The invention provides antibodies to a mutant form of the epidermal growth factor receptor known as EGFRvIII found only or primarily on the surface of glioblastoma cells, and on cells of breast, ovarian and non-small cell lung carcinomas. The antibodies provided by the invention have the complementarity determining regions ("CDRs") of the scFv designated MR1, but with mutations at positions 98 and 99 in the CDR3 of the heavy chain variable region and, optionally, in other CDRs. In particular, the invention provides an antibody, designated MR1-1, which mutates MR1 in the CDR3 of the VH and VL chains. The invention provides additional antibodies in which MR1 is mutated in the CDR1 and 2 of VH or VL, or both.

DETAILED DESCRIPTION OF THE INVENTION

I. Introduction

The present invention provides scFv antibodies and other antibodies with higher affinity for EGFRvIII than that of MR1. It further provides immunotoxins with higher cytotoxicity for EGFRvIII-expressing cells than that of the same immunotoxins using MR1 as the targeting portion of the molecule. The antibodies are created from MR1, but are mutated in hot spot regions of their complementarity determining regions (CDRs). Surprisingly, immunotoxins incorporating these mutated forms of MR1 as the targeting moiety not only have high affinity for EGFRvIII, but also can be produced at significantly higher yields than similar immunotoxins employing MR1 as the targeting portion of the molecule.

The limiting factor in the construction of antibody libraries with randomizations in the CDRs is the large number of residues that constitute the CDRs. Because x-ray structures are not available for most antibodies, usually there is no attempt to identify the few CDR residues where mutations are likely to yield higher affinity variants. Consequently, it is necessary to construct extremely large randomized libraries to ensure the isolation of higher affinity variants.

In exemplary studies, mutations were made in CDR3 of the heavy chain variable region (V.sub.H) of MR1. The V.sub.H CDR3 of MR1 has eleven amino acid residues. The final two residues of the V.sub.H region, D101 and Y102, were excluded from the mutagenesis studies because they were considered unlikely to participate in antigen binding, for two reasons. First, these two residues usually lie at the interface with the V.sub.L chain. As a result, they are not exposed. Second, these residues are contributed to by the J segment and follow the 3'-junctional region of the V.sub.HCDR2. As a result, they are relatively more conserved than the rest of the V.sub.HCDR3.

The other nine amino acids at each position of the CDR3 V.sub.H were substituted by each of the other natural amino acids. The studies showed that only mutations in regions known to be hot spots (regions known to undergo hypermutation during antibody affinity maturation, best characterized by the motifs RGYW and AGY, where R is A or G, Y is C or T, and W is A or T) resulted in antibodies with higher cytotoxicity than the parental antibody MR1. The mutations were also found to enhance the yield of immunotoxin when these mutated scFv were incorporated into immunotoxins, compared to a like immunotoxin made with the parental MR1 scFv. In further exemplary studies, mutations were made in the MR1 CDR3 of the light chain variable region. In these studies, mutations were made only to amino acids within a hot spot. Once again, the mutations resulted in antibodies with higher cytotoxicity and yield than the parental antibody MR1.

Based on these results, it is not necessary to randomize every residue in a CDR to identify those where mutations will yield higher affinity variants. It is expected that mutations in the hot spots of V.sub.H and V.sub.L CDRs 1 and 2 will likewise result in antibodies with improved affinity for EGFRvIII, improved cytotoxicity for EGFRvIII-expressing cells when incorporated into immunotoxins, and improved yield for immunotoxins incorporating such scFvs compared to MR1-based immunotoxins. Since the improved properties are due to the substitutions of the amino acids, the same positive effects will also be found when these mutated forms of MR1 (including those mutated in the V.sub.H and V.sub.L CDR3) are incorporated into disulfide stabilized Fvs (dsFvs), or into Fab', F(ab').sub.2, or Fab.

Increased cytotoxic activity does not necessarily correlate with increased affinity (Tables 4 and 5, infra). There is no obvious explanation for this lack of correlation. Besides binding affinity, which is typically measured at 22.degree. C., there are many aspects in the toxicity process which could be affected by the "hot spot" mutations. Such aspects include stability at 37.degree. C., rate of internalization, proteolytic processing and transfer to the compartment required for translocation. It is possible that one or more of these aspects is affected. Antibodies with higher affinity for EGFRvIII are, however, useful for a variety of purposes, and especially for diagnostic uses and in vitro assays to determine the presence or absence of EGFRvIII-expressing cells in a sample. For example, for in vitro uses, antibodies such as scFv with higher affinity for EGFRvIII can be conjugated to radionuclides or to any of a number of other detectable labels and used to detect the presence of cells expressing EGFRvIII in a biopsy sample from a patient to determine whether the patient has a cancer characterized by the presence of such cells, or to determine that the cancer has not yet been eradicated from a patient known to have such a cancer. Similarly, in in vivo uses, scFv, dsFv, or other antibodies of the invention can be conjugated to radionuclides or other detectable labels and used to detect the presence of cells expressing EGFRvIII in the patient, thereby again diagnosing whether the patient has a cancer characterized by the presence of such cells, or that the cancer has not yet been eradicated from a patient known to have such a cancer.

Finally, one striking difference observed among the CDR mutants was the final yield of active monomeric protein. Recombinant toxins accumulate in inclusion bodies as insoluble aggregated protein (immunotoxin). Active monomers are produced by dissolving the inclusion bodies in 6M Guanidine HCl, followed by controlled renaturation in a redox system and separation of monomers from multimers and aggregates. The studies showed that mutations in 1 or 2 amino acids in the CDRs can greatly increase yields (Table 6) of immunotoxin, as defined and explained below. The yield of MR1(Fv)-PE38 is only 2%, but it was dramatically increased to 17% with heavy chain CDR3 S98P-T99S mutations. Presumably, these mutations have a profound effect on the folding pathway. In general, all the heavy chain mutants isolated in the initial mutagenesis of the heavy chain had a better yield than the parental MR1. Thus, the heavy chain of CDR3 is very important for proper folding and the phage expression system may select in some way for proteins that fold more efficiently. Consequently, phage containing better folding Fvs appear to be present in larger numbers and are preferentially enriched during panning on antigen.

Based on these results, it can be expected that phage display can be used to select Fvs mutated in the CDR1 or 2 of the V.sub.H and V.sub.L chains which likewise exhibit an increased production yield compared to the parental MR1 scFv.

In vivo tests were conducted to determine the effect of exemplary immunotoxins targeted by MR1-1 on animal models of human tumors. To establish the tumors, athymic rats and mice were injected intracranially, intrathecally, or subcutaneously with cells of a human gliablastoma cell line (U87MG) which was transfected to express EGFRvIII (the transfected EGFRvIII-expressing cell line is designated U87MG..DELTA.EGFRvIII and is described by Nishikawa et al., Proc Natl Acad Sci (USA) 91(16):7727 7731 (1994)). Once the tumors were established, the rats or mice were treated with bolus injections or continuous infusions of various doses of immunotoxin. As described in the Examples, below, and FIGS. 3 6, regardless of the tumor location and method of administering the immunotoxin, animals given the immunotoxin at doses of 1 or 2 .mu.g showed markedly less tumor growth, and longer survival times, than that of animals given a saline control.

III. Creation of Antibodies with Higher Affinity for EGFRvIII than that of MR1

Antibodies bind to antigens via residues in their CDRs. Consequently, mutagenesis of CDRs is widely used to improve the affinity of Fab and Fv fragments of antibodies. There are a number of different approaches to CDR mutagenesis. Most of these, such as codon-based mutagenesis (Yelton et al., J. Immunol. 155:1994 2004 (1995)), CDR walking (Barbas et al., Trends Biotech. 14:230 234 (1996); Yang et al., J. Mol. Biol. 254:392 403 (1995)), error prone replication (Low et al., J. Mol. Biol. 260:359 368 (1996)) and synthetic CDR construction (de Kruif et al., J. Mol. Biol. 248:97 105 (1995)), require the construction of large libraries that are technically difficult to make and are hard to handle. The trend in antibody affinity maturation has been towards the isolation of high affinity binders from relatively smaller sized libraries (Pini et al., J. Biol. Chem. 273:21769 21776 (1998); Wu et al., Proc. Natl. Acad. Sci. USA 95:6037 6042 (1998); Chowdhury et al., Nature Biotechnol. 17:568 572 (1999)). All of these approaches involve the construction of expression libraries of antibodies with mutations in the CDRs and selection for better binders.

Phage display technology has become a useful tool for screening large peptide or protein libraries (Winter et al., Annu. Rev. Immunol. 12:433 455 (1994); McCafferty, J., Nature 348:552 554 (1990); Barbas et al., Proc. Natl. Acad. Sci. USA 88:7978 7982 (1991)). Single chain Fvs can be expressed on phage particles as fusions with M13 gene 3 protein in a phagemid vector. The fusion proteins are expressed in E. coli and, in the presence of helper phage, are displayed on the tips of M13 phage which can be collected from culture media. Phage that display scFv fusion proteins, which bind to specific antigen, are selected by panning the phage libraries on cells expressing the antigen or on a surface to which the antigen is coupled, such as magnetic beads. Phage which do not bind are washed away, bound phage are eluted and amplified by re-infecting E. coli. Several rounds of panning result in an enrichment of specific binders. By making the panning conditions more stringent, better binders can be more effectively separated from poor binders.

Phage display technology can be exploited to develop antibodies which bind to EGFRvIII with higher affinity than MR1. As is well known in the art, an intact antibody comprises two heavy chains and two light chains; each chain has three CDRs, designated 1, 2, and 3, respectively. Each CDR is known to contribute to antigen binding, but they do so unequally. See generally, Kuby, J., Immunology, W.H. Freeman & Co., New York (3.sup.rd Ed. 1998). While the amino acids of any of the CDRs can be mutated to find mutations which increase affinity, on average, the CDR3 of each chain makes a greater contribution to antigen binding than does the CDR 1 or CDR2 of that chain. See generally, Kuby, supra, at page 117. Thus, mutations of the CDR3 V.sub.H and V.sub.L chains can be particularly advantageous. Moreover, the CDR3 V.sub.H of MR1 is relatively long for a mouse CDR3, which may contribute in part to the relative stability of MR1 scFv relative to other scFvs.

After excluding the last two amino acids of the V.sub.H CDR3 as unlikely to contribute to antigen binding for the reasons discussed in the Introduction, the remaining nine residues in V.sub.H CDR3 was substituted with each of the other 19 natural amino acids. Only mutations of the serine and threonine at amino acid positions 98 and 99, respectively, resulted in improved cytotoxicity of the resulting immunotoxin (the amino acids sequence of the V.sub.H and the V.sub.L CDR3s are set forth in single letter code in Table 5, 6, and 7, infra. The sequence setting forth the V.sub.H does not show the two residues, D101 and Y102, which were considered not likely to contribute to antigen binding.) The preferred substitutions in V.sub.H CDR3 were P98-Y99, which gave an IC.sub.50 in a PE38 immunotoxin of 3.5 ng/ml, P98-N99, P98-W99, and P98-I99, all of which had IC.sub.50s of 4.5 ng, P98-F99, with an IC.sub.50 of 6 ng, and P98-V99, which had an IC.sub.50 of 6.5 ng. Four clones, P98-S99, W98-V99, S98-W99, and P98-T99, had IC.sub.50s equal to the parental clone. (By convention, a term such as "P98-Y99" denotes that the amino acid proline appears at position 98 of the designated polypeptide, in this case MR1 V.sub.H CDR3, and that tyrosine appears at position 99, but that the rest of the molecule is that of the normal polypeptide, in this case, the parental antibody MR1.)

With respect to the mutation of the V.sub.L CDR3, only one mutation, F92W, gave an immunotoxin more active than the parent. Its IC.sub.50 was 1.3 ng/ml. In this case, the "parent" molecule was MR1 with the substitutions P98-Y99 in the VH CDR3, which without the added substitution in the V.sub.L CDR3 had an IC.sub.50 of 3.5 ng/ml. This mutated MR1, which combined mutations in the CDR3 of both the V.sub.H and the V.sub.L chains (V.sub.H S98P-T99Y-V.sub.LF92W), was the most cytotoxic form tested when employed as the targeting portion of an immunotoxin, and is now termed "MR1-1."

These results show that mutations in various CDRs can have an additive effect and can increase the cytotoxicity of the resulting immunotoxin. In view of these results, mutations of the amino acids in the hot spots of CDRs 1 and 2 of the V.sub.H and V.sub.L chains of MR1 will likewise result in antibodies with further improved affinity for EGFRvIII and in immunotoxins with further improved cytotoxicity when compared to MR1.

The best time to analyze clones is early in the process. Panning after the enrichment peaks can be deleterious because of the risk of losing clones. It is possible that Fvs with low affinities but high expression may be preferentially enriched while good binders may be lost. Evidence supporting this was observed while panning the light chain CDR3 libraries: mutant F92S with a low affinity (K.sub.d 22 nM) was found in 7 of 10 clones examined after the third round, whereas, the best binder, F92W, was present only once. In contrast, in the second round F92S was only found in 2 of 17 clones, whereas F92W was present in 6 of 17 clones.

IV. Anti-EGFRvIII Antibodies

The present invention provides antibodies which bind to EGFRvIII with higher affinity than prior art antibodies and which selectively react with EGFRvIII. In particular, the invention provides antibodies which have a lower Kd with regard to EGFRvIII than does MR1, the best previously known scFv targeted to this antigen. Moreover, these antibodies form immunotoxins which have higher cytotoxicity for EGFRvIII-expressing cells than does the same immunotoxin made with MR1. The invention further provides a method for generating antibodies with higher affinity and greater cytotoxicity against EGFRvIII than MR1 has. The immunoconjugates disclosed below target EGFRvIII using antibodies of the present invention. These antibodies are selectively reactive under immunological conditions to those determinants of EGFRvIII displayed on the surface of mammalian cells and accessible to the antibody from the extracellular milieu.

In preferred embodiments of the present invention, the anti-EGFRvIII antibody is a recombinant antibody such as a scFv or a disulfide stabilized Fv antibody. Fv antibodies are typically about 25 kDa and contain a complete antigen-binding site with 3 CDRs on both the heavy and light chains. If the V.sub.H and the V.sub.L chain are expressed non-contiguously, the chains of the Fv antibody are typically held together by noncovalent interactions. However, these chains tend to dissociate upon dilution, so methods have been developed to crosslink the chains through glutaraldehyde, intermolecular disulfides, or a peptide linker. Disulfide stabilized Fvs are taught, for example, in U.S. Pat. No. 5,747,654.

In a particularly preferred embodiment, the antibody is a single chain Fv (scFv). The V.sub.H and the V.sub.L regions of a scFv antibody comprise a single chain which is folded to create an antigen binding site similar to that found in two chain antibodies. Once folded, noncovalent interactions stabilize the single chain antibody. In a more preferred embodiment, the scFv is recombinantly produced. One of skill will realize that conservative variants of the antibodies of the instant invention can be made. Such conservative variants employed in scFv fragments will retain critical amino acid residues necessary for correct folding and stabilizing between the V.sub.H and the V.sub.L regions.

In some embodiments of the present invention, the scFv antibody is directly linked to the effector molecule (EM) through the light chain. However, scFv antibodies can be linked to the EM via its amino or carboxyl terminus.

While the V.sub.H and V.sub.L regions of some antibody embodiments can be directly joined together, one of skill will appreciate that the regions may be separated by a peptide linker consisting of one or more amino acids. Peptide linkers and their use are well-known in the art. See, e.g., Huston et al., Proc. Nat'l. Acad. Sci. USA 8:5879 (1988); Bird et al., Science 242:4236 (1988); Glockshuber et al., Biochemistry 29:1362 (1990); U.S. Pat. No. 4,946,778, U.S. Pat. No. 5,132,405 and Stemmer et al., Biotechniques 14:256 265 (1993), all incorporated herein by reference. Generally the peptide linker will have no specific biological activity other than to join the regions or to preserve some minimum distance or other spatial relationship between the V.sub.H and V.sub.L. However, the constituent amino acids of the peptide linker may be selected to influence some property of the molecule such as the folding, net charge, or hydrophobicity. Single chain Fv (scFv) antibodies optionally include a peptide linker of no more than 50 amino acids, generally no more than 40 amino acids, preferably no more than 30 amino acids, and more preferably no more than 20 amino acids in length. In some embodiments, the peptide linker is a concatamer of the sequence Gly-Gly-Gly-Gly-Ser (SEQ ID NO:12), preferably 2, 3, 4, 5, or 6 such sequences. However, it is to be appreciated that some amino acid substitutions within the linker can be made. For example, a valine can be substituted for a glycine.

A. Production of scFvs

As described above, in preferred embodiments, the antibody (for example, the targeting moiety of an immunotoxin) is a scFv. Methods of making scFv antibodies have been described. See, Huse et al., supra; Ward et al. Nature 341:544 546 (1989); and Vaughan et al., supra. In brief, mRNA from B-cells is isolated and cDNA is prepared. The cDNA is amplified by well known techniques, such as PCR, with primers specific for the variable regions of heavy and light chains of immunoglobulins. The PCR products are purified by, for example, agarose gel electrophoresis, and the nucleic acid sequences are joined. If a linker peptide is desired, nucleic acid sequences that encode the peptide are inserted between the heavy and light chain nucleic acid sequences. The sequences can be joined by techniques known in the art, such as blunt end ligation, insertion of restriction sites at the ends of the PCR products or by splicing by overlap extension (Chowdhury et al., Mol. Immunol 34:9 (1997)). After amplification, the nucleic acid which encodes the scFv is inserted into a vector, again by techniques well known in the art. Preferably, the vector is capable of replicating in prokaryotes and of being expressed in both eukaryotes and prokaryotes. In a preferred embodiment, the scFv genes are joined with the PE38 gene by a short linker and cloned into a T7-based expression vector. In particularly preferred embodiments, the scFv is expressed under control of the T7 promoter in E. coli BL21 (.lamda. DE3).

As noted in preceding sections, scFv that specifically bind to EGFRvIII are found by panning. Panning can be performed by any of several methods. In a preferred method with respect to the present invention, panning can conveniently be performed using cells expressing EGFRvIII on their surface. A protocol for performing panning using cells is set forth in the Examples, below. Panning can also be performed on a solid surface by coating the solid surface with EGFRvIII and incubating the phage on the surface for a suitable time under suitable conditions. Conveniently, the surface can be a magnetic bead. The unbound phage are washed off the solid surface and the bound phage are eluted.

Finding the antibody with the highest affinity is dictated by the efficiency of the selection process and depends on the number of clones that can be screened and the stringency with which it is done. Typically, higher stringency corresponds to more selective panning. If the conditions are too stringent, however, the phage will not bind. After one round of panning, the phage that bind to EGFRvIII coated plates or to cells expressing EGFRvIII on their surface are expanded in E. coli and subjected to another round of panning. In this way, an enrichment of many fold occurs in 3 rounds of panning. Thus, even when enrichment in each round is low, multiple rounds of panning will lead to the isolation of rare phage and the genetic material contained within which encodes the scFv with the highest affinity or one which is better expressed on phage.

Regardless of the method of panning chosen, the physical link between genotype and phenotype provided by phage display makes it possible to test every member of a cDNA library for binding to antigen, even with large libraries of clones.

B. Binding Affinity of Antibodies

The antibodies of this invention bind to an epitope of EGFRvIII with a Kd at least 1 nM lower than that of parental antibody MR1. Binding affinity for a target antigen is typically measured or determined by standard antibody-antigen assays, such as competitive assays, saturation assays, or immunoassays such as ELISA or RIA.

Such assays can be used to determine the dissociation constant of the antibody. The phrase "dissociation constant" refers to a measure of the affinity of an antibody for an antigen. Specificity of binding between an antibody and an antigen exists if the dissociation constant (Kd=1/K, where K is the affinity constant) of the antibody is in the micromolar range, preferably <100 nM, and most preferably <0.1 nM. Antibody molecules will typically have a Kd in the lower ranges. Kd=[Ab-Ag]/[Ab][Ag] where [Ab] is the concentration at equilibrium of the antibody, [Ag] is the concentration at equilibrium of the antigen and [Ab-Ag] is the concentration at equilibrium of the antibody-antigen complex. Typically, the binding interactions between antigen and antibody include reversible noncovalent associations such as electrostatic attraction, Van der Waals forces and hydrogen bonds. This method of defining binding specificity applies to single heavy and/or light chains, CDRs, fusion proteins or fragments of heavy and/or light chains, that are specific for EGFRvIII.

C. Immunoassays

The antibodies can be detected and/or quantified using any of a number of well recognized immunological binding assays (see, e.g., U.S. Pat. Nos. 4,366,241; 4,376,110; 4,517,288; and 4,837,168). For a review of the general immunoassays, see also METHODS IN CELL BIOLOGY, VOL. 37, Asai, ed. Academic Press, Inc. New York (1993); BASIC AND CLINICAL IMMUNOLOGY 7TH EDITION, Stites & Terr, eds. (1991). Immunological binding assays (or immunoassays) typically utilize a ligand (e.g., EGFRvIII) to specifically bind to and often immobilize an antibody. The antibodies employed in immunoassays of the present invention are discussed in greater detail supra.

Immunoassays also often utilize a labeling agent to specifically bind to and label the binding complex formed by the ligand and the antibody. The labeling agent may itself be one of the moieties comprising the antibody/analyte complex, i.e., the anti-EGFRvIII antibody. Alternatively, the labeling agent may be a third moiety, such as another antibody, that specifically binds to the antibody/EGFRvIII protein complex.

In one aspect, a competitive assay is contemplated wherein the labeling agent is a second anti-EGFRvIII antibody bearing a label. The two antibodies then compete for binding to the immobilized EGFRvIII. In embodiments where the question to be answered is to compare the affinity of the first antibody to that of MR1, the second antibody can be MR1. Alternatively, in a non-competitive format, the EGFRvIII antibody lacks a label, but a second antibody specific to antibodies of the species from which the anti-EGFRvIII antibody is derived, e.g., murine, and which binds the anti-EGFRvIII antibody, is labeled.

Other proteins capable of specifically binding immunoglobulin constant regions, such as Protein A or Protein G may also be used as the label agent. These proteins are normal constituents of the cell walls of staphylococcal bacteria. They exhibit a strong non-immunogenic reactivity with immunoglobulin constant regions from a variety of species (see generally, Kronval et al., J. Immunol. 111:1401 1406 (1973); and Akerstrom et al., J. Immunol. 135:2589 2542 (1985)).

Throughout the assays, incubation and/or washing steps may be required after each combination of reagents. Incubation steps can vary from about 5 seconds to several hours, preferably from about 5 minutes to about 24 hours. However, the incubation time will depend upon the assay format, antibody, volume of solution, concentrations, and the like. Usually, the assays will be carried out at ambient temperature, although they can be conducted over a range of temperatures, such as 4.degree. C. to 40.degree. C.

While the details of the immunoassays of the present invention may vary with the particular format employed, the method of detecting anti-EGFRvIII antibodies in a sample containing the antibodies generally comprises the steps of contacting the sample with an antibody which specifically reacts, under immunologically reactive conditions, to the EGFRvIII/antibody complex.

V. Production of Immunoconjugates

The anti-EGFRvIII antibodies generated in the present invention can be linked to effector molecules (EM) through the EM carboxyl terminus, the EM amino terminus, through an interior amino acid residue of the EM such as cysteine, or any combination thereof. Similarly, the EM can be linked directly to the heavy, light, Fc (constant region) or framework regions of the antibody. Linkage can occur through the antibody's amino or carboxyl termini, or through an interior amino acid residue. If PE or a cytotoxic fragment thereof is used as the EM, linkages at or near the carboxyl terminus should be made in a manner to maintain, or to add (if the linkage is at the C-terminus) a sequence which functions as a signal sequence to direct the molecule into the cytosol. Appropriate signal amino acid sequences such as REDLK (SEQ ID NO:13) (the sequence of native PE, in single letter code), KDEL (SEQ ID NO:14), RDEL (SEQ ID NO:15), and repeats of KDEL (SEQ ID NO:16), are known in the art. See, e.g., WO 91/18099. Further, multiple EM molecules (e.g., any one of from 2 10) can be linked to the anti-EGFRvIII antibody and/or multiple antibodies (e.g., any one of from 2 5) can be linked to an EM. In a particularly preferred embodiment, KDEL (SEQ ID NO:16) is the C-terminal sequence. The molecule formed by the linking of an effector molecule to an antibody is known as an immunoconjugate.

A therapeutic agent is an agent with a particular biological activity directed against a particular target molecule or a cell bearing a target molecule. One of skill in the art will appreciate that therapeutic agents may include various drugs such as vinblastine, daunomycin and the like, cytotoxins such as native or modified Pseudomonas exotoxin or Diphtheria toxin, encapsulating agents, (e.g., liposomes) which themselves contain pharmacological compositions, radioactive agents such as .sup.125I, .sup.32P, .sup.14C, .sup.3H and .sup.35S and other labels, target moieties and ligands.

The choice of a particular therapeutic agent depends on the particular target molecule or cell and the biological effect is desired to evoke. Thus, for example, the therapeutic agent may be a cytotoxin which is used to bring about the death of a particular target cell. Conversely, where it is merely desired to invoke a non-lethal biological response, the therapeutic agent may be conjugated to a non-lethal pharmacological agent or a liposome containing a non-lethal pharmacological agent.

With the therapeutic agents and antibodies herein provided, one of skill can readily construct a variety of clones containing functionally equivalent nucleic acids, such as nucleic acids which differ in sequence but which encode the same EM or antibody sequence. Thus, the present invention provides nucleic acids encoding antibodies and conjugates and fusion proteins thereof.

A. Recombinant Methods

The nucleic acid sequences of the present invention can be prepared by any suitable method including, for example, cloning of appropriate sequences or by direct chemical synthesis by methods such as the phosphotriester method of Narang et al., Meth. Enzymol. 68:90 99 (1979); the phosphodiester method of Brown et al., Meth. Enzymol. 68:109 151 (1979); the diethylphosphoramidite method of Beaucage et al., Tetra. Lett. 22:1859 1862 (1981); the solid phase phosphoramidite triester method described by Beaucage & Caruthers, Tetra. Letts. 22(20):1859 1862 (1981), e.g., using an automated synthesizer as described in, for example, Needham-VanDevanter et al. Nucl. Acids Res. 12:6159 6168 (1984); and, the solid support method of U.S. Pat. No. 4,458,066. Chemical synthesis produces a single stranded oligonucleotide. This may be converted into double stranded DNA by hybridization with a complementary sequence, or by polymerization with a DNA polymerase using the single strand as a template. One of skill would recognize that while chemical synthesis of DNA is limited to sequences of about 100 bases, longer sequences may be obtained by the ligation of shorter sequences.

In a preferred embodiment, the nucleic acid sequences of this invention are prepared by cloning techniques. Examples of appropriate cloning and sequencing techniques, and instructions sufficient to direct persons of skill through many cloning exercises are found in Sambrook et al., MOLECULAR CLONING: A LABORATORY MANUAL (2ND ED.), Vols. 1 3, Cold Spring Harbor Laboratory (1989)), Berger and Kimmel (eds.), GUIDE TO MOLECULAR CLONING TECHNIQUES, Academic Press, Inc., San Diego Calif. (1987)), or Ausubel et al. (eds.), CURRENT PROTOCOLS IN MOLECULAR BIOLOGY, Greene Publishing and Wiley-Interscience, NY (1987). Product information from manufacturers of biological reagents and experimental equipment also provide useful information. Such manufacturers include the SIGMA chemical company (Saint Louis, Mo.), R&D systems (Minneapolis, Minn.), Pharmacia LKB Biotechnology (Piscataway, N.J.), CLONTECH Laboratories, Inc. (Palo Alto, Calif.), Chem Genes Corp., Aldrich Chemical Company (Milwaukee, Wis.), Glen Research, Inc., GIBCO BRL Life Technologies, Inc. (Gaithersberg, Md.), Fluka Chemica-Biochemika Analytika (Fluka Chemie AG, Buchs, Switzerland), Invitrogen (San Diego, Calif.), and PE Applied Biosystems (Foster City, Calif.), as well as many other commercial sources known to one of skill.

Nucleic acids encoding native effector molecules (EM) or anti-EGFRvIII antibodies can be modified to form the EM, antibodies, or immunoconjugates of the present invention. Modification by site-directed mutagenesis is well known in the art. Nucleic acids encoding EM or anti-EGFRvIII antibodies can be amplified by in vitro methods. Amplification methods include the polymerase chain reaction (PCR), the ligase chain reaction (LCR), the transcription-based amplification system (TAS), and the self-sustained sequence replication system (3SR). A wide variety of cloning methods, host cells, and in vitro amplification methodologies are well known to persons of skill.

In a preferred embodiment, immunoconjugates are prepared by inserting the cDNA which encodes an anti-EGFRvIII scFv antibody into a vector which comprises the cDNA encoding the EM. The insertion is made so that the scFv and the EM are read in frame, that is in one continuous polypeptide which contains a functional Fv region and a functional EM region. In a particularly preferred embodiment, cDNA encoding a Diphtheria toxin fragment is ligated to a scFv so that the toxin is located at the amino terminus of the scFv. In a most preferred embodiment, cDNA encoding PE or a cytotoxic fragment thereof is ligated to a scFv so that the toxin is located at the carboxyl terminus of the scFv. In the most preferred embodiment, the cytotoxic fragment is PE38.

Once the nucleic acids encoding an EM, an anti-EGFRvIII antibody, or an immunoconjugate of the present invention are isolated and cloned, one may express the desired protein in a recombinantly engineered cell such as bacteria, plant, yeast, insect and mammalian cells. It is expected that those of skill in the art are knowledgeable in the numerous expression systems available for expression of proteins including E. coli, other bacterial hosts, yeast, and various higher eucaryotic cells such as the COS, CHO, HeLa and myeloma cell lines. No attempt to describe in detail the various methods known for the expression of proteins in prokaryotes or eukaryotes will be made. In brief, the expression of natural or synthetic nucleic acids encoding the isolated proteins of the invention will typically be achieved by operably linking the DNA or cDNA to a promoter (which is either constitutive or inducible), followed by incorporation into an expression cassette. The cassettes can be suitable for replication and integration in either prokaryotes or eukaryotes. Typical expression cassettes contain transcription and translation terminators, initiation sequences, and promoters useful for regulation of the expression of the DNA encoding the protein. Conveniently, the cassettes can be placed in plasmids which also contain one or more antibiotic resistance genes. To obtain high level expression of a cloned gene, it is desirable to construct expression cassettes which contain, at the minimum, a strong promoter to direct transcription, a ribosome binding site for translational initiation, and a transcription/translation terminator. For E. coli this includes a promoter such as the T7, trp, lac, or lambda promoters, a ribosome binding site and preferably a transcription termination signal. For eukaryotic cells, the control sequences can include a promoter and preferably an enhancer derived from immunoglobulin genes, SV40, cytomegalovirus, and a polyadenylation sequence, and may include splice donor and acceptor sequences. The cassettes of the invention can be transferred into the chosen host cell by well-known methods such as calcium chloride transformation or electroporation for E. coli and calcium phosphate treatment, electroporation or lipofection for mammalian cells. Cells transformed by the cassettes can be selected by resistance to antibiotics conferred by resistance genes contained in the cassettes, such as the amp, kan, gpt, neo and hyg genes. Kanomycin resistance and ampicillin resistance are preferred embodiments for working with phage.

One of skill would recognize that modifications can be made to a nucleic acid encoding a polypeptide of the present invention (i.e., anti-EGFRvIII antibody, PE, or an immunoconjugate formed from their combination) without diminishing its biological activity. Some modifications may be made to facilitate the cloning, expression, or incorporation of the targeting molecule into a fusion protein. Such modifications are well known to those of skill in the art and include, for example, a methionine added at the amino terminus to provide an initiation site, or additional amino acids (e.g., poly His) placed on either terminus to create conveniently located restriction sites or termination codons or purification sequences.

In addition to recombinant methods, the immunoconjugates, EM, and antibodies of the present invention can also be constructed in whole or in part using standard peptide synthesis. Solid phase synthesis of the polypeptides of the present invention of less than about 50 amino acids in length may be accomplished by attaching the C-terminal amino acid of the sequence to an insoluble support followed by sequential addition of the remaining amino acids in the sequence. Techniques for solid phase synthesis are described by Barany & Merrifield, THE PEPTIDES: ANALYSIS, SYNTHESIS, BIOLOGY, VOL. 2: SPECIAL METHODS IN PEPTIDE SYNTHESIS, PART A. pp. 3 284; Merrifield et al. J. Am. Chem. Soc. 85:2149 2156 (1963), and Stewart et al., SOLID PHASE PEPTIDE SYNTHESIS, 2ND ED., Pierce Chem. Co., Rockford, Ill. (1984). Proteins of greater length may be synthesized by condensation of the amino and carboxyl termini of shorter fragments. Methods of forming peptide bonds by activation of a carboxyl terminal end (e.g., by the use of the coupling reagent N,N'-dicyclohexylcarbodiimide) are known to those of skill.

B. Purification

Once expressed, the recombinant immunoconjugates, antibodies, and/or effector molecules of the present invention can be purified according to standard procedures of the art, including ammonium sulfate precipitation, affinity columns, ion exchange and gel filtration columns and batch chromatography, and the like (see, generally, R. Scopes, PROTEIN PURIFICATION, Springer-Verlag, N.Y. (1982)). Substantially pure compositions of at least about 90 to 95% homogeneity are preferred, and 98 to 99% or more homogeneity are most preferred for pharmaceutical uses. Once purified, partially or to homogeneity as desired, if to be used therapeutically, the polypeptides should be substantially free of endotoxin.

Methods for expression of single chain antibodies and/or refolding to an appropriate active form, including single chain antibodies, from bacteria such as E. coli have been described and are well-known and are applicable to the antibodies of this invention. See, Buchner et al., Anal. Biochem. 205:263 270 (1992); Pluckthun, Biotechnology 9:545 (1991); Huse et al., Science 246:1275 (1989) and Ward et al., Nature 341:544 (1989), all incorporated by reference herein.

Often, functional heterologous proteins from E. coli or other bacteria are isolated from inclusion bodies and require solubilization using strong denaturants, and subsequent refolding. During the solubilization step, as is well-known in the art, a reducing agent must be present to reduce disulfide bonds. An exemplary buffer with a reducing agent is: 0.1 M Tris pH 8, 6 M guanidine, 2 mM EDTA, 0.3 M DTE (dithioerythritol). Reoxidation of the disulfide bonds can occur in the presence of low molecular weight thiol reagents in reduced and oxidized form, as described in Saxena et al., Biochemistry 9: 5015 5021 (1970), incorporated by reference herein, and especially described by Buchner et al., supra.

Renaturation is typically accomplished by dilution (e.g., 100-fold) of the denatured and reduced protein into refolding buffer. An exemplary buffer is 0.1 M Tris, pH 8.0, 0.5 M L-arginine, 8 mM oxidized glutathione (GSSG), and 2 mM EDTA.

As a modification to the two chain antibody purification protocol, the heavy and light chain regions are separately solubilized and reduced and then combined in the refolding solution. A preferred yield is obtained when these two proteins are mixed in a molar ratio such that a 5 fold molar excess of one protein over the other is not exceeded. It is desirable to add excess oxidized glutathione or other oxidizing low molecular weight compounds to the refolding solution after the redox-shuffling is completed.

VI. Pseudomonas Exotoxin and Other Toxins

Toxins can be employed with antibodies of the present invention to yield immunotoxins. Exemplary toxins include ricin, abrin, diphtheria toxin and subunits thereof, as well as botulinum toxins A through F. These toxins are readily available from commercial sources (e.g., Sigma Chemical Company, St. Louis, Mo.). Diphtheria toxin is isolated from Corynebacterium diphtheriae. Ricin is the lectin RCA60 from Ricinus communis (Castor bean). The term also references toxic variants thereof. For example, see, U.S. Pat. Nos. 5,079,163 and 4,689,401. Ricinus communis agglutinin (RCA) occurs in two forms designated RCA.sub.60 and RCA.sub.120 according to their molecular weights of approximately 65 and 120 kD, respectively (Nicholson & Blaustein, J. Biochim. Biophys. Acta 266:543 (1972)). The A chain is responsible for inactivating protein synthesis and killing cells. The B chain binds ricin to cell-surface galactose residues and facilitates transport of the A chain into the cytosol (Olsnes et al., Nature 249:627 631 (1974) and U.S. Pat. No. 3,060,165).

Abrin includes toxic lectins from Abrus precatorius. The toxic principles, abrin a, b, c, and d, have a molecular weight of from about 63 and 67 kD and are composed of two disulfide-linked polypeptide chains A and B. The A chain inhibits protein synthesis; the B-chain (abrin-b) binds to D-galactose residues (see, Funatsu et al., Agr. Biol. Chem. 52:1095 (1988); and Olsnes, Methods Enzymol. 50:330 335 (1978)).

In preferred embodiments of the present invention, the toxin is a Pseudomonas exotoxin A (PE). Native PE is an extremely active monomeric protein (molecular weight 66 kD), secreted by Pseudomonas aeruginosa, which inhibits protein synthesis in eukaryotic cells. The native PE sequence is provided in U.S. Pat. No. 5,602,095, incorporated herein by reference. The method of action is inactivation of elongation factor 2 (EF-2) by ADP-ribosylation. The exotoxin contains three structural domains that act in concert to cause cytotoxicity. Domain Ia (amino acids 1 252) mediates cell binding. Domain II (amino acids 253 364) is responsible for translocation into the cytosol and domain III (amino acids 400 613) mediates ADP ribosylation of elongation factor 2. The function of domain Ib (amino acids 365 399) remains undefined, although a large part of it, amino acids 365 380, can be deleted without loss of cytotoxicity.

The term "Pseudomonas exotoxin" as used herein refers to the native sequence, cytotoxic fragments of the native sequence, and conservatively modified variants of native PE and its cytotoxic fragments. In preferred embodiments, the PE molecule has been modified to delete domain Ia, to reduce or eliminate non-specific binding of the toxin. Cytotoxic fragments of PE include those which are cytotoxic with or without subsequent proteolytic or other processing in the target cell (e.g., as a protein or pre-protein). Cytotoxic fragments of PE include PE40, PE38, and PE35. PE40 is a truncated derivative of PE as previously described in the art. See, Pai et al., Proc. Nat'l Acad. Sci. USA 88:3358 62 (1991); and Kondo et al., J. Biol. Chem. 263:9470 9475 (1988). PE35 is a 35 kD carboxyl-terminal fragment of PE composed of a met at position 280 followed by amino acids 281 364 and 381 613 of native PE. PE38 is a truncated PE pro-protein composed of amino acids 253 364 and 381 613 of PE. PE38 is activated to its cytotoxic form upon processing within a cell (see U.S. Pat. No. 5,608,039, incorporated herein by reference).

In particularly preferred embodiments, PE38 is the toxic moiety of the immunotoxin of this invention. The cytotoxic fragments PE35 and PE40, however, can also be used; these fragments are disclosed in U.S. Pat. Nos. 5,602,095 and 4,892,827, each of which is incorporated herein by reference. Based on work performed with MR1-based immunotoxins, KDEL is a preferred modification of the C-terminal sequence (see Lorimer et al., Proc Natl Acad Sci USA 93:14815 14820 at 14818 (1996).

A. Conservatively Modified Variants of PE

Conservatively modified variants of PE or cytotoxic fragments thereof have at least 80% sequence similarity, preferably at least 85% sequence similarity, more preferably at least 90% sequence similarity, and most preferably at least 95% sequence similarity at the amino acid level, with the PE of interest, such as PE38.

The term "conservatively modified variants" applies to both amino acid and nucleic acid sequences. With respect to particular nucleic acid sequences, conservatively modified variants refer to those nucleic acid sequences which encode identical or essentially identical amino acid sequences, or if the nucleic acid does not encode an amino acid sequence, to essentially identical nucleic acid sequences. Because of the degeneracy of the genetic code, a large number of functionally identical nucleic acids encode any given polypeptide. For instance, the codons GCA, GCC, GCG and GCU all encode the amino acid alanine. Thus, at every position where an alanine is specified by a codon, the codon can be altered to any of the corresponding codons described without altering the encoded polypeptide. Such nucleic acid variations are "silent variations," which are one species of conservatively modified variations. Every nucleic acid sequence herein which encodes a polypeptide also describes every possible silent variation of the nucleic acid. One of skill will recognize that each codon in a nucleic acid (except AUG, which is ordinarily the only codon for methionine, and UGG, which is the only codon for tryptophan) can be modified to yield a functionally identical molecule. Accordingly, each silent variation of a nucleic acid which encodes a polypeptide is implicit in each described sequence.

As to amino acid sequences, one of skill will recognize that individual substitutions, deletions or additions to a nucleic acid, peptide, polypeptide, or protein sequence which alters, adds or deletes a single amino acid or a small percentage of amino acids in the encoded sequence is a "conservatively modified variant" where the alteration results in the substitution of an amino acid with a chemically similar amino acid.

B. Assaying for Cytotoxicity of PE

Pseudomonas exotoxins employed in the invention can be assayed for the desired level of cytotoxicity by assays well known to those of skill in the art. For example, cytotoxicity is often measured using inhibition by Pseudomonas exotoxin of protein synthesis as a surrogate measure. See, e.g., Lorimer et al., Proc. Natl. Acad. Sci. USA 93:14815 14820 (1996) at 14818. Conveniently, this can be done by measuring the uptake of a radiolabeled amino acid (as taught in Prior et al., Cell 64: 1017 1023 (1991)) by cells which do not normally express EGFRvIII but which have been transfected with EGFRvIII cDNA. In the Examples below, uptake of tritiated leucine was measured in NR6M cells (Swiss 3T3 cells selected for lack of expression of mouse EGFR and transfected with human EGFRvIII cDNA). Thus, cytotoxic fragments of PE and conservatively modified variants of such fragments can be readily assayed for cytotoxicity.

A large number of candidate PE molecules can be assayed simultaneously for cytotoxicity by methods well known in the art. For example, subgroups of the candidate molecules can be assayed for cytotoxicity. Positively reacting subgroups of the candidate molecules can be continually subdivided and reassayed until the desired cytotoxic fragment(s) is identified. Such methods allow rapid screening of large numbers of cytotoxic fragments or conservative variants of PE.

C. Other Therapeutic Moieties

Antibodies of the present invention can also be used to target any number of different diagnostic or therapeutic compounds to cells expressing EGFRvIII on their surface. Thus, an antibody of the present invention, such as an anti-EGFRvIII scFv, may be attached directly or via a linker to a drug that is to be delivered directly to cells bearing EGFRvIII. Therapeutic agents include such compounds as nucleic acids, proteins, peptides, amino acids or derivatives, glycoproteins, radioisotopes, lipids, carbohydrates, or recombinant viruses. Nucleic acid therapeutic and diagnostic moieties include antisense nucleic acids, derivatized oligonucleotides for covalent cross-linking with single or duplex DNA, and triplex forming oligonucleotides.

Alternatively, the molecule linked to an anti-EGFRvIII antibody may be an encapsulation system, such as a liposome or micelle that contains a therapeutic composition such as a drug, a nucleic acid (e.g. an antisense nucleic acid), or another therapeutic moiety that is preferably shielded from direct exposure to the circulatory system. Means of preparing liposomes attached to antibodies are well known to those of skill in the art. See, for example, U.S. Pat. No. 4,957,735; and Connor et al., Pharm. Ther. 28:341 365 (1985).

D. Detectable Labels

Antibodies of the present invention may optionally be covalently or non-covalently linked to a detectable label. Detectable labels suitable for such use include any composition detectable by spectroscopic, photochemical, biochemical, immunochemical, electrical, optical or chemical means. Useful labels in the present invention include magnetic beads (e.g. DYNABEADS.RTM.), fluorescent dyes (e.g., fluorescein isothiocyanate, Texas red, rhodamine, green fluorescent protein, and the like), radiolabels (e.g., .sup.3H, .sup.125I, .sup.35S, .sup.14C, or .sup.32P), enzymes (e.g., horse radish peroxidase, alkaline phosphatase, luciferase, and others commonly used in an ELISA), and colorimetric labels such as colloidal gold or colored glass or plastic (e.g. polystyrene, polypropylene, latex, etc.) beads.

Means of detecting such labels are well known to those of skill in the art. Thus, for example, radiolabels may be detected using photographic film or scintillation counters, fluorescent markers may be detected using a photodetector to detect emitted illumination. Enzymatic labels are typically detected by providing the enzyme with a substrate and detecting the reaction product produced by the action of the enzyme on the substrate, and colorimetric labels are detected by simply visualizing the colored label.

E. Conjugation to the Antibody

In a non-recombinant embodiment of the invention, effector molecules, e.g., therapeutic, diagnostic, or detection moieties, are linked to the anti-EGFRvIII antibodies of the present invention using any number of means known to those of skill in the art. Both covalent and noncovalent attachment means may be used with anti-EGFRvIII antibodies of the present invention.

The procedure for attaching an effector molecule to an antibody will vary according to the chemical structure of the EM. Polypeptides typically contain variety of functional groups; e.g., carboxylic acid (COOH), free amine (--NH.sub.2) or sulfhydryl (--SH) groups, which are available for reaction with a suitable functional group on an antibody to result in the binding of the effector molecule.

Alternatively, the antibody is derivatized to expose or attach additional reactive functional groups. The derivatization may involve attachment of any of a number of linker molecules such as those available from Pierce Chemical Company, Rockford Ill.

A "linker", as used herein, is a molecule that is used to join the antibody to the effector molecule. The linker is capable of forming covalent bonds to both the antibody and to the effector molecule. Suitable linkers are well known to those of skill in the art and include, but are not limited to, straight or branched-chain carbon linkers, heterocyclic carbon linkers, or peptide linkers. Where the antibody and the effector molecule are polypeptides, the linkers may be joined to the constituent amino acids through their side groups (e.g., through a disulfide linkage to cysteine). However, in a preferred embodiment, the linkers will be joined to the alpha carbon amino and carboxyl groups of the terminal amino acids.

In some circumstances, it is desirable to free the effector molecule from the antibody when the immunoconjugate has reached its target site. Therefore, in these circumstances, immunoconjugates will comprise linkages which are cleavable in the vicinity of the target site. Cleavage of the linker to release the effector molecule from the antibody may be prompted by enzymatic activity or conditions to which the immunoconjugate is subjected either inside the target cell or in the vicinity of the target site. When the target site is a tumor, a linker which is cleavable under conditions present at the tumor site (e.g. when exposed to tumor-associated enzymes or acidic pH) may be used.

In view of the large number of methods that have been reported for attaching a variety of radiodiagnostic compounds, radiotherapeutic compounds, drugs, toxins, and other agents to antibodies one skilled in the art will be able to determine a suitable method for attaching a given agent to an antibody or other polypeptide.

VII. Pharmaceutical Compositions and Administration

The antibody and/or immunoconjugate compositions of this invention (i.e., PE linked to an antibody with a Kd for an epitope of EGFRvIII at least 1 nM lower than that of MR1), are useful for administration into the brain. Use of immunotoxins for brain tumor therapy was recently reviewed by Oldfield, E., and Youle, R., Curr. Top. Microbiol. Immunol. 234:97 114 (1998). Small polypeptides cross the blood brain barrier. For longer polypeptides that do not the cross blood brain barrier, methods of administering proteins to the brain are well known. For example, proteins, polypeptides, other compounds and cells can be delivered to the mammalian brain via intracerebroventricular (ICV) injection or via a cannula (see, e.g., Motta & Martini, Proc. Soc. Exp. Biol. Med. 168:62 64 (1981); Peterson et al., Biochem. Pharamacol. 31:2807 2810 (1982); Rzepczynski et al., Metab. Brain Dis. 3:211 216 (1988); Leibowitz et al., Brain Res. Bull. 21:905 912 (1988); Sramka et al., Stereotact. Funct. Neurosurg. 58:79 83 (1992); Peng et al., Brain Res. 632:57 67 (1993); Chem et al., Exp. Neurol. 125:72 81 (1994); Nikkhah et al., Neuroscience 63:57 72 (1994); Anderson et al., J. Comp. Neurol. 357:296 317 (1995); and Brecknell & Fawcett, Exp. Neurol. 138:338 344 (1996)).

Thus, for example, glioblastoma may be treated by localized delivery by cannula or by syringe to the tissue surrounding the tumor, or more generally within the central nervous system compartment by ICV. Additionally, the immunoconjugates can be administered systemically where, for example, the patient's glioblastoma has damaged the epithelial cells sufficiently to permit breach of the blood-brain barrier.

The antibody or immunoconjugates of the invention can also be administered locally or systemically for treating breast, ovarian, and lung carcinomas. For example, these malignancies can be treated by direct injection into tumors which cannot be surgically excised. These carcinomas can also be treated by parenteral administration of the immunoconjugates to, for example, locate and kill any metastatic cells which have not yet formed tumors of sufficient size to be treated with radiation or surgery or, indeed, to be readily detected.

The compositions for administration will commonly comprise a solution of the antibody and/or immunoconjugate dissolved in a pharmaceutically acceptable carrier, preferably an aqueous carrier. A variety of aqueous carriers can be used, e.g., buffered saline and the like. These solutions are sterile and generally free of undesirable matter. These compositions may be sterilized by conventional, well known sterilization techniques. The compositions may contain pharmaceutically acceptable auxiliary substances as required to approximate physiological conditions such as pH adjusting and buffering agents, toxicity adjusting agents and the like, for example, sodium acetate, sodium chloride, potassium chloride, calcium chloride, sodium lactate and the like. The concentration of fusion protein in these formulations can vary widely, and will be selected primarily based on fluid volumes, viscosities, body weight and the like in accordance with the particular mode of administration selected and the patient's needs.

Thus, a typical pharmaceutical immunotoxin composition of the present invention for administration to the brain would be about 1.2 to 1200 .mu.g per day. A typical composition for intravenous administration to treat breast, ovarian, or lung carcinoma are about 0.1 to 10 mg per patient per day. Dosages from 0.1 up to about 100 mg per patient per day may be used, particularly if the drug is administered to a secluded site and not into the circulatory or lymph system, such as into a body cavity or into a lumen of an organ. Actual methods for preparing administrable compositions will be known or apparent to those skilled in the art and are described in more detail in such publications as REMINGTON'S PHARMACEUTICAL SCIENCE, 19TH ED., Mack Publishing Company, Easton, Pa. (1995).

The compositions of the present invention can be administered for therapeutic treatments. In therapeutic applications, compositions are administered to a patient suffering from a disease, such as a glioblastoma, breast carcinoma, ovarian carcinoma, or lung carcinoma, in an amount sufficient to at least slow or partially arrest the disease or its complications. An amount adequate to accomplish this is defined as a "therapeutically effective dose." Amounts effective for this use will depend upon the severity of the disease and the general state of the patient's health. An effective amount of the compound is that which provides either subjective relief of a symptom(s) or an objectively identifiable improvement as noted by the clinician or other qualified observer.

Single or multiple administrations of the compositions are administered depending on the dosage and frequency as required and tolerated by the patient. In any event, the composition should provide a sufficient quantity of the proteins of this invention to effectively treat the patient. Preferably, the dosage is administered once but may be applied periodically until either a therapeutic result is achieved or until side effects warrant discontinuation of therapy. Generally, the dose is sufficient to treat or ameliorate symptoms or signs of disease without producing unacceptable toxicity to the patient.

Controlled release parenteral formulations of the immunoconjugate compositions of the present invention can be made as implants, oily injections, or as particulate systems. For a broad overview of protein delivery systems see, Banga, A. J., THERAPEUTIC PEPTIDES AND PROTEINS: FORMULATION, PROCESSING, AND DELIVERY SYSTEMS, Technomic Publishing Company, Inc., Lancaster, Pa., (1995) incorporated herein by reference. Particulate systems include microspheres, microparticles, microcapsules, nanocapsules, nanospheres, and nanoparticles. Microcapsules contain the therapeutic protein as a central core. In microspheres the therapeutic is dispersed throughout the particle. Particles, microspheres, and microcapsules smaller than about 1 .mu.m are generally referred to as nanoparticles, nanospheres, and nanocapsules, respectively. Capillaries have a diameter of approximately 5 .mu.m, so only nanoparticles are administered intravenously. Microparticles are typically around 100 .mu.m in diameter and are administered subcutaneously or intramuscularly. See, e.g., Kreuter, J., COLLOIDAL DRUG DELIVERY SYSTEMS, J. Kreuter, ed., Marcel Dekker, Inc., New York, N.Y., pp. 219 342 (1994); and Tice & Tabibi, TREATISE ON CONTROLLED DRUG DELIVERY, A. Kydonieus, ed., Marcel Dekker, Inc. New York, N.Y., pp. 315 339, (1992) both of which are incorporated herein by reference.

Polymers can be used for ion-controlled release of immunoconjugate compositions of the present invention. Various degradable and nondegradable polymeric matrices for use in controlled drug delivery are known in the art (Langer, R., Accounts Chem. Res. 26:537 542 (1993)). For example, the block copolymer, polaxamer 407 exists as a viscous yet mobile liquid at low temperatures but forms a semisolid gel at body temperature. It has shown to be an effective vehicle for formulation and sustained delivery of recombinant interleukin-2 and urease (Johnston et al., Pharm. Res. 9:425 434 (1992); and Pee et al., J. Parent. Sci. Tech. 44(2):58 65 (1990)). Alternatively, hydroxyapatite has been used as a microcarrier for controlled release of proteins (Ijntema et al., Int. J. Pharm. 112:215 224 (1994)). In yet another aspect, liposomes are used for controlled release as well as drug targeting of the lipid-capsulated drug (Betageri et al., LIPOSOME DRUG DELIVERY SYSTEMS, Technomic Publishing Co., Inc., Lancaster, Pa. (1993)). Numerous additional systems for controlled delivery of therapeutic proteins are known. See, e.g., U.S. Pat. Nos. 5,055,303, 5,188,837, 4,235,871, 4,501,728, 4,837,028 4,957,735 and 5,019,369, 5,055,303; 5,514,670; 5,413,797; 5,268,164; 5,004,697; 4,902,505; 5,506,206, 5,271,961; 5,254,342 and 5,534,496, each of which is incorporated herein by reference.

Among various uses of the immunotoxins of the present invention are included disease conditions caused by specific human cells expressing EGFRvIII that may be eliminated by the toxic action of the immunoconjugate. One preferred application for the immunotoxins of the invention is the treatment of malignant cells expressing EGFRvIII. Exemplary malignant cells include cells of glioblastoma, breast carcinoma, ovarian carcinoma, and lung carcinoma.

VIII. Diagnostic Kits and In Vitro Uses

In another embodiment, this invention provides for kits for the detection of EGFRvIII in a biological sample. A "biological sample" as used herein is a sample of biological tissue or fluid that contains EGFRvIII. Such samples include, but are not limited to, tissue from biopsy, sputum, amniotic fluid, blood, and blood cells (e.g., white cells). Fluid samples may be of some interest, but are generally not preferred herein since detectable concentrations of EGFRvIII are rarely found in such a sample. Biological samples also include sections of tissues, such as frozen sections taken for histological purposes. A biological sample is typically obtained from a multicellular eukaryote, preferably a mammal such as rat, mouse, cow, dog, guinea pig, or rabbit, and more preferably a primate, such as a macaque, chimpanzee. Most preferably, the biological sample is from a human.

Kits will typically comprise an anti-EGFRvIII scFv of the present invention, which has a higher affinity for EGFRvIII than does MR1 scFv.

In addition the kits will typically include instructional materials disclosing means of use of an antibody of the present invention (e.g. for detection of EGFRvIII-containing cells in a sample). The kits may also include additional components to facilitate the particular application for which the kit is designed. Thus, for example, the kit may additionally contain means of detecting the label (e.g. enzyme substrates for enzymatic labels, filter sets to detect fluorescent labels, appropriate secondary labels such as a sheep anti-mouse-HRP, or the like). The kits may additionally include buffers and other reagents routinely used for the practice of a particular method. Such kits and appropriate contents are well known to those of skill in the art.

In one embodiment of the present invention, the diagnostic kit comprises an immunoassay. As described above, although the details of the immunoassays of the present invention may vary with the particular format employed, the method of detecting EGFRvIII in a biological sample generally comprises the steps of contacting the biological sample with an antibody which specifically reacts, under immunologically reactive conditions, to EGFRvIII with higher affinity to EGFRvIII than does MR1 scFv. The antibody is allowed to bind to EGFRvIII under immunologically reactive conditions, and the presence of the bound antibody is detected directly or indirectly.

Due to the increased affinity of antibodies developed by the methods taught herein, and of the scFv designated MR1-1, in particular, the antibodies provided herein will be especially useful as diagnostic agents and in in vitro assays to detect the presence of EGFRvIII in biological samples. For example, MR1-1, and other antibodies made by the methods taught herein, can be used as the targeting moieties of immunoconjugates in immunohistochemical assays to determine whether a sample contains cells expressing EGFRvIII. If the sample is one taken from a tissue of a patient which should not normally express EGFRvIII, detection of EGFRvIII would indicate either that the patient has a cancer characterized by the presence of EGFRvIII-expressing cells, or that a treatment for such a cancer has not yet been successful at eradicating the cancer. Persons of skill in the art will also appreciate that the anti-EGFRvIII antibodies of the invention, coupled to an appropriate label, can likewise be used in vivo to detect the presence of EGFRvIII expressing cells, thereby indicating either that the patient has a cancer characterized by the presence of EGFRvIII-expressing cells, or that a treatment for such a cancer has not yet been successful at eradicating the cancer.
 

Claim 1 of 30 Claims

1. An isolated polypeptide comprising an antibody heavy chain variable region ("VH") and an antibody light chain variable region ("VL"), each region comprising three complementarity determining regions ("CDRs"), which CDRs of each region are numbered sequentially CDR1 to CDR3 starting from the amino terminus, the polypeptide when made into an immunotoxin with a Pseudomonas exotoxin A or cytotoxic fragment thereof ("PE") forming an immunotoxin which binds to epidermal growth factor receptor type III ("EGFRvIII") antigen and which has a cytotoxicity to cells expressing said antigen at least equal to the cytotoxicity to said cells of an immunotoxin of parental antibody MR1 (SEQ ID NO.:18) and said PE, and a higher yield, when made into an immunotoxin with said PE, than that of MR1 when made into an immunotoxin with said PE, wherein: (a) CDRs 1 3, respectively of the V.sub.H of the polypeptide have the sequence of CDRs 1 3, respectively of parental antibody MR1 V.sub.H, except for: (i) substitution of an amino acid selected from the group consisting of proline and tryptophan for the serine at position 98 of the CDR3 of the heavy chain variable region of antibody MR1, and (ii) substitution of an amino acid selected from the group consisting of: tyrosine, asparagine, tryptophan, isoleucine, phenylalanine, serine, and valine for the threonine at position 99 of the CDR3 of the heavy chain variable region of antibody MR 1 and, optionally, (iii) a substitution in CDR1 or CDR2 of said heavy chain variable region of at least one amino acid encoded by a codon that comprises a nucleotide belonging to a hot spot motif selected from AGY or RGYW, wherein R is A or G, Y is C or T and W is A or T; and (b) CDRs 1 3, respectively, of the V.sub.L of the polypeptide have: (i) the sequence of CDRs 1 3, respectively, of antibody MR1 V.sub.L or, (ii) the sequence of CDRs 1 3, respectively, of antibody MR1 V.sub.L except for a substitution in one or more of said CDRs 1 3 of said polypeptide at least one amino acid encoded by a codon that comprises a nucleotide belonging to a hot spot motif selected from AGY or RGYW, wherein R is A or G, Y is C or T and W is A or T.

____________________________________________
If you want to learn more about this patent, please go directly to the U.S. Patent and Trademark Office Web site to access the full patent.