Membrane-bound IL-12 for cellular immunotherapy

Abstract

The present disclosure relates, inter alia, to cells expressing both a chimeric antigen receptor (CAR) and membrane-bound IL-12, and the use of such cells to target and treat cancer (e.g., solid tumors and cancers expressing CD19, CD22, BCMA, PSCA, HER2, TAG-72, and PSCA). The present disclosure also relates to cells expressing membrane-bound IL-12.

Description

[Technical field]

[0001] This application claims the benefit of U.S. Provisional Application No. 63 / 318,753, filed March 10, 2022, and U.S. Provisional Application No. 63 / 423,479, filed November 7, 2022, the entire contents of which are incorporated herein by reference. The present disclosure relates to immune cells expressing both membrane-bound IL-12 and optionally a chimeric antigen receptor (CAR). [Background technology]

[0002] Chimeric antigen receptor (CAR) engineered T cells have promoted cancer immunotherapy since their demonstrated therapeutic potential for hematological malignancies, but CAR T cells have had limited success against solid tumors. The relative lack of success of CAR T cell therapy against solid tumors is likely due to a variety of factors, including the antigenic heterogeneity of solid tumors, the difficulty of delivering CAR T cells to solid tumors, and the paucity of tumor-selective targets. Thus, there is a need for effective CAR T cell therapy against solid tumors. Summary of the Invention

[0003] The present disclosure is based, at least in part, on the discovery that T cells expressing a chimeric antigen receptor (CAR T cells) and membrane-bound IL-12 have improved function compared to otherwise identical CAR T cells that do not express membrane-bound IL-12, and that CAR T cells administered cells expressing membrane-bound IL-12 have improved function compared to CAR T cells not administered cells expressing membrane-bound IL-12. For example, administration of cells expressing membrane-bound IL-12 can enhance the function of the CAR T cells to eliminate target cells (e.g., cancer cells) that are relatively distant from the site where the CAR T cells are administered. Thus, aspects of the disclosure provide nucleic acid molecules encoding membrane-bound IL-12 with a human CD28 transmembrane domain ("mbIL-12" or "mb(28)IL-12"). In certain embodiments, the mature membrane-bound IL-12 comprises, from amino to carboxy terminus, a mature human IL-12β (p40) subunit, optionally a first peptide linker, a mature human IL-12α (p35) subunit, optionally a second peptide linker, and a human CD28 transmembrane domain. In certain embodiments, all or a portion of the cytoplasmic domain of human CD28 follows the transmembrane domain. The human IL-12β subunit may be preceded by a signal sequence. In certain embodiments, useful mbIL-12 constructs comprise, from amino to carboxy terminus, a mature human IL-12β (p40) subunit, optionally a first peptide linker, and a mature human IL-12α (p35) subunit, optionally a second peptide linker, and a human CD28 transmembrane domain. In certain embodiments, useful mbIL-12 constructs further comprise an intracellular portion (e.g., at least four (e.g., 4, 5, 6, 7, 8, 9, or 10, or no more than 4, 5, 6, 7, 8, 9, or 10 contiguous amino acids) of the cytoplasmic domain of human CD28 following the human CD28 transmembrane domain. In certain embodiments, all or a portion of the cytoplasmic domain of human CD28 follows the transmembrane domain. A signal sequence may precede the human IL-12β subunit. The p40 subunit portion has the following sequence:

[0004] [ka] (SEQ ID NO: 1). Useful p40 subunits may contain 1, 2, 3, 4, or 5 amino acid modifications (e.g., substitutions). In certain embodiments, the amino acid modifications are conservative amino acid substitutions.

[0005] The p35 subunit portion has the following sequence:

[0006] [ka] (SEQ ID NO:2). The p35 subunit portion has the following sequence:

[0007] [ka] (SEQ ID NO: 3). Useful p35 subunits may contain one, two, three, four, or five amino acid modifications (e.g., substitutions). For example, the amino acid S at position 110 of SEQ ID NO:2 (underlined above) may be modified and replaced with a polar, uncharged amino acid, in some embodiments, the amino acid may be selected from the following: cysteine, threonine, tyrosine, asparagine, and glutamine. In some embodiments, the amino acid L at position 110 of SEQ ID NO:3 (underlined above) may be modified and replaced with a non-polar amino acid, in some embodiments, the amino acid may be selected from the following: glycine, alanine, valine, isoleucine, proline, phenylalanine, methionine, tryptophan. The human CD28 transmembrane domain has the following sequence:

[0008] [ka] (SEQ ID NO: 16). The human CD28 transmembrane domain has the following sequence:

[0009] [ka] (SEQ ID NO: 17). The human CD28 transmembrane domain has the following sequence:

[0010] [ka] (SEQ ID NO: 76). The first peptide linker may, for example, have the following sequence:

[0011] [ka] (SEQ ID NO: 4). The second peptide linker can, for example, comprise or consist of the sequence GGG.

[0012] The first and second linker peptide linkers may have any suitable sequence, depending on the circumstances. For example, the first peptide linker may consist of 3-24 amino acids, 3-20 amino acids, 3-15 amino acids, 3-10 amino acids (e.g., 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 amino acids). Preferably, none of the amino acids is proline. For example, the second peptide linker may consist of 3-24 amino acids, 3-20 amino acids, 3-15 amino acids, 3-10 amino acids (e.g., 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 amino acids). Preferably, none of the amino acids is proline. Suitable second peptide linkers include glycine and / or a mixture of glycine and serine (e.g.,

[0013] [ka] (SEQ ID NO: 57), GGGG (SEQ ID NO: 58), GGGGS (SEQ ID NO: 14), GGGSGG (SEQ ID NO: 59) and GGGSGGGS (SEQ ID NO: 60). The first and second linker peptide linkers may have any suitable sequence, depending on the circumstances. For example, the first peptide linker may consist of 3 to 24 amino acids, 3 to 20 amino acids, 3 to 15 amino acids, 3 to 10 amino acids (e.g., 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 amino acids). Preferably, none of the amino acids is proline. For example, the second peptide linker may consist of 3 to 24 amino acids, 3 to 20 amino acids, 3 to 15 amino acids, 3 to 10 amino acids (e.g., 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 amino acids). Preferably, none of the amino acids is proline. Suitable second peptide linkers include glycine and / or a mixture of glycine and serine (e.g.,

[0014] [ka] (SEQ ID NO:57), GGGG (SEQ ID NO:58), GGGGS (SEQ ID NO:14), GGGSGG (SEQ ID NO:59) and GGGSGGGS (SEQ ID NO:60). In one embodiment, a useful flexible linker is 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 repeats of the sequence GGGS (SEQ ID NO:13). In one embodiment, a useful flexible linker is 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 repeats of the sequence GGGGS (SEQ ID NO:14) or SEQ ID NO:13. In one embodiment, the linker is a linker having the following sequence:

[0015] [ka] (SEQ ID NO: 12). The human CD28 cytoplasmic domain portion may comprise or consist of the sequence RSKR (SEQ ID NO:5). The portion of the human cytoplasmic domain, if present, preferably has no signaling activity.

[0016] In one embodiment, mature mb(28)IL-12 has the following sequence:

[0017] [ka] (SEQ ID NO: 6). In one embodiment, mature mb(28)IL-12 has the following sequence:

[0018] [ka] (SEQ ID NO: 7). In the mature mb(28)IL-12 sequence above, the first and second peptide linker sequences are underlined; the human CD28 transmembrane domain sequence is shown in bold, and a portion of the human CD28 cytoplasmic domain sequence is shown in italics. Useful mb(28)IL-12 may include SEQ ID NO:6 with 1, 2, 3, 4, 5 amino acid modifications, or contain 80, 85, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100% identity to any one of the sequences of SEQ ID NO:7. Useful mb(28)IL-12 may comprise SEQ ID NO:7 with 1, 2, 3, 4, 5 amino acid modifications, or may contain 80, 85, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100% identity to any one of the sequences of SEQ ID NO:7. The mature sequence may be preceded by a signal sequence suitable for directing secretion to the surface of a human cell. For example, the signal sequence may be the following sequence:

[0019] [ka] (SEQ ID NO: 10).

[0020] In one embodiment, the mbIL-12 with a signal sequence has the following sequence:

[0021] [ka] (SEQ ID NO: 8). In the above sequence, the signal sequence is underlined.

[0022] In one embodiment, the signal sequence mb(DC28)IL-12 has the following sequence:

[0023] [ka] (SEQ ID NO: 9). In the above sequence, the signal sequence is underlined. In other embodiments, the signal sequence is

[0024] [ka] (SEQ ID NO: 11).

[0025] In one embodiment, the mbIL-12 with a signal sequence has the following sequence:

[0026] [ka] (SEQ ID NO: 48). In the above sequence, the signal sequence is underlined.

[0027] In one embodiment, the mbIL-12 with a signal sequence has the following sequence:

[0028] [ka] (SEQ ID NO: 49). In the above sequence, the signal sequence is underlined.

[0029] In some embodiments, the mb(28)IL12 described herein may comprise a sequence selected from SEQ ID NOs: 6, 7, 8, 9, 48, 49, or variants thereof each having 1, 2, 3, 4, 5 amino acid modifications. In some embodiments, the mb(28)IL12 described herein may comprise 80, 85, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100% identity to any one of SEQ ID NOs: 6, 7, 8, 9, 48, and 49. In some embodiments, the mb(28)IL12 described herein may comprise the following sequence:

[0030] [ka] (SEQ ID NO: 16) is

[0031] [ka] (SEQ ID NO: 17) or

[0032] [ka] (SEQ ID NO: 76). In one embodiment, the portion of the human cytoplasmic domain, if present, preferably has no signaling activity. The nucleic acids described herein may comprise a sequence encoding a sequence selected from SEQ ID NO: 6, 7, 8, 9, 48, 49, or variants thereof having 1, 2, 3, 4, 5 amino acid modifications. In some embodiments, the nucleic acid may comprise a sequence encoding a sequence having 80, 85, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100% identity to any one of SEQ ID NOs: 6, 7, 8, 9, 48, and 49. In some embodiments, the nucleic acid may comprise a sequence encoding the following sequence:

[0033] [ka] (SEQ ID NO: 16) has the following sequence:

[0034] [ka] (SEQ ID NO: 17) or the following sequence:

[0035] [ka] (SEQ ID NO: 76).

[0036] Disclosed herein, inter alia, is a nucleic acid molecule encoding membrane-bound IL-12 (mbIL-12), comprising, from amino terminus to carboxy terminus, a mature human IL-12p40 subunit, optionally a first peptide linker, a mature human IL-12p35 subunit, optionally a second peptide linker, and a human CD28 transmembrane domain. In one embodiment, mbIL-12 further comprises at least four consecutive amino acids of the cytoplasmic domain of human CD28 following the human CD28 transmembrane domain. In one embodiment, the amino acid sequence represented by the mature human IL-12p40 subunit comprises the amino acid sequence represented by SEQ ID NO:1. In one embodiment, the amino acid sequence of mature human IL-12p35 comprises the amino acid sequence represented by SEQ ID NO:2 or 3. In one embodiment, the first peptide linker comprises the amino acid sequence represented by SEQ ID NO:4. In one embodiment, the second peptide linker comprises the amino acid sequence GGG. In one embodiment, the human CD28 transmembrane domain comprises the amino acid sequence represented by SEQ ID NO:16, 17, or 76. In one embodiment, at least four contiguous amino acids of the cytoplasmic domain of human CD28 following the human CD28 transmembrane domain comprise the amino acid sequence set forth in SEQ ID NO: 5. In one embodiment, mbIL-12 comprises the amino acid sequence set forth in SEQ ID NO: 6 or 7.

[0037] Also disclosed herein is a population of human immune cells (e.g., T cells, NK cells, dendritic cells, macrophages, or monocytes) comprising a nucleic acid encoding mb(28)IL12 as described herein. In some embodiments, the population of human immune cells comprises a nucleic acid molecule encoding membrane-bound IL-12 (mbIL-12), comprising, from amino to carboxy terminus, a mature human IL-12p40 subunit, an optional first peptide linker, a mature human IL-12p35 subunit, an optional second peptide linker, and a human CD28 transmembrane domain. In some embodiments, mbIL-12 further comprises the human CD28 transmembrane domain followed by at least a portion of the cytoplasmic domain of human CD28. In some embodiments, the amino acid sequence represented by the mature human IL-12p40 subunit comprises the amino acid sequence represented by SEQ ID NO:1. In some embodiments, the amino acid sequence of the mature human IL-12p35 comprises the amino acid sequence represented by SEQ ID NO:2 or 3. In some embodiments, the first peptide linker comprises the amino acid sequence represented by SEQ ID NO:4. In some embodiments, the second peptide linker comprises the amino acid sequence GGG. In one embodiment, the human CD28 transmembrane domain comprises the amino acid sequence represented by SEQ ID NO: 16, 17 or 76. In one embodiment, the cytoplasmic domain of the human CD28 transmembrane domain comprises at least four contiguous amino acids of the cytoplasmic domain of human CD28 following the human CD28 transmembrane domain, or comprises the amino acid sequence represented by SEQ ID NO: 5. In one embodiment, mbIL-12 comprises the amino acid sequence represented by SEQ ID NO: 6 or 7. In one embodiment, the population of immune cells is a population of T cells, a population of NK cells, a population of dendritic cells, a population of macrophages, or a population of monocytes.

[0038] In one embodiment, the population of immune cells comprises a nucleic acid comprising a sequence encoding mb(28)IL12 and comprises a nucleic acid comprising a sequence encoding a chimeric antigen receptor (CAR). In one embodiment, the chimeric antigen receptor targets a human antigen selected from the group consisting of CD19, CD22, BCMA, PSCA, HER2, and TAG-72. In one embodiment, the chimeric antigen receptor targets human TAG-72. In one embodiment, the chimeric antigen receptor comprises an scFv targeting Tag-72, a spacer, a transmembrane domain, a CD28 costimulatory domain or a 41-BB costimulatory domain, and a human CD3 zeta signaling domain. In one embodiment, the transmembrane domain is selected from a CD4 transmembrane domain, a CD8 transmembrane domain, and a CD28 transmembrane domain. In one embodiment, the TAG72 scFV is selected from IDEC, V15, V59, and V59_V15. In one embodiment, the transmembrane domain is a human CD28 transmembrane domain. In one embodiment, the spacer region comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 2 to 12. In one embodiment, the spacer comprises an IgG hinge region. In one embodiment, the spacer consists of 10 to 50 amino acids. In one embodiment, the 4-1BB costimulatory domain comprises the amino acid sequence represented by SEQ ID NO: 38. In one embodiment, the CD3ζ signaling domain comprises the amino acid sequence represented by SEQ ID NO: 35. In one embodiment, a linker of 3 to 15 amino acids is located between the 4-1BB costimulatory domain and the CD3ζ signaling domain or a variant thereof. In some cases, the mbIL-12 constructs described herein are expressed in conjunction with a CAR, e.g., a CAR that targets a solid tumor antigen, e.g., HER2 or PSCA.

[0039] Described herein, inter alia, is a population of human immune cells (e.g., T cells, NK cells, dendritic cells, macrophages, or monocytes) expressing membrane-bound IL-12 (mbIL-12) comprising, inter alia, from amino to carboxy terminus, a mature human IL-12p40 subunit, a first peptide linker, a mature human IL-12p35 subunit, a second peptide linker, and a human CD28 transmembrane domain. In certain embodiments, the mbIL-12 further comprises the human CD28 transmembrane domain followed by at least a portion of the cytoplasmic domain of human CD28. In certain embodiments, the amino acid sequence represented by the mature human IL-12p40 subunit comprises the amino acid sequence represented by SEQ ID NO:1. In certain embodiments, the amino acid sequence represented by the mature human IL-12p35 comprises the amino acid sequence represented by SEQ ID NO:2 or 3. In certain embodiments, the first peptide linker comprises the amino acid sequence represented by SEQ ID NO:4. In certain embodiments, the second peptide linker comprises the amino acid sequence GGG. In one embodiment, the human CD28 transmembrane domain comprises the amino acid sequence represented by SEQ ID NO: 16 or 17 or 76. In one embodiment, the cytoplasmic domain of the human CD28 transmembrane domain comprises at least 4 contiguous amino acids of the cytoplasmic domain of human CD28 following the human CD28 transmembrane domain, or comprises the amino acid sequence represented by SEQ ID NO: 5. In one embodiment, the mbIL-12 comprises the amino acid sequence represented by SEQ ID NO: 6 or 7. In one embodiment, the population of immune cells is a population of T cells, a population of NK cells, a population of dendritic cells, a population of macrophages, or a population of monocytes.

[0040] Also described herein are methods for treating cancer, comprising administering human immune cells comprising a nucleic acid encoding mb(28)IL12 as described herein, and human immune cells expressing a chimeric antigen receptor. In some embodiments, the immune cells comprising a nucleic acid encoding mb(28)IL12 as described herein are administered prior to administration of immune cells expressing a chimeric antigen receptor. In some embodiments, the immune cells comprising a nucleic acid encoding mb(28)IL12 as described herein are administered together with immune cells expressing a chimeric antigen receptor. In some embodiments, the cancer is a solid tumor. In some embodiments, the cancer expresses CD19, CD22, BCMA, PSCA, HER2, TAG-72, or PSCA.

[0041] Also described herein are methods for treating cancer, comprising administering human immune cells expressing mb(28)IL12 as described herein and human immune cells expressing a chimeric antigen receptor. In some embodiments, the immune cells expressing mb(28)IL12 as described herein are administered prior to administration of immune cells expressing a chimeric antigen receptor. In some embodiments, the immune cells expressing mb(28)IL12 as described herein are administered together with immune cells expressing a chimeric antigen receptor. In some embodiments, the cancer is a solid tumor. In some embodiments, the cancer expresses CD19, CD22, BCMA, PSCA, HER2, TAG-72, or PSCA. Also described herein, inter alia, is a population of human immune cells (e.g., T cells, NK cells, dendritic cells, macrophages, or monocytes) expressing membrane-bound IL-12 (mbIL-12) and a CAR targeting a cancer antigen, where mbIL-12 comprises, from amino terminus to carboxy terminus, a mature human IL-12p40 subunit, a first peptide linker, a mature human IL-12p35 subunit, a second peptide linker, a human CD28 transmembrane domain, and the CAR comprises a cancer antigen-targeting scFv, a spacer, a transmembrane domain, a CD28 costimulatory domain or a 4-1BB costimulatory domain, and a human CD3 zeta signaling domain. In one embodiment, mbIL-12 further comprises a human CD28 transmembrane domain followed by a cytoplasmic domain of human CD28. In one embodiment, the amino acid sequence represented by the mature human IL-12p40 subunit comprises the amino acid sequence represented by SEQ ID NO:1. In one embodiment, the amino acid sequence represented by the mature human IL-12p35 comprises the amino acid sequence represented by SEQ ID NO:2 or 3. In one embodiment, the first peptide linker comprises the amino acid sequence represented by SEQ ID NO: 4. In one embodiment, the second peptide linker comprises the amino acid sequence GGG. In one embodiment, the human CD28 transmembrane domain comprises the amino acid sequence represented by SEQ ID NO: 16 or 17 or 76. In one embodiment, the cytoplasmic domain of the human CD28 transmembrane domain comprises at least 4 consecutive amino acids of the cytoplasmic domain of human CD28 following the human CD28 transmembrane domain or comprises the amino acid sequence represented by SEQ ID NO: 5. In one embodiment, the mbIL-12 comprises the amino acid sequence represented by SEQ ID NO: 6 or 7. In one embodiment, the population of human immune cells is a population of human T cells or human NK cells. In one embodiment, the cancer antigen is from a solid tumor or a metastasis of a solid tumor. In one embodiment, the cancer antigen is selected from the group consisting of CD19, CD22, BCMA, PSCA, HER2, and TAG-72. In one embodiment, the transmembrane domain is selected from the CD4 transmembrane domain, the CD8 transmembrane domain, and the CD28 transmembrane domain. In one embodiment, the TAG72 scFV is selected from IDEC, V15, V59, and V59_V15.In one embodiment, the transmembrane domain is a human CD28 transmembrane domain. In one embodiment, the spacer region comprises an amino acid sequence selected from the group consisting of SEQ ID NOs: 2 to 12. In one embodiment, the spacer comprises an IgG hinge region. In one embodiment, the spacer consists of 10 to 50 amino acids. In one embodiment, the 4-1BB costimulatory domain comprises the amino acid sequence represented by SEQ ID NO: 38. In one embodiment, the CD3ζ signaling domain comprises the amino acid sequence represented by SEQ ID NO: 35. In one embodiment, a linker of 3 to 15 amino acids is located between the 4-1BB costimulatory domain and the CD3ζ signaling domain or a variant thereof.

[0042] In some embodiments, the CAR may comprise any one of SEQ ID NOs: 61, 62, 72, 73, 74, 75, or variants thereof having 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 amino acid modifications. In some embodiments, the CAR may comprise 80, 85, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100% identity to the sequence of any one of SEQ ID NOs: 61, 62, 72, 73, 74, and 75.

[0043] Also described herein are methods for generating populations of T cells (including, e.g., γ / ΔT cells), monocytes, macrophages, dendritic cells or NK cells that express an mb(28)IL-12 construct and, optionally, a CAR.

[0044] Also described herein is a vector comprising any of the nucleic acid constructs described herein. In some embodiments, the vector is a viral vector. In some embodiments, the vector is a lentiviral vector. In some embodiments, the vector comprises a sequence encoding any of the mb(28)IL-12 constructs disclosed herein. In some embodiments, the vector comprises a sequence encoding a sequence selected from SEQ ID NO: 6, 7, 8, 9, 48, 49, or a variant thereof having 1, 2, 3, 4, 5 amino acid modifications. In some embodiments, the vector comprises a sequence encoding a sequence comprising 80, 85, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, or 100% identity to any one of SEQ ID NO: 6, 7, 8, 9, 48, and 49.

[0045] Also disclosed herein is a method for treating cancer, comprising administering any of the population of human immune cells (e.g., macrophages, monocytes, T cells, or NK cells) described herein to a patient in need thereof. In some embodiments, the patient is afflicted with a solid tumor. In some embodiments, the solid tumor is metastatic. In some embodiments, the population of human immune cells (e.g., macrophages, monocytes, T cells, or NK cells) is administered intravenously. In some embodiments, the population of human immune cells (e.g., macrophages, monocytes, T cells, or NK cells) (e.g., macrophages, monocytes, T cells, or NK cells) is administered intraperitoneally to the area of ​​the solid tumor. In some embodiments, the solid tumor is ovarian cancer, and the chimeric antigen receptor targets TAG72. In some embodiments, the population of human immune cells (e.g., macrophages, monocytes, T cells, or NK cells) is allogeneic. In some embodiments, the population of human immune cells (e.g., macrophages, monocytes, T cells, or NK cells) is autologous. In some embodiments, a therapeutically effective amount of a population of immune cells (e.g., macrophages, monocytes, T cells, or NK cells) is administered to a patient. In some embodiments, the cancer is a solid tumor. In some embodiments, the cancer expresses CD19, CD22, BCMA, PSCA, HER2, TAG-72, or PSCA.

[0046] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art of the present invention. Methods and materials for use in the present invention are described herein, and other suitable methods and materials known in the art may also be used. Materials, methods, and examples are merely illustrative and are not intended to be limiting. All publications, patent applications, patents, sequences, database entries, and other references mentioned herein are incorporated by reference in their entirety. In case of conflict, the present specification, including definitions, will control. Other features and advantages of the invention will be apparent from the following detailed description and drawings, and from the claims. Hereinafter, an embodiment of the present invention will be described with reference to the drawings. [Brief description of the drawings]

[0047] [Figure 1] 1 is a graph showing TAG72 expression in ovarian tumor cell lines. Flow cytometry analysis of TAG72 expression was performed using the tumor cell lines human OVCAR8 (TAG72 negative), human OVCAR3 (TAG72 positive), human OV90 (TAG72 positive), and human OVCAR8-sTn (TAG72 positive). [Diagram 2]Graph showing in vitro and in vivo analysis of scFv-diverse TAG72-CAR T cells. (a) Diagram of lentiviral construct of TAG72-CAR containing three different humanized scFvs (based on CC49 clone) targeting TAG72, dCH2 extracellular spacer domain (dCH2), CD4 transmembrane domain (4tm), and intracellular 4-1BB costimulatory domain (BBz) followed by cytolytic domain (CD3z). Lentiviral transduced T cells were identified by expressing truncated non-signaling CD19 (CD19t) separated from the CAR sequence by a ribosomal skip sequence (T2A). (bc) Untransduced (UTD) and three TAG72-CAR T cells positively enriched for CD19t were assessed by flow cytometry for CD19t expression (b) or protein L for scFv detection (c) to detect lentiviral transduction of CAR. (d-g) Quantification of tumor cell killing (d), IFNγ production (e), T cell proliferation (f), and tumor cell killing (g) of UTD T cells and three TAG72-CAR T cells at an E:T ratio of 1:2 after 24 or 72 h of co-culture with antigen-positive and antigen-negative tumor targets as described in Materials and Methods. (h) Quantification of flux from intraperitoneal OV90 (eGFP / ffluc) tumor-bearing mice treated intraperitoneally with UTD or TAG72-CAR T cells. n=6-7 per group. [Diagram 3]Graphical representation of in vitro optimization of TAG72-CAR T cells. (a) Lentiviral constructs with humanized scFv targeting TAG72, different extracellular spacer domains (EQ, dCH2, CD8h, HL, L), transmembrane domains (CD4tm, CD8tm, CD28tm), and intracellular co-stimulatory domains (4-1BB, CD28) followed by a cytolytic domain (CD3z). Truncated non-signaling CD19 (CD19t) separated from the CAR sequence with a ribosomal skip sequence (T2A) was expressed to identify lentiviral transduced T cells. (b-d) Flow cytometry was performed for CD19t expression on untransduced (UTD) and seven different TAG72-CAR T cells positively enriched for CD19t to detect lentiviral transduction of CAR (b), fragment constant (Fc) derived spacer-containing CAR (c), or lentiviral transduction of protein L to detect scFv (d). (e-j) In vitro tumor cell killing activity of CAR T cells against tumor targets (TAG72-OVCAR8; TAG72+OVCAR3, OV90, and OVCAR8-sTn) after 72 h of co-culture at an effector:target (E:T) ratio of 1:4, in terms of expression of UTD (e), CD137 activation and PD-1 depletion (f-g), T cell proliferation by fold change by flow cytometry (h), and IFNγ and IL-2 cytokine production by ELISA (i-j). (k) TAG72-CAR T cell killing of OV90 cells measured by xCELLigence (E:T=1:20) over a 10-day period. (l) Schematic of repeated tumor cell challenge assay (top). TAG72-CAR T cells were co-cultured with OV90 cells (E:T=1:2) and re-challenged with OV90 cells every 2 days. The fold change of remaining viable tumor cells and TAG72-CAR T cells was quantified as described in Materials and Methods before each tumor cell re-challenge. [Figure 4]Graphs showing that locoregional intraperitoneal delivery of TAG72-dCH2(28tm)BBz CAR T cells reduced tumor burden and improved overall survival in vivo. (a) Bioluminescence flux imaging of intraperitoneally (ip) OVCAR3 (eGFP / ffluc) tumor-bearing female NSG mice treated intraperitoneally (ip) with UTD or the indicated TAG72-CAR T cells. (b) Quantification of flux (individual mice from each group) from treated OVCAR3 tumor-bearing mice. UTD (n=8 / group); TAG72-CAR T cells (n=10 / group). (c) Flow cytometry analysis of the frequency of human CD45+ (hCD45) and mouse CD45+ (mCD45) cells in the peritoneal cavity of tumor-bearing mice 14 days after treatment (top); human epithelial cell adhesion molecule+ (EpCAM) and tumor-associated glycoprotein-72+ (TAG72) tumor cells in the peritoneal cavity of tumor-bearing mice 62 days after treatment (bottom). (d) Quantification of TAG72-CAR T cells per μL of peripheral blood 7, 14, and 28 days after treatment. N=5 / group. [Diagram 5] Graphs and figures showing that local intraperitoneal delivery of TAG72-dCH2(28tm)BBz CAR T cells reduced tumor burden and increased overall survival in vivo. (a) Bioluminescence flux imaging of mice treated intraperitoneally with UTD or TAG72-CAR T cells. (b-c) Quantification of flux from the peritoneal cavity of mice bearing OV90(eGFP / ffluc) tumors treated intraperitoneally with UTD or TAG72-CAR T cells (b, individual mice per group; c, average). n=9 mice per group. (d) Kaplan-Meier survival of UTD, scFv-deficient and TAG72-CAR T cell treated mice. n=8-9 mice / group. (e) Quantification of TAG72-CAR T cells per μL of peripheral blood at 7, 14 and 28 days post treatment. n=5 / group. Graphs and figures [Figure 6]Graph showing in vitro safety of TAG72-CAR T cells against normal human cell lines. (a) Flow cytometry analysis of TAG72 expression on the cell surface of TAG72-negative (OVCAR8) tumor cells, TAG72-positive (OV90) tumor cells, and the indicated primary human normal cells. (b-c) Quantification of CD137 activation (b) and tumor and normal cell killing (c) by TAG72-dCH2(28tm)BBz CAR T cells compared to UTD T cells at various E:T ratios, assessed by flow cytometry after 48 h of co-culture with the indicated cells, as described in Materials and Methods. Data are from 5 independent donors. [Figure 7] Graphs showing that membrane-bound IL-12 modified TAG72-CAR T cells induce higher IFNγ, T cell proliferation, and anti-tumor activity in vitro. (a-c) Tumor cell killing of OV90 cells by various concentrations of anti-IFNγR1 blocking antibody, isotype control, and recombinant human IL-12 cytokine-added TAG72-CAR T cells (E:T=1:20) measured in xCELLigence for 10 days (a), IFNγ levels in the supernatant on day 10 were quantified by ELISA (b), and the number of remaining T cells was analyzed by flow cytometry (c). (d-h) Diagrams of two lentiviral constructs of membrane-bound IL-12 (mbIL12) containing B7.1 or CD28 transmembrane domains (d). Flow cytometric analysis of IL-12 surface expression on the indicated T cells (e). TAG72-CAR / mbIL12 T cell killing in OV90 cells (E:T=1:20) measured over 10 days in xCELLigence (f), the number of remaining T cells on day 10 was analyzed by flow cytometry (g), and IFNγ levels in the supernatant were quantified by ELISA (h). (i) Schematic of repeated tumor cell challenge assay (top). TAG72-CAR / mbIL12(CD28tm) T cells were co-cultured with OV90 cells (E:T=1:3) and re-challenged with OV90 cells every 2 days. The fold change in remaining viable tumor cells and TAG72-CAR T cells was quantified before each tumor cell re-challenge as described in Materials and Methods. [Figure 8]Figure 8. Graphs showing that TAG72-CAR T cell anti-tumor activity is regulated by IFNγ. Tumor cell killing of OVCAR3 cells by TAG72-CAR T cells (E:T=1:50) was measured using various concentrations of anti-IFNγR1 blocking antibody, isotype control, and recombinant human IL-12 cytokine in xCELLigence. [Figure 9] Supplement to Figure 6. Graph showing that inhibition of IFNγ signaling attenuated repeated tumor cell killing by TAG72-CAR T cells. Schematic of repeated tumor cell challenge assay (top). TAG72-CAR T cells were co-cultured with OV90 cells (E:T=1:2) and re-challenged with OV90 cells every 3 days. Remaining viable tumor cells and TAG72-CAR T cells were quantified before each tumor cell re-challenge as described in Materials and Methods. [Figure 10]1A-1D are graphs depicting antigen-dependent IL-12 signaling in TAG72-CAR T cells. (a) Intracellular flow cytometric analysis of phosphorylated STAT4 (pSTAT4, pY693) in response to TAG72 and / or recombinant huIL12 (10 ng / mL) at the indicated time points. (b) Quantification of pSTAT4 in (a). (c) Flow cytometric analysis of surface expression of mbIL12 on TAG72-CAR T cells stimulated with plate-bound TAG72 (100 U / mL) or control antigen (PSCA, 2.5 ug / mL). (d) Flow cytometric analysis of surface or intracellular expression of mbIL12 in TAG72-CAR T cells stimulated with different concentrations of plate-bound TAG72. (e) Flow cytometric analysis of mbIL12 surface expression of TAG72-CAR T cells stimulated with plate-bound TAG72 (100 U / mL) either in total or gated on CD137+ or CD137- populations. Inset: CD137+ expression on TAG72-CAR T cells after stimulation with different concentrations of plate-bound TAG72. (f) Intracellular flow cytometric analysis of phosphorylated STAT3 (pSTAT3, pY705) (left) and pSTAT4 (right) in TAG72-CAR and TAG72-CAR / mbIL12 T cells stimulated with various concentrations of plate-bound TAG72 or recombinant huIL12 (10 ng / mL). (g) Intracellular flow cytometric analysis of pSTAT4 in TAG72-CAR T cells co-cultured with mbIL12-transduced HT1080 (TAG72-) cells. Cells were stimulated with Immunocult CD3 / CD28 as recommended by the manufacturer. Cells were targeted to CAR T cells and assessed for pSTAT4. (h) IL-12 mRNA levels in TAG72-CAR and TAG72-CAR / mbIL12 T cells when stimulated with TAG72-OVCAR8 or TAG72+OV90 tumors overnight prior to RNA isolation and bulk RNAseq. [Figure 11]Graphs and diagrams showing that locoregional intraperitoneal delivery of TAG72-CAR / mbIL12 T cells reduced tumor burden and enhanced local and systemic CAR T cell persistence in vivo. (a) Bioluminescence flux imaging of intraperitoneal OVCAR3 (eGFP / ffluc) tumor-bearing mice treated intraperitoneally with CD19-CAR, CD19-CAR / mbIL12, TAG72-CAR or TAG72-CAR / mbIL12 T cells. (b) Quantification of flux (individual mice per group) from mice treated intraperitoneally with CD19-CAR T cells (n=6 / group), CD19-CAR / mbIL12 T cells (n=6 / group), TAG72-CAR T cells (n=7 / group) and TAG72-CAR / mbIL12 T cells (n=7 / group). (c) Flow cytometry analysis of human CD45+ (hCD45) and mouse CD45+ (mCD45) cells in the peritoneal cavity of tumor-bearing mice at 2 (top) or 4 (bottom) weeks of treatment. (d) Quantification of TAG72-CAR T cells per μL of peritoneal ascites fluid (left) at 2 and 4 weeks of treatment. n=2 per group. Quantification of TAG72-CAR T cells per μL of peripheral blood (right) at 1, 2, and 4 weeks after treatment. n=5 / group. (e) Quantification of flux (individual mice per group) from OV90 (eGFP / ffluc) tumor-bearing mice treated intraperitoneally with CD19-CAR T cells (n=5 / group), CD19-CAR / mbIL12 T cells (n=5 / group), TAG72-CAR T cells (n=9-10 / group) and TAG72-CAR / mbIL12 T cells (n=9-10 / group). Combined data are from two independent studies. (f) Quantification of TAG72-CAR T cells per μL of peritoneal ascites fluid (left) at weeks 2 and 4 of treatment. n=2 per group. Quantification of TAG72-CAR T cells per μL of peripheral blood (right) at weeks 1, 2, and 4 of treatment. n=5 / group. [Figure 12]Graphs showing that locoregional intraperitoneal delivery of TAG72-CAR / mbIL12 T cells eradicates subcutaneous tumors in dual-tumor-bearing mice: (a) Schematic of subcutaneous (sc) and ip OV90 dual-tumor model and treatment; (b) Bioluminescence flux imaging of dual-tumor-bearing mice treated intraperitoneally with TAG72-CAR or TAG72-CAR / mbIL12 T cells; (c) Quantification of flux (individual mice per group) from mice treated intraperitoneally with TAG72-CAR T cells (n=8 / group) and TAG72-CAR / mbIL12 T cells (n=8 / group); (d) Quantification of subcutaneous tumor volume (individual mice per group) from mice treated intraperitoneally with TAG72-CAR T cells; (e) Quantification of TAG72-CAR T cells per μL of peripheral blood at days 7, 14, 21, and 28 post-treatment. n=5 / group; (f) Quantification of TAG72-CAR T cells per μL of peritoneal ascites fluid 2 weeks after treatment. n=2 / group; (g) Immunohistochemistry analysis of CD3+ T cells in subcutaneous tumors 1 day after treatment. [Figure 13] Figure 1: TAG72-CAR T cells expressing mbIL-12 do not proliferate and survive in the absence of exogenous cytokines. (a-b) Quantification of T cell numbers (left) and percentage of surviving cells (right) during ex vivo culture in the absence (a) or presence (b) of exogenous cytokines as described in Materials and Methods. [Figure 14]Figure 2: mbIL-12 signaling improves in vitro function of PSCAand HER2-CAR T cells. (a-f) Quantification of tumor cells (left), T cells (middle), and IFN levels (right) in the supernatants for PSCA-CAR T cells (a-c) and HER2-CAR T cells (d-f) after 6 days of co-culture with γ-antigen positive targets (PC3-PSCA or 468-HER2) at an E:T ratio of 1:10. (g) HER2-CAR T cells with or without mbIL12 were co-cultured with 468-HER2 cells (E:T=1:10) and re-challenged with 468-HER2 cells every 3 days. Residual viable tumor cells (g) and T cells (h) were quantified by flow cytometry before each re-challenge with 468-HER2 cells and 2 or 3 days after the last re-challenge. (ik) The percentages of CD4+ (i), CD8+ (j), and CD62L+CCR7+ T cells were measured by flow cytometry from the re-challenge assay. [Figure 15] Graphs showing that the TAG72-dCH2(CD28tm)BBz CAR exhibits equivalent pSLP76 expression but attenuated pPLCy and pERK signaling compared to an otherwise identical TAG72 CAR containing a CD28 costimulatory domain (a-c). [Figure 16]Graphs showing that mbIL-12 improves in vivo local brain and systemic bone metastatic disease targeting by HER2 CAR. Locoregional intraventricular delivery of HER2-CAR / mbIL12 T cells reduces tumor burden and enhances regional and systemic CAR T cell persistence in vivo. (a) Schematic of intratibial (i.ti.) and intracranial (ic) BBM1 dual tumor model and treatment. (b) Bioluminescence flux imaging of mice bearing dual tumors left untreated (no tx) or treated by intraventricular (icv) injection of HER2-CAR or HER2-CAR / mbIL12 T cells. (c) Quantification of brain (top) or bone (bottom) flux from individual mice treated icv with HER2-CAR T cells (n≧10 / group), HER2-CAR / mbIL12 T cells (n≧10 / group), or no tx (n=4 / group). Representative flow cytometric analysis (d) and quantification (e) of HER2-CAR T cells per μL of blood at 1, 2, and 5 weeks post-treatment. n>3 / group. (f) Immunohistochemistry of CD3+ T cells in peritoneal and ic tumors 7 days after treatment. (g-h) Increase in central memory CAR T cells in peripheral blood of HER2-CAR / mbIL12 T cell-treated mice. (g) Flow cytometric analysis of the frequency of HER2-CAR and HER2-CAR / mbIL12 T cells in peripheral blood. (h) Quantification of CAR T cells per μL of peripheral blood at 35 days post-treatment. n=5-7 / group. (i) Quantification of brain (top) or bone (bottom) flux from individual mice treated icv with HER2-CAR T cells (n>10 / group), HER2-CAR / mbIL12 T cells (n>10 / group). (j) Quantification of HER2-CAR T cells per μL of blood at weeks 1 and 2 post-treatment. n≧3 / group. [Figure 17]Figure 1 shows that mbIL-12 cytokine signaling in CAR T cells appears to be safer than recombinant IL-12 injections and demonstrates potent antitumor activity. (a-c) Locoregional intraperitoneal delivery of TAG72-CAR / mbIL12 T cells safely and effectively targets ovarian cancer peritoneal metastases in an immunocompetent syngeneic mouse model. (a) Mean tumor flux (left) and percent body weight change (right) for the indicated T cell treatments relative to pre-treatment body weight (n>7 / group). (b) Percentages of lymphocytes, monocytes, and neutrophils (n>4 / group) from complete blood count analysis taken on day 7 post-treatment (top) and quantification of serum levels of ALT from mice on day 6 post-treatment (bottom). (c) ELISA quantification of IFNγ (left) and IL-12 (right) cytokines in mouse serum on days 18 and 20 (days 4 and 6 post-treatment, respectively; n>5 / group). [Figure 18] Figure 18 shows in vitro analysis of CAR T cells with various extracellular spacers, transmembrane domains and costimulatory domains. In vitro T cell proliferation in fold change compared to UTD (a), CD137 (b) and PD-1 (c) expression by flow cytometry, and IL-2 production by ELISA (d) of CAR T cells against tumor targets (TAG72-OVCAR8; TAG72+OVCAR3, OV90, and OVCAR8-sTn) after 24 hours (for ELISA) or 72 hours of co-culture at an effector:target (E:T) ratio of 1:4 to support the CAR T cell function data in Figure 1 c-d. Statistical analysis reflects a comparison between TAG72-dCH2(28 tm)28z and TAG72-dCH2(28 tm)BBz. (e) HER2-CAR T cell killing of HER2+SKOV3 cells (E:T=1:20) measured by xCELLigence over a 5-day period. Remaining T cells (f) along with expression of CD137 (g) and PD-1 (h) by flow cytometry. [Figure 19]Graphs showing that membrane-bound IL-12 modified TAG72-CAR T cells induce higher IFNγ, T cell proliferation, and anti-tumor activity in vitro. (a-b) Tumor cell killing of OV90 cells by TAG72-CAR T cells (E:T=1:20) supplemented with various concentrations of anti-IFNγR1 blocking antibody, isotype control, and recombinant human IL-12 cytokine as measured by xCELLigence over 10 days (a-b). Locoregional intraperitoneal delivery of TAG72-CAR / mbIL12 T cells reduces tumor burden and enhances local and systemic CAR T cell persistence in vivo. Mice bearing OV90 (eGFP / ffluc) tumors were treated intraperitoneally 1018 times with CD19-CAR, CD19-CAR / mbIL12, TAG72-CAR, or TAG72-CAR / mbIL12 T cells. (c) Flow cytometry analysis of TAG72-CAR T cells per μL of ascites. (d) Quantification of TAG72-CAR T cells per μL of peritoneal ascites (left) at weeks 2 and 4 post-treatment. n=2 per group. Quantification of TAG72-CAR T cells per μL of peripheral blood (right) at weeks 1, 2, and 4 post-treatment. n=5 / group. [Figure 20] Graphs depicting the in vitro generation and functional characterization of murine TAG72-CAR expressing murine mbIL12. (a) Schematic of dual expression of murine TAG72-CAR and murine mbIL12 on the surface of murine T cells. (b) Flow cytometry analysis of surface TAG72-CAR expression detected by expression of truncated murine CD19t (left) and surface murine mbIL12 detected by anti-mIL12 PE-conjugated antibody. (d-g) Functional characterization of 72-hour long-term co-culture killing assay tumor killing against ID80-mSTn tumor cells at 1:1 E:T (e), T cell numbers (f), CAR T cell CD137 (g), and post-co-culture CD8 and CD4 ratios (h) for TAG72-CAR versus TAG72-CAR / mbIL12 T cells. [Figure 21]Graphs showing that locoregional intraperitoneal delivery of TAG72-CAR / mbIL12 T cells safely and effectively targets ovarian cancer peritoneal metastases in an immunocompetent syngeneic mouse model. (a) Schematic of intraperitoneal (ip) ID8-mSTn tumor model and treatment. (b) Bioluminescence flux imaging of tumor-bearing mice treated with intraperitoneal (ip) injection of the indicated T cells. (c) Mean tumor flux and (d) percent mass change upon indicated T cell treatment relative to pre-treatment mass (n≧7 / group). (e) Spleen photographs (n=3 / group) and (f) CD3 IHC on liver harvested 7 days after treatment. (g) Quantification of serum levels of ALT (left) and AST (right) from mice 6 days after treatment (n≧4 / group). (h) Percentages of lymphocytes, monocytes, and neutrophils from complete blood count analysis collected 7 days after treatment. ELISA quantification of IFNγ (i) and IL-12 (j) cytokines in mouse serum on days 18 and 20 after tumor injection (days 4 and 6 after treatment, respectively) (n≧5 / group). (k) Systemic effect of sIL-12 and TAG72-CAR / mbIL12 T cells. H&E finger showing spleen and liver harvested from ip ID8-mSTn tumor-bearing mice treated with T cells. [Figure 22]Graphs showing that TAG72-CAR / mbIL12 T cell therapy induces TME alterations in ovarian cancer peritoneal metastases in an immunocompetent syngeneic mouse model. (a) H&E and CD3 IHC on solid tumor masses collected from the superior omental region of mice bearing ip ID8-mSTn tumors treated with the indicated T cells. (b) Flow cytometry analysis of tumor cells (CD3-CD45- double negative) in ascites. (c) Quantification of tumor cells (CD3-CD45- double negative) and (d) immune subsets (CD45+, CD3+, CD11b+, and NK+) (as cells / mL in ascites). (e) Flow cytometry analysis of the proportion of CAR T cells (CD3+CD19t+) and (f) quantification counts of CD4+ and CD8+ CAR T cells / mL in ascites. (g) Quantification of mean fluorescence intensity (MFI) of CD137+ in CAR T cells in ascites. (h) Quantification of the proportion of CD62L+CD44+ (Tcm) in both CAR+ and CART cells in ascites. (i) Quantification of myeloid cell counts (Ly6G+, Ly6C+, Ly6G- / Cdouble negative tumor associated macrophages (TAMs) and CD11c+CD 1073+ dendritic cells (DCs) as cells / mL in peritoneal ascites gated from total CD11b+103 cells. Flow cytometry analysis of proportion (j) and quantification of MFI (k) on CD103+MHC class II+ double positive DCs in peritoneal ascites. All analyses represent data collected from ascites of ID8-mSTn tumor-bearing mice at day 7 post-treatment (n=3 / group). (l) Flow cytometry gating strategy. (m) Total CD4+ and CD8+ T cells mL in peritoneal ascites. [Diagram 23]Figure 1. In vitro safety of TAG72-CAR / mbIL12 T cells against normal human cell lines. (a) Flow cytometry analysis of TAG72 expression on the cell surface of TAG72-negative (OVCAR8), TAG72-positive (OV90) tumor cells, and gastric (GAS1) and ovarian (OAS3 and 4) peritoneal ascites from the indicated patients. (b-c) TAG72-CAR T cells with or without mbIL12 were co-cultured with tumor and normal cells at an E:T ratio of 1:4 for 48 hours. (b) Quantification of CD137 expression on CAR T cells was measured by flow cytometry. (c) IFNγ levels in the supernatants were quantified by ELISA. (d) IFNγ signaling drives antitumor activity in TAG72-CAR / mbIL12 T cells in vitro. Tumor cell killing of OV90 cells by TAG72-CAR / mbIL12 T cells (E:T=1:20) supplemented with various concentrations of anti-IFNγR1 blocking antibodies or isotypes by xCELLigence over a 10-day period. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0048] Membrane-bound IL-12 Membrane-bound IL-12 contains the human CD28 transmembrane domain, which is important for providing membrane-bound IL-12 with the ability to improve CAR T cell function. Thus, mb(28)IL-12 contains, from amino to carboxy terminus, a mature human IL-12β (p40) subunit, optionally a first peptide linker, a mature human IL-12α (p35) subunit, optionally a second peptide linker, and a human CD28 transmembrane domain. In an embodiment, all or a portion of the cytoplasmic domain of human CD28 follows the transmembrane domain. A signal sequence may precede the human IL-12β subunit. The p40 subunit portion has the following sequence:

[0049] [ka] (Sequence No. 1). The first peptide linker may, for example, have the following sequence:

[0050] [ka] Represented by sequence number 4. The p35 subunit portion has the following sequence:

[0051] [ka] (Sequence No. 2). The p35 subunit portion has the following sequence:

[0052] [ka] (SEQ ID NO: 3). The second peptide linker can have, for example, the following sequence:

[0053] [ka] It is expressed as:

[0054] The human CD28 transmembrane domain has the following sequence:

[0055] [ka] Represented by (SEQ ID NO: 16). The human CD28 cytoplasmic domain portion has the following sequence:

[0056] [ka] It is represented by RSKR (sequence number 5). The human CD28 transmembrane domain containing a portion of the human CD28 cytoplasmic domain has the following sequence:

[0057] [ka] Represented by (SEQ ID NO: 76). Mature mb(28)IL-12 has the following sequence:

[0058] [ka] (SEQ ID NO: 6). Mature mb(28)IL-12 has the following sequence:

[0059] [ka] (SEQ ID NO: 7). In the above sequence, the first and second peptide linker sequences are underlined, the second peptide; human CD28 transmembrane domain sequence is shown in bold, and a portion of the human CD28 cytoplasmic domain sequence is shown in italics. The mature sequence may be preceded by a signal sequence suitable for directing secretion to the surface of human cells. For example, the signal sequence may be the following sequence:

[0060] [ka] (SEQ ID NO: 10). In one example, the signal sequence mb(CD28)IL-12 has the sequence:

[0061] [ka] (SEQ ID NO: 8). In this sequence, the signal sequence is underlined. In one example, mb(CD28)IL-12, which has a signal sequence, has the following sequence:

[0062] [ka] (SEQ ID NO: 9). In this sequence, the signal sequence is underlined. In other cases, the signal sequence has the following sequence:

[0063] [ka] (SEQ ID NO: 11) or may consist of: In one example, mbIL-12 with a signal sequence has the following sequence:

[0064] [ka] Represented by (SEQ ID NO: 48). In one example, mbIL-12 with a signal sequence has the following sequence:

[0065] [ka] Represented by (SEQ ID NO: 49). In some cases, the following sequence:

[0066] [ka] (SEQ ID NO: 16) is

[0067] [ka] (SEQ ID NO: 17) or

[0068] [ka] (SEQ ID NO: 76).

[0069] The first and second linker peptide linkers may have any suitable sequence, as the case may be. For example, the first peptide linker may consist of 3-24 amino acids, 3-20 amino acids, 3-15 amino acids, 3-10 amino acids (e.g., 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 amino acids). Preferably, none of the amino acids is proline. For example, the second peptide linker may consist of 3-24 amino acids, 3-20 amino acids, 3-15 amino acids, 3-10 amino acids (e.g., 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 amino acids). Preferably, none of the amino acids is proline. A suitable second peptide linker may consist of glycine and / or a mixture of glycine and serine (e.g.,

[0070] [ka] Includes. The portion of the human cytoplasmic domain, if present, preferably has no signaling activity. Amino acid modification refers to amino acid substitutions, insertions, and / or deletions in a protein or peptide sequence. An "amino acid substitution" or "substitution" refers to the replacement of an amino acid at a particular position in a parent peptide or protein sequence with another amino acid. Substitutions can be made to change the amino acids in the resulting protein in a non-conservative manner (i.e., by changing a codon from an amino acid belonging to a grouping of amino acids of a particular size or characteristic to an amino acid belonging to another grouping) or in a conservative manner (i.e., by changing a codon from an amino acid belonging to a grouping of amino acids of a particular size or characteristic to an amino acid belonging to the same grouping). Such conservative changes generally result in little change in the structure and function of the resulting protein. The following are examples of various groupings of amino acids: 1) amino acids with non-polar R groups: alanine, valine, leucine, isoleucine, proline, phenylalanine, tryptophan, methionine; 2) amino acids with uncharged polar R groups: glycine, serine, threonine, cysteine, tyrosine, asparagine, glutamine; 3) amino acids with charged polar R groups (negatively charged at pH 6.0): aspartic acid, glutamic acid; 4) basic amino acids (positively charged at pH 6.0): lysine, arginine, histidine (pH 6.0). Another group can be amino acids with phenyl groups: phenylalanine, tryptophan, and tyrosine.

[0071] Chimeric Antigen Receptor Chimeric antigen receptor (CAR) refers to an artificial immune cell receptor engineered to recognize and bind to surface antigens. T cells expressing CAR polypeptides are referred to as CAR T cells. CARs have the ability to redirect T cell specificity and reactivity to selected targets in a non-MHC-restricted manner. Non-MHC-restricted antigen recognition gives CAR T cells the ability to recognize antigens independent of antigen processing, thereby circumventing a major mechanism of tumor escape. CARs can also be expressed by other immune effector cells, including but not limited to natural killer CARs ("NK CARs"), to direct NK cell killing to cells expressing the CAR's target. There are different generations of CARs, each containing different components. First generation CARs link antibody-derived scFv to the CD3ζ intracellular signaling domain of the T cell receptor via a spacer region (also called hinge domain) and a transmembrane domain. Second generation CARs incorporate additional costimulatory domains (e.g., CD28, 4-BB, or ICOS) to provide costimulatory signals. Third generation CARs contain two costimulatory domains (e.g., a combination of CD27, CD28, 4-1BB, ICOS, or OX40) fused to the TCR CD3ζ chain.

[0072] (a) Extracellular binding domain The CAR described herein is a fusion protein that includes an extracellular binding domain that recognizes a tumor antigen, e.g., a solid tumor antigen. This extracellular domain is often an antibody single-chain fragment (scFv) or other antibody fragment, but can be a ligand that binds to a cell receptor. The binding domain is followed by an intracellular domain that includes a spacer, a transmembrane domain, at least one costimulatory domain, and a signaling domain of the T cell receptor (TCR) complex (e.g., CD3ζ). CARs are often fused to a signal peptide at the N-terminus for surface expression. When the binding domain is an scFv, there is a heavy chain variable region and a light chain variable region, which may be present in that order, linked together via a flexible linker of, for example, 5-25 amino acids. In certain embodiments, a useful flexible linker is 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 repeats of the sequence GGGS (SEQ ID NO: 13). In certain embodiments, a useful flexible linker is 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 repeats of the sequence GGGGS (SEQ ID NO: 14) or SEQ ID NO: 13. In certain embodiments, the light chain variable domain is amino-terminal to the heavy chain variable domain, and in other cases, it is carboxy-terminal to the heavy chain variable domain. In some cases, the linker comprises the sequence SSGGGGSGGGGSGGGGGS (SEQ ID NO: 12).

[0073] (b) Transmembrane domain The CAR polypeptides disclosed herein may contain a transmembrane domain, which may be a hydrophobic alpha-helix that spans the membrane. As used herein, a transmembrane domain refers to an optional protein structure that is thermodynamically stable in a cell membrane, preferably a eukaryotic cell membrane. The transmembrane domain of the CAR provided herein has the sequence: FWVLVVVGGVLACYSLLVTVAFIIFWV (SEQ ID NO: 16). Other transmembrane domains can be used, including those shown below. Table 1: Examples of transmembrane domains

[0074] [Table 1] (c) Spacer region The CAR or polypeptide described herein may include a spacer region located between the T1 targeting domain (i.e., the HER2 or TAG-72 targeting scFv or variant thereof) and the transmembrane domain. The spacer region may function to provide flexibility to the CAR or its domains or to prevent steric hindrance of the CAR or its domains. A variety of different spacers can be used. Some of them include at least a portion of a human Fc region, such as a hinge portion or a CH3 domain of a human Fc region or variants thereof. Table 2 below provides various spacers that may be used in the CARs described herein. Table 2: Examples of spacers

[0075] [Table 2] Some spacer regions include all or part of an immunoglobulin (e.g., IgG1, IgG2, IgG3, IgG4) hinge region, i.e., the sequence between the CH1 and CH2 domains of an immunoglobulin, e.g., an IgG4 Fc hinge or a CD8 hinge. Some spacer regions include an immunoglobulin CH3 domain (referred to as CH3 or ΔCH2) or both a CH3 domain and a CH2 domain. The immunoglobulin-derived sequence may include one or more amino acid modifications, e.g., 1, 2, 3, 4 or 5 substitutions, e.g., substitutions that reduce off-target binding. The hinge / linker region may also comprise the following sequence:

[0076] [ka] (SEQ ID NO:26) or

[0077] [ka] The hinge / linker region may also comprise the IgG4 hinge region represented by the following sequence:

[0078] [ka] , followed by a linker sequence

[0079] [ka] followed by the IgG4 CH3 sequence:

[0080] [ka] (SEQ ID NO: 34). Thus, the entire linker / spacer region can include the sequence:

[0081] [ka] (SEQ ID NO: 31). In some cases, the spacer has 1, 2, 3, 4, or 5 single amino acid changes (e.g., conservative changes) compared to SEQ ID NO: 31. In some cases, the IgG4 Fc hinge / linker region is mutated at two positions (e.g., L235E and N297Q) in a manner that reduces binding by the Fc receptor (FcR).

[0082] (d) Intracellular signaling domain Any of the CAR constructs described herein contain one or more intracellular signaling domains (e.g., CD3ζ and, optionally, one or more costimulatory domains) that are functional termini of the receptor. After antigen recognition, the receptors cluster and a signal is transmitted to the cell. CD3ζ is the cytoplasmic signaling domain of the T cell receptor complex. CD3ζ contains three immunoreceptor tyrosine-based activation motifs (ITAMs), which transmit activation signals to T cells after they bind to cognate antigens. In some cases, CD3ζ provides the primary T cell activation signal, but does not provide a fully competent activation signal, and a costimulatory signal is required. Thus, in some examples, the CAR polypeptides disclosed herein may further comprise one or more costimulatory signaling domains in addition to CD3ζ. For example, the costimulatory domains CD28 and / or 4-1BB can be used to transmit a proliferation / survival signal along with the primary signaling mediated by CD3ζ. The costimulatory domain(s) are located between the transmembrane domain and the CD3ζ signaling domain. Table 3 includes examples of suitable costimulatory domains, along with the sequence of the CD3ζ signaling domain. Table 3: Examples of CD3 ζ domains and costimulatory domains

[0083] [Table 3] In some examples, the CD3 zeta signaling domain comprises an amino acid sequence that is at least 90%, at least 95%, at least 98% identical to SEQ ID NO: 35. In such cases, the CD3 zeta signaling domain has 1, 2, 3, 4, or 5 amino acid changes (preferably conservative substitutions) compared to SEQ ID NO: 35. In other examples, the CD3 zeta signaling domain is SEQ ID NO: 35. In various embodiments, the costimulatory domain is selected from the group consisting of a costimulatory domain or variant thereof having one to five (e.g., one or two) amino acid modifications as set forth in Table 3, a CD28 costimulatory domain or variant thereof having one to five (e.g., one or two) amino acid modifications, a 4-1BB costimulatory domain or variant thereof having one to five (e.g., one or two) amino acid modifications, and an OX40 costimulatory domain or variant thereof having one to five (e.g., one or two) amino acid modifications. In certain embodiments, a 4-1BB costimulatory domain or variant thereof having one to five (e.g., one or two) amino acid modifications is present in a CAR polypeptide described herein. In certain embodiments, there are two costimulatory domains, e.g., a CD28 costimulatory domain or variant thereof having 1-5 (e.g., 1 or 2) amino acid modifications and a 4-1BB costimulatory domain or variant thereof having 1-5 (e.g., 1 or 2) amino acid modifications (e.g., substitutions). In various embodiments, the 1-5 (e.g., 1 or 2) amino acid modifications are substitutions. In various embodiments, the costimulatory domain is amino-terminal to the CD3ζ signaling domain, and a short linker of 2-10, e.g., 3 amino acids (e.g., GGG) can be located between the costimulatory domain and the CD3ζ signaling domain.

[0084] In some cases, CARs may be produced using vectors in which the CAR open reading frame is followed by a T2A ribosomal skip sequence and a truncated EGFR (EGFRt) with a missing cytoplasmic signaling tail, or a truncated CD19R (also called CD19t). In this configuration, co-expression of EGFRt or CD19t provides an inert, non-immunogenic surface marker that allows accurate measurement of genetically modified cells, positive selection of genetically modified cells, and efficient cell tracking of therapeutic T cells in vivo after adoptive transfer. Efficient control of proliferation to avoid cytokine storm and off-target toxicity is a key hurdle for successful T cell immunotherapy. EGFRt or CD19t incorporated into the CAR lentiviral vector can act as a suicide gene to eliminate CAR+ T cells in the event of treatment-related toxicity. The CD3ζ signaling domain is followed by a ribosomal skipping sequence (e.g.,

[0085] [ka] and the following sequence:

[0086] [ka] This may be followed by a truncated EGFR having a sequence that is at least 90%, at least 95%, at least 98% identical or identical to (SEQ ID NO: 46). In some cases, the truncated EGFR has 1, 2, 3, 4 or 5 amino acid changes (preferably conservative) compared to SEQ ID NO: 46. Alternatively, the CD3ζ signaling domain may be followed by a ribosomal skipping sequence, e.g.

[0087] [ka] SEQ ID NO: 45) and the following sequence:

[0088] [ka] This may be followed by a truncated CD19R (also referred to as CD19t) having a sequence that is at least 90%, at least 95%, at least 98% identical or identical to (SEQ ID NO: 47).

[0089] The CAR or polypeptide described herein may be produced by any means known in the art, but is preferably produced using recombinant DNA technology. Nucleic acids encoding several regions of the chimeric receptor can be conveniently prepared and assembled into complete coding sequences by standard techniques of molecular cloning known in the art (genomic library screening, overlap PCR, primer-assisted ligation, site-directed mutagenesis, etc.). The resulting coding regions are preferably inserted into an expression vector and used to transform a suitable expression host cell line (preferably T lymphocytes, and most preferably autologous T lymphocytes).

[0090] Various T cell subsets isolated from patients can be transduced with vectors for CAR or polypeptide expression. Central memory T cells are one useful T cell subset. Central memory T cells can be isolated from peripheral blood mononuclear cells (PBMCs) by selecting for CD45RO+ / CD62L+ cells, for example, using a CliniMACS® device to immunomagnetically select cells expressing the desired receptor. Cells enriched for central memory T cells can be activated with, for example, lentiviral vectors directing expression of the CAR, or anti-CD 3 / CD 28 transduced with non-immunogenic surface markers for in vivo detection, ablation, and potential ex vivo selection. Activated / genetically modified central memory T cells can be expanded in vitro with IL-2 / IL-15 and then cryopreserved. Further methods of preparing CAR T cells can be found in International Patent Application No. PCT / US2016 / 043392. Methods for preparing useful T cell populations are described, for example, in WO 2017 / 015490 and WO 2018 / 102761. In some cases, it may be useful to use natural killer (NK) cells, such as allogeneic NK cells derived from peripheral blood or umbilical cord blood. In other cases, NK cells may be derived from human embryonic stem cells (hESCs) or induced pluripotent stem cells (iPSCs).

[0091] In some embodiments, compositions comprising iPSC-derived CAR T cells or CAR NK cells are described herein. In some embodiments, compositions comprising iPSC-derived CAR T cells or CAR NK cells have enhanced therapeutic properties. In some embodiments, iPSC-derived CAR T cells or CAR NK cells have demonstrated enhanced functional activities including potent cytokine production, cytotoxicity, and cytostatic inhibition of tumor growth, e.g., activities that reduce tumor burden. The CAR may be transiently expressed in a T cell population by mRNA encoding the CAR. The mRNA can be introduced into the T cells by electroporation (Wiesinger et al. 2019 Cancers (Basel) 11:1198). In some embodiments, the composition comprising the CAR T cells comprises one or more of helper T cells, cytotoxic T cells, memory T cells, naive T cells, regulatory T cells, natural killer T cells, or a combination thereof. Suitable TAG-72-targeting CARs are described, for example, in WO 2020 / 028721. Suitable HER2-targeting CARs are described, for example, in WO 2017 / 079694; suitable PSCA-targeting CARs are described, for example, in WO 2017 / 062628. In some cases, the TAG-72 CAR has the following amino acid sequence:

[0092] [ka] (SEQ ID NO: 61). Thus, from the amino terminus to the carboxy terminus, there is the scFv (underlined), a spacer (not underlined), a transmembrane domain (

[0093] [ka] SEQ ID NO: 17; underlined), 4-1BB costimulatory domain (

[0094] [ka] SEQ ID NO:38; not underlined), a GGG linker (GGG; underlined) and the ζ domain (

[0095] [ka] SEQ ID NO: 35; not underlined). The scFv contains the following amino acid sequence:

[0096] [ka] (SEQ ID NO: 63). This VH chain comprises the following: VH CDR1 (DHAIH, SEQ ID NO: 64); VH CDR2 (

[0097] [ka] , SEQ ID NO: 65); and VH CDR3 (SLNMAY, SEQ ID NO: 66).

[0098] [ka] (SEQ ID NO: 67). This VL chain comprises a VL CDR1 (

[0099] [ka] SEQ ID NO:68); VL CDR2 (WASTRES, SEQ ID NO:69); and VL CDR3 (QQYYSYPLS, SEQ ID NO:70). The spacer is an IgG4 hinge with an S to P mutation (

[0100] [ka] SEQ ID NO:25; bold mutations, underlined), linker (

[0101] [ka] SEQ ID NO: 24), and IgG4 CH3 domain (

[0102] [ka] SEQ ID NO: 34). In some cases, the mCD28 transmembrane domain is a CD28 transmembrane domain that is:

[0103] [ka] (SEQ ID NO: 16). In some cases, TAG72 is a signal sequence at the amino terminus, e.g., the GMCSF signal sequence (

[0104] [ka] SEQ ID NO: 11). Thus, the TAG72 CAR with signal sequence comprises the following:

[0105] [ka] (SEQ ID NO: 62).

[0106] In some cases, the TAG72 CAR can be followed by a T2A skip sequence and a truncated CD19 can function as a marker:

[0107] [ka] (Sequence number 71).

[0108] HER2 CAR In some cases, the HER2 CAR has the amino acid sequence:

[0109] [ka] (SEQ ID NO:72) Thus, it comprises from amino to carboxy terminus an scFv (underlined) and a spacer (not underlined), a CD8 transmembrane domain (not underlined), a 4-1BB costimulatory domain (not underlined), a GGG linker (not underlined) and a CD3 zeta domain (not underlined). The spacer in this HER2 CAR is IgG4 (S228P, L235E, N297Q).

[0110] PSCA-CAR Suitable PSCA CAR is PSCAscFv-IgG4(HL-CH3)-CD4tm-41BB-Zeta (SEQ ID NO: 73):

[0111] [ka] PSCAscFv-Linker-CD4tm-41BB-ζ (SEQ ID NO: 74)

[0112] [ka] PSCAscFv-IgG4(S228P,L235E,N297Q)-CD4tm-41BB-Zeta (SEQ ID NO: 75)

[0113] [ka] Cancer Treatment Aspects of the present disclosure provide methods for treating a subject with cancer by administering immune cells, e.g., T cells that express mb(28)IL-12 and a chimeric antigen receptor. (a) Target The subject treated by the described methods can be a human patient with a solid tumor (e.g., gastrointestinal cancer, breast cancer, lung cancer, bladder cancer, thyroid cancer, endometrial cancer, esophageal cancer, and ovarian cancer). Non-limiting examples of gastrointestinal cancer include colon cancer, gastric cancer, rectal cancer, colorectal cancer, pancreatic cancer, and combinations thereof. A subject at risk of cancer may exhibit one or more symptoms of cancer, such as unexplained weight loss, fatigue, pain, persistent cough, subcutaneous lumps, or abnormal bleeding. A subject at risk of cancer may have one or more risk factors for cancer, such as family history of cancer, age, smoking, obesity, or exposure to sunlight or carcinogens. A subject in need of the treatment described herein can be identified by routine medical examination, such as clinical examination, biopsy, magnetic resonance imaging (MRI), or ultrasound examination.

[0114] (b) Treatment regimen An aspect of the disclosure provides a method of treating a solid tumor, comprising administering a lymphodepleting therapy (e.g., cyclophosphamide) in combination with mb(28)IL-12 and immune cells expressing an appropriate chimeric antigen receptor. In some cases, the therapy also includes an inhibitor of PD-L1 activity, e.g., an anti-PD-L1 antibody or some other therapeutic agent that reduces immune suppression. The two components may be administered on the same day or on different days. In some cases, subjects suitable for the therapeutic methods described herein may undergo lymphodepletion therapy to reduce or deplete the subject's endogenous lymphocytes. Lymphodepletion generally refers to the destruction of endogenous lymphocytes and / or T cells, which is used prior to immune transplantation and immune therapy. Lymphodepletion may be achieved by administering lymphodepleting agents and / or irradiation (e.g., stereotactic irradiation). Lymphodepleting agents can be any molecule that, when administered to a subject, can reduce, deplete, or eliminate endogenous lymphocytes and / or T cells. In some instances, lymphodepleting agents are cytotoxic agents that specifically kill lymphocytes. Non-limiting examples of lymphodepleting agents include cyclophosphamide, fludarabine, gemcitabine, methotrexate, doxorubicin, and etoposide phosphate. In some instances, lymphodepleting treatment is non-myeloablative.

[0115] The methods described herein may include a pretreatment regimen that includes a single lymphodepleting agent (e.g., cyclophosphamide) or multiple lymphodepleting agents (e.g., cyclophosphamide and fludarabine). Subjects treated with the methods described herein may receive one or more doses of one or more lymphodepleting agents for a period of time suitable to reduce or deplete the subject's endogenous lymphocytes (e.g., 1-5 days). The subject can then be administered any of the CAR immune cells described herein following administration of a lymphodepletion therapy described herein. For example, one or more lymphodepletion agents can be administered to the subject 1-5 days (e.g., 1, 2, 3, 4, or 5 days) prior to administration of the CAR immune cells, e.g., CAR T cells. The methods described herein may include re-administering CAR immune cells to the subject. In some examples, the subject is administered lymphodepletion therapy before re-administration of anti-CAR immune cells. Each dose of CAR immune cells may be the same, or the dose may be gradually increased or decreased. The methods described herein may be used in combination with another anti-cancer therapy (eg, chemotherapy) or with other therapeutic agents that reduce the side effects of the therapies described herein.

[0116] (c) Administration An effective amount of a therapy (e.g., lymphodepleting agent, T cells expressing a CAR) can be administered (e.g., administered locally or systemically to a subject (e.g., a human) in need of treatment, optionally via a suitable route. Suitable modes of administration include injection, infusion, instillation, or ingestion. Injections include, but are not limited to, intravenous, intramuscular, intraarterial, intrathecal, intraventricular, intradermal, intraperitoneal, and subcutaneous injections and infusions. An effective amount refers to the amount of each active agent required to provide a therapeutic effect to a subject, either alone or in combination with one or more other active agents. Effective amounts will vary depending on factors such as the particular condition being treated, the severity of the condition, individual patient parameters (including age, physical condition, size, sex and weight), duration of treatment, the nature of concurrent therapy (if any), and the particular route of administration, as will be recognized by those skilled in the art. The materials, methods, and examples are illustrative only and are not intended to be limiting. All publications, patent applications, patents, sequences, database entries, and other references mentioned herein are incorporated by reference in their entirety, where applicable, and for all purposes. Other features and advantages of the described compositions and methods will become apparent from the following detailed description and drawings, and from the claims.

[0117] [Example] In order that the invention described may be more fully understood, the following examples are presented. The materials and methods used in the following examples are detailed after the examples. The examples described in this application are provided to illustrate the methods and compositions provided herein and should not be construed as limiting the scope thereof in any way. EXAMPLES

[0118] CD28 transmembrane in TAG 72-CAR containing a 4-1BB costimulatory domain enhances antitumor activity in vitro . We previously generated and preclinically evaluated second-generation TAG 72-specific CAR T cells containing the 4-1BB intracellular costimulatory domain, which demonstrated potent antitumor activity using a human xenograft peritoneal ovarian tumor model (C) (Reference 31). Based on our preclinical studies showing a lack of curative antitumor activity, and early Phase I data with first-generation TAG72-CAR T cells (Reference 33). Having demonstrated that the production of anti-idiotypic antibodies in patients likely contributes to the lack of durable therapeutic responses, we decided to redesign the CAR molecule for optimal functionality. First, we reevaluated the antigen-binding single-chain variable fragment (scFv) domain of our TAG72-CAR construct in an attempt to minimize the possibility of anti-CAR immunogenicity and improve T cell persistence. We utilized two additional scFvs (v15, and v59-15: a fusion between v15 and v59) based on the original humanized CC49 scFv (IDEC) that showed reduced anti-idiotypic immunogenic potential through affinity maturation (references 34, 35). Two of the three scFvs showed similar high binding affinity to the TAG72 antigen (IDEC, KD=33±20 nM; v15, KD=35±10 nM; v59-15, not determined). For all relevant in vitro and in vivo studies, we used human ovarian cancer cell lines that are TAG72 negative (OVCAR8) or have varying cell surface expression levels of TAG72 (OVCAR3, OV90, and OVCAR8-sTn) (Figure 1). We incorporated the scFvs into the same CAR scaffold we originally published (31) and evaluated their antitumor activity in vitro and in vivo (Figures 2a-2h). TAG72-CAR T cells with the v15 scFv were the most optimal in terms of antitumor activity compared to IDEC and v59-15.

[0119] Next, we generated seven v15 scFv-based TAG72-CAR constructs with various extracellular spacer domains and lengths (designated EQ, dCH2, CD8h, HL, and L), transmembrane domains (CD4tm, CD8tm, CD28tm), and intracellular costimulatory domains (CD28, 4-1BB) (Figure 3a). Although all seven CAR molecules expressed the CD19t marker equally, we observed higher cell surface CAR expression with Fc-derived spacers (EQ, dCH2) as measured by protein L staining of scFv, which closely matched the staining pattern using an anti-Fc antibody to detect the extracellular spacer domains (Figures 3b-3d and Figures 18a, 18b, and 18d). The three TAG72-CAR leads (dCH2(28tm)28z, dCH2(4tm)BBz, and dCH2(28tm)BBz) showed the highest T cell activation and cytokine production. Furthermore, we showed the highest PD-1 global phenotype in CD28 costimulatory domain-containing CAR T cells, consistent with previous reports with other CARs (Figure 18c). Interestingly, we showed the lowest PD-1 and highest IFNγ in TAG72-CAR T cells containing CD28tm and 4-1BB costimulatory domains. Some of this data was confirmed with HER2-CAR T cells, which showed the highest tumor cell killing, highest CD137 activation, and enhanced antigen-dependent T cell proliferation when CD28tm was coupled with the 4-1BB costimulatory domain (Figures 18e-18h).

[0120] Next, we evaluated the cytotoxic function of the TAG72-CAR T cells using an in vitro co-culture killing assay against cancer cell lines with various TAG72 expressions. In general, we found that TAG72-CAR T cells containing dCH2 spacer domains exhibited superior functionality with the greatest tumor cell killing, highest CD137 activation, reduced PD-1 expression, enhanced antigen-dependent T cell proliferation, and strong IFNγ and IL-2 cytokine production (Figure 3e-i). Three TAG72-CAR leads (dCH2(4tm)BBz, dCH2(28tm)BBz, and dCH2(28tm)28z) showed the highest antigen density-dependent T cell activation and cytokine production. Furthermore, we showed the greatest PD-1-enhancing phenotype in CD28 costimulatory domain-containing CAR T cells, similar to previous reports with other CARs (References 16, 36-38). Unexpectedly, TAG72-CAR variants with CD8h and HL spacers displayed a distinct lack of CAR cell surface expression but showed variable tumor cell killing capacity with some, albeit suboptimal, ability to induce antigen-dependent T cell proliferation. Finally, we showed the lowest PD-1 expression and highest IFNγ production in our TAG72-CAR T cells containing CD28tm and 4-1BB costimulatory domains.

[0121] From this study, we further subjected three TAG72-CAR T cell lead candidates to a "stress test". We first evaluated the cytotoxicity of CAR T cells over a long-term 10-day co-culture assay with OV90 tumor cells. Using xCELLigence as a readout, TAG72-CAR T cells containing CD28tm and 4-1BB costimulatory domains showed the greatest anti-tumor activity (Figure 3k). We then performed a recursive tumor cell killing assay by re-challenging TAG72-CAR T cells with OV90 tumor cells. We observed two interesting patterns; both 4-1BB costimulatory domain-containing CARs showed superior antigen-dependent T cell proliferation profiles, but CD28 transmembrane domain-containing CAR T cells achieved better control of tumors over the re-challenge time point (Figure 3l). Taken together, our in vitro studies identified three TAG72-CAR T cell lead candidates with potent but varying anti-tumor functional profiles, which we selected for further evaluation of their in vivo preclinical therapeutic activity. EXAMPLES

[0122] The CD28 transmembrane domain in 4-1 BB-based TAG72-CAR T cells induces a durable therapeutic response in vivo. The antitumor efficacy of these three TAG72-CAR leads was evaluated using a previously established human peritoneal ovarian tumor xenograft model (A) (ref. 31). Fourteen days after intraperitoneal (i.p.) OVCAR3 tumor injection, mice were transfected with untransduced (UTD), TAG72-dCH2(4tm)BBz, TAG72-dCH2(28tm)28z, or TAG72-dCH2(28tm)BBz CAR T cells (5.0 × 10 6) by local i.p. delivery. Dramatic antitumor responses were demonstrated with all three CAR T cells, which persisted for up to 6 weeks in treated mice (Fig. 4a). However, we observed tumor recurrence 6-8 weeks after treatment in mice treated with TAG72-dCH2(4tm)BBz or TAG72-dCH2(28tm)28z CAR T cells, whereas TAG72-dCH2(28tm)BBz durably controlled tumors, with 4 of 10 mice achieving a complete therapeutic response (Fig. 4b). To clarify potential differences observed between the three TAG72-CAR T cell leads, we quantified CAR T cells in the ascites and peripheral blood of mice after treatment. CAR T cells were found in the peritoneal ascites of treated mice, with TAG72-dCH2(28tm)BBz CAR T cells showing the highest percentage of cells collected from ascites 14 days after CAR T cell injection compared to the other two CAR T cells (Figure 4c). We further collected ascites at day 62 after treatment, at the time of tumor recurrence, and showed the elimination of TAG72+ tumor cells in TAG72-dCH2(28tm)BBz CAR T cell treated mice, but the presence of TAG72+ tumor cells with the other two CAR T cells (Figure 4c). We observed similar numbers of 4-1BB costimulatory domain-containing CAR T cells, higher than CD28 costimulatory domain-containing CAR T cells, in the blood of mice at all three time points (Figure 4d). The data were largely consistent with the T cell proliferation patterns in our in vitro functional assays. We further observed similar antitumor kinetics of the three CAR T cell lead candidates using a second, more aggressive human OV90 peritoneal ovarian tumor xenograft model (Figure 5).

[0123] To better address potential safety concerns of this optimized TAG72-CAR containing the new anti-TAG 72 scFv and CAR backbone, we evaluated the on- and off-target normal cell killing ability of TAG72-dCH2(28tm)BBz CAR T cells. Little or no TAG72 expression was observed across 10 primary human normal cells assessed by flow cytometry, and CD137+ T cell activation and cell killing by TAG72-CAR T cells was minimal, a consistent finding across five independent human CAR T cell donors (Figure 6). In summary, the TAG72-dCH2(28tm)BBz CAR construct showed the most optimal anti-tumor functionality in vitro and had safe, potent and durable anti-tumor efficacy in vivo. EXAMPLES

[0124] IL-12 / IFNγ signaling regulates the antitumor activity of CAR T cells. Based on accumulating evidence that IFNγ signaling affects the therapeutic activity of CAR T cells (references 39-41). Together with our observation that TAG72-dCH2(28tm)BBz CAR T cells secreted the highest levels of IFNγ in our study, we hypothesized that IFNγ signaling contributed to the superior antitumor activity of TAG72-dCH2(28tm)BBz CAR T cells. To test this hypothesis, we inhibited IFNγ signaling using an anti-IFNγR1 blocking antibody or enhanced IFNγ secretion using human recombinant interleukin-12 (IL-12) in our extended in vitro co-culture tumor cell killing assay (Figures 19a and 7a-c). Surprisingly, we observed a dose-dependent attenuation of tumor cell killing by blockade of IFNγ signaling using OV90 tumor 234 cells (Figure 19b, left and Figure 7A) and OVCAR3 tumor cells (Figure 8). We also observed a dose-dependent enhancement of tumor cell killing with increasing concentrations of recombinant huIL12, which resulted in increased production of IFNγ by CAR T cells as determined by ELISA at the assay endpoint (Figure 19b, right and 2c). We also observed a dose-dependent enhancement of tumor cell killing with increasing concentrations of recombinant huIL-12, which resulted in significant secretion of IFNγ by CAR T cells as determined by ELISA at the assay endpoint (Figure 7b). Interestingly, in this assay, only a minor effect was observed on T cell proliferation after IFNγ blockade or addition of huIL-12 (Figure 7c). We further corroborated our findings using a recurrent tumor cell killing assay and demonstrated the requirement for IFNγ in promoting sustained antitumor activity in vitro (FIG. 9). We sought to better understand the signaling dynamics downstream of IL-12 in the context of CAR T cell antigen stimulation. We first optimized the assay to confirm phosphorylated STAT4 (pSTAT4) downstream of huIL12 in TAG72-CAR T cells. Interestingly, we observed that pSTAT4 levels peaked at 1 hour with recombinant huIL12 alone and declined over a 24 hour time course, whereas pSTAT4 persisted over 24 hours in CAR T cells stimulated with a combination of plate-bound TAG72 antigen and huIL12 (Figures 10A-10B). The data suggest that IL-12 signaling enhances CAR T cell247 activation and cytotoxicity, and conversely, CAR T cell activation enhances sustained IL12 / pSTAT4 signaling. EXAMPLES

[0125] Engineered membrane-bound IL-12 signaling drives robust CAR T cell-mediated tumor cell killing and antigen-dependent T cell proliferation . Based on our in vitro findings that IL-12-induced IFNγ signaling is important for CAR T cell antitumor activity, we engineered IL-12 into our TAG72-CAR T cells to prevent potential off-target toxicity induced by secreted IL-12. 42We aimed to spatially restrict the effect of IL-12 by constructing membrane-bound IL-12 (mbIL12) constructs with various transmembrane domains (Figure 7d). When we checked for their cell surface expression, we observed a slightly enhanced expression of mbIL12(CD28tm) (SEQ ID NO: 6 or 7) compared to mbIL12(B7.1tm), but both showed considerable expression on TAG72-CAR T cells (Figure 7e). We then evaluated the tumor cell killing activity of TAG72-CAR T cells expressing either of the two versions of mbIL12 and found that TAG72-CAR / mbIL12(CD28tm) T cells showed the highest tumor cell killing activity, T cell proliferation and IFNγ production (Figures 7f-7h). In the in vitro assay, the T cell cytotoxic advantage of mbIL12 was dependent on IFNγ signaling (Figure 23d). Similar enhancement of activation and IFNγ production by mbIL12 was shown for TAG72+ gastric and ovarian cancer ascites cells from human patients, with little or no effect on targeting of normal primary human cells (Figures 23a-23c). T cell functional benefit was also observed using HER2- and PSCA-targeted CAR T cells engineered with mbIL12(CD28tm) (Figures 14a-14h). mbIL12-modified CAR T cells showed little / no change in CD4 / CD8 ratios or markers of naive / memory phenotype (CD62L, 269 CCR7) in in vitro functional assays (Figures 14i-14k). We further evaluated the activity of CAR T cells with mbIL12(CD28tm) in repeated tumor cell killing, where we again observed enhanced tumor cell killing and CAR antigen-dependent T cell proliferation across multiple re-challenge time points (Figure 7i).

[0126] We then compared the expression of cell surface mbIL12 on TAG72-CAR T cells rested in the absence of serum or exogenous cytokines before stimulation with plate-bound TAG72 antigen or control antigen. Without T cell stimulation, cell surface mbIL12 was detected at low levels on T cells, but rather was found mostly intracellularly (Figure 10c-10d). We showed comparable mRNA transcripts of mbIL12 in stimulated and unstimulated T cells, suggesting that the low cell surface mbIL12 expression is likely due to intracellular localization and not driven by differential EF1α promoter activity (Figure 10h). Interestingly, a significant dose-dependent increase in cell surface expression of mbIL12 was observed in antigen-stimulated CAR T cells, especially in the CD137+ activated T cell subset (Figure 10d-10e). EXAMPLES

[0127] Antigen-dependent IL-12 signaling in CAR T cells. We sought to further understand the downstream signaling dynamics of IL-12 using phospho-flow cytometry. We first optimized the assay to confirm phosphorylated STAT4 (pSTAT4) downstream of recombinant human IL-12 (huIL12) in TAG72-CAR T cells. Interestingly, we observed that with recombinant huIL12 alone, pSTAT4 levels peaked at 1 hour and declined over a 24 hour time course, whereas in CAR T cells stimulated with a combination of plate-bound TAG72 antigen and huIL12, pSTAT4 persisted over 24 hours (Figure 10a-b). We then compared the expression of cell surface mbIL12 on TAG72-CAR T cells rested in the absence of serum or exogenous cytokines prior to stimulation with plate-bound TAG72 antigen or control antigen. We found that without stimulation, mbIL12 was detected at very low levels in engineered T cells. However, a significant increase in cell surface expression of mbIL12 was seen in antigen-stimulated TAG72-CAR T cells (Figure 10c). Next, we examined pSTAT4 expression in TAG72-CAR / mbIL12 T cells in response to CAR stimulation. We observed the expected phosphorylation of STAT3 in response to TAG72 antigen in both TAG72-CAR and TAG72-CAR / mbIL12 T cells, which was only slightly activated by huIL12. However, CAR stimulation showed a dose-dependent increase in pSTAT4 in TAG72-CAR / mbIL12 T cells compared to TAG72-CAR T cells alone (Figure 10d). We further evaluated the possibility of trans-signaling in TAG72-CAR / mbIL12 T cells. We transduced mbIL12 into HT1080(TAG72-) cells and co-cultured them with T cells in the presence of soluble CD3 / CD28 stimulation. We observed an increase in pSTAT4 in T cells when cultured with both HT1080-mbIL12 and soluble CD3 / CD28, but not with HT1080-mbIL12 alone or HT1080-WT with CD3 / CD28 (Figure 10e).Taken together, the data suggest that mbIL12-engineered CAR T cells exhibit improved in vitro antitumor activity, unexpectedly dependent on CAR antigen stimulation, which we term antigen-dependent IL-12 signaling.

[0128] Next, we investigated downstream pSTAT4 expression in TAG72-CAR / mbIL12 T cells in response to CAR stimulation. We observed the expected phosphorylation of STAT3 in response to TAG72 antigen in both TAG72-CAR and TAG72-CAR / mbIL12 T cells, which was only slightly activated by huIL12. However, CAR stimulation showed a dose-dependent increase in pSTAT4 in TAG72-CAR / mbIL12 T cells compared to TAG72-CAR T cells alone (Figure 10f). We further evaluated the possibility of trans-signaling in TAG72-CAR / mbIL12 T cells. We transduced mbIL12 into HT1080 (TAG72-) cells and co-cultured them with T cells in the presence of soluble CD3 / CD28 stimulation. An increase in pSTAT4 was observed in T cells when cultured with both HT1080-mbIL12 and soluble CD3 / CD28, but not with HT1080-mbIL12 alone or HT1080-WT with CD3 / CD28 (Figure 10g). Antitumor activity was further assessed in a recursive tumor cell killing assay, where we again observed enhanced tumor cell killing and CAR antigen-dependent T cell proliferation across multiple rechallenge time points in mbIL12-engineered CAR T cells (Figure 7i). Importantly, no proliferation or survival benefit was observed in the absence of CAR antigen stimulation (Figures 13a-13b). Taken together, the data suggest that mbIL 12-modified CAR T cells exhibit improved in vitro antitumor activity, unexpectedly dependent on CAR antigen stimulation, which we term antigen-dependent IL-12 signaling. EXAMPLES

[0129] Superior antitumor activity through antigen-dependent IL-12 signaling in CAR T cells. Next, we evaluated the therapeutic potential of TAG72-CAR T cells with antigen-dependent IL-12 signaling. Using an ipOVCAR3 tumor xenograft model, mice treated with TAG72-CAR / mbIL12 T cells sustained more durable antitumor responses and achieved a higher incidence of durable complete responses compared to TAG72-CAR T cells (Figures 11a-b). Importantly, tumor-bearing mice treated with CD19-CAR T cells and CD19-CAR / mbIL12 T cells showed little difference in therapy, supporting the antigen-dependent nature of mbIL12. We observed that the frequency of hCD45+ cells in the peritoneal ascites of mice treated with TAG72-CAR / mbIL12 T cells was higher at 2 and 4 weeks after treatment compared to mice treated with TAG72-CAR T cells alone (Figures 11c-11d and 19c-19d). We also observed significantly higher and sustained levels of TAG72-CAR / mbIL12 T cells in peripheral blood compared to TAG72-CAR T cells alone (Figures 11d and 19d). We replicated the findings using the more challenging ipOV90 tumor xenograft model, where the differences in therapeutic and T cell persistence between TAG72-CAR and TAG72-CAR / mbIL12 T cells were even more pronounced (Figures 11e-f). EXAMPLES

[0130] Improved systemic disease targeting by CAR T cells with antigen-dependent IL-12 signaling. One common argument against locoregional administration of CAR T cells is their potential spatial restriction, which prevents systemic therapy in patients with widespread metastatic disease (Ref. 43). However, our data suggest that locally administered intraperitoneally mbIL12 modified CAR T cells may have a greater ability to target disease outside the peritoneum. To test this, we established a xenograft OV90 tumor model bearing both local ip and systemic sc tumors in the same mouse. OV90 (ffluc negative) tumor cells were injected subcutaneously (sc) and tracked by caliper measurements, and OV90 (eGFP / ffluc expressing) tumor cells were injected ip and tracked by bioluminescence flux imaging (Fig. 12a). As we observed in our previous experiments, ip antitumor responses were greater in mice treated locally with TAG72-CAR / mbIL12 T cells compared to TAG72-CAR T cells alone (Fig. 12b-c). Although sc tumors initially regressed similarly in both treatment groups, all tumors recurred after TAG72-CAR T cell treatment alone, whereas sc tumors were completely eradicated in all mice with TAG72-CAR / mbIL12 T cell treatment (Figure 12d). We again confirmed this phenomenon by observing higher levels of CAR T cells in the peripheral blood and peritoneal ascites of mice treated with TAG72-CAR / mbIL12 T cells (Figures 12e-f). Furthermore, immunohistochemistry (IHC) analysis of subcutaneous tumors 12 days after treatment demonstrated significantly greater infiltration of TAG72-CAR / mbIL12 T cells compared to TAG72-CAR T cells alone (Figure 12g). Overall, the data support logoregional delivery of CAR T cells with engineered antibody-dependent IL-12 signaling in persistently targeting both local and systemic disease. EXAMPLES

[0131] Example 8: Membrane-bound IL-12 improves the antitumor activity of various CARs. We further engineered HER2 or PSCA-targeting CAR T cells with the membrane-bound IL-12 molecule mbIL12 (CD28tm) (SEQ ID NO: F) and showed similar improvements in antitumor activity (Figure 14a-f). To extend our findings that mbIL12 signaling enhances local to systemic disease targeting by CAR T cells, we used a patient-derived HER2+ BBM1 tumor model that we previously published in the development of our HER2-CAR T cells. Using this model system, we established a BBM1 tumor xenograft model with both local intracranial (ic) and intratibial (i.ti.) brain metastases. and bone metastases in the same mice. BBM1-ZsGreen-ffluc tumor cells were injected ic and i.ti. and followed by bioluminescence flux imaging (Figure 16a). Mice were treated with a single dose of HER2-CAR T cells by local intracerebroventricular (icv) delivery, corresponding to day 8 and day 23 post-injection in the intratumoral injection. We observed a strong therapeutic response in the brain with either HER2-CAR T cells alone or HER2-CAR T cells engineered with mbIL12 (Figures 16b and 16c, top row). However, only mbIL12-modified HER2-CAR T cells induced bone metastases in the majority of treated mice, compared to the heterogeneous response of i.ti.HER2-CAR T cells alone (Figures 16b and 16c, bottom row). Similar to our observations in the dual tumor ovarian cancer model described above, we observed significant expansion and greater persistence of HER2-CAR / mbIL12 T cells in the peripheral blood of local icv treated mice compared to HER2-CAR T cells alone (Figures 16d-16e), and the persistent HER2-CAR / mbIL12 T cells in the blood displayed a greater central memory phenotype (Figures 16g-16h), along with greater T cell infiltration in both brain and bone metastases (Figure 16f). Together, the two models strongly support the advantage of mbIL12 in targeting systemic disease following local administration of CAR T cells. EXAMPLES

[0132] mbIL-12 modified CAR T cells demonstrate safety and efficacy in an immunocompetent mouse model of ovarian cancer peritoneal metastasis. Safety concerns of IL-12 have limited its therapeutic application in humans to date. Infusion of soluble recombinant IL-12 into patients resulted in undesirable toxicities, including colitis. Furthermore, engineering adoptive T cell therapy with soluble IL-12 to restrict IL-12 production to activated T cells, even under an NFAT-inducible promoter system, resulted in similar toxicity. To test whether mbIL12-modified CAR T cells are a safe therapeutic approach compared to soluble IL-12, we constructed a fully immunocompetent mouse model of TAG72+ ovarian cancer peritoneal metastasis with mouse versions of our TAG72-CAR and mbIL12 constructs (Figure 20a). TAG72-CAR T cells were further efficiently produced as previously described and engineered to express mbIL12 using retroviral transduction of mouse splenic T cells (Figures 20b-20c). Murine ID8 ovarian cancer cells were stably transduced and cloned to express TAG72 using mouse sialyltransferase (mSTn) and used as tumor targets for our studies (Figure 20d). Murine TAG72-CAR / mbIL12 T cells showed increased activity and cytotoxicity against TAG72+ID8 tumor cells compared to TAG72-CAR T cells alone (Figures 20e-20g). We observed a bias towards antigen-dependent proliferation of CD8+CAR T cells compared to the starting product during in vitro tumor cell killing assays (Figures 20c and 20h). We performed a direct comparison of the safety and efficacy of CAR T cells either engineered with mbIL12 or combined with soluble IL-12 injections. TAG72+ID8 (ffluc positive) tumor cells were injected ip, followed by bioluminescence flux imaging, and treated locally ip for 5 days with either TAG72-CAR T cells alone, TAG72-CAR / mbIL12 T cells, or TAG72-CAR T cells with soluble IL-12 (sIL12) injections 14 days after tumor injection (Figure 21a). Robust therapeutic responses were seen in all treated mice, but were greater in mice locally treated with TAG72-CAR / mbIL12 T cells or TAG72-CAR T cells with sIL12 compared to TAG72-CAR T cells alone (Figures 17a, 21b-21c). Interestingly, at later time points, we observed relapse in mice treated with CAR T cells and sIL12 injections, but sustained tumor control in mice treated with mbIL12 engineered CAR T cells. Although IL-12 benefited CAR T cell therapy by injection or with T cell engineering, we observed significant weight loss in mice after TAG72-CAR T cells and sIL12 injection, but not in TAG72-CAR / mbIL12 treated mice (Figures 17a, 21d). We confirmed systemic toxicity associated with sIL12 injection, with mice showing signs of splenomegaly (Fig. 21e, 21k, top), liver abnormalities and greater T cell infiltration in the liver (Fig. 21f and 21k, bottom), along with increased serum ALT and AST levels and altered peripheral blood lymphocyte counts (Fig. 17b, 21g-21h), but no appreciable systemic effects were observed with TAG72-CAR / mbIL12 T cells. Importantly, sIL12 injection led to increased systemic IFNγ levels in the serum of TAG72-CAR T cell-treated mice, which were undetectable in TAG72-CAR / mbIL12-treated mice (Fig. 17c, 21i-21j). The data support the safety and efficacy of locally administered mbIL 12-modified CAR T cells in an immunocompetent mouse tumor model. EXAMPLES

[0133] mbIL-12 modified CAR T cells modify the immunosuppressive tumor microenvironment in peritoneal metastases. IL-12 has multiple effects on the tumor microenvironment, including reshaping the myeloid cell compartment, promoting antigen presentation, and enhancing adoptive T cell therapy. Therefore, we evaluated the impact of mbIL 12 manipulation of CAR T cells and its role in the tumor microenvironment in an ovarian cancer peritoneal metastasis model. Peritoneal tumors harvested from mice treated with TAG72-CAR / mbIL12 T cells showed increased T cell infiltration compared to TAG72-CAR T cells alone (Figure 22a). At this early time point post-treatment, both treatment groups showed comparable efficacy in clearing tumor cells in the ascites (Figures 22b-22c). We further evaluated the ascites for changes in immune cell subsets by flow cytometry (Figure 22l), which showed an increase in overall CD45+ immune cell numbers consisting of T cells, myeloid cells, and NK cells, with little change in total B cell numbers (Figures 22d and 22m). TAG72-CAR T cells engineered with mbIL12 showed increased persistence of CD8+ T cells in the ascites (Figures 22e-22f), and a small but significant increase in CD137 activation (Figure 22g). Interestingly, persisting TAG72-CAR / mbIL12 T cells showed little change in the CD62L+CD44+ naive / memory phenotype, but there was an increase in the naive / memory phenotype of non-CAR T cells in the ascites (Figure 22h). We then evaluate changes in the myeloid cell compartment in the ascites, showing a significant increase in total 17 F4 / 80+ tumor-associated macrophages (TAMs), along with an increase in Ly6C+ inflammatory monocytes, without changes in Ly6G+ neutrophil and CD11c+ DC populations (Figure 22i). However, we observed a significant increase in MHC-II expression among CD11c+CD103+414 DCs, suggesting an improved mature antigen-presenting phenotype. Collectively, the findings demonstrate beneficial modulation of the immunosuppressive tumor microenvironment in this model by mbIL 12-modified CAR T cell(s).

[0134] Materials and Methods Used in the Examples cell line The epithelial ovarian cancer line OVCAR3 (ATCC HTB-161) was cultured in RPMI-1640 (Lonza) (complete RPMI) containing 20% ​​fetal bovine serum (FBS, Hyclone) and 1× antibiotic-antimycotic (1×AA, Gibco). The epithelial ovarian cancer line derived from metastatic ascites OV90 (CRL-11732) was cultured in a 1:1 mixture of MCDB 105 medium (Sigma) and Medium 199 (Thermo) adjusted to pH 7.0 with sodium hydroxide (Sigma) and final 20% FBS and 1×AA. The epithelial ovarian cancer line OVCAR8 was a generous gift from Dr. Carlotta Glackin at City of Hope and was cultured in complete RPMI-1640. All cells were cultured at 37°C and 5% CO2. The line BBM1, derived from a patient with breast-brain metastasis, was cultured as previously described. The mouse ovarian cancer cell line ID8 was cultured in Dulbecco's modified Eagle's medium (DMEM) (cDMEM) supplemented with 10% FBS, 2 mM L-glutamine (Fisher Scientific), and 25 mM HEPES (Irvine Scientific). The human breast cancer cell line MDA-MB-468 (ATCC HTB-132) was engineered to express HER2 (Accession Number: NM_004448.4) under the control of the EF1α promoter via epHIV7 lentiviral transduction (468-HER2). All cells were cultured at 37°C with 5% CO2. Human primary cell lines were obtained from Cell Biologics (human primary colonic epithelial cells H-6047, human primary esophageal epithelial cells H-6046, human primary kidney epithelial cells H-6034, human primary ovarian epithelial cells H-6036, human primary pancreatic epithelial cells H-6037, human primary proximal tubular epithelial cells H-6015, human primary small intestinal epithelial cells H-6051, human primary gastric epithelial cells H-6039), Promocell (human cardiomyocytes C-12810), and Lonza (human bronchial epithelial cells CC-2541) and cultured according to the supplier's specifications.

[0135] DNA constructs, tumor lentiviral transduction, and retroviral production Tumor cells were engineered to express enhanced green fluorescent protein and firefly luciferase (eGFP / ffluc) by transduction with epHIV7 lentivirus carrying an eGFP / ffluc fusion under the control of the EF1α promoter as previously described. 23 The humanized scFv sequence used in the CAR construct was obtained from the monoclonal antibody clone huCC49, which targets TAG72. 31The extracellular spacer domain contained a 129 amino acid intermediate-length CH2 deleted version (ΔCH2) of the IgG4 Fc spacer (Reference 31). The intracellular costimulatory signaling domain contained was 4-1BB with the CD4 transmembrane domain. The CD3ζ cytolytic domain was previously described (Reference 31). Variations in the extracellular spacer domain, transmembrane domain, and intracellular costimulatory signaling domain have been previously described (References 11, 23, 38). The CAR sequence was separated from a truncated CD19 gene (CD19t) by a T2A ribosomal skip sequence and cloned into an epHIV7 lentiviral backbone under the control of the EF1α promoter. The PSCA-BBζ CAR construct was previously described (Reference 38). Membrane-bound IL-12 (mbIL12) constructs were generated using p35 and p40 genes (p35, NC_000003.12; p40, NC_000005.10) separated by a G4S spacer and linked to either B7.1 or CD28 transmembrane domains. Lentiviruses were generated as previously described (ref. 38). Lentivirus titers were quantified using HT1080 cells based on CD19t or IL-12 cell surface expression using flow cytometry. Human and mouse tumor cells were engineered to express enhanced green fluorescent protein and firefly luciferase (eGFP / ffluc), or ffluc alone, by transduction with epHIV7 lentivirus carrying eGFP / ffluc fusion or ffluc alone under the control of the EF1α promoter as previously described23. The mouse ovarian cancer cell line ID8 was also engineered to express the target antigen TAG72 via transduction with epHIV7 lentivirus carrying the mouse st6galnac-I gene (mSTn) under the control of EF1α. mSTn is a unique sialyltransferase responsible for the surface expression of aberrantly glycosylated sialyl-Tn (TAG72). The humanized scFv sequence used in the CAR construct was obtained from the monoclonal antibody clone huCC49 targeting TAG7231.

[0136] The extracellular spacer domain contained a 129 amino acid mid-length CH2 deleted version (ΔCH2) of the IgG4 Fc spacer31. The intracellular costimulatory signaling domain contained was 4-1BB with the CD4 transmembrane domain. The CD3ζ cytolytic domain was previously described31. Variations in the extracellular spacer domain, transmembrane domain, and intracellular costimulatory signaling domain were previously described11,23,38. The CAR sequence was separated from a truncated CD19 gene (CD19t) by a T2A ribosomal skip sequence and cloned into an epHIV7 lentiviral backbone under the control of the EF1α promoter. The PSCA-BBζCAR and HER2-BBζCAR constructs were previously described38. Membrane-bound IL-12 (mbIL12) constructs were generated using p35 and p40 genes (p35, 532 NC_000003.12; p40, NC_000005.10) separated by a G4S spacer and linked to either the B7.1 or CD28 transmembrane domain. Lentiviruses were generated as previously described. Lentivirus titers were quantified using HT1080 cells based on CD19t or IL-12 cell surface expression using flow cytometry. A murine CAR (mTAG72-CAR) construct was developed using scFv sequences derived from the murine anti-human TAG72 antibody clone (CC49). The extracellular spacer domain comprised a murine IgG1 region followed by a murine CD28 transmembrane domain66,67. The intracellular costimulatory signaling domain comprised a murine 4-1BB followed by a murine CD3ζ cytolytic domain as previously described68. The CAR sequence was separated from a truncated murine CD19 gene (mCD19t) by a T2A ribosomal skip sequence. Murine membrane-bound IL12 was generated using murine p40 and p35 subunit sequences linked to a cell membrane-tethered murine CD28 transmembrane domain sequence. All retroviral constructs were cloned into the pMYs retroviral backbone under the control of a hybrid MMLV / MSCV promoter (Cell Biolabs Inc). Production of retroviruses used to transduce primary murine T cells was performed as previously described69. Retrovirus was produced by transfecting the ecotropic retroviral packaging cell line PLAT-E with mTAG72-CAR retroviral backbone plasmid DNA using FuGENE HD transfection reagent (Promega). Viral supernatants were collected after 24, 36, and 48 hours, pooled, and stored in aliquots at -80°C for future T cell transduction. Control non-targeted murine mPSCA-CAR was generated as previously described (ref. 70).

[0137] T cell isolation, lentiviral transduction, and ex vivo expansion Leukapheresis products were obtained from consenting study participants (healthy donors) under a protocol approved by the City of Hope Internal Review Board (IRB) and enriched for T cells as previously described (Reference 38.58). T cell activation and transduction were performed as previously described (Reference 38). Where indicated, we performed a second lentiviral transduction 24 hours after the first transduction. Cells were then produced ex vivo, enriched for CAR, and frozen as previously described (Reference 38). Purity and cell surface phenotype of CAR T cells were analyzed by flow cytometry using the antibodies and methods described below. For mouse T cells, splenocytes were obtained by manual digestion of spleens from female C57BL / 6j mice. Enrichment of T cells was performed with the EasySep™ Mouse T Cell Isolation Kit according to the manufacturer's protocol (StemCell Technologies). Single or dual retroviral transduction with mTAG72-CAR, mPSCA-CAR and / or mouse mbIL12 and subsequent expansion were performed as previously described and cultured in cRPMI69.

[0138] Flow cytometry For flow cytometry analysis, cells were resuspended in FACS buffer (Hank's Balanced Salt Solution without Ca2+, Mg2+, or phenol red (HBSS- / -, Life Technologies) containing 2% FBS and 1x AA). Cells were incubated with primary antibodies for 30 min at 4°C in the dark. For secondary staining, cells were washed twice and then incubated with either Brilliant Violet 510 (BV510), fluorescein isothiocyanate (FITC), phycoerythrin (PE), peridinin chlorophyll protein complex (PerCP), PerCP-Cy5.5, PE-Cy7, allophycocyanin (APC), or APC-Cy7 (or APC-eFluor780) conjugated antibodies for 30 min at 4°C in the dark. Antibodies used included anti-CD3 (BD ​​Biosciences, clone SK3), anti-CD4 (BD Biosciences, clone SK3), anti-CD8 (BD Bosciences, clone SK1), anti-CD19 (BD Biosciences, clone SJ25C1), anti-CD45 (BioLegend, clone 30-F11), anti-CD45 (BD Biosciences, clone 2D1), anti-CD69 (BD Biosciences, clone L78), anti-CD137 (BD Biosciences, clone 4B4-1), biotinylated protein L (GenScript USA) (25), TAG72 (clone muCC49), donkey anti-rabbit Ig (Invitrogen), goat anti-mouse Ig (BD Biosciences), and streptavidin (BD Biosciences). Cell viability was determined using 4',6-diamidino-2-phenylindole (DAPI, Sigma). Flow cytometry was performed on a MACSQuant Analyzer 10 (Miltenyi Biotec) and data were analyzed with FlowJo software (v10, TreeStar). For intracellular flow cytometry, CAR T cells were thawed and cultured at 1 × 10 in IL-2 (50 U / mL) and IL-15 (0.5 ng / mL). 6The next day, CAR T cells were washed twice with 1× PBS and diluted to 1×10 6 Cells were suspended at 1 × 10 cells / mL per well in a 96-well plate. 5 Cells were plated and left overnight. The next day, cells were stimulated with soluble cytokines [IL-2 (50 U / mL), IL-15 (0.5 ng / mL), IL-12 (10 ng / mL)] or transferred to high-binding 96-well plates pre-coated with the indicated amounts of control or TAG72 antigen (BioRad). Reagents and buffers for flow cytometry processing were pre-chilled on ice unless otherwise stated. After antigen stimulation, cells were washed with FACS buffer (supplemented with 0.1% sodium azide) and then fixed in pre-warmed 1x BD PHosflow Lyse / Fix buffer (558049) for 10 min at 37°C. Cells were then washed with FACS buffer and stained with extracellular antibodies on ice for 30 min in the dark, if required. Stained cells were washed and suspended in pre-chilled (-20°C) BD Perm Buffer III (558050) and kept on ice for 30 min. After washing, cells were suspended in human FC block (Miltenyi Biotec Inc., FLP3330), kept on ice for 30 min, washed and stained with intracellular antibodies: PE-pSTAT3, PE-pSTAT4, PE-pSTAT5 (Biolegend). Data were acquired on a MACSQuant Analyzer 16 cytometer (Miltenyi) and analyzed with FlowJo v10.8.

[0139] In vitro tumor killing and T cell functional assays For tumor cell killing assays, CAR T cells and tumor targets were co-cultured in complete X-VIVO (for human T cell assays) or cRPMI (for mouse T cell assays) without cytokines at the indicated effector:tumor (E:T) ratios in 96-well plates for the indicated time points and analyzed by flow cytometry as described above. Tumour cells were plated overnight before adding T cells. Tumor cell killing by CAR T cells was calculated by comparing the number of CD 45 negative DAPI negative (viable) cells for targets co-cultured with untransduced (UTD) T cells. For tumor cell challenge assays, TAG72-CAR T cells engineered with or without mbIL12 were co-cultured with OV90 cells at an E:T ratio of 1:2 and re-challenged with OV90 cells every 2 days up to 5 times. Similarly, HER2-CAR T cells engineered with or without mbIL12 were co-cultured with 468-HER2 tumor cells at an E:T ratio of 1:10 and re-challenged with 468-HER2 cells every 3 days. Remaining viable tumor cells and T cells were quantified by flow cytometry before each re-challenge with tumor cells and 2 or 3 days after the last re-challenge. For xCELLigence tumor cell killing assays, CAR T cells and tumor targets were co-cultured in complete X-VIVO without cytokines at the indicated effector:tumor (E:T) ratios in 96-well plates for up to 10 days and analyzed by flow cytometry as described above. To assess CAR T cell activity against normal tissues, normal tissue cells were co-cultured with CAR T cells at the indicated E:T ratios. Primary gastric cancer ascites (GAS1) and ovarian cancer ascites (OAS3 and OAS4) from patients were immediately thawed and assessed in T cell functional assays. After 48 hours, CD137 expression and cell death were assessed by flow cytometry, and supernatants were collected and IFNγ was quantified by ELISA. For T cell activation assays, CAR T cells and tumor targets were co-cultured in complete X-VIVO without cytokines in 96-well plates at the indicated E:T ratios for the indicated time points and analyzed by flow cytometry for the indicated markers of T cell activation. For T cell activation assays against plate-bound antigens, purified soluble TAG72 antigen (BioRad) was plated in duplicates in 1×PBS in 96-well flat-bottom high affinity plates (Corning) at the indicated TAG72 units overnight at 4° C. Using the Bradford protein assay, a 20,000 units / mL stock solution of soluble TAG72 antigen was determined to be approximately 1.234 mg / mL total protein. The indicated number of TAG72-CAR T cells was then added to each well in a fixed volume of 100 uL and incubated for the indicated times before collection of cells for analysis of activation markers (CD69, CD137) by flow cytometry. Supernatants were also collected for analysis of cytokine production. For T cell survival assays, T cells were cultured at 1 × 10 in X-VIVO 10% FBS with or without cytokines. 6 Cells were plated at 1 × 10 cells / mL and counted every 2 days. After each counting day, the cell concentration was adjusted to 1 × 10 cells / mL with fresh medium. 6 Adjusted to cells / mL.

[0140] ELISA cytokine assays Supernatants from tumor cell killing assays or CAR T cell activation assays against plate-bound TAG72 antigen were collected at the indicated times and frozen at -20°C for further use. Supernatants were then analyzed for secreted human IFNγ and IL-2, respectively, using the Human IFNγ and IL-2 ELISA Ready-SET-GO!®; ELISA kit manufacturer's protocol. Plates were read at 450 nm using a Wallac Victor3 1420 counter (Perkin-Elmer) and Wallac 1420 Workstation software.

[0141] Western blotting analysis The cell pellet was thawed on ice. After thawing, the cell pellet was resuspended in RIPA buffer consisting of 25 mM Tris-HCl (pH 8.5), 150 mM NaCl, 1 mM EDTA (pH 8.0), 1% (v / v) NP-40 substitute, 0.5% (w / v) sodium deoxycholate, 0.1% (w / v) SDS, 10 mM NaF, 1 mM NaOV, 10 mM β-glycerophosphate, and 1× Halt Protease and PHospHotase Inhibitor Cocktail (Thermo Scientific). The lysate was incubated on ice for 30 min and then centrifuged at 17,200 g for 20 min at 4°C. The lysate supernatant was transferred to a new tube and anylized for total protein concentration by Braford protein assay. Laemmli sample buffer (BioRad) containing DTT (Sigma Aldrich) was added to increasing amounts of total protein and samples were boiled at 95°C for 5 min. Proteins were separated on 7.5% Criterion TGX Precast Midi Protein Gel (BioRad) using a Criterion Cell (BioRad) and transferred to a 0.2 μm nitrocellulose blotting membrane (Genesee) in Tris-Glycine Transfer Buffer (Thermo Scientific) using a Trans-Blot Turbo ElectropHoretic Transfer Cell (BioRad). The membrane was washed with deionized water, incubated in Ponceau S solution (Sigma Aldrich) to confirm protein transfer, and then washed for 1 min in Tris-buffered saline (TBST) containing 0.05% Tween 20 (Sigma Aldrich). The membrane was then blocked for 1 h at room temperature in blocking buffer containing 5% PHospHoBLOCKER blocking reagent (Cell Biolabs) in TBST. After blocking, the membrane was transferred to blocking buffer containing the primary antibody and incubated overnight at 4°C.All primary antibodies were supplied by Cell Signaling Technology and included actin (1:2000), p44 / 42 MAPK (ERK1 / 2) (1:1000), pp44 / 42 MAPK (pERK1 / 2) (1:1000), SLP76 (1:1000), pSLP76 (1:1000), PLCγ1 (1:1000), and pPLCγ1 (1:1000). Membranes were washed in TBST and then incubated for 45 min at room temperature in blocking buffer containing either anti-rabbit or anti-mouse HRP-conjugated secondary antibodies (Cell Signaling Technology). Membranes were washed in TBST and imaged on a ChemiDoc Imaging System using SuperSignal chemiluminescent substrate (Thermo 684 Scientific). In vivo studies All animal experiments were performed under protocols approved by the City of Hope Institutional Animal Care and Use Committee. For in vivo intraperitoneal (ip) tumor studies, OVCAR3 and OV90 cells (5.0 × 10 6 ) in a final volume of 500 μL HBSS- / - and implanted into >6-week-old female NSG mice by ip injection. For subcutaneous (sc) tumor studies, OV90 cells (5 × 10 6 ) were prepared in a final volume of 100 μL HBSS- / - and injected subcutaneously into the abdominal region of 6-8 week old female NSG mice. Tumor growth was monitored at least once a week by non-invasive bioluminescence imaging (Xenogen, LagoX) and flux signals were analyzed with Living Image software (Xenogen). For imaging, mice were intraperitoneally injected with 150 μL of D-luciferin potassium salt (Perkin Elmer) suspended in PBS at 4.29 mg / mouse. On day 8 for OV90 and day 14 for OVCAR3, mice were injected with the indicated T cells (5 × 10 6) i.p. Humane endpoints were used to determine survival. Mice were euthanized when there were signs of distress such as peritoneal ascites, labored or dyspnea, obvious weight loss, impaired motility, or abdominal distention as evidence of moribundity. At predetermined time points or in a moribund state, mice were euthanized and tissues and / or peritoneal fluid were harvested and processed for flow cytometry and immunohistochemistry as described herein. Peripheral blood was collected from isoflurane-anesthetized mice by retro-orbital (RO) bleeding through heparinized capillary tubes (Chase Scientific) into polystyrene tubes containing heparin / PBS solution (1000 units / mL, Sagent Pharmaceuticals). The total volume of each RO blood draw (approximately 120 μL / mouse) was recorded. Red blood cells (RBCs) were lysed with 1× red blood cell lysis buffer (Sigma) according to the manufacturer's protocol, then washed, stained, and analyzed by flow cytometry as described above. Cells from peritoneal ascites were collected from euthanized mice by intraperitoneal injection of 5 mL of cold 1× PBS, which was drawn up by a syringe and stored on ice until further processing. RBC-depleted ascites cells were washed, stained, and analyzed by flow cytometry using the antibodies and methods described above. Serum from immune competent mouse studies was collected from non-heparinized blood collected by RO blood draw as described above. Blood was kept at room temperature for 30 minutes, followed by centrifugation at 6000×g for 10 minutes at 4° C., then aliquoted into multiple tubes and frozen at −80° C. until used for serum cytokine ELISA or chemistry analysis. Serum chemistry analysis was performed by running samples on a VETSCAN® VS2 chemistry analyzer (Zoetis) using a phenobarbital chemistry panel rotor (Zoetis) for ALT and AST quantification as described by the manufacturer's protocol. Complete blood count (CBC) analysis on whole blood collected from RO blood draw was performed using a VETSCAN® HM5 Hematology Analyzer according to the manufacturer's protocol, as described previously.

[0142] immunohistochemistry Tumor tissues were fixed in 4% paraformaldehyde (4% PFA, Boston BioProducts) for up to 3 days and stored in 70% ethanol until further processing. Immunohistochemistry was performed by the Pathology Core at City of Hope. Briefly, paraffin-embedded sections (10 μm) were stained with hematoxylin & eosin (H&E, Sigma-Aldrich), mouse anti-human CD3 (DAKO), mouse anti-human CD4 (DAKO), mouse anti-human CD8 (DAKO), and mouse anti-human TAG72 (AB16838, Abcam). Images were acquired using a Nanozoomer 2.0 HT digital slide scanner and associated NDP.view2 software (Hamamatzu).

[0143] statistical analysis Data are presented as mean ± standard error of the mean (SEM) unless otherwise stated. Statistical comparisons between groups were performed using unpaired two-tailed Student's t-tests to calculate p-values ​​unless otherwise stated.

[0144] Further sequences Examples of nucleotide sequences useful in human mbIL-12 constructs are as follows:

[0145] [ka]

[0146] [ka] Examples of Nucleotide Sequences Useful in Mouse mbIL-12 Constructs Retroviral backbone promoter LTR / gag / pol / psi+ (specific for retroviral packaging plasmids):

[0147] [ka]

[0148] [ka] References

[0149] [Table 4]

[0150] [Table 5]

[0151] [Table 6] Other embodiments Although the present invention has been described in conjunction with the detailed description thereof, it should be understood that the above description is intended to be illustrative, and not limiting, of the scope of the invention as defined by the appended claims. Other aspects, advantages, and modifications are within the scope of the following claims.

Claims

1. A nucleic acid molecule encoding membrane-bound IL-12 (mbIL-12), comprising from the amino terminus to the carboxy terminus a mature human IL-12 p40 subunit, optionally a first peptide linker, a mature human IL-12 p35 subunit, optionally a second peptide linker, and a human CD28 transmembrane domain.

2. The nucleic acid molecule of claim 1, wherein mbIL-12 further comprises at least four consecutive amino acids of the cytoplasmic domain of human CD28 following the human CD28 transmembrane domain.

3. The nucleic acid molecule of claim 1, wherein the amino acid sequence represented by the mature human IL-12p40 subunit comprises the amino acid sequence represented by SEQ ID NO:

1.

4. The nucleic acid molecule of claim 1, wherein the amino acid sequence of mature human IL-12p35 comprises the amino acid sequence represented by SEQ ID NO: 2 or 3.

5. The nucleic acid molecule of claim 1 , wherein the first peptide linker comprises the amino acid sequence represented by SEQ ID NO:

4.

6. The nucleic acid molecule of claim 1 , wherein the second peptide linker comprises the amino acid sequence GGG.

7. The nucleic acid molecule of claim 1, wherein the human CD28 transmembrane domain comprises an amino acid sequence represented by SEQ ID NO: 16, 17 or 76.

8. The nucleic acid molecule of claim 1, wherein at least four consecutive amino acids of the cytoplasmic domain of human CD28 following the human CD28 transmembrane domain comprise the amino acid sequence represented by SEQ ID NO:

5.

9. The nucleic acid molecule of claim 1, wherein mbIL-12 comprises the amino acid sequence represented by SEQ ID NO: 6 or 7.

10. The nucleic acid molecule of claim 9, wherein the mbIL-12 further comprises a signal sequence.

11. The nucleic acid molecule of claim 10, wherein the signal sequence comprises the amino acid sequence represented by SEQ ID NO: 10 or SEQ ID NO:

11.

12. The nucleic acid molecule of claim 10, wherein mbIL-12 comprises the amino acid sequence represented by SEQ ID NO:8 or SEQ ID NO:

9.

13. The nucleic acid molecule according to claim 9, wherein mbIL-12 consists of the amino acid sequence represented by SEQ ID NO:6 or SEQ ID NO:

7.

14. The nucleic acid molecule according to claim 9, wherein mbIL-12 consists of the amino acid sequence represented by SEQ ID NO:8 or SEQ ID NO:

9.

15. A population of human immune cells comprising the nucleic acid molecule of claim 1.

16. 16. The population of human immune cells of claim 15, wherein the immune cells are T cells, natural killer cells, dendritic cells, macrophages, or monocytes.

17. The nucleic acid molecule of claim 1, further comprising a nucleotide sequence encoding a chimeric antigen receptor that targets a cancer antigen.

18. 16. The population of claim 15, further comprising a nucleic acid molecule comprising a nucleotide sequence encoding a chimeric antigen receptor that targets a cancer antigen.

19. The nucleic acid molecule of claim 17, wherein the cancer antigen is CD19, CD22, BCMA, PSCA, HER2, or TAG-72.

20. 20. A method for treating cancer comprising administering to a patient in need thereof the population of human T cells of claim 18.

Images