WO1992011031A1

WO1992011031A1 - TARGETTING IgE EFFECTOR CELLS TO TUMOR CELLS

Info

Publication number: WO1992011031A1
Application number: PCT/GB1991/002303
Authority: WO
Inventors: Kevin Stuart Johnson; David John Chiswell
Original assignee: Cambridge Antibody Technology Limited
Priority date: 1990-12-21
Filing date: 1991-12-23
Publication date: 1992-07-09
Also published as: GB9027767D0; JPH06503956A; AU9107191A; EP0563198A1

Abstract

The invention relates to binding molecules capable of recruiting IgE effector functions. The invention provides molecules having a first binding domain capable of binding to a tumour antigen and a second binding domain capable of binding to a Fcε receptor on cells of the immune system, whereby the molecule is able to mediate IgE effector mechanisms against a tumour cell expressing the antigen.

Description

TARGETTING IgE EFFECTOR CELLS TO TUMOR CELLS

The present invention relates to binding molecules. In particular, the invention relates to binding molecules capable of recruiting IgE effector functions, methods for their preparation and their use in the destruction of cancer cells and multicellular tumours in host mammals, particularly humans. In particular, the present invention relates to molecules containing a binding domain capable of binding one or more type of Fc€ receptor.

The invention further relates to pharmaceutical formulations and methods for their preparation in which said binding molecules are active constituents. In order that the utility of IgE effector mechanisms be more fully understood, it is first necessary to describe some aspects of the biology of IgE.

IgE consists of two heavy and two light chains. The light chains are derived from the ubiquitous kappa and lambda gene families which encode the light chains of all antibody isotypes. The epsilon ( 6 ) heavy chain is definitive for IgEs, and comprises an N-terminal variable domain VH, and four constant domains C_ζ-l-Cr-4. As with other antibody isotypes, the variable domains confer antigen specificity and the constant domains recruit the isotype-specific effector functions. IgE differs from the more abundant IgG isotypes, in that it is unable to fix complement and does not bind to the Fc receptors FcRI, RII and RIII expressed on the surfaces of mononuclear cells and neutrophils. However, it is capable of very specific interactions with the 'high affinity' receptor on mast cells and basophils (Fee RI, Ka. 10^ ~l), and with the 'low affinity' receptor, Fee RII (KA. IO¹-⁷ M^~l), expressed on inflammatory cells (macrophages, eosinophils, platelets) and T and B lymphocytes. The sites on IgE responsible for these receptor interactions have been mapped to peptide sequences on the Ce chain, and are distinct. The Fee RI site lies in a cleft created by residues between Gin 301 and Arg 376, and includes the junction between the Ce2 and Ce3 domains [Helm, B. et al. (1988) Nature 331, 180- 183] . The Fee RII binding site is located within Ce3 around residue Val 370 [Vercelli, D. et al. (1989) Nature 338, 649-651]. A major difference distinguishing the two receptors is that Fee RI binds monomeric Ce, whereas Fee RII will only bind dimerised ce, i.e. the two Ce chains must be associated. Although IgE is glycosylated in vivo, this is not necessarily for its binding to Fee RI and Fee RII. Binding is in fact marginally stronger in the absence of glycosylation [Vercelli, D. et al. (1989) et. supra] . In the developed world, IgE isotypes are most frequently associated with allergies, in which IgE antibodies directed against innocuous antigens such as pollen and house dust, mediate Type 1 hypersensitivity reactions leading to the symptoms of asthma and hay fever. Virtually all current research on IgE is directed towards ways of inhibiting these inappropriate immune responses [for example, Helm, B. et al. (1988) et. supra., Vercelli D. et al. (1989) et. supra.]. However, it is generally believed that IgE has evolved to combat parasitic worm infections (largely because it is well-suited to this purpose, and no other). The efficacy of IgEs in combating infections by multicellular pathogens is exemplified by schistosomes, helminth parasites of man and his domestic animals in the developing world. Invading schistosomes (called schistosomula) are resistant to the action of cycotoxic T cells [Butterworth, A.E. et al. (1979) J. Immunol 122, 1314-1321], but can be destroyed by inflammatory cells in antibody-dependent cell-mediated cytotoxicity (ADCC) reactions [Capron, A. et al. (1986) Immunol. Today 7, 15- 18] . Iι vitro, the ADCC reaction with immune sera is at least in part dependent upon IgE, since selective depletion of IgE, inclusion of an irrelevant myeloma IgE competitor and anti-F receptor antibody all dramatically decrease killing capacity by eosinophils, macrophages and platelets. The cytotoxic responses initiated by IgE are extremely violent. The effector properties of IgE are summarised in Tables 1 and 2. Certain IgG subclasses may also direct ADCC by eosinophls (but not platelets or macrophages), but only if the eosinophils are 'activated', for example, from donors with allergic disorders [Khalife, J. et al. (1989) J. Immunol. 142, 4422-4427]. However, inflammatory cells can respond quite differently depending upon which antibody isotype is bound. For example, IgG-triggered eosinophils release cationic proteins, but little eosinophil peroxidase, whereas for IgE-triggered cells, the reverse is true [Capron, A. et al. supra.]. These findings serve to reinforce the concept that IgE as opposed to other isotypes, calls down unique and destructive immune responses that are ideally suited to some therapeutic applications.

A common property shared by worms and allergens, the preferred targets of IgE responses, is that both tend to be larger than the cells of the immune system. Destruction of large foreign objects poses a difficult problem for the immune system, since the usual means of immune disposal, phagocytosis, is, by itself, ineffective. The present applicants have realised that a solid tumour poses a problem similar to that of a large multicellular parasite, and that an IgE-like response to antigens of the tumour would effect the tumours' destruction. The present applicants have further realised that rather than inhibiting IgE effector functions, therapeutic agents capable of actively recruiting these functions in vivo will have considerable utility. One aspect of the present invention relates to the use of such antigen-specific binding molecules incorporating sequences capable of recruiting IgE effector functions in vivo. A second aspect of the present invention is that IgE effector functions will be particulary effective in destroying cancer cells and solid tumours in situ, and in explants of the hosts tissue intended for reintroduction into the body of the host.

The use of monoclonal antibodies to kill cancer cells has met with limited success, as the efficacy of the treatment depends upon accessibility of the cancer cells to circulating antibody and the effectiveness of killing by natural effector mechanisms or toxins or radioisotopes conjugated to the antibody [Reviewed by: Menard, S., Canevari, S. and Colnaghi, M.I. (1990) Int. J. Biol. Markers. 4, 131-134. Hellstrom, I and Hellstrom, K.E. (1989) Int. J Rad. Appl. Instrum. 16, 613-616] . There is also a problem caused by escape of malignant cells which no longer express the target- antigen and are therefore refractory to killing. Most antibodies that have been used to date for cancer therapy are of the isotype IgG.

IgE effector mechanisms have not been used in cancer therapy, but the present applicants believe that they would be surprisingly efficacious, particulary against solid tumours. IgE effector mechanisms are exceedingly cytotoxic and appear to have evolved specifically to deal with large intransigent foreign pathogens or allergens. Moreover, although the inflammatory processes are localised, they are not confined to the surface of the target, rather to its vicinity. This will result in the destruction of some normal cells at the site of the tumour, but an advantageous aspect is that this feature will make breakthrough by variant cells less likely, as long as the majority of cells express the antigen recognised by the antigen-specific binding domain of the molecules herein provided.

To enable IgE effector mechanisms to be targeted to a specific site, it is necessary for a specific binding molecule to comprise or be associated with sequences capable of a specific interaction with Fee receptors. Preferably, an antigen binding domain and sequences capable of a specific interaction with Fee receptors are covalently associated using for example, chemical coupling or recombinant DNA techniques. An antigen binding domain may be naturally associated with sequences capable of a specific interaction with Fee receptors, such as in an IgE antibody.

The binding molecule may be any compound capable of a binding interaction with a desired target e.g. a tumour antigen. For example, the binding molecule can be a peptide, a receptor or fragment thereof, or a ligand. Most preferably, the binding molecule may comprise one or more antibody variable domains. Antibody variable domains may derive from any source, including monoclonal antibodies produced by hybridoma technology, and those produced wholly or in part by recombinant DNA methodology [International Patent Application PCT/GB91/01134; Winter, G. and Milstein, C. (1991) Nature 349, 293-299]. It will be apparent to those skilled in the art that there are several ways in which nucleic acid fragments encoding antibody variable domains can be generated, and linked to nucleic acid fragments encoding sequences capable of a specific interaction with Fee receptors, such that one or more antibody variable domains and sequences capable of a specific interaction with Fee receptors are produced as fusion proteins. Preferably, heavy and light chain variable domains are produced by Polymerase Chain amplification (PCR) of mRNA or DNA, using appropriate oligonueleotide primers and ligated into expression vectors [Orlandi, R. et al. (1989) Proc. Natl. Acad. Sci USA 86 3833-3837; Winter G. and Milstein, C. (1991) Nature 349, 293-299; Marks J. et al. (1991) Eur J. Immunol. 21_^ 985-991]. The expression vectors may be for cloning in prokaryotic or eukaryotic cells and contain DNA sequences encoding domains capable of a specific interaction with Fee receptors. The sequences capable of a specific interaction with Fee receptors in expression vectors are tailored to the specific application. Such changes can be readily introduced by standard DNA manipulation techniques [Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) "Molecular Cloning - A Laboratory Manual". Second edition, Cold Spring Harbor Laboratory Press] and/or by PCR ligation.

The present invention thus provides binding molecules which mimic IgE in recruiting all, or selected, effector functions in vivo. The key event triggering IgE-mediated effector functions is the aggregation of Fee receptors on the surface of the effector cell. This is normally brought about by binding of an antigen specific, IgE isotype antibody, to localised antigen on the target, in turn localising the Fee receptors on the effector cells through binding of IgE Fee by the Fee receptors. This is an IgE dependent mechanism and in one embodiment of the invention, an antibody specific for a suitable antigen (e.g. a tumour cell surface marker such as ErbB2) can be isolated by standard methods (Harlow, E. and Lane, D. (1988) Antibodies, a laboratory manual. Cold Spring Harbor Laboratory Press) and if it is not isolated as an IgE isotype, converted to an IgE isotype by antibody class switching (Harlow and Lane, supra. ) or more preferably by recombinant methods (Winter and Milstein, supra. ).

There are IgE-independent routes through which IgE- specific effector mechanisms can be activated in an antigen-specific manner. For example, a preparation containing bispecific antibody molecules (Winter and Milstein supra. ) having two or more antigen binding specificities can be used. One arm of the bispecific antibody is specific for the target antigen, and the other has specificity for one or more Fee receptors. In this manner, target antigen and IgE effector mechanisms are brought into proximity, and the Fee receptors aggregated through association with the target antigen, via the antibody. Such a strategy is particularly suited to those cancers with over-expressed marker proteins, for example the cerbB-2 proto oncogene in breast cancer [Styles, J.M. et al., (1990) Int. J. Cancer 45, 320-324]. Another embodiment would be the construction of an antibody fusion molecule with specificity for a suitable antigen, and a peptide sequence that binds to one or more Fee receptors. The peptide sequence may be derived from the Fee receptor binding portion of an IgE antibody. The binding molecules disclosed here may be manufactured in prokaryotic or eukaryotic hosts. Preferably the substances are manufactured in prokaryotic hosts such as E.coli. Alternatively, they may be produced in eukaryotic cells such as CHO cells, although production in readily fermented prokaryotes such as E.coli may be advantageous with regard to costs. The molecules may be produced as single-chain fusion proteins or as hetero-multimeric proteins, targeted to the periplasm, secreted into the growth medium or produced in the cytoplasm as an insoluble protein and refolded into an active conformation in vitro [Huston, J.S. et al. (1988) Proc. Natl. Acad. Sci. USA 8<5_ 5879-5883; Glockshuber, R. et al (1990) Biochemistry 29i, 1362-1367; British Patent Number GB 2137631B, European Patent Application Number EP 324162] .

Thus the present invention provides a molecule which comprises a first binding domain capable of binding to a tumour antigen and a second binding domain capable of binding to a Fee receptor on cells of the immune system, whereby the molecule is able to mediate IgE effector mechanisms against a tumour cell expressing said antigen. The first and/or second binding domains may be synthetic analogues in whole or in part of naturally occurring binding domains. The second binding domain may comprise a polypeptide sequence which is an analogue in whole or in part, of the Ce chain of IgE from a human or animal. The molecules may be able to bind either or both of the Fee RI and Fee RII receptors. The molecule may have a second binding domain in which the polypeptide sequence is similar or identical to the polypeptide sequence of the C chain of the human or animal patient to receive therapy. The first and second binding domains may be derivable from immunoglobulins. The first and second binding domains may be derivable from immunoglobulin variable domains. The first binding domain may be derivable from an immunoglobulin variable domain and the second binding domain may be derivable from an immunoglobulin constant region.

The molecule may be an antibody of the IgE isotype. The molecule may be a chimeric molecule in which the first binding domain is derivable from an immunoglobulin of an isotype other than IgE and said second binding domain is derivable from an immunoglobulin of the isotype

IgE. The invention also provides samples of culture- medium, ascites, sera or pharmaceuticals containing molecules as defined above. The invention also provides a method of using a molecule as described above to prepare a pharmaceutical for the treatment of cancer. Such a method may comprise determining a therapeutically active amount of the molecule for a given cancer and mixing the appropriate amount with one or more pharmacologically acceptable excipients.

The invention also provides methods of treating an animal or human patient which comprise administering a pharmaceutical containing a molecule as described to the patient. Similarly provided are methods of treating ex vivo a tissue explanted from an animal or human.

The present invention also provides recombinant methods for making a molecule as described. The methods may comprise obtaining the nucleotide sequence encoding either or both, of said first and second binding domains, inserting the nucleotide sequence(s) in one or more expression vector also incorporating appropriate regulatory sequences, transforming host cells with the one or more expression vectors, growing the host cells under conditions in which they expresse said nucleotide sequence(s) as said first and/or second binding domains and recovering said first and/or second binding domains. The host cell may be E.coli. The nucleotide sequence(s) may be obtained by use of the polymerase chain reaction. In order that the invention is more fully understood, embodiments will now be described in more detail by way of example only and not by way of limitation. Example

Isolation of an Antibody Specificity Directed Against a Cell Surface Component of a Cancer Cell and Its Conversion into an Immunoglobulin E Class Antibody

In most cases a monoclonal antibody with a desired specificity against a cell surface marker isolated using hybridoma techniques will be of a different subclass. For instance, the CAMPATH series of antibodies (G. Hale et al. (1983) Molec. biol. Med. 1,305-319), which are directed against a cell surface marker on lymphocytes and one of which has been used therapeutically in the treatment of lymphoma, contains monoclonal antibodies of several subclasses but none of the IgE subtype. In order to convert a monoclonal antibody with a desired specificity against a cell surface marker, derived by hybridoma technology, to the IgE subtype, the DNA encoding the variable regions of the heavy and light chains of the antibodies are amplified separately using the polymerase chain reaction using suitable primers such as those described by T. Clackson et al (1991) Nature 352, 624-628. The heavy chain variable region is then subcloned into a vector encoding the constant domains of immunoglobulin E such as pSV-V_NPH_e (M.S.Neuberger et al, Nature 314 268-270, 1985) by a technique such as sticky feet utagenesis (T. Clackson & G. Winter, (1989) Nucleic Acids Res. 17 10163-10170) to replace the variable region contained in the vector with the heavy chain variable region specific for the cancer cell surface marker just generated by amplification. Alternatively, suitable restriction sites for exchanging the variable regions could be introduced by site directed mutagenesis (J.R. Sayers & F. Eckstein (1989) in Protein Function: A Practical Approach pp 279-295, ed. T. Creighton, IRL Press, Oxford) and the new variable region inserted. The light chain variable domain is subcloned into an expression vector such as pSV-hyg-HuCK as described by Orlandi et al (1989) supra. The two vectors containing insert are then linearized with a suitable restriction enzyme and cotransfected into the nonsecreting myeloma line NSO (J.F. Kearney et al, (1979) J. Immunol. 123, 1548-1550) by electroporation (H. Potter et al, (1984) Proc. Natl. Acad. Sci. U.S.A. 81, 7161-7163). Following antibiotic selection, the cells would be screened for the secretion of an antibody with the appropriate IgE isotype and for specific binding to the cell surface marker, for example using solid phase radioimmunoassay on monolayers of fixed tumour cells with -*-²⁵l-anti-IgE antibody. Cells secreting IgE antibodies of the appropriate specificity are then cultured to produce antibody molecules in sufficient amount for therapeutic use. Such antibodies which have been derived from mouse hybridoma technology would need to be 'humanised' (L. Riechmann et al Nature (1988) 332,323-327) before being used for in vivo therapy. It would be preferable to use human antibodies with specificities which are directed against cell surface markers. Human monoclonal antibodies have been difficult to make. Moreover, cells expressing antibodies which are directed against cell surface markers on self cells are clonally deleted in mammals and hence unavailable for fusion. Phage antibody technology (J.D. Marks et al, (1991) J. Mol. Biol. 222,581-597) provides a possible solution to this problem. The use of phage display libraries derived from an unimmunised human has allowed the isolation of antibodies against not only foreign antigens to which the humans have not been exposed but also against a circulating self antigen, tumour necrosis factor (J.D. Marks et al (1991) supra.). This raises the prospect of isolating antibodies directed against tumour cell markers directly from these libraries. A phage display library containing a large number of diverse clones would be made as described by Marks et al (1991) supra. They describe the generation of a library of single chain Fv fragments expressed on phage but libraries of Fab or Fv fragments could be used as alternatives. To select antibodies which are specific for tumour cell markers the antibody library would first be bound to tumour cells, washed to remove non- specifically bound phage, for example with phosphate buffered saline, and specifically bound phage eluted, for example using pH2.8 citrate buffer, for enrichment of antibody directed against the rhesus D antigen of blood cells by panning against blood cells expressing the rhesus D antigen. The population of phage could be depleted for those expressing antibodies binding to normal cells by finding the phage population to normal cells and selecting the phage which did not bind to the cells. This could be performed before or after the panning against the tumour cell. If the tumour is caused by an oncogene, encoding a cell surface marker, which is available in cloned form such as the erb-2 protooncogene implicated in breast cancer, an alternative strategy is available. Cell lines can be transformed with the oncogene and the phage library depleted against the untransformed cell and phage specific for the cell surface marker selected using the transformed cell. If the cloned cell surface marker protein is available in purified form the library can be panned directly against the marker protein, for example by panning on coated tubes. The phage clones thus isolated by any of these methods would then be screened individually for specific binding to a tumour cell by a procedure such as solid phase radioimmunoassay using ^12*^l-antibacteriophage fd using a monolayer of cancer cells. The VH and VL domains would then be separately amplified by the polymerase chain reaction from a selected phage clone which binds specifically to tumour cells with the desired characteristics. The VH domain would be cloned into pSV- V_Np-H_e and the VL domain into pSV-hyg-HuCK, transfected into NSO and cells screened as described above. The antibody fragments for phage display could be derived alternatively from mouse immunoglobulin genes, for instance from mice immunised with human tumour material. Antibodies thus generated would need to be humanised for in vivo therapeutic use.

A development of this system would be to express and secrete an antibody with the IgE effector functions, from

E.coli. To date it has not proved possible to express whole antibodies such as immunoglobulin E from E.coli.

However, Fab fragments have been expressed and secreted

(A. Skerra et al (1990) FEBS Letters 271 203-206). It may be possible to construct comparable fragments retaining IgE effector functions by cloning the VH domain from an antibody obtained as above into a fusion vector.

This vector would encode IgE sequences enabling the generation of a fusion directly with the immunoglobulin E domain(s) providing the effector function, CH3 together with CH2 or CH4 in the case of the low affinity receptor (D. Vereeli et al (1989) supra) and CH2 and 3 in the case of the high affinity receptor (B. Helm et al (1988) supra). These VH-IgE domain fusions would then be expressed in combination with the corresponding VL domain fused either with a CL domain or with the same IgE constant domains which are fused to the VH domain with which it is being coexpressed. Alternatively, the VH-IgE domain fusions and the VL-IgE fusions can be assembled together by recombinant DNA techniques and the assembled genes cloned into an expression vector such as pUC119 or pHENl (with expression in HB2151 (eg. as described for Fab fragments in H.R. Hoogenboom et al (1991) Nucleic Acids Res. 19, 4133-4151)). The chances of success of the strategy of expressing IgE fusions in E.coli will be enhanced by the fact that glycosylation is not required for binding of immunoglobulin E to its receptors. Alternatively, it is possible to retain effector functions with smaller fragments of the IgE constant domains. In particular, a peptide spanning the junction of the human IgE CH2 and CH3 domains, Gln301-Arg376 retains approximately 30% of IgE effector function activity for binding to the Fc6 RI (high affinity) receptor (Helm et al, (1988) supra). It would be desirable to use a smaller fragment such as Gln301-Arg375 to fuse to the antibody variable domains. The use of this fragment would be particularly advantageous since it is active in a monomeric form and it has been successfully expressed while retaining effector function from E.coli. It has also been expressed in E.coli as part of a larger fragment containing CH domains 2 to 4. Thus the Gln301-Arg375 fragment should provide an effector function whether fused just to the VH domain or to both the VH and VL domains of an antibody fragment. Moreover, alternatively, it would be possible to express the Gln301-Arg375 peptide as a fusion as an extension from the CHI domain of a Fab fragment.

To construct fusions with a small fragment such as Gln310-Arg375 the same strategies as mentioned above could be used, that is the fragment could be encoded in the vector to allow the generation of fusion proteins or the VH-IgE Gln310-Arg375 and VL genes could be assembled together prior to cloning into the vector. The situation for fragments binding to the low affinity Fee RII receptor is more complex since, although the binding site for the receptor is within the CH3 domain, either the CH2 or the CH4 domain is required to dimerise the region and obtain receptor function. Hence, fusions would need to be made with both the VH and VL domains and the fusions would need to include both the CH3 and the CH2 or CH4 domains.

Claims

1. A molecule which comprises a first binding domain capable of binding to a tumour antigen, and a second binding domain capable of binding to a Fee receptor on cells of the immune system, whereby the molecule is able to mediate IgE effector mechanisms against a tumour cell expressing said antigen.

2. A molecule according to claim 1 wherein the first and second binding domains are derivable from immunoglobulins or are synthetic analogues or part of naturally occuring immunoglobulin domains.

3. A molecule according to claim 1 or claim 2 wherein both the first and second binding domains are derivable from immunoglobulins variable domains.

4. A molecule according to claim 1 or claim 2 wherein the first binding domain is derivable from an immunoglobulin variable domain and the second binding domain is derivable from an immunoglobulin constant region.

5. A molecule according to any one of claims 1, 2 or 4 which is an antibody of the IgE isotype.

6. A molecule according to any one of claims 1 to 4 which is a chimeric molecule, wherein said first binding domain is derivable from an immunoglobulin of an isotype other than IgE and said second binding domain is derivable from an immunoglobulin of the isotype IgE.

7. A sample of culture medium, ascites or sera containing a molecule according to any one of claims 1 to 6.

8. A pharmaceutical which comprises a molecule according to any one of claims 1 to 6.

9. A method which comprises using a molecule according to any one of claims 1 to 6 to prepare a pharmaceutical for the treatment of cancer.

10. A method of treating an animal or human patient which comprises administering a pharmaceuticl according to claim 8 to said patient.

11. A method of treating ex vivo a tissue explanted from an animal or human patient, which comprises applying a molecule according to any one of claims 1 to 6, a sample according to claim 7, or a pharmaceutical according to claim 8, to said tissue.

12. A method of making a molecule according to any one of claims 1 to 6 which comprises obtaining the nucleotide sequence encoding either or both of part or all of said first and second binding domains; inserting said nucleotide sequence(s) into the same or different expression vectors also incorporating appropriate regulatory sequences; transforming one or more host cells with said same or different expression vectors; growing said transformed host cell under conditions in which it expresses said nucleotide sequence(s) as said first and/or second binding domains; and recovering said first and/or second binding domains.

13. A method according to claim 12 wherein the host cell is a prokaryotic cell.

14. A method according to claim 13 wherein the prokaryotic cell is E.coli.

15. A method according to claim 12 wherein the host cell is a eukaryotic cell.

16. A method according to claim 15 wherein the eukaryotic cell is the murine myeloma cell NSO or Chinese hamster ovary cell.

17. A method according to claim 12 or claim 13 wherein said nucleotide sequence(s) are obtained by use of the polymerase chain reactions.