preferred embodiments it could be as high as 500 times. In general, the corresponding reduction in activity in the chemiluminescence or chemical fluorescence assay (described in more detail below) can be as high as 30-300 fold. The reduced complement activity may be in the order of 50 times. The corresponding figure for the ADCC can be "much higher for example10,000 times However, those skilled in the art will appreciate that the combination of these (reduced) activities may still be of benefit in certain applications, with respect to the precise level of reduction. Although the IgGl / IgG2 and IgG1 / IgG4 chimeras have been prepared in the past (see for example Morgan et al. (1995) Im unology 86: 319-324, or Chappel et al. (1991) Proc Nati Acad Sci USA 88: 9036 -9040, or Greenwood et al. (1993) Eur J Immunol 23: 1098-1104) none of these have been shown to have the combination of properties possessed by the binding molecules of the present invention. The various functions of the linker molecule can be assessed without weight by those skilled in the art, for example by using the methods described below, or methods analogous thereto. For example, the binding properties of FC? R can be assessed directly, or indirectly for example through the inability to activate the chemical fluorescence of monocytes. Specifically, the inability to initiate significant, complement-dependent lysis (which will generally be through reduced affinity for the Clq molecule) can be measured by the release of CR-51 from the target cells in the presence of the components of the complement for example in the form of a serum (as described below) because of that the binding molecule causes less than 5%, preferably less than 2% lysis of specific target cells. mediated by the target cells can be assessed by the release of CR-51 from the target cells in the presence of cytotoxic cells, suitable for example the mononuclear, blood effector cells (as described below) whereby the binding molecule causes less than 5%, preferably less than 2% objective cell lysis.As an alternative for direct measurement, functionality can be inferred for the ability to inhibit these attributes in functional immunoglobulins. For example, by providing a protective effect against complement lysis of cells, or destruction of cells (for example, by ADCC), or by inhibiting the response of monocytes to sensitize cells. In a preferred embodiment of this aspect of the invention, the effector domain comprises an amino acid sequence substantially homologous to the CH2 sequence of human IgG1, G2 or G4, the sequence comprising one or more of the following modifications (substitutions or deletions) of amino acids) in the established positions, numbered with respect to the US numbering system (see Kabat et al. Sequences of proteins of immunological interest., Bethesda, US Department of Health and Human Services, NIH, 1991): Posn Amino Acid 233 P 234 V 235 A 236 (without residue) or G 327 G 330 S 331 S In a preferred embodiment, these substitutions are made in 'blocks' of 233-236 and / or 327,330,331. In this way, the region mutated in the CH2 domain will be 100% homologous to the subclass from which the substituted residues originated, due to which the probability that the region represents an epitope of the B cells or the T cells for the immune system . The various mutant immunoglobulins based on IgG1, IgG2, or IgG4 having the established characteristics have been prepared and shown to have the required properties. Although some of the mutations of individual residues have been prepared in the binding molecules of the prior art, the specific combinations are new as are the functionalities achieved. The preferred forms of the link molecule will now be discussed in more detail: The effector domain The peptide comprises an effector domain having an amino acid sequence substantially homologous to all or part of a constant region of the human immunoglobulin, preferably a C domain of IgG. The numerous sequences for human C regions have been published; see for example Clark (1997) supra. Other sequences for the human immunoglobulin heavy chains can be obtained from the SwissProt and PIR databases "using the Lasergene programming elements (DNAStar Limited, London UK) under the access numbers A93433, B90563, A90564, B91668, A91723 and A02146 for region C of the human Ig? -1 chain, A93906, A92809, A90752, A93132, A02148 for region C of human Ig? -2 chain, A90933, A90249, A02150 for region C of the Ig? -4 chain of human and S23511 for region C of the human Ig? -3 chain.The homology (or identity, or similarity) can be assessed by any convenient method. of the coding nucleotide sequence or the encoded amino acid sequence. "Substantially homologous" is intended to mean that the amino acid sequence comprised shares at least about 50%, or 60% or 70% or 80% homology, in most preferably at least about 9 0% 95%, 96%, 97%, 98% or 99% homology with the reference immunoglobulin. The similarity or homology can be as defined and determined by the TBLASTN program, by Altschul et al. (1990) J. Mol. Biol. 215: 403-10, which is in normal use in the art or, and this may be preferred, the normal BestFit program, which is part of the Wisconsin Package, Version 8, September 1994, (Genetics Computer Group, 575 Science Drive, Madison, Wisconsin, USA, Wisconsin 53711). The BestFit program makes an optimal alignment of the best segment of similarity between two sequences. Optimal alignments are found by inserting separations to maximize the number of matches using the local homology algorithm of Smith and Waterman. This assessment can be made weightless by a person of ordinary skill in the art, in conjunction with the assessment of the required combination of activities, in order to recognize a molecule of the present invention. In addition to having reduced affinity for FcγRI, FcγRIIa, FcγRIIIa and FcγRIIIb, it may be desirable to retain or possess a binding capacity of the 'inhibitory' FcγRIIb receptor to some degree by the molecule effector, and which is preferably higher than its affinity for the Fc? RIIa receptor, and more preferably of equal measure with that of an Ig domain of origin from which it is derived. The results obtained by the present inventors indicate that the binding molecules which they have developed have this property. To date it has not been appreciated in the art that the binding of the Fe to Fc? RIIa and Fc? RIIb regions could be manipulated independently. This capacity can complement the other functions required (as indicated by the ability to bind protein A) in the increase of the therapeutic potential of the binding molecule. In particular, a number of publications have highlighted the important role that Fc? RIIb can play in the inhibition of cellular processes (see Daeron et al., 1995 Immunity 3 (5): 635-46; Van den Herik et al., 1995 Blood 85 (8): 2201-11; Sarmay et al., 1996 I munol Lett 54 (2-3): 93-100; Fong et al., 1996 Immunol Lett 54 (2-3); 83-91; Sarmay et al., 1996 J Biol Chem 271 (48): 30499-504; Unkeless & Jin, 1997 Curr Opin Immunol 9 (3): 338-43; Isakov, 1997 Immunol Res 16 (1): 85-100; Hunter et al., 1998 Blood 91 (5): 1762-8; Malbec et al., 1998 J Immunol 160 (4): 1647-58; Clynes et al., 1999 J Exp Med 189 (1): 179-85). These workers showed that Fc? RIIb, when cross-linked to other receptors, could inhibit signaling of these, inhibiting due to that processes such as the activation of B cells, degranulation of mast cells or Mastzelle and phagocytosis by macrophages. In this manner, the binding molecules of the present invention which retain this activity could be used not "solely to compete with, and competitively inhibit, the undesirable interactions of antibodies-antigens (such as autoantigens or alloantigens), but also to inhibit non-competitively these processes, for example, by additionally preventing the production of autoantibodies or alloantibodies by inhibiting the activation of B cells. Other exemplary applications for this inhibitory effect are discussed below-in relation to therapeutic substances for allergy and asthma (inhibition of degranulation of mast cells or Mastzelle) and anti-RhD molecules (inhibition of phagocytosis). Preferably, the effector domain is itself derived from a human immunoglobulin constant region, more preferably a C domain of IgG.Preferably, the amino acid sequence c omped is substantially homologous to the CH2 sequence (ie, approximately residues 231-340) of human IgGl, G2 or G4, which has the modified amino acids discussed above. More preferred CH2 sequences are shown in Figure 17, particularly those designated Gl? Ab, G2? A, or Gl? Ac, respectively.

Any of these sequences can be combined with (for example, contiguously placed with) the natural or modified CH3 sequences and the natural or modified articulation region, more optionally CH1, in the molecules of the present invention. However, it will be appreciated by those skilled in the art that there is no requirement that other portions of the effector domain (or other domains of the molecule) comprise natural sequences - in particular it may be desirable to combine the sequence modifications described herein with others, for example selected from the literature, only with the condition that the required activities are retained. The skilled person will appreciate that binding molecules comprising such additionally modified effector domains (e.g. as an addition, insertion, deletion or substitution of amino acids) are within the scope of the present invention. Particularly, the sequences of 'null allotypes' can be preferred, such as sequences derived from heavy chains of IgG (see WO 92/16562) wherein the allotypic residues are mutated to match those found in other molecules of the IgG subclass of human. This can minimize the sequences that are seen as strange by any individual.

The binding domain and the target molecule The polypeptide molecule comprises a binding domain capable of binding a target molecule. The linkage domain will have an ability to interact with a target molecule which will preferably be another polypeptide, but can be any target (eg, carbohydrate, lipid (such as phospholipid) or nucleic acid). Preferably, the interaction will be specific. The link domain can be derived from the same source or a different source from the effector domain. For example, while the effector domain will generally be derived from an antibody, the binding domain can be derived from any molecule with specificity by another molecule eg, an enzyme, a hormone, a receptor (cell or circulation link) cytokine or an antigen (which specifically binds an antibody). Preferably, it comprises all or part of an antibody or a derivative thereof, particularly a variable, natural or modified domain of an antibody. In this manner, a binding molecule according to the present invention can provide a binding domain of rodent or camel idae antibodies (see WO 94/25591) and a human immunoglobulin heavy chain discussed above. Also preferred are molecules that have more than one type of the binding domain, such as bispecific antibodies (see for example PCT / US92 / 09965). In these cases an 'arm' is linked to a target cell and the other is linked to a second cell to initiate target destruction. In such cases, it may be desirable to minimize the impact of the effector portion, which could otherwise activate the additional cells which interfere with the desired result. The 'arms' themselves (ie, the binding domain) can be based on the Ig domains (for example Fab) or can be of other proteins as in a fusion protein, as discussed in more detail below. The linker molecule may comprise more than one polypeptide chain in association, for example, covalent or otherwise (eg, hydrophobic interaction, ionic interaction, or linked via sulfide bridges). For example, it may comprise a light chain in conjunction with a heavy chain comprising the effector domain. Any appropriate light chain can be used, for example, the allotype of light chain kappa, more common is Km (3) in the general population. Therefore, it may be desirable to use this common kappa light chain allotype, since relatively few members of the population would observe it as foreign. Typically, the target will be an antigen present in a cell, or a receptor with a soluble ligand by which the antibody competes. This can be selected as being a therapeutic goal, because of that it is desirable to link it with a molecule having the properties discussed above, for example to compete with or displace undesirable antibodies from it. Alternatively, it may be desirable to bind the target molecule per se, without causing destruction mediated by the cells, inflammation initiated by antibodies or complement lysis. Similarly, the effector domain may function primarily in the mediation of transport and / or the improved life span of the serum - in such cases the binding domain and the target molecule may be any system that would benefit from these qualities.

A selection of applications where the binding molecules of the present invention could be used as therapeutic antibodies having inert Fe regions (in some respects) are set forth below: 1) The competition with the alloantibodies of maternal IgG for the antigenic epitope in the blood cells of the fetus / neonate Alloimmune disorders of blood, fetal cells have a common pathogenesis. There is a synthesis of alloantibodies of IgG by the mother for a paternally inherited antigen in erythrocytes, granulocytes or fetal platelets. This is followed by the transplacental transport of the alloantibody. In the fetus or neoanto, there is a destruction of blood cells, fetal, coated with antibodies, which can lead to a clinically significant fall in the circulation levels of the relevant cells. Therapeutic antibodies for the relevant epitope, but with a Fe which does not initiate destruction, could compete with the maternal antibody for binding to the fetal cells, thereby inhibiting their destruction.

Antibodies to the alloantigens of the arteries lead to hemolytic disease of the fetus and neonate.

The alloantigens of the erythrocytes, most important are in the systems of the blood group Rhesus and Kell. The incidence of hemolytic disease due to RhD antigen has fallen dramatically since the introduction of postnatal prophylaxis, but cases still occur due to maternal sensitization during the first pregnancy. Other Rhesus antigens (C, c, E, e) can also cause hemolytic disease, as can antibodies to the Kell antigen (Kl), which also worsen erythropoiesis in the bone marrow, fetal. The current therapy for severely affected fetuses consists of intrauterine, regular transfusion of antigen negative erythrocytes. It has been shown that infusions of non-specific immunoglobulin are effective in this condition. Anemia and hyperbilirubinemia in the neonate may require intercavity transfusion and / or phototherapy. Experiments using inert Fe constructs with RhD specificity (designated Fog-1) have demonstrated their failure to activate effector mechanisms (activation of monocytes detected by chemical fluorescence and ADCC), and they have also been shown to inhibit Importantly, chemical fluorescence and ADCC activated by human sera containing polyclonal anti-D. It has previously been shown that ADCC and chemical fluorescence predict the destruction of erythrocytes in vivo. Previously published work has also demonstrated the ability of Fog-1 to compete with most anti-D human sera by epitopes on the RhD protein.

Antibodies to allogeneic platelets lead to fetal and neonatal alloimmune thrombocytopenia The most relevant antigen is the human platelet antigen (HPA) -la. Antibodies to HPA-la complicate 1 in 350 normal pregnancies, and lead to severe thrombocytopenia in 1 in 1200 fetuses. The most severely affected cases result in intracranial hemorrhage or death. Current options for therapy are weekly transfusions of negative HPA-platelets (which carry a 0.5% fetal death risk / procedure), and intravenous immunoglobulin at high dosages given to the mother, which has variable and unpredictable efficacy . HPA-la is defined by an individual epitope on the platelet glycoprotein Illa (GPIIIa), and an individual Fv chain recognizing this epitope is available within the University of Cambridge Division of Transfusion Medicine (Griffin HM, Ouwehand WH. monoclonal antibody specific for the leucine-33 (PAA1, HPA-la) form of platelet glycoprotein Illa from a V gene phaqge display library, Blood 1995; 86: 4430-4436). It has been shown that the binding of an antibody based on this construction to human platelets is inhibited by anti-HPA-human sera. Inhibition was more consistent for sera with the highest titer of specific antibodies, which were associated with the most serious disease. This indicates that the recombinant antibody and antibodies from the sera bind to the same epitope on the platelets. In the above and subsequent applications, in addition to a competitive binding effect, the therapeutic antibodies of the present invention may also initiate a beneficial inhibitory effect through the FcγRIIb. 2) Competence with the autoantibody by the epitope in the autoantigen Destruction of blood cells mediated by au toan ti bodies Hemolytic anemia due to autoantibodies of hot type IgG and thrombocytopenia due to autoantibodies have a common mechanism of destruction of blood cells. In both, autoantibodies target a selected repertoire of autoantigens (Rh and K in erythrocytes, and GPIIb / IIIa, GPIb / IX / V in platelets). The autoantibody binding shortens the duration of blood cells which leads to anemia or thrombocytopenia, respectively. It is not unlikely that the autoantibodies of erythrocytes and platelets target a limited number of B cell epitopes in their respective autoantigens. The variable domain antibodies, recombinant against these epitopes can be generated by the bacteriophage exposure technology of gene V. Therapeutic antibodies for the relevant epitopes, but with inert Fe, could compete with the autoantibodies of the patient's blood cells by the link to the autoantigen, to inhibit in this way the destruction of the blood cell.

Goodpasture syndrome (anti-glomerular base membrane disease [GBM]) This is a major cause of rapidly progressive glomerulonephritis, which leads to lung hemorrhage and end-stage renal failure within weeks or months of onset. Conventional therapy depends on dialysis in combination with intensive plasma exchange and immunosuppressive therapy, which itself can be complicated by life-threatening opportunistic viral and fungal infections. There is overwhelming evidence that this disease is mediated by autoantibodies, and autoantigen has been localized for type IV collagen, a major component of GBM. It has been shown that autoantibodies in GBM disease bind to the non-collagenous domain (NC1) of the a3 (IV) chain. The gene that codes for this sequence (COL4A3) has been cloned and ordered in series. It is hypothesized that the effect of dangerous anti-GMB autoantibodies can be neutralized by a competing molecule of monoclonal IgG which targets the immunodominant epitope in oc3 (IV) NCl and has been equipped, by design, with a Fe domain biologically inactive. A chimeric, recombinant IgG antibody will be developed which binds the immunodominant epitope a3 (IV) NCl but lacks the classical effector functions. It will be possible to achieve this when the genes coding for the variable domains of murine anti-a3 (IV) NCl have been • developed and characterized (Pusey CD et al, Lab Invest 1987, 56; 23-31 and Ross CN et al., Lab Invest 1996, 74; 1051-1059). Again, in addition to a competitive binding effect, the therapeutic antibodies of the present invention may also activate an inhibitory, beneficial effect through FcγRIIb. 3) Allergy and Asthma Allergies and asthma result from immune responses, inappropriate to common environmental antigens such as lawn pollen proteins, house dust mites and many other sources of common antigens, an example being the Der P 1 protein Dust mite of the house Derma tophagoides pteronyssinus. Affected individuals produce high levels of immunoglobulins particularly of the IgE class. These IgE antibodies are capable of binding to the high affinity Fc-epsilon Rl receptor in mast cells or Mastzelle and Eosinophil cells. The cross-linking of the IgE linked to the receptor by the allergen results in the activation of the cells and degranulation. This releases a number of inflammatory mediators which can cause severe symptoms or even death as a result of an anaphylactic reaction. Two mechanisms of action of a blocking antibody could be considered. First, an IgG antibody with an inert Fe region could compete for the binding of the allergen to IgE. This would prevent the cross-linking of the IgE and therefore prevent the activation of the cells. For this mechanism, the IgG antibody with the inert Fe would have to compete directly for the link of the allergen with the IgE. A second significant mechanism would involve the role of negative signaling through the Fc? RIIb receptor. It has been shown that the cross-linking of the Fe gamma RIIB and the Fe epsilon R1 results in an inhibition of the activation signals normally observed when only the Fe epsilon Rl receptors are cross-linked. Thus, the introduction of an IgG antibody with an Fe binding capacity for the Fe gamma Rllb and an antigen specificity for an allergen could result in an inhibition of mast cell or mast cell activation. IgE and Eosinophil cells. Thus, the IgG antibody would also mediate its negative, strong affect if it binds the allergen to a site other than IgE such that both could bind to the allergen at the same time. 4) Inflammatory disorders such as Crohn's disease There are a number of disorders of the immune system, which seem to cause pathology as a result of the chronic state of activation of immune cells (leukocytes), including T-lymphocytes, neutrophils and NK cells. This chronic activation is normally observed as a state of inflammation with a continuous migration of activated cells in the affected tissues. To migrate into the tissue, the cells must receive and respond to the inflammatory mediators and then regulate the adhesion molecules to make it possible for them to adhere first to the cells that cover the walls of the blood vessels and then migrate between the cells of the cells. walls of the vessels and inside the tissue. It should be possible to stop this cycle of inflammation by blocking either the adhesion molecules on the surface of the leukocytes or the corresponding ligands in the activated epithelial cells that cover the walls of the vessel. This activation ant is VAP-1 and an antibody with an inert Fe which binds to this molecule must prevent the adhesion of leukocytes and the migration to the sites of inflammation, in order to break the cycle of chronic activation.

) Inhibition of the ligand / receptor interaction Sickle cell disease Homozygosity for the human hemoglobin variant characterized by a substitution of valine with glutamic acid (HbSS) leads to chronic hemolysis and a tendency for the molecule to undergo tactile formation in the deoxygenated state. This leads to the erythrocytes adopting a form of drepamositos in the microcirculation leading to 'crisis' of drepanositos in localized areas. These can be thrombotic (in bone, lung, brain or abdomen), aplastic, hemolytic or associated with massive sequestration of erythrocytes in the spleen and liver. It is assumed that during these crises, the erythrocytes adhere to the endothelial cells. This adhesion process is based on the interaction of various receptors with their respective ligands. Two of the dominant adhesion pathways are the interaction between Lutheran and laminin and between thrombospondin and an undefined membrane lipid of erythrocytes. In animal experiments, evidence has been obtained that human-variable, recombinant antibodies against thrombospondin decrease the adhesion of sickle cells to endothelial cells. It is presumed that recombinant, variable domain antibodies, similar to the laminin binding domain of lutene (the near domain of the membrane) which blocks the interaction with laminin can be developed by exposure of V gene bacteriophages. These fragments of variable domain antibodies can be equipped with inert Fe domains to produce therapeutic antibodies capable of interfering with the adhesion of sickle erythrocytes to endothelial cells, without causing the destruction of erythrocytes.

Blockade mediated by anti bodies of platelet collagen receptors There is substantial evidence that two receptors are crucial for the activation of platelets by subendothelial collagens, a case that initiates thrombosis; the integrin a2ß? (platelet glycoprotein Ia / IIa) which is seen primarily as an adhesive in function, and the integrin VI different glycoprotein (GpVI) as essential for activation, preceding secretion and aggregation. Human, recombinant antibodies can be generated by exposure of V gene bacteriophages that recognize the different domains within each receptor, and these can be used to produce major antibodies with an inert Fe domain for anti-thrombotic therapy based on to collagen. These can be used to relieve coronary thrombosis, restenosis after angioplasty and thrombotic complications associated with the implantation of a deviation. 6) Monoclonal antibodies are sometimes used to block cell functions, for example OKT3 is used to immunosuppress T cells by blocking the T-cell receptor and CD18 antibodies are used to prevent cell-cell adhesion. Through integrin molecules, however, the binding of Fe to Fe receptors can initiate serious side effects through the stimulation of cytokine release and inflammation. 7) The Fe regions of the antibodies sometimes bind to other recombinant proteins to give the fusion molecules long biological lifetimes. In this way, the TNF receptor has been bound to human IgG4 Fe to form a molecule which inhibits the effects of soluble TNF, and CTLA4 has been made as a fusion protein with the IgG Fe and was used to block signaling through the B7 co-receptor molecule (a ligand for CTLA4) on cell surfaces. However, the activation of cytokines by the Fe of the fusion protein is again undesirable. The V domains, or other link regions, appropriate for the types of application discussed above, where specifically discussed, will be well known to those skilled in the art. For example, a CD3 binding domain (eg, YTH12.5) is described by Routledge et al. (1991) Eur J Immunol 21, 2717-2725 and Bolt et al. (1993) Eur J immunol 23, 403-411. A CD52 binding domain (e.g., CAMPATH-1) is described by Riechmann et al. (1988) Nature 332, 323-327. A linkage domain of VAP-1 is described by Salmi et al. (1993) J Exp Med 178: 2250-60 and Smith et al. (1998) J Exp Med 188: 17-27. A domain of Der p I (for example 2C7) is described by McElveen et al. (1998) Clin Exp Allergy 28, 1427-1434. In this way, a binding molecule which did not bind to the Fe receptors and initiated the destruction, and did not activate the complement, but which did not bind to a target molecule, could be used in all the previous examples to minimize some side effects Specifically, this 'blocking' antibody could be introduced in situations 1-5 above and prevent undesirable destruction by naturally occurring antibodies. The same Fe regions of the blocking type would be the Fe regions of the selection for use for recombinant antibodies such as the CD3 or CD18 antibodies in the above situation 6 or as the Fe for the fusions in the situation 7 above. The linker domains and effectors can be combined by any suitable method. For example, domains can be covalently linked through side chains. Alternatively, the sulfhydryl groups generated by the chemical reduction of the cysteine residues have been used to cross-link the antibody domains (Rhind, S K (1990) EP 0385601, Cross-linked antibodies and processes for their preparation). Finally, the chemical modification of the carbohydrate groups has been used to generate reactive groups for the purposes of crosslinking. These methods are normal techniques, available to those skilled in the art. These may be particularly applicable in embodiments wherein the binding polypeptide contains non-protein portions or groups. In general, it may be more appropriate to use recombinant techniques to express the binding molecule in the form of a fusion protein. The methods and materials employed in this approach form additional aspects of the present invention, as discussed below.

Nucleic acids In one aspect of the present invention, a nucleic acid encoding a linker molecule is described as described above.

The nucleic acid according to the present invention may include cDNA, RNA, genomic DNA (including introns) and modified nucleic acids or nucleic acid analogs (eg, peptide nucleic acid). Where a DNA sequence is specified, for example, with reference to a Figure, unless the context otherwise requires the RNA equivalent, with U replaced by T where this occurs, is included. The nucleic acid molecules according to the present invention can be provided isolated and / or purified from their natural environment, in substantially pure or homogeneous form, or free or substantially free of other nucleic acids of the species of origin. used herein, the term "isolated" encompasses all of these possibilities.Nucleic acid molecules can be completely or partially synthetic.In particular, these can be recombinant in that the nucleic acid sequences which are not found together in nature (they do not extend contiguously) they have been linked or otherwise artificially combined, alternatively, they may have been synthesized directly, for example, using an automated synthesizer.

In a further aspect a nucleic construct is described, for example, a duplicatable vector, comprising the nucleic acid sequence. . A vector that includes nucleic acid according to the present invention need not include a promoter or other regulatory sequence, particularly if the vector is to be used to introduce the nucleic acid into the cells for recombination in the genome. Preferably, the nucleic acid in the vector is under the control of, and operably linked to, an appropriate promoter or other regulatory elements for transcription in a host cell such as a microbial cell (eg, bacterial, yeast, filamentous fungi) or eukaryotic (for example, insects, plants, mammals). Particularly, the vector may contain a gene (for example gpt) to allow selection in a host or a host cell, and one or more appropriate enhancers for the host. The vector can be a bi-functional expression vector which functions in multiple hosts. In the case of genomic DNA, this may contain its own promoter or other regulatory elements and in the case of the cDNA it may be under the control of an appropriate promoter or other regulatory elements for expression in the host cells. By "promoter" is meant a nucleotide sequence from which the transcription of the DNA operably linked downstream can be initiated (ie, in the 3 'direction in the direction of the double stranded DNA strand). The promoter may optionally be an inducible promoter. "Operably linked" means linked as part of the same nucleic acid molecule, suitably positioned and oriented so that transcription is initiated from the promoter. The DNA operably linked to a promoter is "under transcriptional initiation regulation" of the promoter. Thus, this aspect of the invention provides a gene construct, preferably a duplicatable vector, comprising a promoter operably linked to a nucleotide sequence provided by the present invention. Generally speaking, those skilled in the art are capable of constructing vectors and design protocols for the expression of recombinant genes. Suitable vectors can be selected or constructed, containing appropriate regulatory sequences, including promoter sequences, terminator fragments, polyadenylation sequences, enhancer sequences, marker genes and other sequences as appropriate. For additional details see, for example, Molecular Cloning: a Labora tory Manual: 2nd edition, Sambrook et al., 1989, Cold Spring Harbor Laboratory Press. Many known techniques and protocols for the manipulation of nucleic acid, for example in the preparation of nucleic acid constructs, mutagenesis, sequence formation, introduction of DNA into cells and gene expression, and protein analysis are describe in detail in Current Protocols in Molecular Biology, Second Edition, Ausubel et al. eds., John Wiley & Sons, 1992. The descriptions of Sambrook et al. And Ausubel et al. Are incorporated herein by reference.

Hosting Cells & methods Cells transformed by the expression vectors defined above are also encompassed by the present invention. Cell cultures (preferably from rodents) and cell culture products containing the binding molecules are also provided. Also provided are methods for making the linker molecules according to the present invention comprising: (i) combining a nucleic acid encoding a binding domain with a nucleic acid encoding an effector domain to form a nucleic acid construct; (ii) cause or allow the expression of the construction in a host cell, adequate. The combination, to produce a construction, can be by any convenient method, known to those skilled in the art, for example, by ligating fragments (eg, restriction fragments) or using different models in one or more steps of amplification, for example, the use of a PCR. Methods for the production of antibodies (and therefore the binding domains) include immunization of a mammal (e.g., human, mouse, rat, rabbit, horse, goat, sheep, camel or monkey) with a target protein, adequate or a fragment of it. The antibodies can be obtained from animals immunized using any of a variety of techniques known in the art, and can be selected, preferably by using the binding of the antibodies to the antigen of interest. For example, western blotting techniques or immunoprecipitation can be used (Armitage et al., 1992, Nature 357: 80-82). The cloning and expression of chimeric antibodies are described in European Patent Application No. 0120694 and European Patent Application No. 0125023. The nucleic acid encoding the effector domain can be generated, in view of the present disclosure, by Site-directed mutagenesis, for example by the methods described herein or in the published art (see for example, WO 92/16562 or WO 95/05468 both by Lynxvale Ltd.).

Other aspects The use of binding molecules of the present invention is also provided to prevent, inhibit or otherwise interfere with the binding of a second molecule to a target molecule. This may involve competing with, or displacing, an antibody from a therapeutically relevant target or antigen. The present invention also provides a reagent which comprises a linker molecule as above, if it is produced recombinantly or otherwise. The present invention also provides a pharmaceutical preparation which comprises a linker molecule as above, plus a pharmaceutically acceptable carrier. The present invention also provides a method for the treatment of a patient which comprises administering a pharmaceutical preparation as before to the patient, or a sample (for example a blood sample) separated from that patient, which is subsequently returned to the patient. Particularly, a method of treatment for the following diseases: graft-versus-host disease; host-versus-graft disease; rejection of organ transplantation; rejection of bone marrow transplantation; autoimmunity; alloimmunity; allergy; Chronic or acute inflammatory diseases. The present invention also provides a method for the treatment of a patient which comprises causing or allowing the expression of a nucleic acid encoding a linker molecule as described above, whereby the linker molecule exerts its effects in vivo in the patient. In general, the expression will occur in the patient, or in certain specialized circumstances where the patient is an unborn infant, in the patient's mother. The use of a linker molecule as above is also provided in the preparation of a pharmaceutical substance for modifying an immune response, particularly a pharmaceutical substance for the treatment of the diseases discussed above. In order that the present invention be more fully understood, the modes will now be described in more detail, by way of example only, and not by way of limitation. Other embodiments that are within the scope of the invention may occur for those skilled in the art in view thereof.

FIGURES Figure 1 Formation of rosettes of cells bearing FcγRI by RBCs (erythrocytes) coated with Fog-1 antibodies. RBC R2R2 were coated with Fog-1 antibodies in a range of antibody concentrations, incubated with B2KA cells that were cultured in a 96-well plate and the percentage of B2KA cells were determined with RBC rosettes. The error bars indicate the normal deviation values for the wells in triplicate. For Gl? B, Gl? C, Gl? Ab, Gl? Ac, G2? A, G4? B and G4? C of the Fog-1 series, mutants, as for G2 (shown), there was no rosetting between B2KA cells and RBCs at any of the coating concentrations.

Figure 2 Fluorescent staining of the cells bearing the FcγRI. The transfectant cell lines of Fc? RI, B2KA (a and b) and the chain 3T3 + Fc? RI +? (c and d) were incubated sequentially with the antibodies of the CAMPATH-1 (a and c) or Fog-1 (b and d) series, the anti-human, biotinylated K antibodies and ExtrAvidin-FITC. The fluorescence intensities were measured for 10,000 cases and the intermediate, geometric channel of fluorescence was recorded in a diagram.

Figure 3 Representation in a histogram of fluorescently stained cells carrying the FcγRI. B2KA cells were stained as in Figure 2 using 100 μg / ml of CAMPTH-1 series antibodies. The records in the histogram showing the number of cells found in each fluorescence channel were covered by the representative antibodies.

Figure 4 QL response of human monocytes to RBCs sensitized with the Fog-1 series antibodies. RBC RiRi were coated with antibodies over a range of concentrations. The number of antibody molecules bound per cell and the QL response of the monocytes to the RBCs were determined for each sample described.

Figure 5 Inhibition of QL due to the Gl of the series Fog-1 by other antibodies of the Fog-1 series. The RBCs were sensitized with 2 μg / ml Gl of the Fog-1 series and different Ab concentrations of the indicated Fog-1 series. These Ab gave a low QL response in Figure 4. The QL response of the monocytes was measured.

The response due to 2 μg / ml Gl is only taken as 100%.

Figure 6 Inhibition of the QL response to clinical sera by the G2α of the Fog-1 series. The RBCs were synthesized with a constant amount of Gl from the Fog-1 series (20 μg / ml) or clinically relevant sera and different amounts of G2α from the Fog-1 series. The 100% response was achieved with a normal amount of BRAD 5. In the absence of G2? A from the Fog-1 series,% of the responses were Gl: 150%, serum A: 142%, serum B: 265 %, serum C: 200%, serum D: 163%, serum E: 94%, antiserum C + D: 259% and anti-serum K: 119%.

Figure 7 Complement lysis mediated by the CAMPATH-1 series antibodies. Peripheral blood mononuclear cells, PBMC, from human were labeled with 51 Cr and incubated with the antibodies in the presence of serum as a source of complement. The% of specific Cr release is recorded in a diagram as a measure of the lysis that occurred.

Figure 8 Inhibition by G2? To CAMPTH-1 of the lysis of the complement mediated by Gl of the CAMPATH-1 series. Complement lysis was carried out as in Figure 7 but the samples contained a constant amount (6.25 μg / ml final concentration) of Gl from the CAMPATH-1 series and increasing amounts of G2α from the CAMPATH-1 series .

Figure 9 ADCC mediated by antibodies of the CAMPATH-1 series. Human PBMCs were labeled with 51 Cr and incubated with the antibody. After washing, the cells were incubated with additional PBMCs, which acted as effector cells, in an effector: target ratio of 20: 1. The% of specific Cr release is recorded in a diagram as a measure of the lysis that occurred.

Figure 10a ADCC of RBC RhD + mediated by antibodies of the Fog-1 series.

Figure 10b ADCC of RBC RhD + mediated by antibodies of the Fog-1 series Figure 1. Inhibition by the antibodies of the FCC-1 ADCC series of RBC RhD + mediated by the Gl of the Fog-1 series at 2 ng / mg.

Figure 11b Inhibition by the FCC-1 ADCC series of RBC RhD + mediated Gl antibodies of the Fog-1 series. The RBCs were sensitized in a mixture of antibodies consisting of a constant amount of Gl of the Fog-1 series (2 ng / ml) and different concentrations of inhibitory antibodies.

Figure 12 Inhibition by antibodies of the ADCC Fog-1 series of RBC RhD + mediated by polyclonal anti-RhD at 3 ng / mg.

Figure 13a Fluorescent staining of cells bearing FcγRIIa 131H / H. The cells of the 3T6 + FcγRIIa 131H / H transfectant line were incubated with the Fog-1 series antibodies made complex with goat F (ab ') 2 antihuman and then with FITC conjugated donkey anti-goat IgG. The fluorescence intensities were measured per 10000 cases and the intermediate, geometric channel of the fluorescence was recorded in a diagram.

Figure 13b Fluorescent staining of cells bearing FcγRIIa 131R / R. The cells of the 3T6 + FcγRIIa 131R / R transfectant line were incubated with Fog-1 series antibodies made complex with FITC-conjugated goat anti-human F (ab ') 2 K. The fluorescence intensities were measured per 10000 cases and the intermediate, geometric channel of the fluorescence was recorded in a diagram.

Figure 14a Fluorescent staining of cells bearing Fc? RIIbl *. The experiment was carried out as in Figure 13b using the 3T6 + Fc? RIIbl * transfectant line and complexing the Fog-1 series antibodies using a FITC conjugated goat anti-human F (ab ') 2 mixture. and not marked.

Figure 14b Fluorescent staining of cells bearing Fc? RIIIb NAl. The experiment was carried out as in Figure 13 using the CHO + Fc? RIIIb NAl transfectant line.

Figure 14c Fluorescent staining of cells bearing Fc? RIIIb NA2. The experiment was carried out as in Figure 13 using the CHO + Fc? RIIIb NA2 transfectant line.

Figure 15 This shows Table 1, which compares the mutations made to the non-cultured Gl, G2 and G4 antibody sequences.

Figure 16 This shows Table 2, which is a summary of the activities of the antibodies.

Figure 17 This shows the sequences of certain modified and uncultivated CH2 sequences, including those designated Gl? Ab, "G2? A, Gl? Ac.

EXAMPLES General Materials and Methods Construction of expression vectors The starting point for the constant region of the IgG1 was the gene of the constant region of the human IgG1 of the Glm allotype (1.17) in a version of the M13tgl31 vector which contains a modified polylinker (Clark, MR: WO 92 / 16562). Thus, the 2.3kb IgGl insert has a BamHI site at the 5 'end and contains an idIIIII site adjacent to the BamHI site. At the 3 'end, downstream of the polyadenylation signal, the following sites occur in the order 5' to 3 ': Sphl, Notl, BglII, BamHI. The human IgG2 constant region gene has been obtained as a fragment of idindIII-Sphl in M13tgl31 and the HindIII site has been destroyed by digestion with HindIII, insertion at the projecting ends and ligation of the ends together again. The Asll-Sphl fragment of this vector was cloned to replace the equivalent fragment in the IgG1 vector described above. The human IgG4 constant region gene has been obtained as a HindlII-Smal fragment in M13tgl31 and the HindlII site has been destroyed. The Smal site occurs between the 3 'end of the CH3 exon and the polyadenylation site so that the polyadenylation site was restored by adding the Smal fragment of the IgG1 vector, which comprises DNA from the Sjnal equivalent site in the gene of IgGl and the Smal site downstream of the gene in the polylinker. The first procedure was to introduce a Xjbal restriction site between the CH1 and articulation exons, an Xhol site between the articulation exons and CH2 and a Kpnl site between the CH2 and CH3 exons in order to facilitate the exchange of exon mutant sequences. This was similar to the manipulation of the IgG1 and IgG4 genes carried out previously (Greenwood, J., Clark, M. and Waldmann, H. (1993) Structural motifs involved in human IgG antibody effector functions, Eur. J. Immunol 23, 1098-1104). To provide the model DNAs, the RZ1032 of E. col i was infected with the M13 described above and the ssDNA was prepared. The strain is du t'ung 'so that the ssDNA produced must contain some uridine instead of thymidine. The oligonucleotides used to introduce the mutations were: between the exons of articulation and CH2 MO10 5 'GGA TGC AGG CTA CTC GAG GGC ACC TG 3' between exons CH2 and CH3 MOll 5 'TGT CCA TGT GGC CCT GGT ACC CCA CGG GT 3 'between exons CH1 and articulation M012 5' GAG CCT GCT TCC TCT AGA CAC CCT CCC T 3 'The restriction sites are underlined. The oligonucleotides were phosphorylated in 50μl reactions containing 25 pmol of oligonucleotide and 5u of T4 polynucleotide kinase (nbl) in 70 mM Tris HCl pH 7.6, MgCl210 mM, 100 mM KCl, 5 mM DTT, 0.5 mg / ml BSA, ATP 1 mM. The reactions were incubated at 37C for 1 hour and heated at 70C for 5 minutes. To strengthen the mutagenic oligonucleotides for the model DNA, 500 ng of uridine-containing DNA and 1 pmol of each phosphorylated oligonucleotide were incubated in 20 μl of 40 mM Tris HCl pH 7.5, 20 mM MgCl 2, 50 mM NaCl at 80 ° C for 5 minutes and it was allowed to cool slowly to 37C. The volume was increased to 30 μl with the same buffer and DTT was added to 7 mM, ATP to 1 M and dATP, dCTP, DGTP and dTTP each to 250 μM. Five u of T7 DNA polymerase (unmodified, United States Biochemical) and 0.5 u of T4 DNA ligase (Gibco BRL) were added and the reaction was incubated at room temperature for 16 hours to synthesize the mutant chain. The DNA was precipitated with ethanol, 50 μl of 20 M Tris HCl pH 8.0, 1 mM EDTA, 1 mM DTT, 0.1 mg / ml of BSA were dissolved and 1 u of uracil DNA glycosylase (New England Biolabs) was added. After incubation at 37 C for 2 hours, 50 μl of 400 mM NaOH was added and the reaction was left at room temperature for 5 minutes to fragment the model DNA strand. The DNA was precipitated with ethanol, dissolved in H20 and transformed into TG1 of E. coli The duplicative (RF) form of the DNA was made for a selection of the resulting M13 clones and digested to find the clones which contained the required Xbal, Xhol and Kpnl restriction sites. Suitable clones were obtained for IgGl and 4 vectors but M012 appeared to be unstrengthened in the IgG2 vector so that mutagenesis was repeated for IgG2 without this oligonucleotide since the site between the CH1 and the articulation exons was not necessary for these experiments. For each vector, the DNA sequences of the exons were confirmed by sequencing. Changes in CH2 at amino acid positions 327, 330 and 331 (mutation? A) had to be introduced using the oligonucleotides: -M022BACK (coding strand): 5 'TCT CCA ACA AAG GCC TCC CGT CCT CCA TCG AGA AAA 3 ' M022 (complementary chain): 5 'TTT TCT CGA TGG AGG ACG GGA GGC CTT TGT TGG AGA 3' The changes in CH2 at positions 233 to 236 (mutation? By? C) had to be introduced using the oligonucleotides: - M07BACK ( encoding and mutation coding? c coding): 5 'TCC TCA GCA CCT CCA GTC GCG GGG GGA CCG TCA GTC 3' M021 (complementary chain and mutation? b coding): 5 'GAC TGA CGG TCC CGC GAC TGG AGG TGC TGA GGA 3 'Mutations had to be introduced by an overlay extension or coating PCR that also required the oligonucleotides MOll and MO10BACK: 5 'CAG GTG CCC TCG AGT AGC CTG CAT CC 3' The Xhol restriction site is underlined. For the mutation? A, the first set of PCRs used the IgGl and IgG2 models amplified with M022 and MO10BACK with the M022BACK and MOll. For mutations? A and? C, the first set of PCRs used the IgGl and IgG4 models with M021 and MO10BACK and with M07BACK and MOll. In the product . Finally, the DNAs that originate from a barley chain with M021 would have the mutation? b and those that originate from M022BACK would carry the mutation? c. Each PCR contained approximately 30 ng of M13tgl31 + constant region ssDNA, 25 pmo of each oligonucleotide and 1 u of Pwo DNA polymerase (Boehringer Mannheim) in 50 ul of 10 mM Tris HCl, pH 8.85, 25 mM KCl, (NH4) 2S0 5mM, 2mM MgSO4 and 250μM each of dATG, dCTP, dGTP and DTTP. The reactions were subjected to 14 cycles of 94 C, 30 s; 50C, 30 s; 72C, 60s, followed by 72C, 5 minutes until the end. The bands representing the product DNAs of the expected sizes were removed from low melting point agarose and melted in 100 μl H20. For each mutation, the two initial PCR products were joined together by overlapping extension PCR. Approximately 4 μl total of melted gel cuts, such that the initial products of the PCR were in equimolar amounts, were mixed with 25 pmoles of each of MO10BACK and MOll and other components as before. The PCR was carried out for 18 cycles as before, except that the strengthening or tempering temperature was reduced from 50C to 48C. The obtained products, which contained the total CH2 exon, were purified and digested with Xhol and Kpnl. The RF DNAs of the M13tgl31 + constant region vectors, which contain the extra restriction sites described above, were digested with XhoI and KpnI to separate existing CH2 DNAs and mutant CH2 regions ligated therein. The DNA samples were transformed into TG1 of E. coli. The DNA of the representative clones was sequenced to identify the correctly mutated constant regions. In order to obtain the IgGl vectors with both? A and? B or? C, the DNA, representing a mutant of? A, was used as the model for a second round of PCRs to introduce the mutations? B and? C described above. The constant region genes, non-cultured and mutated type 2 and 4, of IgG1 were each deleted from RF DNA as a BamHI-Notl fragment and cloned into the CAMPATH vector Hu4VH modified pSVgpt HuIgGl (Clark, MR: Lynxvale Binding Molecules as before) to replace the constant, existing region. The resulting vectors were designated pSVgptCAMPATHHu4VHHuIgGl? A, etc. The vector also contains the gpt gene to allow selection in mammalian cells, the murine immunoglobulin heavy chain enhancer and the AD? of variable region CAMPATH1 Hu4VH so that it carries a complete heavy chain gene which can be expressed in mammalian cells. The humanized light chain gene CAMPATH-1 exists in the CAMPATH expression vector HuVL pSVneo (Reichmann, L., Clark, M.R., Waldman, H. and Winter, G (1988) Nature 332, 323-327). The variable region ADNS of Fogl (Bye, J.

M., Carter, C, Cui, Y., Gorick, BD, Songsivilai, S., Winter, G., Hunghes-Jones, NC and Marks, JD (1992) Germline variable region gene segment derivation of human monoclonal anti-Rh (D) antibodies. J. Clin. Invest. 90, 2481-2490) were obtained in the vector pHENl. These were amplified by PCR, using the oligonucleotides: - FOG1VHBACK 5 'TCC ACA GGT GTC CAC TCC CAG GTG CAT CTA CAG CAG 3' FOG1VHFOR 5 'GAG GTT GTA AGG ACT CAC CTG AGG AGA CGG TGA CCG T 3' FOG1VKBACK 5 ' TCC ACA GGT GTC CAC TCC GAC ATC CAG ATG ACC CAG 3 'FOG1VKFOR 5' GAG GTT GTA AGG ACT CAC GTT TGA TCT CCA GCT TGG T 3 'The position 5' of the insert in the vector M13CHPCR1 (Orlandi, R., Gussow, DH, Jones, PT and Winter, G. (1989) Proc. Nati, Acad. Sci. USA 86, 3833), which comprises the promoter and the DNA encoding the signal peptide was amplified using the reverse primer M13, universal and V03: 5 'GGA GTG GAC ACC TGT GGA GA 3' the DNA, 3 'of the VH in M13VHPCR1 and representing the 5' end of the intron VH-CH, was obtained by PCR using the universal primer M13-40 and V04: 5 'GTG AGT CCT TAC AAC CTC TC 3' These two DNA segments are sequentially linked to the amplified DNA of both VH of the Fog-1 and V series? of the Fog-1 series by an overlap extension PCR as described above. The internal BajnHI restriction site for the VH of the Fog-1 series was deleted by the same method using olinucleotides which separated the recognition site without changing the encoded amino acids. The complete PCR products were cloned into M13mpl9 as the HindIII-BainHI fragments and their confirmed DNA sequences. The ífindlll-Bai? IHI fragment containing the VH of the Fog-1 series was used to replace the VH-containing fragment of the CAMPATH-1 series in the pSVgpt vectors described above, to give the expression vectors designated pSVgptFoglVHHuIgG2, et cetera . For the IgG1 vectors, the extra HindIII restriction site at the 5 'end of the constant region DNAs meant that it was not possible to simply exchange the HindIII-BamHI variable region fragment. In contrast, the relevant pSVgptCAMPATHHu4HHuIgGl vectors were digested with HindlII. The linkers, designed to suppress the HindlII site and to add a BamHI site, were ligated at the cut ends. The DNAs were then digested with BamHI and NotI so that the constant regions could be isolated and these were cloned into pSVgptFoglVHHuIgG2 to replace the constant region of IgG2. The FindIII-BainHI fragment containing the V of the Fog-1 series was transferred to the pSVhyg-HuCK vector (Orlandi et al., 1989) which already contains the murine immunoglobulin heavy chain enhancer and the constant region K gene. of human. The resulting expression vector was called pSVhygFoglVKHuCK.

Production of bodies μg of each heavy chain expression vector and 20 μg of the relevant chain expression vector were linearized by means of the Pvul digestion and were combined in 50 μl of H20. Cells from the non-secreting rat myeloma line, YB2 / 0, were grown to semi-confluence in an Iscove-modified Dulbecco's medium (IMDM) with 5% fetal bovine serum (FBS). 107 cells were collected by centrifugation, resuspended in 0.5 ml of the medium and transferred to a GenePulser probe (BioRad). The DNA was added and the mixture was incubated on ice for 5 minutes. The cells were boosted at 960 μF / 170 V and returned to the ice for 15 minutes before being placed in a flask in 20 ml of IMDM + 10% FBS. These were incubated at 37C, C02 5% in a humid atmosphere. After 24 hours, the volume was doubled and the medium was made selective by the addition of mycophenolic acid at 0.8 μg / ml and xanthine at 250 μg / ml. The cells were prorated on two 96-well plates. Approximately 18 days after the selection was applied, the colonies were visible and the supernatants were subjected to an assay for the presence of IgG by ELISA. In summary, wells of microtiter plates were coated with antibodies specific for Fe, goat antihuman IgG (Sigma) and then incubated with 5-fold dilutions of the supernatants. The bound antibody was detected by incubation with conjugated goat antihuman antibodies, with HRPO (Seralab) and the development of the assay with the o-phenylenediamine substrate. The cells in the wells containing the highest amounts of antibodies were expanded and the stock solutions were cryopreserved. The cell line that secretes the highest amounts of Ab was expanded to 500 ml in IMDM + 2% FBS to provide a saturated supernatant for the purification of antibodies. The supernatant was purified by centrifugation and 0.1 M Tris HCl, pH 8.0 made. Protein A-agarose (Sigma) was added and the mixture was stirred at 4C for 16 hours. The agarose beads were collected on a column and washed with 0.1 M Tris HCl pH 8.0, followed by 10 mM Tris HCl pH 8.0. The antibody was eluted with 1 ml aliquots of 0.1 M glycine pH 3.0 in samples of 100 μl of 1 M Tris HCl pH 8.0 and the fractions containing significant amounts of protein were identified from the A2ßonm readings. These fractions were dialyzed against PBS, they were sterilized by a filter and the A28onm was re-measured to give the approximate antibody concentration (concentration = A28onm x 0.714 mg / ml). The purity and integrity of the antibodies were established by reducing SDS-PAGE, using 12.5% acrylamide. The concentrations were verified in an ELISA which used goat antihuman, K antibodies (Seralab) as the capture reagent and goat antihuman, biotinylated antibodies (Sigma) followed by ExtrAvidin-HRPO (Sigma) for detection. This means that it was unlikely that the nature of the heavy chain would influence the level of bond obtained.

Formation of rosettes of Fc? RI transfectants Washed RBC R2R2 were incubated with Ab samples in 100 ml of PBS in 96-well plates at room temperature for 1 hour. The RBCs were washed three times, resuspended in PBS and incubated at 37 C for 40 minutes with transfectants expressing the FcγRI, B2KA cDNA (S. Gorman and G. Hale, unpublished), cultured in 96 well plates. wells. The supernatant was discarded and the wells were washed one, the excess RBCs were separated. For each well, 200 B2KA cells were examined and the number with RBC rosettes was observed. The average percentage and the normal deviation for the wells in triplicate were recorded in a diagram. Alternatively, sensitized RBCs and B2KA cells were mixed in microcentrifuge tubes, pelletized and gently resuspended before being transferred to a microscope slide.

Fluorescent staining of Fc? R transfectants The transfectants that express the Fc? R, B2KA cDNA and the 3T3-Fc? RIa +? (van Urgt, J., Jeijnen, IAFM, Capel, PJA, Park, S. And, Ra, C, Saito, T. Verbeek, JS and van de Winkel, JGJ (1996) FcR? -chain is essential for both surface expression and function of human FcγRI (CD64) in vivo, Blood 87, 3593-3599), were obtained as individual cell suspensions in phosphate buffered saline containing 0.1% NaN3 (w / v), 0.1% BSA ( p / v) (wash buffer) after treatment with the cell dissociation buffer (Gibco BRL). Cell pellets were formed at 10 5 cells / well in 96-well plates, resuspended in 100 μl dilutions of CAMPATH-1 or Fog-1 series Ab and incubated on ice for 30 minutes. The cells were washed three times with 150 μl / well of wash buffer and similarly incubated with 20 μg / ml of goat antihuman K chain Ab, conjugated with biotin (Sigma) and then with 20 μg / ml of ExtrAvidin- FITC (Sigma). After the final wash, the cells were fixed in 100 μl of wash buffer containing 1% formaldehyde (v / v). Surface expression of FcγRI was confirmed by staining with the mAB CD64 (Serotec) and Ab of goat and mouse IgG, conjugated with FITC (Sigma). The fluorescence intensities were measured in a FACScan (Becton Dickinson). For transfectants carrying FcγRII, 3T6 + FcγRIIa 131H / H, 3T6 + FcγRIIa 131R / R (Warmerdam, PAM, van de Winkel, JGJ, Gosselin, EJ, and Capel, PJA (1990) Molecular basis for a polymorphism of human Fc? receptor II (CD32) J. Exp. Med. 172, 19-25; Warmerdam, PAM, van de Winkel, J GJ, Vlug, A., Westerdaal, NAC and Capel, PJA ( 1991) A single amino acid in the second Ig-like domain of the human Fc? Receptor II is critical for human IgG2 binding, J. Immunol., 147, 1338-1343) and 3T6 + Fc? RIIb'l * (Warmerdam, PAM , van den Herik-Oudij k, IE, Parren, PWHI, Westerdaal, NA C, van de Winkel, JG, J. and Capel, PJA (1993) Int. Immunol., 5, 239-247) the antibodies became complex before of being incubated with the cells. For FcγRIIa 131H / H, the antibodies were mixed with equimolar amounts of goat anti-human F (ab ') 2 (Seralab) and incubated at 37C for 1 hour. The complexes were then mixed with the cells and the assay was continued as before, except that the detection of antibodies was anti-goat IgG, conjugated with FITC (Serotec). For FcγRIIa 131R / R, the complexes were made using equimolar amounts of goat F (ab ') 2 anti-human K, conjugated with FITC (Seralab) and for FcγRIIbl *, complexes were made using amounts equimolar of a 1: 1 mixture of F (ab ') 2 goat antihuman FITC-conjugated and unlabeled. In this way, for these receivers only one incubation step was needed. for transfectants that carry Fc? RIIIb, CHO + Fc? RIIIb NAl or NA2 (Bux, J., Kissel, K., Hofmann, C. and Santoso, S. (1999) .The use of allele-specific recombinant Fe gamma receptor Illb antigens for the detection of granulocyte antibodies, Blood 93, 357-362), the staining was carried out as described for the above 3T6 + FcRIIa 131H / H cells.

Sensitization of the Eri troci tos The RBC RiRi of group O were washed in PBS and resuspended in RPMl + 10% FBS in a final concentration of 5% v / v. 10 μl of cells were added to 50 μl of mAb or RPMI / FBS in bottom well plates V and incubated for 60 minutes at 37 ° C. In some experiments, the mAbs were serially diluted in RPMI / FBS to achieve a range of IgG linked to erythrocytes. In the competition experiments, the erythrocytes were sensitized in a 25 μl mixture competing with the mAbs and 25 μl of uncultured type mAb or 25 μl of serum containing alloantibodies. After sensitization, the cells were washed 4 times with 200 μl of PBS volumes and resuspended in 50 μl of RPMI / FBS (final concentration = 1% v / v). In all experiments, an aliquot of cells (E-IgG) was used in the chemiluminescence or chemical fluorescence (QL) assay and an aliquot was assayed by flow cytometry to determine the level of IgG bound to erythrocytes .

Qumiol uminiscence assay or Chemical Fluorescence The PBMC were isolated by density gradient centrifugation of blood anticoagulated with EDTA collected from 6 normal donors. The PBMCs were washed 4 times with PBS containing 1% BSA-free globulin and then resuspended at 2 x 10 6 / ml in Hank's Balanced Salt Solution (HBSS) containing 25% RPMl and 2.5% FBS. The aliquots (100 μl) were distributed in opaque, white, 96-well flat-bottomed plates and incubated for 2 hours at 37 ° C in a humidified atmosphere of 5% CO 2 in air. The plates were then placed in a luminometer (Anthos Lucy 1, Labtech International, Uckfield, UK) and 100 μl of HBSS containing 4 x 10 ~ 4 M luminol (Sigma) and 20 μL of E-IgG was added to each well. The QL response was then monitored at 37C for 60 minutes.

Determination of IgG Linked to the Eri troci Aliquots of 25 μl of E-IgG were transferred to a V-bottom well plate, washed once with PBS, centrifuged to a pellet and resuspended in 50 μl FITC-anti-IgG F (ab) 2 (diluted 1/30 in PBS / BSA 1%). After 30 minutes at room temperature, the cells were washed once with 200 μl of PBS / BSA and kept on ice until analyzed by flow cytometry (EPICS XL-MCL, Coulter Electronics, Luton, UK). The fluorescence of the intermediate channel was recorded. Fluorescence of the intermediate channel was converted to IgG / cell molecules by using a normal curve which was prepared by adding 100 μl of 5% RiRi v / v cells to 900 μl of serially 2-fold dilutions of monoclonal IgGl anti-D of human (BRAD-5). The sensitized erythrocytes were washed 3 times with PBS / BSA and resuspended at 1% v / v in PBS / BSA. Aliquots of 25 μl were separated and analyzed by flow cytometry as described above. The remaining erythrocytes were counted, centrifuged in a pellet, lysed in a buffer containing Triton X-100 and the IgG in those used was determined by an ELISA as described by Kumpel (Kumpel, BM (1990). -isotopic method for the quantitation of red cell-bond immunoglobulin, Vox Sanguinis, 59, 34-39). The number of IgG molecules bound by erythrocyte was deduced from the IgG concentration and the number of erythrocytes from which each lysate was prepared. A normal curve was then plotted by comparing the fluorescence intensity with the number of IgG molecules bound by erythrocytes.

Li sis of the complement mediated by the bodies of the CAMPATH-1 series 100 ml of blood was defibrated from the veins of a healthy volunteer and the components were separated by density gradient centrifugation using Ficoll-Paque Plus (Pharmacia). The serum and the mononuclear cell layers were separated to new tubes. The cells were diluted in Iscove's modified Dulbecco's medium (IMDM) and collected by centrifugation. The cells were washed twice in IMDM while they were combined in a pellet which was resuspended in 200 μl of IMDM. 900 μCi of sodium chromate [51 Cr] were added and the cells were incubated at 37 ° C for 40 minutes. 10 ml of IMDM were added and the cells formed into pellets. The cells were washed twice and resuspended in IMDM at approximately 6 x 10 6 cells / ml. Aliquots of 50 μl of the labeled cells were added to the antibody samples in 50 μl of IMDM in wells of 96-well plates. 100 μl of retained serum diluted 1: 1 with IMDM was added to each well and the plates were incubated at 37 ° C for 1 hour. Plates were centrifuged and supernatants were sampled and the relative amounts of 51 Cr released in a? -counter were measured. The level of spontaneous release was obtained from samples where no antibodies were added and a measurement of the total amount of 51 Cr available for the release was found from similar samples taken after resuspending the cells. The specific 51Cr release% was calculated from the formula: (sample accounts - spontaneous accounts) x 100 (total accounts - spontaneous accounts) The intermediate and normal deviations of the samples in triplicate were recorded in a diagram. For the inhibition of complement lysis, the antibody samples contained a constant amount (6.25 μg / ml final concentration) of Gl from the CAMPATH-1 series and increasing amounts of G2? A from the CAMPATH-1 series.

ADCC mediated by the antibodies of the CAMPATH-1 series Peripheral blood mononuclear cells were prepared as described above. After washing, the cells were resuspended in IMDM supplemented with 5% FBS and transferred to the flask which has been coated with the CD3 antibody. Cells were cultured at 37C, C02 5% for three days. 5% of the cells were labeled with 51 Cr for use as target cells, washed and resuspended to 6 x 105 cells / ml in IMDM + 5% FBS. Aliquots of 50 μl were added to the wells of the 96-well plates containing 50 μl of samples of the antibodies in IMDM + 5% FBS. The target cells and antibodies were incubated at 37C for 1 hour, the RBCs were added as carriers and the cells were pelletized. The cells were washed twice in IMDM. The remaining mononuclear cells were collected by centrifugation and resuspended at 4 x 106 cells / ml in IMDM + 5% FBS and 150 μl was added to each well to resuspend the target cells in the process. This gives an effector ratio: goal of 20: 1. The cells were gently centrifuged and placed in a tissue culture incubator for 6 hours. Samples of the supernatant were formed and the specific 51Cr release was determined as described above. The average specific release values for the duplicate samples were plotted against the final concentrations of the antibodies.

Example 1 - Generation and basic characterization of antibodies Mutations selected to eliminate effector functions are shown in Table 1 (Figure 15). The mutation? A made in the IgG1 and IgG2 genes introduces the IgG4 residues at positions 327, 330 and 331. Similarly, the IgG2 residues at positions 233-236 were introduced into IgG1 and IgG4 but, since the IgG2 has a deletion in 236 where the other subclasses have a glycine residue, the mutation was made by omitting (? B) or including (? C) G236. The vectors that allow the expression of the VH DNA of the CAMPATH-1 series or of the Fog-1 series in conjunction with the non-cultured or mutant constant region genes were co-transfected with the appropriate light chain expression vectors in the cells of rat myeloma. Stable transfectants were isolated, expanded and Ab purified from the supernatant in protein A-agarose. The CAMPATH-1H was selected as providing a token targeting system to study complement-mediated lysis and in vitro cells. For Ab of the Fog-1 series, a precipitate formed after purification but, once this has been removed by filter sterilization, no additional precipitation was observed. Ab concentrations were estimated from the absorbance at 280 nm and adjusted where necessary after an ELISA which measures the relative amounts of the K chain present. The Ab were subjected to reduction with SDS-PAGE. Each sample showed two bands with apparent molecular weights of approximately 25 and 55 kDa which represent the expected sizes of the light and heavy chains. There was no discernible difference in the size between the heavy chains of each series of the Ab but both chains of the Ab of the Fog-1 series appeared to be slightly smaller than their counterparts of the CAMPATH-1 series. The fact that the heavy chain within each series appeared to have the same apparent molecular weight indicates that the mutations did not cause any extensive difference in the glycosylation of the proteins. For Ab with specificity by CAMPATH-1, the yield after purification varied from 0.6 to 9 μg / ml of supernatant while the yield of the soluble Fog-1 Ab were between 3 and 20 μg / ml. There was no correlation in the condition of the purification yields for the two sets of antibodies suggesting that none of the mutations affected the production of Ab or their ability to bind protein A. The specificities of the two series of Ab Then they underwent a test. The Ab of the CAMPATH-1 series were shown to compete with the clinical grade CAMPATH-1H in the binding of the anti-CAMPATH-1 idiotype mAb, YID13.9. The Ab of the Fog-1 series were able to bind the RBC RhD + in the presence of the anti-human IgG Ab as crosslinking reagents. Similarly, the IgG subclasses of the Ab of the Fog-1 series were examined by coating the RBC RhD + with the different Ab and observing the agglutination pattern using anti-Glm (a), anti-IgG2 or anti-IgG4 Ab. as the crosslinking Ab. The result indicated that the antibodies were from the correct subclasses. The agglutination of the RBC RhD + by the IgGl of the Fog-1 and anti-Glm (a) series, by the IgG2 of the Fog-1 and anti-IgG2 series and by the IgG4 of the Fog-1 and anti-IgG4 series then it was carried out in the presence of Ab in excess of the CAMPATH-1 series. The Ab of the CAMPATH-1 series were able to inhibit the agglutination, by competing for the cross-linking reagent, only where they were of the same subclass as the Ab of the Fog-1 series, to verify in this way their subclasses.

Example 2 - Fc? Rl binding RBCs with approximately 30,000 RhD sites per cell (R2R2) were coated with each of the 11 Ab of the Fog-1 series over a range of concentrations and added to the expressing transfectants the Fc? RI of human, B2KA, cultivating in the wells. After incubation, the excess RBCs were washed and the percentage of B2KA cells in the form of rosettes was recorded by the RBCs (Figure 1). For Gl and Gl? A, where the IgG4 residues are included in portions 327, 330 and 331, similar levels of rosette formation were achieved, with the maximum possible rosette formation occurring when the RBCs were coated with the Ab in approximately 0.1 μg / ml, a concentration at which the Ab of the Fog-1 series would be expected to occupy approximately one third of the RhD sites. Slightly higher concentrations of G4 were necessary to obtain the same levels of rosette formation. No rosettes were formed when RBC coated with Ab Gl and G4 containing the β and βc mutations or Ab G2 were used. In the experiment shown in Figure 1, the coating concentration tested was 10 mg / ml, predictable to correspond to approximately 90% occupancy of the RhD sites. The experiment was repeated using coating concentrations of up to 80 mg / ml, to essentially saturate the RhD sites, and still no rosettes were observed for G2 and the Ab containing the Db or De mutations and to thereby incorporate residues of IgG2 in the lower joint region.

This indicates that, even when the RBCs were coated with these Ab at the maximum density for this antigen, there was insufficient interaction of IgG / Fc? GRI for the formation of rosettes. Centrifugation of sensitized RBC and B2KA cells, together, before observing rosettes on a microscope slide was found to give a higher proportion of rosettes than incubation of the cells in the wells so that this method was used to investigate the inhibition of rosette formation. The RBC R2R2 were coated with a mixture of 1 mg / ml Gl of the Fog-1 series and different amounts of G2Da of the Fog-1 or G4Db series of the Fog-1 series before mixing with the B2KA cells. When 1 μg / ml of Gl of the Fog-1 series alone was used, the coated RBCs formed rosettes in 95% of the B2KA cells while the sensitization in the presence of 64 mg / ml of G2α or G4βb completely suppressed the formation of rosettes (data not shown). Ab binding of both series to two different cell lines, which express the FcγRI cDNA on its surface, was measured by fluorescent staining. Figure 2 shows the representative experiments. The level of Fc? RI expressed on the surface, detected using Ab CD64, was higher for the 3T3 transfectants than for the B2KA line and this is reflected in the upper signals obtained when the link is measured by Fe. For both series , the Ab of Gl and Gl? a were bound to the receptor with the same apparent affinity indicating that the mutations at positions 327, 330 and 331 did not significantly affect the interaction. The bond of the Ab G3 was about three times lower than that of the Ab Gl and Gl? A. Little binding was observed for the Ab G2 or any other Ab mutant, suggesting that the? B and? C mutations in IgG1 and IgG4 were sufficient to reduce the binding to Fc? RI by at least 104 fold. The Ab containing the? C mutation, especially Gl? C, showed a small degree of binding to the Fc? RI at the highest concentrations subjected to a test if the fluorescence level is compared to the basic level or to the Ab equivalents with the mutation? b. If the histograms of the fluorescence intensity are covered, as shown in Figure 3 by the highest concentrations of the Ab of the CAMPATH-1 series and the B2KA cells, the records for Gl and Gl? A coincide. There is a clear difference between the histograms for Ab Gl? B and Gl? C.

Example 3 - Activation of Fc? RI as measured by chemiluminescence In order to measure the functional activity through the Fc? RI / II, the chemiluminescence (QL) response of the monocytes to the RBCs sensitized with the Ab of the Fog-1 series was measured and recorded in a diagram in relation to the number of Ab molecules joined by the RBCs (Figure 4). A difference between the Ab Gl and Gl? A is observed with higher amounts of Ab but both give higher responses than the Ab G4 through the range of Ab concentrations. A significant activation is achieved by Ab Gl? C and, to a lesser degree, by Gl? Ac and G4? C but the other Ab do not give any response. The ab, which are known to be deficient in the activation of the FcγRI from the previous section, were mixed in increasing concentrations with a constant amount of Gl from the Fog-1 series and used to sensitize the RBCs. The response of QL to the RBCs is shown in Figure 5. When comparing the response of QL to that obtained when the Gl is activated alone, it seems that six of the eight Ab inhibit the reaction to a degree that predicts assuming that the mutants displace the active Gl of the RBCs in proportion to their relative concentrations. For G2, the inhibitory effect is retained in that it is necessary approximately three times more G2 to give the same amount of inhibition. The Gl? C inhibits to approximately the same degree as the other mutants, except that the response is not reduced to zero. Two papers, which have discussed the usefulness of chemiluminescence in predicting the severity of in-vivo pathology, are Hadley (1995) Transfusion Medicine Revi9: 302-313 and Hadley et al. (1998) Br J Obstet Gynaecol 105 : 231-234. In these trials, a result above 30% of chemiluminescence produced by the control of monoclonal antibodies BRAD-5 of in vivo pathology in HDN would be predictable. In this way, those antibodies which can block levels below 30% should be suitable for therapy. One of the Ab mutants, G2α of the Fog-1 series, was tested for its ability to inhibit the QL response to sera containing the clinically significant Ab. The sera contained anti-RhD or antiC + D Ab and, in the absence of an inhibitor, gave the QL responses greater than 30% on this scale which is indicative of the serious haemolytic disease of the newborn and the need for intrauterine transfusions. . The sera were mixed with different concentrations of G2α, the mixtures were used to sensitize the RBCs and the responses of the monocytes were measured (Figure 6). The addition of Ab G2? A reduced the QL signals due to all five anti-RhD sera below the 30% limit. The amount of Ab needed to achieve this varied from 16 - 260 μg / ml, the range that presumably reflects the difference in amounts and affinities of the anti-RhD Ab in the serum. There are two control sera. The anti-K serum can not be blocked at all by the G2α as its reactivity is directed towards a different antigen in the RBCs. Only part of the anti-C + D serum activity could be inhibited by G2? A.

Example 4 - Activity in complement lysis Figure 7 shows that all mutations made to antibodies Gl and G2 of the CAMPATH-1 series dramatically reduced their ability to mediate complement lysis. When the assay was carried out using a constant amount of Gl and different amounts of G2α (Figure 8), the G2α antibody was able to block the destruction of the PBMCs by Gl of the CAMPATH-1 series.

Example 5 - Activity in the ADCC The ability to mediate ADCC was measured by CAMPATH-1 series antibodies using human PBMC as target cells (Figure 9) and for Fog-1 series antibodies using RBC RhD + as target cells (Figures 10 and 10b) . Figure 9 shows the mixed capacities of the CAMPATH-1 series antibodies in the ADCC, with some of the mutants that have very low activities. Figures 10 and 10b show that the mutants of Gl? Ab, Gl? Ac, G2? A, G4? B and G4? C of the Fog-1 series were unable to withstand any destruction of the RBCs. In Figure 10, some lysis of the RBCs sensitized with G2 and G4 was observed, but these antibodies have no apparent activity in the assay of Figure 10b. This demonstrates the observation that the degree of lysis can be dependent on the donor of the effector cells and can even vary when the effector cells taken from the same donor at different times are used. However, for the mutants listed above, no activity above the basic levels has been observed although a range of effector cell donors has been tested.

Some of the antibodies of the Fog-1 series were used to try to inhibit the ADCC of the RBC RhD + by the Gl of the Fog-1 series (Figures 11 and 11b) and by a clinical sample of anti-RhD serum (Figure 12 ). The figures show that all the antibodies subjected to a test were able to inhibit the ADCC when they were mixed with the active antibodies before the sensitization of the RBCs. The Gl? B, Gl? Ab, Gl? Ac, G4? B and G4? C mutant antibodies of the Fog-1 series were particularly effective in blocking ADCC.

Example 6 - Link of Fc? RII Figures 13, 13b and 14 show the binding of antibody complexes of the Fog-1 series to the cells carrying the FcγRIIa 131H / H, FcγRIIa 131R / R and Fc? RIIbl * respectively. It is necessary to form antibody complexes when measuring the binding to these receptors due to their low affinity for the individual antibody molecules. Fc? RIIa 131H / H is an allotype of Fc? RIIa which is expected to strongly bind antibodies to IgG2 and, in fact, Gl and G2 show strong binding activity (Figure 13). The addition of the mutations to these two antibodies seems to give a gradual reduction in the binding levels and the Gl? C and Gl? Ac antibodies have only basic levels of binding exhibited by the G4 antibodies. Figure 13b shows that the antibodies have different relative activities when they bind to the 131R alotype of FcγRIIa but the mutations made to the non-cultured Gl antibody again decrease the binding to the receptor. All antibodies show significantly more binding to the inhibitory receptor, FcγRIIbl *, than the negative control samples of crosslinking F (ab ') 2 alone or an agglucosyl IgGl antibody made complex with F (ab') 2 ( Figure 14). Although the binding of most of the mutants is reduced relative to the corresponding uncultivated antibodies, some mutants show a link within two times of that exhibited by the non-cultured Gl antibodies.

Example 6b - FcgRIII linkage Figures 14b and 14c show the binding of antibody complexes of the Fog-1 series to cells carrying the Fc? RIIIb of the NAl and NA2 allotypes respectively. For both allotypes, the bond is observed by the Gl antibody and, to a lesser degree, the Gl? A and Gl? C antibodies. No binding is observed by the other mutant antibodies since they show similar levels of fluorescence to the negative control samples of the crosslinking F (ab ') 2 alone or an agglucosyl IgGl antibody made complex with the F (ab') 2 .

Example 7 - Production of anti-HPA-la antibodies The VH and the V? of the anti-HPA-la scFv (Griffin, HM and Ouwehand, WH (1995) A human monoclonal antibody specific for the leucine-33 form of the platelet glycoprotein Illa from a V gene phage display library, Blood 86, 4430-4436) they amplified and each one was linked to the leader sequence of the vector M13VHPVR1 (Orlandi et al., 1989) by the overlap extension PCR described previously. The DNA, 3 'of the VH in M13VHPCR1 and representing the 5' end of the intron VH-CH, was ligated in a manner similar to the leader DNA / VH. The product was cloned as a HindlII-BamHI fragment into the expression vector of IgG1 and IgG2 to replace the existing variable region fragment and to give the vectors pSVgptB2VHHuIgGl and pSVgptB2VHHuIgG2. The DNA leader / V? was ligated in the structure to the constant region DNA of the chain? of human of the Kern "Oz" allotype (Rabbitts, T. H .. Forster, H. and Matthews, JG 1983. Mol. Biol. Med ".1: 11), taken from an existing expression vector (Routledge, EG, Lloyd, I, Gorman, SD, Clark, M. and Waldmann, H. 1991, Eur. J. Immunol, 21: 2111.) The complete gene was cloned into M13 as a fragment i? IndIII-BamHI and the enhancer of murine heavy chain of pSVhyg-HuCK (Orlandi et al., 1989) added 5 'of the gene using adapters so that the total insert could be transferred to pSV2neo (Southern, P. J. and Berg. P. 1982. J. Mol. Appl. Genet 1: 327) as a Bamñl fragment. The vector was designated pSVneoB2V? HuC ?. The expression vectors were transfected into the rat myeloma cell line YB2 / 0, the transfectants were selected and the antibody was purified as described above. These antibodies B2IgGl and G2IgG2 can be used as control antibodies. Once the constant, null, preferred regions have been selected, the B2 VH HindI11-BaJ? HI fragment can be introduced into the expression vectors carrying the appropriate constant region genes to replace the existing variable region fragment. The heavy chain expression vectors can then be co-transfected with pSVneoB2V? HuC? In myeloma cells and antibodies can be purified for use.

Example 8 - Therapeutic use of the linker molecule A therapeutic molecule according to the present invention can be used to treat pregnancies complicated by the alloimmunization of HPA-la, for example by intravenous administration to the mother, thereby relying on placental transfer (for example, by means of FcRn ) to provide a therapeutic dosage to the fetus. An alternative is the direct administration to the fetus by sampling umbilical vessels, percutaneously. This procedure is currently performed at FAIT to provide compatible platelet transfusions. Due to the short survival of transfusion platelets, the procedure may have to be repeated many times during the course of a pregnancy. However, this is dangerous, with a risk of fetal loss of 0.5% / procedure. However, fetal administration of a therapeutic antibody would have the advantage that a much lower dosage is likely to be required, and therefore a combined approach using the molecules of the present invention in conjunction with platelet transfusion can be considered as a first step in therapy. This approach can reduce or eliminate the need for additional platelet transfutions prior to delivery.

Summary The activities of the antibodies are summarized in Table 2 (Figure 16). As can be seen, binding molecules have been produced which have reduced capacity to bind to FcγRI, FcγRIIa 131H / H, FcγRIIa 131R / R, FcγRIIIb NAl and FcγRIIIb NA2; are unable to activate chemiluminescence of monocytes; they can not mediate complement lysis and are not activated in the ADCC. However, the binding molecules retain the binding to the inhibitory receptor, FcγRIIb. Other mutations previously used to sensitize the effector functions, such as the separation of the glycosylation site in the CH2 domain to make the aglycosyl antibodies, can also eliminate the binding to this receptor, which may not be desirable. It has been shown that the selected mutants are capable of completely inhibiting rosetting of the cells carrying the FcγRI by the Gl of the Fog-1 series; the response of monocytes to RBCs sensitized with the Gl of the Fog-1 series; the response of monocytes to RBCs sensitized with polyclonal anti-RhD; the destruction of PBMC by lysis of the complement with Gl of the CAMPATH-1 series; the destruction of RBCs through the ADCC with Gl from the Fog-1 series; the destruction of RBCs by ADCC with anti-RhD polyclonal serum. The results herein demonstrate that alteration of even an individual residue in a CH2 domain of IgG to correspond to a different subclass can lead to different activities. Thus, for the three pairs of mutants Db and De: Gl? B and Gl? C, Gl? Ab and Gl? Ac, G4? B and G4? C. Within each pair, the antibodies differ only by the absence (? B) or the presence (? C) of G236. However, for most of the functions measured here, the? B and? C antibodies have different activities. The Db mutants are more active in the binding to FcγRIIa 131H / H whereas the De mutants are more active in the binding to the FcγRI, the binding to the FcγRIIIb NAl and NA2, the activation of monocytes and the ADCC. The region where the β and c mutations are made is known as the lower joint or the joint junction region and is likely to have an extended structure, which connects the joint to the rest of the CH2 domain. The addition or deletion of a residue from this region presumably alters the alignment of the lower joint residues relative to the sites of receptor interaction in the rest of the CH2 domain. However, it should be emphasized that the effect of mutations may not always be predictable from the activities of uncultivated antibodies, but it will depend on the new context (based on the 'mixed' subclasses of the IgG) in which the mutation occurs. An example is in the complement lysis assay where IgG2 antibody activity is only about three times lower than that of IgG1 but the introduction of IgG2 residues in IgGl (Gl? By Gl? C) eliminates the lysis Similarly, IgGl and IgG2 show a bond equal to Fc? RIIa 131H but the activities of Gl? B and Gl? C are 50 and 10 times lower, respectively. In the ADCC assays of Figures 9 and 10, IgG2 and IgG4 gave low but measurable, similar lysis levels. The substitution of the residues between IgG2 and IgG4, as well as in IgGl, reduces the activity. These data suggest that uncultivated antibodies from different subclasses of human IgG and, presumably, mutant antibodies can use different residues in binding to other molecules to activate activities.

It is noted that in relation to this date, the best method known to the applicant to carry out the aforementioned invention, is that which is clear from the present description of the invention.

Having described the invention as above, the content of the following claims is claimed as property.

Claims (31)

1. A binding molecule which is a recombinant polypeptide comprising: (i) a binding domain capable of binding a target molecule, and (ii) an effector domain having an amino acid sequence substantially homologous to all or part of a constant domain of a human immunoglobulin heavy chain; wherein the binding molecule is capable of binding the target molecule without activating complement-dependent lysis, significant, or destruction mediated by target cells, characterized in that the effector domain is - capable of specifically binding the FcRn and / or Fc? RIIb, and - a chimeric effector domain which is derived from two or more heavy chain CH2 domains of human immunoglobulin including a first heavy chain CH2 domain, human immunoglobulin wherein 2, 3 or 4 amino acids have been modified in at least 1 region of the CH2 domain to the corresponding amino acids of a second heavy chain, human immunoglobulin CH2 domain, wherein the region is selected from the 2 discrete regions numbered residues 233-236 and 327-331 according to the US numbering system, and wherein in each case human immunoglobulin is selected from IgGl, -IgG2 and IgG4.

2. A binding molecule according to claim 1, characterized in that the first human immunoglobulin is selected from IgG1, IgG2 and IgG4 and the second human immunoglobulin is selected from IgG2 and IgG4.

3. A linker molecule according to claim 1 or claim 2, characterized in that 2 amino acids are modified in 1 region of the CH2 domain to the corresponding amino acids of a second heavy chain CH2 domain, of human immunoglobulin.

4. A linker molecule according to any of the preceding claims, characterized in that at least 2 amino acids are modified in each of the 2 discrete regions of the CH2 domain to the corresponding amino acids in the corresponding region in a second and a third chain CH2 domain heavy, human immunoglobulin respectively.

5. A linker molecule according to any of the preceding claims, characterized in that the effector domain shares at least about 90% sequence identity with the first heavy chain CH2 domain of human immunoglobulin.

6. A linker molecule according to any of the preceding claims, characterized in that it comprises a heavy chain, human immunoglobulin CH2 domain having one or more of the following amino acids or deletions at the positions established according to the numbering system of USA: Posn Amino Acid 233 P 234 V 235 A 236 (without residue) or G 327 G 330 S 331 S

7. A linker molecule according to any of the preceding claims, characterized in that it comprises a heavy chain, human immunoglobulin CH2 domain having one or more of the following blocks of amino acids or deletions at the positions established according to the system of US numbering: 233P, 234V, 235A and without residue at 236; or 233P, 234V, 235A and 236G; and / or 327G, 330S and 331S.

8. A linker molecule according to any of claims 5 to 7, characterized in that the effector domain is selected from Gl? Ab, G2? A or Gl? Ac.

9. A linker molecule according to any of the preceding claims, characterized in that it further comprises modifications to render the allotypic molecule substantially nil.

10. A linker molecule according to any of the preceding claims, characterized in that the effector domain has a reduced affinity for FcγRI, FCαRIIa or FcγRIII and a reduced ability to mediate complement lysis by comparison with the first or the second heavy chain CH2 domain, of human immunoglobulin.

11. A linker molecule according to claim 10, characterized in that the effector domain has retained an affinity for Fc? RIIb.

12. A binding molecule according to any one of the preceding claims, characterized in that the binding domain is derived from a source different from the effector domain.

13. A linker molecule according to any of the preceding claims, characterized in that the binding domain is selected from the binding site of an antibody; an enzyme; a hormone; a receiver; a cytokine or an antigen; a ligand or an adhesion molecule.

14. A binding molecule according to any of the preceding claims, characterized in that the binding domain is capable of binding any of: the RhD antigen of the lymphocytes; an HPA alloantigen of platelets; a neutrophil antigen; a T cell receptor; integrin; GBM collagen; Der Pl; HPA-la; VAP-1; laminin; Lutheran; platelet glycoprotein VI; platelet glycoprotein Ia / IIa.

15. A linker molecule according to claim 14, characterized in that the binding domain is selected from that of CAMPATH-1 and FOG1; OKT3; B2 (anti-HPA-la); VAP-1; murine anti-a3 (IV) NC1; YTH12.5 (CD3); 2C7 (anti-Der p I); anti-laminin; anti-Lutheran

16. An isolated nucleic acid, characterized in that it comprises a nucleotide sequence encoding the effector domain of the binding molecule according to any of the preceding claims.

17. A nucleic acid, according to claim 16, characterized in that the nucleotide sequence encodes a linker molecule according to any of the preceding claims.

18. A nucleic acid according to claim 16 or claim 17, characterized in that it is a duplicatable vector.

19. A nucleic acid according to claim 18, characterized in that the nucleotide sequence is operably linked to a promoter.

20. A host cell, characterized in that it comprises or transformed with the vector of claim 19 or claim 20.

21. A process for the production of a linker molecule according to any of claims 1 to 15, the process is characterized in that it comprises the step of modifying a nucleotide sequence encoding a first heavy chain CH2 domain, of human immunoglobulin such that 2, 3 or 4 amino acids in at least 1 region of the CH2 domain correspond to an amino acid of a second heavy chain CH2 domain, of human immunoglobulin, wherein the region is selected from the 2 discrete regions the numbered residues 233-236 and 327-331 according to the US numbering system, and wherein in each case the human immunoglobulin is selected from IgG1, IgG2 and IgG4.

22. A process according to claim 21, characterized in that 2 amino acids are modified in 1 region of the CH2 domain to the corresponding amino acids of a second human immunoglobulin heavy chain CH2 domain.

23. The use of a binding molecule or a nucleic acid according to any of claims 1 to 19 for linking an objective molecule with the binding molecule.

24. The use according to claim 23, wherein the target molecule is FcγRIIb, the linkage causing the inhibition of one or more of: activation of B cells; degranulation of mast cells or Mastzelle; phagocytosis

25. The use according to claim 24 to prevent, inhibit or otherwise interfere with the binding of a second molecule to the target molecule.

26. The use according to claim 25, wherein the second binding molecule is an antibody.

27. The use according to claim 25 or claim 26 wherein the target molecule is selected from: the RhD antigen of the lymphocytes; an HPA alloantigen of platelets; a neutrophil antigen; a T cell receptor; an integrin; a GBM collagen; Der Pl; HPA-la; VAP-1; laminin; Lutheran; platelet glycoprotein VI; platelet glycoprotein Ia / IIa.

28. The use according to any of claims 24 to 27 for the treatment of a patient for a disorder selected from: graft-versus-host disease; host disease versus graft; rejection of organ transplantation; rejection of bone marrow transplantation; autoimmunity such as vasculitis, autoimmune hemolytic anemia, autoimmune thrombocytopenia and arthritis; alloimmunity such as fetal / neoantal alloimmune thrombocytopenia; asthma and allergy; inflammatory, chronic or acute diseases such as Crohn's; HDN; Goodpastures, sickle cell anemia, coronary artery occlusion.

29. The use according to any of claims 23 to 28 wherein the binding molecule is administered to a patient, or optionally in cases where the patient is an unborn infant, to the patient's mother.

30. A pharmaceutical preparation, characterized in that it comprises a linker molecule according to one of claims 1 to 15, or a nucleic acid according to any of claims 17 to 19, plus a pharmaceutically acceptable carrier.

31. An oligonucleotide, characterized in that it is selected from: M022BACK: 5 'TCT CCA ACA AAG GCC TCC CGT CCT CCA TCG AGA AAA 3' M022: 5 'TTT TCT CGA TGG AGG ACG GGA GGC CTT TGT TGG AGA 3' M07BACK: 5 'TCC TCA GCA CCT CCA GTC GCG GGG GGA CCG TCA GTC 3 'M021: 5' GAC TGA CGG TCC CGC GAC TGG AGG TGC TGA GGA 3 'SUMMARY OF THE INVENTION Binding molecules are described which are recombinant polypeptides comprising: (i) a binding domain capable of binding a target molecule, and (ii) an effector domain having a substantially homologous amino acid sequence of all or part of a constant domain of a human immunoglobulin heavy chain; characterized in that the binding molecule is capable of binding the target molecule without activating complement-dependent lysis, significant, or destruction mediated by target cells, and more preferably wherein the effector domain is capable of specifically binding the FcRn and / or Fc? RIIb. These are generally based on chimeric domains which are derived from two or more human immunoglobulin heavy chain CH2 domains. In the preferred embodiments, regions 233-236 and 327-331, are modified as are the additional residues to render the alotypic molecule, null. The link domain can be derived from any source appropriate to the application (usually clinical) for the molecule and may be of, for example, an antibody; an enzyme; a hormone; a receiver; a cytokine or an antigen; a ligand and an adhesion molecule. Nucleic acids, host cells, production processes and materials and uses are also described, for example to inhibit the activation of B cells; degranulation of mast cells or Mastozelle; phagocytosis, or to inhibit the binding of a second molecule binding to the target molecule. Pharmaceutical preparations are also described. "