EA049204B1 - ANTIBODIES TO TGF-BETA AND THEIR APPLICATION - Google Patents

Abstract

The present invention provides an improved pan-specific TGF-β antibody for the treatment of TGF-β-mediated conditions, including autoimmune diseases, fibrotic conditions, and cancers. Methods and routes of using the antibody in combination with other immunomodulatory agents, such as an anti-PD-1 antibody, are also provided.

Description

Cross-reference to related applications

This application claims priority to U.S. Provisional Patent Application No. 62/448,800 and European Patent Application No. 17305061.8, both filed January 20, 2017. The disclosures of these two priority applications are incorporated herein by reference in their entireties.

List of sequences

The present invention contains a sequence listing that was filed electronically in ASCII format and is hereby incorporated by reference in its entirety. A copy of said ASCII file created on January 11, 2018 is named 022548_WO011_SL.txt and is 30458 bytes in size.

Background of the invention

Transforming growth factor beta (TGF-β) is a cytokine that controls many key cellular functions, including proliferation, differentiation, survival, migration, and epithelial-mesenchymal transition. It regulates various biological processes, such as extracellular matrix formation, wound healing, embryonic development, bone development, hematopoiesis, immune and inflammatory responses, and malignant transformation. Deregulation of TGFβ leads to pathological conditions, such as birth defects, cancer, chronic inflammation, and autoimmune and fibrotic diseases.

TGF-β has three known isoforms, TGF-β 1, 2, and 3. All three isoforms are initially translated as a propeptide. Following cleavage, the mature C-terminus remains associated with the N-terminus (termed the latency-associated peptide, or LAP), forming the small latency complex (SLC), which is secreted from the cell. Failure of SLC to bind the TGF-β receptor type II (TGFeRII) prevents receptor recruitment. Activation upon dissociation of the N- and C-termini occurs by one of several mechanisms, including proteolytic cleavage, exposure to acidic pH, or structural changes in the integrin (Connolly et al., Int J Biol Sci (2012) 8(7):964–78).

TGF-β 1, 2, and 3 are pleiotropic in function and have distinct expression profiles in different cell types and tissues. They have similar activities in vitro, but distinct knockouts in specific cell types suggest non-identical roles in vivo despite their shared ability to bind the same receptor (Akhurst et al., Nat Rev Drug Discov (2012) 11(10):790–811). Upon binding of TGF-β to TGFβRII, the constitutive kinase activity of the receptor mediates the phosphorylation and activation of TGFβRI, which phosphorylates SMAD2/3, allowing its association with SMAD4, nuclear localization, and transcription of TGF-β-responsive genes. Ibid. In addition to this canonical signaling cascade, noncanonical signaling pathways transmit signals through other factors, including p38 MAPK, PI3K, AKT, JUN, JNK, and NF-kB. TGF-β signaling is also modulated by other signaling pathways, including WNT, Hedgehog, Notch, INF, TNF, and RAS. Thus, the end result of TGF-β signaling is crosstalk among all of these signaling pathways that integrate the cell's state and environment. Ibid.

Given the diverse functions of TGF-β, there is a need for pan-specific therapeutic TGF-β antibodies that are safe for human patients (Bedinger et al., mAbs. (2016) 8(2):389–404). However, TGF-β is highly conserved across species. Therefore, generating human TGF-β antibodies in animals such as mice is challenging.

There is also a medical need for patients for whom there is currently no effective treatment. For example, more than 50% of patients with advanced melanoma in the phase III Checkmate-067 trial treated with monotherapy with the anti-PD1 antibody nivolumab did not experience a complete or partial response to therapy (Larkin et al., N Engl J Med (2015) 373:23–34; Redman et al., BMC Med (2016) 14:20–30).

Brief description of the invention

The present invention provides improved monoclonal antibodies that specifically bind to human TGF-β1, TGF-βΣ and TGF-β3 (i.e., pan-specific for TGF-β). These antibodies are less likely to form half-antibodies (i.e., a dimeric complex having one heavy chain and one light chain) during manufacture. They also have better pharmacokinetic profiles, such as increased half-life, and thus may provide improved clinical benefit to patients. The present inventors have also found that inhibition of TGF-β, such as by the antibodies and antigen-binding fragments thereof of the present invention, attenuates the immunosuppressive microenvironment of tumors and enhances the efficacy of immunotherapy, such as therapy targeting programmed cell death protein 1 (PD-1), PD-1 ligands 1 (PD-L1) and 2 (PD-L2).

In one aspect, the present invention provides an isolated monoclonal antibody that specifically binds to human TGF-β 1, TGF-βΣ and TGF^, comprising the heavy chain complementarity determining regions (CDRs) 1-3 of SEQ ID NO: 1 and the light chain CDRs 1-3 of SEQ ID NO:

- 1 049204

NO: 2, wherein the antibody comprises a human IgG ₄ constant region having a mutation at position 228 (EU numbering). In some embodiments, the mutation is a serine to proline mutation (S228P). In some embodiments, the antibody comprises a heavy chain variable domain (VH) amino acid sequence corresponding to residues 1-120 of SEQ ID NO: 1 and a light chain variable domain (VL) amino acid sequence corresponding to residues 1-108 of SEQ ID NO: 2. In further embodiments, the antibody comprises the heavy chain amino acid sequence shown as SEQ ID NO: 1 (with or without a C-terminal lysine) and the light chain amino acid sequence shown as SEQ ID NO: 2. The present invention also provides an antigen-binding F(ab') ₂ fragment of the aforementioned antibody.

In preferred embodiments, the antibody or fragment thereof of the present invention has an increased half-life, increased exposure, or both, compared to fresolimumab. For example, the increase is a 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, or 100% or greater increase. The exposure of a drug, such as an antibody or fragment thereof of the present invention, depends on the concentration of the drug in the body over time. The concentration of the drug in the body is often indicated by the level of the drug in the blood, blood plasma, or blood serum. The half-life and exposure (biological exposure) of a drug can be measured by well-known methods, as illustrated in Example 7 below.

The present invention further provides a composition comprising an antibody of the present invention, wherein the composition comprises less than 1% half-antibodies. The formation of half-antibodies can be determined by analyzing the purity of monoclonal antibody preparations using, for example, non-reducing SDS capillary electrophoresis, or non-reducing SDS-PAGE followed by densitometry, or RP-HPLC (Angal et al., Mol Immunol (1993) 30(1):105-8; Bloom et al., Protein Science (1997) 6:407-415; Schuurman et al., (2001) 38(1):l-8; and Solanos et al., Anal Chem (2006) 78:6583-94). In some embodiments, the composition is a pharmaceutical composition further comprising a pharmaceutically acceptable excipient.

In another aspect, the present invention provides a method of inhibiting TGF-β signaling in a human patient in need thereof, comprising administering to the patient a therapeutic amount of an antibody or fragment thereof of the present invention. In some embodiments, the patient has an immune-mediated disease (e.g., scleroderma), a fibrotic condition (e.g., a fibrotic condition of the kidneys such as focal segmental glomerulosclerosis (FSGS) or a fibrotic condition of the lungs such as idiopathic pulmonary fibrosis), or a congenital or bone defect (e.g., osteogenesis imperfecta). In some embodiments, the patient has cancer. In some embodiments, the antibody or fragment thereof used in the method inhibits the differentiation of CD4+ T cells into inducible regulatory T cells (iTreg). The antibody or fragment thereof can attenuate the immunosuppressive tumor microenvironment. This action of the antibody or fragment thereof helps activate the immune system and enhances the effectiveness of immunotherapy. Efficacy of the treatment methods described herein may be indicated by, for example, one or more of the following in a patient (e.g., in the patient's tumor tissue): (1) increases in MIP2 and/or KC/GRO levels, (2) activation or infiltration into tumor tissue of CD8+ T cells, such as INF-Y-positive CD8+ T cells, and (3) increases in the formation of natural killer (NK) cell aggregates.

The present invention further provides a method of treating cancer in a patient (human), comprising administering to the patient (1) a therapeutically effective amount of an antibody or fragment thereof of the present invention and (2) a therapeutically effective amount of an inhibitor of a protein that is a checkpoint of an immune response. These two agents can be administered simultaneously (e.g., in a single composition or in separate compositions) or sequentially in any order. The two agents can, for example, be administered on the same day. In some embodiments, therapeutic agent (1) is administered to the patient before (e.g., one or more days before) the administration of therapeutic agent (2).

In some embodiments, the immune checkpoint protein is PD-1, PD-L1, or PD-L2. In further embodiments, the inhibitor of the immune checkpoint protein is an anti-PD-1 antibody. In further embodiments, the anti-PD-1 antibody comprises (1) the heavy chain CDR1-3 of SEQ ID NO: 5 and the light chain CDR1-3 of SEQ ID NO: 6, (2) the VH amino acid sequence corresponding to residues 1-117 of SEQ ID NO: 5 and the VL amino acid sequence corresponding to residues 1-107 of SEQ ID NO: 6, or (3) the heavy chain amino acid sequence shown as SEQ ID NO: 5 (with or without a C-terminal lysine) and the light chain amino acid sequence shown as SEQ ID NO: 6. In one particular embodiment,

- 2 049204 embodiments, the method includes administering to a patient with cancer an antibody to TGF-β comprising the amino acid sequence of the heavy chain presented under SEQ ID NO: 1 (with or without a C-terminal lysine) and the amino acid sequence of the light chain presented under SEQ ID NO: 2, and an antibody to PD-1 comprising the amino acid sequence of the heavy chain presented under SEQ ID NO: 5 (with or without a C-terminal lysine) and the amino acid sequence of the light chain presented under SEQ ID NO: 6. In some embodiments, the patient is refractory to monotherapy with an antibody to PD-1. The patient may have advanced or metastatic melanoma or squamous cell carcinoma of the skin.

In some regimens, the anti-TGF-β antibody and the anti-PD-1 antibody are administered to the patient every 2 weeks or every 3 weeks. In some regimens, the two agents are administered at a dose of 0.01-40 (e.g., 0.02-20, 0.0515, or 0.05-20) mg/kg of body weight, respectively.

The present invention also provides a method of increasing the intensity of an immune response in a patient in need thereof, comprising administering to the patient an immune checkpoint inhibitor and an antibody or fragment thereof of the present invention. In some embodiments, the immune checkpoint inhibitor is an anti-PD-1 antibody, such as an antibody comprising (1) HCDR1-3 of SEQ ID NO: 5 and LCDR1-3 of SEQ ID NO: 6; (2) a VH and VL corresponding to residues 1-117 of SEQ ID NO: 5 and residues 1-107 of SEQ ID NO: 6, respectively; or (3) a heavy chain having the amino acid sequence of SEQ ID NO: 5 (with or without a C-terminal lysine) and a light chain having the amino acid sequence of SEQ ID NO: 6.

The methods of the present invention can be used to treat various forms of cancer, including but not limited to melanoma (e.g., metastatic melanoma or advanced melanoma), lung cancer (e.g., non-small cell lung cancer), squamous cell carcinoma of the skin, colorectal cancer, breast cancer, ovarian cancer, fallopian tube cancer, uterine cancer, head and neck cancer (e.g., squamous cell carcinoma of the head and neck), liver cancer (e.g., hepatocarcinoma), urothelial cancer, and kidney cancer (e.g., renal cell carcinoma). In some embodiments, the patient has a mesenchymal tumor or a solid tumor of the mesenchymal subtype. Examples of such a solid tumor include tumors in the colon (e.g., colorectal cancer), ovary, head and neck (e.g., squamous cell carcinoma of the head and neck), liver (e.g., hepatocellular carcinoma), and urothelium of the genitourinary system.

In some embodiments, a cancer, including a mesenchymal tumor, may be characterized by overexpression of one or more of ACTA2 (a2-smooth muscle actin), VIM (vimentin), MGP (Matrix Gla protein), ZWINT (kinetochore protein interacting with ZW10) and ZEB2 (zinc finger homeobox protein 2, binding to E-box). The expression levels of such biomarkers can be determined, for example, at the mRNA level or the protein level in a biological sample from a patient, such as a tumor biopsy or circulating tumor cells.

The present invention also provides the above-mentioned antibodies, fragments or compositions for use in the treatment of the conditions described herein, as well as the use of the above-mentioned antibodies, fragments or compositions in the manufacture of medicaments intended for the treatment of the conditions described herein.

Also included in the present invention are nucleic acid expression vectors encoding a heavy chain, a light chain, or both of an antibody of the present invention; host cells comprising sequences encoding a heavy chain and a light chain of the antibody; and methods for producing an antibody using host cells, comprising the steps of culturing the host cells in a suitable culture medium to allow expression of the antibody genes and then harvesting the antibody.

Brief description of graphic materials

Fig. 1A-E are graphs showing the effects of Ab1, fresolimumab, and 1D11 on the proliferation of mink lung cells (Mv 1 Lu) treated with 1 ng/ml human TGF-β 1 (A), human TGF-β2 (B), human TGF-β3 (C), mouse TGF-β 1 (D), or mouse TGF^2 (E). Antibody concentrations are expressed as μg/ml.

Fig. 2 is a bar graph showing the effect of 50 μg/ml Ab1 on differentiation into human inducible regulatory T cells (iTreg). The stimulation received by the T cells was the exposure to anti-CD3 and anti-CD28 antibodies together with IL-2.

Fig. 3 is a bar graph showing the effect of Ab1 on differentiation into human inducible regulatory T cells (iTreg) in human CD4+ T cell cultures treated with 2 ng/ml human TGF-β 1. The stimulation received by the T cells was the treatment with anti-CD3 and anti-CD28 antibodies together with IL-2.

Fig. 4 is a bar graph showing the effect of Ab1 (30 μg/ml) and human TGF-β1 (18 ng/ml) on NFATc-driven luciferase expression in Jurkat T cells after T cell stimulation and anti-PD-1 antibody treatment.

Fig. 5 is a graph showing the median tumor volumes with absolute

- 3,049,204 mean average deviations (MAD) in the indicated treatment groups using the MC38 colon tumor model in C57BL/6 mice. Vehicle: PBS. Anti-PD-1 antibody: Mab x-anti-mPD-1 (see detailed description of the invention below). Isotype control for Ab1: hIgG4 antibody to HEL.

Fig. 6 is a scatter plot showing the changes in tumor volume from baseline to day 27 after the indicated treatment using the MC38 colon tumor model in C57BL/6 mice. Control: PBS. Anti-PD-1 mIgG1 antibody RPM114: Mab x-anti-mPD-1.

Fig. 7A-F are graphs showing the change in tumor volumes over time for each indicated treatment group in the C57BL/6 mouse MC38 colon tumor model. Each line in the graphs represents one animal. mpk: mg/kg. Isotype control for Ab1: hIgG4 antibody to HEL. aPD1: Mab x-anti-mPD-1.

Fig. 8 is a graph showing the effects of Ab1 on the concentration of active TGF-e1 in tumor lysates from LoVo.

Fig. 9A is a graph showing the change in serum Ab1 and fresolimumab concentrations over time in five groups of rats given a single 5 mg/kg dose of either antibody. Groups (Gr) 1-3 were given three different lots (B1, B2, and B3) of fresolimumab. Groups 4 and 5 were given two different lots (B1 and B2) of Ab1.

Fig. 9B is a graph showing the change in serum Ab1 and fresolimumab concentrations over time in monkeys given a single 1 mg/kg dose of either antibody.

Fig. 9C is a graph showing the change in serum Ab1 and fresolimumab concentrations over time in monkeys given five weekly doses of Ab1 at 1 mg/kg per dose or biweekly doses of fresolimumab at 1 mg/kg per dose for the indicated study durations.

Fig. 9D is a graph showing the change in serum Ab1 and fresolimumab concentrations over time in monkeys given a single 10 mg/kg dose of either antibody.

Fig. 9E is a graph showing the change in serum Ab1 and fresolimumab concentrations over time in monkeys given five weekly doses of Ab1 at 10 mg/kg per dose or biweekly doses of fresolimumab at 10 mg/kg per dose for the indicated study durations.

Fig. 10A is a graph showing changes in TGF-e1 levels in MC38 tumors after treatment with Ab1 (+/- anti-PD1 antibody).

Fig. 10B is a graph showing changes in MIP-2 levels in MC38 tumors after treatment with Ab1 (+/- anti-PD1 antibody).

Fig. 10C is a graph showing the changes in KC/GRO levels in MC38 tumors after treatment with Ab1 (+/- anti-PD1 antibody).

Fig. 11A are graphs quantifying CellTrace violet staining and IFN-γ staining in CD8 positive cells.

Fig. 11B is a graph showing that Ab1 restored both proliferation and IFN-γ production in TGFe-treated CD8+ T cells.

Fig. 12A is a graph showing the relative abundance of CD8+ T cells (log2-transformed) within a panel of syngeneic mouse tumor models for colon cancer, leukemia, lung cancer, lymphoma, breast cancer, melanoma, mesothelioma, and renal cell carcinoma.

Fig. 12B is a graph showing activation of the TGFe signaling pathway within a panel of syngeneic mouse tumor models for colon cancer, leukemia, lung cancer, lymphoma, breast cancer, melanoma, mesothelioma, and renal cancer.

Detailed description of the invention

The present invention provides improved pan-specific monoclonal antibodies to TGF-β that are less prone to forming semi-antibodies and also have better pharmacokinetic profiles, such as higher in vivo exposure, than previously known antibodies. The antibodies of the present invention are collectively referred to as Ab1 and related antibodies and share common structural characteristics: they comprise heavy chain CDRs (HCDRs) 1-3 of SEQ ID NO: 1 and light chain CDRs (LCDRs) 1-3 of SEQ ID NO: 2 and comprise a human IgG ₄ constant region wherein residue 228 (EU numbering) in the hinge region has been mutated from serine to proline. P228 is boxed and highlighted in bold in the sequence of SEQ ID NO: 1 shown below.

The Ab1 antibody has an estimated molecular weight of 144 kDa in its unglycosylated form. The amino acid sequences of its heavy and light chains are SEQ ID NOs: 1 and 2, respectively. The two sequences are shown below. The variable domains are in italics. The CDRs are shown in boxes. The glycosylation site in the constant domain of the heavy chain is in bold.

- 4 049204

font (N297). QVQLVQSGAE \VIPIVDIAN}] VKKPGSSVKV SCKASGYTFS \SNVI^WVRQA PGQGLEWMG^ \AQRFKd[RVTI TADESTSTTY MELSSLRSED TAVYYCA^TL GLVLDAMD^W GQGTLVTVSS ASTKGPSVFP LAPCSRSTSE STAALGCLVK DYFPEPVTVS WNSGALTSGV HTFPAVLQSS GLYSLSSWT VPSSSLGTKT YTCNVDHKPS NTKVDKRVES kygppcp|p|cp APEFGLGPSV FLFPPKPKDT LMISRTPEVT CVWDVSQED PEVQFNWYVD GVEVHNAKTK PREEQFNSTY RWSVLTVLH QDWLNGKEYK CKVSNKGLPS SIEKTISKAK GQPREPQVYT LPPSQEEMTK NQVSLTCLVK GFYPSDIAVE WESNGQPENN YKTTPPVLDS DGSFFLYSRL TVDKSRWQEG NVFSCSVMHE ALHNHYTQKS LSLSLGK (SEQ ID NO :D ETVLTQSPGT LSLSPGERAT LSd[RASQSLG SSYL^WYQQK PGQAPRLLIY \gassral\gip DRFSGSGSGT DFTLTISRLE PEDFAVYYdQ__ QYADSPI1\FG QGTRLEIKRY VAAPSVFIFP PSDEQLKSGT ASWCLLNNF YPREAKVQWK VDNALQSGNS QESVTEQDSK DSTYSLSSTL

TLSKADYEKH KVYACEVTHQ GLSSPVTKSF NRGEC (SEQ ID NO: 2)

In some embodiments, antibodies of the present invention, such as antibodies to TGF-β, lack a C-terminal lysine in the heavy chain. The C-terminal lysine may be removed during manufacture or by recombinant technology (i.e., the sequence encoding the heavy chain does not contain a codon for a C-terminal lysine). Thus, antibodies comprising the amino acid sequence of the heavy chain under SEQ ID NO: 1 without the C-terminal lysine are also contemplated within the scope of the present invention.

Ab1 and related antibodies bind specifically to human TGF-β 1, -β2 and -β3. By specifically is meant that the binding is characterized by a K _D of less than 10 ^-7 M, such as less than 10 ^-8 M (e.g., 1-5 nM), as determined by, for example, surface plasmon resonance (see, for example, Example 1 below) or biolayer interferometry. Ab1 and related antibodies may also be characterized by potent TGF-β neutralizing potency as measured in a mink lung epithelial cell assay (see, e.g., Example 2 below) or an EC50 of about 0.05 to 1 μg/mL as determined in an IL-11 induction assay in A549 cells (see, e.g., Example 6 of PCT publication WO 2006/086469, the disclosure of which is incorporated herein by reference in its entirety).

These antigen binding and neutralizing properties of Ab1 and related antibodies are comparable to the previously known anti-TGF-β antibody frezolimumab (an IgG ₄ antibody PET1073G12 in a germline configuration, described in WO 2006/086469). The heavy and light chain sequences of frezolimumab, including the leader sequences, are shown as SEQ ID NOs: 3 and 4, respectively. As shown in SEQ ID NO: 3, frezolimumab lacks a proline at position 228 (EU numbering, which corresponds to the actual position 247 in SEQ ID NO: 3). Ab1 and related antibodies have several improved characteristics compared to frezolimumab.

During manufacture, frezolimumab can form up to 6-18% half-antibodies (i.e., dimers having a heavy chain and a light chain, rather than tetramers having two heavy chains complexed with two light chains) under non-reducing denaturing conditions. In contrast, Ab1 produces significantly fewer half-antibodies (< 1%). Thus, Ab1 and related antibodies yield purer drug during manufacture.

In addition, Ab1 and related antibodies may have improved pharmacokinetic (PK) profiles compared to fresolimumab. They may have a linear PK profile with a much longer half-life and a slower clearance rate than fresolimumab, resulting in approximately 1.7-fold increased in vivo exposure compared to fresolimumab. For example, in rats, Ab1 was shown to have a mean half-life of 7.1 days compared to 4.3 days for fresolimumab, and its clearance rate (CL) was 0.30 mL/hr/kg compared to 0.51 mL/hr/kg for fresolimumab (Example 7, below). In cynomolgus macaques, Ab1 was shown to have a mean half-life of 13 days compared to 4.5 days for fresolimumab, and its clearance rate (CL) was 0.40 mL/hr/kg compared to

- 5 049204

0.66 ml/h/kg for frezolimumab. Ibid. These improved PK properties suggest that Ab1 and related antibodies could be given to patients at a lower dose and/or less frequently than frezolimumab to achieve similar or better clinical efficacy, while causing fewer undesirable side effects and a less intense antibody response to the drug, allowing for longer treatment periods if needed.

In addition, toxicology studies with frezolimumab in non-human primates have shown a correlation between drug exposure and adverse events such as anemia. However, similar studies conducted with Ab1 at equivalent or even higher exposure levels have not shown such events.

Without being limited by any theory, the present inventors postulate that mutation of residue 228 in the heavy chain of Ab1 and related antibodies results in increased stability as well as improved PK and toxicology profiles.

The constant domain of Ab1 and related antibodies can be further modified if necessary, for example at residue L248 according to Kabat (e.g. by introducing the L248E mutation), to reduce any unwanted effector function of the molecule.

As used herein, the term antibody (Ab) or immunoglobulin (Ig) refers to a tetrameric protein comprising two heavy (H) chains (approximately 50-70 kDa) and two light (L) chains (approximately 25 kDa) linked together by disulfide bonds. Each heavy chain consists of a heavy chain variable domain (VH) and a heavy chain constant region (CH). Each light chain consists of a light chain variable domain (VL) and a light chain constant region ( _CL ). The V _H and VL domains can be further subdivided into regions of hypervariability, termed complementarity determining regions (CDRs), interspersed with more conserved regions, termed framework regions (FRs). Each VH or V _L consists of three CDRs and four FRs, arranged from amino terminus to carboxyl terminus in the following order: FR1, CDR1, FR2, CDR2, FR3, CDR3, and FR4. The assignment of amino acids to each region can be made according to the IMGT® definitions (Lefranc et al., Dev Comp Immunol 27(1):55–77 (2003)) or the definitions of Kabat, Sequences of Proteins of Immunological Interest (National Institutes of Health, Bethesda, MD (1987 and 1991); Chothia & Lesk, J. Mol. Biol. 196:901–917 (1987) or Chothia et al., Nature 342:878–883 (1989).

The term "human antibody" refers to an antibody in which the variable domain and constant region sequences are derived from human sequences. The term includes antibodies with sequences derived from human genes, wherein these sequences have been modified, for example, to reduce immunogenicity, increase affinity, and increase stability. The term includes antibodies produced recombinantly in non-human cells that may impart glycosylation that is not characteristic of human cells.

The term chimeric antibody refers to an antibody that contains sequences from two different animal species. For example, a chimeric antibody may contain the VH and VL of a mouse antibody (i.e., an antibody encoded by mouse antibody genes, such as an antibody produced from an immunized mouse using hybridoma technology) linked to the constant regions of an antibody from another species (e.g., human, rabbit, or rat).

The term antigen-binding fragment of an antibody refers to a fragment of an antibody that retains the ability to specifically bind to an antigen. In some embodiments, the antigen-binding fragment of the present invention is an F(ab') ₂ fragment, which is a bivalent fragment comprising two Fab fragments linked by a disulfide bridge at the hinge region (Fab is a monovalent fragment of an antibody consisting of VL, VH, CL, and CHi domains). In some embodiments, the antigen-binding fragment of the present invention may also comprise a _CH2 or _CH3 domain.

The antibodies and antigen-binding fragments described herein may be isolated. The term isolated protein, isolated polypeptide, or isolated antibody refers to a protein, polypeptide, or antibody that, by virtue of its origin or source of production, is (1) not associated with naturally associated components that accompany it in its native state, (2) is substantially free of other proteins from the same species, (3) is expressed by a cell from a different species, or (4) is not naturally occurring. Thus, a polypeptide that is chemically synthesized or synthesized in a cellular system different from the cell from which it naturally originates will be isolated from its naturally associated components. A protein can also be rendered substantially free of naturally associated components by isolation using protein purification techniques well known in the art.

I. Use of Ab1 and related antibodies.

The TGF-β receptor is widely expressed on immune cells, resulting in extensive effects of TGF-β in both the innate and adaptive immune systems. TGF-β has been associated with many diseases

- 6 049204 conditions, such as congenital defects, cancer, chronic inflammation, autoimmunity and fibrotic diseases. The therapeutic amount of Ab1 or a related antibody can be used to treat these conditions. A therapeutically effective amount refers to the amount of Ab1, a related antibody or other therapeutic agent mentioned herein that alleviates one or more symptoms of the condition being treated. This amount may vary depending on the condition or patient being treated and can be determined by a healthcare professional using well-established principles.

In some embodiments, Ab1 or a related antibody can be administered at a dose of 40, 20, or 15 mg/kg or less (such as 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, or 1 mg/kg). In some further embodiments, the doses can be 0.01, 0.02, 0.03, 0.04, 0.05, 0.1, 0.2, 0.3, 0.4, or 0.5 mg/kg. The frequency of dosing can be, for example, once a day, once every two, three, four, or five days, once a week, once every two weeks, or once every three weeks, once a month, or once every two months. The antibody may be administered intravenously (eg, by intravenous infusion over 0.5-8 hours), subcutaneously, locally, or by any other route of administration that is appropriate for the condition and dosage form.

Ab1 and related antibodies are derived from human antibody genes and thus have low immunogenicity in humans. Toxicology studies of Ab1 are described in detail in Example 8 below. Certain cardiac and pulmonary side effects have been observed in rats. Thus, adverse events can be monitored in patients when treating patients with Ab1 or related antibodies.

In some embodiments, the efficacy of the antibodies of the present invention may be indicated by one or more of the following in a patient (e.g., in diseased tissue, such as tumor tissue in a patient): (1) a decrease in the level or activity of TGF-β, (2) an increase in the levels of MIP2 and/or KC/GRO, (3) activation or infiltration into tumor tissue of CD8+ T cells, such as INF-Y-positive CD8+ T cells, and (4) an increase in the formation of natural killer (NK) cell aggregates.

A. Non-oncologic disease states.

Conditions that can be treated with Ab1 and related antibodies may include, but are not limited to, bone defects (e.g., osteogenesis imperfecta), glomerulonephritis, scarring of the nervous tissue or dermis, pulmonary fibrosis or pulmonary fibrosis (e.g., idiopathic pulmonary fibrosis), radiation-induced fibrosis, liver fibrosis, myelofibrosis, scleroderma, immune-mediated diseases (including rheumatoid arthritis, multiple sclerosis, systemic lupus erythematosus, Sjogren's syndrome, Berger's disease, and transplant rejection), and Dupuytren's contracture.

They may also be useful in the treatment, prevention, and reduction of the risk of renal failure, including but not limited to focal segmental glomerulosclerosis (FSGS), diabetic (type I and type II) nephropathy, radiation nephropathy, obstructive nephropathy, diffuse systemic sclerosis, inherited kidney disease (e.g., polycystic kidney disease, medullary sponge kidney, horseshoe kidney), glomerulonephritis, nephrosclerosis, nephrocalcinosis, systemic or glomerular hypertension, tubulointerstitial nephropathy, renal tubular acidosis, renal tuberculosis, and renal infarction. In particular, they are useful in combination with antagonists of the renin-angiotensin-aldosterone system, including but not limited to: renin inhibitors, angiotensin-converting enzyme (ACE) inhibitors, Ang II receptor antagonists (also known as Ang II receptor blockers) and aldosterone antagonists. See, for example, WO 2004/098637, the disclosure of which is incorporated herein by reference in its entirety.

Ab1 and related antibodies are useful for the treatment of diseases and conditions associated with ECM deposition, such as systemic sclerosis, postsurgical adhesions, keloid and hypertrophic scars, proliferative vitreoretinopathy, glaucoma drainage surgery, corneal injury, cataract, Peyronie's disease, adult respiratory distress syndrome, liver cirrhosis, post-myocardial infarction scarring, post-angioplasty restenosis, post-subarachnoid hemorrhage scarring, post-laminectomy fibrosis, post-tendon and post-repair fibrosis, biliary cirrhosis (including sclerosing cholangitis), pericarditis, pleurisy, tracheostomy, penetrating CNS injury, eosinophilia-myalgia syndrome, vascular restenosis, veno-occlusive disease, pancreatitis and psoriatic arthropathy.

Ab1 and related antibodies are additionally useful in conditions where stimulation of re-epithelialization is desirable. Such conditions include, but are not limited to, skin diseases such as venous ulcers, ischemic ulcers (bedsores), diabetic ulcers, transplant sites, transplant donor sites, abrasions and burns, bronchial epithelial diseases such as asthma, ARDS, intestinal epithelial diseases such as mucositis associated with cytotoxic treatment, esophageal ulcers (reflux disease), gastroesophageal reflux disease, gastric ulcers, small bowel and colon lesions (inflammatory bowel disease).

- 7 049204

Other uses of Ab1 and related antibodies include conditions in which endothelial cell proliferation is desirable, such as stabilization of atherosclerotic plaques, stimulation of vascular anastomotic healing, or conditions in which inhibition of smooth muscle cell proliferation is desirable, such as arterial disease, restenosis, and asthma.

Ab1 and related antibodies are also useful for enhancing the immune response to macrophage-mediated infections such as those caused by Leishmania spp., Trypanosoma cruzi, Mycobacterium tuberculosis and Mycobacterium leprae, as well as the protozoa Toxoplasma gondii, the fungi Histoplasma capsulatum, Candida albicans, Candida parapsilosis and Cryptococcus neoformans. They are also useful for reducing immunosuppression caused by, for example, tumors, AIDS or granulomatous diseases.

Ab1 and related antibodies are also useful for the prevention and/or treatment of ophthalmological conditions such as glaucoma and post-trabeculectomy scarring.

B. Oncological disease states.

TGF-β regulates several biological processes including cell proliferation, epithelial-mesenchymal transition (EMT), matrix remodeling, angiogenesis and immune functions. Each of these processes contributes to tumor progression. The broad detrimental role of TGF-β in patients with cancer across all indications is further supported by its increased levels both in the tumor microenvironment and systemically. See, e.g., Kadam et al., Mol. Biomark. Diagn. (2013) 4(3). Studies have shown that in the malignant state, TGF-β can induce EMT and the resulting mesenchymal phenotype leads to increased cell migration and invasion.

Ab1 and related antibodies are useful in the treatment of hyperproliferative diseases such as cancers including, but not limited to, skin cancer (e.g., melanoma, including unresectable or metastatic melanoma, squamous cell carcinoma of the skin, and keratoacanthoma), lung cancer (e.g., non-small cell lung cancer), esophageal cancer, gastric cancer, colorectal cancer, pancreatic cancer, liver cancer (e.g., hepatocellular carcinoma), primary peritoneal cancer, bladder cancer, renal or kidney cancer (e.g., renal cell carcinoma), urothelial carcinoma, breast cancer, ovarian cancer, fallopian tube cancer, cervical cancer, uterine cancer, prostate cancer, testicular cancer, head and neck cancer (e.g., squamous cell carcinoma of the head and neck), brain cancer, glioblastoma, glioma, mesothelioma, leukemia and lymphoma.

In some embodiments, Ab1 and related antibodies are useful in treating cancers in patients in whom prior therapy using an anti-PD-1 antibody, an anti-PD-L1 antibody, or an anti-PD-L2 antibody as a therapeutic agent has failed or is expected to fail, i.e., in patients who do not respond or are not expected to respond to therapy with an anti-PD-1 antibody, an anti-PD-L1 antibody, or an anti-PD-L2 antibody. In some embodiments, Ab1 and related antibodies are useful in treating cancers in patients who have relapsed after prior therapy with an anti-PD-1 antibody, an anti-PD-L1 antibody, or an anti-PD-L2 antibody. As used herein, the term "expected" means that a person skilled in the art can predict, without prescribing the therapy, whether the patient will respond or not respond and whether the therapy will fail or be ineffective, based on his/her general medical knowledge and the specific condition of the patient.

In some embodiments, the cancers are mesenchymal subtypes of solid tumors, including but not limited to mesenchymal colorectal cancer, mesenchymal ovarian cancer, mesenchymal lung cancer, mesenchymal head cancer, and mesenchymal neck cancer. Epithelial-mesenchymal transition (EMT) promotes cellular migration and invasive properties by downregulating epithelial cell gene expression and upregulating mesenchymal cell gene expression. EMT is a hallmark of tumor progression and invasion. Up to a quarter of colorectal and ovarian cancers are mesenchymal. Thus, by inhibiting TGF-β and inducing EMT, Ab1 or a related antibody can be used to treat mesenchymal solid tumors. Mesenchymal subtypes of solid tumors can be identified using a variety of genetic markers and pathological tests. Markers include ACTA2, VIM, MGP, ZEB2 and ZWINT, which can be detected by qRT-PCR or immunohistochemistry. Such markers can be used in selecting patients for monotherapy with an antibody to TGF-β or combination therapy of the present invention.

In some embodiments, Ab1 and related antibodies are useful in the treatment of patients with advanced solid tumors.

Ab1 and related antibodies may also be used in the treatment of hematopoietic disorders or malignancies such as multiple myeloma, myelodysplastic syndrome (MDS), Hodgkin lymphoma, non-Hodgkin lymphoma and leukemia, as well as various forms of sarcoma such as Kaposi's sarcoma.

Ab1 and related antibodies may also be useful for inhibiting cycle-mediated

- 8 049204 losporin progression of malignant neoplasm or cancer (eg, metastasis).

Of course, it is understood that in the context of cancer therapy, treatment includes any medical intervention that results in slowing the rate of cancer growth, delaying the progression or recurrence of cancer, or reducing the rate of cancer metastasis, as well as partial remission of cancer with the aim of increasing the patient's life expectancy.

S. Combination therapy in oncology.

The level of cytotoxic T cell infiltration in cancer has been observed to correlate with favorable clinical outcome (Fridman et al., Nat Rev Cancer (2012) 12(4):298–306; and Galon et al., Immunity (2013) 39(1):11–26). Furthermore, levels of T helper cells that promote cytotoxic T cells (CD4 ⁺ T _H1 ) and the cytokines they produce (eg, IFN-γ) often also correlate with favorable patient outcomes. In contrast, the presence of Treg cells has been shown to correlate with poor patient prognosis (Fridman, supra).

TGF-β suppresses virtually all aspects of the antitumor immune response. This cytokine promotes iTreg differentiation and reduces the proliferation and infiltration of cytotoxic (CD8+) cells. When TGF-β is inhibited by Ab1 or a related antibody, the immunosuppressive tumor microenvironment is attenuated as described above, leading to positive outcomes for cancer patients.

In addition, the inventors have found that by attenuating the immunosuppressive tumor microenvironment, Ab1 and related antibodies can provide checkpoint modulators, such as anti-PD-1 antibody, with the ability to better induce immune responses. As a result, more patients can benefit from immunotherapy, such as treatment with anti-PD-1 antibody, anti-PD-L1 antibody, or anti-PD-L2 antibody.

With or without therapeutic agents targeting immune checkpoint molecules, Ab1 and related antibodies can also be used in combination with other types of cancer therapies, such as chemotherapy (e.g., platinum- or taxoid-based therapies), radiation therapy, and therapies targeting cancer antigens or oncogenic factors.

Cancers that can be treated with a combination of an Ab1 or related antibody and an immune checkpoint inhibitor such as an anti-PD-1 antibody include the cancers listed in the subsection above.

In some embodiments, the cancers are refractory to prior therapy with an anti-PD-1 antibody, an anti-PD-L1 antibody, or an anti-PD-L2 antibody, such as advanced or metastatic melanoma, non-small cell lung cancer, renal cell carcinoma, squamous cell carcinoma of the head and neck, and Hodgkin's lymphoma. Refractory patients are patients whose disease progresses, as evidenced, for example, by radiography, within 12 weeks of starting treatment without any evidence of a response.

In some embodiments, Ab1 or a related antibody may be used in combination with another type of cancer therapy, such as anti-PD-1 antibody therapy, to treat mesenchymal cancers such as colorectal cancer, non-small cell lung cancer, ovarian cancer, bladder cancer, squamous cell carcinoma of the head and neck, renal cell carcinoma, hepatocellular carcinoma, and squamous cell carcinoma of the skin. See also discussions above.

Examples of anti-PD-1 antibodies include nivolumab, pembrolizumab, pidilizumab, MEDI0608 (formerly AMP-514; see, e.g., WO 2012/145493 and U.S. Patent No. 9,205,148), PDR001 (see, e.g., WO 2015/112900), PF-06801591 (see, e.g., WO 2016/092419), and BGB-A317 (see, e.g., WO 2015/035606). In some embodiments, anti-PD-1 antibodies include antibodies disclosed in WO 2015/112800 (such as antibodies referred to as H1M7789N, H1M7799N, H1M7800N, H2M7780N, H2M7788N, H2M7790N, H2M7791N, H2M7794N, H2M7795N, H2M7796N, H2M7798N, H4H9019P, H4xH9034P2, H4xH9035P2, H4xH9037P2, H4xH9045P2, H4xH9048P2, H4H9057P2, H4H9068P2, H4xH9119P2, H4xH9120P2, H4xH9128P2, H4xH9135P2, H4xH9145P2, H4xH8992P, H4xH8999P and H4xH9008P in Table 1 of the PCT publication, and the antibodies referred to as H4H7798N, H4H7795N2, H4H9008P and H4H9048P2 in Table 3 of the PCT publication). The disclosure of WO 2015/112800 is incorporated herein by reference in its entirety.

For example, the antibodies disclosed in WO 2015/112800 and related antibodies, including antibodies and antigen-binding fragments comprising the CDRs, V _H and V _L sequences, or heavy and light chain sequences disclosed in this PCT publication, as well as antibodies and antigen-binding fragments that bind to the same epitope of PD-1 as the antibodies disclosed in this PCT publication, can be used in combination with Ab1 or a related antibody of the present invention to treat cancer. In related embodiments, a useful anti-PD-1 antibody may comprise the heavy and light chain amino acid sequences shown below as SEQ ID NOs: 5 and 6, respectively; the V _H and V _L sequences of SEQ ID NOs: 5 and 6 (shown in italics); or one or more (e.g., all six) CDRs of SEQ ID NOs: 5 and 6 (shown in boxes).

- 9 049204

EVQLLESGGV LVQPGGSLRL SCAA^GFTFS NFdMTWVRQA PGKGLEWVSG \isgggrdi]yf ADSVKGRFTI SRDNSKNTLY LQMNSLKGED TAVYYC\VKW$ \niyfd^wgqg TLVTVSSASY KGPSVFPLAP CSRSTSESTA ALGCLVKDYF PEPVTVSWNS GALTSGVHTF PAVLQSSGLY SLSSWTVPS SSLGTKTYTC NVDHKPSNTK VDKRVESKYG PPCPPCPAPE FLGGPSVFLF PPKPKDTLMI SRTPEVTCW VDVSQEDPEV QFNWYVDGVE VHNAKTKPRE EQFNSTYRW SVLTVLHQDW LNGKEYKCKV SNKGLPSSIE KTISKAKGQP REPQVYTLPP SQEEMTKNQV SLTCLVKGFY PSDIAVEWES NGQPENNYKT TPPVLDSDGS

FFLYSRLTVD

KSRWQEGNVF SCSVMHEALH NHYTQKSLSL SLGK (SEQ

ID NO:5)

DIQMTQSPSS LSASVGDSIT ITCRA^LSIN _T^LNWYQQKP GKAPNLLI^ \a^slhggvps RFSGSGSGTD FTLTIRTLQP EDFATYY^QQ ssntpfy\fgp GTVVDFRRYV AAPSVFIFPP SDEQLKSGTA SWCLLNNFY PREAKVQWKV DNALQSGNSQ ESVTEQDSKD STYSLSSTLT LSKADYEKHK VYACEVTHQG LSSPVTKSFN RGEC (SEQ ID NO: :6)

In some embodiments, antibodies of the present invention, such as anti-PD-1 antibodies, lack a C-terminal lysine in the heavy chain. The C-terminal lysine may be removed during manufacture or by recombinant technology (i.e., the heavy chain encoding sequence does not contain a codon for a C-terminal lysine). Thus, antibodies comprising the heavy chain amino acid sequence of SEQ ID NO: 5 without the C-terminal lysine are also contemplated within the scope of the present invention.

In some embodiments, an anti-TGF-β antibody or fragment thereof of the present invention may also be used in combination with an antibody to an immunomodulatory antigen, such as PD-L1 and CTLA-4. Exemplary anti-PD-L1 antibodies include atezolizumab, avelumab, durvalumab, LY3300054, and BMS-936559. Exemplary anti-CTLA-4 antibodies include ipilimumab or tremelimumab.

D. Biomarkers of treatment effectiveness.

The efficacy of Ab1 and related antibodies can be determined by biomarkers or by target occupancy. For example, in tumor tissues, target occupancy can be analyzed by assessing the levels of active TGF-β in biopsies using the Meso Scale Discovery (MSD) assay. In blood, target binding can be analyzed by assessing the effect of reduced circulating TGF-β on peripheral blood mononuclear cells such as lymphocytes (T cells, B cells, NK cells) and monocytes. For example, increased proliferation of circulating CD8+ T cells can be assessed using CD45+RO+CCR7+CD28+ Ki67 ⁺ as flow cytometric markers. Activation of circulating NK cells can be assessed using CD3 ^- CD56 ^{, CD16} +, or CD137+ as flow cytometric markers. Additionally, Ki-67, PD-1 and ICOS can be used as PD markers associated with T cell activation.

Immune modulation following treatment with Ab1 or a related antibody can be analyzed by assessing changes in infiltrating immune cells and immune markers using multiplex immunohistochemistry (IHC) assays using, for example, the NeoGenomics platform. Specifically, the NeoGenomics MultiOmyx TIL panel stains for a panel of immune markers, allowing for the quantification of the density and localization of various immune cells. Immune markers can indicate iTreg differentiation; CD8 ⁺ T cell infiltration and proliferation; and IFN-γ production by CD8 ⁺ T cells. Ab1 has been shown to inhibit the differentiation of CD4 ⁺ T cells into iTreg (see, for example, Example 3, below) and to increase CD8 ⁺ T cell proliferation and IFN-γ production (as demonstrated in a mixed lymphocyte culture reaction assay; data not shown). Thus, the efficacy of treatment with Ab1 or a related antibody may be indicated by inhibition of iTreg, induction of CD8 ⁺ T cell proliferation and infiltration into tumor or other diseased tissues, increased IFN-γ production, and/or an increased ratio of CD8 ⁺ T cells to Treg cells. Immune modulation by treatment with Ab1 or a related antibody can also be analyzed in peripheral blood by quantitative immune cell counting using methylation-sensitive PCR for CD8 ⁺ T cells, Tregs, NK cells, and other immune cells. Treatment efficacy may manifest clinically as a delay or reversal of disease progression, such as tumor progression.

II. Methods for obtaining antibodies.

Ab1 and related antibodies, as well as antibodies targeting other co-targets such as PD-1, PD-L1 or PD-L2, can be produced using methods well known in the art.

- 10 049204 nicks. DNA sequences encoding heavy and light chains of antibodies can be inserted into expression vectors such that the genes are operably linked to necessary expression control sequences, such as transcription and translation control sequences. Expression vectors include plasmids, retroviruses, adenoviruses, adeno-associated viruses (AAV), plant viruses such as cauliflower mosaic virus, tobacco mosaic virus, cosmids, YACs, EBV-derived episomes, etc. The antibody light chain encoding sequence and the antibody heavy chain encoding sequence can be inserted into separate vectors and can be operably linked to the same or different expression control sequences (e.g., promoters). In one embodiment, both coding sequences are inserted into the same expression vector and may be operably linked to the same expression control sequence (e.g., a common promoter), to separate identical expression control sequences (e.g., promoters), or to different expression control sequences (e.g., promoters). The antibody encoding sequences may be inserted into the expression vector by standard methods (e.g., by ligation at complementary restriction sites in the antibody gene fragment and the vector, or by blunt-end ligation if restriction sites are not present).

In addition to the antibody chain genes, the recombinant expression vectors may carry regulatory sequences that control the expression of the antibody chain genes in the host cell. Examples of regulatory sequences for expression in mammalian host cells include viral elements that direct high levels of protein expression in mammalian cells, such as promoters and/or enhancers derived from the LTRs of retroviruses, cytomegalovirus (CMV) (such as the CMV promoter/enhancer), simian virus 40 (SV40) (such as the SV40 promoter/enhancer), adenovirus (such as the adenovirus major late promoter (AdMLP)), polyoma virus, and strong mammalian promoters such as the native immunoglobulin and actin gene promoters.

In addition to the antibody chain genes and regulatory sequences, the recombinant expression vectors of the present invention may carry additional sequences, such as sequences that regulate replication of the vector in host cells (e.g., origins of replication) and selectable marker genes. For example, a selectable marker gene confers resistance to drugs, such as G418, hygromycin, or methotrexate, to a host cell into which the vector has been introduced. Selectable marker genes may include the dihydrofolate reductase (DHFR) gene (for use in dhfr- host cells for methotrexate selection/amplification), the neo gene (for G418 selection), and the glutamate synthetase gene.

Expression vectors encoding the antibodies of the present invention are introduced into host cells for expression. The host cells are cultured under conditions suitable for expression of the antibody, which is then harvested and isolated. Host cells include mammalian, plant, bacterial, or yeast host cells. Mammalian cell lines available as expression hosts are well known in the art and include a variety of immortalized cell lines available from the American Type Culture Collection (ATCC). These include, but are not limited to, Chinese hamster ovary (CHO) cells, NS0 cells, SP2 cells, HEK-293T cells, 293 Freestyle cells (Invitrogen), NIH-3T3 cells, HeLa cells, baby hamster kidney (BHK) cells, African green monkey kidney (COS) cells, human hepatocellular carcinoma cells (e.g., Hep G2), A549 cells, and a number of other cell lines. Cell lines can be selected based on their expression levels. Other cell lines that can be used are insect cell lines such as Sf9 or Sf21 cells.

Additionally, antibody expression can be enhanced using a number of known techniques. For example, the glutamine synthetase gene expression system (GS system) is a common approach for enhancing expression under certain conditions.

The tissue culture medium for the host cells may or may not contain animal derived components (ADCs), such as bovine serum albumin. In some embodiments, ADC-free culture media are preferred for human safety. The tissue culture may be performed using a fed-batch culture method, a continuous perfusion method, or any other method suitable for the host cells and the desired yield.

III. Pharmaceutical compositions.

The antibody of the present invention can be formulated to provide suitable storage stability. For example, the antibody can be lyophilized, or stored, or diluted for use with pharmaceutically acceptable excipients. For combination therapy, two or more therapeutic agents, such as antibodies, can be formulated together, such as mixed and provided in a single composition.

The term excipient or carrier is used herein to describe any ingredient other than the compound(s) of the present invention. The choice of excipient(s) will depend to a large extent on factors such as the particular route of administration.

- 11 049204 conditions, the effect of the filler on solubility and stability and the nature of the dosage form. A pharmaceutically acceptable filler includes any possible solvents, dispersion media, coatings, antibacterial and antifungal agents, isotonic and absorption delaying agents, etc., which are physiologically compatible. Some examples of pharmaceutically acceptable fillers are water, saline, phosphate-buffered saline, dextrose, glycerol, ethanol, etc., as well as combinations thereof. In some cases, isotonic agents will be included in the composition, for example, sugars, polyhydric alcohols such as mannitol, sorbitol, or sodium chloride. Further examples of pharmaceutically acceptable substances are wetting agents or minor amounts of auxiliary substances such as wetting or emulsifying agents, preservatives or buffers that enhance the stability or effectiveness of the antibody.

The pharmaceutical composition of the present invention may be prepared, packaged, or sold in bulk form, as a single unit dose, and/or as a plurality of single unit doses. As used herein, a unit dose is a discrete amount of a pharmaceutical composition containing a predetermined amount of an active ingredient. The amount of active ingredient is typically equal to the dose of the active ingredient administered to a subject, or a convenient fraction of such a dose, such as, for example, one-half or one-third of such a dose.

The pharmaceutical compositions of the present invention are generally suitable for parenteral administration. As used herein, parenteral administration of a pharmaceutical composition includes any route of administration characterized by physically breaking the tissue of a subject and introducing the pharmaceutical composition through the tissue break, which typically results in direct administration into the bloodstream, into a muscle, or into an internal organ. Parenteral administration thus includes, without limitation, administering the pharmaceutical composition by injection of the composition, by introducing the composition through a surgical incision, by introducing the composition through a penetrating non-surgical wound of tissue, and the like. In particular, parenteral administration is contemplated to include, without limitation, subcutaneous, intraperitoneal, intramuscular, intrasternal, intravenous, intraarterial, intrathecal, intraventricular, intraurethral, intracranial, intratumoral, and intraarticular injection or infusion; as well as renal dialysis infusion techniques. Regional perfusion is also contemplated. Preferred embodiments may include intravenous and subcutaneous routes.

Formulations based on a pharmaceutical composition suitable for parenteral administration typically comprise the active ingredient in combination with a pharmaceutically acceptable carrier, such as sterile water or sterile isotonic saline solution. Such formulations can be prepared, packaged or sold in a form suitable for bolus administration or for continuous administration. Injectable formulations can be prepared, packaged or sold in unit dosage form, such as ampoules or multi-dose containers containing a preservative. Formulations for parenteral administration include, but are not limited to, suspensions, solutions, emulsions in oily or aqueous vehicles, pastes and the like. Such formulations can further contain one or more accessory ingredients, including, but not limited to, suspending, stabilizing or dispersing agents. In one embodiment of the parenteral formulation, the active ingredient is provided in dry (i.e., powder or granular) form for reconstitution with a suitable carrier medium (e.g., sterile, pyrogen-free water) prior to parenteral administration of the diluted composition.

Formulations for parenteral administration also include aqueous solutions which may contain excipients such as salts, carbohydrates and buffering agents (e.g., at a pH of 3 to 9), but for some routes of administration they may be more conveniently formulated as a sterile non-aqueous solution or as a dried form for use in conjunction with a suitable carrier medium such as sterile, pyrogen-free water. Illustrative parenteral forms include solutions or suspensions in sterile aqueous solutions, for example, aqueous propylene glycol or dextrose solutions. Such dosage forms may be suitably buffered if necessary. Other parenteral formulations which are useful include those which contain the active ingredient in microcrystalline form or as a liposomal preparation. Formulations for parenteral administration may be formulated for immediate and/or modified release. Modified release formulations include delayed, sustained, intermittent, controlled, targeted and programmed release formulations.

IV. Illustrative embodiments.

Additional specific embodiments of the present invention are described as follows.

1. An isolated monoclonal antibody that specifically binds to TGF-βΓ TGF-e2 and

- 12 049204

Human TGF-β3 comprising the heavy chain complementarity determining regions (CDRs) 1-3 of SEQ ID NO: 1 and the light chain CDRs 1-3 of SEQ ID NO: 2, wherein the antibody comprises a human IgG ₄ constant region containing a proline at position 228 (EU numbering).

2. The antibody of embodiment 1, wherein the antibody comprises an amino acid sequence of a heavy chain variable domain (VH) corresponding to residues 1-120 in SEQ ID NO: 1 and an amino acid sequence of a light chain variable domain (VL) corresponding to residues 1-108 in SEQ ID NO: 2.

3. The antibody of embodiment 2, wherein the antibody comprises the heavy chain amino acid sequence shown as SEQ ID NO: 1 (with or without a C-terminal lysine) and the light chain amino acid sequence shown as SEQ ID NO: 2.

4. An antigen-binding fragment of an antibody according to embodiment 3, wherein the fragment is F(ab')2.

5. An antibody or fragment thereof according to any one of embodiments 1-4, wherein the antibody or fragment thereof has an increased half-life or increased exposure compared to fresolimumab.

6. An antibody or fragment thereof according to any one of embodiments 1-5, wherein the antibody or fragment thereof is characterized by one or more of the following properties:

a) inhibit the differentiation of CD4+ T cells into inducible regulatory T cells (iTreg);

b) increase the proliferation of CD8+ T cells;

c) increase the formation of natural killer (NK) cell clusters;

d) increase the level of MIP-2 and

e) increase the level of KC/GRO.

7. A composition comprising an antibody or fragment thereof according to any one of embodiments 1-6, wherein the composition comprises less than 1% semi-antibodies.

8. An antibody or fragment thereof according to any one of embodiments 1-6 as a medicinal product.

9. A method for inhibiting TGF-β signaling in a patient in need thereof, comprising administering to the patient a therapeutic amount of an antibody or fragment thereof according to any one of embodiments 1-6.

10. The method according to embodiment 9, wherein the patient has cancer.

11. The method according to embodiment 10, wherein the cancer is selected from the group consisting of melanoma, lung cancer, squamous cell carcinoma of the skin, colorectal cancer, breast cancer, ovarian cancer, head and neck cancer, hepatocellular carcinoma, urothelial cancer, and renal cell carcinoma.

12. The method according to embodiments 10 or 11, wherein the cancer is characterized by overexpression of one or more of ACTA2, VIM, MGP, and ZWINT.

13. The method according to any one of embodiments 10-12, wherein the cancer is a mesenchymal tumor.

14. The method according to any one of embodiments 10-13, wherein the antibody or fragment thereof attenuates the immunosuppressive microenvironment of a tumor.

15. A method of treating cancer in a patient, comprising administering to the patient (1) an antibody or fragment thereof according to any one of embodiments 1-6 and (2) an inhibitor of a protein that is an immune checkpoint.

16. The method according to embodiment 15, wherein the immune checkpoint protein is PD-1, PD-L1, or PD-L2.

17. The method of embodiment 16, wherein the immune checkpoint protein inhibitor is an anti-PD-1 antibody.

18. The method according to embodiment 17, wherein the anti-PD-1 antibody comprises the heavy chain CDR1-3 of SEQ ID NO: 5 and the light chain CDR1-3 of SEQ ID NO: 6.

19. The method according to embodiment 17, wherein the anti-PD-1 antibody comprises a V _H amino acid sequence corresponding to residues 1-117 of SEQ ID NO: 5 and a V _L amino acid sequence corresponding to residues 1-107 of SEQ ID NO: 6.

20. The method according to embodiment 17, wherein the anti-PD-1 antibody comprises the heavy chain amino acid sequence shown as SEQ ID NO: 5 (with or without a C-terminal lysine) and the light chain amino acid sequence shown as SEQ ID NO: 6.

21. The method according to any one of embodiments 15-20, wherein the TGF-β antibody comprises the heavy chain amino acid sequence shown as SEQ ID NO: 1 (with or without a C-terminal lysine) and the light chain amino acid sequence shown as SEQ ID NO: 2.

22. The method according to any one of embodiments 15-21, wherein the cancer is refractory to treatment with an anti-PD-1 antibody.

23. The method according to any one of embodiments 15-22, wherein the cancer is chalk

- 13 049204 late stage or metastatic melanoma or squamous cell carcinoma of the skin.

24. The method according to any one of embodiments 15-23, wherein the cancer is a mesenchymal subtype of a solid tumor.

25. The method according to any one of embodiments 15-24, wherein the cancer is characterized by overexpression of one or more of ACTA2, VIM, MGP, and ZWINT.

26. The method according to any one of embodiments 15-25, wherein the cancer is selected from the group consisting of melanoma, lung cancer, squamous cell carcinoma of the skin, colorectal cancer, breast cancer, ovarian cancer, head and neck cancer, hepatocellular carcinoma, urothelial cancer, and renal cell carcinoma.

27. The method according to any one of embodiments 15-26, wherein the antibody or fragment thereof attenuates the immunosuppressive microenvironment of a tumor.

28. The method according to any one of embodiments 15-27, wherein the anti-TGF-β antibody and the anti-PD-1 antibody are administered to the patient on the same day.

29. The method according to any one of embodiments 15-28, wherein the anti-TGF-β antibody and the anti-PD-1 antibody are administered to the patient once every two weeks.

30. The method according to any one of embodiments 15-29, wherein the anti-TGF-β antibody and the anti-PD-1 antibody are administered, respectively, at a dose of 0.05-20 mg/kg body weight.

31. A method for increasing the intensity of an immune response in a patient in need thereof, comprising administering to the patient an immune checkpoint inhibitor and an antibody or fragment thereof according to any one of embodiments 1-6.

32. The method of embodiment 31, wherein the checkpoint inhibitor is an anti-PD-1 antibody.

33. The method of embodiment 32, wherein the anti-PD-1 antibody comprises:

a) HCDR1-3 from SEQ ID NO: 5 and LCDR1-3 from SEQ ID NO: 6;

b) VH and Vl corresponding to residues 1-117 of SEQ ID NO: 5 and residues 1-107 of SEQ ID NO: 6, respectively; or

c) a heavy chain with the amino acid sequence presented under SEQ ID NO: 5 (with or without C-terminal lysine) and a light chain with the amino acid sequence presented under SEQ ID NO: 6.

34. The method according to any one of embodiments 31-33, wherein the TGF-β antibody comprises the heavy chain amino acid sequence shown as SEQ ID NO: 1 (with or without a C-terminal lysine) and the light chain amino acid sequence shown as SEQ ID NO: 2.

35. The method according to any one of embodiments 31-34, wherein the patient has cancer.

36. The method of embodiment 35, wherein the patient is refractory to prior treatment with an immune checkpoint inhibitor and/or has a solid tumor of the mesenchymal subtype.

37. The method according to embodiments 35 or 36, wherein the cancer is selected from the group consisting of melanoma, lung cancer, squamous cell carcinoma of the skin, colorectal cancer, breast cancer, ovarian cancer, head and neck cancer, hepatocellular carcinoma, urothelial cancer, and renal cell carcinoma.

38. The method according to any one of embodiments 35-37, wherein the cancer is characterized by overexpression of one or more of ACTA2, VIM, MGP, and ZWINT.

39. The method according to any one of embodiments 35-38, wherein the antibody or fragment thereof attenuates the immunosuppressive microenvironment of a tumor.

40. An antibody or fragment thereof according to any one of embodiments 1-6 for use in treating a patient using any of the above methods.

41. Use of an antibody or fragment thereof according to any one of embodiments 1-6 for the manufacture of a medicinal product intended for treating a patient using any of the above methods.

42. An isolated nucleic acid molecule comprising a nucleotide sequence encoding the heavy chain, the light chain, or both of an antibody or fragment thereof according to any one of embodiments 1-6.

43. An expression vector comprising an isolated nucleic acid molecule according to embodiment 42.

44. A host cell comprising an expression vector according to embodiment 43.

45. A method for producing an antibody or antigen-binding fragment thereof according to any one of embodiments 1-6, wherein the method comprises:

obtaining a host cell containing a first and second nucleotide sequences encoding, respectively, a heavy chain and a light chain of an antibody or an antigen-binding fragment thereof, growing the host cell under conditions that ensure the production of the antibody or an antigen-binding fragment thereof, and

- 14 049204 isolation of an antibody or its antigen-binding fragment.

46. A method for producing a pharmaceutical composition, comprising producing an antibody or antigen-binding fragment thereof according to any one of embodiments 1-6 and mixing the antibody or antigen-binding fragment thereof with a pharmaceutically acceptable excipient.

47. An article or kit comprising an antibody or antigen-binding fragment thereof according to any one of embodiments 1-6 and another therapeutic agent.

48. The article or kit of embodiment 47, wherein the other therapeutic agent is an immune checkpoint inhibitor described herein.

The present invention is further described in the following examples, which do not limit the scope of the present invention described in the claims.

Examples

For a better understanding of the present invention, the following examples are provided. These examples are provided for illustrative purposes only and should in no way be construed as limiting the scope of the present invention.

Example 1. TGF-β binding properties of Ab1.

The affinity of Ab1 for all human and mouse TGF-β isoforms was determined by surface plasmon resonance on a Biacore T200 Biosensor (GE Healthcare) using a carboxymethylated dextran (CM5)-coated S-series chip. Ab1 at a series of concentrations (1.11, 3.33, 10, and 30 nM) was loaded onto immobilized recombinant TGF-β to measure the binding interaction in real time. TGF-β homodimers were immobilized at low density to reduce avidity effects. Loadings were performed in triplicate and binding assays were repeated three times. Data from kinetic experiments were analyzed using Biacore T200 BIAevaluation v2.0 software. The resulting sensorgrams were interpolated to zero, equalized, double-matched to the standard, and trimmed for curve-fitting analysis using a 1:1 binding model to determine the association rate constant ( _ka ), dissociation rate constant (kd), and equilibrium dissociation constant (KD).

Recombinant proteins were produced in-house (human TGF-β 1, 2, and 3) or purchased from R&D Systems (mouse TGF-β 1 and 2). Table 1 below shows the amino acid sequence homology of the three active TGF-β isoforms between rhesus monkeys, mice, or rats and humans (homology is given as the percentage of conserved amino acids relative to the total number of amino acids).

Table 1

Homology of active TGF-β isoforms to human

Homology percentage

TGF-β1 TGF-β2 TGF-β3 Rhesus macaque 100% (112/112) 100% (112/112) 100% (112/112) Mouse 99, 1% 97% (109/112)* 100% (112/112) (111/112)* Rat 99, 1% 97% (109/112)* 99, 1% (111/112)* (111/112)

*Rat and mouse TGFβ1 are 100% homologous to each other, and rat and mouse TGFβ2 are 100% homologous to each other

Because human TGF-β)3 and mouse TGF-β)3 are identical in amino acid sequence, separate affinity measurements were not made for these two proteins. Similarly, mouse and rat TGF-β1 and 2 are identical in amino acid sequence, and separate affinity measurements were also not made.

The _ka , kd, and KD values of Ab1 determined by the above method are presented in Table 2 below. The KD values of Ab1 for human TGF-β 1, 2, and 3 were determined to be 1.48, 3.00, and 1.65 nM, respectively. The KD values of Ab1 for mouse/rat TGF-β 1 and 2 were determined to be 2.80 and 1.88 nM, respectively. These binding properties are similar to those of fresolimumab.

Table 2

Equilibrium constants and affinity of Ab1 for TGF-β

Human Mice and rats Human and mouse

TGF-βΙ TGF-32 TGF-βΙ Τ6Ε-β2 Τ6Ε-β3

- 15 049204

ka (x 10 ⁵ M~ 3, 11 ± 2.86 ± 2.98 ± 3.48 ± 2.23 ± ic- ¹ ) 0,3 0,3 0, 07 0.24 0, 6 kd (x 10- 3.43 ± 8.23 ± 8.35 ± 6.45 ± 3.72 ± ⁴ C-!) 2.9 5, 0 0.38 1.44 2,3 K _D (nM) 1.48 ± 3, 00 ± 2.80 ± 1, 88 ± 1, 65 ± 1,1 2.0 0, 08 0.56 1,1

The above data demonstrate that Ab1 is a potent and selective pan-specific inhibitor of TGF-β. Surface plasmon resonance measurements demonstrate that Ab1 has an affinity of 1 to 5 nM for all human and murine TGF-β isoforms. The high level of specificity was confirmed in tissue cross-reactivity studies using GLP-compliant immunohistochemistry (IHC) using normal rat, cynomolgus monkey and human tissues.

Example 2. TGF-β neutralizing potency of Ab1.

The in vitro potency of Ab1 in terms of TGF-β neutralizing activity was measured in a cell-based assay. This assay measured the ability of TGF-β to inhibit the proliferation of non-transformed mink lung epithelial cells (Mv 1 Lu cells). See, e.g., WO 2006/086469 and Mazzieri et al., Eds, Methods in Molecular Biology, Vol. 142, Transforming Growth Factor-β Protocols. The ability of Ab1, fresolimumab, and 1D11 (a mouse anti-TGF-β antibody whose heavy and light chain sequences are disclosed herein under SEQ ID NOs: 9 and 10) to neutralize human TGF-β 1, 2, 3 and mouse TGF-β 1 and 2 was assessed. Recombinant TGF-β proteins were produced in-house (human TGF-β 1, 2, and 3) or purchased from R&D Systems (mouse TGF-β 1 and 2).

All human and mouse TGF-β isoforms inhibited mink lung cell proliferation in a dose-dependent manner over the range of 0.02 pg/mL to 10 ng/mL. To quantify the potency of Ab1, fresolimumab, and 1D11, 1 ng/mL of the indicated TGF-β and antibody at serial dilutions were incubated with mink lung cells. After three days of incubation, cell proliferation was quantified using CyQUANT, a dye that fluoresces upon binding to DNA (Fig. 1A–E). The data demonstrate that Ab1, fresolimumab, and their mouse mimic 1D11 inhibited all human and mouse TGF-β isoforms to similar extents.

Example 3. Inhibition of differentiation into inducible regulatory T cells by Ab1.

Regulatory T cells (Tregs) are immunosuppressive, and their levels correlate with adverse outcomes in patients with cancer. In the study described below, we examined whether Ab1 could inhibit TGF-β-induced differentiation of human CD4+ T cells into inducible regulatory T cells (iTregs). Primary human CD4+ T cells were isolated from healthy normal donors. Human TGF-β 1 was purchased from R&D Systems.

To study the antagonist activity of Ab1 on TGF-β produced endogenously by cultured cells, total CD4+ T cells without addition of exogenous TGF-β were treated with 50 μg/ml isotype control (human IgG4-kappa anti-hen egg white lysozyme (HEL), Crown Bioscience), Ab1, or fresolimumab in the presence or absence of stimulation (anti-CD3, anti-CD28, and IL-2) for 6 days, followed by flow cytometric analysis. The mean percentage and standard deviation of the CD25+FOXP3+ population were calculated in the initial population (lymphocytes/live/single cells/CD4 ^{+ CD127-} ) in triplicate. When total human CD4+ T cells were stimulated with anti-CD3, anti-CD28, and IL-2, the percentage of FOXP3+CD25+ (iTreg) in culture increased from 0% to 15%. Treatment with 50 μg/ml Ab1 or 50 μg/ml fresolimumab reduced the percentage of iTreg to a similar extent (to 8% and 7%, respectively; Fig. 2). In contrast, treatment with human IgG ₄ (hIgG ₄ ) as an isotype control had a minimal effect on iTreg differentiation (20% iTreg) (Fig. 2). CD4 ⁺ T cells isolated from a second healthy normal volunteer gave similar results.

To study the antagonist activity of Ab1 against exogenous TGF-β, total CD4 ⁺ T cells incubated with 2 ng/ml human TGF-β 1 were treated with isotype control, Ab1, or fresolimumab at different antibody concentrations in the presence or absence of stimulation (with anti-CD3, anti-CD28, and IL-2) for 6 days and analyzed by flow cytometry. The mean percentage and standard deviation of the CD25 ⁺ FOXP3 ⁺ population were calculated in the parent population (lymphocytes/live/single cells/CD4 ⁺ CD127 ^- ) in triplicate except where noted. When exogenous TGF-β 1 (2 ng/ml) was added to the stimulated total CD4 ⁺ T cell population, the percentage of iTreg in the culture increased from 15% to 55%. When treated with increasing concentrations of Ab1, the percentage of iTreg decreased in parallel from 55% to 15% at

- 16 049204

200 μg/ml and to 43% at 6.25 μg/ml. Frezolimumab treatment reduced the percentage of iTreg to the same extent as Ab1 (from 55% to 16% at 200 μg/ml and to 32% at 6.25 μg/ml). Treatment with isotype control antibody at various concentrations had no effect on the percentage of iTreg: 60% at 200 μg/ml and 6.25 μg/ml. See Fig. 3. CD4' T cells isolated from a second healthy normal volunteer gave similar results.

This study demonstrates that Ab1 inhibits TGF-β-induced iTreg differentiation and thus may provide clinical benefit by attenuating the immunosuppressive tumor microenvironment.

Example 4. Effects of the combination of Ab1 antibody and anti-PD-1 antibody in vitro.

In this study, we investigated whether TGF-β could prevent maximal T cell stimulation in vitro following treatment with an anti-PD-1 antibody, and if so, whether Ab1 could counteract this prevention. Luciferase expression from an expression construct under the transcriptional control of NFATc (cytoplasmic nuclear factor of activated T cells 1) regulatory sequences was used to measure the level of T cell activation.

For this study, we used a cell-based assay system purchased from Promega. This system contains two cell types: 1) Jurkat T cells expressing human PD-1 and a luciferase reporter gene driven by an NFAT response element, and 2) CHO-K1 cells expressing human PD-L1 and an engineered cell surface protein designed to activate a cognate T cell receptor in an antigen-independent manner. Following co-culture, Jurkat T cells interact with CHO-K1 cells, causing T cell receptor stimulation and translocation of NFATc to the nucleus, where it drives luciferase expression. However, upon engagement of PD-1/PD-L1 (SHP2), non-receptor tyrosine protein phosphatase type 11 is recruited to the T cell receptor complex, inhibiting the nuclear translocation of NFATc and subsequent luciferase expression. Blocking PD-1 signaling attenuates SHP2-dependent suppression and thus allows maximal luciferase expression. Thus, this system provides a functional method to determine the effect of TGF-β on T cell signaling and the effect of Ab1 on T cell processing by anti-PD-1.

Due to the low kinetic parameters associated with TGF-β-dependent effects, Jurkat T cells were pretreated with TGF-β prior to TCR stimulation. Human TGFβ1 was purchased from R&D Systems. The isotype control antibody for Ab1 (hIgG ₄ anti-HEL) was purchased from Crown Bioscience (Cat# C0004-5). Mouse IgG anti-hPD-1 and its isotype control were purchased from BioLegend (Cat# 329912). Fourteen replicates were analyzed for each sample.

Results showed that addition of hPD-1 antibody to Jurkat T cells co-cultured with CHO-K1 cells for 24 h induced luciferase activity (865,794 relative luminescence units [RLU]) to a greater extent than addition of isotype control (234,963 RLU, fold change=3.685, p-value<0.0001) or no antibody at all (206,043 RLU, fold change=4.202, p-value<0.0001). Pretreatment of Jurkat T cells with 18 ng/ml TGF-β1 for 12 days induced less luciferase activity (638,866 RLU) in co-culture with CHO-K1 cells in the presence of hPD-1 antibody compared to Jurkat T cells not treated with TGF-β1 (865,794 RLU, fold change = -1.355, p-value < 0.0001) (Fig. 4).

To assess the antagonist potency of Ab1, Jurkat T cells were pretreated with 18 ng/ml TGF-β1 in the presence of Ab1, isotype control Ab, or no Ab for 12 days and then co-cultured with CHO-K1 cells in the presence of anti-PD-1 antibody for 24 h. The presence of Ab1 (924,186 RLU) attenuated TGF-β-dependent suppression of luciferase activity compared with isotype control Ab (639,440 RLU, fold change=1.445, p-value<0.0001) and no Ab control (638,866 RLU, fold change=1.447, p-value<0.0001). In the Ab1 (975,654 RLU) or isotype control (955,717 RLU) control groups added to TGF-β1-naïve Jurkat T cells co-cultured with CHO-K1 cells in the presence of PD-1 Ab, statistically significantly increased luciferase activity was observed compared to the no-Ab control but with minimal fold change (865,794 RLU, fold change=1.127 and 1.104, p-value=0.0023 and 0.001284, respectively) (Fig. 4). RLU values are also presented in Table 3 below.

- 17 049204

Table 3

Relative luminescence units in T cell activation assays

AЬ to PD-1 Isotype control for Ab to PD-1 Carrier medium (PBS) Without pre-treatment with TGF-βΙ With pretreatment with TGF-βΙ Without AY 865794 638866 234963 206043 AЫ 975654 924186 - - Antibody h!gG ₄ to HEL 955717 639440 - -

To exclude the possibility that pretreatment of Jurkat T cells with TGF-β1 might result in decreased proliferation or viability and hence decreased luciferase activity during 24 h co-culture with CHO-K1 cells, we incubated Jurkat T cells with 18 ng/ml TGF-β1 or PBS in the presence of Ab1, anti-HEL hIgG ₄ antibody, or no antibody (with vehicle) for 7 days. Every 2-3 days, an equal volume from each group was used to seed a new flask, during which both TGF-e1 and antibody were refreshed (for a total of two subculture events). Final culture evaluation demonstrated that viability was unchanged across treatment groups (ranging from 94% to 96%). Furthermore, the total number of Jurkat T cells in the final culture in each treatment group was very similar (ranging from 25 to 29 million).

The above study demonstrates that the increase in T cell receptor downstream signaling following anti-PD-1 antibody treatment is suppressed by TGF-β, resulting in suboptimal T cell stimulation. Our data suggest that inhibition of TGF-β attenuates the immunosuppressive tumor microenvironment and allows checkpoint modulators such as anti-PD-1 antibody agents to better induce immune responses, thereby increasing the proportion of patients benefiting from immuno-oncology treatment.

Example 5. Effects of the combination of Ab1 antibody and anti-PD-1 antibody in vivo

The present inventors then examined the effects of combination treatment with an anti-TGF-β antibody and an anti-PD-1 antibody in a C57BL/6 mouse cancer model.

Tolerability/preliminary safety assessment.

Materials and methods.

The tolerability of Ab1 and monoclonal antibody (mAb) to mouse PD-1 (mPD-1) as monotherapy and in combination was assessed in female C57BL/6 mice. Ab1 (10, 20, and 50 mg/kg) or isotype control Ab (hIgG ₄ anti-HEL antibody purchased from Crown Bioscience; administered at 10 and 20 mg/kg) were administered IV every three days (Q3D) as monotherapy or in combination with anti-PD1 mAb at 5 mg/kg IV twice weekly. The anti-PD1 Ab used in this study was designated antimPD1_hyb_RMP114_mIgG1LCfullrat (or Mab x-anti-mPD-1). It was a chimeric rat antibody to mPD-1 generated by replacing the rat Fc region of rat IgG _2a clone RMP1-14 (BioXCell, Cat#BE0146) with the Fc region of mouse IgG _1. The amino acid sequences of the heavy and light chains of this chimeric antibody are shown under SEQ ID NOs: 7 and 8. Tolerability was assessed by measuring animal body weights and clinical observations. At the end of the three-week treatment period, final sampling was performed four hours after the last treatment and tissues (heart, kidney, liver, lung, and spleen) were fixed in formaldehyde and submitted for histopathological analyses.

A dose was considered excessively toxic if it caused 15% body weight loss over three consecutive days in an individual mouse, 20% body weight loss in one day, or 10% or more drug-related deaths, unless tumor-induced cachexia resulting in body weight loss was observed in the vehicle-treated control group. Animal body weight included tumor weight.

Toxicity/safety analysis results.

A tolerability study in C57BL/6 mice showed that all tested dose levels of Ab1 and Mab x-anti-mPD-1 as monotherapy and in combination were well tolerated. No significant changes in body weight were observed in any treatment group at any dose tested. No severe or serious clinical events were observed during the study. Histopathological analyses revealed increased lymphocyte counts in the spleen (in the white pulp) in all

- 18,049,204 treatment groups, including those treated with the isotype control antibody, regardless of the combination and without any dose dependence. No other significant microscopic findings were observed. Two mice in the isotype control Ab (10 mg/kg) and anti-PD-1 Mab (5 mg/kg) combination group were found dead after the final administration on the last day of the study. Histopathological analyses did not reveal any drug-related causes of death.

Efficiency research.

The effects of combination treatment with an anti-TGF-β antibody and an anti-PD-1 antibody were evaluated in C57BL/6 mice bearing subcutaneous syngeneic MC38 colon tumors. Mice were treated with Ab1 at 25 mg/kg, mAb x-αnti-mPD-1 at 5 mg/kg, or both Q3D for three weeks. This study demonstrates that the combination of Ab1 and mAb to mPD-1 had significantly greater antitumor activity than monotherapy alone. The materials and methods and data from this study are described in detail below.

Materials and methods.

Animals.

Female C57BL/6 mice were obtained from Charles River Labs (Wilmington, MA, USA). Animals were allowed to acclimate for at least three days prior to study inclusion. Mice were 11 weeks old and weighed between 17.0 and 20.9 g at baseline. They had free access to food (Harlan 2916 rodent chow, MA, USA) and sterile water and were maintained on a 12-hour light/dark cycle.

Tumor cells.

MC38 is a colon adenocarcinoma cell line. Cells were obtained from the National Cancer Institute (Bethesda, MD, USA) and cultured in 5% _CO2 at 37°C in complete medium (CM) containing Roswell Park Memorial Institute medium (RPMI)-1640 with L-glutamine (Gibco, Cat. #11875) supplemented with 10% heat-inactivated fetal bovine serum (HI FBS) (Gibco, Cat. #10438026). Cells were harvested and resuspended in Dulbecco's phosphate-buffered saline (DPBS) (Gibco, Cat. #14190), and 1x106 cells/200 µl per mouse were implanted subcutaneously (SC) into the right flank of female C57BL/6 mice.

Connections.

Ab1 was administered to animals in an aqueous solution. It was filtered through PES with a pore diameter of 0.22 μm and stored as sterile aliquots at 2-10°C. The antibody was given to animals intraperitoneally (IP) at a dose of 10 ml/kg, 25 mg/kg.

Anti-HEL hIgG ₄ antibody (Crown Bioscience) was used as an isotype control for Ab1. This antibody was given to control animals IP at 10 ml/kg, 25 mg/kg.

Mab x-anti-mPD-1 (above) was provided in DPBS (Gibco, Cat. #14190-094) and given to animals IP at 10 ml/kg, 5 mg/kg.

Research plan.

On day 0, 60 animals were implanted with MC38 tumor cells. On day 8 post-implantation, mice with an average tumor size of 50-75 ^mm3 were pooled and randomly assigned to control and treatment groups (ten mice per group). Treatments with vehicle medium (PBS, pH 7.2), anti-HEL hIgG ₄ , Ab1, and anti-mPD-1 Mab at the doses described above were initiated on day 9 and repeated on days 12, 15, 18, 21, and 27. Animals treated with vehicle and anti-HEL hIgG ₄ served as controls. Mice were examined daily and adverse clinical reactions were noted. Individual mice were weighed three to four times per week until the end of the experiment.

Mice were euthanized if morbidity or weight loss >20% was observed. Tumors were measured with calipers twice weekly until sacrificed at the end. When tumor size reached approximately 2000 ^mm3 or animal health problems occurred (ulceration of 20% of tumor area), animals were euthanized and the date of death was recorded. Solid tumor volumes were estimated from the two tumor measurements and calculated according to the following equation:

Tumor volume ( ^mm3 )=[length (mm) ^xwidth2 ( ^mm2 )]/2

The median percent regression for the group on a given day was then obtained by taking the median of the individual percent regressions calculated for each animal in the group on that day. The calculation day was defined as the day on which the RT/LS (i.e., the ratio of the median changes in tumor volume from baseline between the treatment and control groups) was calculated, unless the median percent regression did not reflect activity in the group. In that case, the day was defined as the first day on which the median percent regression was highest. Regression was defined as partial (PR) if tumor volume was reduced to 50% of the tumor volume at the start of treatment. Complete regression (CR) was achieved if tumor volume was less than 14 ^mm3 or could not be recorded.

- 19 049204

Efficiency.

The primary efficacy endpoints were changes from baseline in tumor volume as indicated by LT/LS, median percent regression, partial regression, and complete regression. Changes in tumor volume for each treatment group (T) and control group (C) were calculated for each animal each day by subtracting the tumor volume on the day of the first treatment (day of staging) from the tumor volume on the specified observation day. The median LT was calculated for the treatment group, and the median LS was calculated for the control group. The LT/LS ratio was calculated and expressed as a percentage:

LT/LS=(median delta-T/median delta-C)x100

An AT/AC ratio of <40% was considered indicative of therapeutic activity. An AT/AC ratio of 0% was considered indicative of tumor growth cessation. An AT/AC ratio of <0% was considered indicative of tumor regression.

Percent tumor regression was defined as the percentage reduction in tumor volume in the treatment group on a given observation day compared to the tumor volume at baseline (to). At a given time point (t) and for each animal, percent regression was calculated using the following formula:

% regression (at time 1) = |(volume _y - volume t/volume _t0 ]x 100

The median % regression for the group on a given day was then calculated by taking the median of the individual % regression values calculated for each animal in the group. The calculation day was determined by the day on which AT/AC was calculated, except in cases where the median % regression did not reflect the activity in the group. In this case, the day was determined by the first day on which the median % regression was highest.

Statistical analysis.

For changes from baseline in tumor volume, a two-way ANOVA with treatment and day as factors (repeated measures) was performed. In case of statistically significant treatment*day interaction or statistically significant treatment effects, a Holm-Bonferroni-corrected multiplicity contrast analysis was then performed comparing all treatment groups to the control group on each day from day 8 to day 27. Changes from baseline in tumor volume were calculated for each animal and each day by subtracting the tumor volume on the day of the first treatment (day 8) from the tumor volume on the specified observation day.

Because there was heterogeneity in variances between groups, a compound symmetry (CS) covariance structure with group= function was chosen for the ANOVA model (SAS Institute Inc. (2008), SAS/STAT 9.2 User's Guide, Cary, NC). In Figs. 5 and 6, the medians and absolute deviation of the median (MAD) for each group are presented for each measurement day. In Tables 4–6 below, the medians and normalized MAD values (nMAD=1.4826*MAD) for each group are presented for each measurement day. All statistical analyses were performed using SAS software version v9.2. A probability of less than 5% (p<0.05) was considered statistically significant.

Results of the efficiency analysis.

Treatment of tumor-bearing C57BL/6 mice with Ab1, anti-PD-1 Mab, or a combination of the two antibodies was also well tolerated and nontoxic, as evidenced by the general health and activity of the animals and the absence of significant changes in body weight. As monotherapy, Ab1 at 25 mg/kg Q3D and anti-PD-1 Mab at 5 mg/kg Q3D induced body weight loss with maximum reductions of only 3.4% (day 9) and 2.1% (day 9), respectively. The combination of Ab1 (25 mg/kg Q3D) and anti-PD-1 Mab (5 mg/kg Q3D) was also well tolerated, showing body weight loss with maximum reductions of 1.3% (day 9) (Table 4).

As monotherapy, Ab1 (25 mg/kg Q3D) and anti-PD-1 Mab (5 mg/kg Q3D) did not demonstrate tumor growth impairment compared to animals treated with the isotype control for Ab1 (anti-HEL hIgG ₄ antibody). The AT/AC ratio on day 27 of treatment was 93% and 109%, respectively (Table 4). The combination of anti-PD-1 Mab and anti-HEL hIgG ₄ antibody demonstrated minimal antitumor activity with an AT/AC of 31% on day 27 of treatment (which was not statistically different from the control group) and complete regression observed in only 2 of 10 mice. However, the combination of anti-PD-1 and Ab1 MAb demonstrated superior antitumor activity from day 15 to day 27 with AT/AC -1 at day 27 of treatment (which was statistically significantly different from the control group) and complete regression observed in 6 out of 10 mice (Table 4).

- 20 049204

Table 4

Ab1 and Mab X-anti-mPD-1 activity in the MC38 cancer model in C57BL/6

Remedy AW* (day of maximum decline) AT/AC, % (day 27) Median regression (day 27) Value* * (day 27) Regression (day 27) PR CR Carrier medium -1.8 (D9) 100 - - 0/10 0/10 Antibody h!gG ₄ to HEL - 104 - 0.7914 1/10 1/10 (25 mg/kg) AЫ -3.4 (D9) 93 — 0.9953 0/10 0/10 (25 mg/kg) x-anti-mPD-1 -2.1 (D9) 109 — 0,5435 2/10 1/10 (5 mg/kg) AЫ (25 mg/kg) + -1.3 (D9) -1 24.5 <0.0001 6/10 6/10 x-anti-mPD-1 (5 mg/kg) Antibody h!gG ₄ to HEL (25 mg/kg) + -3.0 (D9) 31 - 0,4020 2/10 2/10 x-anti-mPD-1 (5 mg/kg)

* ΔW denotes the average change in body weight in % per group at maximum loss.

* * P-values were obtained by contrast analysis comparing each treatment group with the control using Holm-Bonferroni correction for multiplicity after two-way ANOVA with repeated measures of changes in tumor volume from baseline. A probability of less than 5% (p < 0.05) was considered statistically significant.

Tables 5 and 6 and Figs. 5–7 provide additional data demonstrating the activity of antibodies alone or in combination on tumor volumes in the mouse model.

Table 5

Comparison of treatment groups with control group

Changes in tumor volume from baseline, ^mm3 : median (nMAD), η and p-value* Processing group Generally Day 12 Day 15 Day 19 Day 2 3 Day 27 Carrier medium 26.0 (20.02) n=10 71.5 (53.37) n=10 220, 5 (117,87) n=10 528.0 (157.16) n=10 1610.0 (413.65) n=9

- 21 049204

Antibody h!gG4 to HEL (25 mg/kg) 0.9629 17.0 (17.79) n=10 0.5197 65.5 (51, 89) n=10 1.0000 182.5 (113.42) n=10 0.8858 491.5 (358.05) n=10 1.0000 1675.0 (621.21) n=9 0.7914 Abi (25 mg/kg) 0.9629 17.0 (15.57) n=10 0.4754 79.0 (43.74) n=10 1.0000 218.0 (83.03) n=10 0.8858 567.0 (344.70) n=10 1.0000 1496.5 (472.21) n=8 0.9953 Mab X-anti- mPDl (5 mg/kg) 0.2439 -2.0 (21.50) n=10 0.0224 38.5 (68.20) n=10 0.1235 147.5 (149.74) n=10 0.2259 458.0 (280.21) n=10 0.2228 1748.0 (851.01) n=9 0.5435 AY (25 mg/kg) + Mab x-anti- mPD-1 (5 mg/kg) 0,0007 -3.5 (10.38) n=10 0.0743 20.0 (20.02) n=10 0.0121 -0.5 (69.68) n=10 <.0001 -17.0 (51.15) n=10 <.0001 -9.0 (86.73) n=10 <.0001 Antibody h!gG4 to HEL (25 mg/kg) + x-anti-m-PD-1 (5 mg/kg) 0.2531 19, 0 (37.81) n=10 0.5197 61.5 (65.98) n=10 0.2848 145.0 (160.12) n=10 0.2259 266.0 (398.82) n=10 0.1979 503, 0 (841.38) n=8 0.4020

** P-values were obtained using contrast analysis for comparison with control on each day with Holm-Bonferroni correction for multiplicity after two-way ANOVA for changes in tumor volume from baseline

Table 6

Ab1 and Mab X-anti-mPD-1 as monotherapy versus combination

Changes in tumor volume relative to baseline, ^mm3 : (nMAD), η and p-value* median Processing group Generally Day 12 Day 15 Day 19 Day 23 Day 27 AA (25 mg/kg) + Mab x-anti-mPD-1 (5 mg/kg) - -3.5 (10.38) n=10 20.0 (20.02) n=10 -0.5 (69.68) n=10 -17, 0 (51.15) n=10 -9.0 (86.73) n=10 X-anti-mPD-1 - -2.0 38,5 147.5 458, 0 1748,0

- 22 049204

(5 mg/kg) 0,040 5 (21.50) n=10 1.0000 (68.20) n=10 0.7631 (149.74) n=1 0.0276 (280.21) n=10 0.0004 (851.01) n=9 0.0024 Abi (25 mg/kg) 0,001 7 17.0 (15.57) n=10 1.0000 79.0 (43.74) n=10 0.0694 218.0 (83.03) n=10 0.0007 567.0 (344.70) n=10 <.0001 1496.5 (472.21) n=8 <.0001 Antibody h!gG ₄ to HEL (25 mg/kg) + Mab x-anti-mPD-1 (5 mg/kg) - 19, 0 (37.81) n=10 61.5 (65.98) n=10 145.0 (160.12) n=10 266, 0 (398,82) n=10 503, 0 (841,38) n=8 Mab X-anti-mPD-1 (5 mg/kg) 0, 868 0 -2.0 (21.50) n=10 0.2024 38.5 (68.20) n=10 0.7631 147.5 (149.74) n=10 0.9476 458.0 (280.21) n=10 0.8576 1748.0 (851.01) n=9 0.8491 Antibody h!gG ₄ to HEL (25 mg/kg) 0.569 1 17.0 (17.79) n=10 1.0000 65.5 (51, 89) n=10 0.5743 182.5 (113.42) n=10 0.4761 491.5 (358.05) n=10 0.2357 1675.0 (621.21) n=9 0.8491

* P-values were obtained by contrast analysis comparing the combinations of Ab1, hIgG4 antibody to HEL, and x-anti-mPD-1 with each of these agents as monotherapy at the dose included in the combination on each day with Holm-Bonferroni correction for multiplicity after two-way ANOVA for changes in tumor volume from baseline

The data in the tables and figures show that the combination of Ab1 at 25 mg/kg Q3D and Mab x-anti-mPD-1 at 5 mg/kg Q3D produced greater antitumor effect than either antibody at these doses. This difference was statistically significant with p-values at days 19, 23, and 27 of 0.0007, <0.0001, and <0.0001, respectively, when comparing the combination with Ab1 alone. This difference was also statistically significant with p-values at days 19, 23, and 27 of 0.0276, 0.0004, and 0.0024, respectively, when comparing the combination with Mab x-anti-mPD-1 alone (Table 6). For the group receiving the combination of hIgG ₄ antibody to HEL at a dose of 25 mg/kg Q3D and Mab x-anti-mPD-1 at a dose of 5 mg/kg Q3D, the treatment effect on changes in tumor volume from baseline was not statistically significantly different from the effect of either agent alone on any day of measurement.

Thus, the combination of Ab1 at 25 mg/kg Q3D and mPD-1 Mab at 5 mg/kg Q3D produced significantly greater antitumor effect than either agent alone from day 15 to day 27.

In another study, we evaluated the antitumor activity of a combination of Ab1 at 1, 10, or 25 mg/kg and mouse anti-PD-1 at 5 mg/kg against a subcutaneous MC38 murine colon cancer model in C57BL/6J mice. Exponentially growing MC38 colon adenocarcinoma cells (NCI, Frederick, MD) were cultured in RPMI-1640 supplemented with 10% FBS in a humidified 5% CO2 incubator and then implanted subcutaneously (1 X 10 ⁶ cells) in the flank of female C57BL/6J mice (Jackson Laboratory, Bar Harbor, ME). After tumors reached an average size of 50-75 mm ³ , mice were pooled and randomly assigned to control and treatment groups (10 mice per group). Tumor-bearing mice were then treated by intraperitoneal administration of PBS, IgG4 isotype control antibody (25 mg/kg), or Ab1 (1, 10, and 25 mg/kg) three times weekly until each animal had received a total of 6 to 7 doses. Tumors were measured with digital calipers twice weekly, and tumor volumes were calculated ( ^mm3 L x W x H) and plotted using GraphPad Prism. Mice were euthanized with CO2 at the end of the study if tumors grew to >2000 ^mm3 or if tumors demonstrated ulceration of >20% of the tumor surface.

As monotherapy, Ab1 at 25 mg/kg Q3D and mouse a-PD-1 antibody at 5 mg/kg demonstrated partial activity with complete regression in 2/8 and 4/8 mice bearing PD-1, respectively.

- 23 049204 MC38. The combination of Ab1 at 1, 10, or 25 mg/kg Q3D and murine a-PD-1 antibody at 5 mg/kg Q3D was therapeutically active. At day 24 post-implantation, when comparing changes in tumor volume from baseline, the effect of the combination of Ab1 at all doses tested and murine a-PD-1 antibody at 5 mg/kg Q3D was superior to the effect of either agent as monotherapy, with complete regression in 5/8, 6/8, and 7/8 mice for 1, 10, and 25 mg/kg Ab1, respectively. Table 6A provides a summary of the results.

Table 6A

Antitumor effects of Ab1+mAb combinations to mPD-1

Group Processing Total number of mice Number of cases of complete response/regression (frequency of cases of complete response/regression) 1. PBS 25 mg/kg isotype control antibody for Abi 8 0 (0%) 2 . + 5 mg/kg isotype control antibody for Mab to mPD-1 25 mg/kg Abi + 8 0 (0%) 3 . 5 mg/kg antibody isotype control for Mab to mPD-1 25 mg/kg antibody isotype control 8 2 (25%) 4 . for Abi+ 5 mg/kg Mab x-anti-mPD-1 1 mg/kg Abi 8 4 (50%) 5. + 5 mg/kg Mab x-anti-mPD-1 10 mg/kg Abi 8 5 (62.5%) 6. + 5 mg/kg Mab x-anti-mPD-1 25 mg/kg Abi 8 6 (75%) 7 . + 5 mg/kg Mab x-anti-mPD-1 8 7 (87.5%)

Thus, these preclinical data demonstrate that the combination of PD-1 inhibition and TGF-β inhibition is capable of providing tumor growth inhibition to a greater extent than checkpoint inhibitor blockade alone.

Example 6. Intratumoral TGF-β levels 1.

Intratumoral TGF-β1 levels were studied in a subcutaneous colorectal cancer xenograft model of LoVo in BALB/c mice. Mice were injected with Ab1 or isotype control Mab at 10, 25, or 50 mg/kg intravenously every 3 days for a total of eight IV administrations, starting when tumor volume was less than 100 mm ³ .

- 24 049204

Tumor samples stored at -80°C in 2-mL plastic tubes with 2.8-mm ceramic beads (MoBio 13114-50) were thawed at room temperature. One milliliter (mL) of cold Meso Scale Diagnostic (MSD) Tris-Lysis Buffer (R60TX-2) supplemented with 1x Halt™ Protease and Phosphatase Inhibitor Cocktail (Thermo 78440) was added to the tissue, which was then homogenized using a Precellys® 24 Dual homogenizer (Bertin Instruments) for two cycles at 6,500 rpm for 20 s each at 4°C. Lysates were cleared by centrifugation for 10 min at 20,000 x g in an Eppendorf 5417C centrifuge at 4°C. Supernatant samples were transferred to clean, chilled Eppendorf tubes and further clarified by centrifugation for an additional 20 min as described above. Supernatant samples were then transferred to a 96-well plastic storage block, flash frozen in liquid nitrogen and stored at -80°C.

The following day, samples were thawed at room temperature and placed on ice. Protein concentrations in lysates were measured using a bicinchoninic acid (BCA) protein assay kit (Thermo 23225) according to the manufacturer's instructions. Lysates were normalized to a protein concentration of approximately 8 mg/mL using MSD Tris lysis buffer with protease and phosphatase inhibitors (see above) and aliquoted into plastic microtubes.

The concentration of TGF-e1 in normalized tumor lysates was measured according to the manufacturer's instructions using the Human TGF-e1 Kit (MSD, K151IUC-2) by electrochemiluminescence assay. Recombinant mouse TGF-e1 (R&D Systems, Cat. #7666MB-005) serially diluted in MSD lysis buffer was used as a calibration standard. Samples were loaded onto the plate in duplicate. The electrochemiluminescence signal was measured using a MESO SECTOR S 600 plate reader (MSD), and the concentration of TGF-e1 in the samples was quantified using MSD Discovery Workbench v4.0 software based on the calibration curve.

The mean concentration for duplicate samples was calculated using the software. Concentration values determined by the software to be below the curve fitting range or below the detection range were replaced by zero values. To calculate the TGF-e1 concentration per mg total protein, the concentration measured in the assay (pg/mL) was divided by the protein concentration (mg/mL) in the samples.

The results showed that intratumoral TGF-e1 levels in isotype control-injected mice had a median of 21.4 pg/mg total protein, while the corresponding levels in Ab1-injected mice were undetectable (Fig. 8).

To demonstrate the significance of the above results in humans, the inventors tested ten human colorectal tumor samples and ten human melanoma tumor samples for their intratumoral TGF-β 1 levels, as described above, using the method described above. In human CRC samples, TGF-β 1 levels ranged from about 7 to 25 pg/mg. In human melanoma samples, TGF-β 1 levels ranged from about 1 pg/mg to 43 pg/mg. These data further support the use of anti-TGF-β 1 antibody therapeutics, such as Ab1, in the treatment of tumors alone or in combination with other immune checkpoint inhibitors, such as anti-PD-1 antibodies.

Example 7. Pharmacokinetic studies of Ab1.

This example describes studies that characterized the pharmacokinetic (PK) profile of Ab1 and compared it to that of fresolimumab. In one study, five groups of cannulated Sprague-Dawley rats were given a single 5 mg/kg dose of Ab1 or fresolimumab intravenously. Each group consisted of five females and five males. Blood was collected from the rats at 0.25, 6, 24, 48, 72, 144, 192, and 240 hours post-dose. Serum concentrations of Ab1 and fresolimumab were determined by ELISA. Comparability was established if the 90% confidence intervals for the AUC (test to reference) ratios were between 80% and 125%.

The change in serum antibody concentrations over time in the five groups of rats is shown in Figure 9A. The PK parameters for groups 2, 4, and 5 (see Figure 9A legend) are shown in Table 7 below. This study showed that Ab1 had a linear PK profile with a significantly longer half-life (mean T _1/2 7.1 days vs. 4.3 days) and a slower clearance rate (CL 0.30 mL/h/kg vs. 0.51 mL/h/kg) than fresolimumab. The data showed that Ab1 had 1.7-fold higher exposure in rats than fresolimumab.

- 25 049204

Table 7

Comparison of PK parameters of frezolimumab and Ab1

RK parameters Fresolimumab (batch 2) АЫ (Party 1) АЫ (Party 2) tl/2 (h) 103.21 ± 12.98 181.83 ± 80.92* 158.38 ± 66.73 Stakh (mcg/ml) 114.56 ± 16.42 116.60 ± 16.78 109.72 ± 18.49 Vz (ml/kg) 75.22 ± 15.21 78.35 ± 26.33 63.47 ± 10.65 CL (ml/h/kg) 0.51 ± 0.10 0.31 ± 0.04* 0.30 ± 0.05* AUCiast (h*mcg/ml) 8046 ± 1210 10007 ± 1180* 11223 ± 1035* AUCo-inf (h*mcg/ml) 10107 ± 1613 16322 ± 2523* 17553 ± 4130*

* Mean values ±SD in the group are statistically significantly different from the frezolimumab (B2) group (group 2 in Fig. 9A)

An additional PK study of Ab1 (Study 2) was conducted in groups of cynomolgus macaques. Each group contained five females and five males that were given a single dose of Ab1 by intravenous infusion at 1 mg/kg (Fig. 9B) or 10 mg/kg (Fig. 9D) or five weekly doses of Ab1 at 1 mg/kg (Fig. 9C) or 10 mg/kg (Fig. 9E) per dose. The changes in serum Ab1 concentrations over time in the monkeys are shown in Figs. 9B–E. The changes in serum concentrations over time of fresolimumab, which was given to monkeys as single or repeated doses Q2W (every two weeks) in the previous studies, are also shown in the figures for comparison. These data showed that Ab1 also had a linear PK profile in monkeys and demonstrated higher exposure after both single and repeated dosing than fresolimumab at both 1 mg/kg and 10 mg/kg per dose. Following a single 10 mg/kg dose, Ab1 had a half-life of 13 days, whereas fresolimumab had a half-life of 4.5 days; Ab1 had a CL of approximately 0.40 mL/h/kg, whereas fresolimumab had a CL of 0.66 mL/h/kg. Similar to the rat study, the monkey study also showed that Ab1 had approximately 1.7-fold higher exposure than fresolimumab.

The above studies demonstrate that Ab1 has a statistically significantly longer half-life, longer clearance time, and higher in vivo biological exposure than fresolimumab.

Furthermore, a study in Balb/C mice bearing Ab12 tumors showed that Ab1 had similar PK profiles whether administered intravenously or intraperitoneally.

Using allometric scaling in a two-compartment model, we predicted the following PK parameters for a 70 kg human based on monkey data (Table 8).

Table 8 Allometric modeling of PK parameters

RK parameters Ay in monkeys AЫ in people T1/2 (day) 13, 1 20, 9 CL (ml/h/kg) 0, 392 5, 7 VI (central chamber) (l) 0, 104 2.43 Q (ml/h) 3, 18 46 V2 (peripheral chamber) (l) 0,0693 1, 62

The predicted PK parameters for Ab1 in humans are also more favorable than those for fresolimumab. For example, fresolimumab, with a CL of 12.3 mL/h/kg in humans, has a higher clearance rate than Ab1.

- 26 049204

Example 8. Toxicological studies Ab1.

Toxicology studies with Ab1 were performed in rats and cynomolgus monkeys. Safety pharmacologic endpoints were assessed in a GLP (Good Laboratory Practice) repeat dose study administered weekly for 5 weeks. No Ab1-related histopathology was observed at injection sites at doses up to 10 mg/kg/dose (2 mg/mL concentration) in monkeys and up to 30 mg/kg/dose (6 mg/mL concentration) in rats. In this study, no Ab1-related effects were observed in neurologic examinations, body temperature, respiratory rate, blood pressure, or ECG parameters at any dose level tested.

The NOAEL (no observed adverse effect level) in rats was found to be 3 mg/kg/dose with weekly repeated dosing for 5 weeks, and the STD10 (toxic dose causing death or irreversible severe toxicity in 10% of animals) in rats was found to be 3 to 10 mg/kg/dose. Toxicity included cardiac valve cell proliferation characterized by multiple thickened nodules; and abnormal lung findings such as mixed cellular alveolar exudates, mixed cellular perivascular infiltrates, muscular arterial hypertrophy, hemorrhage, and/or increased lung weight.

The NOAEL and HNSTD (i.e., the highest no-fail toxic dose above which lethality, life-threatening toxicity, or irreversible toxicity occurs) doses in monkeys with weekly repeat dosing for 5 weeks were found to be 10 mg/kg/dose (compared to frezolimumab, which has been shown to have a NOAEL in monkeys of 1 mg/kg when administered biweekly for 7 or 13 doses or Q3D for 4 weeks). See also the data shown in Table 9 below.

Table 9

Summary of toxicology studies in rats and monkeys

Toxicological parameters Rats Monkeys LD (lethal dose) 50 mg/kg/dose ¹ > 10 mg/kg/dose HNLD (highest non-lethal dose) 30 mg/kg/dose 10 mg/kg/dose STD10 3-10 mg/kg/dose NA HNSTD NA 10 mg/kg/dose NOAEL < 3 mg/kg/dose 10 mg/kg/dose Major target organs (histopathology) Heart, lungs, bones, teeth Not identified E based on exploratory research NA: no data

Based on the above toxicological data, it is expected that Ab1 can be safely administered to human patients at a dose level of approximately 0.05 mg/kg to 0.5 mg/kg once weekly or less frequently, such as once every two weeks.

Example 9. Efficacy of monotherapy with an antibody to TGF-β in vivo.

In this study, the present inventors examined the effect of 1D11, a mouse IgG1 antibody to bovine TGF-β that cross-reacts with human and mouse TGF-βΓ 2 and 3, in a metastatic syngeneic tumor model. In this model, mouse melanoma B16-F10 cells were injected into the footpad of C57BL/6 IV mice and formed metastases in the draining lymph node of the mice. While treatment with the control antibody 13C4 had no effect, treatment with 50 mg/kg 1D11 three times a week, starting one day after tumor inoculation, completely eliminated the metastases.

To investigate the role of the immune response, mice deficient in the β2 microglobulin gene and therefore lacking a CD8+ cytotoxic T cell immune response were implanted with B16-F10 in the footpad and treated as described previously. In contrast to the results observed in immunocompetent mice, 1D11 had no effect on the number of draining lymph node metastases in these mice. These results suggest that the mechanism of action of TGF-β inhibition is dependent on adaptive cellular immunity.

Example 10. TGF-β signatures in different forms of cancer.

Previous studies have shown that melanoma patients who failed to respond to anti-PD-1 antibody therapy have the IPRES transcriptional signature (Hugo

- 27 049204 et al., Cell (2016) 165:35–44). To investigate the mechanism of intrinsic resistance to anti-PD-1 monotherapy, we examined the transcriptional signatures of non-responders versus responders. We found that comparison of these profiles using gene array abundance analyses against a database of over 1 million profiles revealed a strong correlation between anti-PD-1 response and activation of TGF-β signaling in tumors. These data suggest that baseline TGF-β is associated with intrinsic resistance to anti-PD-1 monotherapy in melanoma.

Furthermore, we found that not only was there a correlation between the response to anti-PD-1 antibody and the activation of TGF-β signaling, but that the correlation was strong (R=0.59, t-test p-value <9E-4). We therefore arrived at Input 1: TGF-β-mediated immunosuppression in melanoma (e.g., metastatic melanoma) may contribute to innate resistance. Furthermore, we found that TGF-β-induced gene expression changes were suppressed by treatment with 1D11, confirming the specificity of the TGF-β activation signature. These results support the benefit of using anti-TGF-β antibody and anti-PD-1 antibody therapeutics in combination to treat patients with cancer who do not respond to anti-PD-1 antibody monotherapy.

Analysis of this correlation in tumor types other than melanoma revealed that mesenchymal tumors (e.g., CRC, HCC, head and neck squamous cell carcinoma, and ovarian cancer) also had both increased TGF-β activation and increased predicted resistance to anti-PD-1. This finding was consistent with a role for TGF-β signaling in EMT. We therefore generated Insight 2: mesenchymal tumors, especially those infiltrated by immune cells, may benefit from combination therapy with anti-TGF-β and anti-PD-1. Machine learning was used to identify a smaller subset of over 30 EMT marker genes, such as ACTA2, VIM, MGP, ZEB2, and ZWINT, that could be used to select mesenchymal tumors. For example, ACTA2 and VIM have been found to be compatible across tumor types. Accordingly, the transcriptional signature of TGF-β activation and the genes in this expression signature may serve as useful biomarkers for selecting patients with baseline cancer for combination therapy with an anti-TGF-β antibody and an anti-PD-1 antibody.

To investigate biomarkers in the tumor microenvironment, the immune environment of patient-derived tumors in CRC and melanoma was assessed using the MultiOmyx multiplexed IHC assay. Multiplexing was performed with 12 biomarkers (collectively describing 22 immune cell types) on a single FFPE section from each tumor sample. Studies included a range of inflammation to assess how well these assays assess each tumor type and correlate with potential treatment effects. Statistical methods were developed to assess differences at the cell population level, including consistency analysis of replicates, volcano plots for ANOVAs, and correlation matrices. The MultiOmyx assay demonstrated excellent technical reproducibility and precision, favorable dynamic range, differences in inflammation status within individual immune cells and regions of interest, and included both positive and negative correlations between cell populations.

Example 11. Changes in TGF-β1, MIP-2, and KC/GRO levels in MC38 tumors after treatment with Ab1 with or without anti-PD1 antibody.

To demonstrate TGF-β neutralization, the ability of Ab1 (with or without anti-PD1 antibody) to influence cytokine expression in tumors was assessed.

MC38 tumor-bearing mice were treated with a single dose of PBS or anti-PD-1 antibody alone (5 mg/kg) or increasing doses of Ab1 (10, 25, or 50 mg/kg, ip) in combination with anti-PD-1 antibody (5 mg/kg) at tumor volumes ranging from 61 to 110 mm ³ . Tumors were harvested at 1 h, 6 h, 10 h, 24 h, 72 h, and 168 h after treatment, snap frozen in 2 mL plastic tubes with 2.8 mm ceramic beads (Precellys KT3961-1007.2), and stored at -80°C. Tumor lysates were thawed at room temperature. One ml of cold Meso Scale Diagnostics (MSD) Tris Lysis Buffer (R60TX-2) supplemented with 1x Halt™ Protease and Phosphatase Inhibitor Cocktail (Thermo 78440) was added to the tissue, which was then homogenized using a Precellys®24 Dual Homogenizer (Bertin Instruments) for 2 cycles at 6,500 rpm for 20 sec each at 4°C. Lysates were cleared by centrifugation for 10 min at 20,000 x g in an Eppendorf 5417C centrifuge at 4°C. Supernatant samples were transferred to fresh chilled Eppendorf tubes and further cleared by centrifugation for an additional 30 min as described above. Supernatant samples were transferred to a 96-well plastic storage block and placed on ice. Protein concentration in lysates was measured using a bicinchoninic acid (BCA) protein assay kit (Thermo 23225) in

- 28 049204 according to the manufacturer's instructions. Lysates were normalized to a protein concentration of approximately 5 mg/mL using MSD Tris lysis buffer with protease and phosphatase inhibitors (see above), aliquoted into plastic microtubes, flash frozen in liquid nitrogen, and stored at -80°C.

Active TGFe-1 concentrations in tumor lysates were measured using the Human TGFe1 Kit (MSD, K151IUC-2) by electrochemiluminescence assay. Recombinant mouse TGFe-1 (R&D Systems, 7666-MB-005) serially diluted in MSD lysis buffer was used as a calibration standard. Normalized tumor lysates prepared as described above were thawed and assayed according to the manufacturer's instructions. Acid treatment of samples was not performed to quantify only the active form of TGFe-1 present in tumors and not total TGFe-1, which includes TGFe-1 complexed with latency-associated peptide. Samples were loaded per plate in duplicate. The electrochemiluminescent signal was measured using a MESO SECTOR S 600 microplate reader (MSD), and the TGF-β 1 concentration in the samples was quantified using MSD Discovery Workbench version 4.0 software based on a calibration curve.

Compared with animals treated with PBS or anti-PD-1 antibody alone, animals treated with Ab1 at all dose levels (10, 25, or 50 mg/kg) plus anti-PD-1 antibody (5 mg/kg) were shown to have reduced levels of active TGF-β1 in tumors, demonstrating binding of Ab1 to its target in vivo (Fig. 10A). Reduced levels of active TGF-β1 were observed within one hour and were maintained for at least 168 hours.

MIP-2 (CXCL2) and KC/GRO (CXCL1) are chemotactic chemokines for granulocytes, including neutrophils. MIP-2 and KC/GRO levels were also assessed in the same samples. Following treatment with Ab1 plus anti-PD-1, intratumoral MIP-2 levels were shown to be increased at least 4-fold in animals treated with Ab1 plus anti-PD-1 compared to animals treated with PBS or anti-PD-1 alone; and the increase in MIP-2 levels was shown to be sustained for at least 168 hours (Fig. 10B). Similarly, KC/GRO levels were shown to be increased, albeit at later time points of 72 and 168 hours compared to MIP-2 (Fig. 10C). Thus, the combination of Ab1 and anti-PD-1 mAb induced a decrease in active TGF-β levels before an increase in MIP-2 and KC/GRO levels. These results demonstrate that Ab1 can decrease TGF-β levels and inhibit it in the tumor microenvironment. Furthermore, the observed increase in MIP-2 and KC/GRO levels indicates that they are cytokines affected by TGFβ neutralization and thus may serve as potential biomarkers in patients treated with Ab1.

Example 12. Restoration of NK cell cluster formation upon treatment with Ab1.

TGF-β is known to affect the immune system by inhibiting the activity of various immune cell types. TGF-β has been reported to inhibit natural killer (NK) cell activity and NK cell-mediated ADCC (Trotta et al., Journal of immunology (2008) 181:3784–3792). Recently, NK cells have been reported to form dense aggregates as a mechanism to enhance their activity and activation via localization of IL-2 to these densely packed aggregates (Kim et al., Scientific Reports (2017) 7:40623). Purified human NK cells cultured in vitro in the presence of IL2 have been shown to form these densely packed aggregates.

In the present study, we evaluated the effects of TGF-β in the absence or presence of Ab1 on NK cell aggregate formation. NK cells were de novo isolated from healthy donor blood by negative selection with RosetteSep NK cell reagent according to the manufacturer's protocols (Stem Cell Technologies). NK cells were cultured at 1.2 x 105 cells/well in MyeloCult medium (Stem Cell Technologies) supplemented with IL-2 (100 IU/mL) in round bottom assay plates (Costar). TGF-β1 was added to a final concentration of 0.1, 1, or 10 ng/mL in the presence of irrelevant IgG4 or Ab1 at 100 μg/mL, as indicated. Cells were cultured for 72 hours and NK cell aggregate formation was visualized by image capture on a Nikon microscope.

It was shown that the formation of NK cell aggregates was inhibited by the addition of increasing doses of TGF-β 1. It was shown that NK cell aggregates were formed when Ab1, but not the control antibody IgG4, was added to NK cell cultures. This result demonstrates that neutralization of TGF-β affects NK cell activation, leading to increased NK cell activity and proliferation, promoting the antitumor immune response.

Example 13. Reversal of TGF-β-mediated suppression of IFN-γ production in proliferating CD8 ⁺ T cells by treatment with Ab1.

In addition to the innate immune system, TGF-β has been reported to inhibit CD8 ⁺ T cell activity (Flavell et al., Nature Reviews Immunology (2010) 10:554–567). To investigate the role of TGF-β

- 29 049204 β and Ab1 in CD8+ T cell activity, a mixed lymphocyte reaction (MLR) assay system was established in which purified human CD3 ⁺ cells were mixed with BLCL cells. CD8 ⁺ cell proliferation and IFN-γ production were first assessed in the presence of TGF-β. Specifically, CD3 ⁺ cells were isolated using the EasySep T Cell Enrichment Kit (StemCell Technologies) from PBMCs from normal healthy donors fractionated after Ficoll gradient isolation. CD3 ⁺ cells were then labeled with CellTrace violet according to the manufacturer's protocol (ThermoFisher). MLR analysis was performed by mixing labeled CD3 ⁺ cells (2 X ¹⁰⁵ cells) with irradiated BLCL cells (Astarte Bio) (2 X ¹⁰⁴ cells; 2 min) in RPMI supplemented with 10% FBS. As indicated, TGF-e1, IgG4 control antibody, and/or Ab1 were added to the cultures and the cultures were incubated at 37°C with 5% _CO2 for 4 days. Cells were then stimulated for 4 hours in the presence of PMA cell stimulation cocktail (eBioscience) and protein transport inhibitor cocktail (eBioscience). Viable cells were detected by Zombie NIR viability staining (BioLegend) on ice and washed with FACS buffer. Cells were fixed in True-Nuclear buffer (BioLegend), washed, pelleted, and resuspended in FACS buffer. Cells were prepared for flow cytometry by staining with BV650-conjugated anti-huCD4, PERCP/Cy5.5-conjugated anti-huCD8, FITC-conjugated anti-huCD3, and PE-conjugated anti-huIFNY (BioLegend). Flow cytometry was performed on a BD Canto and results were analyzed using FlowJo software with gating on live cells, singlets, and CD3 ⁺ cells. The percentage of IFNy ⁺ CD8 ⁺ T cells was quantified by gating on CD8 ⁺ cells that were proliferating based on decreased CellTrace violet staining intensity and were positive for IFN-γ. FMO was used as a control for all antibody staining.

Inclusion of TGF-β in the MLR assay was shown to reduce the percentage of CD8 ⁺ T cells that were positive for IFN-γ by approximately 4-fold (Fig. 11A). Inclusion of Ab1 or control Ab had no effect on the generation of these proliferating IFNy ⁺ CD8 ⁺ cells in the absence of TGF-β (Fig. 11B). However, inclusion of Ab1, but not control Ab, restored IFNy ⁺ CD8 ⁺ cell proliferation in a dose-dependent manner. These results demonstrate that neutralization of TGF-β is able to influence the adaptive immune system by blocking the immunosuppressive effects of TGF-β on the proliferation of effector CD8 ⁺ cells that express IFN-γ. These IFNy ⁺ CD8 ⁺ T cells are suggested to play an important role in antitumor immunity (Ikeda et al., Cytokine Growth Factor Rev (2002) 13:95-109).

Example 14. Responses to TGF-β antibody therapy in syngeneic mouse models.

In this study, we investigated which syngeneic mouse models could be used to predict response to treatment with anti-TGF-β Ab1 and anti-PD-1. To stratify the mouse models, we assessed CD8 ⁺ T cell infiltration into mouse tumors and activation of the TGF-β signaling pathway. CD8 ⁺ T cell infiltration was analyzed based on the CD8 ⁺ T cell signature as determined by RNA-sequencing. Seventeen different syngeneic mouse models with tumor cells arising in multiple indications (as shown below the x-axis in Figs. 12A and 12B) were transcriptionally profiled using whole-transcriptome RNA-sequencing. This set of syngeneic models was generated in the genetic background of the widely used C57/BL6 line using 5-7 biological replicates per model. Following Illumina 2000 sequencing, gene expression profiles were obtained in transcripts per million (TPM) using standard processing of raw sequence reads using the STAR alignment algorithm and the Cufflinks transcript abundance estimation algorithm. The resulting multi-sample data matrix was finally subjected to quantile normalization.

Figure 12A shows the relative abundance of CD8 ⁺ T cells (log2-transformed) within the array. Relative abundance of CD8 ⁺ T cells was estimated using the unique marker gene CD8B, which has been shown to be a highly specific indicator of the presence of CD8 T cells (Becht et al., Curr Opin Immunol (2016) 39:7–13 and Becht et al., Genome Biol (2016) 17:218). Each box plot summarizes the range of values across biological replicates. The MC38 model showed approximately 2-fold greater CD8 ⁺ T cell infiltration than the EMT6 model (left and right boxes, respectively). The A20 and EL4 lymphoma models showed the highest and lowest overall levels of CD8 ⁺ T cell infiltration, respectively, with negligible numbers of CD8 ⁺ T cells in EL4.

The mouse cell lines MC38, MC38.ova, CT26 and L1210 exhibited the highest levels of CD8 signature gene expression. In addition, the breast cancer cell line EMT-6 was shown to have T cell infiltration close to baseline levels, consistent with recent reports that EMT6 tumors have an immune exclusion phenotype (S. Mariathasan et al., 2017, ESMO Immuno-Oncology Congress, Geneva, Switzerland).

- 30 049204

Figure 12B shows the activation of the TGF-β signaling pathway within the array. The 170-gene transcriptional signature of TGFe signaling activation, determined by in vitro stimulation of MCF7 cells with TGF-β and validated by comparison with several other TGF-β signatures, was used to assign a pathway activation score to each profile in the array. Scores were calculated using regulated group of genes abundance analysis (rGSEA, Theilhaber et al. 2014) and expressed as the log2 overrepresentation of signature genes relative to the genetic environment. While the MC38 models showed moderate activation, the EMT6 model showed very high activation of the TGF-β signaling pathway (left and right boxes, respectively).

Example 15. Effects of the combination of Ab1 and anti-PD-1 antibody in a mouse model of breast cancer

In this study, we investigated the therapeutic effects of Ab1 with or without anti-PD-1 antibody. Exponentially growing EMT-6 breast cancer cells (CRL2755, ATCC) were cultured in RPMI-1640 supplemented with 10% FBS in a humidified incubator with 5% _CO2 and then implanted subcutaneously (0.5 X ¹⁰⁶ cells/mouse) into the flank of female BALB/c mice (Shanghai Lingchang Bio-Technology Co. Ltd, Shanghai, China). After the tumors reached an average size of 68-116 mm, ³ mice were pooled and randomly divided into control and treatment groups (10 mice per group). Tumor-bearing mice were then treated by intraperitoneal administration of PBS, Ab1 (10 and 25 mg/kg) three times a week for each animal for a total of 6 doses. Tumors were measured with digital calipers twice weekly and tumor volumes were calculated ( ^mm3 = L x W x H) and plotted using GraphPad Prism. Mice were euthanized with CO2 at the end of the study if tumors grew to >3000 ^mm3 or if tumors demonstrated ulceration of >20% of the tumor surface.

As monotherapy, Ab1 at 10 or 25 mg/kg Q3D and murine a-PD-1 antibody at 5 mg/kg demonstrated partial activity with complete regression in 1/10, 2/10, and 2/10 mice bearing EMT-6 tumors, respectively. The combination of Ab1 at 10 or 25 mg/kg Q3D and murine a-PD-1 antibody at 5 mg/kg Q3D was therapeutically active. At day 31 post-implantation, when comparing changes in tumor volume from baseline, the combination of Ab1 at all doses tested and murine a-PD-1 antibody at 5 mg/kg Q3D was superior to either agent as monotherapy with complete regression in 7/10 and 4/10 mice for 10 and 25 mg/kg Ab1, respectively. Table 10 provides a summary of the results.

Table 10

Effects of LY/anti-mPD-1 autobody combination in the EMT-6 mouse model

Group Processing Total number of mice Number of complete response cases (complete response rate) 1. PBS 10 0 (0%) 2 . 5 mg/kg Mab x-anti-mPD-1 10 2 (20%) 3. 10 mg/kg AY 10 1 (10%) 4 . 2 5 mg/kg AY 10 2 (20%) 5. 10 mg/kg AA + 5 mg/kg Mab x-anti-mPD-1 10 7 (70%) 6. 2 5 mg/kg AA + 5 mg/kg Mab x-anti-mPD-1 10 4 (40%)

Unless otherwise defined herein, scientific and technical terms used in connection with the present invention shall have the meanings that are commonly understood by those of ordinary skill in the art. Illustrative methods and materials are described below, although methods and materials similar or equivalent to those described herein may also be used in the practice or testing of the present invention. All publications and other literature references mentioned herein are incorporated by reference in their entireties. In the event of conflict, the present description, including definitions, will control. Although a number of documents are cited in this document, such citation is not an admission that any of these documents forms part of the common knowledge in the art. Furthermore, unless the context otherwise requires, singular terms include the plural and plural terms include the singular. In general,

- 31 049204 the nomenclature used in connection with cell and tissue culture, molecular biology, immunology, microbiology, genetics, analytical chemistry, synthetic organic chemistry, medicinal and pharmaceutical chemistry, and protein and nucleic acid chemistry and hybridization, as well as the corresponding techniques described herein, are well known and widely used in the art. Enzymatic reactions and purification techniques are performed in accordance with the manufacturer's descriptions, as commonly performed in the art, or as described herein. Throughout the present description and embodiments, the words have and contain, or variations thereof such as has, having, comprises, or containing, should be understood to imply the inclusion of the stated integer or group of integers, but not the exclusion of any other integer or group of integers.

The sequences described in this document are listed below.

List of sequences.

SEQ ID NO:1 (heavy chain Abi)

QVQLVQSGAE VKKPGSSVKV SCKASGYTFS SNVISWVRQA PGQGLEWMGG VIPIVDIANY AQRFKGRVTI TADESTSTTY MELSSLRSED TAVYYCASTL GLVLDAMDYW GQGTLVTVSS ASTKGPSVFP LAPCSRSTSE STAALGCLVK DYFPEPVTVS WNSGALTSGV HTFPAVLQSS GLYSLSSWT VPSSSLGTKT YTCNVDHKPS NTKVDKRVES KYGPPCPPCP APEFGLGPSV FLFPPKPKDT LMISRTPEVT CVWDVSQED PEVQFNWYVD GVEVHNAKTK PREEQFNSTY RWSVLTVLH QDWLNGKEYK CKVSNKGLPS SIEKTISKAK GQPREPQVYT LPPSQEEMTK NQVSLTCLVK GFYPSDIAVE WESNGQPENN YKTTPPVLDS DGSFFLYSRL

TVDKSRWQEG

NVFSCSVMHE ALHNHYTQKS LSLSLGK SEQ ID NO:2 ETVLTQSPGT (Light chain LSLSPGERAT Idol) LSCRASQSLG SSYLAWYQQK PGQAPRLLIY GASSRAPGIP DRFSGSGSGT DFTLTISRLE PEDFAVYYCQ QYADSPITFG QGTRLEIKRT VAAPSVFIFP PSDEQLKSGT ASWCLLNNF YPREAKVQWK VDNALQSGNS QESVTEQDSK

DSTYSLSSTL

TLSKADYEKH

KVYACEVTHQ GLSSPVTKSF NRGEC

SEQ ID NO:3 (frezolimumab heavy chain including leader sequence residues 1-19)

MGWSCIILFL VATATGVHSQ VQLVQSGAEV

KKPGSSVKVS

CKASGYTFSS

NVISWVRQAP

GQGLEWMGGV

IPIVDIANYA

QRFKGRVTIT

ADESTSTTYM

ELSSLRSEDT AVYYCASTLG

LVLDAMDYWG QGTLVTVSSA STKGPSVFPL

APCSRSTSES

TAALGCLVKD

YFPEPVTVSW

NSGALTSGVH

TFPAVLQSSG

LYSLSSWTV

PSSSLGTKTY

TCNVDHKPSN TKVDKRVESK

YGPPCPSCPA PEFLGGPSVF LFPPKPKDTL

MISRTPEVTC

VWDVSQEDP

EVQFNWYVDG

VEVHNAKTKP

REEQFNSTYR

WSVLTVLHQ

DWLNGKEYKC

KVSNKGLPSS IEKTISKAKG

QPREPQVYTL PPSQEEMTKN QVSLTCLVKG

FYPSDIAVEW

ESNGQPENNY

KTTPPVLDSD GSFFLYSRLT VDKSRWQEGN VFSCSVMHEA LHNHYTQKSL SLSLGK

SEQ ID NO:4 (frezolimumab light chain, including leader sequence - residues 1-19)

- 32 049204

MGWSCIILFL VATATGVHSE TVLTQSPGTL

SLSPGERATL

SCRASQSLGS

SYLAWYQQKP

GQAPRLLIYG

ASSRAPGIPD

RFSGSGSGTD

FTLTISRLEP

EDFAVYYCQQ YADSPITFGQ

GTRLEIKRTV AAPSVFIFPP SDEQLKSGTA

SWCLLNNFY

PREAKVQWKV

DNALQSGNSQ

ESVTEQDSKD

STYSLSSLT

LSKADYEKHK

VYACEVTHQG

LSSPVTKSFN RGEC

SEQ ID NO:5 (heavy chain Mab to PD-1)

EVQLLESGGV LVQPGGSLRL SCAASGFTFS

NGFMTVRQA

PGKGLEWVSG

ISGGGRDTYF

ADSVKGRFTI

SRDNSKNTLY

LQMNSLKGED

TAVYYCVKWG

NIYFDYWGQG

TLVTVSSAST

KGPsVFPLAP

CSRSTSESTA

ALGCLVKDYF

PEPVTVSWNS

GALTSGVHTF

PAVLQSSGLY

SLSSWTVPS

SSLGTKTYTC

NVDHKPSNTK

VDKRVESKYG

PPCPPCPAPE

FLGGPSVFLF

PPKPKDTLMI

SRTPEVTCW

VDVSQEDPEV

QFNWYVDGVE

VHNAKTKPRE

EQFNSTYRW

SVLTVLHQDW

LNGKEYKCKV

SNKGLPSSIE

KTISKAKGQP

REPQVYTLPP

SQEEMTKNQV

SLTCLVKGFY

PSDIAVES

NGQPENNYKT

TPVLDSDGS

FFLYSRLTVD

KSRWQEGNVF SCSVMHEALH NHYTQKSLSL SLGK

SEQ ID NO:6 (light chain of Mab to PD-1)

DIQMTQSPSS LSASVGDSIT ITCRASLSIN

TFLNWYQQKP

GKAPNLLIYA

ASSLHGGVPS

RFSGSGSGTD

FTLTIRTLQP

EDFATYYCQQ

SSNTPFTFGP

GTWDFRRTV

AAPSVFIFPP

SDEQLKSGTA

SWCLLNNFY

PREAKVQWKV

DNALQSGNSQ

ESVTEQDSKD

STYSLSSLT

LSKADYEKHK

VYACEVTHQG

LSSPVTKSFN

RGE

SEQ ID NO:7 (heavy chain Mab x-anti-mPD-1)

EVQLQESGPG LVKPSQSLSL TCSVTGYSIT

SSYRWNWIRK

FPGNRLEWMG

YINSAGISNY

NPSLKRRISI

TRDTSKNQFF

LQVNSVTTED

AATYYCARSD

NMGTTPFTYW

GQGTLVTVSS

AKTTPPSVYP

LAPGSAAQTN

SMVTLGCLVK

GYFPEPVTVT

WNSGSLSSGV

HTFPAVLQSD

LYTLSSVTV

P.S.T.W.P.S.T.W.S.T.V.

TCNVAHPASS

TKVDKKIVPR

DCGCKPCICT

VPEVSSVFIF

PPKPKDVLTI

TLTPKVTCW

VDISKDDPEV

QFSWFVDDVE

VHTAQTQPRE

EQFNSTFRSV

SELPIMHQDW

LNGKEFKCRV

NSAAFPAPIE

KTISKTKGRP

KAPQVYTIPP

PKEQMAKDKV

SLTCMITDFF

PEDITVEWQW

NGQPAENYKN

TQPIMDTDGS

YFVYSKLNVQ

KSNWEAGNTF TCSVLHEGLH NHHTEKSLSH SPG

SEQ ID NO:8 (light chain Mab x-anti-mPD-1)

DIVMTQGTLP NPVPSGESVS ITCRSSKSLL

YSDGKTYLNW

YLQRPGQSPQ

LLIYWMSTRA

SGVSDRFSGS

GSGTDFTLKI

SGVEAEDVGI

YYCQQGLEFP

TFGGGTKLEL

KRADAAPTVS

IFPPSTEQLA

TGGASWCLM

NNFYPRDISV

KWKIDGTERR

DGVLDSVTDQ

DSKDSTYSMS

STLSLTKADY

ESHNLYTCEV

VHKTSSSSPW

KSFNRNEC

SEQ ID NO:9 (heavy chain 1D11)

HVQLQQSGPE LVRPGASVKL SCKASGYIFI

TYWMNWVKQR

PGQGLEWIGQ

-

Claims (6)

IFPASGSTNY NEMFEGKATL TVDTSSSTAY MQLSSLTSED SAVYYCARGD GNYALDAMDY WGQGTSVTVS SAKTTPPSVY PLAPGSAAQT NSMVTLGCLV KGYFPEPVTV TWNSGSLSSG VHTFPAVLQS DLYTLSSSSVTVPSSTWPSQT VTCNVAHPAS STKVDKKIVP RDCGCKPCIC TVPEVSSVFI FPPKPKDVLT ITLTPKVTCV WDISKDDPE VQFSWFVDDV EVHTAQTKPR EEQFNSTFRS VSELPIMHQD WLNGKEFKCR VNSAAFPAPI EKTISKTKGR PKAPQVYTIP PPKEQMAKDK VSLTCMITDF FPEDITVEWQ WNGQPAENYK NTQPIMDTDG SYFVYSKLNV QKSNWEAGNT FTCSVLHEGL HNHHTEKSLS HSPGK SEQ ID NO:10 (light chain 1D11) NIVLTQSPAS LAVSLGQRAT ISCRASESVD SYGNSFMHWY QQKSGQPPKL LIYLASNLES GVPARFSGSG SRTDFTLTID PVEADDAATY YCQQNNEDPL TFGAGTKLEL KRADAAPTVS IFPPSSEQLT SGGASWCFL NNFYPKDINV KWKIDGSERQ NGVLNSWTDQ DSKDSTYSMS STLTLTKDEY ERHNSYTCEA THKTSTSPIV KSFNRNEC CLAUSE OF INVENTION 1. An isolated nucleic acid molecule encoding a monoclonal antibody that specifically binds to human TGF-βΚ TGF^2 and TGF^3, wherein the antibody comprises complementarity determining regions (CDRs) 1-3 of the heavy chain comprising the sequences SNVIS, GVIPIVDIANYAQRFKG and TLGLVLDAMDY, respectively, and CDR1-3 of the light chain comprising the sequences RASQSLGSSYLA, GASSRAP and QQYADSPIT, respectively, and wherein the antibody comprises a human IgG4 constant region with a proline at position 228 (EU numbering). 2. Two isolated nucleic acid molecules encoding a monoclonal antibody that specifically binds to human TGF-βΚ TGF^2 and TGF^3, wherein the antibody comprises complementarity determining regions (CDRs) 1-3 of the heavy chain comprising the sequences SNVIS, GVIPIVDIANYAQRFKG and TLGLVLDAMDY, respectively, and CDR1-3 of the light chain comprising the sequences RASQSLGSSYLA, GASSRAP and QQYADSPIT, respectively, and wherein the antibody comprises a human IgG4 constant region with a proline at position 228 (EU numbering). 3. The isolated nucleic acid molecule(s) of claim 1 or 2, wherein the antibody comprises an amino acid sequence of a heavy chain variable domain (VH) corresponding to residues 1-120 of SEQ ID NO: 1 and an amino acid sequence of a light chain variable domain (VL) corresponding to residues 1-108 of SEQ ID NO: 2. 4. The isolated nucleic acid molecule(s) according to claim 1 or 2, wherein the antibody comprises the heavy chain amino acid sequence of SEQ ID NO: 1 and the light chain amino acid sequence of SEQ ID NO: 2. 5. An isolated nucleic acid molecule encoding the F(ab') ₂ fragment of a monoclonal antibody that specifically binds to human TGF-β1, TGF^2 and TGF^3, wherein the antibody comprises complementarity determining regions (CDRs) 1-3 of the heavy chain comprising the sequences SNVIS, GVIPIVDIANYAQRFKG and TLGLVLDAMDY, respectively, and CDR1-3 of the light chain comprising the sequences RASQSLGSSYLA, GASSRAP and QQYADSPIT, respectively, and wherein the antibody comprises a human IgG4 constant region with a proline at position 228 (EU numbering). 6. Two isolated nucleic acid molecules encoding the F(ab') ₂ fragment of a monoclonal antibody that specifically binds to human TGF-βΚ TGF^2 and TGF^3, -