Anti-variable MUC1* antibodies and uses thereof

[0001] The instant application contains a Sequence Listing which has been submitted electronically in ASCII format and is hereby incorporated by reference in its entirety. Said ASCII copy, created on Mar. 19, 2024, is named 56699-741_301_SL.txt and is 1,025,433 bytes in size.

[0002] The present application relates to human, humanized and non-human anti-MUC1* antibodies and methods of making and using them. The present application also relates to using an immune cell transfected or transduced with a cleavage enzyme for the treatment of cancer. The present invention also relates to using an immune cell transfected or transduced with a CAR and another protein for the treatment of cancer.

• • [0003] We previously discovered that a cleaved form of the MUC1 (SEQ ID NO:1) transmembrane protein is a growth factor receptor that drives the growth of over 75% of all human cancers. The cleaved form of MUC1, which we called MUC1* (pronounced muk 1 star), is a powerful growth factor receptor. Cleavage and release of the bulk of the extracellular domain of MUC1 unmasks a binding site for activating ligands dimeric NME1, NME6, NME7, NME7_AB, NME7-X1 or NME8. It is an ideal target for cancer drugs as it is aberrantly expressed on over 75% of all cancers and is likely overexpressed on an even higher percentage of metastatic cancers (Mahanta et al. (2008) A Minimal Fragment of MUC1 Mediates Growth of Cancer Cells. PLOS ONE 3 (4): e2054. doi: 10.1371/journal.pone.0002054; Fessler et al. (2009), “MUC1* is a determinant of trastuzumab (Herceptin) resistance in breast cancer cells,” Breast Cancer Res Treat. 118 (1): 113-124). After MUC1 cleavage most of its extracellular domain is shed from the cell surface. The remaining portion has a truncated extracellular domain that comprises most or all of the primary growth factor receptor sequence called PSMGFR (SEQ ID NO: 2).

[0004] Antibodies are increasingly used to treat human diseases. Antibodies generated in non-human species have historically been used as therapeutics in humans, such as horse antibodies. More recently, antibodies are engineered or selected so that they contain mostly, or all, human sequences in order to avoid a generalized rejection of the foreign antibody. The process of engineering recognition fragments of a non-human antibody into a human antibody is generally called ‘humanizing’. The amount of non-human sequences that are used to replace the human antibody sequences determines whether they are called chimeric, humanized or fully human.

[0005] Alternative technologies exist that enable generation of humanized or fully human antibodies. These strategies involve screening libraries of human antibodies or antibody fragments and identifying those that bind to the target antigen, rather than immunizing an animal with the antigen. Another approach is to engineer the variable region(s) of an antibody into an antibody-like molecule. Another approach involves immunizing a humanized animal. The present invention is intended to also encompass these approaches for use with recognition fragments of antibodies that the inventors have determined bind to the extracellular domain of MUC1*.

[0006] In addition to treating patients with an antibody, cancer immunotherapies have recently been shown to be effective in the treatment of blood cancers. One cancer immunotherapy, called CAR T (chimeric antigen receptor T cell) therapy, engineers a T cell so that it expresses a chimeric receptor having an extra cellular domain that recognizes a tumor antigen, a transmembrane domain and cytoplasmic tail comprising T cell signaling and co-stimulatory components (Dai H, Wang Y, Lu X, Han W. (2016) Chimeric Antigen Receptors Modified T-Cells for Cancer Therapy. J Natl Cancer Inst. 108 (7): djv439). Such receptor is composed of a single chain antibody fragment (scFv) that recognizes a tumor antigen, linked to a T cell transmembrane, signaling domain and co-stimulatory domain or domains. Upon binding of the receptor to a cancer associated antigen, a signal is transmitted resulting in T-cell activation, propagation and the targeted killing of the cancer cells. In practice, T cells are isolated from a patient or donor and transduced with a CAR, expanded and then injected back into the patient. If from a donor, the immune cells may be mutated or engineered such that they do not induce graft versus host disease in the recipient. When the CAR T cells bind to the antigen on a cancer cell, the CAR T cells attack the cancer cells and then expand that population of T cells.

[0007] Thus far, CAR T therapies have been very successful in the treatment of blood cancers but as yet have not shown efficacy against solid tumors in humans. Because most blood cancers are B cell malignancies, the CAR T cells can just eliminate all of the patient's B cells without causing serious harm to the patient. There is no B cell equivalent in solid tumors. Most tumor associated antigens are also expressed on normal tissues; they are just expressed at a higher level in cancerous tissues. Thus, the challenge is to develop an antibody that recognizes an epitope on a tumor associated antigen that is somehow different in the context of the tumor compared to normal tissue. To further minimize the risk of off-tumor/on-target killing of normal tissues, the antibody should recognize and bind to cancerous tissues at least two-times more than normal tissues. Antibodies that are not so cancer selective may be used therapeutically if they are inducibly expressed at the tumor site.

[0008] Another cancer therapy that incorporates cancer selective antibodies is Bi-specific T cell Engagers, also called BiTEs™. The BiTE™ approach attempts to eliminate the CAR T associated risk of off-tumor/on-target effects. Unlike CAR T, BiTES™ are bispecific antibodies that should not pose any greater risk than regular antibody-based therapies. However, unlike typical anti-cancer antibodies that bind to and block a cancer antigen, BiTES™ are designed to bind to an antigen on the tumor cell and simultaneously bind to an antigen on an immune cell, such as a T cell. In this way, a BiTE™ recruits the T cell to the tumor. BiTES™ are engineered proteins that simultaneously bind to a cancer associated antigen and a T-cell surface protein such as CD3-epsilon. BiTES™ are antibodies made by genetically linking the scFv's of an antibody that binds to a T cell antigen, like anti-CD3-epsilon to a scFv of a therapeutic monoclonal antibody that binds to a cancer antigen (Patrick A. Baeuerle, and Carsten Reinhardt (2009) Bispecific T-cell engaging antibodies for cancer therapy. Cancer Res. 69 (12): 4941-4944). A drawback of BiTE™ technology is that, unlike CAR T cells, they do not expand in the patient, so have limited persistence.

[0009] Yet another cancer therapy that incorporates cancer selective antibodies is antibody drug conjugate, also called ADC, technology. In this case, a toxin, or a precursor to a toxin, is linked to a cancer selective antibody. Unlike CAR T cells that use the CD8 positive T cell's natural killing to kill cancer cells, ADCs carry a toxic payload to the tumor. Drawbacks of ADCs include the potential of delivering the toxic payload to normal cells and that most ADCs require binding to a cell surface molecule which then gets internalized after binding, with an approximate 10,000 surface molecule required for resultant cell death.

[0010] In one aspect, the present invention is directed to a non-human, human or humanized anti-MUC1* antibody or antibody fragment or antibody-like protein that binds to a region on extracellular domain of MUC1 isoform or cleavage product that is devoid of the tandem repeat domains. The non-human, human or humanized anti-MUC1* antibody or antibody fragment or antibody-like protein may specifically bind to

• • [0011] (i) PSMGFR region of MUC1;
  • [0012] (ii) PSMGFR peptide;
  • [0013] (iii) a peptide having amino acid sequence of QFNQYKTEAASRYNLTISDVSVSDVPFPFSAQSGA (N-10) (SEQ ID NO:3)
  • [0014] (iv) a peptide having amino acid sequence of
  • [0015] ASRYNLTISDVSVSDVPFPFSAQSGA (N-19) (SEQ ID NO:4)
  • [0016] (v) a peptide having amino acid sequence of
  • [0017] NLTISDVSVSDVPFPFSAQSGA (N-23) (SEQ ID NO:5)
  • [0018] (vi) a peptide having amino acid sequence of
  • [0019] ISDVSVSDVPFPFSAQSGA (N-26) (SEQ ID NO:6)
  • [0020] (vii) a peptide having amino acid sequence of
  • [0021] SVSDVPFPFSAQSGA (N-30) (SEQ ID NO:7)
  • [0022] (viii) a peptide having amino acid sequence of
  • [0023] QFNQYKTEAASRYNLTISDVSVSDVPFPFS (N-10/C-5) (SEQ ID NO:8)
  • [0024] (ix) a peptide having amino acid sequence of
  • [0025] ASRYNLTISDVSVSDVPFPFS (N-19/C-5) (SEQ ID NO:9)
  • [0026] (x) a peptide having amino acid sequence of
  • [0027] FPFSAQSGA (SEQ ID NO:10)

[0028] The non-human, human or humanized antibody may be IgG1, IgG2, IgG3, IgG4 or IgM. The human or humanized antibody fragment or antibody-like protein may be scFv or scFv-Fc.

[0029] The murine, camelid, human or humanized antibody, antibody fragment or antibody-like protein as in above may comprise a heavy chain variable region and light chain variable region which is derived from mouse monoclonal MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, and H11 antibody, and has at least 80%, 90% or 95% or 98% sequence identity to the mouse monoclonal MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, and H11 antibody. The heavy chain variable region of CDR1 and CDR2 may have at least 90% or 95% or 98% sequence identity to the particularly indicated antibody heavy chain variable region sequence set forth in the present application in the sequence listing, and the light chain variable region of CDR1 and CDR2 may have at least 90% or 95% or 98% sequence identity to the particularly indicated antibody heavy chain variable region sequence set forth in the present application in the sequence listing section. The heavy chain variable region of CDR3 may have at least 80% or 85% or 90% sequence identity to the particularly indicated antibody heavy chain variable region sequence set forth in the present application in the sequence listing, and the light chain variable region of CDR3 may have at least 80% or 85% or 90% sequence identity to the particularly indicated antibody heavy chain variable region sequence set forth in the present application in the sequence listing section.

[0030] The murine, camelid, human or humanized antibody, antibody fragment or antibody-like protein according to above may include complementarity determining regions (CDRs) in the heavy chain variable region and light chain variable region having at least 90% or 95% or 98% sequence identity to the particularly indicated antibody heavy chain CDR1, CDR2 or CDR3 region and light chain CDR1, CDR2 or CDR3 region sequences set forth in the present application in the sequence listing section.

[0031] In another aspect, the present invention is directed to an anti-MUC1* extracellular domain antibody or anti-N-10 antibody, which may be any of the antibodies described above, comprised of sequences represented by humanized IgG2 heavy chain, or humanized IgG1 heavy chain, paired with humanized Kappa light chain, or humanized Lambda light chain. The humanized IgG2 heavy chain may be SEQ ID NOS: 53, humanized IgG1 heavy chain may be SEQ ID NO:57, humanized Kappa light chain may be SEQ ID NO:108, and humanized Lambda light chain may be SEQ ID NO:112, or a sequence having 90%, 95% or 98% sequence identity thereof.

[0032] In another aspect, the invention is directed to an anti-MUC1* extracellular domain antibody or anti-N-10 antibody comprised of sequences of a humanized MN-C2 represented by humanized IgG1 heavy chain, humanized IgG2 heavy chain, paired with humanized Lambda light chain, and humanized Kappa light chain.

[0033] In another aspect, the invention is directed to an anti-MUC1* extracellular domain antibody or anti-N-10 antibody comprised of sequences of a humanized MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11 represented by humanized IgG1 heavy chain or humanized IgG2 heavy chain, paired with humanized Lambda light chain, or humanized Kappa light chain.

[0034] In another aspect, the invention is directed to an antibody that is “like” MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11 in that they have the same or very similar pattern of binding to subsets of peptides derived from the PSMGFR peptide, also do not recognize a linear epitope, competitively inhibit the binding of NME1 or NME7_AB to MUC1*, recognize a MUC1 transmembrane cleavage product produced by cleavage by MMP9 or contain CDR sequences that are at least 80% homologous to the MN-E6, MN-C2, MN-18G12, MN-20A10, MN-25E6, MN-28F9, MN-5C6F3, MN-3C2B1, and MN-1E4 CDR consensus sequences.

[0035] In another aspect, the invention is directed to an antibody that binds to the extra cellular domain of a MUC1 that is devoid of the tandem repeat domain, which may be a cleavage product. In one aspect of the invention, the antibody binds to a peptide having the sequence of QFNQYKTEAASRYNLTISDVSVSDVPFPFSAQSGA (SEQ ID NO: 3) (N-10). In one aspect of the invention, the antibody binds to a peptide having the sequence of ASRYNLTISDVSVSDVPFPFSAQSGA (SEQ ID NO: 4) (N-19). In one aspect of the invention, the antibody binds to a peptide having the sequence of SVSDVPFPFSAQSGA (SEQ ID NO: 7) (N-30). In one aspect of the invention, the antibody binds to a peptide having the sequence of FPFSAQSGA (SEQ ID NO: 10) (N-36). Examples of such antibodies include but are not limited to monoclonal antibodies MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, and H11. The heavy chain and light chain complementary determining region sequences for these antibodies are set forth in the present application in the sequence listing section.

[0036] In one aspect of the invention, one or more of these antibodies is administered to a patient diagnosed with or at risk of developing a cancer. The antibody may be human or humanized. The antibody may be murine or camelid. The antibody may be bivalent or monovalent. The antibody may be a fragment, including a single chain fragment, scFv, of one of the antibodies. The antibody or antibody fragment may be administered directly to the patient or incorporated into a bispecific antibody, a bispecific T cell engager, BiTE, or an antibody drug conjugate, ADC. The antibody or antibody fragment may be incorporated into a T cell receptor, TCR. The sequence of the antibody or antibody fragment may be incorporated into a chimeric antigen receptor, a “CAR”, or other similar entity, then introduced into an immune cell, ex vivo, then administered to a patient diagnosed with or at risk of developing a cancer. The immune cell, which may be a T cell or natural killer cell, may be derived from a donor or from the patient. In one aspect the immune cell is derived from a stem cell that has been directed to differentiate to that immune cell type in vitro. In one aspect, the antibody or a CAR containing sequences of the antibody may be expressed off of an inducible promoter. In one case the antibody or the CAR is expressed upon activation of the T cell or other immune cell. In one instance, the antibody or the CAR of the invention is expressed off of an NFAT response element. In another instance, CAR recognition of a target tumor cell activates the immune cell, leading to NFAT inducible expression of a cytokine, such as IL-12 or IL-18, or expression of a checkpoint inhibitor such as a PD1 inhibitor or a PDL-1 inhibitor. In yet another aspect, CAR recognition of a target tumor cell activates the immune cell, leading to NFAT inducible expression of a second CAR that contains sequences of a second antibody.

[0037] In another aspect, the invention is directed to a murine, camelid, human, humanized anti-MUC1* antibody or antibody fragment or antibody-like protein that binds to the N-10 peptide, according to above, which inhibits the binding of NME protein to MUC1*. The NME may be NME1, NME6, NME7_AB, NME7-X1, NME7 or NME8.

[0038] In yet another aspect, the invention is directed to a single chain variable fragment (scFv) comprising a heavy and light chain variable regions connected via a linker, further comprising CDRs of antibodies that bind to MUC1* extracellular domain. The CDRs may be derived from MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, and H11. The scFv may be one that possesses the SEQ ID NOS: 233, 235 and 237 (MN-E6); SEQ ID NOS: 239, 241, and 243 (MN-C2)

[0039] In still another aspect, the invention is directed to a chimeric antigen receptor (CAR) comprising a scFv or a humanized variable region that binds to the extracellular domain of a MUC1 that is devoid of tandem repeats, a linker molecule, a transmembrane domain and a cytoplasmic domain. The single chain antibody fragment may bind to

• • [0040] (i) PSMGFR region of MUC1;
  • [0041] (ii) PSMGFR peptide;
  • [0042] (iii) a peptide having amino acid sequence of QFNQYKTEAASRYNLTISDVSVSDVPFPFSAQSGA (N-10) (SEQ ID NO:3)
  • [0043] (iv) a peptide having amino acid sequence of
  • [0044] ASRYNLTISDVSVSDVPFPFSAQSGA (N-19) (SEQ ID NO:4)
  • [0045] (v) a peptide having amino acid sequence of
  • [0046] NLTISDVSVSDVPFPFSAQSGA (N-23) (SEQ ID NO:5)
  • [0047] (vi) a peptide having amino acid sequence of
  • [0048] ISDVSVSDVPFPFSAQSGA (N-26) (SEQ ID NO:6)
  • [0049] (vii) a peptide having amino acid sequence of
  • [0050] SVSDVPFPFSAQSGA (N-30) (SEQ ID NO:7)
  • [0051] (viii) a peptide having amino acid sequence of
  • [0052] QFNQYKTEAASRYNLTISDVSVSDVPFPFS (N-10/C-5) (SEQ ID NO:8)
  • [0053] (ix) a peptide having amino acid sequence of
  • [0054] ASRYNLTISDVSVSDVPFPFS (N-19/C-5) (SEQ ID NO:9)
  • [0055] (x) a peptide having amino acid sequence of
  • [0056] FPFSAQSGA (N-36) (SEQ ID NO:10)

[0057] In the CAR as described above, portions of any of the variable regions set forth and described above, or combination thereof may be used in the extracellular domain of the CAR. The CAR also comprises a transmembrane region and a cytoplasmic tail that comprises sequence motifs that signal immune system activation. The extracellular domain may be comprised of murine, camelid, human, non-human, or humanized single chain antibody fragments of an MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, and H11. Additional antibodies from which single chain antibody fragments may made include but are not limited to monoclonal antibodies that are like MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, and H11 in that they have the same or very similar pattern of binding to subsets of peptides derived from the PSMGFR peptide, may not recognize a linear epitope or competitively inhibit the binding of NME1 or NME7_AB to MUC1*, or recognize a MUC1 transmembrane cleavage product produced by cleavage by MMP9 or contain CDR sequences that are at least 80% homologous to the MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, and H11 CDR consensus sequences.

[0058] In the CARs as described above, the extracellular domain may include a murine, camelid, human, non-human or humanized single chain antibody fragments of an MN-E6 scFv set forth as SEQ ID NOS: 233, 235, or 237), MN-C2 scFv (SEQ ID NOS: 239, 241, or 243), or 20A10 scFv as set forth as SEQ ID NOS: 1574-1575, 25E6 scFv as set forth as SEQ ID NOS: 1598-1599.

[0059] In any of the CARs described above, the cytoplasmic tail may be comprised of one or more of signaling sequence motifs CD3-zeta, CD27, CD28, 4-1BB, OX40, CD30, CD40, ICAm-1, LFA-1, ICOS, CD2, CD5, or CD7. In any of the CARs described above, the cytoplasmic tails may include mutations that dampen signaling. Such mutations include but are not limited to Tyrosines that are mutated to inhibit phosphorylation and signaling (Salter et al, 2018). In any of the CARs described above, the ITAMs of CD3-zeta may be mutated to inhibit or dampen signaling (Feucht et al 2019). In any of the CARs described above, the CD3 of the cytoplasmic tail may comprise mutations in the ITAMs including those referred to as 1XX. In any of the CARs described above, the T cell may be engineered to overexpress c-Jun as a method to inhibit T cell exhaustion (Lynn et al 2019).

[0060] In any of the CARs described above, the sequence may be CAR MN-E6 CD28/CD3z (SEQ ID NOS: 298); CAR MN-E6 4-1BB/CD3z (SEQ ID NOS: 301); CAR MN-E6 OX40/CD3z (SEQ ID NOS: 617); CAR MN-E6 CD28/4-1BB/CD3z (SEQ ID NOS: 304); CAR MN-E6 CD28/OX40/CD3z (SEQ ID NOS: 619); CAR MN-C2 CD3z (SEQ ID NOS: 607); CAR MN-C2 CD28/CD3z SEQ ID NOS: 609); CAR MN-C2 4-1BB/CD3z (SEQ ID NOS: 611 and SEQ ID NOS: 719); CAR MN-C2 OX40/CD3z (SEQ ID NOS: 613); CAR MN-C2 CD28/4-1BB/CD3z (SEQ ID NOS: 307); CAR MN-C2 CD28/OX40/CD3z (SEQ ID NOS: 615) or CAR MN-C3 4-1BB/CD3z (SEQ ID NOS: 601).

[0061] In another aspect, the invention is directed to a composition that includes at least two CARs with different extracellular domain units transfected into the same cell, which may be an immune cell, which may be derived from the patient requiring treatment for a cancer. The expression of the second CAR may be inducible and driven by the recognition of a target by the first CAR. The nucleic acid encoding the second CAR may be linked to an inducible promoter. The expression of the second CAR may be induced by an event that occurs specifically when the immune cell mounts an immune response to a target tumor cell. The antibody fragments of one or both of the CARs may direct the cell to a MUC1* positive tumor. The antibody fragments of the first and second CARs may bind to a MUC1* that is produced when MUC1 is cleaved by two different cleavage enzymes. Expression of the second CAR by the inducible promoter may be induced when the antibody fragment of the first CAR engages or binds to a MUC1 or MUC1* on the tumor. One way to do this is to induce expression of the second CAR when, or shortly after, an NFAT protein is expressed or translocated to the nucleus. For example, a sequence derived from an NFAT promoter region is put upstream of the gene for the second CAR. In this way, when the transcription factors that bind to the promoter of the NFAT protein are present in sufficient concentration to bind to and induce transcription of the NFAT protein, they will also bind to that same promoter that is engineered in front of the sequence for transcription of the second CAR. The NFAT protein may be NFAT1 also known as NFATc2, NFAT2 also known as NFATc or NFATc1, NFAT3 also known as NFATc4, NFAT4 also known as NFATc3, or NFAT5. In one aspect of the invention, the NFAT is NFATc1, NFATc3 or NFATc2. In one aspect of the invention, the NFAT is NFAT2 also known as NFATc1. SEQ ID NO:646 shows nucleic acid sequence of the upstream transcriptional regulatory region for NFAT2. The recognition unit of the second CAR may be an antibody fragment or a peptide, wherein the recognition units may bind to NME7, PD-1, PDL-1, or a checkpoint inhibitor.

[0062] The at least two CARs may have one CAR that does not have a tumor antigen targeting recognition unit and the other CAR does have a tumor antigen targeting recognition unit. In another aspect of the invention, one of the extracellular domain recognition units may bind to MUC1* extracellular domain. In another aspect of the invention, one of the extracellular domain recognition units may be an antibody fragment and the other is a peptide, which may be devoid of transmembrane and signaling motifs; the peptide may be a single chain antibody fragment or antibody. In another aspect of the invention, one of the recognition units may bind PD-1 or PDL-1. In another aspect of the invention, one extra cellular domain recognition unit is an anti-MUC1* antibody, antibody fragment or scFv chosen from the group consisting of MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, and H11. The other recognition unit may be a CAR or may be an anti-NME7 antibody.

[0063] In another aspect, the invention is directed to a cell comprising a CAR with an extracellular domain that binds to the extra cellular domain of a MUC1 molecule that is devoid of tandem repeats. In another aspect, the invention is directed to a cell comprising a CAR with an extracellular domain that binds to a MUC1* transfected or transduced cell. The cell that includes the CAR may be an immune system cell, preferably a T cell, a natural killer cell (NK), a dendritic cell or mast cell.

[0064] In another aspect, the invention is directed to an engineered antibody-like protein.

[0065] In another aspect, the invention is directed to a method for treating a disease in a subject comprising administering an antibody according to any claim above, to a person suffering from the disease, wherein the subject expresses MUC1 aberrantly. The disease may be cancer, such as breast cancer, ovarian cancer, pancreatic cancer, lung cancer, colon cancer, gastric cancer or esophageal cancer.

[0066] In another aspect, the invention is directed to an antibody, antibody fragment or scFv comprising variable domain fragments derived from an antibody that binds to an extracellular domain of MUC1 isoform or cleavage product that is devoid of the tandem repeat domains. In a preferred embodiment, the antibody or antibody fragment binds to the N-10 peptide. The variable domain fragments may be derived from mouse monoclonal antibody MN-E6 (SEQ ID NO: 13 and 66) or from the humanized MN-E6 (SEQ ID NO: 39 and 94), or from MN-E6 scFv (SEQ ID NO: 233, 235 and 237). Or, the variable domain fragments may be derived from mouse monoclonal antibody MN-C2 (SEQ ID NO: 119 and 169) or from the humanized MN-C2 (SEQ ID NO: 145 and 195), or from MN-C2 scFv (SEQ ID NO: 239, 241 and 243). Or, the variable domain may be derived from monoclonal antibodies MN-18G12, MN-20A10, MN-25E6, MN-28F9, MN-5C6F3, MN-3C2B1, or MN-1E4. The heavy chain and light chain complementary determining region sequences for these antibodies are also set forth in the sequence listing herein.

[0067] In another aspect, the invention is directed to a method for the treatment of a person diagnosed with, suspected of having or at risk of developing a MUC1 or MUC1* positive cancer involving administering to the person an effective amount of the antibody, antibody fragment or scFv described above, wherein the species may be murine, camelid, human or humanized.

[0068] In another aspect, the invention is directed to a polypeptide comprising at least two different scFv sequences, wherein one of the scFv sequences is a sequence that binds to extracellular domain of MUC1 isoform or cleavage product that is devoid of the tandem repeat domains. The polypeptide may bind to

• • [0069] (i) PSMGFR region of MUC1;
  • [0070] (ii) PSMGFR peptide;
  • [0071] (iii) a peptide having amino acid sequence of QFNQYKTEAASRYNLTISDVSVSDVPFPFSAQSGA (N-10) (SEQ ID NO:3)
  • [0072] (iv) a peptide having amino acid sequence of
  • [0073] ASRYNLTISDVSVSDVPFPFSAQSGA (N-19) (SEQ ID NO:4)
  • [0074] (v) a peptide having amino acid sequence of
  • [0075] NLTISDVSVSDVPFPFSAQSGA (N-23) (SEQ ID NO:5)
  • [0076] (vi) a peptide having amino acid sequence of
  • [0077] ISDVSVSDVPFPFSAQSGA (N-26) (SEQ ID NO:6)
  • [0078] (vii) a peptide having amino acid sequence of
  • [0079] SVSDVPFPFSAQSGA (N-30) (SEQ ID NO:7)
  • [0080] (viii) a peptide having amino acid sequence of
  • [0081] QFNQYKTEAASRYNLTISDVSVSDVPFPFS (N-10/C-5) (SEQ ID NO:8)
  • [0082] (ix) a peptide having amino acid sequence of
  • [0083] ASRYNLTISDVSVSDVPFPFS (N-19/C-5) (SEQ ID NO:9)
  • [0084] (x) a peptide having amino acid sequence of
  • [0085] FPFSAQSGA (N-36) (SEQ ID NO:10)

[0086] The polypeptide may bind to a receptor on an immune cell, such as T cell, and in particular, CD3 on T-cell.

[0087] In another aspect, the invention is directed to a method of detecting presence of a cell that expresses MUC1* aberrantly, comprising contacting a sample of cells with the scFv-Fc described above and detecting for the presence of the binding of scFv-Fc to the cell. The cell may be cancer cell.

[0088] In another aspect, the invention is directed to a method for testing a subject's cancer for suitability of treatment with a composition comprising antibodies of the invention, which may be murine, camelid, human or humanized, or fragments thereof, or portions of the variable regions of antibodies MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11, comprising the steps of contacting a bodily specimen from the patient, in vitro, ex-vivo, or in vivo, with the antibody and determining that the patient exhibits aberrant expression of MUC1* compared to normal tissue or specimen. The antibody used in these diagnostics may be conjugated to an imaging agent.

[0089] In another aspect, the invention is directed to a method of treating a subject suffering from a disease comprising, exposing T cells from the subject, or from a donor, to MUC1* peptides wherein through various rounds of maturation, T cells develop MUC1* specific receptors, creating adapted T cells, and expanding and administering the adapted T cells to the donor patient who is diagnosed with, suspected of having, or is at risk of developing a MUC1* positive cancer. The MUC1* peptide is chosen from among the group:

• • [0090] (i) PSMGFR region of MUC1;
  • [0091] (ii) PSMGFR peptide;
  • [0092] (iii) a peptide having amino acid sequence of QFNQYKTEAASRYNLTISDVSVSDVPFPFSAQSGA (N-10) (SEQ ID NO: 3)
  • [0093] (iv) a peptide having amino acid sequence of
  • [0094] ASRYNLTISDVSVSDVPFPFSAQSGA (N-19) (SEQ ID NO: 4)
  • [0095] (v) a peptide having amino acid sequence of
  • [0096] NLTISDVSVSDVPFPFSAQSGA (N-23) (SEQ ID NO: 5)
  • [0097] (vi) a peptide having amino acid sequence of
  • [0098] ISDVSVSDVPFPFSAQSGA (N-26) (SEQ ID NO: 6)
  • [0099] (vii) a peptide having amino acid sequence of
  • [0100] SVSDVPFPFSAQSGA (N-30) (SEQ ID NO: 7)
  • [0101] (viii) a peptide having amino acid sequence of
  • [0102] QFNQYKTEAASRYNLTISDVSVSDVPFPFS (N-10/C-5) (SEQ ID NO: 8)
  • [0103] (ix) a peptide having amino acid sequence of
  • [0104] ASRYNLTISDVSVSDVPFPFS (N-19/C-5) (SEQ ID NO: 9)
  • [0105] (x) a peptide having amino acid sequence of
  • [0106] FPFSAQSGA (N-36) (SEQ ID NO: 10)

[0107] In one aspect of the invention, the antibody that is administered to a patient for the treatment or prevention of a MUC1 or MUC1* positive cancer is selected for its ability to bind to the N-10 peptide of the PSMGFR. The antibody can be administered alone, as a monovalent antibody, as an scFv, or a fragment of the antibody can be incorporated into a CAR, a BiTE or an ADC.

[0108] In one aspect of the invention, the antibody that is administered to a patient for the treatment or prevention of a MUC1 or MUC1* positive cancer is selected for its inability to recognize a linear epitope of MUC1 or MUC1*. The antibody can be administered alone, as a monovalent antibody, as an scFv, or a fragment of the antibody can be incorporated into a CAR, a BiTE or an ADC.

[0109] In one aspect of the invention, the antibody that is administered to a patient for the treatment or prevention of a MUC1 or MUC1* positive cancer is selected for its ability to recognize the MUC1 transmembrane cleavage product after it has been cleaved by MMP9. The antibody can be administered alone, as a monovalent antibody, as an scFv, or a fragment of the antibody can be incorporated into a CAR, a BiTE or an ADC.

[0110] In one aspect of the invention, the antibody that is administered to a patient for the treatment or prevention of a MUC1 or MUC1* positive cancer is selected for its ability to competitively inhibit the binding of NME7_AB or NME7-X1 to the extra cellular domain of a MUC1 that is devoid of tandem repeats. The antibody can be administered alone, as a monovalent antibody, as an scFv, or a fragment of the antibody can be incorporated into a CAR, a BiTE or an ADC.

[0111] In another aspect, the invention is directed to a method of treating cancer in a patient comprising administering to the patient the immune cell of any of the above, in combination with a checkpoint inhibitor.

[0112] In the method above, any of the antibodies, or variable regions thereof, set forth in the following may be used: MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11.

[0113] In the method above, any of the variable regions set forth in the following may be used:

• • [0114] (i) an anti-MUC1* extracellular domain antibody or anti-N-10 antibody comprised of sequences of a humanized MN-E6 represented by humanized IgG2 heavy chain, or humanized IgG1 heavy chain, paired with humanized Kappa light chain, or humanized Lambda light chain;
  • [0115] (ii) an antibody of (i), wherein the humanized IgG2 heavy chain is SEQ ID NOS: 53, humanized IgG1 heavy chain is SEQ ID NO:57, humanized Kappa light chain is SEQ ID NO: 108, and humanized Lambda light chain is SEQ ID NO:112, or a sequence having 90%, 95% or 98% sequence identity thereof;
  • [0116] (iii) an anti-MUC1* extracellular domain antibody or anti-N-10 antibody comprised of sequences of a humanized MN-C2 represented by humanized IgG1 heavy chain, humanized IgG2 heavy chain, paired with humanized Lambda light chain, and humanized Kappa light chain;
  • [0117] (iv) an antibody of (iii), wherein the humanized IgG1 heavy chain MN-C2 (SEQ ID NOS: 159) or IgG2 heavy chain (SEQ ID NOS: 164) paired with Lambda light chain (SEQ ID NO: 219) or Kappa light chain (SEQ ID NO:213), or a sequence having 90%, 95% or 98% sequence identity thereof,

[0118] In the method above, in the CAR, the extracellular domain may be comprised of humanized single chain antibody fragments of MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11. The extracellular domain may be comprised of humanized single chain antibody fragments of an MN-E6 scFv set forth as SEQ ID NOS: 233, 235, or 237), MN-C2 scFv (SEQ ID NOS: 239, 241, or 243). In the CAR, the cytoplasmic tail may be comprised of one or more of signaling sequence motifs CD3-zeta, CD27, CD28, 4-1BB, OX40, CD30, CD40, ICAm-1, LFA-1, ICOS, CD2, CD5, or CD7.

[0119] The method above may include at least two CARs with different extracellular domain units transfected into the same cell. One of the extracellular domain recognition units may bind to MUC1* extracellular domain. One of the extracellular domain recognition units may bind to PD-1. One of the extracellular domain recognition units may be an antibody fragment and the other may be a peptide or an anti-MUC1* antibody fragment.

[0120] The method may include an immune cell transfected or transduced with a plasmid encoding a CAR and a plasmid encoding a non-CAR species that is expressed from an inducible promoter. The non-CAR species may be expressed from an inducible promoter that is activated by elements of an activated immune cell. The non-CAR species may be expressed from an NFAT inducible promoter. The NFAT may be NFATc1, NFATc3 or NFATc2. The cleavage enzyme may be MMP2, MMP3, MMP9, MMP13, MMP14, MMP16, ADAM10, ADAM17, or ADAM28, or a catalytically active fragment thereof. The non-CAR species may be a cytokine. The cytokine may be IL-7, IL-12, IL-15 or IL-18.

[0121] The present invention is directed to an antibody, or fragment thereof, for the diagnosis, treatment or prevention of cancers wherein the antibody specifically binds to the PSMGFR peptide (SEQ ID NO: 2) or a fragment thereof of the peptide.

[0122] The antibody binds to the N-10 peptide (SEQ ID NO:3), N-19 peptide (SEQ ID NO:4), N-23 peptide (SEQ ID NO:5), N-26 peptide (SEQ ID NO:6), N-30 peptide (SEQ ID NO:7), N-10/C-5 peptide (SEQ ID NO:8), N-19/C-5 peptide (SEQ ID NO:9), or C-5 peptide (SEQ ID NO:825).

[0123] The antibody interacts with a peptide comprising conformational epitope SVSDV (SEQ ID NO: 1751) and FPSA (SEQ ID NO:1791) within N-26 sequence ISDVSVSDVPFPFSAQSGA (SEQ ID NO: 6), wherein mutation or deletion of FPFS (SEQ ID NO:1747) destroys binding of the antibody or fragment thereof to the N-26 peptide.

[0124] The antibody interacts with a peptide comprising conformational epitope ASRYNLT (SEQ ID NO: 1745), SVSDV (SEQ ID NO:1751), and FPSA (SEQ ID NO:1791) within the N-19 sequence ASRYNLT ISDVSVSDVPFPFSAQSGA (SEQ ID NO:4), wherein mutation or deletion of ASRYNLT (SEQ ID NO: 1745) destroys binding of the antibody or fragment thereof to the N-26 peptide.

[0125] The antibody does not bind to the C-10 peptide (SEQ ID NO:825).

[0126] The antibody binds to the N-10 peptide (SEQ ID NO:3), but not to the C-10 peptide (SEQ ID NO: 825).

[0127] The antibody inhibits interaction between NME7_AB and MUC1*.

[0128] The antibody inhibits interaction between NME7_AB and PSMGFR peptide (SEQ ID NO:2).

[0129] The antibody inhibits interaction between NME7_AB and N-10 peptide (SEQ ID NO:3), N-19 peptide (SEQ ID NO:4), N-23 peptide (SEQ ID NO:5), N-26 peptide (SEQ ID NO:6), N-30 peptide (SEQ ID NO: 7), N-10/C-5 peptide (SEQ ID NO:8), N-19/C-5 peptide (SEQ ID NO:9), or C-5 peptide (SEQ ID NO: 825).

[0130] The antibody recognizes a MUC1 transmembrane enzymatic cleavage product.

[0131] In the above, the cleavage enzyme is MMP14 or MMP9 or a catalytically active fragment thereof of the enzyme.

[0132] The antibody binds to PSMGFR (SEQ ID NO:2) or fragment thereof in which presence of an amino acid sequence within PSMGFR (SEQ ID NO:2) induces binding of the antibody to the PSMGFR.

[0133] The amino acid sequence of the binding conformationally inducing peptide is present in N-10 peptide (SEQ ID NO:3).

[0134] The antibody does not bind to a linear form of the binding conformationally inducing peptide sequence wherein the linear form of the peptide is a denatured form.

[0135] The binding conformationally inducing peptide sequence is in the N-26 peptide sequence ISDVSVSDVPFPFSAQSGA (SEQ ID NO:6), wherein mutation or deletion of FPFS (SEQ ID NO:1747) destroys binding of the antibody or fragment thereof to the N-26 peptide.

[0136] The binding conformationally inducing peptide sequence is located within the N-19 sequence ASRYNLTISDVSVSDVPFPFSAQSGA (SEQ ID NO:4), wherein mutation or deletion of ASRYNLT (SEQ ID NO: 1745) destroys binding of the antibody or fragment thereof to the N-19 peptide.

[0137] The binding inducing peptide sequence may be located within the N-26 sequence ISDVSVSDVPFPFSAQSGA (SEQ ID NO:6), wherein mutation or deletion within FPFS (SEQ ID NO: 1747) destroys binding of the antibody or fragment thereof to PSMGFR.

[0138] The antibodies may have a consensus sequence.

• • [0139] heavy chain CDR1 comprises consensus sequence at least 90% identical to sequence: F or I at position 1, T at position 2, F at position 3, S at position 4, T, G, or R at position 5, Y at position 6, A, G or T at position 7, M at position 8 and S at position 9;
  • [0140] heavy Chain CDR2 comprises consensus sequence at least 90% identical to sequence: T at position 1, I or S at position 2, I or S at position 3, G or R at position 5, G or A at position 6, T or I at position 9, Y at position 10, Y at position 11, P or S at position 12 and DSVKG (SEQ ID NO: 1793) for positions 13-17;
  • [0141] heavy chain CDR3 comprises consensus sequence at least 90% identical to sequence: _G, L, or N at position 2, G or T at position 4, Y at position 7, D or E at position 12, A at position 14, and Y at position 15;
  • [0142] light chain CDR1 comprises consensus sequence at least 90% identical to sequence: K or R at position 1, A or S at position 2, S at position 3, K or Q at position 4, S at position 5, L or V at position 6, L at position 7, T or S at position 10, Y at position 15, and I, L or M at position 16;
  • [0143] light Chain CDR2 comprises consensus sequence at least 90% identical to sequence: L or W, or S at position 1, A or T at position 2, S at position 3, N or T at position 4, L or R at position 5, E or A at position 6, and S at position 7; and
  • [0144] light chain CDR3 comprises consensus sequence at least 90% identical to sequence: Q at position 1, H or Q at position 2, S, Q or R at position 3, R, S or Y at position 4, E, L, or S at position 5, L or S at position 6, P or S at position 7, F or L at position 8 and T at position 9.

[0145] An antibody binding conformationally inducing peptide is within the N-26 sequence ISDVSVSDVPFPFSAQSGA (SEQ ID NO:6), wherein mutation or deletion within FPFS (SEQ ID NO: 1747), SVSDV (SEQ ID NO:1751), or ASRYNLT (SEQ ID NO:1745) destroys binding of the antibody or fragment thereof to PSMGFR.

[0146] The antibody may have a further consensus sequence,

• • [0147] wherein
    • [0148] heavy chain CDR1 comprises consensus sequence at least 90% identical to sequence: F or I at position 1, T or A at position 2, F at position 3, S at position 4, T, G, or R at position 5, Y or F at position 6, A, G or T at position 7, M at position 8 and S at position 9;
    • [0149] heavy Chain CDR2 comprises consensus sequence at least 90% identical to sequence: T or A at position 1, I or S at position 2, I or S at position 3, N, S, T or G at position 4, G or R at position 5, G or A at position 6, G, T, or D at position 7, Y, K, H or S at position 8, T or I at position 9, Y or F at position 10, Y at position 11, P or S at position 12 and D at position 13, S or T at position 14, V or L at position 15 and KG for positions 16-17;
    • [0150] heavy chain CDR3 comprises consensus sequence at least 90% identical to sequence: G, L, or N at position 2, G, T, or Y at position 3, G or T at position 4, Y at position 7, Y, A, or G at position 10, M, D or F at position 11, D or E at position 12 and AY at position 14-15;
    • [0151] light chain CDR1 comprises consensus sequence _ at least 90% identical to sequence: K or R at position 1, A or S at position 2, S or R at position 3, S, Y, I or V at position 8, T or S at position 10, G, S, D, or Q at position 12, V, Y, K or N at position 13, N, S, or T at position 14, Y or F at position 15, and I, L or M at position 16;
    • [0152] light Chain CDR2 comprises consensus sequence at least 90% identical to sequence: A, T or V at position 2, S at position 3, N, T, or K at position 4, L or R at position 5, E, A, F or D at position 6, and S at position 7; and
    • [0153] light chain CDR3 comprises consensus sequence at least 90% identical to sequence: Q, F or W at position 1, H or Q at position 2, R, S, T, Y or N at position 4, E, L, S or H at position 5, L, S, V, D or Y at position 6, P or S at position 7, and T at position 9.

[0154] The antibody above which may be MNC2, having

• • [0155] heavy chain CDR1 comprises consensus sequence FTFSGYAMS (SEQ ID NO: 1794);
  • [0156] heavy Chain CDR2 comprises consensus sequence TISSGGTYIYYPDSVKG (SEQ ID NO: 127);
  • [0157] heavy chain CDR3 comprises consensus sequence -LGGDNYYEYFDV-- (SEQ ID NO: 131);
  • [0158] light chain CDR1 comprises consensus sequence RASKS--VSTSGYSYMH (SEQ ID NO: 173);
  • [0159] light Chain CDR2 comprises consensus sequence LASNLES (SEQ ID NO: 177); and
  • [0160] light chain CDR3 comprises consensus sequence QHSRELPFT (SEQ ID NO: 181).

[0161] MNE6, having

• • [0162] heavy chain CDR1 comprises consensus sequence FTFSRYGMS (SEQ ID NO: 1795);
  • [0163] heavy Chain CDR2 comprises consensus sequence TISGGGTYIYYPDSVKG (SEQ ID NO: 21);
  • [0164] heavy chain CDR3 comprises consensus sequence DNYGRNYDYGMDY-- (SEQ ID NO: 25);
  • [0165] light chain CDR1 comprises consensus sequence -------SATSSVSYIH (SEQ ID NO: 70);
  • [0166] light Chain CDR2 comprises consensus sequence STSNLAS (SEQ ID NO: 74); and
  • [0167] light chain CDR3 comprises consensus sequence QQRSSSPFT (SEQ ID NO: 78).

[0168] B2, having

• • [0169] heavy chain CDR1 comprises consensus sequence FAFSTFAMS (SEQ ID NO: 1796);
  • [0170] heavy Chain CDR2 comprises consensus sequence AISNGGGYTYYPDTLKG (SEQ ID NO: 1439);
  • [0171] heavy chain CDR3 comprises consensus sequence ----RYYDLYFDL-- (SEQ ID NO: 1443);
  • [0172] light chain CDR1 comprises consensus sequence RSSQNIV-HSNGNTYLE (SEQ ID NO: 1449);
  • [0173] light Chain CDR2 comprises consensus sequence KVSNRFS (SEQ ID NO: 467); and
  • [0174] light chain CDR3 comprises consensus sequence FQDSHVPLT (SEQ ID NO: 1381).

[0175] B7, having

• • [0176] heavy chain CDR1 comprises consensus sequence FTFSRYGMS (SEQ ID NO: 1795);
  • [0177] heavy Chain CDR2 comprises consensus sequence TISSGGTYIYYPDSVKG (SEQ ID NO: 127);
  • [0178] heavy chain CDR3 comprises consensus sequence DNYGSSYDYAMDY--; (SEQ ID NO: 1471) light chain CDR1 comprises consensus sequence RSSQTIV-HSNGNTYLE (SEQ ID NO: 463);
  • [0179] light Chain CDR2 comprises consensus sequence KVSNRFS (SEQ ID NO: 467); and
  • [0180] light chain CDR3 comprises consensus sequence FQDSHVPLT (SEQ ID NO: 1381).

[0181] B9, having

• • [0182] heavy chain CDR1 comprises consensus sequence FTFSRYGMS (SEQ ID NO: 1795);
  • [0183] heavy Chain CDR2 comprises consensus sequence TISSGGTYIYYPDSVKG (SEQ ID NO: 127);
  • [0184] heavy chain CDR3 comprises consensus sequence DNYGSSYDYAMDY-- (SEQ ID NO: 1471);
  • [0185] light chain CDR1 comprises consensus sequence -------SASSSVSYMH (SEQ ID NO: 1561);
  • [0186] light Chain CDR2 comprises consensus sequence TTSNLAS (SEQ ID NO: 1565); and
  • [0187] light chain CDR3 comprises consensus sequence QQRSSYPF—(SEQ ID NO: 1569).

[0188] 8C7F3, having

• • [0189] heavy chain CDR1 comprises consensus sequence FTFSTYAMS (SEQ ID NO: 1797);
  • [0190] heavy Chain CDR2 comprises consensus sequence AISNGGGYTYYPDSLKG (SEQ ID NO: 1363);
  • [0191] heavy chain CDR3 comprises consensus sequence ----RYYDHYFDY-- (SEQ ID NO: 1367);
  • [0192] light chain CDR1 comprises consensus sequence --RASESVATYGNNFMQ (SEQ ID NO: 1431);
  • [0193] light Chain CDR2 comprises consensus sequence LASTLDS (SEQ ID NO: 1509); and
  • [0194] light chain CDR3 comprises consensus sequence QQNNEDPPT (SEQ ID NO: 1513).

[0195] H11, having

• • [0196] heavy chain CDR1 comprises consensus sequence FAFSTFAMS (SEQ ID NO: 1796);
  • [0197] heavy Chain CDR2 comprises consensus sequence AISNGGGYTYYPDTLKG (SEQ ID NO: 1439);
  • [0198] heavy chain CDR3 comprises consensus sequence ----RYYDLYFDL-- (SEQ ID NO: 1443);
  • [0199] light chain CDR1 comprises consensus sequence RSSQNIV-HSNGNTYLE (SEQ ID NO: 1449);
  • [0200] light Chain CDR2 comprises consensus sequence KVSNRFS (SEQ ID NO: 467); and
  • [0201] light chain CDR3 comprises consensus sequence FQDSHVPLT (SEQ ID NO: 1381).

[0202] B12, having

• • [0203] heavy chain CDR1 comprises consensus sequence SYGVH (SEQ ID NO: 1417);
  • [0204] heavy Chain CDR2 comprises consensus sequence VIWPGGSTNYNSTLMSRM (SEQ ID NO: 1421);
  • [0205] heavy chain CDR3 comprises consensus sequence DRTPRVGAWFAY (SEQ ID NO: 1425); and
  • [0206] light chain CDR1 comprises consensus sequence RASESVATYGNNFMQ (SEQ ID NO: 1431);
  • [0207] light Chain CDR2 comprises consensus sequence LASTLDS (SEQ ID NO: 1509); and
  • [0208] light chain CDR3 comprises consensus sequence QQNNEDPPT (SEQ ID NO: 1513). 20A10, having
  • [0209] heavy chain CDR1 comprises consensus sequence FTFSTYAMS (SEQ ID NO: 1797);
  • [0210] heavy Chain CDR2 comprises consensus sequence -SIGRAGSTYYSDSVKG (SEQ ID NO: 997);
  • [0211] heavy chain CDR3 comprises consensus sequence ---GPIYNDYDEFAY (SEQ ID NO: 1001);
  • [0212] light chain CDR1 comprises consensus sequence KSSQSVLYSSNQKNYLA (SEQ ID NO: 1009);
  • [0213] light Chain CDR2 comprises consensus sequence WASTRES (SEQ ID NO: 1013); and
  • [0214] light chain CDR3 comprises consensus sequence HQYLSSLT (SEQ ID NO: 1017).

[0215] 3C2B1, having

• • [0216] heavy chain CDR1 comprises consensus sequence ITFSTYTMS (SEQ ID NO: 1798);
  • [0217] heavy Chain CDR2 comprises consensus sequence TISTGGDKTYYSDSVKG (SEQ ID NO: 1393);
  • [0218] heavy chain CDR3 comprises consensus sequence -GTTAMYYYAMDY (SEQ ID NO: 1397);
  • [0219] light chain CDR1 comprises consensus sequence RASKS---ISTSDYNYIH (SEQ ID NO: 1803);
  • [0220] light Chain CDR2 comprises consensus sequence LASNLES (SEQ ID NO: 177); and
  • [0221] light chain CDR3 comprises consensus sequence QHSRELPLT (SEQ ID NO: 1411).

[0222] In another aspect, the invention is directed to an antibody, or fragment thereof, for the diagnosis, treatment or prevention of cancers that requires presence of antibody binding conformationally inducing peptide ASRYNLT (SEQ ID NO:1745) of PSMGFR (SEQ ID NO:2). The antibody may be 25E6, having

• • [0223] heavy chain CDR1 comprises consensus sequence FTFSSYGMS (SEQ ID NO: 1799);
  • [0224] heavy Chain CDR2 comprises consensus sequence TISNGGRHTFYPDSVKG (SEQ ID NO: 1029);
  • [0225] heavy chain CDR3 comprises consensus sequence QTGTEGWFAY (SEQ ID NO: 1033);
  • [0226] light chain CDR1 comprises consensus sequence KSSQSLLDSDGKTYLN (SEQ ID NO: 1041);
  • [0227] light Chain CDR2 comprises consensus sequence LVSKLDS (SEQ ID NO: 981); and
  • [0228] light chain CDR3 comprises consensus sequence WQGTHFPQT (SEQ ID NO: 1049).

[0229] In another aspect, the invention is directed to an antibody, or fragment thereof, for the diagnosis, treatment or prevention of cancers that requires presence of antibody binding conformationally inducing peptide SVSDV (SEQ ID NO: 1751) of PSMGFR (SEQ ID NO:2). The antibody may be 5C6F3, having

• • [0230] heavy chain CDR1 comprises consensus sequence FTFSTYAMS (SEQ ID NO: 1797);
  • [0231] heavy Chain CDR2 comprises consensus sequence AISNGGGYTYYPDSLKG (SEQ ID NO: 1363);
  • [0232] heavy chain CDR3 comprises consensus sequence RYYDHYFDY (SEQ ID NO: 1367);
  • [0233] light chain CDR1 comprises consensus sequence RSSQTIVHSNGNTYLE (SEQ ID NO: 463);
  • [0234] light Chain CDR2 comprises consensus sequence KVSNRFS (SEQ ID NO: 467); and
  • [0235] light chain CDR3 comprises consensus sequence FQDSHVPLT (SEQ ID NO: 1381).

[0236] The antibody or fragment thereof according all of the above may be murine, camelid, human or humanized. The antibody fragment may be scFv or scFv-Fc, which variable regions thereof may be murine, camelid, human or humanized.

[0237] In another aspect, the invention is directed to a chimeric antigen receptor (CAR) comprising the antibody fragments of above, and may further comprise mutations in the co-stimulatory domain or CD3-zeta signaling domain. Tyrosines may be mutated in CD28 or 4-1BB. CD3-zeta may contain 1XX mutations.

[0238] In another aspect, the invention is directed to an immune cell comprising the CAR of above. Immune cell may be T cell, NK cell, dendritic cell, or mast cell.

[0239] In another aspect, the invention is directed to a cell composition expressed in a cell comprising a CARs of above, and second entity having a biological recognition unit that has a specificity that is different from that of the CAR. The second entity may bind PD-1, PDL-1, or other checkpoint inhibitor, or NME7, or a cytokine such as IL-12 or IL-18, or c-Jun.

[0240] In yet another aspect, the invention is directed to an immune cell engineered to express a nucleic encoding a CAR of above and a nucleic acid encoding a second entity as in any of the claims above wherein the second entity expressed from an inducible promoter. The second entity may be expressed from an inducible promoter that is activated by elements of an activated immune cell. The second entity may be expressed from an NFAT inducible promoter. NFAT may be NFATc1, NFATc3 or NFATc2. The second entity may be a cytokine such as IL-7, IL-15, or IL-18. The nucleic acids encoding the second entity may be inserted into a Foxp3 promoter or enhancer region, wherein the cytokine is IL-18. The cytokine may be expressed from an NFAT inducible promoter.

[0241] In another aspect, the invention is directed to a BiTE construct comprising the antibody fragment of above.

[0242] In yet another aspect, the invention is directed to an antibody drug conjugate (ADC) comprising the antibody or antibody fragment of above.

[0243] The present invention is directed to an antibody or fragment thereof that specifically binds to PSMGFR (SEQ ID NO:2) and N-10 (SEQ ID NO:3); and

• • [0244] does not bind to full-length MUC1;
  • [0245] does not bind to C-10 (SEQ ID NO:825);
  • [0246] competitively inhibits binding of NME1 or NME7_AB to MUC1* extra cellular domain or a PSMGFR peptide;
  • [0247] recognizes a MUC1* generated by cleavage by a cleavage enzyme;
  • [0248] recognizes a conformational epitope and not a linear epitope; or
  • [0249] is cancer selective by immunohistochemistry on tissues.

[0250] Four of the criteria (i)-(vi) may be satisfied. Five of the criteria (i)-(vi) may be satisfied. Six of the criteria (i)-(vi) may be satisfied. At least criteria (vi) may be satisfied. Cleavage enzyme may be MMP-9.

[0251] In all of the above, the cancer may be breast cancer, pancreatic cancer, ovarian cancer, lung cancer, colon cancer, gastric cancer or esophageal cancer.

[0252] The present invention is also directed to a method of diagnosing, treating or preventing cancer by administering the antibodies and fragments disclosed herein to a cancer patient in need thereof that has been identified as expressing MUC1 aberrantly and expressing truncated MUC1, such as MUC1*.

[0253] These and other objects of the invention will be more fully understood from the following description of the invention, the referenced drawings attached hereto and the claims appended hereto.

[0479] In the present application, “a” and “an” are used to refer to both single and a plurality of objects.

[0480] As used herein, occasionally, in short hand, a polypeptide is indicated as being “transduced or transfected” into a cell. In these occurrences, it is understood that the nucleic acid encoding the polypeptide sequence is transduced or transfected into the cell, as it is an impossibility that a polypeptide could be transduced or transfected into a cell.

[0481] As used herein, occasionally when referring to number of cells injected into an animal or otherwise contextually wherein the number of cells is referred to, “M” refers to millions, and “K” refers to thousands.

[0482] As used herein, interchangeable designations for various monoclonal antibodies are used, such as, “MN-C2”, which is interchangeable with “C2”, “Min-C2” and “MNC2”; “MN-E6”, which is interchangeable with “E6”, “Min-E6” and “MNE6”; “MN-C3”, which is interchangeable with “C3”, “Min-C3” and “MNC3”; and “MN-C8”, which is interchangeable with “C8”, “Min-C8” and “MNC8”. The monoclonal antibodies provided herein follow the same convention.

[0483] As used herein, “h” or “hu” placed before an antibody construct is short-hand for humanized.

[0484] As used herein, the term “antibody-like” means a molecule that may be engineered such that it contains portions of antibodies but is not an antibody that would naturally occur in nature. Examples include but are not limited to CAR (chimeric antigen receptor) T cell technology and the Ylanthia® technology. The CAR technology uses an antibody epitope fused to a portion of a T cell so that the body's immune system is directed to attack a specific target protein or cell. The Ylanthia® technology consists of an “antibody-like” library that is a collection of synthetic human Fabs that are then screened for binding to peptide epitopes from target proteins. The selected Fab regions can then be engineered into a scaffold or framework so that they resemble antibodies.

[0485] As used herein, “PSMGFR” is abbreviation for Primary Sequence of the MUC1 Growth Factor Receptor which is identified by SEQ ID NO:2, and thus is not to be confused with a six amino acid sequence. “PSMGFR peptide” or “PSMGFR region” refers to a peptide or region that incorporates the Primary Sequence of the MUC1 Growth Factor Receptor (SEQ ID NO: 2).

[0486] As used herein, the “MUC1*” extra cellular domain is defined primarily by the PSMGFR sequence (GTINVHDVETQFNQYKTEAASRYNLTISDVSVSDVPFPFSAQSGA (SEQ ID NO:2)). Because the exact site of MUC1 cleavage depends on the enzyme that clips it, and that the cleavage enzyme varies depending on cell type, tissue type or the time in the evolution of the cell, the exact sequence of the MUC1* extra cellular domain may vary at the N-terminus.

[0487] Other clipped amino acid sequences may include SNIKFRPGSVVVQLTLAFREGTINVHDVETQFNQYKTEAASRY (SEQ ID NO:620); or SVVVQLTLAFREGTINVHDVETQFNQYKTEAASRY (SEQ ID NO:621).

[0488] As used herein, the term “PSMGFR” is an acronym for Primary Sequence of MUC1 Growth Factor Receptor as set forth as GTINVHDVETQFNQYKTEAASRYNLTISDVSVSDVPFPFSAQSGA (SEQ ID NO:2). In this regard, the “N-number” as in “N-10 PSMGFR” or simply “N-10”, “N-15 PSMGFR” or simply “N-15”, or “N-20 PSMGFR” or simply “N-20” refers to the number of amino acid residues that have been deleted at the N-terminal end of PSMGFR. Likewise “C-number” as in “C-10 PSMGFR” or simply “C-10”, “C-15 PSMGFR” or simply “C-15”, or “C-20 PSMGFR” or simply “C-20” refers to the number of amino acid residues that have been deleted at the C-terminal end of PSMGFR. A mixture of deletions and additions is also possible. For instance, N+20/C-27 refers to a peptide fragment of wild-type MUC1 in which 20 amino acids are added to the PSMGFR at the N-terminus and 27 amino acids are deleted from the C-terminus.

[0489] As used herein, the “extracellular domain of MUC1*” refers to the extracellular portion of a MUC1 protein that is devoid of the tandem repeat domain. In most cases, MUC1* is a cleavage product wherein the MUC1* portion consists of a short extracellular domain devoid of tandem repeats, a transmembrane domain and a cytoplasmic tail. The precise location of cleavage of MUC1 is not known perhaps because it appears that it can be cleaved by more than one enzyme. The extracellular domain of MUC1* will include most of the PSMGFR sequence but may have an additional 10-20 N-terminal amino acids.

[0490] As used herein “sequence identity” means homology in sequence of a particular polypeptide or nucleic acid to a reference sequence of nucleic acid or amino acid such that the function of the homologous peptide is the same as the reference peptide or nucleic acid. Such homology can be so close with the reference peptide such that at times the two sequences may be 90%, 95% or 98% identical yet possess the same function in binding or other biological activities.

[0491] As used herein, “MUC1 positive” cell refers to a cell that expresses a gene for MUC1, MUC1-Y or MUC1-Z or other MUC1 variant.

[0492] As used herein, “MUC1 negative” cell refers to a cell that does not express a gene for MUC1.

[0493] As used herein, “MUC1* positive” cell refers to a cell that expresses a gene for MUC1, wherein that gene's expressed protein is a transmembrane protein that is devoid of tandem repeats, which may be a consequence of post-translational modification, cleavage, alternative splicing, or transfecting or transducing a cell with a MUC1 protein that is devoid of tandem repeats.

[0494] As used herein, “MUC1* negative” cell refers to a cell that may or may not express a gene for MUC1 but does not express a MUC1 transmembrane protein that is devoid of tandem repeats.

[0495] As used herein, “MUC1 positive” cancer cell refers to a cancer cell that overexpresses the gene for MUC1, expresses MUC1 in an aberrant pattern, wherein its expression is not restricted to the apical border and/or expresses a MUC1 that is devoid of tandem repeats.

[0496] As used herein, “MUC1 negative” cancer cell refers to a cancer cell that may or may not express a gene for MUC1 but does not overexpress MUC1 or does not overexpress a MUC1 transmembrane protein that is devoid of tandem repeats.

[0497] As used herein, “MUC1* positive” cancer cell refers to a cancer cell that overexpresses a MUC1 transmembrane protein that is devoid of tandem repeats.

[0498] As used herein, “MUC1* negative” cancer cell refers to a cancer cell that may or may not express a gene for MUC1 but does not overexpress a MUC1 transmembrane protein that is devoid of tandem repeats.

[0499] As used herein “conformational epitope” refers to a peptide sequence that is required to be present in a specific three-dimensional structure or conformation for an antibody to bind. However the antibody binds when the peptide sequence is in the three-dimensional structure or conformation and is not bound when linear. A common technique for determining whether an antibody binds to a linear stretch or a conformational epitope is to use the antibody to probe a denaturing Western blot. Traveling through a denaturing gel linearizes proteins and peptides. Antibodies that do not work in a denaturing Western but do recognize the native target, for example expressed on an intact cell, are determined to recognize a conformational epitope. As used herein, the antibody may or may not actually bind to the “conformational epitope”, however the presence of the “conformational epitope” sequence is required to render a three dimensional structure so that the MUC1* region on cancer cells is able to be bound by the antibody that is specific for cancer treatment. Thus, the conformational epitope is an amino acid sequence that induces the binding of the antibody to the MUC1* region on cancer cells. Thus, a term “conformational inducing peptide sequence” may be used, which indicates that a peptide sequence is present within a larger peptide not as a binding site but that induces binding of an antibody to the larger peptide by causing a three-dimensional structure to form that facilitates the binding of the antibody to the larger peptide.

MUC1* Antibodies (Anti-PSMGFR) for Treatment or Prevention of Cancers

[0500] We discovered that a cleaved form of the MUC1 (SEQ ID NO:1) transmembrane protein is a growth factor receptor that drives the growth of over 75% of all human solid tumor cancers. The cleaved form of MUC1, which we called MUC1* (pronounced muk 1 star), is a powerful growth factor receptor. Enzymatic cleavage releases the bulk of the MUC1 extracellular domain. It is the remaining portion comprising a truncated extracellular domain, transmembrane domain and cytoplasmic tail that is called MUC1*. Cleavage and release of the bulk of the extracellular domain of MUC1 unmasks a binding site for activating ligands dimeric NME1, NME6, NME8, NME7_AB, NME7-X1 or NME7. Cell growth assays show that it is ligand-induced dimerization of the MUC1* extracellular domain that promotes growth (FIG. 1A-1D). MUC1* positive cells treated with either bivalent ‘bv’ anti-MUC1* antibody, monovalent ‘mv’ or Fab, NM23-H1 dimers or NME7-AB. Bivalent anti-MUC1* antibodies stimulate growth of cancer cells whereas the monovalent Fab inhibits growth. Classic bell-shaped curve indicates ligand induced dimerization stimulates growth. Dimeric NM23-H1, aka NME1, stimulates growth of MUC1* positive cancer cells but siRNA to suppress MUC1 expression eliminate its effect (FIG. 1C). NME7-AB also stimulates the growth of MUC1* positive cells (FIG. 1D).

[0501] MUC1* is an excellent target for cancer drugs as it is aberrantly expressed on over 75% of all cancers and is likely overexpressed on an even higher percentage of metastatic cancers. After MUC1 cleavage, most of its extracellular domain is shed from the cell surface. The remaining portion has a truncated extracellular domain that at least comprises the primary growth factor receptor sequence, PSMGFR (SEQ ID NO:2). Antibodies that bind to the PSMGFR sequence and especially those that competitively inhibit the binding of activating ligands such as NME proteins, including NME1, NME6, NME8, NME7_AB, NME7-X1 and NME7, are ideal therapeutics and can be used to treat or prevent MUC1 positive or MUC1* positive cancers, as stand-alone antibodies, antibody fragments or variable region fragments thereof incorporated into bispecific antibodies, or chimeric antigen receptors also called CARs, which are then transfected or transduced into immune cells, then administered to a patient.

[0502] Therapeutic anti-MUC1* antibodies can be monoclonal, polyclonal, antibody mimics, engineered antibody-like molecules, full antibodies or antibody fragments. Examples of antibody fragments include but are not limited to Fabs, scFv, and scFv-Fc. Human or humanized antibodies are preferred for use in the treatment or prevention of cancers. In any of these antibody-like molecules, mutations can be introduced to prevent or minimize dimer formation. Anti-MUC1* antibodies that are monovalent or bispecific are preferred because MUC1* function is activated by ligand induced dimerization. Typical binding assays show that NME1 and NME7_AB bind to the PSMGFR peptide portion of MUC1* (FIG. 2A, 2D). Further, they show that these activating growth factors bind to the membrane proximal portion of MUC1*, as they do not bind to the PSMGFR peptide if the 10 C-terminal amino acids are missing. Similarly, anti-MUC1* antibodies MN-C2 and MN-E6 bind to the PSMGFR peptide if an only if the 10 C-terminal amino acids are present (FIG. 2B, 2C). Antibodies MN-C3 and MN-C8 bind to epitopes that are different from MN-C2 and MN-E6, as they do not depend on the presence of the 10 C-terminal amino acids of the PSMGFR peptide (FIG. 2E, 2F). Antibodies MN-C2, MN-E6, or fragments derived from them, can be administered to a patient for the treatment or prevention of cancers, as stand-alone antibodies or incorporated into bispecific antibodies, BiTEs or chimeric antigen receptors also called CARs that have been transduced into immune cells. MNC2 and MNE6 and other anti-MUC1* antibodies that competitively inhibit the binding of NME1 and NME7_AB are preferred for use as stand alone antibody therapeutics.

[0503] Therapeutic anti-MUC1* antibodies for use as a stand alone antibody therapeutic or for integration into a BiTE or a CAR can be selected based on specific criteria. The parent antibody can be generated using typical methods for generating monoclonal antibodies in animals. Alternatively, they can be selected by screening antibody and antibody fragment libraries for their ability to bind to a MUC1* peptide, which can be:

• • [0504] (i) PSMGFR region of MUC1;
  • [0505] (ii) PSMGFR peptide;
  • [0506] (iii) a peptide having amino acid sequence of QFNQYKTEAASRYNLTISDVSVSDVPFPFSAQSGA (N-10) (SEQ ID NO: 3)
  • [0507] (iv) a peptide having amino acid sequence of
  • [0508] ASRYNLTISDVSVSDVPFPFSAQSGA (N-19) (SEQ ID NO: 4)
  • [0509] (v) a peptide having amino acid sequence of
  • [0510] NLTISDVSVSDVPFPFSAQSGA (N-23) (SEQ ID NO: 5)
  • [0511] (vi) a peptide having amino acid sequence of
  • [0512] ISDVSVSDVPFPFSAQSGA (N-26) (SEQ ID NO: 6)
  • [0513] (vii) a peptide having amino acid sequence of
  • [0514] SVSDVPFPFSAQSGA (N-30) (SEQ ID NO: 7)
  • [0515] (viii) a peptide having amino acid sequence of
  • [0516] QFNQYKTEAASRYNLTISDVSVSDVPFPFS (N-10/C-5) (SEQ ID NO: 8)
  • [0517] (ix) a peptide having amino acid sequence of
  • [0518] ASRYNLTISDVSVSDVPFPFS (N-19/C-5) (SEQ ID NO: 9) or
  • [0519] (x) a peptide having amino acid sequence of
  • [0520] FPFSAQSGA (N-36) (SEQ ID NO: 10).

[0521] Resultant antibodies or antibody fragments generated or selected in this way can then be further selected by passing additional screens. For example, antibodies or antibody fragments become more preferred based on their ability to bind to MUC1* positive cancer cells or tissues but not to MUC1 negative cancer cells or to normal tissues. Further, anti-MUC1* antibodies or antibody fragments may be de-selected as anti-cancer therapeutics if they bind to stem or progenitor cells. Anti-MUC1* antibodies or antibody fragments become more preferred if they have the ability to competitively inhibit the binding of activating ligands to MUC1*. FIGS. 3A-3C shows that MN-E6 and MN-C2 competitively inhibit the binding of activating ligands NME1 and NME7 to MUC1*.

[0522] A process for selecting anti-MUC1* antibodies for use in treating a patient diagnosed with a MUC1 positive cancer, at risk of developing a MUC1 positive cancer or suspected of having a MUC1 positive cancer comprises one or more of the following steps of selecting antibodies or antibody fragments that 1) bind to the PSMGFR peptide; 2) bind to the N-10 PSMGFR peptide; 3) bind to cancer cells; 4) do not bind to stem or progenitor cells; and 5) competitively inhibited the binding of dimeric NME1 or NME7-AB to the PSMGFR peptide. For example, FIGS. 3A-3C show that monoclonals MN-E6 and MN-C2 satisfy all five criteria, while monoclonals MN-C3 and MN-C8 do not competitively inhibit the binding of activating ligands NME1 and NME7 (FIG. 3C). Recall that the MUC1* growth factor receptor is activated by ligand-induced dimerization of its extracellular domain. Therefore, the ideal antibody therapeutic should not dimerize the MUC1* extracellular domain. Preferably, suitable antibodies in this regard include monovalent antibodies such as those generated in lamas and camels, Fabs, scFv's, single domain antibodies (sdAb), scFv-Fc as long as the Fc portion is constructed such that it does not homo-dimerize.

[0523] FACS scans show that anti-MUC1* antibodies MN-C2 and MN-E6 specifically bind to MUC1* positive solid tumor cancer cells and MUC1* transfected cells but not MUC1* negative or MUC1 negative cells. In one example, a humanized MN-C2 scFv is shown to bind to ZR-75-1, aka 1500, MUC1* positive breast cancer cells (FIG. 4A-4C). MN-E6 was shown to bind to MUC1 negative HCT-116 colon cancer cells if an only if they were transfected with MUC1*. MN-E6 also bound to MUC1* positive cancer cells such as ZR-75-1, aka 1500, MUC1* positive breast cancer cells (FIG. 4D-4F). Binding assays such as ELISAs, immunofluorescence, and the like all confirm that MN-C2 and MN-E6 bind to the PSMGFR peptide and to live MUC1 positive cancer cells. Humanized anti-MUC1* antibodies are selected based on their ability to also bind to the PSMGFR peptide or to MUC1 positive cancer cells. FIG. 5 shows that humanized MN-C2 scFv binds with high affinity to the MUC1* peptide PSMGFR with an EC-50 of about 333 nM. Humanized MN-C2 scFv, like Fabs, potently inhibits the growth of MUC1* positive cancer cells as is shown in one example in FIGS. 6A, 6B. Like the parent antibodies, humanized scFv's show the same binding pattern. huMNE6-scFv binds to the PSMGFR peptide, binds to the N-10 peptide but does not bind to the C-10 peptide (SEQ ID NO: 825) (FIG. 8). Murine or humanized MNC3-scFv binds to the, PSMGFR peptide, binds to the N-10 peptide and binds to the C-10 peptide (FIG. 9).

[0524] The Fabs of MN-E6 and MN-C2 or the comparable single chain variable regions derived from them potently inhibit the growth of MUC1* positive cancers in vitro and in vivo. In several examples, the Fabs of Anti-MUC1* antibodies inhibited the growth of human MUC1* positive cancers in vivo. In one case, immune-compromised mice were implanted with human breast tumors then treated with MN-E6 Fab after tumor engraftment. FIG. 7A shows that MN-E6 Fab potently inhibited the growth of MUC1* positive breast cancers. Female nu/nu mice implanted with 90-day estrogen pellets were implanted with 6 million T47D human breast cancer cells that had been mixed 50/50 with Matrigel. Mice bearing tumors that were at least 150 mm³ and had three successive increases in tumor volume were selected for treatment. Animals were injected sub-cutaneously twice per week with 80 mg/kg MN-E6 Fab and an equal number of mice fitting the same selection criteria were injected with vehicle alone (FIG. 7A).

[0525] In another aspect, MN-E6 was shown to halt the growth of prostate cancer. FIG. 7B shows that MN-E6 Fab potently inhibited the growth of MUC1* positive prostate cancers. Male NOD/SCID mice were implanted with 6 million DU-145 human prostate cancer cells that had been mixed 50/50 with Matrigel. Mice bearing tumors that were at least 150 mm{circumflex over ( )}3 and had three successive increases in tumor volume were selected for treatment. Animals were injected sub-cutaneously every 48 hours with 160 mg/kg MN-E6 Fab and an equal number of mice fitting the same selection criteria were injected with vehicle alone (FIG. 7B). Tumors were measured independently by two researchers twice per week and recorded. Statistics were blindly calculated by independent statistician, giving a P value of 0.0001 for each. Anti-MUC1* Fab inhibited breast cancer growth and prostate cancer growth. Treatment had no effect on weight, bone marrow cell type or number. The MN-E6 Fab effectively inhibited the growth of the tumors, while the control group's tumors continued to grow until sacrifice. No adverse effects of treatment were observed or detected.

[0526] Recombinant forms of MN-E6 and MNC2 were constructed that like the Fab are monomeric. In this case, MN-E6 was humanized and MN-C2 was humanized. There are a number of methods known to those skilled in the art for humanizing antibodies. In addition to humanizing, libraries of human antibodies can be screened to identify other fully human antibodies that bind to the PSMGFR.

[0527] A single chain of the humanized MN-E6 variable region, called an scFv, was genetically engineered such that it was connected to the Fc portion of the antibody (SEQ ID NO: 256 and 257). Fc regions impart certain benefits to antibody fragments for use as therapeutics. The Fc portion of an antibody recruits complement, which in general means it can recruit other aspects of the immune system and thus amplify the anti-tumor response beyond just inhibiting the target. The addition of the Fc portion also increases the half-life of the antibody fragment (Czajkowsky D M, Hu J, Shao Z and Pleass R J. (2012) Fc-fusion proteins: new developments and future perspectives. EMBO Mol Med. 4 (10): 1015-1028). However, the Fc portion of an antibody homo-dimerizes, which in the case of anti-MUC1* antibody based therapeutics is not optimal since ligand-induced dimerization of the MUC1* receptor stimulates growth. Therefore, mutations in the Fc region that resist dimer formation are preferred for anti-MUC1* anti-cancer therapeutics. Deletion of the hinge region and other mutations in the Fc region that make the Fc-mutant resistant to dimerization were made and could be used as therapeutics.

[0528] A human or humanized MN-E6 antibody or antibody fragment, Fab, MN-E6 scFv or hu MN-E6 scFv-Fc_mut are effective anti-cancer agents that can be administered to a person diagnosed with a MUC1 or MUC1* positive cancer, suspected of having a MUC1 or MUC1* positive cancer or is at risk of developing a MUC1 or MUC1* positive cancer.

Humanizing

[0529] Humanized antibodies or antibody fragments or fully human antibodies that bind to the extracellular domain of −MUC1* are preferred for therapeutic use. The techniques described herein for humanizing antibodies are but a few of a variety of methods known to those skilled in the art. The invention is not meant to be limited by the technique used to humanize the antibody.

[0530] Humanization is the process of replacing the non-human regions of a therapeutic antibody (usually mouse monoclonal antibody) by human one without changing its binding specificity and affinity. The main goal of humanization is to reduce immunogenicity of the therapeutic monoclonal antibody when administered to human. Three distinct types of humanization are possible. First, a chimeric antibody is made by replacing the non-human constant region of the antibody by the human constant region. Such antibody will contain the mouse Fab region and will contain about 80-90% of human sequence. Second, a humanized antibody is made by grafting of the mouse CDR regions (responsible of the binding specificity) onto the variable region of a human antibody, replacing the human CDR (CDR-grafting method). Such antibody will contain about 90-95% of human sequence. Third and last, a full human antibody (100% human sequence) can be created by phage display, where a library of human antibodies is screened to select antigen specific human antibody or by immunizing transgenic mice expressing human antibody.

[0531] A general technique for humanizing an antibody is practiced approximately as follows. Monoclonal antibodies are generated in a host animal, typically in mice. Monoclonal antibodies are then screened for affinity and specificity of binding to the target. Once a monoclonal antibody that has the desired effect and desired characteristics is identified, it is sequenced. The sequence of the animal-generated antibody is then aligned with the sequences of many human antibodies in order to find human antibodies with sequences that are the most homologous to the animal antibody. Biochemistry techniques are employed to paste together the human antibody sequences and the animal antibody sequences. Typically, the non-human CDRs are grafted into the human antibodies that have the highest homology to the non-human antibody. This process can generate many candidate humanized antibodies that need to be tested to identify which antibody or antibodies has the desired affinity and specificity.

[0532] Once a human antibody or a humanized antibody has been generated it can be further modified for use as an Fab fragment, as a full antibody, or as an antibody-like entity such as a single chain molecule containing the variable regions, such as scFv or an scFv-Fc. In some cases it is desirable to have Fc region of the antibody or antibody-like molecule mutated such that it does not dimerize.

[0533] In addition to methods that introduce human sequences into antibodies generated in non-human species, fully human antibodies can be obtained by screening human antibody libraries with a peptide fragment of an antigen. A fully human antibody that functions like MN-E6 or MN-C2 is generated by screening a human antibody library with a peptide having the sequence of the PSMGFR N-10 peptide. Humanized anti-MUC1* antibodies were generated based on the sequences of the mouse monoclonal antibodies MN-E6 and MN-C2. In one aspect of the invention, a patient diagnosed with a MUC1* positive cancer is treated with an effective amount of a murine or camelid MNC2, MNE6, 20A10 (SEQ ID NOS: 1574-1581), 3C2B1 (SEQ ID NOS: 1572-1573), 5C6F3, 25E6 (SEQ ID NO:1598-1601), 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11. In another aspect of the invention, a patient diagnosed with a MUC1* positive cancer is treated with an effective amount of humanized MN-E6 or MN-C2. In a preferred embodiment, a patient diagnosed with a MUC1* positive cancer is treated with an effective amount of humanized MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11. In another aspect of the invention, a patient diagnosed with a MUC1* positive cancer is treated with an effective amount of humanized monovalent MNC2, MNE6, 20A10 (SEQ ID NOS: 1574-1581), 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11, wherein monovalent means the corresponding Fab fragment, the corresponding scFv or the corresponding scFv-Fc fusion. In a preferred embodiment, a patient diagnosed with a MUC1* positive cancer is treated with an effective amount of a humanized scFv or monomeric humanized scFv-Fc of MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11. Since the MUC1* growth factor receptor is activated by ligand induced dimerization of its extracellular domain, and because the Fc portion of an antibody homo-dimerizes, it is preferable that a construct that includes an Fc portion uses a mutated Fc region that prevents or minimizes dimerization.

[0534] Antibodies that bind to PSMGFR (SEQ ID NO:2) peptide, and more specifically to the N-10 peptide, of the extracellular domain of the MUC1* receptor are potent anti-cancer therapeutics that are effective for the treatment or prevention of MUC1* positive cancers. They have been shown to inhibit the binding of activating ligands dimeric NME1 (SEQ ID NO:1781) and NME7_AB (SEQ ID NOS: 827) to the extracellular domain of MUC1*. Anti-MUC1* antibodies that bind to the PSMGFR sequence inhibit the growth of MUC1*-positive cancer cells, specifically if they inhibit ligand-induced receptor dimerization. Fabs of anti-MUC1* antibodies have been demonstrated to block tumor growth in animals. Thus, antibodies or antibody fragments that bind to the extracellular domain of MUC1* would be beneficial for the treatment of cancers wherein the cancerous tissues express MUC1*.

[0535] Antibodies that bind to PSMGFR region of MUC1* or bind to a synthetic PSMGFR peptide are preferred. We have identified several monoclonal antibodies that bind to the extracellular domain of MUC1*. Among this group are mouse monoclonal antibodies MNC2 (SEQ ID NOS: 118-131, 144-158, 163-164, 168-181, 194-209), MNE6 (SEQ ID NOS: 12-25, 39-59, 65-78, 93-114), 20A10 (SEQ ID NOS: 988-1019, 1574-1597, 1659-1666); 3C2B1 (SEQ ID NOS: 1386-1413, 1572-1573), 5C6F3 (SEQ ID NOS: 1356-1385), 25E6 (SEQ ID NOS: 1020-1051, 1598-1617, 1667-1674), 18G12 (SEQ ID NOS: 956-987), 28F9 (SEQ ID NOS: 1052-1083), 1E4 (SEQ ID NOS: 1116-1227), B12 (SEQ ID NOS: 1414-1431, 1733-1742), B2 (SEQ ID NOS: 1432-1459), B7 (SEQ ID NOS: 1460-1487), B9 SEQ ID NOS: 1544-1571), 8C7F3 (SEQ ID NOS: 1488-1515), or H11 (SEQ ID NOS: 1516-1543), the variable regions of which were sequenced and are given as for MN-E6 SEQ ID NOS: 12-13 and 65-66, for MN-C2 SEQ ID NOS: 118-119 and 168-169. The CDRs of these antibodies make up the recognition units of the antibodies and are the most important parts of the mouse antibody that should be retained when grafting into a human antibody. The sequences of the CDRs for each mouse monoclonal are as follows, heavy chain sequence followed by light chain: MN-E6 CDR1 (SEQ ID NO:16-17 and 69-70) CDR2 (SEQ ID NO:20-21 and 73-74) CDR3 (SEQ ID NO: 24-25 and 77-78), MN-C2 CDR1 (SEQ ID NO:122-123 and 172-173) CDR2 (SEQ ID NO:126-127 and 176-177) CDR3 (SEQ ID NO:130-131 and 180-181). In some cases, portions of the framework regions that by modeling are thought to be important for the 3-dimensional structure of the CDRs, are also imported from the mouse sequence.

[0536] Monoclonal antibodies MN-E6 and MN-C2 have greater affinity for MUC1* as it appears on cancer cells. Monoclonal antibodies MN-C3 and MN-C8 have greater affinity for MUC1* as it appears on stem cells.

[0537] All four antibodies have been humanized, which process has resulted in several humanized forms of each antibody. CDRs derived from the variable regions of the mouse antibodies were biochemically grafted into a homologous human antibody variable region sequence. Humanized variable regions of MN-E6 (SEQ ID NOS: 38-39 and 93-94), MN-C2 (SEQ ID NOS: 144-145 and 194-195), MN-C3 (SEQ ID NOS: 439-440 and 486-487) and MN-C8 (SEQ ID NOS: 525-526 and 543-544) were generated by grafting the mouse CDRs into the variable region of a homologous human antibody. The humanized heavy chain variable constructs were then fused into constant regions of either human IgG1 heavy chain constant region (SEQ ID NOS: 58-59) or human IgG2 heavy chain constant region (SEQ ID NO:54-55), which are then paired with either humanized light chain variable constructs fused to a human kappa chain (SEQ ID NO: 109-110) or human lambda chain (SEQ ID NO: 113-114) constant region. Other IgG isotypes could be used as constant region including IgG3 or IgG4.

[0538] Examples of humanized MN-E6 variable region into an IgG2 heavy chain (SEQ ID NOS: 52-53) and into an IgG1 heavy chain (SEQ ID NOS: 56-57), humanized MN-C2 variable into an IgG1 heavy chain (SEQ ID NOS: 157-158) or into an IgG2 heavy chain (SEQ ID NOS: 163-164) paired with either Lambda light chain (SEQ ID NO: 111-112 and 216-219) or Kappa chain (SEQ ID NO:107-108 and 210-213) and, humanized MN-C3 (SEQ ID NOS: 455-456, 453-454 and 500-501, 502-503) and MN-C8 (SEQ ID NOS: 541-542, 539-540 and 579-580, 581-582) antibodies were generated. Which IgG constant region is fused to the humanized variable region depends on the desired effect since each isotype has its own characteristic activity. The isotype of the human constant region is selected on the basis of things such as whether antibody dependent cell cytotoxicity (ADCC) or complement dependent cytotoxicity (CDC) is desired but can also depend on the yield of antibody that is generated in cell-based protein expression systems. In a preferred embodiment, humanized anti-MUC1* antibodies or antibody fragments are administered to a person diagnosed with or at risk of developing a MUC1-positive cancer.

[0539] One method for testing and selecting the humanized anti-MUC1* antibodies that would be most useful for the treatment of persons with cancer or at risk of developing cancers is to test them for their ability to inhibit the binding of activating ligands to the MUC1* extracellular domain. Dimeric NME1 can bind to and dimerize the MUC1* extracellular domain and in so doing stimulates cancer cell growth. Antibodies and antibody fragments that compete with NME1 for binding to the MUC1* extracellular domain are therefore anti-cancer agents. NME7_AB is another activating ligand of MUC1*. In some cases, it is preferable to identify antibodies that block the binding of NME7, or an NME7_AB truncation or cleavage product of NME7-X1, to the MUC1* extracellular domain. Antibodies and antibody fragments that compete with NME7 and NME7 variants for binding to the MUC1* extracellular domain are effective as anti-cancer therapeutics. These antibodies include but are not limited to MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11 as well as single chain versions, such as scFv, of these antibodies and humanized version thereof. Other NME proteins also bind to MUC1 or MUC1* including NME6 and NME8. Antibodies that compete with these proteins for binding to MUC1* may also be useful as therapeutics. In a preferred embodiment, murine, camelid, human or humanized anti-MUC1* antibodies or antibody fragments are administered to a person diagnosed with or at risk of developing a MUC1-positive cancer. In a more preferred embodiment, single chain antibody fragments, or monomeric scFv-Fc fusions, derived from humanized sequences of MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11 are administered to a person diagnosed with or at risk of developing a MUC1-positive cancer.

[0540] Single chain variable fragments, scFv, or other forms that result in a monovalent antibody or antibody-like protein are also useful. In some cases it is desired to prevent dimerization of the MUC1* extracellular domain. Single chain variable fragments, Fabs and other monovalent antibody-like proteins have been shown to be effective in binding to the extracellular domain of MUC1* and blocking MUC1* dimerization. These single chain variable fragments, Fabs and other monovalent antibody-like molecules effectively blocked cancer growth in vitro and in animals xenografted with human MUC1-positive cancer cells. Thus, humanized single chain variable fragments or monovalent anti-MUC1* antibodies or antibody-like molecules would be very effective as an anti-cancer therapeutic. Such humanized single chain antibodies, Fabs and other monovalent antibody-like molecules that bind to the MUC1* extracellular domain or to a PSMGFR peptide are therefore useful as anti-cancer therapeutics. Anti-MUC1* single chain variable fragments are generated by grafting non-human CDRs of antibodies, which bind to extracellular domain of MUC1* or bind to PSMGFR peptide, into a framework of a homologous variable region human antibody. The resultant humanized heavy and light chain variable regions are then connected to each other via a suitable linker, wherein the linker should be flexible and of length that it allows heavy chain binding to light chain but discourages heavy chain of one molecule binding to the light chain of another. For example a linker of about 10-15 residues. Preferably, the linker includes [(Glycine)₄ (Serine)₁]₃ (SEQ ID NOS: 401-402), but is not limited to this sequence as other sequences are possible.

[0541] In one aspect, the humanized variable regions of MN-E6 (SEQ ID NOS: 38-39 and 93-94), MN-C2 (SEQ ID NOS: 144-145 and 194-195), or other antibodies of the invention are biochemically grafted into a construct that connects heavy and light chains via a linker. Examples of humanized single chain anti-MUC1* antibodies comprising humanized sequences from the variable regions of MN-E6 and MN-C2, were generated. Several humanized MN-E6 single chain proteins were generated (SEQ ID NOS: 232-237). Several humanized MN-C2 single chain proteins were generated (SEQ ID NOS: 238-243). In a preferred embodiment, humanized anti-MUC1* antibody fragments, including variable fragments, scFv antibody fragments MN-E6 scFv, MN-C2 scFv, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11 scFv are administered to a person diagnosed with or at risk of developing a MUC1-positive cancer.

[0542] One aspect of the invention is a method for treating a patient diagnosed with, suspected of having, or at risk of developing a MUC1 positive or MUC1* positive cancer, wherein the patient is administered an effective amount of a monomeric MN-E6 scFv, MN-C2 scFv, or MN-E6 scFv-Fc, MN-C2 scFv-Fc, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11, wherein the antibody variable fragment portions are human or have been humanized and wherein the Fc portion of the antibody-like protein has been mutated such that it resists dimer formation.

CAR T and Cancer Immunotherapy Techniques

[0543] In another aspect of the invention, some or all of the single chain portions of anti-MUC1* antibody fragments are biochemically fused onto immune system molecules, using several different chimeric antigen receptor, ‘CAR’ strategies. The idea is to fuse the recognition portion of an antibody, typically as a single chain variable fragment, to an immune system molecule that has a transmembrane domain and a cytoplasmic tail that is able to transmit signals that activate the immune system. The recognition unit can be an antibody fragment, a single chain variable fragment, scFv, or a peptide. In one aspect, the recognition portion of the extracellular domain of the CAR is comprised of sequences from the humanized variable region of MN-E6 (SEQ ID NOS: 38-39 and 93-94), MN-C2 (SEQ ID NOS: 144-145 and 194-195), 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11. Examples of murine or humanized antibodies of the invention, or their single chain fragments, scFv's, which can be incorporated into CARs, BiTEs or ADCs are given as: 3C2B1 (SEQ ID NOS: 1572-1573), 20A10 (SEQ ID NOS: 1574-1581), 25E6 (SEQ ID NOS: 1598-1601). In another aspect, it is comprised of sequences from a single chain variable fragment. Examples of single chain constructs are given. Several humanized MN-E6 single chain proteins, scFv, were generated (SEQ ID NOS: 232-237). Several humanized MN-C2 single chain proteins, scFv, were generated (SEQ ID NOS: 238-243). The transmembrane region of the CAR can be derived from CD8, CD4, antibody domains or other transmembrane region, including the transmembrane region of the proximal cytoplasmic co-stimulatory domain, such as CD28, 4-1BB or other. The cytoplasmic tail of the CAR can be comprised of one or more motifs that signal immune system activation. This group of cytoplasmic signaling motifs, sometimes referred to as, co-stimulatory cytoplasmic domains, includes but is not limited to CD3-zeta, CD27, CD28, 4-1BB, OX40, CD30, CD40, ICAm-1, LFA-1, ICOS, CD2, CD5, CD7 and Fc receptor gamma domain. A minimal CAR may have the CD3-zeta or an Fc receptor gamma domain then one or two of the above domains in tandem on the cytoplasmic tail. In one aspect, the cytoplasmic tail comprises CD3-zeta, CD28, 4-1BB and/or OX40.

[0544] The extracellular domain recognition unit of a MUC1* targeting CAR can comprise variable regions of any non-human, humanized or human antibody that is able to bind to at least 12 contiguous amino acids of the PSMGFR peptide (SEQ ID NO:2) or the N-10 peptide. In one aspect, the MUC1* targeting portion of the CAR comprises variable regions from non-human, humanized or human MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11. Examples of a few antibodies of the invention, incorporated into CARs as either murine or humanized are given as 20A10 (SEQ ID NOS: 1582-1597) and 25E6 (SEQ ID NOS: 1602-1617). In the humanization process, the antibody CDRs can be inserted into a number of different framework regions; as a demonstration we generated three versions of a humanized 20A10 which differ only in the framework regions. These have been incorporated into CARs (SEQ ID NOS: 1675, 1678, 1685) that when transduced into human T cells are able to recognize target MUC1* expressing cells and kill them. In one aspect, the extracellular domain recognition unit of a CAR is comprised essentially of a humanized MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11 single chain variable fragment scFv. The transmembrane region of the CAR can be derived from CD8 (SEQ ID NOS: 363-364), or can be the transmembrane domain of CD3-zeta, CD28, 41bb, OX40 or other transmembrane region (SEQ ID NOS: 361-372) and the cytoplasmic domain of a CAR with antibody fragment targeting MUC1* extracellular domain can be comprised of one or more selected from the group comprising an immune system co-stimulatory cytoplasmic domain. The group of immune system co-stimulatory domains includes but is not limited to CD3-zeta, CD27, CD28, 4-1BB, OX40, CD30, CD40, ICAm-1, LFA-1, ICOS, CD2, CD5, CD7 and Fc receptor gamma domain (SEQ ID NOS: 373-382).

[0545] The CARs described can be transfected or transduced into a cell of the immune system. In a preferred embodiment, a MUC1* targeting CAR is transfected or transduced into a T cell. In one aspect, the T cell is a CD3+/CD28+ T cell. In another case it is a dendritic cell. In another case it is a B cell. In another case it is a mast cell. In yet another case it is a Natural Killer, NK, cell. The recipient cell can be from a patient or from a donor. If from a donor, it can be engineered to remove molecules that would trigger rejection. Cells transfected or transduced with a CAR of the invention can be expanded ex vivo or in vitro then administered to a patient. Administrative routes are chosen from a group containing but not limited to bone marrow transplant, intravenous injection, in situ injection or transplant. In a preferred embodiment, the MUC1* targeting CAR is administered to a person diagnosed with or at risk of developing a MUC1-positive cancer.

[0546] There are many possible anti-MUC1* CAR constructs that can be transduced into T cells or other immune cells for the treatment or prevention of MUC1* positive cancers. CARs are made up of modules and the identity of some of the modules is relatively unimportant, while the identity of other modules is critically important.

[0547] We and others have shown that intracellular signaling modules, such as CD3-zeta (SEQ ID NOS: 373-376), CD28 (SEQ ID NOS: 377-378) and 41BB (SEQ ID NOS: 379-380), alone or in combinations stimulate immune cell expansion, cytokine secretion and immune cell mediated killing of the targeted tumor cells (Pule M A, Straathof K C, Dotti G, Heslop H E, Rooney C M and Brenner M K (2005) A chimeric T cell antigen receptor that augments cytokine release and supports clonal expansion of primary human T cells. Mol Ther. 12 (5): 933-941; Hombach A A, Heiders J, Foppe M, Chmielewski M and Abken H. (2012) OX40 costimulation by a chimeric antigen receptor abrogates CD28 and IL-2 induced IL-10 secretion by redirected CD4 (+) T cells. Oncoimmunology. 1 (4): 458-466; Kowolik C M, Topp M S, Gonzalez S, Pfeiffer T, Olivares S, Gonzalez N, Smith D D, Forman S J, Jensen M C and Cooper L J. (2006) CD28 costimulation provided through a CD19-specific chimeric antigen receptor enhances in vivo persistence and antitumor efficacy of adoptively transferred T cells. Cancer Res. 66 (22): 10995-11004; Loskog A, Giandomenico V, Rossig C, Pule M, Dotti G and Brenner MK. (2006) Addition of the CD28 signaling domain to chimeric T-cell receptors enhances chimeric T-cell resistance to T regulatory cells. Leukemia. 20 (10): 1819-1828; Milone M C, Fish J D, Carpenito C, Carroll R G, Binder G K, Teachey D, Samanta M, Lakhal M, Gloss B, Danet-Desnoyers G, Campana D, Riley J L, Grupp S A and June C H. (2009) Chimeric receptors containing CD137 signal transduction domains mediate enhanced survival of T cells and increased antileukemic efficacy in vivo. Mol Ther. 17 (8): 1453-1464; Song D G, Ye Q, Carpenito C, Poussin M, Wang LP, Ji C, Figini M, June C H, Coukos G, Powell DJ Jr. (2011) In vivo persistence, tumor localization, and antitumor activity of CAR-engineered T cells is enhanced by costimulatory signaling through CD137 (4-1BB). Cancer Res. 71 (13): 4617-4627). Antibodies of the invention including but not limited to fragments of MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11 can also be incorporated into CARs that have mutated cytoplasmic tails, such as mutated tyrosines or ITAMs. In any of the CARs described above, the cytoplasmic tails may include mutations that dampen signaling. Such mutations include but are not limited to Tyrosines that are mutated to inhibit phosphorylation and signaling (Salter et al, 2018;). In any of the CARs described above, the ITAMs of CD3-zeta may be mutated to inhibit or dampen signaling (Feucht et al 2019). In any of the CARs described above, the CD3 of the cytoplasmic tail may comprise mutations in the ITAMs including those referred to as 1XX. Examples of antibodies of the invention incorporated into CARs with 1XX mutations in ITAMs of CD3-zeta are given in the following sequences: MNC2 (SEQ ID NOS: 1618-1625), MNE6 (SEQ ID NOS: 1626-1633), 20A10 (SEQ ID NOS: 1590-1595), 25E6 (SEQ ID NOS: 1610-1617). We note that the CDRs of antibodies can be inserted into a background of a number of different framework regions. As an example, 20A10 CDRs were inserted into three different sets of framework regions (SEQ ID NOS: 1692, 1699 and 1706) and all were able to function when transduced into T cells. In any of the CARs described above, the T cell may be engineered to overexpress c-Jun as a method to inhibit T cell exhaustion (Lynn et al 2019). A variety of promoters can be used upstream of the genes for CARs and other compositions of the invention, including insertion into a naturally occurring promoter in the cell, such as the TRAC locus, using CRISPR, Sleeping Beauty or similar technology for site directed insertion of a gene. Among the promoters commonly used are the CMV promoter, or a mini CMV (SEQ ID NO: 1634), a minimal IL-2 promoter (SEQ ID NO: 1635), or Minimal Promoter minip (SEQ ID NO: 1636).

[0548] Single chain antibody fragments that included the variable domain of the monoclonal anti-MUC1* antibodies called MN-E6 or MN-C2 were engineered into a panel of CARs. The MUC1* targeting CARs were then transduced, separately or in combinations, into immune cells. When challenged with surfaces presenting a MUC1* peptide, an antigen presenting cell transfected with MUC1*, or MUC1* positive cancer cells, the immune cells that were transduced with MUC1* targeting CARs elicited immune responses, including cytokine release, killing of the targeted cells and expansion of the immune cells.

[0549] For example, the gene encoding the CARs and activated T cell induced genes described herein can be virally transduced into an immune cell using viruses, or inserted into a region downstream of one of the cell's promoters or enhancers, such as the TRAC (T cell receptor alpha chain) locus. Virus delivery systems and viral vectors including but not limited to retroviruses, including gamma-retroviruses, lentivirus, adenoviruses, adeno-associated viruses, baculoviruses, poxvirus, herpes simplex viruses, oncolytic viruses, HF10, T-Vec and the like can be used. In addition to viral transduction, CARs and activated T cell induced genes described herein can be directly spliced into the genome of the recipient cell using methods such as CRISPR technology, CRISPR-Cas9 and -CPF1, TALEN, Sleeping Beauty transposon system, and SB 100×.

[0550] Similarly, the identity of molecules that make up the non-targeting portions of the CAR such as the extracellular domain, transmembrane domain and membrane proximal portion of the cytoplasmic domain, are not essential to the function of a MUC1*-targeting CAR. For example, the extracellular domain, transmembrane domain and membrane proximal portion of the cytoplasmic domain can be comprised of portions of CD8, CD4, CD28, or generic antibody domains such as Fc, CH2CH3, or CH3. Further, the non-targeting portions of a CAR can be a composite of portions of one or more of these molecules or other family members.

[0551] One aspect of the invention is a method for treating a patient diagnosed with, suspected of having, or at risk of developing a MUC1 positive or MUC1* positive cancer, wherein the patient is administered an effective amount of immune cells that have been transduced with a MUC1* targeting CAR. In another aspect of the invention, the immune cells are T cells isolated from a patient, which are then transduced with CARs wherein the targeting head of the CAR binds to MUC1*, and after expansion of transduced T cells, the CAR T cells are administered in an effective amount to the patient. In yet another aspect of the invention, the immune cells are T cells isolated from a patient, which are then transduced with CARs wherein the targeting head of the CAR comprises portions of MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11, and after optional expansion of transduced T cells, the CAR T cells are administered in an effective amount to the patient.

Specificity of Anti-MUC1* Targeting Antibodies

[0552] As these experiments demonstrate, the critical portion of a CAR is the antibody fragment that directs the immune cell to the tumor cell. As we will show in the following section, MN-E6 and MN-C2 are specific for the form of MUC1* that is expressed on tumor cells. The next most important part of a CAR is the cytoplasmic tail bearing immune system co-stimulatory domains. The identity of these domains modulates the degree of immune response but does not affect the specificity. As shown, the identity of the transmembrane portion of a CAR is the least important. It appears that as long as the transmembrane portion has some flexibility and is long enough to allow the antibody fragment to reach its cognate receptor on the tumor cell, it will suffice. CARs comprising the MN-E6 targeting antibody fragment, and intracellular co-stimulatory domains 41BB and CD3-zeta but having a variety of different extracellular, transmembrane and short cytoplasmic tail all worked in that they specifically killed the targeted cells while stimulating the expansion of the host T cells.

[0553] The most accurate way of demonstrating antibody specificity is testing the antibody on normal human tissue specimens compared to cancerous tissue specimens. MN-C2 and MN-E6 were shown to specifically bind to MUC1 or MUC1* positive cancer cells. Several breast tumor arrays were assayed using several anti-MUC1 or MUC1* antibodies. Essentially the studies involving serial sections of breast cancer tissue specimens from over 1,200 different breast cancer patients showed that very little full-length MUC1 remains on breast cancer tissues. The vast majority of the MUC1 expressed is MUC1* and is stained by MN-C2. The analysis was performed by Clarient Diagnostics and tissue staining was scored using the Allred method. For example, FIG. 10 shows serial sections of breast cancer tissue arrays that were stained with either VU4H5, a commercially available anti-MUC1 antibody that binds to the tandem repeats, or MN-C2 that binds to MUC1*. FIGS. 10 and 11 are photographs of breast cancer tissue arrays stained with either VU4H5 which recognizes MUC1-FL (full length) or MN-C2 which recognizes cancerous MUC1*. Tissue staining was scored using Allred scoring method which combines an intensity score and a distribution score. Below the photographs of the tissue arrays are color-coded graphs displaying the results. As can be seen, the arrays stained with VU4H5 are very light and many tissues do not stain at all despite the published reports that MUC1 is aberrantly expressed on over 96% of all breast cancers as evidenced by nucleic acid based diagnostics. In contrast, the arrays stained with MN-C2 are very dark (red versus yellow or white in graph). Additionally, many tissues did not stain at all with anti-full-length MUC1 but stained very dark with MN-C2, (see green boxes in graph). Similarly, we stained normal or cancerous breast tissues with humanized MN-E6 scFv-Fc. The antibody fragment was biotinylated so it could be visualized by a secondary streptavidin based secondary. As can be seen in FIG. 12, hMN-E6 scFv-Fc does not stain normal breast tissue but stains cancerous breast tissue. Further, the intensity and homogeneity of staining increases with tumor grade and/or metastatic grade of the patient (FIG. 12-13). Similarly, hMN-E6 scFv-Fc did not stain normal lung tissue but did stain lung cancer tissue (FIG. 14-18) and the intensity and distribution of staining increased as tumor grade or metastatic grade increased. FIG. 19 shows photographs of normal small intestine and cancerous small intestine tissues stained with humanized MN-E6-scFv-Fc biotinylated anti-MUC1* antibody at 5 ug/mL, then stained with a secondary streptavidin HRP antibody. A) is a normal small intestine tissue. B) is small intestine cancer from patient as denoted in the figure. C,D are photographs of the corresponding serial sections that were stained with the secondary antibody alone. FIG. 20 shows photographs of normal small intestine tissues stained with humanized MN-E6-scFv-Fc anti-MUC1* antibody at 50 ug/mL, then stained with a secondary goat-anti-human HRP antibody. A-D are normal small intestine tissue. E-H are photographs of the corresponding serial sections that were stained with the secondary antibody alone. FIG. 21 shows photographs of cancerous small intestine tissues stained with humanized MN-E6-scFv-Fc anti-MUC1* antibody at 50 ug/mL, then stained with a secondary goat-anti-human HRP antibody. A-D are cancerous small intestine tissue from a patient as denoted in figure. E-H are photographs of the corresponding serial sections that were stained with the secondary antibody alone. FIG. 22 shows photographs of cancerous small intestine tissues stained with humanized MN-E6-scFv-Fc anti-MUC1* antibody at 50 ug/mL, then stained with a secondary goat-anti-human HRP antibody. A-D are cancerous small intestine tissue from a patient as denoted in figure. E-H are photographs of the corresponding serial sections that were stained with the secondary antibody alone. FIG. 23 shows photographs of normal colon tissues stained with humanized MN-E6-scFv-Fc anti-MUC1* antibody at 50 ug/mL, then stained with a secondary goat-anti-human HRP antibody. A-D are normal colon. E-H are photographs of the corresponding serial sections that were stained with the secondary antibody alone. FIG. 24 shows photographs of colon cancer tissues stained with humanized MN-E6-scFv-Fc anti-MUC1* antibody at 50 ug/mL, then stained with a secondary goat-anti-human HRP antibody. A-D are colon cancer tissue from a metastatic patient as denoted in figure. E-H are photographs of the corresponding serial sections that were stained with the secondary antibody alone. FIG. 25 shows photographs of colon cancer tissues stained with humanized MN-E6-scFv-Fc anti-MUC1* antibody at 50 ug/mL, then stained with a secondary goat-anti-human HRP antibody. A-D are colon cancer tissue from a Grade 2 patient as denoted in figure. E-H are photographs of the corresponding serial sections that were stained with the secondary antibody alone. FIG. 26 shows photographs of colon cancer tissues stained with humanized MN-E6-scFv-Fc anti-MUC1* antibody at 50 ug/mL, then stained with a secondary goat-anti-human HRP antibody. A-D are colon cancer tissue from a metastatic patient as denoted in figure. E-H are photographs of the corresponding serial sections that were stained with the secondary antibody alone. FIG. 27 shows photographs of prostate cancer tissues stained with humanized MN-E6-scFv-Fc anti-MUC1* antibody at 50 ug/mL, then stained with a secondary goat-anti-human HRP antibody. A-D are prostate cancer tissue from a patient as denoted in figure. E-H are photographs of the corresponding serial sections that were stained with the secondary antibody alone. FIG. 28 shows photographs of prostate cancer tissues stained with humanized MN-E6-scFv-Fc anti-MUC1* antibody at 50 ug/mL, then stained with a secondary goat-anti-human HRP antibody. A-D are prostate cancer tissue from a patient as denoted in figure. E-H are photographs of the corresponding serial sections that were stained with the secondary antibody alone. FIG. 29 shows photographs of prostate cancer tissues stained with humanized MN-E6-scFv-Fc anti-MUC1* antibody at 50 ug/mL, then stained with a secondary goat-anti-human HRP antibody. A-D are prostate cancer tissue from a patient as denoted in figure. E-H are photographs of the corresponding serial sections that were stained with the secondary antibody alone.

[0554] One aspect of the invention is a method for treating a patient diagnosed with, suspected of having, or at risk of developing a MUC1 positive or MUC1* positive cancer, wherein a specimen is obtained from the patient's cancer and is tested for reactivity with an antibody that binds to PSMGFR SEQ ID NO:2, or more specifically to the N-10 peptide. The patient is then treated with an scFv, scFv-Fc or CAR T that comprises antibody variable fragments from the antibody that reacted with their cancer specimen or can be chosen from among MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11. Another aspect of the invention is a method for treating a patient diagnosed with, suspected of having, or at risk of developing a MUC1 positive or MUC1* positive cancer, wherein a specimen is obtained from the patient's cancer and is tested for reactivity with MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11; the patient is then treated with the antibody, antibody fragment, scFv, scFv-Fc-mut, BiTE or CAR T that comprises portions of the antibody that reacted with their cancer specimen.

[0555] As we previously reported, it is MUC1*, the transmembrane cleavage product, not full-length MUC1, the is a growth factor receptor that drives tumor growth. The growth factors that activate MUC1* bind to ectopic sites that are only exposed after cleavage and release of the tandem repeat portion of MUC1. Antibodies of the invention, like the activating growth factors, cannot bind to full-length MUC1. FACS analysis clearly shows that anti-MUC1* antibody MNC2 is unable to bind to HCT-116, MUC1 negative cells (FIG. 35A), binds robustly to those cells if they are transfected with MUC1* (FIG. 35B), but will not bind to HCT cells transfected with full-length MUC1 (FIG. 35C). A commercially available anti-tandem repeat antibody VU4H5 clearly recognizes full-length MUC1 (FIG. 35D).

[0556] We discovered that MUC1 can be cleaved to MUC1* by more than one cleavage enzyme and that the site of cleavage affects its fold and consequently affects which monoclonal antibody is able to recognize that form of MUC1*. Different cancer cells or cancerous tissues express different cleavage enzymes. We tested various cleavage enzyme inhibitors on different cancer cell lines and found that an inhibitor that inhibits cleavage of MUC1 in one cancer cell line did not inhibit its cleavage in another cancer cell line. Similarly, PCR experiments showed that cleavage enzymes are expressed at different levels in different cells or cell lines. For example, hematopoietic stem cells of the bone marrow express a MUC1* that is recognized by monoclonal antibody MNC3 but not MNE6 or MNC2 (FIG. 39). The growth of DU145 prostate cancer cells and T47D breast cancer cells is inhibited by the Fabs of MNC2 and MNE6 but not by the Fabs of MNC3 or MNC8, indicating that the cancer cell lines express a MUC1* that is recognized by MNE6 and MNC2 but not by MNC3 or MNC8 (FIG. 42). PCR experiments show that CD34 positive cells of the bone marrow express about 2,500-times more MMP2 and about 350-times more ADAM28 than T47D breast cancer cells, while DU145 prostate cancer cells express about 2,000-times more ADAM TS16, about 400-times more MMP14 and about 100-times more MMP1 than T47D breast cancer cells (FIG. 43 and FIG. 44). Conversely, T47D breast cancer cells express about 80-times more MMP9 than the bone marrow cells and about twice as much as DU145 prostate cancer cells. Various cleavage enzyme inhibitors were tested for their ability to inhibit cleavage in different kinds of cancer cells.

General Strategy for Using Antibodies, Antibody Fragments and CARs that Target the Extracellular Domain of MUC1*

[0557] In one aspect of the invention, a second factor, which may be a cleavage enzyme, an antibody, a cytokine, or a second CAR, and a CAR are transduced into the same T cell. In another aspect of the invention, the second factor is on an inducible promoter such that its expression is activated when the CAR engages the targeted cancer cells. In some cases, the expression of the second factor is controlled by an inducible promoter. In one aspect of the invention, expression of the second factor is induced when the immune cell is activated, for example when it recognizes or engages its target. In one example, a T cell is transfected or transduced with a second factor whose expression is induced when the T cell recognizes a target cancer cell. One way to do this is to induce expression of the second factor when, or shortly after, an NFAT protein is expressed or translocated to the nucleus. For example, a sequence derived from an NFAT promoter region is put upstream of the gene for the second factor. In this way, when the transcription factors that bind to the promoter of the NFAT protein are present in sufficient concentration to bind to and induce transcription of the NFAT protein, they will also bind to that same promoter that is engineered in front of the sequence for transcription of the second factor. The NFAT protein may be NFAT1 also known as NFATc2, NFAT2 also known as NFATc or NFATc1, NFAT3 also known as NFATc4, NFAT4 also known as NFATc3, or NFAT5. In one aspect of the invention, the NFAT is NFATc1, NFATc3 or NFATc2. In one aspect of the invention, the NFAT is NFAT2 also known as NFATc1. SEQ ID NO:646 shows nucleic acid sequence of the upstream transcriptional regulatory region for NFAT2. The promoter sequence for NFAT gene may include the nucleic acid sequence of SEQ ID NO:781-783 or SEQ ID NO:815 as examples, but it can be seen that the optimal sequence or minimal sequence for expression of the second factor may be obtained by making fragments, extensions or mutations of the promoter and testing for the strength of the promoter with respect to expression of the second factor. In one aspect of the invention, the transcriptional regulatory region for NFAT2 is engineered upstream of the gene encoding the second factor, which if for cleavage enzyme MMP9 (SEQ ID NO:647) or the catalytic sub-unit of MMP9 (SEQ ID NO: 648). In one aspect of the invention, the NFAT is NFATc3 and the promoter sequence of NFATc3 includes nucleic acid sequences from SEQ ID NO:816. In one aspect of the invention, the transcriptional regulatory region for NFATc3 is engineered upstream of the gene encoding the second factor, here as an example is MMP9. In another aspect of the invention, the NFAT is NFATc2. SEQ ID NO:817-818 shows nucleic acid sequence of the upstream transcriptional regulatory region for NFATc2. In one aspect of the invention, the transcriptional regulatory region for NFATc2 is engineered upstream of the gene encoding the second factor, which may be cleavage enzyme MMP9 (SEQ ID NO:647) or the catalytic sub-unit of MMP9 (SEQ ID NO: 648).

[0558] Another method for having the expression of the second factor induced when the T cell or CAR T cell is activated is to have the gene for the second factor on an inducible promoter where the NFAT protein itself binds to and induces transcription of the second factor. In this case, an NFAT response element (NFAT RE) may be positioned upstream of the gene for the second factor or fragment of the second factor. The NFAT may bind to its responsive element upstream of the second factor alone or as part of a complex. The NFAT protein may be NFATc1, NFATc2, NFATc3, NFATc4, or NFAT5. In a preferred embodiment, the NFAT protein is NFAT2 aka NFATc1, aka NFATc. The gene of the second factor or fragment thereof is cloned downstream of an NFAT-response element (SEQ ID NO:649), which may be repeats of the response element (SEQ ID NO:650) and CMV minimal promoter (mCMV) (SEQ ID NO:651) to induce expression of second factor by NFAT protein. The NFAT response element may include nucleic acid sequence of NFAT consensus sequence (SEQ ID NO:804). The NFAT response element may include the nucleic acid sequence of SEQ ID NOS: 805-814 as examples, but it can be seen that the optimal sequence or minimal sequence for expression of the second factor may be obtained by making fragments, extensions or mutations of the responsive element nucleic acid and testing for the strength of the responsive element with respect to expression of the second factor. The enhancer region of Foxp3 also contains NFAT response elements within the 120-bp from 2079 to 2098 (SEQ ID NO:821). The NFAT response element may include nucleic acid NFAT consensus sequence of (5′-cattttttccat-3′) (SEQ ID NO:819) or (5′-tttttcca-3′) (SEQ ID NO: 820), which NFATc1 specifically binds to (Xu et al., Closely related T-memory stem cells correlate with in vivo expansion of CAR. CD19-T cells and are preserved by IL-7 and IL-15, Blood 2014 123:3750-3759), or repeats thereof. The NFAT response elements may also be separated by nucleic acid spacer sequences. Other NFAT responsive elements may exist and may further be discovered, and a skilled artisan in the art when directed to determine NFAT responsive element may do so by carrying out molecular biological assays to obtain it given the guidance of at least the responsive elements as set forth as SEQ ID NOS: 804-814 albeit as only mere examples. In one aspect of the invention, the cleavage enzyme that is downstream of the NFAT-response element and CMV minimal promoter is MMP9 (SEQ ID NO:652). In another aspect of the invention, the cleavage enzyme is a catalytic sub-unit of MMP9 (SEQ ID NO:653).

[0559] Because NFATs 1˜4 are regulated by the calcineurin pathway, potential toxicities that may arise in a patient can be stopped by treatment with an immunosuppressive agent such as FK506, Cyclosporin, Cyclosporin A, or Tacrolimus that block calcineurin activity and inhibit NFAT translocation to the nucleus. The T cell transduced or transfected with a cleavage enzyme on an inducible promoter may also be transfected or transduced with a CAR that recognizes a protein or molecule on the cancer cell. In a specific example, the cleavage enzyme is one that is able to cleave MUC1 full-length and the CAR bears an antibody fragment that directs it to MUC1* on the surface of cancer cells.

[0560] To determine which cleavage enzymes cleave MUC1 on cancer cells, we tested a series of MMP and ADAM enzyme inhibitors. These experiments pointed to MMP9 as being an important cleavage enzyme in cancer cells. To confirm that MMP9 cleaves MUC1 on cancer cells, we transfected HCT-116 MUC1 negative colon cancer cells with a mimic of full-length MUC1 having 41 tandem repeat domains: HCT-MUC1-41TR. Through single cell cloning we were able to establish this cell line wherein MUC1 only minimally gets cleaved to MUC1*. FIGS. 36A-36D show Western blots and FACS analysis showing that HCT-MUC1-41TR is 95% positive for full-length MUC1 and only 5-10% positive for the cleaved form, MUC1*. HCT-MUC1-41TR cells were incubated with MMP9 at varying concentrations and then assayed by immunofluorescence to measure binding of MNC2 monoclonal antibody to the resultant cells. As can be seen in FIGS. 37A-37C binding of MNC2 increased as the concentration of MMP9 added to the cells increased. These experiments show that MMP9 cleaves MUC1 to a form that is recognized by MNC2. The human cancer tissue array studies we performed (FIG. 30A-30F, FIG. 31A-31F, FIG. 32A-32F, FIG. 33A-33F) show that MNC2 recognizes the form of cleaved MUC1 that is present on cancerous tissue but not on healthy cells or tissues (FIG. 34A-34I). Importantly, MNC2 does not recognize the form of cleaved MUC1 that is expressed on healthy hematopoietic stem cells of the bone marrow (FIGS. 39-41).

[0561] In one aspect of the invention, an immune cell is transduced with both a CAR to target the immune cell to the tumor, and a cleavage enzyme. The CAR and the cleavage enzyme can be encoded on the same plasmid or on two different plasmids. In one aspect, the cleavage enzyme is on an inducible promoter. In another aspect, expression of the cleavage enzyme is induced by a protein that is expressed when the immune cell is activated. In one case, expression of the cleavage enzyme is induced by an NFAT protein. In another aspect, expression of the cleavage enzyme is induced by NFATc1. In another aspect, expression of the cleavage enzyme is induced when one of the NFAT proteins binds to an NFAT response element that is inserted upstream of the gene for the cleavage enzyme or a catalytically active fragment thereof. In one aspect, the cleavage enzyme is MMP9 or a fragment of MMP9 that is catalytically active.

[0562] In one aspect of the invention, the cleavage enzyme is MMP9 (SEQ ID NO:643). Some cleavage enzymes are naturally expressed as pro-enzymes that need to be activated. This can be accomplished by biochemical means, by expressing a co-enzyme that activates a cleavage enzyme or by engineering the enzyme in an activated form. The invention anticipates overcoming this problem by co-expressing the cleavage enzyme with its activator. In one aspect of the invention, the cleavage enzyme is MMP9 and the co-activator is MMP3. In another aspect of the invention, the cleavage enzyme is expressed in a form that is already active, for example by expressing a fragment of the cleavage enzyme that still has catalytic function. In one case, the cleavage enzyme is an MMP9 fragment that is catalytically active. One example of an MMP9 catalytic fragment is given as SEQ ID NO:645.

[0563] MMP9, which must be activated by MMP3, is overexpressed in a large percentage of solid tumors. Further, it is known that MNC2 anti-MUC1* monoclonal antibody recognizes MUC1 after it is cleaved by MMP9. The various breast, ovarian, pancreatic and lung cancer tissue arrays that were shown in FIGS. 30-33 were probed with MNC2-scFv, further indicating that MUC1 in these cancers is being cleaved by MMP9. To see if cleavage of tumors by MMP9 would increase T cell access to the tumor, we did a series of experiments using a cell line that expresses full-length MUC1, HCT-MUC1-41TR, a breast cancer cell line that is a high expresser of both full-length MUC1 and MUC1* and a MUC1 negative cell line that we transfect with MUC1*45. We transfected cells with MMP9 and MMP3, which activates MMP9. We took the supernatant of those cells, which contained activated MMP9, and added it to the various cells, which were then co-cultured with T cells transduced with an anti-MUC1* CAR: huMNC2-CAR44. The result was greatly increased CAR T cell killing of the targeted MUC1/MUC1* positive cancer cells, compared to the control cells that were not incubated with a MUC1 cleavage enzyme.

[0564] APMA is a biochemical that activates MMPs. We used APMA along with the conditioned media of cells that we transfected with either MMP9 or ADAM17 to see if any of these cleavage enzymes would cleave MUC1 on the HCT-MUC1-41TR cell line that only expresses full-length MUC1. As controls, we also tested the enzymes on HCT-MUC1* cells. The MUC1 and MUC1* expressing cells were stained with a red dye, CMTMR. Human T cells that were transduced with an anti-MUC1* CARs, CAR44 or CAR50 were co-cultured with the cancer cells. Untransduced T cells were used as a control (FIG. 45A-45P). As can be seen in FIG. 45B, FIG. 45C, and FIG. 45D, the anti-MUC1* CAR T cells effectively recognized and clustered the HCT-MUC1* cancer cells, which is a sign of T cell activation and killing. However, no CAR T cell induced clustering is visible in the wells containing HCT-MUC1-41TR, the full-length MUC1 expressing cells (FIG. 45F, FIG. 45G, and FIG. 45H). However, the cells that were incubated with activated MMP9 show dramatic increase in CAR T cell induced clustering (FIG. 45J, FIG. 45K, and FIG. 45L), indicating that MMP9 cleaved the full-length MUC1 to a form of MUC1* that is recognized by MNC2 monoclonal antibody and more specifically by huMNC2-scFv. ADAM17 had no apparent effect. ADAM17 either did not cleave MUC1 or cleaved it at a position that is not recognized by MNC2, which is more likely (FIG. 45N-45P).

[0565] We performed the same experiment, this time using T47D breast cancer cells that were hard to kill using anti-MUC1* CAR T cells presumably because they express high levels of full-length MUC1 as well as MUC1* (FIG. 46A-46T). As can be seen in FIGS. 46B, 46C, and 46D, anti-MUC1* CAR44 and CAR50 have little effect on the T47D cancer cells. Only in FIG. 46D, which is CAR44 at the highest level of CAR expression in the T cells, do we see a small amount of CAR T cell induced clustering. However, the presence of activated MMP2 (FIG. 46J, 46K, 46L) or activated MMP9 (FIG. 46R, 46S, 46T) shows a dramatic increase in CAR T cell recognition, clustering and killing, showing that cleavage of full-length MUC1 increases T cell access to the cancer cells. To ensure that the addition of the APMA was not inducing cleavage or anti-MUC1* CAR T recognition by some other mechanism, we made a catalytically active form of MMP9 and added it to T47D cells that were then co-cultured with MNC2-CAR44 T cells (FIG. 47A-47I). As can be seen in the figure, MNC2-CAR T cells recognize and cluster cells transfected with MUC1* (FIG. 47B-47C), poorly cluster T47D breast cancer cells that express both full-length MUC1 and MUC1* (FIG. 47E-47F), but robustly bind to and cluster the T47D cells after the addition of the catalytically active MMP9 (FIG. 47H-47I). This results supports the claim that MNC2 does not recognize full-length MUC1 but does recognize the growth factor receptor MUC1*. Note that the full-length MUC1 expressed on this cell line may sterically hinder the binding of CAR T cells near the cell membrane.

[0566] In another example, T47D MUC1 positive tumor cells were incubated with a recombinant catalytic domain of MMP9 (Enzo Life Sciences, Inc., Farmingdale, NY) at either 100 ng/mL or 500 ng/mL. Western blot analysis showed that the MUC1/MUC1* positive cancer cells underwent extensive cleavage of MUC1 to MUC1*. In another example, T47D breast cancer cells were pre-incubated with a human recombinant MMP9 catalytic domain protein then co-cultured with anti-MUC1* CAR44 T cells. The specific killing of the T47D cells by CAR44 T cells was monitored in real-time on an xCelligence instrument that measures impedance as a function of time. This analysis uses electrode arrays upon which cancer cells are plated. The adherent cancer cells insulate the electrode and cause an increase in impedance as they grow. Conversely, T cells are not adherent and remain in suspension so do not increase or decrease impedance. However, if the T cells or CAR T cells kill the cancer cells on the electrode plate, the cancer cells ball up and float as they die, which causes the impedance to decrease. The addition of MMP9 catalytic domain dramatically increased the killing of T47D cancer cells. FIG. 48 shows an xCelligence graph of T47D breast cancer cells in co-culture with either untransduced T cells, as a control, or huMNC2-CAR44 T cells over a 45 hour period. After 18 hours of cancer cell growth, a catalytic sub-unit MMP9 was added to some of the cells. At 25 hours, T cells were added. As can be seen, huMNC2-CAR44 T cell killing is greatly improved when the T47D cells are pre-incubated with cleavage enzyme MMP9. In the xCelligence system, target cancer cells, which are adherent, are plated onto electrode array plates. Adherent cells insulate the electrode and increase the impedance. The number of adherent cancer cells is directly proportional to impedance. T cells are not adherent and do not contribute to impedance. Therefore, increasing impedance reflects growth of cancer cells and decreasing impedance reflects killing of cancer cells. Prostate cancer cell line DU145 expresses both MUC1 and MUC1* but at a much lower level of expression than T47D cells. DU145 cells are efficiently killed by anti-MUC1* CAR T cells in the presence or absence of a cleavage enzyme.

[0567] FIG. 49 shows an xCelligence graph of DU145 prostate cancer cells in co-culture with either untransduced T cells, as a control, or huMNC2-CAR44 T cells over a 45 hour period. After 18 hours of cancer cell growth, a catalytic sub-unit MMP9 was added to some of the cells. At 25 hours, T cells were added. As can be seen, huMNC2-CAR44 T cell killing of low density MUC1/MUC1* positive cancer cells is not affected by pre-incubation with cleavage enzyme MMP9. DU145 cancer cells express a significantly lower amount of MUC1 which includes the full-length form as well as MUC1*. The lower density of full-length MUC1 does not sterically hinder T cell access to the membrane proximal MUC1*. DU145 cells represent an early stage cancer that expresses both full length and cleaved MUC1 but at lower levels so that T cell access is not sterically hindered. T47D cells represent mid-stage cancers that express high levels of both MUC1 and MUC1*, wherein the density of MUC1 full-length sterically hinders access of T cells to the tumor. HCT-MUC1* cells are a MUC1 negative cell line that has been stably transfected with MUC1*45, and they represent late stage cancer cells. It is significant that MUC1 cleaved to MUC1* by MMP9 is recognized by the anti-MUC1* antibody MNC2, which is the targeting head of the CAR. Immune cell access to tumor antigens on the cancer cell surface can be sterically hindered by the presence of bulky extra cellular domain proteins or other obstructing elements also known as the tumor micro-environment. The aforementioned serve as an example that can be extended to improve the efficacy of CAR T therapies that target other tumor antigens. In one aspect of the invention, an immune cell is transfected or transduced with both a CAR comprising an antibody fragment that targets a tumor antigen and a cleavage enzyme. In another aspect of the invention, an immune cell is transfected or transduced with both a CAR comprising an antibody fragment that targets a tumor antigen and a cleavage enzyme that cleaves a tumor antigen to a form recognized by the antibody fragment of the CAR. In one aspect, an immune cell is transfected or transduced with both a CAR comprising an antibody fragment that targets a tumor antigen and a cleavage enzyme that cleaves a tumor antigen to a form recognized by the antibody fragment of the CAR, wherein the antibody fragment of the CAR recognizes MUC1* extra cellular domain and the cleavage enzyme cleaves MUC1 to MUC1*. In one aspect, an immune cell, which may be a T cell or an NK cell, is transfected or transduced with a CAR comprising an antibody fragment derived from MNC2, MNE6, MNC3 or MNC8 and a cleavage enzyme chosen from the group comprising MMP1, MMP2, MMP3, MMP7, MMP8, MMP9, MMP11, MMP12, MMP13, MMP14, MMP16, ADAM9, ADAM10, ADAM17, ADAM 19, ADAMTS16, ADAM28 or a catalytically active fragment thereof.

[0568] In one aspect of the invention, a person diagnosed with cancer or at risk of developing cancer is administered a sufficient amount of an immune cell transduced with both a CAR and a cleavage enzyme. In another aspect of the invention, a person diagnosed with cancer or at risk of developing cancer is administered a sufficient amount of an immune cell transduced with both a CAR and a cleavage enzyme, wherein the cleavage enzyme is on an inducible promoter that is activated by proteins that are expressed when the immune cell becomes activated. In another aspect of the invention, a person diagnosed with cancer or at risk of developing cancer is administered a sufficient amount of an immune cell transduced with both a CAR and a cleavage enzyme, wherein the cleavage enzyme is on an inducible promoter that is activated by one or more NFAT. In one case the NFAT is NFATc1. In another aspect, the NFAT is NFATc3. In another aspect, the NFAT is NFATc2. In any of the instances above, the extra cellular domain of the CAR comprises a fragment of an anti-MUC1* antibody. In one aspect, the anti-MUC1* antibody is MNC2scFv or a humanized form of MNC2scFv. In another aspect, the anti-MUC1* antibody is MNE6scFv or a humanized form of MNE6scFv. In any of the instances above, the immune cell can be a T cell, an NK cell, a mast cell, or a dendritic cell.

[0569] It is not intended that the present invention be limited to one or two specific methods of having expression of a cleavage enzyme induced by an activated T cell. We have demonstrated specific expression of a cleavage enzyme only upon T cell activation by constructing a plasmid with the cleavage enzyme gene downstream of an NFAT promoter sequence or downstream of one or more repeats of NFAT response elements. In another aspect of the invention, expression of the cleavage enzyme is induced by constructing a plasmid where the cleavage enzyme gene is inserted downstream of an IL-2 promoter sequence or downstream of an IL-2 response element, then inserting the plasmid into an immune cell. In another aspect of the invention, expression of the cleavage enzyme is induced by constructing a plasmid where the cleavage enzyme gene is inserted downstream of a Calcineurin promoter sequence or downstream of a Calcineurin response element, then inserting the plasmid into an immune cell and then administering to a patient for the treatment or prevention of cancers. There are also drug-inducible plasmids that can be used to induce expression of the cleavage enzyme or used to stop expression induced by an element of an activated T cell. These drug inducible systems may include tetracycline-inducible systems, Tet-on, Tet-off, tetracycline response elements, doxycycline, tamoxifen inducible systems, ecdysone inducible systems and the like.

[0570] It is not intended that the present invention be limited to one or two specific promoters used in the plasmids encoding the CARs or inducible cleavage enzymes. As is known by those skilled in the art, many promoters can be interchanged including SV40, PGK1, Ubc, CAG, TRE, UAS, Ac5, polyhedron, CaMKIIa, GAL1, GAL10, TEF1, GDS, ADH1, CaMV35S, Ubi, H1 and U6. Another solution to the problem of steric hindrance of CAR T cell access, caused by bulky cell surface proteins such as MUC1-FL, is to increase the length of the linker region of the CAR that is expressed by the T cell. In standard design CARs, the length of the extracellular linker region between the transmembrane portion and the antibody fragment is about 45-50 amino acids in length. We made long-arm CARs where the length of the extracellular linker is extended from about 50 amino acids to 217-290 amino acids. Co-culture assays show that CARs with longer extracellular linkers have improved access to the tumor-associated antigen on the target cancer cells.

BiTEs

[0571] Divalent (or bivalent) single-chain variable fragments (di-scFvs, bi-scFvs) can be engineered by linking two scFvs. This can be done by producing a single peptide chain with two V_H and two V_L regions, yielding tandem scFvs. Another possibility is the creation of scFvs with linker peptides that are too short for the two variable regions to fold together (about five amino acids), forcing scFvs to dimerize. This type is known as diabodies. Diabodies have been shown to have dissociation constants up to 40-fold lower than corresponding scFvs, meaning that they have a much higher affinity to their target. Consequently, diabody drugs could be dosed much lower than other therapeutic antibodies and are capable of highly specific targeting of tumors in vivo. Still shorter linkers (one or two amino acids) lead to the formation of trimers, so-called triabodies or tribodies. Tetrabodies have also been produced. They exhibit an even higher affinity to their targets than diabodies.

[0572] All of these formats can be composed from variable fragments with specificity for two different antigens, in which case they are types of bispecific antibodies. The furthest developed of these are bispecific tandem di-scFvs, known as bi-specific T-cell engagers (BiTE antibody constructs). BiTEs are fusion proteins consisting of two scFvs of different antibodies, on a single peptide chain of about 55 kilodaltons. One of the scFvs may bind to T cells such as via the CD3 receptor, and the other to a tumor cell via a tumor specific molecule, such aberrantly expressed MUC1*.

[0573] Another aspect of the invention is a method for treating a patient diagnosed with, suspected of having, or at risk of developing a MUC1 positive or MUC1* positive cancer, wherein the patient is administered an effective amount of a BiTE wherein one antibody variable fragment of the BiTE binds to a T cell surface antigen and the other antibody variable fragment of the BiTE binds to PSMGFR (SEQ ID NO:2), or more specifically to N-10 peptide. In one case, the antibody variable fragment of the BiTE that binds to MUC1* comprises portions of MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11.

[0574] In another aspect of the invention, MUC1* peptides including PSMGFR (SEQ ID NO: 2), or most or all of N-10 peptide are used in adoptive T cell approaches. In this case, a patient's T cells are exposed to the MUC1* peptides and through various rounds of maturation, the T cells develop MUC1* specific receptors. The adapted T cells are then expanded and administered to the donor patient who is diagnosed with, suspected of having, or is at risk of developing a MUC1* positive cancer.

[0575] A series of CARs were also made that had MNC2 and humanized MNC2 as the extra cellular, targeting head of the CAR. The constructs for these CARs were inserted into a plasmid that was then inserted into a Lenti viral vector. Human T cells were then transduced with the lenti viral vector carrying the MNC2 CARs and huMNC2 CARs. MNC2-scFv-CARs that were mouse sequence or humanized were generated. In one aspect of the invention, the CAR comprised huMNC2-scFv-short hinge region-transmembrane domain derived from CD8-short intracellular piece-4-1BB-3zeta. In another aspect, the transmembrane domain was derived from CD4 transmembrane sequence. In another aspect, the intracellular co-stimulatory domain was CD28-3zeta. In yet another aspect, the intracellular co-stimulatory domain was CD28-4-1BB-3zeta.

[0576] There are a variety of methods for assessing whether or not T cells recognize a target cell and are in the process of mounting an immune response. T cells cluster when they recognize a target or foreign cell. This can be readily seen with the naked eye or at low magnification. The appearance of CAR T cell clustering when co-cultured with target cancer cells is one measure of: a) whether or not they recognize the cells as target cells; and b) whether or not they are getting activated to attack the targeted cells, which in this case are cancer cells. FIGS. 45-47 show photographs of MUC1* positive T47D breast cancer cells that were either stably transfected with mCherry or dyed with CMTMR, so are red, which were co-cultured with either human T cells without a CAR or human T cells transduced with huMNC2-scFv-CAR44, or with huMNC2-scFv-CAR50. The CAR T cells are clear. As can be seen, there is no T cell induced clustering of the cancer cells when the T cell does not carry a CAR. However, when T cells carrying a MUC1* targeting CAR, there is dramatic clustering of the MUC1* positive cancer cells.

[0577] After T cells recognize and cluster target cells, they overexpress perforin and granzyme B. Together these two molecules activate a cell death pathway in the targeted cell. It is thought that the perforin makes a hole in the target cell into which the T cell injects granzyme B which then activates apoptotic proteases, causing the target cell to lyse. FIG. 55 and FIG. 56 show huMNC2-scFV-CAR44 T cells binding to target MUC1* positive prostate cancer and pancreatic cancer cells and injecting granzyme B.

[0578] Another measure of whether or not a T cell has recognized a target cell and is activated to kill that cell, is the upregulation and secretion of cytokines, interferon gamma (IFN-g) and interleukin-2 (IL-2), by the T cell. Activation of CAR T cells, as evidenced by IFN-g and IL-2 secretion, can be readily measured in vitro. CAR T cells are co-cultured with target cells and after an incubation period, the conditioned media is assayed by ELISA to detect secreted IFN-g and IL-2. In order to determine the cancer-specificity of CAR T cells wherein the targeting head of the CAR was either huMNC2 or huMNE6, these experiments were performed with huMNC2-CAR44 T cells and huMNE6-CAR44 T cells in co-culture with MUC1* positive cancer cells and normal cells. Table 1 details the MUC1 positive normal or primary cells that were tested.

[0579]

TABLE 1
Normal Cell Lines and Primary Cells
	ATCC
Cell Line	Designation	Tissue	Origin
Hep.G2		Liver
THLE-3	CRL-11233	Liver	The THLE-2 (ATCC CRL-10149 and the
			THLE-3 (ATCC CRL-11233) cell lines
			were derived from primary normal liver
			cells by infection with SV40 large T
			antigen. THLE-2 and THLE-3 cells express
			phenotypic characteristics of normal adult
			liver epithelial cells. They are
			nontumorigenic when injected into athymic
			nude mice, have near-diploid karyotypes,
			and do not express alpha-fetoprotein.
Lonza	HUM181141	Liver	Male, Caucasian
Primary			2.0 months old
Hepatocytes			Induction Fold CYP1A2 (a) 14.0
			Induction Fold CYP2B6 (b) 13.0
			Induction Fold CYP3A4 (c) 44.0
			Basal Activity CYP1A2 2.6
			Basal Activity CYP2B6 0.7
			Basal Activity CYP3A4 14.0
			Additional Information:
			Inducer/Marker Metabolite
			(a) 0.05 mM Omeprazole/Acetaminophen
			(b) 1 mM Phenobarbital/
			Hydroxybupropion
			(c) 0.01 mM Rifampicin/6-Beta-
			Hydroxytestosterone
			Basal activity is expressed as: pmol/million
			cells/minute
T/G HA-	CRL-1999	Aortic Smooth	11 months
VSMC		Muscle	Female, Caucasian
CCD-18Lu	CCL-205	Lung	This fibroblast-like cell line was derived
			from the lung tissue of a 2 month, 17-day-
			old Black female.
			The donor had cerebral anoxia, cardiac
			anomaly, sepsis, endocardial cushion defect
			and fetal alcoholic syndrome.
			Female, Black
			2.5 months
HBEC-5i	CRL-3245	Brain	Derived from small fragments of human
		endothelium	cerebral cortex obtained from patients who
			had died of various causes.
Hs	CRL-7869	Stomach/Intestine	18 weeks gestation fetus
738.St/Int			Male, Caucasian
			Part of the NBL Cell Line Collection. This
			cell line is neither produced nor fully
			characterized by ATCC. We do not
			guarantee that it will maintain a specific
			morphology, purity, or any other property
			upon passage.
MCF-12A	CRL-10782	Breast	The MCF-12A cell line is a non-
			tumorigenic epithelial cell line established
			from tissue taken at reduction mammoplasty
			from a nulliparous patient with fibrocystic
			breast disease that contained focal areas of
			intraductal hyperplasia. The line was
			produced by long term culture in serum free
			medium with low Ca++ concentration.
			MCF-12A was derived from adherent cells
			in the population.
Hs 1.Tes	CRL-7002	Testis	Male, Caucasian
			second trimester
			Part of the NBL Cell Line Collection. This
			cell line is neither produced nor fully
			characterized by ATCC. We do not
			guarantee that it will maintain a specific
			morphology, purity, or any other property
			upon passage.
HRCE	Lonza:	Kidney	Human Renal Cortical Cells (HRCE) are
	catalogue		from proximal and distal tubules.
	#CC-2554		Donor info: 49 year old female, passage 2,
	Lot		95% viability, doubling time (hours) 24 hrs
	#0000542104

[0580] FIG. 50 is a graph of PCR measurement of the various cell lines tested, wherein mRNA levels of MUC1 are measured. The cancer cell lines that were tested in these assays were HCT-MUC1* and T47D breast cancer cells. These cells were co-cultured with huMNC2-CAR44 human T cells. Co-culture of huMNC2-CAR44 T cells with the cancer cells induced the CAR T cells to secrete large amounts of IFN-g and IL-2 into the surrounding media, yet co-culture with the MUC1 positive normal cells induced no secretion of the cytokines (FIG. 51 and FIG. 52). In addition to testing for IFN-g and IL-2 secretion by the CAR T cells, the normal cells were assayed for signs of cell death, which could have been induced by the CAR T cells if the antibody targeting head were not extremely cancer-specific. After co-culture with huMNC2-CAR44 T cells, the cells were incubated with a cell death marker, then assayed by FACS. huMNC2-CAR44 T cells induced no cell death in the normal cells (FIG. 53A-53J).

[0581] In addition to FACS analysis, many researchers now use an xCELLigence instrument to measure CAR T killing of cancer cells. FACS is not the best method for tracking T cell induced cell killing because the T cells lyse the target cell. By FACS it is difficult to measure dead cells because they are excluded as cell debris, so one must infer an amount of cell killing and by various methods determine if the missing cells are T cells or cancer cells.

[0582] The xCELLigence instrument uses electrode arrays upon which cancer cells are plated. The adherent cancer cells insulate the electrode and so cause an increase in impedance as they grow. Conversely, T cells are not adherent and remain in suspension so do not contribute to insulation of the electrode which would increase impedance. However, if the T cells or CAR T cells kill the cancer cells on the electrode plate, the cancer cells ball up and float off as they die, which causes the impedance to decrease. The xCELLigence instrument measures impedance as a function of time, which is correlated to cancer cell killing. In addition, the electrode plates also have a viewing window. When CAR T cells effectively kill the adsorbed target cancer cells, there is a decrease in impedance but also one can see that there are no cancer cells left on the plate surface.

[0583] FIGS. 55A-55H show the cytotoxic effect of huMNC2-CAR44 T cells on MUC1* positive DU145 prostate cancer cells as measured by a variety of assays. FIG. 55A is a fluorescent photograph of untransduced T cells co-cultured with the prostate cancer cells, wherein granzyme B is stained with a red fluorophore. FIG. 55C is a fluorescent photograph of huMNC2-CAR44 T cells co-cultured with the prostate cancer cells, wherein granzyme B is stained with a red fluorophore. FIG. 55D is the DAPI and granzyme B merge. FIG. 55E is a FACS scan for fluorescently labeled granzyme B for untransduced T cells incubated with the cancer cells. FIG. 55F is a FACS scan showing a positive increase in fluorescently labeled granzyme B for huMNC2-CAR44 T cells incubated with the cancer cells. FIG. 55G is a graph of the mean fluorescent intensity. FIG. 55H is an xCELLigence scan tracking the real-time killing of DU145 cancer cells by huMNC2-CAR44 T cells (blue trace) but not by untransduced T cells (green). FIGS. 56A-56H show the cytotoxic effect of huMNC2-CAR44 T cells on MUC1* positive CAPAN-2 pancreatic cancer cells as measured by a variety of assays. FIG. 56A is a fluorescent photograph of untransduced T cells co-cultured with the pancreatic cancer cells, wherein granzyme B is stained with a red fluorophore. FIG. 56B is the DAPI and granzyme B merge. FIG. 56C is a fluorescent photograph of huMNC2-CAR44 T cells co-cultured with the pancreatic cancer cells, wherein granzyme B is stained with a red fluorophore. FIG. 56D is the DAPI and granzyme B merge. FIG. 56E is a FACS scan for fluorescently labeled granzyme B for untransduced T cells incubated with the cancer cells. FIG. 56F is a FACS scan showing a positive increase in fluorescently labeled granzyme B for huMNC2-CAR44 T cells incubated with the cancer cells. FIG. 56G is a graph of the mean fluorescent intensity. FIG. 56H is an xCELLigence scan tracking the real-time killing of CAPAN-2 cancer cells by huMNC2-CAR44 T cells (blue trace) but not by untransduced T cells (green). FIGS. 57A-57C show xCELLigence scans tracking the real-time killing of MUC1* positive cancer cells, but not MUC1* negative cells, by huMNC2-CAR44 T cells. FIG. 57A shows that huMNC2-CAR44 T cells effectively kill HCT colon cancer cells that have been stably transfected with MUC1*. FIG. 57B shows that huMNC2-CAR44 T cells have almost no effect on HCT-MUC1-41TR, which is a MUC1 negative cancer cell that has been stably transfected with a MUC1 full-length. In this cell line only about 10% of the cell have MUC1 cleaved to MUC1*. FIG. 57C shows that huMNC2-CAR44 T cells have no effect on HCT-116 cells, which is a MUC1 negative colon cancer cell line.

[0584] These data demonstrate that T cells transduced with a CAR wherein the antibody fragment targeting head is MNC2, effectively kill MUC1* positive cancer cells. These data specifically show that huMNC2-scFV-CAR44 transduced into human T cells effectively kill MUC1* positive cancer cells. Because we and others have now demonstrated that the most important aspect of CAR T function is the targeting antibody fragment, it follows that an immune cell or a T cell transduced with any CAR having the antibody fragment MNC2-scFV or huMNC2-scFV would have similar efficacy against MUC1 or MUC1* positive tumors. For example, the hinge region that connects the scFv to the transmembrane portion could be any flexible linker. The intracellular co-stimulatory domains could be CD28-3zeta, CD28-4-1BB-3zeta or any combination of immune cell co-stimulatory domains.

[0585] FIG. 61 shows an experiment in which huMNC2-scFv-CAR44 transduced human T cell that were bead stimulated (Protocol 1) or cancer cell stimulated (Protocol 2) were tested for their ability to inhibit tumor growth in animals. Human cancer cells that had been stably transfected with Luciferase were injected into female NOD/SCID/GAMMA (NSG) mice between 11 and 15 weeks of age. 500,000 BT-20 breast cancer cells were injected sub-cutaneously into a rear flank. Tumor engraftment was verified by injecting the animals with Luciferin and then imaging the fluorescent cancer cells using an IVIS instrument. IVIS images taken Day 5 post implantation showed the presence of tumor cells. On Day 6 after IVIS measurement, animals were given a one-time injection of 10 million of either human T cells transduced with huMNC2-scFv-CAR44 or untransduced T cells. 5 million T cells were injected intra-tumor and 5 million were injected into the tail vein. 10 minutes prior to IVIS photographs, mice were IP injected with Luciferin, which fluoresces after cleavage by Luciferase, thus making tumor cells fluoresce. FIGS. 61A, 61D, 61G show photographs of mice that were treated with huMNC2-scFv-CAR44 T cells that had been pre-stimulated by co-culturing for 24 hours with 4 μm beads to which was attached a synthetic MUC1*, PSMGFR peptide 24 hours prior to administration, “Protocol 1”. FIGS. 61B, 61E, 61H show photographs of mice that were treated with huMNC2-scFv-CAR44 T cells that had been pre-stimulated by twice co-culturing for 24 hours with MUC1* positive cancer cells 24 hours prior to administration, “Protocol 2”. As can be seen in FIG. 61, huMNC2-CAR44 T cells that were peptide-bead stimulated inhibited tumor growth better than cells pre-stimulated by incubation with live cancer cells, which likely contaminated the target cells and increased the tumor volume.

[0586] huMNC2-scFv-CAR44 transduced human T cell that were bead stimulated (Protocol 1) or cancer cell stimulated (Protocol 2) were also tested for their ability to inhibit tumor growth in animals. Human cancer cells that had been stably transfected with Luciferase were injected into female NOD/SCID/GAMMA (NSG) mice between 11 and 15 weeks of age. In another experiment, 500,000 BT-20 MUC1* positive triple negative breast cancer cells were injected sub-cutaneously into a rear flank. Tumor engraftment was verified by injecting the animals with Luciferin and then imaging the fluorescent cancer cells using an IVIS instrument. IVIS images taken Day 6 post implantation showed the presence of tumor cells. On Day 6, after IVIS imaging, 10M huMNC2-scFv-CAR44 T cells were administered to the animals. 5M of the CAR T cells were administered by intratumor injection and the other 5M were administered by tail vein injection. Control group was injected by same administration routes with the same number of untransduced T cells. IVIS measurements of tumor burden were taken on Days 6, 8, and 12. As can be seen in FIGS. 61A-61J, both groups of mice treated with huMNC2-CAR44 T cells showed a decrease in tumor burden compared to the control group.

[0587] huMNC2-scFv-CAR44 transduced human T cell that were bead stimulated (Protocol 1) were also tested for their ability to inhibit ovarian cancer growth in animals. Human SKOV-3 MUC1* positive ovarian cancer cells that had been stably transfected with Luciferase were injected into female NOD/SCID/GAMMA (NSG) mice between 11 and 15 weeks of age. In one experiment, 500,000 SKOV-3 cancer cells were injected into the intraperitoneal cavity to mimic metastatic ovarian cancer in humans. Tumor engraftment was verified by injecting the animals with Luciferin and then imaging the fluorescent cancer cells using an IVIS instrument. IVIS images taken Day 3 post implantation showed the presence of tumor cells. On Day 4 and Day 11, post tumor implantation, 10M huMNC2-scFv-CAR44 T cells were IP administered to the animals. On Day 4, CAR T cells were IP injected. On Day 11 half the CAR T cells were injected into the intraperitoneal space and the other half was injected into the tail vein. Control groups were injected by same administration routes with either the same number of untransduced T cells or same volume of PBS. Subsequent IVIS measurements of tumor burden were taken on Day 7, Day 10 and Day 15. As can be seen in FIGS. 62A-62L, control mice have tumors that are growing at a much faster rate than the huMNC2-CAR44 T cell treated mice. FIG. 62M shows the IVIS color bar correlating photons/second to color.

[0588] One aspect of the invention is a method for treating a patient diagnosed with, suspected of having, or at risk of developing a MUC1 positive or MUC1* positive cancer, wherein the patient is administered an effective amount of immune cells that have been transduced with a MUC1* targeting CAR, wherein the CAR is chosen from among the group consisting of MN-E6-CD8-CD28-3z (SEQ ID NOS: 297-298); MN-E6-CD4-CD28-3z (SEQ ID NOS: 748-749); MN-E6-CD8-41BB-3z (SEQ ID NOS: 300-301); MN-E6-CD4-41BB-3z (SEQ ID NOS: 750-751); MN-E6-CD8-CD28-41BB-3z (SEQ ID NOS: 303-304); MN-E6-CD4-CD28-41BB-3z (SEQ ID NOS: 754-755); MN-E6scFv-Fc-8-41BB-CD3z (SEQ ID NOS: 310-311); MN-E6scFv-IgD-Fc-8-41BB-CD3z (SEQ ID NOS: 770-771); MN-E6scFv-FcH-8-41BB-CD3z (SEQ ID NOS: 315-316); MN-E6scFv-IgD-FcH-8-41BB-CD3z (SEQ ID NOS: 772-773); MN-E6scFv-Fc-4-41BB-CD3z (SEQ ID NOS: 318-319); MN-E6scFv-FcH-4-41BB-CD3z (SEQ ID NOS: 321-322); MN-E6scFv-IgD-8-41BB-CD3z (SEQ ID NOS: 323-324); MN-E6scFv-IgD-4-41BB-CD3z (SEQ ID NOS: 327-328); MN-E6scFv-X4-8-41BB-CD3z (SEQ ID NOS: 330-331); MN-E6scFv-X4-4-41BB-CD3z (SEQ ID NOS: 333-334); MN-E6scFv-8-4-41BB-CD3z (SEQ ID NOS: 336-337), or any of the aforementioned CARs wherein the MN-E6 is replaced by fragment derived from MNC2, MNE6, 20A10, 3C2B1, 5C6F3, 25E6, 18G12, 28F9, 1E4, B12, B2, B7, B9, 8C7F3, or H11. Another aspect of the invention is a method for treating a patient diagnosed with, suspected of having, or at risk of developing a cancer, wherein the patient is administered an effective amount of immune cells that have been transduced with one of the aforementioned CARs wherein the MN-E6 is replaced by a peptide comprising antibody variable domain fragments that are specific for a cancer antigen. In any of the above methods, the immune cell may be a T cell and may further be isolated from the patient to be treated.

Other MUC1 Cleavage Sites

[0589] It is known that MUC1 is cleaved to the growth factor receptor form, MUC1*, on some healthy cells in addition to cancer cells. For example, MUC1 is cleaved to MUC1* on healthy stem and progenitor cells. A large percentage of bone marrow cells are MUC1* positive. Portions of the intestine are MUC1* positive.

[0590] The inventors have discovered that MUC1 can be cleaved at different positions that are relatively close to each other but the location of cleavage changes the fold of the remaining portion of the extracellular domain. As a result, monoclonal antibodies can be identified that bind to MUC1* cleaved at a first position but do not bind to MUC1* that has been cleaved at a second position. This discovery is disclosed in WO2014/028668, filed Aug. 14, 2013, the contents of which are incorporated by reference herein its entirety. We identified a set of anti-MUC1* monoclonal antibodies that bind to MUC1* as it appears on cancer cells but do not bind to MUC1* as it appears on stem and progenitor cells. Conversely, we identified a second set of monoclonal antibodies that bind to stem and progenitor cells but do not bind to cancer cells. One method used to identify stem specific antibodies is as follows: supernatants from monoclonal hybridomas were separately adsorbed onto 2 multi-well plates. Stem cells, which are non-adherent cells, were put into one plate and cancer cells which are adherent were put into an identical plate. After an incubation period, the plates were rinsed and inverted. If the non-adherent stem cells stuck to the plate, then the monoclonal antibody in that particular well recognizes stem cells and will not recognize cancer cells. Antibodies that did not capture stem cells or antibodies that captured cancer cells were identified as cancer specific antibodies. FACS analysis has confirmed this method works.

[0591] Antibodies MN-E6 and MN-C2 are examples of cancer-specific antibodies. Antibodies MN-C3 and MN-C8 are examples of stem-specific antibodies. Although both sets of antibodies are able to bind to a peptide having the PSMGFR sequence, FACS analysis shows that the anti-MUC1* polyclonal antibody and MN-C3 bind to MUC1* positive bone marrow cells but MN-E6 does not. The MUC1* polyclonal antibody was generated by immunizing a rabbit with the PSMGFR peptide. Similarly, MN-C3 binds to stem cells of the intestinal crypts but MN-E6 does not. Conversely, MN-E6 antibody binds to cancerous tissue while the stem-specific MN-C3 does not. Competition ELISA experiments indicate that the C-terminal 10 amino acids of the PSMGFR peptide are required for MN-E6 and MN-C2 binding, but not for MN-C3 and MN-C8. Therefore, another method for identifying antibodies that are cancer specific is to immunize with a peptide having the sequence of the PSMGFR peptide minus the 10 N-terminal amino acids or use that peptide to screen for antibodies or antibody fragments that will be cancer specific. Antibodies that bind to a peptide with a sequence of PSMGFR peptide minus the N-terminal 10 amino acids, referred to herein as N-10 peptide, but do not bind to a peptide with a sequence of PSMGFR peptide minus the C-terminal 10 amino acids, C-10 peptide, are cancer specific antibodies for use in the treatment or prevention of cancers.

[0592] The extracellular domain of MUC1 is also cleaved on stem cells and some progenitor cells, where activation of cleaved MUC1 by ligands NME1 in dimer form or NME7 promotes growth and pluripotency and inhibits differentiation. The transmembrane portion of MUC1 that remains after cleavage is called MUC1* and the extracellular domain is comprised essentially of the Primary Sequence of MUC1 Growth Factor Receptor (PSMGFR) sequence. However, the exact site of cleavage can vary depending on cell type, tissue type, or which cleavage enzyme a particular person expresses or overexpresses. In addition to the cleavage site that we previously identified which leaves the transmembrane portion of MUC1* comprising most or all of the PSMGFR (SEQ ID NO:2), other cleavage sites could possibly result in an extended MUC1* comprised of most or all of SNIKFRPGSVVVQLTLAFREGTINVHDVETQFNQYKTEAASRY (SEQ ID NO:620); or SVVVQLTLAFREGTINVHDVETQFNQYKTEAASRY (SEQ ID NO:621).

[0593] To test this hypothesis, and to determine if antibodies to an N-terminally extended PSMGFR, would generated more cancer-specific antibodies than antibodies that bind to the PSMGFR, we generated monoclonal antibodies by immunization with peptides:

[0594]

	(PSMGFR)
	(SEQ ID NO: 2)
	GTINVHDVETQFNQYKTEAASRYNLTISDVSVSDVPFPFSAQSGA,
	(N+20/C−27)
	(SEQ ID NO: 822)
	SNIKFRPGSVVVQLTLAFREGTINVHDVETQFNQYKTE,
	or
	(N+9/C−9)
	(SEQ ID NO: 824)
	VQLTLAFREGTINVHDVETQFNQYKTEAASRYNLTISDVSVSDVP

[0595] Monoclonal antibodies generated from immunization with the same peptide can also show differences in reactivity to the same cancerous tissue specimen. These results indicate that the monoclonal antibodies recognize different conformations of the truncated MUC1 extra cellular domain produced by immunizing with different length peptides, mimicking different cleavage sites, or from cleavage at different sites in the host animal. Antibodies that recognize different cleavage site conformations may be cancer sub-type specific or patient specific, depending on which cleavage enzyme their tumor expresses. In one aspect of the invention, a patient diagnosed with a certain type of cancer is treated with an antibody of the invention that recognizes a cleaved MUC1 wherein the antibody is specific for cleavage by a specific enzyme that is known to be typically expressed by that sub-type of cancer. In another aspect, a patient tumor is analyzed to determine which enzyme his or her tumor expresses and an antibody that recognizes a MUC1 cleaved by that enzyme is then administered to the patient for the treatment of their cancer. The antibody may be in the form of a CAR, a BiTE, an ADC, or a bispecific antibody.

[0596] We previously reported that it is the MUC1 transmembrane cleavage product, called MUC1* (muk1 star), that mediates tumor growth and not full-length MUC1 (Mahanta et al 2008). MUC1* is a growth factor receptor that is activated by ligand induced dimerization of its short extra cellular domain (FIG. 1A). Dimerization of the MUC1* extra cellular domain activates the MAP kinase signaling cascade and stimulates growth and survival of cancer cells (Fessler et al 2009). Bivalent antibodies that dimerize the MUC1* extra cellular domain stimulate cancer cell growth while the monovalent Fab of the same antibody, which cannot dimerize, inhibits cancer cell growth. We demonstrated this in vitro (FIG. 1B) and in vivo (FIG. 7A-7B).

[0597] We then identified the natural ligands that dimerize and activate MUC1* growth factor receptor function. Dimers of NME1 bind to and dimerize the MUC1* extra cellular domain and stimulate growth (FIG. 1C and Smagghe et al 2013). NME1 can turn its growth factor properties off. NME1 is secreted by MUC1* positive cells. Dimeric NME1 binds to MUC1* to stimulate growth. However, as the cell population grows, more and more NME1 is secreted from the cells. At high concentrations, the NME1 dimers multimerize and form hexamers, which do not bind to MUC1*, but likely bind to some unknown receptor, as the addition of NME1 hexamers turns off growth. NME1 is an adult form. The embryonic form is NME7_AB (Carter et al 2016). Each NME7_AB monomer has two binding sites for MUC1* so as a monomer it dimerizes MUC1* (FIG. 1D), stimulates growth and cannot turn itself off. In the developing embryo, BRD4 turns off NME7 and its co-factor JMJD6 turns on the self-regulating form, NME1. However, in cancers, NME7, which should be silenced in adult life, is aberrantly expressed again, where is renders the MUC1* growth factor receptor constitutively active.

[0598] In vitro, NME1 (SEQ ID NO:4) and NME7_AB (SEQ ID NO:827) bind to the PSMGFR portion of the MUC1* extra cellular domain. Both growth factors can bind to the PSMGFR peptide (SEQ ID NO:2) even if the 10 N-terminal amino acids are deleted, referred to herein as N-10 (SEQ ID NO:3). However, neither NME1 nor NME7_AB can bind to the PSMGFR peptide if the 10 membrane proximal amino acids are deleted (FIG. 2A-2D), referred to herein as C-10 (SEQ ID NO:825). In summary, the epitope to which NME1 and NME7_AB bind includes all or part of the 10 membrane proximal amino acids: PFPFSAQSGA (SEQ ID NO:1743). We tested various antibodies that were generated in animals by immunizing with the PSMGFR peptide for their ability to recognize cancer cells but not healthy cells. Among the most cancer selective were the MNC2 and MNE6 monoclonal anti-MUC1* antibodies. Two other monoclonal antibodies that were generated from immunizing animals with the PSMGFR peptide are MNC3 and MNC8. Although MNC2, MNE6, MNC3 and MNC8 all bind to the PSMGFR peptide, like NME1 and NME7_AB, MNC2 and MNE6 bind strongly to the N-10 peptide but not to the C-10 peptide (FIG. 2B-2C). In fact, MNC2 and MNE6 competitively inhibit the binding of NME1 and NME7_AB to PSMGFR (FIG. 3A-3C). Conversely, MNC3 and MNC8 bind to the C-10 peptide, bind less well to the N-10 peptide and do not compete with NME1 nor NME7_AB for binding to MUC1* peptides, including PSMGFR (FIG. 2E-2F). MNC3 and MNC8 are far less cancer specific than MNC2 and MNE6. MNC3 and MNC8 recognize stem and progenitor cells, such as hematopoietic stem cells, whereas MNC2 and MNE6 do not (FIG. 39-41). Because hematopoietic stem cells are the progenitor cells for the blood cells, it would be problematic to have a cancer therapeutic that would also target such an important normal cell type.

[0599] Because MUC1* is generated by enzymatic cleavage of MUC1, we researched which cleavage enzymes cleave MUC1 to a MUC1* and whether or not we could identify antibodies that would recognize a MUC1* generated by a first cleavage enzyme but not MUC1* generated by a second cleavage enzyme. We found that MNC2 and MNE6 recognized a MUC1* generated by cleavage of MUC1 by MMP9 but not by cleavage by other enzymes such as MMP2 (FIG. 37 and FIG. 75). We note that MMP9 is overexpressed in cancers and is a predictor of poor prognosis (vant Veer et al 2002; Dufour et al 2011) and has been implicated in metastasis (Owyong et al, 2019), whereas MMP2 is expressed in bone marrow. One antibody binding to a MUC1* generated by cleavage by a first enzyme but not by cleavage by a second enzyme implies that the antibody recognizes a conformational epitope rather than a linear epitope.

[0600] We reasoned that the most cancer specific antibodies would be those antibodies that are characterized by some combination of most or all of the following:

• • [0601] Antibody binds to PSMGFR peptide;
  • [0602] Antibody does not bind to full-length MUC1;
  • [0603] Antibody binds to N-10;
  • [0604] Antibody does not bind to C-10;
  • [0605] Antibody competitively inhibits binding of NME1 or NME7_AB to MUC1* extra cellular domain or a PSMGFR peptide;
  • [0606] Antibody recognizes a MUC1* generated by cleavage by MMP9;
  • [0607] Antibody recognizes a conformational epitope not a linear epitope.
    MNC2 and MNE6 are Cancer Specific.

[0608] Our experiments show that both MNC2 and MNE6: a) Bind to tumor cells; b) monovalent forms block tumor growth in vitro and in vivo; c) have minimal to no binding of normal tissue while having robust binding to a wide panel of tumor tissues; d) when incorporated into CAR T cells, MNC2 and MNE6 directed CAR T cells do not recognize full-length MUC1 and do not kill cells that only express full-length MUC1; e) MNC2 and MNE6 directed CAR T cells cluster then kill tumor cells expressing MUC1*; and f) MNC2 and MNE6 recognize a MUC1 cleavage product when it is cleaved by MMP9.

[0609] MNC2 directed CAR T cells do not recognize normal, healthy cells that are MUC1* positive. A panel of normal cell lines, as well as primary cells, were co-cultured with huMNC2-CAR44 T cells. The normal cell populations were analyzed to determine whether or not the MNC2 directed CAR T cells killed them. The CAR T cells were analyzed to see if co-culture with the MUC1 positive normal cells activated the killing function of the CAR T cells, as measured by secretion of IL-2 or interferon gamma. As FIGS. 50-52 show, the MNC2 directed CAR T cells did not kill the normal cells, nor was there cytokine secretion, indicative of T cell activation. In addition, over 2,000 human tissue specimens were analyzed. The results showed that neither MNC2 nor MNE6 showed any significant binding to normal tissues but showed robust staining of a wide panel of cancerous tissues. For example, MNC2 stained 93% breast cancer specimens, 83% ovarian, 78% pancreatic and 71% lung cancer specimens. In addition, patient-matched primary tumors (FIG. 54) and subsequent metastases showed that the amount of MNC2-reactive MUC1* increased with tumor progression and metastases. In summary, MNC2 is a highly cancer specific antibody.

Characterization of MNC2 and MNE6

[0610] Our gold standard, cancer-specific antibodies MNC2 and MNE6: 1) bind to N-10 peptide but not to the C-10 peptide; 2) compete with NME7_AB and dimeric NME1 for the same binding site near the C-terminus of the PSMGFR peptide, which is the membrane proximal portion of MUC1* on cells; 3) do not work in a Western blot assay indicating that they recognize a conformational rather than linear epitope; 4) recognize a MUC1* generated when MUC1 is cleaved by MMP9; 5) do not bind to full-length MUC1 but only to the cleaved form, MUC1*, in model cell lines as well as cancer cell lines; 6) show little to no binding to normal tissues but robustly stain a wide variety of tumor tissues; and 7) share some consensus sequences in their Complementarity Determining Regions, CDRs.

[0611] In an effort to identify other antibodies that are highly cancer-specific, like MNC2 and MNE6, we subjected new antibodies to a set of seven (7) characterization experiments: 1) epitope binding assays; 2) functional assays such as the ability to displace activating growth factor NME7_AB or dimeric NME1 from binding to MUC1* peptides PSMGFR or N-10; 3) Western blots to determine whether or not the antibodies recognized a linear epitope versus a conformational epitope, in which case the antibodies would not work in a Western; 4) binding assays to see if the antibodies recognized a cleaved MUC1 that was dependent on cleavage by MMP9; 5) FACS analysis to measure the ability of the antibodies to recognize MUC1* positive cells but not full-length MUC1; and FACS analysis to measure the ability of the antibodies to recognize MUC1/MUC1* positive cancer cells; 6) immunohistochemistry, IHC, assays of normal tissues versus cancerous tissues to determine true cancer specificity; and 7) aligning antibody sequences to determine if subsets of antibodies shared consensus sequences that could predict their cancer specificity or lack thereof.

[0612] Monoclonal antibodies were produced by immunizing animals with peptides derived from a MUC1 that is devoid of tandem repeats. These antibodies included PSMGFR and peptides that were extended at the N-terminus of PSMGFR. Immunizing peptides were:

[0613] PSMGFR (SEQ ID NO:2))

[0614] N+9/C-9 (9 amino acids added onto the N-terminus and 9 amino acids deleted from the C-terminus) (SEQ ID NO:824)

[0615] N+20/C-27 (20 amino acids added onto the N-terminus and 27 amino acids deleted from the C-terminus) (SEQ ID NO:823)

[0616] These monoclonal antibodies were then tested to determine which satisfied the seven (7) characterization criteria cited above, which we reasoned would identify the most cancer specific antibodies.

Epitope Binding Assays

[0617] ELISA assays were performed to determine if, in addition to recognizing their immunizing peptide, they recognized PSMGFR, N-10 or C-10. In addition, they were tested for their ability to bind to N+20/C-27, N+9/C-9. We first did the ELISA assay on our set of reference antibodies, MNC2, MNE6, which we know are cancer-specific plus MNC3, which we know recognizes stem cells and progenitor cells (FIG. 63A-63B). None of the reference antibodies bound to the N+20/C-27 peptide. MNC2 and MNE6 cannot bind to PSMGFR peptides with 27, 10 or 9 C-terminal deletions, however, MNC3 binds to C-10 and to N+9/C-9 peptides.

[0618] This same ELISA assay was performed on the antibodies of the invention (FIG. 64-66 and FIG. 201). The binding patterns of the antibodies that were generated by immunizing with the PSMGFR peptide are shown in FIG. 64A-64B. Note that only 20A10 exactly matches the binding profile of MNC2 and MNE6. 25E6, 28F9 and 18G12 are all able to bind to the N-10 peptide. 18B4 is the only antibody raised against the PSMGFR peptide that requires the 10 most N-terminal amino acids of the peptide. The color of the bars for each antibody in the ELISA graph are color coded to match the deductive cognate sequence, or a portion thereof, of that antibody. In addition, another set of antibodies was assayed by ELISA (FIG. 201). Of this set, B12, B2, B7, B9, 8C7F3, and H11 bound to the PSMGFR peptide, bound to the N-10 peptide, but not to the C-10 peptide (FIG. 201). The binding patterns of the antibodies that were generated by immunizing with the N+20/C-27 peptide are shown in FIG. 65A-65B. Although these antibodies were raised against the N+20/C-27 peptide, all but one, 45C11, still bind to the PSMGFR peptide, albeit at the N-terminal portion of PSMGFR. The binding of 45C11 is weak but deductive reasoning shows that all or some of the cognate epitope must lie within SNIKFRPGSVV (SEQ ID NO: 1744).

[0619] Of the antibodies generated by immunizing with the N+9/C-9 peptide, 8A9 and 17H6 do not bind to the PSMGFR peptide, so must bind to the 9 additional N-terminal amino acids. Antibodies 3C5 and 39H5 appear to bind to the 10 most N-terminal amino acids of the PSMGFR peptide.

[0620] In order to further refine the epitopes to which each antibody binds, a series of smaller peptides derived from the PSMGFR sequence were synthesized: N-30 (SEQ ID NO:7), N-26 (SEQ ID NO:6), N-19 (SEQ ID NO:4), N-10/C-5 (SEQ ID NO:8), N-19/C-5 (SEQ ID NO: 9). Each of the antibodies was tested in an ELISA assay for their ability to bind to this refined set of peptides, plus PSMGFR, N-10 and C-10 peptides (FIG. 67-69).

[0621] In FIG. 67A-67D, antibodies generated by immunization with the PSMGFR peptide were assayed. As can be seen in the figure, amino acids ASRYNLT (SEQ ID NO:1745), which are essentially in the middle of the PSMGFR peptide, are important or essential for the binding of 28F9, 18G12, 25E6, and MNC3 antibodies. Amino acids GTINVHDVET (SEQ ID NO:1746), which comprise the most N-terminal part of the PSMGFR peptide are important or essential for the binding of the 18B4 antibody. Amino acids FPFS (SEQ ID NO:1747) are important or essential for the binding of 20A10, MNC2 and MNE6. We note that these three antibodies recognize a conformational epitope, not a linear epitope. Because the proline in the FPFS (SEQ ID NO: 1747) sequence significantly alters the conformation of nearby portions of the PSMGFR peptide, it is also possible that the antibodies do not bind directly to these four amino acids, but that the absence of the proline alters the fold of the remaining peptide such that the conformation to which 20A10, MNC2 and MNE6 bind, is no longer present.

[0622] In FIG. 68A-68D, antibodies generated by immunization with the N+20/C-27 peptide were assayed. As can be seen in the figure, amino acids GTINVHDVET (SEQ ID NO: 1746), which comprise the most N-terminal part of the PSMGFR peptide are important or essential for the binding of the 29H1, 32C1, and 31A1 antibodies. Amino acids SNIKFRPGSVVVQLTLAFRE (SEQ ID NO:1748), which is 20 additional amino acids N-terminal to the PSMGFR peptide and outside of the PSMGFR peptide, are important or essential for the binding of antibody 45C11. However, referring back to FIG. 65, antibody 45C11 was not able to bind to the N+9/C-9 peptide, therefore we conclude that amino acids within the SNIKFRPGSVV sequence (SEQ ID NO: 1744) are essential for the binding of 45C11. Amino acids QFNQYKTEA (SEQ ID NO:1749), which are still within the sequence of PSMGFR, are important or essential for the binding of antibody 1E4.

[0623] In FIG. 69A-69D, antibodies generated by immunization with the N+9/C-9 peptide were assayed. As can be seen in the figure, amino acids GTINVHDVET (SEQ ID NO: 1746), which comprise the most N-terminal part of the PSMGFR peptide are important or essential for the binding of the 39H5 and 3C5 antibodies. As can be seen in the figure, amino acids VOLTLAFRE (SEQ ID NO:1750), which is 9 additional amino acids N-terminal to the PSMGFR peptide and outside of the PSMGFR peptide, are important or essential for the binding of antibodies 17H6 and 8A9. Because the 17H6 and 8A9 antibodies do not bind to any of the smaller peptides shown in this figure, refer to FIG. 66A-66C, which shows that these two antibodies only bind to the peptide that has 9 additional amino acids N-terminal to the PSMGFR peptide.

[0624] Table 2 below lists antibodies of the invention and their cognate epitopes.

[0625]

TABLE 2
Immunizing	Antibody
Peptide	Name	Cognate Sequence
PSMGFR	MNC2	FPFS (SEQ ID NO: 1747) or
		PFPFSAQSGA (SEQ ID NO: 1743)
	MNE6	FPFS (SEQ ID NO: 1747) or
		PFPFSAQSGA (SEQ ID NO: 1743)
	20A10	FPFS (SEQ ID NO: 1747) or
		PFPFSAQSGA (SEQ ID NO: 1743)
	3C2B1	FPFS (SEQ ID NO: 1747) or
		PFPFSAQSGA (SEQ ID NO: 1743)
	5C6F3	SVSDV (SEQ ID NO: 1751)
	MNC3	ASRYNLT (SEQ ID NO: 1745)
	25E6	ASRYNLT (SEQ ID NO: 1745)
	28F9	ASRYNLT (SEQ ID NO: 1745)
	18G12	ASRYNLT (SEQ ID NO: 1745)
	18B4	GTINVHDVET (SEQ ID NO: 1746)
N+20/C-27	45C11	SNIKFRPGSVV (SEQ ID NO: 1744)
	29H1	GTINVHDVET (SEQ ID NO: 1746)
	32C1	GTINVHDVET (SEQ ID NO: 1746)
	31A1	GTINVHDVET (SEQ ID NO: 1746)
	1E4	QFNQYKTEA (SEQ ID NO: 1749)
N+9/C-9	17H6	VQLTLAFRE (SEQ ID NO: 1750)
	8A9	VQLTLAFRE (SEQ ID NO: 1750)
	39H5	GTINVHDVET (SEQ ID NO: 1746)
	3C5	GTINVHDVET (SEQ ID NO: 1746)

Ability to Displace NME7_AB Binding to the MUC1* Extra Cellular Domain Peptide Psmgfr

[0626] We previously reported that dimeric NME1 dimerizes MUC1* extra cellular domain and stimulates growth. Monomeric NME7_AB has two binding sites for MUC1* so that as a monomer it dimerizes MUC1* and mediates cancer cell growth. We showed that NME1 and NME7_AB can bind to the MUC1* extra cellular domain. In vitro, NME1 and NME7_AB bind to the PSMGFR peptide even if the 10 N-terminal amino acids are deleted, referred to herein as N-10 (SEQ ID NO:3). However, neither NME1 nor NME7_AB can bind to the PSMGFR peptide if the 10 membrane proximal amino acids are deleted, referred to herein as C-10 (SEQ ID NO:825). In summary, the epitope to which NME1 and NME7_AB bind includes all or part of the 10 membrane proximal amino acids: PFPFSAQSGA (SEQ ID NO:1743). We tested various antibodies that were generated in animals by immunizing with the PSMGFR peptide for their ability to recognize cancer cells but not healthy cells. Among the most cancer selective were the MNC2 and MNE6 monoclonal anti-MUC1* antibodies. Two other monoclonal antibodies that were generated from immunizing animals with the PSMGFR peptide are MNC3 and MNC8. Although MNC2, MNE6, MNC3 and MNC8 all bind to the PSMGFR peptide, like NME1 and NME7_AB, MNC2 and MNE6 bind strongly to the N-10 peptide but not to the C-10 peptide. In fact, MNC2 and MNE6 competitively inhibit the binding of NME1 and NME7_AB to PSMGFR. Conversely, MNC3 and MNC8 are able to bind to the C-10 peptide, bind less well to the N-10 peptide and do not compete with NME1 nor NME7_AB for binding to MUC1* peptides, including PSMGFR (FIG. 70). MNC3 and MNC8 are less cancer specific than MNC2 and MNE6. MNC3 and MNC8 recognize stem and progenitor cells, such as hematopoietic stem cells, whereas MNC2 and MNE6 do not. Because hematopoietic stem cells are the progenitor cells for the blood cells, it would be problematic to have a cancer therapeutic that would also target such an important normal cell type.

[0627] In this experiment, antibodies of the invention were tested for their ability to displace NME7_AB from binding to the PSMGFR peptide. In this experiment, a multi-well plate was coated with the PSMGFR peptide. Recombinant NME7_AB was allowed to bind to the surface-immobilized PSMGFR peptide. Wash steps followed. Various antibodies were added, followed by wash steps. The amount of NME7_AB that remained attached to the PSMGFR coated plate, after antibody competition, was measured by detecting a tag on the NME7_AB. As a control, anti-NME7_AB antibodies were also tested for their ability to displace NME7_AB from the PSMGFR. FIG. 70 shows a graph of an ELISA displacement assay. The bar graph is color coded to indicate the cognate epitope to which each antibody binds. As can be seen in the figure, the antibodies that bind to the more C-terminal portions of PSMGFR are the most potent at disrupting the binding of onco-embryonic growth factor NME7_AB to the MUC1* extra cellular domain or the PSMGFR peptide. The rank order of potency for disrupting binding of NME7_AB to PSMGFR according to their cognate epitope is as follows: FPFS (SEQ ID NO: 1747)>ASRYNLT (SEQ ID NO: 1745)>QFNQYKTEA (SEQ ID NO: 1749)>GTINVHDVET (SEQ ID NO: 1746). Antibodies that bind to epitopes outside of the PSMGFR peptide, such as 45C11, 8A9 and 17H6 did not compete with NME7_AB for binding.

Western Blot Assay to Determine Linear Versus Conformational Cognate Epitope

[0628] Antibodies were tested to determine whether they recognize a linear or a conformational epitope. Only antibodies that recognize a linear epitope work in Western blots when using denaturing gels. For comparison, known antibodies were tested for their ability to bind to HCT-116, a MUC1 negative cancer cell line, HCT-MUC1-18, which is a cleavage resistant clone of HCTs transfected with full-length MUC1, and HCTs transfected with MUC1*, wherein the extra cellular domain comprises only the PSMGFR sequence. The antibodies tested for comparison are MNC2 and MNE6, which were known to only recognize a conformational epitope, SDIX which is a polyclonal antibody raised against PSMGFR and VU4H5, which is a commercially available monoclonal antibody that recognizes the tandem repeats of full-length MUC1 (FIG. 71A-71D). As can be seen, neither MNC2 nor MNE6 recognize a MUC1 or MUC1* specific linear epitope. The SDIX polyclonal antibody recognizes HCT-MUC1* but not full-length MUC1 and VU4H5 only recognizes full-length MUC1. These same antibodies were also tested for their ability to work in Western blots of two breast cancer cell lines 1500, aka Zr-75-1, and T47D cells and show the same binding pattern (FIG. 71E-71H).

[0629] Antibodies that were raised against the PSMGFR peptide were tested the same way in Western blots (FIG. 72A-72P). As can be seen, antibodies 25E6 and 18B4 recognize linear epitopes but 20A10, 3C2B1, 5C6F3, 18G12 and 28F9 do not, indicating that they bind to a conformational epitope. Antibodies that were raised against the N+20/C-27 peptide were tested the same way in Western blots (FIG. 73A-73J). As can be seen, antibodies 31A1 and 32C1 recognize linear epitopes. Antibodies 1E4 and 45C11 may recognize a conformational epitope. Antibodies that were raised against the N+9/C-9 peptide were tested the same way in Western blots (FIG. 74A-74H). As can be seen, none of these antibodies recognize linear MUC1 or MUC1* specific epitopes. These antibodies may recognize a conformational epitope. However, an alternative interpretation is that the lack of binding in a Western blot means that they do not specifically recognize MUC1 or a MUC1 cleavage product or that the concentration used in this assay was insufficient.

Recognition of a MUC1 Cleavage Product after Cleavage by MMP9

[0630] We previously demonstrated that MNC2 recognizes a MUC1* that is generated when full-length MUC1 is cleaved by matrix metalloprotease 9, MMP9 (FIG. 37). MMP9 is expressed by tumor tissues and is a predictor of poor prognosis for breast cancers (vant Veer et al 2002; Dufour et al 2011). MMP9 has also been implicated in metastasis (Owyong et al 2019). Recall also that MNC2 competitively inhibits the binding of onco-embryonic growth factor NME7_AB to the MUC1* extra cellular domain (FIG. 3). Therefore, it follows that onco-embryonic growth factor, which activates growth and survival functions of MUC1*, also recognizes a MUC1* generated by cleavage by MMP9. It then follows that the most cancer specific antibodies are those that recognize a conformational epitope formed when MUC1 is cleaved to MUC1* by MMP9.