EA043093B1 - METHOD FOR TREATING CANCER USING HUMAN ANTIBODIES TO LAG-3 - Google Patents

Description

State of the art

Lymphocyte activation gene-3, or LAG-3 (also known as CD223), is a member of the immunoglobulin supergene (immunoglobulin supergene) family and is structurally and genetically related to CD4. LAG-3 is not expressed on dormant peripheral blood lymphocytes, but is expressed on activated T cells and NK cells. LAG-3 is a membrane protein encoded by a gene located on the distal short arm of chromosome 12, near the CD4 gene, suggesting that the LAG-3 gene can be isolated by gene duplication (Triebel et al. (1990) J. Exp. Med 171: 1393-1405).

It has been shown that, similar to CD4, LAG-3 does not interact with MHC Class II molecules, but, unlike CD4, LAG-3 does not interact with the gp120 protein of the human immunodeficiency virus (Baixeras et al. (1992) J. Exp Med 176: 327-337). Studies using LAG-3 immunoglobulin fusion protein (sLAG-3Ig) demonstrate direct and specific binding of LAG-3 to MHC Class II molecules on the cell surface (Huard et al. (1996) Eur. J. Immunol. 26: 1180-1186).

In in vitro studies of antigen-specific T cell responses, the addition of anti-LAG-3 antibodies results in increased T cell proliferation, increased expression of activation antigens such as CD25, and increased concentrations of cytokines such as interferon-gamma and interleukin-4. supporting a role for LAG-/MHC class II interaction in downregulation of antigen-dependent stimulation of CD4+ T lymphocytes (Huard et al. (1994) Eur. J. Immunol. 24: 3216-3221). The intraplasmic region of LAG-3 has been shown to interact with a protein called LAP, which is proposed to be a signal transduction molecule involved in the downregulation of the CD3/TCR activation pathway (Iouzalen et al. (2001) Eur. J. Immunol. 31: 2885- 2891). In addition, CD4+CD25+ regulatory T cells ( _Treg ) have been shown to express LAG-3 upon activation, and antibodies to LAG-3 inhibit suppression by induced _Treg cells, both in vitro and in vivo, suggesting that LAG-3 contributes to the suppressor activity of T _reg cells (Huang, S. et al. (2004) Immunity 21: 503-513).

Moreover, LAG-3 has been shown to negatively regulate T cell homeostasis through regulatory T cells through both T cell-dependent and T cell-independent mechanisms (Workman, C. J. and Vignali, D. A. (2005 ) J Immunol 174: 688-695).

LAG-3 has also been shown to have an immunostimulatory effect under certain conditions. For example, LAG-3-transfected tumor cells transplanted into syngeneic mice exhibit a marked increase in reduction or complete regression compared with untransfected tumor cells, suggesting that LAG-3 expression on tumor cells stimulates an antitumor response by triggering , inclusion) of antigen-presenting cells by MHC class II molecules (Prigent et al. (1999) Eur. J. Immunol. 29: 3867-3876). In addition, soluble LAG-3 Ig fusion protein has been shown to stimulate both humoral and cellular immune responses when administered along with antigen to mice, indicating that soluble LAG-3Ig can act as a vaccine adjuvant (El Mir and Triebel (2000) ) J Immunol 164: 5583–5589). Soluble human LAG-3Ig has further been shown to enhance in vitro production of type I tumor-specific immunity (Casati et al. (2006) Cancer Res. 66: 4450-4460). A review of the functional activity of LAG-3 is given in Triebel (2003) Trends Immunol. 24: 619-622. Based on the above, it is clear that additional agents to modulate LAG3 activity are of interest.

The essence of the invention

The present invention includes monoclonal antibodies, in particular human monoclonal antibodies, that specifically bind LAG-3 and have the desired functional properties. These properties include high-affinity binding to human LAG-3, binding to human and monkey LAG-3 (e.g., cynomolgus and/or rhesus LAG-3) but not murine LAG3, and the ability to inhibit LAG-3 binding to major histocompatibility complex molecules ( MHC, MHC) class II and/or the ability to stimulate antigen-specific T-cell reactions (responses). Antibodies of the invention can be used, for example, to detect LAG-3 protein or to stimulate antigen-specific T-cell responses, such as in a tumor host or a viral host.

In one aspect, the invention provides an isolated human monoclonal antibody, or an antigen binding region thereof, characterized in that the antibody binds human LAG-3 and exhibits at least one of the following properties:

a) binds monkey LAG-3;

(b) does not bind mouse LAG-3;

(c) inhibits the binding of LAG-3 to class II molecules of the major histocompatibility complex (MHC, MHC) and (d) stimulates the immune response.

Preferably, the antibody exhibits at least two of these properties (a), (b), (c) and (d). More preferably, the antibody exhibits at least three of these properties (a), (b), (c) and (d). Even more preferably, the antibody exhibits all four of these properties (a), (b), (c) and (d).

- 1 043093

In a preferred embodiment of the invention, the antibody stimulates an antigen-specific T cell response, such as the production of interleukin-2 (IL-2) during the antigen-specific T cell response. In other embodiments, the antibody stimulates an immune response, such as an antitumor response (eg, inhibits tumor growth in an in vivo tumor graft model) or an autoimmune response (eg, stimulates the development of diabetes in NOD mice). In another preferred embodiment, the antibody binds an epitope of human LAG-3 containing the amino acid sequence PGHPLAPG (SEQ ID NO: 76). In yet another preferred embodiment, the antibody binds an epitope of human LAG-3 containing the amino acid sequence HPAAPSSW (SEQ ID NO: 77) or PAAPSSWG (SEQ ID NO: 78). In other embodiments, the antibody binds to human LAG-3 with a KD _of 1x10 ^-7 M or less, or binds to human LAG-3 with a KD of 1x10 ^-8 M or less, or binds to human LAG-3 with a KD of 5x10 ^-9 M or less, or binds to human LAG-3 with a K _D of 1x10 ^-9 M or less. In one embodiment, the antibody stains mucosal tissue in an immunohistochemical assay, while in another embodiment, the antibody does not stain mucosal tissue in an immunohistochemical assay.

In another aspect, the invention provides an isolated human monoclonal antibody, or antigen binding region thereof, characterized in that the antibody competes for binding to human LAG-3 with a reference antibody, wherein the reference antibody comprises:

(a) a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 37, and a light chain variable region containing the amino acid sequence of SEQ ID NO: 43;

(b) a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 38, and a light chain variable region containing the amino acid sequence of SEQ ID NO: 44;

(c) a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 39, and a light chain variable region containing the amino acid sequence of SEQ ID NO: 45;

(d) a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 40, and a light chain variable region containing the amino acid sequence of SEQ ID NO: 46;

(e) a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 41, and a light chain variable region containing the amino acid sequence of SEQ ID NO: 47; or (e) a heavy chain variable region comprising the amino acid sequence of SEQ ID NO: 42 and a light chain variable region comprising the amino acid sequence of SEQ ID NO: 48.

In a preferred embodiment, the reference antibody comprises a heavy chain variable region comprising the amino acid sequence of SEQ ID NO: 37 and a light chain variable region comprising the amino acid sequence of SEQ ID NO: 43. In another preferred embodiment of the invention, the reference antibody comprises a heavy chain variable region comprising the amino acid sequence of SEQ ID NO: 38, and a light chain variable region comprising the amino acid sequence of SEQ ID NO: 44. In another preferred embodiment of the invention, the reference antibody comprises a heavy chain variable region comprising the amino acid sequence of SEQ ID NO: 39, and a light chain variable region comprising the amino acid sequence of SEQ ID NO: 45. In another preferred embodiment, the reference antibody comprises a heavy chain variable region comprising the amino acid sequence of SEQ ID NO: 40 and a light chain variable region comprising the amino acid sequence of SEQ ID NO: 46. In another preferred embodiment In an embodiment of the invention, the reference antibody comprises a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 41, and a light chain variable region containing the amino acid sequence of SEQ ID NO: 47. In another preferred embodiment of the invention, the reference antibody comprises a heavy chain variable region containing the amino acid sequence SEQ ID NO: 42, and a light chain variable region containing the amino acid sequence of SEQ ID NO: 48.

In another aspect, the invention provides an isolated monoclonal antibody, or an antigen binding region thereof, comprising a heavy chain variable region that is the product of or derived from the human _VH 3-20 gene, the human _VH 4-34 gene, the human _VH 3 gene -33 or human V _H 1-24 gene, wherein the antibody specifically binds human LAG-3. In another aspect, the invention provides an isolated monoclonal antibody, or an antigen binding region thereof, comprising a light chain variable region that is the product of or derived from the human _VK L18 gene, the human _VK L6 gene, or the human _VK A27 gene, wherein the antibody specifically binds human LAG-3. IN

- 2043093 In a preferred embodiment, the invention includes an isolated monoclonal antibody, or an antigen-binding portion thereof, comprising:

(a) a heavy chain variable region that is the product of or derived from the human VH 4-34 gene, and a light chain variable region that is the product of or derived from the human V _K L6 gene;

(b) a heavy chain variable region that is the product of or derived from the human VH 3-33 gene and a light chain variable region that is the product of or derived from the human V _K A27 gene;

(c) a heavy chain variable region that is the product of or derived from the human VH 3-20 gene and a light chain variable region that is the product of or derived from the human V _K L18 gene;

(d) a heavy chain variable region that is the product of or derived from the human VH 1-24 gene, and a light chain variable region that is the product of or derived from the human V _K L6 gene; or (e) a heavy chain variable region that is the product of or derived from the human VH 3-33 gene and a light chain variable region that is the product of or derived from the human V _K L6 gene;

wherein the antibody specifically binds human LAG-3. In another aspect, the invention relates to an isolated monoclonal antibody, or an antigen binding region thereof, comprising:

(a) a heavy chain variable region comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 37-42;

(b) a light chain variable region comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 43-48; wherein the antibody specifically binds human LAG-3. Preferred combination includes:

(a) a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 37; and (b) a light chain variable region comprising the amino acid sequence of SEQ ID NO: 43.

Another preferred combination includes:

(a) a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 38; and (b) a light chain variable region comprising the amino acid sequence of SEQ ID NO: 44.

Another preferred combination includes:

(a) a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 39; and (b) a light chain variable region comprising the amino acid sequence of SEQ ID NO: 45.

Another preferred combination includes:

(a) a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 40; and (b) a light chain variable region comprising the amino acid sequence of SEQ ID NO: 46.

Another preferred combination includes:

(a) a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 41; and (b) a light chain variable region comprising the amino acid sequence of SEQ ID NO: 47.

Another preferred combination includes:

(a) a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 42; and (b) a light chain variable region comprising the amino acid sequence of SEQ ID NO: 48.

The antibodies of the invention may be, for example, full-length antibodies, for example of the IgG1, IgG2 or IgG4 isotype. In a preferred embodiment of the invention, the antibody is an IgG4 isotype antibody. In another preferred embodiment, the antibody is an IgG4 isotype antibody containing a serine to proline mutation at the heavy chain constant region hinge region (at a position corresponding to position 241 as described in Angal et al. (1993) Mol. Immunol. 30: 105-108), so that heterogeneity due to the disulfide bridge between the heavy chains is reduced or eliminated. Or antibodies

- 3043093 may be antibody fragments such as Fab, Fab' or Fab'2 fragments, or single chain antibodies.

This disclosure also includes an immunoconjugate comprising an antibody of the invention, or an antigen-binding site thereof, coupled to a therapeutic agent, such as a cytotoxin or radioactive isotope. This disclosure also includes a bispecific molecule comprising an antibody, or antigen binding site thereof, of the invention linked to a second functional entity having a binding specificity different from that of said antibody, or antigen binding site thereof.

The invention also includes compositions containing an antibody, or an antigen-binding fragment thereof, or an immunoconjugate, or a bispecific molecule of the invention and a pharmaceutically acceptable carrier.

Also included in this disclosure are nucleotide molecules encoding antibodies or antigen-binding regions thereof, as well as expression vectors containing such nucleic acids and host cells containing such expression vectors. The invention also includes methods for producing anti-LAG-3 antibodies using such expression vectors, these methods may include the steps of:

(i) expressing the antibody in the host cell; and (ii) releasing the antibody from the host cell.

Another aspect of the invention relates to methods of stimulating immune response(s) using anti-LAG-3 antibodies of the invention. For example, in one embodiment, the invention includes a method of stimulating an antigen-specific T-cell response by contacting said T-cell with an antibody of the invention such that an antigen-specific T-cell response is stimulated. In a preferred embodiment of the invention, the production of interleukin-2 by an antigen-specific T cell is stimulated. Another embodiment of the invention includes a method of stimulating an immune response (eg, an antigen-specific T-cell response) in a subject, comprising administering to the subject an antibody of the invention such that the subject's body's immune response (eg, an antigen-specific T-cell response) is stimulated. In a preferred embodiment of the invention, the subject is a tumor carrier and an immune response against the tumor is stimulated. In another preferred embodiment of the invention, the subject is a carrier of the virus and an immune response against the virus is stimulated.

Another aspect of the invention includes a method of inhibiting the growth of tumor cells in a subject, comprising administering to the subject an antibody of the invention such that tumor growth in the subject is inhibited. Another aspect of the invention includes a method of treating a viral infection in a subject, comprising administering to the subject an antibody of the invention so as to treat the viral infection in the subject.

Another embodiment of the invention includes a method of stimulating an immune response in a subject, comprising administering to the subject an anti-LAG-3 antibody and at least one additional immunostimulatory antibody, such as an anti-PD-1 antibody, an anti-PD-L1 antibody, and/or an anti-CTLA-antibody. 4 so that an immune response is stimulated in the subject, for example, to inhibit tumor growth or to stimulate an antiviral response. In one embodiment, the subject is administered an anti-LAG-3 antibody and an anti-PD-1 antibody. In another embodiment, the subject is administered an anti-LAG-3 antibody and an anti-PD-L1 antibody. In yet another embodiment, the subject is administered an anti-LAG-3 antibody and an anti-PD antibody; in another embodiment, the subject is administered an anti-LAG-3 antibody and an anti-CTLA-4 antibody. In one embodiment, the anti-LAG-3 antibody is a human antibody, such as an antibody as described herein. Alternatively, the anti-LAG-3 antibody may be, for example, a chimeric or humanized antibody. In another embodiment, the at least one additional immunostimulatory antibody (eg, anti-PD-1, anti-PD-L1, and/or anti-CTLA-4) is a human antibody. Or, the at least one additional immunostimulatory antibody may be, for example, a chimeric or humanized antibody.

In yet another aspect, the invention relates to a method for producing an anti-LAG-3 antibody. This method includes:

(a) provision:

(i) a heavy chain variable region sequence comprising a CDR1 sequence selected from the group consisting of SEQ ID NOs: 1-6, a CDR2 sequence selected from the group consisting of SEQ ID NOs: 7-12, and/or a CDR3 sequence, selected from the group consisting of SEQ ID NO: 13-14, GGY and 16-18; and/or (ii) a light chain variable region sequence comprising a CDR1 sequence selected from the group consisting of SEQ ID NOs: 19-24, a CDR2 sequence selected from the group consisting of SEQ ID NOs: 25-30, and/or a CDR3 sequence selected from the group consisting of SEQ ID NO: 31-36;

(b) a change in at least one amino acid residue in the variable sequence

- 4 043093 a new region of the antibody heavy chain and/or in the sequence of the variable region of the light chain of the antibody to create at least one altered antibody sequence; and (c) expression of the sequence-altered antibody as a protein.

Other features and advantages of the present disclosure will be apparent from the following detailed description and examples, which should not be construed as limiting. The contents of all references, Genbank data, patents and published patent applications cited in this application are expressly incorporated herein by reference.

Brief description of the figures

In fig. 1A shows the nucleotide sequence (SEQ ID NO: 49) and amino acid sequence (SEQ ID NO: 37) of the heavy chain variable region of the human monoclonal antibody 25F7. The CDR1 (SEQ ID NO: 1), CDR2 (SEQ ID NO: 7) and CDR3 (SEQ ID NO: 13) regions are highlighted and the germline V, D and J origins are indicated.

In fig. 1B shows the nucleotide sequence (SEQ ID NO: 55) and amino acid sequence (SEQ ID NO: 43) of the kappa light chain variable region of the human monoclonal antibody 25F7. The CDR1 (SEQ ID NO: 19), CDR2 (SEQ ID NO: 25) and CDR3 (SEQ ID NO: 31) regions are highlighted, and the germline V and D derivations are indicated.

In fig. 2A shows the nucleotide sequence (SEQ ID NO: 50) and amino acid sequence (SEQ ID NO: 38) of the heavy chain variable region of the human monoclonal antibody 26H10. The CDR1 (SEQ ID NO: 2), CDR2 (SEQ ID NO: 8) and CDR3 (SEQ ID NO: 14) regions are highlighted and the germline V, D and J origins are indicated.

In fig. 2B shows the nucleotide sequence (SEQ ID NO: 56) and amino acid sequence (SEQ ID NO: 44) of the kappa light chain variable region of the human monoclonal antibody 26H10. The CDR1 (SEQ ID NO: 20), CDR2 (SEQ ID NO: 26) and CDR3 (SEQ ID NO: 32) regions are highlighted, and the germline V and D derivations are indicated.

In fig. 3A shows the nucleotide sequence (SEQ ID NO: 51) and amino acid sequence (SEQ ID NO: 39) of the heavy chain variable region of the human monoclonal antibody 25E3. The CDR1 (SeQ ID NO: 3), CDR2 (SEQ ID NO: 9) and CDR3 regions are highlighted, and the germline derivations V, D and J are indicated.

In fig. 3B shows the nucleotide sequence (SEQ ID NO: 57) and amino acid sequence (SEQ ID NO: 45) of the kappa light chain variable region of the human monoclonal antibody 25E3. The CDR1 (SEQ ID NO: 21), CDR2 (SEQ ID NO: 27) and CDR3 (SEQ ID NO: 33) regions are highlighted, and the germline V and D derivations are indicated.

In fig. 4A shows the nucleotide sequence (SEQ ID NO: 52) and amino acid sequence (SEQ ID NO: 40) of the heavy chain variable region of the human monoclonal antibody 8B7. The CDR1 (SEQ ID NO: 4), CDR2 (SEQ ID NO: 10) and CDR3 (SEQ ID NO: 16) regions are highlighted and the germline derivations V, D and J are indicated.

In fig. 4B shows the nucleotide sequence (SEQ ID NO: 58) and amino acid sequence (SEQ ID NO: 46) of the kappa light chain variable region of the human monoclonal antibody 8B7. The CDR1 (SEQ ID NO: 22), CDR2 (SEQ ID NO: 28) and CDR3 (SEQ ID NO: 34) regions are highlighted, and the germline V and D derivations are indicated.

In fig. 5A shows the nucleotide sequence (SEQ ID NO: 53) and amino acid sequence (SEQ ID NO: 41) of the heavy chain variable region of the human monoclonal antibody 11F2. The CDR1 (SEQ ID NO: 5), CDR2 (SEQ ID NO: 11) and CDR3 (SEQ ID NO: 17) regions are highlighted and the germline derivations V, D and J are indicated.

In fig. 5B shows the nucleotide sequence (SEQ ID NO: 59) and amino acid sequence (SEQ ID NO: 47) of the kappa light chain variable region of the human monoclonal antibody 11F2. The CDR1 (SEQ ID NO: 23), CDR2 (SEQ ID NO: 29) and CDR3 (SEQ ID NO: 35) regions are highlighted and the germline V and D derivations are indicated.

In fig. 6A shows the nucleotide sequence (SEQ ID NO: 54) and amino acid sequence (SEQ ID NO: 42) of the heavy chain variable region of the human monoclonal antibody 17E5. The CDR1 (SEQ ID NO: 6), CDR2 (SEQ ID NO: 12) and CDR3 (SEQ ID NO: 18) regions are highlighted and the germline derivations V, D and J are indicated.

In fig. 6B shows the nucleotide sequence (SEQ ID NO: 60) and amino acid sequence (SEQ ID NO: 48) of the kappa light chain variable region of the human monoclonal antibody 17E5. The CDR1 (SEQ ID NO: 24), CDR2 (SEQ ID NO: 30) and CDR3 (SEQ ID NO: 36) regions are highlighted, and the germline V and D derivations are indicated.

In fig. 7 shows an alignment of the amino acid sequence of the heavy chain variable regions of antibody 25F7 (SEQ ID NO: 37) with the amino acid sequences of human germline VH 4-34 and JH5b (SEQ ID NO: 61 and 62, respectively).

In fig. 8 shows an alignment of the amino acid sequence of the light chain variable region of antibody 25F7 (SEQ ID NO: 43) with the amino acid sequences of human germline _VK L6 and JK2 (SEQ ID NO: 63 and 64, respectively).

- 5 043093

In fig. 9 shows an alignment of the amino acid sequence of the heavy chain variable regions of antibody 26H10 (SEQ ID NO: 38) with the amino acid sequences of human germline VH 3-33 and JH6B (SEQ ID NO: 65 and 66, respectively).

In fig. 10 shows an alignment of the amino acid sequence of the light chain variable region of antibody 26H10 (SEQ ID NO: 44) with the amino acid sequences of human germline _VK A27 and JK3 (SEQ ID NO: 67 and 68, respectively).

In fig. 11 shows an alignment of the amino acid sequence of the heavy chain variable regions of antibody 25E3 (SEQ ID NO: 39) with the amino acid sequences of human germline VH 3-20 and JH4b (SEQ ID NO: 69 and 70, respectively).

In fig. 12 shows an alignment of the amino acid sequence of the light chain variable region of antibody 25E3 (SEQ ID NO: 45) with the amino acid sequences of human germline _VK L18 and JK2 (SEQ ID NO: 71 and 64, respectively).

In fig. 13 shows an alignment of the amino acid sequence of the heavy chain variable regions of antibody 8B7 (SEQ ID NO: 40) with the amino acid sequences of human germline VH 4-34 and JH5b (SEQ ID NO: 61 and 62, respectively).

In fig. 14 shows an alignment of the amino acid sequence of the light chain variable region of antibody 8B7 (SEQ ID NO: 46) with the amino acid sequences of human germline _VK L6 and JK4 (SEQ ID NO: 63 and 72, respectively).

In fig. 15 shows an alignment of the amino acid sequence of the heavy chain variable regions of antibody 11F2 (SEQ ID NO: 41) with the amino acid sequences of human germline VH 1-24 and JH4b (SEQ ID NO: 73 and 70, respectively).

In fig. 16 shows an alignment of the amino acid sequence of the light chain variable region of antibody 11F2 (SEQ ID NO: 47) with the amino acid sequences of human germline VK L6 and JK1 (SEQ ID NO: 63 and 74, respectively).

In fig. 17 shows an alignment of the amino acid sequence of the heavy chain variable regions of antibody 17E5 (SEQ ID NO: 42) with the amino acid sequences of human germline _VH 3-33 and 2-2 (SEQ ID NO: 65 and 70, respectively).

In fig. 18 shows an alignment of the amino acid sequence of the light chain variable region of antibody 17E5 (SEQ ID NO: 48) with the amino acid sequences of human germline _VK L6 (SEQ ID NO: 63 and 75, respectively).

In fig. 19 shows an alignment of the protein sequence encoded by the monkey LAG-3 cDNA clone pa23-5 (SEQ ID NO: 93) with the rhesus macaque LAG-3 protein sequence deposited in Genbank (SEQ ID NO: 94) (Genbank Accession No. XM_001108923 ). The outer peptide loop region and transmembrane domain are underlined. One amino acid difference between the two sequences (amino acid position 419) is shown in bold.

Detailed Description of the Invention

The present disclosure relates to isolated monoclonal antibodies, in particular, to human monoclonal antibodies that bind to human LAG-3 and have the desired functional properties. In some embodiments, the antibodies of the invention are formed from specific germline heavy and light chain sequences and/or have specific structural features, such as CDR regions containing specific amino acid sequences. This disclosure includes isolated antibodies, methods for producing such antibodies, immunoconjugates and bispecific molecules containing such antibodies, and pharmaceutical compositions containing the antibodies, immunoconjugates or bispecific molecules of the invention. This disclosure also relates to methods of using antibodies, such as detection of LAG-3 protein, as well as methods of using anti-LAG-3 antibodies of the invention to stimulate an immune response, alone or in combination with other immunostimulatory antibodies. Accordingly, this disclosure also includes methods of using anti-LAG-3 antibodies of the invention, for example, to inhibit tumor growth or treat a viral infection.

In order to make the present disclosure clearer, certain terms are first defined. Additional definitions are provided throughout the detailed description.

The term LAG-3 refers to the lymphocyte activator gene-3. The term LAG-3 includes variants, isoforms, homologs, orthologs and paralogs. For example, antibodies specific for human LAG-3 protein may, in some cases, cross-react with a LAG-3 protein of a species other than human LAG-3 protein. In other embodiments, antibodies specific for human LAG-3 protein may be completely specific for human LAG-3 protein and may not exhibit species or other type of cross-reactivity, or may cross-react with some but not all LAG-3 ( e.g., cross-reacts with monkey LAG-3 but not with mouse LAG-3). The term human LAG-3 refers to human LAG-3 sequence, such as the complete amino acid sequence of human LAG-3 having Genbank Accession Number NP_002277. The term mouse LAG-3 refers to a mouse LAG-3 sequence, such as the complete amino acid sequence of mouse LAG-3 having Accession No.

- 6 043093

Genbank NP_032505. In the prior art, LAG-3 is also known, for example, as CD223. The human LAG-3 sequence may differ from the human LAG-3 sequence having Genbank accession number NP_002277 in that it may have, for example, conserved mutations or mutations in non-conserved regions and LAG-3 has essentially the same biological function as human LAG-3 from Genbank Registration No. NP_002277. For example, the biological function of human LAG-3 is that it has an epitope in the extracellular domain of LAG-3 that is specifically bound to an antibody of the present disclosure, or the biological function of LAG-3 is to bind to MHC Class II molecules.

The term simian LAG-3 (monkey LAG-3) is intended to encompass LAG-3 proteins expressed in cells of Old and New World monkeys, including, but not limited to, cynomolgus monkey LAG-3 and rhesus monkey LAG-3. A representative amino acid sequence of simian LAG-3 is the amino acid sequence of rhesus monkey LAG-3 shown in FIG. 19 and SEQ ID NO: 85, which is also deposited in Genbank, Registration No. ХМ_001108923. Another representative amino acid sequence of simian LAG-3 is the alternative sequence of the rhesus monkey clone pa23-5 shown in FIG. 19 and SEQ ID NO: 84, isolated as described in Example 3A, subsection 3. This alternative rhesus amino acid sequence has the only difference at position 419 from the amino acid sequence deposited in Genbank.

The specific sequence of human LAG-3 is typically at least 90% identical to the amino acid sequence of human LAG-3 deposited in Genbank, Accession No. NP_002277, and contains amino acid residues that define the amino acid sequence as a human sequence when compared with the LAG-3 amino acid sequences of other species (eg, mouse LAG-3). In some cases, the amino acid sequence of human LAG-3 may be at least 95%, or even at least 96, 97, 98, or 99% identical to the amino acid sequence of LAG-3 from Genbank Accession No. NP_002277. In some embodiments, there are no more than 10 amino acid differences between the human LAG-3 sequence and the LAG-3 sequence from Genbank Accession No. NP_002277. In some embodiments, there is no more than 5, or even no more than 4, 3, 2, or 1 amino acid difference between the human LAG-3 sequence and the LAG-3 sequence from Genbank Accession No. NP_002277. Percentage identity is determined as described herein.

The term immune response (immune response) refers to the action of, for example, lymphocytes, antigen presenting cells, phagocytic cells, granulocytes and soluble macromolecules produced by the above cells or the liver (including antibodies, cytokines and complement), which leads to selective damage, destruction or removal from human body invasive pathogens, cells or tissues infected with pathogens, cancer cells or, in the case of autoimmunity or pathological inflammation, normal human cells or tissues.

The expression antigen-specific T cell response refers to the T cell response(s) that result from stimulation of the T cell with an antigen to which the T cell is specific. Non-limiting examples of T cell responses to antigen-specific stimulation include proliferation and cytokine production (eg, IL-2 production).

The term antibody as used herein includes whole antibodies and any antigen-binding fragment (ie, antigen-binding region) or single chains thereof. Whole antibodies are glycoproteins containing at least two heavy (H) chains and two light (L) chains linked by disulfide bonds. Each heavy chain consists of a heavy chain variable region (abbreviated herein as VH) and a heavy chain constant region. The heavy chain constant region consists of three domains, CH1, CH2 and _CH3 . Each light chain consists of a light chain variable region (abbreviated herein as VL) and a light chain constant region. The light chain constant region consists of a single domain, CL. The V _H and V _L regions may also be subdivided into hypervariable regions called complementarity determining regions (CDRs), interspersed with more conserved regions called framework regions (FRs). Each V _H and V L region consists of three CDRs and four FRs, located from the amino terminus to the carboxy terminus in the following order: FR1, CDR1, FR2, CDR2, FR3, CDR3, FR4. The variable regions of the heavy and light chains contain a binding domain that interacts with the antigen. Antibody constant regions can mediate binding of immunoglobulin to host tissues or factors, including various cells of the immune system (eg, effector cells) and the first component (C1q) of the classical complement pathway.

The term antigen-binding region of an antibody (or simply antibody region) as used herein refers to one or more antibody fragments that retain the ability to specifically bind an antigen (eg, LAG-3 protein). It has been shown that the antigen-binding function of an antibody can be carried out by fragments of a full-length antibody. Examples of binding moieties included within the term antigen binding region of an antibody include:

- 7 043093 (i) Fab fragment, a monovalent fragment consisting of V _L , V H , C _L and CH1 domains;

(ii) the F(ab')2 fragment, a divalent fragment that is essentially a Fab with a hinge region portion (see, Fundamental immunology (Paul ed., 3. sup.rd ed. 1993);

(iv) Fd fragment consisting of VH and CH1 domains;

(v) an Fv fragment consisting of the V _L and V _H domains of a single arm (Fab fragment) of the antibody, (vi) a dAb fragment (Ward et al. (1989) Nature 341: 544-546), which consists of a VH domain;

(vii) a dedicated hypervariable region (CDR); and (viii) a nanobody, a heavy chain variable region containing one variable domain and two constant domains.

In addition, although the two domains of the Fv fragment, V _L and VH, are encoded by separate genes, they can be linked using recombinant DNA methods with a synthetic linker that allows them to be produced as a single protein chain in which the V _L and VH regions are paired to form monovalent molecules (known as single chain fragment Fv (scFv); see, for example, Bird et al. (1988) Science 242: 423-426; and Huston et al. (1988) Proc. Natl. Acad. Sci. USA 85; 5879-5883). It is intended that such single chain antibodies are also encompassed by the term antigen binding region of an antibody. These antibody fragments are prepared by conventional methods known to those skilled in the art, and these fragments are screened for utility by the same methods as intact antibodies.

The expression isolated antibody as used herein is intended to refer to an antibody that is substantially free of other antibodies having distinct antigen specificity (e.g., an isolated antibody that specifically binds the LAG-3 protein is substantially free of antibodies that specifically bind antigens other than LAG-3 proteins). An isolated antibody that specifically binds human LAG-3 protein may, however, cross-react with other antigens, such as LAG-3 protein from other species. In addition, the isolated antibody may contain substantially no other cellular material and/or chemical agents.

The terms monoclonal antibody or monoclonal antibody composition as used herein refer to a preparation of antibody molecules of a single molecular composition. A monoclonal antibody composition exhibits a single binding specificity and affinity for a particular epitope.

The term human antibody as used herein is intended to include antibodies having variable regions in which both framework and CDR regions are derived from germline immunoglobulin sequences. In addition, if the antibody contains a constant region, the constant region is also derived from human germline immunoglobulin sequences. Human antibodies of the invention may include amino acid residues not encoded by genes for human germline immunoglobulin sequences (eg, mutations introduced by random or site-specific mutagenesis in vitro or somatic mutation in vivo. However, the term human antibody as used herein is not intended to includes antibodies in which CDR sequences derived from the germ line of another mammalian species, such as a mouse, are grafted onto human framework sequences.

The expression human monoclonal antibody refers to antibodies exhibiting a single binding specificity having variable regions in which both framework and CDR regions are derived from human germline immunoglobulin sequences. In one embodiment, human monoclonal antibodies are produced by a hybridoma that includes a B cell derived from a transgenic non-human animal, for example, a transgenic mouse, having a genome containing a human heavy chain transgene and a light chain transgene fused to an immortalized cell.

The expression recombinant human antibody as used herein includes all human antibodies that are produced, expressed, created or isolated by a recombinant method, such as (a) antibodies isolated from an animal (eg, mouse) that are transgenic or trans-chromosomal for human immunoglobulin genes , or hybridomas produced using them (described in more detail below), (b) antibodies isolated using a host cell transformed to express a human antibody, such as transfectomes, (c) antibodies isolated from a recombinant combinatorial library of human antibodies, and ( d) antibodies produced, expressed, created or isolated by any other methods that involve splicing human immunoglobulin gene sequences with other DNA sequences.

Such recombinant human antibodies have variable regions in which the framework and CDR regions are derived from human germline immunoglobulin sequences.

- 8 043093 lines. However, in some embodiments, such recombinant human antibodies may be subject to in vitro mutagenesis (or, if the animal is transgenic for human Ig sequences, in vivo mutagenesis) and, therefore, the amino acid sequences of the VH and _VL regions of the recombinant antibodies are sequences that, being derived from human germline VH and _VL sequences may not naturally exist in the human germline antibody spectrum in vivo.

The term isotype refers to a class of antibodies (such as IgM or IgG1) that is encoded by heavy chain constant region genes.

The expressions antibody that recognizes an antigen and an antibody specific to an antigen are used interchangeably in this specification with the expression antibody that specifically binds an antigen.

The expression human antibody derivatives refers to any modified form of a human antibody, such as a conjugate of the antibody and another agent or antibody.

The term humanized antibody is intended to refer to antibodies in which CDR sequences derived from germline sequences of a mammal of another species, such as a mouse, are grafted onto human framework sequences. In human framework sequences, other modifications to the framework region can be made.

The term chimeric antibody is intended to refer to antibodies in which the variable region sequences are derived from one species and the constant region sequences are derived from another species, for example, an antibody in which the variable region sequences are derived from a murine antibody and the constant region sequences are derived from a human antibody. .

The term antibody that specifically binds human LAG-3 as used herein is intended to refer to an antibody that binds to the human LAG-3 protein (and possibly to the LAG-3 protein of one or more other non-human species), but practically does not bind to proteins other than the LAG-3 protein. Preferably, the antibody binds to human LAG-3 protein with high affinity, namely, a KD of 1x10 ^-7 M or less, more preferably 5x10 ^-8 M or less, more preferably 3x10 ^-8 M or less, more preferably 1x10 ' ⁸ M or less, more preferably 5x10 ^-9 M or less, or even more preferably 1x10 ^-9 M or less.

The expression as used herein essentially does not bind to a protein or cell means: does not bind or does not bind with high affinity to a protein or cell, i.e. binds to a protein or cell with a KD of 1x10 ^-6 M or more, more preferably 1x10 ^-5 M or more, more preferably 1x10 ^-4 M or more, more preferably 1x10 ^-3 M or more, even more preferably 1x10 ^{- 2} M or more.

The term K _assoc or K _a as used herein is intended to refer to a particular antibody-antigen interaction, while the term K _dis or Kd as used herein is intended to refer to the dissociation rate constant of a particular antibody-antigen interaction. The term KD as used herein is intended to refer to the dissociation constant, which is derived from the ratio of Kd to K _a (ie, Kd/Ka) and expressed as molar concentration (M). KD values for antibodies can be determined by methods generally accepted in the art. The preferred method for determining the KD of an antibody is surface plasmon resonance, preferably using a biosensor system such as the Biacore® system.

The term high-affinity IgG antibody refers to an antibody having a KD of 1x10 ^-7 M or less, more preferably 5x10 ^-8 M or less, more preferably 1x10 ^-8 M or less, more preferably 5x10 ^-9 M or less , more preferably 1x10 ^-9 M or less relative to the target antigen. However, high affinity binding may vary for other antibody isotypes. For example, high affinity for the IgM isotype refers to an antibody having a KD of 10 ^-6 M or less, more preferably 10 ^-7 M or less, even more preferably 10 ^-8 M or less.

The term subject includes a human or non-human animal. The term non-human animal includes all vertebrates, for example mammals and non-mammals such as non-human primates, sheep, dogs, cattle, cats, horses, chickens, amphibians and reptiles, although mammals such as non-human primates, sheep are preferred , dogs, cats, cattle and horses.

Various aspects of the invention are described in more detail in the subsections below.

Antibodies against LAG-3 with specific functional properties

The antibodies of the invention are characterized by specific functional characteristics or properties of the antibodies. For example, antibodies specifically bind to human LAG-3 and may bind to LAG-3 of some other species, such as simian LAG-3 (e.g., cynomolgus monkeys, rhesus monkeys), but have virtually no binding to LAG-3 of some other species, for example, mouse LAG-3. Preferably, the antibody of the invention binds to human LAG-3 with high affinity.

- 9 043093

The ability of an antibody to stimulate an immune response, such as an antigen-specific T-cell response, may be indicated, for example, by the ability of the antibody to stimulate interleukin-2 (IL-2) production in an antigen-specific T-cell response. In some embodiments, an antibody of the invention binds to human LAG-3 and exhibits the ability to stimulate an antigen-specific T cell response. In other embodiments, an antibody of the invention binds to human LAG-3 but does not exhibit the ability to stimulate an antigen-specific T cell response. Other methods for determining the ability of an antibody to stimulate an immune response include the ability of the antibody to inhibit tumor growth, such as in an in vivo tumor graft model (see, for example, Example 6), or the ability of an antibody to stimulate an autoimmune response, such as the ability to initiate (promote) the development of an autoimmune disease in an autoimmune model, for example, the ability to promote the development of diabetes in the NOD mouse model (see, for example, Example 7).

The binding of an antibody of the invention to LAG-3 can be assessed by one or more methods conventional in the art. For example, in a preferred embodiment of the invention, the antibody can be tested by flow cytometry in which the antibody is reacted with a cell line expressing human LAG-3, such as CHO cells transfected to express LAG-3 (e.g., human LAG-3 or simian LAG-3). LAG-3 (eg rhesus macaque or cynomolgus monkey) or mouse LAG-3) on its cell surface (for the corresponding analysis see, for example, example 3A). Other suitable cells for use in flow cytometry techniques include anti-CD3 antibody-stimulated activated CD4+ T cells that express native LAG-3. Additionally, or alternatively, antibody binding, including binding kinetics (eg, K _D value), can be determined using BIAcore binding assays (see, for example, Example 3B for appropriate assays). Other suitable binding assays include ELISA methods, for example using recombinant LAG-3 protein (see, for example, Example 1 for suitable assays).

Preferably, the antibody of the invention binds to a LAG-3 protein with a KD _of 5x10 ^-8 M or less, binds to a LAG-3 protein with a KD of 2x10 ^-8 M or less, binds to a LAG-3 protein with a KD of 5x10 ^-9 M or less , binds to LAG-3 protein with a K _D of 4x10 ^-9 M or less, binds to LAG-3 protein with a K _D of 3x10 ^-9 M or less, binds to LAG-3 protein with a K _{D of} 2x10 ^-9 M or less, binds with a LAG-3 protein with a K _D of 1x10 ^-9 M or less, binds to a LAG-3 protein with a K _{D of} 5x10 ^-10 M or less, or binds with a LAG-3 protein with a K _D of 1x10 ^-10 M or less.

Typically, the antibody of the invention binds to LAG-3 in lymphoid tissues such as tonsils, spleen or thymus, which can be detected by immunohistochemical methods. In addition, as further described in Example 8, some anti-LAG-3 antibodies of the invention stain pituitary tissue (e.g., retained in the pituitary gland) when immunohistochemically examined, whereas other anti-LAG-3 antibodies of the invention do not stain pituitary tissue (e.g., are not retained in the pituitary gland), with immunohistochemical examination. Thus, in one embodiment, the invention includes a human anti-LAG-3 antibody that stains pituitary tissue when immunohistochemically examined, while in another embodiment, the invention includes a human anti-LAG-3 antibody that does not stain pituitary tissue when immunohistochemically examined.

Preferred antibodies of the invention are human monoclonal antibodies. In addition, or alternatively, the antibodies may be, for example, chimeric or humanized antibodies.

Monoclonal antibodies 25F7, 26H10, 25E3, 8B7, 11F2 and 17E5 Preferred antibodies of the invention are human monoclonal antibodies 25F7, 26H10, 25E3, 8B7, 11F2 and 17E5, isolated and structurally characterized as described in examples 1 and 2. VH amino acid sequences of antibodies 2 5F7 , 26H10, 25E3, 8B7, 11F2 and 17E5 are shown in SEQ ID NO: 37-42, respectively. The _VK amino acid sequences of antibodies 25F7, 26H10, 25E3, 8B7, 11F2 and 17E5 are shown in SEQ ID NO: 43-48, respectively.

Given that each of these antibodies can bind to human LAG-3, the V _H and V _L sequences can be mixed and matched to create other anti-LAG-3 binding antibody molecules of the invention. Preferably, when V _H and V _L chains are mixed, the V _H sequence in a particular VH/VL pair is replaced by a structurally similar V _H sequence. Likewise, preferably, the VL sequence in a particular VH/VL pair is replaced by a structurally similar VL sequence.

Accordingly, in one aspect, the present invention includes an isolated monoclonal antibody, or antigen binding region thereof, comprising:

(a) a heavy chain variable region comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 47-42; and (b) a light chain variable region containing an amino acid sequence is selected

- 10 043093 from the group consisting of SEQ ID NO: 43-48;

wherein the antibody specifically binds human LAG-3. Preferred combinations of heavy and light chain include:

(a) a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 37, and a light chain variable region containing the amino acid sequence of SEQ ID NO: 43;

(b) a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 38, and a light chain variable region containing the amino acid sequence of SEQ ID NO: 44;

(c) a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 39, and a light chain variable region containing the amino acid sequence of SEQ ID NO: 45;

(d) a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 40, and a light chain variable region containing the amino acid sequence of SEQ ID NO: 46;

(e) a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 41, and a light chain variable region containing the amino acid sequence of SEQ ID NO: 47; or (e) a heavy chain variable region comprising the amino acid sequence of SEQ ID NO: 42 and a light chain variable region comprising the amino acid sequence of SEQ ID NO: 48.

In another aspect, this disclosure includes antibodies that comprise the CDR1, CDR2 and CDR3 regions of the heavy and light chains of antibodies 25F7, 26H10, 25E3, 8B7, 11F2 or 17E5, or combinations thereof. The VH CDR1 amino acid sequences of antibodies 25F7, 26H10, 25E3, 8B7, 11F2 and 17E5 are shown in SEQ ID NO: 37-42, respectively. The VH CDR2 amino acid sequences of antibodies 25F7, 26H10, 25E3, 8B7, 11F2 and 17E5 are shown in SEQ ID NO: 43-48, respectively. The VH CDR3 amino acid sequences of antibodies 25F7, 26H10, 25E3, 8B7, 11F2 and 17E5 are shown in SEQ ID NO: 13-14, GGY and 16-18, respectively. The _VK CDR1 amino acid sequences of antibodies 25F7, 26H10, 25E3, 8B7, 11F2 and 17E5 are shown in SEQ ID NO: 19-24, respectively. The amino acid sequences of the V _K CDR2s of antibodies 25F7, 26H10, 25E3, 8B7, 11F2 and 17E5 are shown in SEQ ID NO: 25-30. The amino acid sequences of V _K CDR3 antibodies 25F7, 26H10, 25E3, 8B7, 11F2 and 17E5 are shown in SEQ ID NO: 31-36, respectively. CDR regions are highlighted (outlined) using the Kabat system (Kabat et al. (1991) Sequences of Proteins of Immunological Interest, Fifth Edition, US Department of Health and Human Services, NIH Publication No. 91-3242).

Given that each of the antibodies can bind to human LAG-3 and that specific binding to the antigen is mediated primarily by the CDR1, CDR2 and CDR3 regions, the V _H CDR1, CDR2 and CDR3 sequences and the V _L CDR1, CDR2 and CDR3 sequences can be mixed (mixed and matched) (i.e., CDR regions from different antibodies can be mixed, although each antibody must contain V _H CDR1, CDR2 and CDR3 and V _L CDR1, CDR2 and CDR3) to create other anti-LAG binding antibody molecules 3 according to the invention. LAG-3 binding by such mixed and matched antibodies can be assayed using the assay methods and Examples described above (eg, ELISA assays, Biacore®). Preferably, when VH CDR sequences are mixed, the CDR1, CDR2 and/or CDR3 sequence from a particular VH sequence is replaced with structurally similar CDR sequence(s). Likewise, when VL CDR sequences are mixed, the CDR1, CDR2 and/or CDR3 sequence from a particular VL sequence is preferably replaced with structurally similar CDR sequence(s). One skilled in the art will appreciate that new VH and VL sequences can be created by replacing the sequences of one or more CDR regions of the VH and/or VL chains with structurally similar sequences from the CDR sequences disclosed herein for monoclonal antibodies 25F7, 26H10, 25E3, 8B7, 11F2 and 17E5.

Accordingly, in another aspect, this disclosure includes an isolated monoclonal antibody, or antigen binding region thereof, comprising:

(a) a heavy chain variable region CDR1 domain comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 1-6;

(b) a heavy chain variable region CDR2 domain comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 7-12;

(c) a heavy chain variable region CDR3 domain comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 13-14, GGY and 16-18;

(d) a light chain variable region CDR1 domain comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 19-24;

(e) a light chain variable region CDR2 domain comprising an amino acid sequence selected from the group consisting of SEQ ID NO: 25-30; And

- 11 043093 (e) a light chain variable region CDR3 domain containing an amino acid sequence selected from the group consisting of SEQ ID NO: 31-36;

wherein the antibody specifically binds human LAG-3.

In a preferred embodiment of the invention, the antibody contains:

(a) a heavy chain variable region CDR1 domain containing SEQ ID NO: 1;

(b) a heavy chain variable region CDR2 domain containing SEQ ID NO: 7;

(c) the heavy chain variable region CDR3 domain containing SEQ ID NO: 13;

(d) a light chain variable region CDR1 domain containing SEQ ID NO: 19;

(e) a light chain variable region CDR2 domain containing SEQ ID NO: 25; and (e) a light chain variable region CDR3 domain comprising SEQ ID NO: 31.

In another preferred embodiment of the invention, the antibody contains:

(a) a heavy chain variable region CDR1 domain containing SEQ ID NO: 2;

(b) a heavy chain variable region CDR2 domain containing SEQ ID NO: 8;

(c) the heavy chain variable region CDR3 domain containing SEQ ID NO: 14;

(d) a light chain variable region CDR1 domain containing SEQ ID NO: 20;

(e) a light chain variable region CDR2 domain containing SEQ ID NO: 26; and (e) a light chain variable region CDR3 domain comprising SEQ ID NO: 32.

In another preferred embodiment of the invention, the antibody contains:

(a) heavy chain variable region CDR1 domain containing SEQ ID NO: 3 (b) heavy chain variable region CDR2 domain containing SEQ ID NO: 9;

(c) heavy chain variable region CDR3 domain containing GGY;

(d) a light chain variable region CDR1 domain containing SEQ ID NO: 21;

(e) a light chain variable region CDR2 domain containing SEQ ID NO: 27; and (e) a light chain variable region CDR3 domain comprising SEQ ID NO: 33.

In another preferred embodiment of the invention, the antibody contains:

(a) a heavy chain variable region CDR1 domain containing SEQ ID NO: 4;

(b) a heavy chain variable region CDR2 domain containing SEQ ID NO: 10;

(c) the heavy chain variable region CDR3 domain containing SEQ ID NO: 16;

(d) a light chain variable region CDR1 domain containing SEQ ID NO: 22;

(e) a light chain variable region CDR2 domain containing SEQ ID NO: 28; and (e) a light chain variable region CDR3 domain comprising SEQ ID NO: 34.

In another preferred embodiment of the invention, the antibody contains:

(a) a heavy chain variable region CDR1 domain containing SEQ ID NO: 5;

(b) a heavy chain variable region CDR2 domain containing SEQ ID NO: 11;

(c) the heavy chain variable region CDR3 domain containing SEQ ID NO: 17;

(d) a light chain variable region CDR1 domain containing SEQ ID NO: 23;

(e) a light chain variable region CDR2 domain containing SEQ ID NO: 29; and (e) a light chain variable region CDR3 domain comprising SEQ ID NO: 35.

In another preferred embodiment of the invention, the antibody contains:

(a) a heavy chain variable region CDR1 domain containing SEQ ID NO: 6;

(b) a heavy chain variable region CDR2 domain containing SEQ ID NO: 12;

(c) the heavy chain variable region CDR3 domain containing SEQ ID NO: 18;

(d) a light chain variable region CDR1 domain containing SEQ ID NO: 24;

(e) a light chain variable region CDR2 domain containing SEQ ID NO: 30; and (e) a light chain variable region CDR3 domain comprising SEQ ID NO: 36.

It is well known in the art that the CDR3 domain, independent of the CDR1 and/or CDR2 domain(s), alone can determine the binding specificity of an antibody to its antigen and that based on a common CDR3 sequence, multiple antibodies with the same binding specificity can be confidently produced. See, for example,

Klimka et al., British J. of Cancer 83(2): 252-260 (2000); Beiboer et aL, J.

Mol. Biol. 296: 833-849 (2000); Rader et al, Proc. Natl. Acad. Sci. U.S.A. 95: 8910-8915 (1998); Barbas et al., J. Am. Chem. Soc. 116: 2161-2162 (1994); Barbas et al., Proc. Natl.

Acad. Sci. U.S.A. 92: 2529-2533 (1995); Ditzel et al., J. Immunol. 157: 739-749 (1996);

Berezov et al, BIAjournal 8: Scientific Review 8 (2001); Igarashi et al, J. Biochem (Tokyo)

117: 452-7 (1995); Bourgeois et al, J. Virol 72: 807-10 (1998); Levi et aL, Proc. Natl.

Acad. Sci. U.S.A. 90: 4374-8 (1993); Polymenis and Stoller, J. Immunol. 152: 5218-5329 (1994) and Xu and Davis, Immunity 13: 37-45 (2000). See also US Patent Nos.

6951646; 6914128; 6090382; 6818216; 6156313; 6827925; 5833943; 5762905 and 5760185.

Each of these references is incorporated herein by reference in its entirety.

- 12 043093

Accordingly, the present disclosure includes monoclonal antibodies comprising one or more heavy and/or light chain CDR3 domains from a human or non-human animal antibody, wherein the monoclonal antibody is capable of specifically binding to human LAG-3. In some aspects, the present disclosure includes monoclonal antibodies comprising one or more heavy and/or light chain CDR3 domains from a non-human antibody, such as a mouse or rat antibody, wherein the monoclonal antibody is capable of specifically binding to human LAG3. In some embodiments, such antibodies of the invention comprising one or more heavy and/or light chain CDR3 domains from a non-human antibody:

(a) capable of competing for binding with the corresponding parent non-human antibody;

(b) retain the functional characteristics of the corresponding parent non-human antibody;

(c) bind to the same epitope as the corresponding parent non-human antibody; and/or (d) have a binding affinity similar to the binding affinity of the corresponding parent non-human antibody.

In other aspects, the present disclosure includes monoclonal antibodies comprising one or more heavy and/or light chain CDR3 domains from a human antibody, such as a human antibody produced using a non-human animal, wherein the human antibody is capable of specifically binding to human LAG-3 . In other aspects, the present disclosure includes monoclonal antibodies comprising one or more heavy and/or light chain CDR3 domains of a first human antibody, such as a human antibody produced using a non-human animal, wherein the human antibody is capable of specifically binding to human LAG-3 and the CDR3 domain from the first human animal replaces the CDR3 domain in the human antibody that lacks binding specificity for LAG-3, thereby producing a second human antibody that is capable of specifically binding to human LAG-3. In some embodiments, such antibodies of the invention comprising one or more heavy and/or light chain CDR3 domains from a first human antibody are (a) capable of competing for binding with the corresponding parent first human antibody;

(b) retain the functional characteristics of the corresponding parent first human antibody;

(c) bind to the same epitope as the corresponding parent first human antibody and/or (d) have a binding affinity similar to the binding affinity of the corresponding parent first human antibody.

Antibodies having specific germline sequences

In some embodiments, the antibodies of the invention comprise a heavy chain variable region derived from a germline-specific immunoglobulin heavy chain gene and/or a germline-specific immunoglobulin light chain gene.

For example, a preferred embodiment of the invention includes an isolated monoclonal antibody, or antigen binding region thereof, comprising a heavy chain variable region that is the product of or derived from the human VH 3-20 gene, the human VH 4-34 gene, the human VH 3- 33 or the human VH 1-24 gene, wherein the antibody specifically binds human LAG-3. Another preferred embodiment of the invention includes an isolated monoclonal antibody, or an antigen-binding region thereof, comprising a light chain variable region that is the product of or derived from the human VK L18 gene, the human VK L6 gene, or the human VK A27 gene, wherein the antibody specifically binds human LAG-3 protein. Another preferred embodiment of the invention includes an isolated monoclonal antibody, or an antigen binding region thereof, comprising a heavy chain variable region that is the product of or derived from the human V _H 3-20 gene and includes an isolated monoclonal antibody, or an antigen binding region thereof, containing (ii) a light chain variable region that is the product of or derived from the human VK L18 gene, wherein the antibody specifically binds the human LAG-3 protein. Another preferred embodiment of the invention includes an isolated monoclonal antibody, or an antigen binding region thereof, comprising a heavy chain variable region that is the product of or derived from the human VH 4-34 gene, and includes an isolated monoclonal antibody, or an antigen binding region thereof, comprising (ii) a light chain variable region that is the product of or derived from the human VK L6 gene, wherein the antibody specifically binds the human LAG-3 protein. Another preferred embodiment of the invention includes an isolated monoclonal antibody, or antigen-binding region thereof, comprising a heavy chain variable region that is the product of or derived from

- 13 043093 when using the human VH 3-33 gene, and includes an isolated monoclonal antibody, or an antigen-binding region thereof, containing a light chain variable region that is the product of or formed using the human _VK A27 gene, wherein the antibody specifically binds human LAG-3 protein. Another preferred embodiment of the invention includes an isolated monoclonal antibody, or an antigen binding region thereof, comprising a heavy chain variable region that is the product of or derived from the human VH 1-24 gene, and includes an isolated monoclonal antibody, or an antigen binding region thereof, comprising (ii) a light chain variable region that is the product of or derived from the human _VK L6 gene, wherein the antibody specifically binds the human LAG-3 protein. Another preferred embodiment of the invention includes an isolated monoclonal antibody, or an antigen binding region thereof, comprising a heavy chain variable region that is the product of or derived from the human VH 3-33 gene, and includes an isolated monoclonal antibody, or an antigen binding region thereof, comprising (ii) a light chain variable region that is the product of or derived from the human _VK L6 gene, wherein the antibody specifically binds the human LAG-3 protein.

Such antibodies may also have one or more of the functional characteristics described in detail above, such as high affinity binding to human LAG-3, binding to simian LAG-3, lack of binding to murine LAG-3, the ability to inhibit the binding of LAG-3 to MHC molecules ( MHC) Class II and/or the ability to stimulate antigen-specific T-cell responses (responses).

An example of an antibody whose VH and V _L are VH 3-20 and VK L18, respectively, is antibody 25E3. Examples of antibodies in which VH and VL are VH 4-34 and V _K L6, respectively, are antibodies 25F7 and 8B7. An example of an antibody whose VH and VL are VH 3-33 and _VK A27, respectively, is antibody 26H10. An example of an antibody whose VH and VL are VH 1-24 and VK L6, respectively, is antibody 11F2. An example of an antibody whose VH and VL are VH 3-33 and V _K L6, respectively, is antibody 17E5.

As used herein, a human antibody comprises heavy or light chain variable regions that are the product of or generated using a specific germline sequence if the antibody variable regions are produced using a system that uses human immunoglobulin genes. Such systems include immunization of transgenic mice carrying human immunoglobulin genes with the target antigen or screening of a human immunoglobulin phage display library visualized by the target antigen. A human antibody that is the product of or generated by using germline genes for human immunoglobulin sequences can be identified by comparing the amino acid sequence of the human antibody to the amino acid sequences of human germline immunoglobulins and selecting the human germline immunoglobulin sequence that is most similar (i.e., having the greatest % identity) to the sequence of a human antibody. A human antibody that is the product of or generated using germline genes for human immunoglobulin sequences may contain amino acid differences compared to the germline sequence due to, for example, naturally occurring somatic mutations or targeted introduction of a site-directed mutation. Typically, however, the amino acid sequence of the selected human antibody is at least 90% identical to the amino acid sequence encoded by the germline immunoglobulin gene and contains amino acid residues that identify the human antibody as human in comparison to the amino acid sequences of germline immunoglobulins of other species (eg , mouse germline sequences). In some cases, the amino acid sequence of a human antibody may be at least 95% identical, or even at least 96%, 97%, 98%, or 99% identical to the amino acid sequence encoded by a germline immunoglobulin gene. Typically, a human antibody generated from a particular human germline sequence exhibits no more than 10 amino acid differences compared to the amino acid sequence encoded by the human germline immunoglobulin gene. In some cases, a human antibody may have no more than 5 or even no more than 4, 3, 2 or 1 amino acid difference compared to the amino acid sequence encoded by the human germline immunoglobulin gene.

Homologous antibodies

In yet another embodiment, the antibody of the invention contains heavy and light chain variable regions having amino acid sequences homologous to the amino acid sequences of the preferred antibodies herein, and the antibodies retain the necessary functional properties of the anti-LAG-3 antibodies of the invention. For example, this disclosure includes an isolated monoclonal antibody, or antigen binding region thereof, comprising a heavy chain variable region and a light chain variable region, different

- 14 043093 because:

(a) the heavy chain variable region contains an amino acid sequence that is at least 80% homologous to an amino acid sequence selected from the group consisting of SEQ ID NOs: 37-42;

(b) the light chain variable region contains an amino acid sequence that is at least 80% homologous to an amino acid sequence selected from the group consisting of SEQ ID NOs: 43-48; and (c) the antibody specifically binds to human LAG-3.

In addition or alternatively, the antibody may have one or more of the following functional properties discussed above, such as high affinity binding to human LAG-3, binding to simian LAG-3, non-binding to murine LAG-3, ability to inhibit LAG-3 binding with MHC Class II molecules and/or the ability to stimulate antigen-specific T-cell responses (responses).

In various embodiments, the antibody may be, for example, a human antibody, a humanized antibody, or a chimeric antibody.

In other embodiments, the VH and/or _VL amino acid sequences may be 85, 90, 95, 96, 97, 98, or 99% homologous to the sequences presented above. An antibody having sequences of the VH and _VL regions having high homology (i.e., 80% or greater) with the sequences of the _VH and _VL regions presented above can be produced by mutagenesis (e.g., site-directed or PCR-mediated mutagenesis) of nucleotide molecules encoding SEQ ID NO: 49-54 or 55-60, followed by testing the encoded altered antibody for retained function (i.e., the functions presented above) using the functional assays presented herein.

As used herein, the percentage of homology (degree of homology in percentage) between two amino acid sequences is equivalent to the percentage of identity (degree of identity in percentage) between the two sequences. The percentage identity between two sequences is a function of the number of identical positions shared by the sequences (i.e., % homology = number of identical positions/total number of positions x 100), taking into account the number of gaps and the length of each gap that must be introduced for optimal alignment of the two sequences . Sequence comparisons and determination of percent identity between two sequences can be performed using a mathematical algorithm as described in the non-limiting examples below.

The percentage identity between two amino acid sequences can be determined using the algorithm of E. Meyers and W. Miller (Comput. Appl. Biosci., 4: 11-17 (1988)), which is included in the ALIGN program (version 2.0) using the PAM120 residue weight table, a gap length penalty of 12 and a gap penalty of 4. In addition, the percent identity between two amino acid sequences can be determined using the Needleman and Wunsch algorithm (J. Mol. Biol 48: 444-453 (1970)), which is included in the GAP program in GCG software (available at http://www.gcg.com), using either a Blossum 62 matrix or a PAM250 matrix, gap weight (weighted average) 16, 14, 12, 10, 8, 6 or 4 and a weighted average of the gap length of 1, 2, 3, 4, 5 or 6.

In addition, or alternatively, the protein sequences of the present disclosure can further be used as query sequences to conduct searches of publicly available databases, for example, to identify related sequences. Such searches can be performed using the XBLAST program (version 2.0) of Altschul et al. (1990) J. Mol. Biol. 215:403-10. A BLAST search for proteins can be performed using the XBLAST program, score=50, word length=3, obtaining amino acid sequences homologous to the antibody molecules of the invention. To perform gap comparisons, you can use the Gapped BLAST program described by Altschul et al. (1997) Nucleic Acids Res. 25(17): 3389-3402. When using the BLAST and Gapped BLAST programs, the default settings of the corresponding programs (for example, XBLAST and NBLAST) are used. See www.ncbi.nlm.nih.gov.

Antibodies with conservative modifications

In some embodiments, the antibodies of the invention comprise a heavy chain variable region comprising CDR1, CDR2 and CDR3 sequences, a light chain variable region comprising CDR1, CDR2 and CDR3 sequences, wherein one or more of these CDR sequences comprise specific amino acid sequences based on preferred antibodies , presented in this description (for example, 25F7, 26H10, 25E3, 8B7, 11F2, 17E5), or their conservative modifications, and the antibodies retain the necessary functional properties of the anti-LAG-3 antibodies of the invention. It is clear from the prior art that certain modifications can be made to the conserved sequence without affecting antigen binding. See, for example,

- 15 043093

Brummell et al. (1993) Biochem 32: 1180-8; de Wildt et al. (1997) Prot.

Eng. 10: 835-41; Komissarov et al. (1997) J Biol. Chem. 272: 26864-26870; Hall et al.

(1992) J. Immunol. 149: 1605-12; Kelley and O'Connell (1993) Biochem. 32: 6862-35;

Adib-Conquy et al. (1998) Int. Immunol. 10: 341-6 and Beers et al. (2000) Clin. Can. Res. 6: 2835-43.

Accordingly, this disclosure includes an isolated monoclonal antibody, or an antigen binding region thereof, comprising a heavy chain variable region comprising the sequences CDR1, CDR2 and CDR3, and a light chain variable region comprising the sequences CDR1, CDR2 and CDR3, wherein:

(a) the heavy chain variable region CDR3 sequence is an amino acid sequence selected from the group consisting of the amino acid sequences of SEQ ID NOs: 13-14, GGY and 16-18 and conservative modifications thereof;

(b) the light chain variable region CDR3 sequence is an amino acid sequence selected from the group consisting of the amino acid sequences of SEQ ID NOs: 31-36 and conservative modifications thereof; and (c) the antibody specifically binds human LAG-3.

In addition or alternatively, the antibody may have one or more of the following functional properties described above, such as high affinity binding to human LAG-3, binding to simian LAG-3, non-binding to murine LAG-3, ability to inhibit LAG-3 binding with MHC Class II molecules and/or the ability to stimulate antigen-specific T-cell responses (responses).

In a preferred embodiment of the invention, the heavy chain variable region CDR2 sequence is an amino acid sequence selected from the group consisting of the amino acid sequences of SEQ ID NO: 7-12 and conservative modifications thereof; a CDR2 light chain variable region sequence is an amino acid sequence selected from the group consisting of the amino acid sequences of SEQ ID NO: 25-30 and conservative modifications thereof. In another preferred embodiment of the invention, the heavy chain variable region CDR1 sequence is an amino acid sequence selected from the group consisting of the amino acid sequences of SEQ ID NOs: 1-6 and conservative modifications thereof; a The light chain variable region CDR1 sequence is an amino acid sequence selected from the group consisting of the amino acid sequences of SEQ ID NO: 19-24.

In various embodiments, the antibody may be, for example, a human antibody, a humanized antibody, or a chimeric antibody.

Conservative sequence modification expressions (conservative modifications) as used herein are intended to refer to amino acid modifications that do not significantly affect or slightly alter the binding characteristics of an antibody containing that amino acid sequence. Such conservative modifications include amino acid substitutions, additions and deletions. Modifications can be introduced into the antibody of the invention by conventional methods known in the art, such as site-directed mutagenesis and PCR-mediated mutagenesis. Conservative amino acid substitutions are those substitutions in which an amino acid residue is replaced by an amino acid residue having a similar side chain. Families of amino acid residues with similar side chains are defined in the prior art. These families include amino acids with basic side chains (eg, lysine, arginine, histidine). acidic side chains (e.g. aspartic acid, glutamic acid), uncharged polar side chains (e.g. glycine, asparagine, glutamine, serine, threonine, tyrosine, cysteine, tryptophan), non-polar side chains (e.g. alanine, valine, leucine, isoleucine , proline, phenylalanine, methionine), beta-branched side chains (e.g. threonine, valine, isoleucine) and aromatic side chains (e.g. tyrosine, phenylalanine, tryptophan, histidine). Thus, one or more amino acid residues in the CDR regions of an antibody of the invention can be replaced with other amino acid residues from the same side chain family, and the altered antibody can be tested for retained function (i.e., the functions presented above) using functional assays according to this description.

Antibodies that bind to the same epitope as anti-LAG-3 antibodies

In another embodiment, the invention includes antibodies that bind to the same epitope on LAG-3 as any of the anti-LAG-3 monoclonal antibodies of the invention (i.e., antibodies that have the ability to compete for binding to human LAG-3 with any from monoclonal antibodies according to the invention). In preferred embodiments of the invention, the reference antibody for studying competitive reactions may be one of the monoclonal antibodies 25F7, 26H10, 25E3, 8B7, 11F2 or 17E5.

- 16 043093

Such competitive antibodies can be identified based on their ability to compete with 25F7, 26H10, 25E3, 8B7, 11F2 and/or 17E5 in standard LAG-3 binding assays. For example, standard ELISA assays can be used in which recombinant human LAG-3 protein is immobilized on a plate, one of the antibodies is fluorescently labeled, and the ability of the unlabeled antibody to compete with the unlabeled antibody for binding is assessed. Additionally or alternatively, the BIAcore assay can be used to assess the ability of antibodies to compete. The ability of the test antibody to inhibit the binding of, for example, 25F7, 26H10, 25E3, 8B7, 11F2 and/or 17E5 antibodies to human LAG-3 indicates that the test antibody can compete with the 25F7, 26H10, 25E3, 8B7, 11F2 and/or 17E5 antibody for binding to human LAG-3 and therefore binds to the same epitope on human LAG-3 as 25F7, 26H10, 25E3, 8B7, 11F2 and/or 17E5. In a preferred embodiment of the invention, the antibody that binds to the same epitope on human LAG-3 as 25E3, 25F7, 8B7, 26H10, 11F2 or 17E5 is a human monoclonal antibody. Such human monoclonal antibodies can be prepared and isolated as described in the examples.

As discussed further in Example 3C, binding of 25E3, 25F7 and 8B7 to human LAG-3 mapped to the extra loop region of the first extracellular domain of human LAG-3. The sequence of the extracellular (outer) loop region is shown in SEQ ID NO: 79. Using peptide scanning, the binding of 25E3 to the outer loop region is mapped to the following amino acid sequence: PGHPLAPG (SEQ ID NO: 76), while the binding of 25F7 to the outer loop region is mapped to the following amino acid sequence: HPAAPSSW (SEQ ID NO: 77), and the binding of 8B7 to the outer loop region is mapped to the following amino acid sequence: PAAPSSWG (SEQ ID NO: 78). Accordingly, in a preferred embodiment, the invention includes an anti-LAG-3 antibody that binds an epitope of human LAG-3 containing the amino acid sequence PGHPLAPG (SEQ ID NO: 76). In another preferred embodiment, the invention includes an anti-LAG-3 antibody that binds an epitope of human LAG-3 containing the amino acid sequence HPAAPSSW (SEQ ID NO: 77) or PAAPSSWG (SEQ ID NO: 78).

Genetically engineered and modified antibodies

In addition, it is possible to produce an antibody of the invention using an antibody having one or more V _H and/or V _L sequences as described herein as starting material to create a modified antibody, which modified antibody may have altered properties compared to the original antibody. . An antibody can be designed (created) by modifying one or more residues in one or both variable regions (i.e., VH and/or VL), for example, in one or more CDR domains and/or in one or more framework regions . In addition, or alternatively, an antibody can be created by modifying residues in the constant region(s), for example, to change the effector function(s) of the antibody.

In some embodiments, CDR grafting can be used to create antibody variable regions. Antibodies interact with target antigens primarily through amino acid residues localized in six hypervariable domains (CDRs) of the heavy and light chains. For this reason, the amino acid sequences in the CDRs of individual antibodies are more different from each other than the sequences outside the CDRs. Since CDR sequences are responsible for most antibody-antigen interactions, it is possible to express recombinant antibodies that mimic the properties of specific natural antibodies by constructing expression vectors that include CDR sequences from a specific natural antibody grafted onto framework sequences from another antibody with different properties (see ., for example, Riechmann et al. (1998) Nature 332: 323-327; Jones et al. (1986) Nature 321: 522-525; Queen et al. (1989) Proc. Natl. Acad. Sci. U.S.A. 86: 10029-10033; US Patent Nos. 5225539; 5530101; 5585089; 5693762 and 6180370.)

Accordingly, another embodiment of the invention provides an isolated monoclonal antibody, or antigen binding region thereof, comprising a heavy chain variable region comprising CDR1, CDR2, and CDR3 sequences having an amino acid sequence selected from the group consisting of SEQ ID NOs: 1-6, SEQ ID NOs: 7-12, and SEQ ID NOs: 13-14, GGY and 16-18, respectively, and a light chain variable region comprising the sequences CDR1, CDR2, and CDR3 having an amino acid sequence selected from the group consisting of SEQ ID NO: 19-24, SEQ ID NO: 25-30, and SEQ ID NO: 31-36, respectively. Thus, these antibodies containing the VH VL CDR sequences of monoclonal antibodies 25F7, 26H10, 25E3, 8B7, 11F2 or 17E5 may contain framework sequences different from these antibodies.

Such framework sequences can be obtained from a publicly available DNA database or from published reference materials that include germline antibody gene sequences. For example, germline DNA sequences for the heavy and light chain variable region genes can be found in the human germline sequence database VBase (available online at www.mrc-cpe.cam.ac.uk/vbase) as well as in Kabat et al. (1991), see above; Tomlinson et al. (1992) The Repertoire of Human Germline V _H Sequences Reveals about Fifty

- 17 043093

Groups of VH Segments with Different Hypervariable Loops J. Mol. Biol. 227-776-798; and Cox et al. (1994) A Directory of Human Germ- line V _H Segments Reveals a Strong Bias in their Usage Eur. J. Immunol. 24: 827836; the contents of each of these materials are expressly incorporated herein by reference. As another example, germline DNA sequences for human heavy and light chain variable region genes can be found in the Genbank database. For example, the following germline heavy chain sequences found in mouse HCo7 HuMAb are available in Genbank under the accompanying accession nos: 1-69 (NG 0010109, NT_024637 & BC070333), 3-33 (NG_0010109 & NT_024637) and 3-7 ( NG_0010109 & NT_024637).

As another example, the following germline heavy chain sequences found in mouse HCo7 HuMAb are available in Genbank under the accompanying accession

No.: 1- 69 (NG_0010109, NT_024637 & ВС070333), 551 (NG_0010109 & NT_024637), 4- 34 (NG_0010109 & NT_024637), 3- 30.3 (CAJ556644) & 3- 23 (AJ406 678).

The antibody protein sequences are compared to a compiled protein sequence database using one of the sequence similarity search methods called Gapped BLAST (Altschul et al. (1997), supra), well known in the art.

Preferred framework sequences for use in antibodies of the invention are sequences similar in structure to the framework sequences used in selected antibodies of the invention, for example, similar to the framework sequences _VH 3-20 (SEQ ID NO: 69), _VH 4-34 ( SEQ ID NO: 61), V _H 3-33 (SEQ ID NO: 65) or V _H 1-24 framework sequences (SEQ ID NO: 73) and/or V _K L18 framework sequences (SEQ ID NO: 71), _VK L6 (SEQ ID NO: 63) or _VK A27 (SEQ ID NO: 67) used in preferred monoclonal antibodies of the invention. VH CDR1, CDR2 and CDR3 sequences and V _K sequences CDR1, CDR2 and CDR3 can be grafted onto framework regions having a sequence identical to that found in the germline immunoglobulin gene from which the framework sequence is derived, or CDR sequences can be grafted onto framework regions , which contain one or more mutations compared to germline sequences. For example, it has been found that in some cases it is advantageous to mutate residues in framework regions to maintain or increase the antigen-binding ability of an antibody (see, for example, US Pat. Nos. 5,530,101; 5,585,089; 5,693,762 and 6,180,370).

Another type of variable region modification is the mutation of amino acid residues in the V _H and/or V _L CDR1, CDR2 and/or CDR3 regions to increase one or more binding properties (eg, the affinity of the target antibody). Site-directed or PCR-mediated mutagenesis can be performed to introduce mutation(s), and the effect on binding or other target functional property of the antibody can be determined using in vitro or in vivo assays provided herein and included in the examples. Preferably, conservative mutations (discussed above) are introduced. Mutations may be amino acid substitutions, additions or deletions, but are preferably substitutions. In addition, typically no more than one, two, three, four, or five residues are changed in a CDR region.

Accordingly, another embodiment of the invention includes isolated anti-LAG-3 monoclonal antibodies, or antigen binding regions thereof, comprising a heavy chain variable region having:

(a) a V _H CDR1 domain comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 1-6, or an amino acid sequence having one, two, three, four or five amino acid substitutions, deletions or additions compared to SEQ ID NO: 1-6;

(b) a VH CDR2 domain containing an amino acid sequence selected from the group consisting of SEQ ID NOs: 7-12, or an amino acid sequence having one, two, three, four or five amino acid substitutions, deletions or additions compared to SEQ ID NO: 7-12;

(c) a VH CDR3 domain containing an amino acid sequence selected from the group consisting of SEQ ID NOs: 13-14, GGY and 16-18, or an amino acid sequence having one, two, three, four or five amino acid substitutions, deletions or additions compared to SEQ ID NO: 13-14, GGY and 16-18;

(d) a VL CDR1 domain comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 19-24, or an amino acid sequence having one, two, three, four or five amino acid substitutions, deletions or additions compared to SEQ ID NO: 19-24;

(e) V _L CDR2 domain containing an amino acid sequence selected from the group consisting of SEQ ID NO: 25-30, or an amino acid sequence having one, two, three, four or five amino acid substitutions, deletions or additions compared to SEQ ID NO: 25-30; and (e) a V _L CDR3 domain comprising an amino acid sequence selected from the group consisting of SEQ ID NOs: 31-36, or an amino acid sequence having one, two, three, four or five amino acid substitutions, deletions or additions compared to SEQ ID NO: 31-36.

- 18 043093

Genetically engineered antibodies of the invention include those in which modifications are made to the framework residues in VH and/or VL, for example, to improve the properties of the antibody. Typically, such modifications to the framework sequence are made to reduce the immunogenicity of the antibody. For example, one method is to backmutate one or more framework residues into the corresponding germline sequence. More specifically, an antibody that undergoes a somatic mutation may contain framework residues that differ from the germline sequence from which the antibody is derived. Such residues can be identified by comparing the framework sequences of the antibody with the germline sequences from which the antibody is formed.

For example, in table. And shows the regions where the amino acid position in the framework region (using the Kabat numbering system) differs from the germline and what germline back mutation can be made at this position using the indicated substitutions:

Table A. Typical back mutations

Region Amino acid position in the framework region (Kabat numbering) Back mutation 25EZ V _H 72 G72R 25EZ V _H 95 Y95H 25EZ V _H 97 T97A 25EZ V _H 98 T98R _25F7VH 69 L69I _25F7VH 71 L71V _25F7VH 83 R83S _25F7VH 97 F97R _8B7VH 76 K76N _8B7VH 79 A79S _8B7VH 83 N83S 8B7 V _h 112 P112Q _11F2VH 3 D3A _17E5VH 3 H3Q 8B7 V _h 59 C59Y 8B7 V _h 59 C59S

Another type of framework modification involves mutation of one or more residues in the framework region, or even in one or more CDR regions, to remove T cell epitopes to reduce the potential immunogenicity of the antibody. This method is also called deimmunization and is described in more detail in US Patent Application Published No. 20030153043.

In addition to, or as an alternative to, modifications in the framework or CDR regions, antibodies of the invention can be engineered to include modifications in the Fc region, typically to alter one or more functional properties of the antibody, such as serum half-life, fixation complement, Fc receptor binding and/or antigen-dependent cellular cytotoxicity. In addition, the antibody of the invention can be modified by chemical methods (eg, one or more chemical groups can be associated with the antibody), or it can be modified by changing its glycosylation to change one or more functional properties of the antibody.

Each of these embodiments of the invention is described in more detail below. The numbering of residues in the Fc region corresponds to the Kabat EU index.

In a preferred embodiment, the antibody is an IgG4 isotype antibody containing a serine to proline mutation at a position corresponding to position 228 (S228P; EU identifier) on the heavy chain constant region hinge region. This mutation has been reported to abolish the heterogeneity of interchain disulfide bridges in the hinge region (Angal et al. supra; position 241 corresponds to the Kabat numbering system). For example, in various embodiments, an anti-LAG-3 antibody of the invention may comprise an antibody heavy chain variable region 25F7 (SEQ ID NO: 37) or 26H10 (SEQ ID NO: 38) linked to a human IgG4 constant domain in which the serine at position , corresponding to position 241 described by Angal et al., supra, is mutated to proline. Thus, in the variable regions of the heavy chain of antibodies 25F7 and 26H10, associated with the constant region of IgG4, this mutation corresponds to the S228P mutation according to the EU index.

In one embodiment of the invention, the hinge region of the C _H 1 domain is modified such that the number of cysteine residues in the hinge region is changed, for example, increased or decreased. This method is described in more detail in US Pat. No. 5,677,425. The number of cysteine residues in the hinge region of the C _H 1 domain is changed, for example, to promote the assembly of light and heavy chains or to increase or decrease the stability of the antibody.

In another embodiment of the invention, mutations are made in the Fc hinge region of an antibody to increase or decrease the biological half-life of the antibody. More specifically, one or more

- 19 043093 amino acid mutations are introduced at the interface region of the CH2-CH3 domains of the Fc-hinge fragment so that the binding of the antibody to staphylococcal protein A (SpA) is weakened compared to the binding of the native Fc-hinge domain to SpA. This method is described in more detail in US Patent No. 6,165,745.

In another embodiment of the invention, the antibody is modified to increase its biological half-life. Various approaches are possible. For example, one or more of the following mutations can be introduced: T252L, T254S, T256F, as described in US Pat. No. 6,277,375. Or, to increase the biological half-life, the antibody can be modified in the C _H 1 or C _L region so that it contains the binding epitope with a scavenger receptor formed by two C _H 2 loops of the IgG Fc domain, as described in US Patent Nos. 5,869,046 and 6,121,022.

In other embodiments, the Fc region is changed by replacing at least one amino acid residue with another amino acid residue to change the effector function(s). For example, one or more amino acids selected from amino acid residues 234, 235, 236, 237, 297, 318, 320 and 322 can be replaced with another amino acid residue such that the antibody changes affinity for the effector ligand molecule but retains the antigen-binding ability of the original antibody . The effector ligand for which the affinity is changed may be, for example, an Fc receptor or a C1 complement component. This approach is described in more detail in US Patent Nos. 5,624,821 and 5,648,260.

In another example, one or more amino acids selected from amino acid residues 329, 331 and 322 can be replaced by another amino acid residue such that the antibody alters C1q binding and/or reduces or abolishes complement-dependent cytotoxicity (CDC). This approach is described in more detail in US Patent No. 6194551.

In another example, one or more amino acid residues at amino acid positions 231 and 239 are changed, thereby altering the ability of the antibody to fix complement. This method is described in more detail in published PCT International Application WO 94/29351.

In yet another example, the Fc region is modified to enhance the ability of the antibody to mediate antibody-dependent cellular cytotoxicity (ADCC) and/or to increase the affinity of the antibody for the Fcy receptor by modifying one or more amino acids at the following positions: 238, 239, 248,

249, 252, 254, 255, 256, 258, 265, 267, 268, 269, 270, 272, 276, 278, 280, 283, 285,286,

289, 290, 292, 293, 294, 295, 296, 298, 301, 303, 305, 307, 309, 312, 315, 320, 322,324,

326, 327, 329, 330, 331, 333, 334, 335, 337, 338, 340, 360, 373, 376, 378, 382, 388,389,

398, 414, 416, 419, 430, 434, 435, 437, 438 or 439.

This method is described in more detail in the published International Application PCT WO 00/42072. In addition, binding sites for FcyRI, FcyRII, FcyRIII and FcRn have been mapped on human IgG1 and variants with increased binding have been described (see Shields et al. (2001) J. Biol. Chem. 276: 6591-6604). Specific mutations at positions 256, 290, 298, 333, 334, and 339 have been shown to increase binding to FcyRIII. In addition, the following combination mutants have been shown to increase FcyRIII binding: T256A/S298A, S298A/E333A, S298A/K224A, and S298A/E333A/K334A.

In yet another embodiment, the glycosylation of the antibody is modified. For example, a non-glycosylated (aglycosylated) antibody can be produced (ie, there is no glycosylation of the antibody). Glycosylation can be altered, for example, to increase the affinity of an antibody for an antigen. Such carbohydrate modifications can be accomplished, for example, by changing one or more glycosylation sites in the amino acid sequence. For example, one or more amino acid substitutions can be made that eliminate one or more glycosylation sites in the framework region of the variable region, thereby eliminating glycosylation at that site. This aglycosylation (lack of glycosylation) can increase the affinity of the antibody for the antigen. See, for example, US Patent Nos. 5,714,350 and 6,350,861.

In addition, or alternatively, an antibody with an altered glycosylation pattern can be produced, such as a hypofucosylated antibody having a reduced number of fucosyl residues, or an antibody having improved bisector GlcNac structures. These altered glycosylation patterns have been shown to enhance the ADCC ability of antibodies. Such carbohydrate modifications can be accomplished, for example, by expressing the antibody in a host cell with an altered glycosylation mechanism. Cells with an altered glycosylation pattern are described in the prior art and can be used as host cells to express recombinant antibodies of the invention, thereby producing an antibody with an altered glycosylation pattern. For example, the Ms704, Ms705 and Ms709 cell lines lack the fucosyltransferase gene, FUT8 (a(1,6)-fucosyltransferase), and there is no fucose among the antibody carbohydrates expressed in the Ms704, Ms705 and Ms709 cell lines. Ms704, Ms705 and Ms709 FUT8 ^- / ^- cell lines are generated by targeted disruption of the FUT8 gene in CHO/DG44 cells using two replacement

- 20 043093 vector (US Patent Application Published No. 20040110704 and Yamane-Ohnuki et al. (2004) Biotechnol Bioeng 87: 614-22). Another example, European patent EP 1176195 describes a cell line with impaired function of the FUT8 gene, which encodes a fucosyltransferase, therefore hypofucosylation of antibodies expressed in such a cell line is observed due to the reduction or elimination of the enzyme with an α-1,6 linkage. European patent EP 1176195 also describes cell lines with reduced enzymatic activity in the reaction of addition of fucose to N-glucosamine, which binds to the Fc region of the antibody, or no activity, for example, the myeloma cell line YB2/0 (ATCC CRL 1662). Published International Patent Application PCT WO 03/035835 describes a variant of the CHO cell line, Lec13 cells, with a reduced ability to attach fucose to Asn(297)-linked carbohydrates, which also results in hypofucosylation of antibodies expressed in that host cell (see also Shields et al (2002) J Biol Chem 277-2673326740). Antibodies with a modified glycosylation profile can also be produced in chicken eggs, as described in published International Patent Application PCT WO 06/089231. Alternatively, antibodies with a modified glycosylation profile can be produced in plant cells such as Lemna. Methods for producing antibodies in a plant system are disclosed in the US patent application corresponding to Alston & Bird LLP Patent Registration No. 040989/314911, filed August 11, 2006. Published PCT International Patent Application WO 99/54342 describes cell lines engineered to express glycoprotein-modifying glycosyltransferases (e.g., β(1,4)-N-acetylglucosaminyltransferase III (GnTIII )), so that these antibodies expressed in genetically engineered cell lines exhibit enhanced GlcNac structures with a bisector plane, which causes increased ADCC activity of the antibodies (see also Umana et al. (1999) Nat. Biotech. 17: 176- 180). Alternatively, the fucose residues of the antibody can be cleaved using the enzyme glucosidase; for example, fucosidase a-L-fucosidase removes fucosyl residues from antibodies (Tarentino et al. (1975) Biochem. 14: 5516-23).

Another modification of the antibodies described herein discussed in this disclosure is PEGylation. The antibody can be pegylated, for example, to increase the biological (eg, serum) half-life. For PEGylation, the antibody, or fragment thereof, is typically reacted with polyethylene glycol (PEG), such as a reactive ester or aldehyde derivative of PEG, under conditions in which one or more PEG groups are attached to the antibody or antibody fragment. Preferably, PEGylation is accomplished by an acylation or alkylation reaction with a reactive PEG molecule (or similar reactive water-soluble polymer). The term polyethylene glycol as used herein is intended to cover any form of PEG that is used to derivatize other proteins, such as mono(C ₁ -C ₁₀ ) alkoxy or aryloxy polyethylene glycol or polyethylene glycol maleimide. In some embodiments, what is pegylated is a non-glycosylated (aglycosylated) antibody. Methods for PEGylation of proteins are known in the art and can be used to PEGylate antibodies according to the invention. See, for example, European patents EP 0154316 and EP 0401384.

Physical properties of antibodies

Antibodies of this invention can be characterized by various physical properties in order to detect them and/or differentiate between different classes of antibodies.

Antibodies of the present invention may contain one or more glycosylation sites in each light or heavy chain variable region. Such glycosylation sites may cause increased antibody immunogenicity or a change in antibody pK due to altered antigen binding (Marshall et al. (1972) App Rev Biochem 41: 673-702; Gala and Morrison (2004) J Immunol 172: 5489-94; Wallick et al (1988) J Exp Med 168:1099-109; Spiro (2002) Glycobiology 12: 43R-56R; Parekh et al (1985) Nature 316: 452-7; Mimura et al (2000) Mol Immunol 37: 697-706). Glycosylation is known to occur in motifs containing the sequence N-X-S/T. In some cases, it is preferable to have an anti-LAG-3 antibody that is not glycosylated in the variable region. This can be achieved by selecting antibodies that do not contain a glycosylation motif in the variable region, or by mutating residues in the glycosylation region.

In a preferred embodiment, the antibodies of the present disclosure do not contain asparagine isomerism sites. Deamidation of asparagine can occur along the N-G or D-G sequences, leading to the formation of an isoaspartic acid residue, which causes a bend (loop) in the polypeptide chain and reduces its stability (isoaspartic acid effect).

Each antibody has a unique isoelectric point (pI), which typically falls between pH 6 and 9.5. The pI for an IgG1 antibody typically falls in the pH range of 7-9.5, and the pI for an IgG4 antibody typically falls in the pH range of 6-8. It is hypothesized that antibodies with a pI outside the normal pH range may experience some degree of unwinding and instability under in vivo conditions. Thus, it is preferable to have an anti-LAG-3 antibody whose pI value falls within the normal pH range. This can be achieved either by selecting antibodies with a pI value in the normal pH range or by mutating charged surface residues.

- 21 043093

Each antibody has a characteristic melting point, with a higher melting point indicating increased overall stability in vivo (Krishnamurthy R and Manning MC (2002) Curr Pharm Biotechnol 3: 361-71). In general, it is preferable that T _M1 (initial spin-up temperature) be above 60°C, preferably above 65°C, even more preferably above 70°C. The melting point of an antibody can be determined by differential scanning calorimetry (Chen et al. (2003) Pharm Res 20: 1952-60; Ghirlando et al. (1999) Immunol Lett 68: 47-52) or by circular dichroism (Murray et al. ( 2002) J Chromatogr Sci 40: 343-9).

In a preferred embodiment of the invention, antibodies are selected that do not decay rapidly. Antibody degradation can be quantified by capillary electrophoresis (CE) and MALDI-MS (Alexander AJ and Hughes DE (1995) Anal Chem 67. 3626-32).

In another preferred embodiment of the invention, antibodies are selected with minimal aggregation effect, which could lead to the triggering of an undesirable immune response and/or altered or undesirable pharmacokinetic properties. Generally, antibodies with an aggregation of 25% or less, preferably 20% or less, even more preferably 15% or less, even more preferably 10% or less, and even more preferably 5% or less are acceptable. Aggregation is quantified by several methods, including high performance gel chromatography (SEC) columns, high performance liquid chromatography (HPLC) and light scattering.

Antibody engineering methods

As discussed above, anti-LAG-3 antibodies having V _H and V _L sequences as described herein can be used to create new anti-LAG-3 antibodies by modifying the VH and/or VL sequences or their associated constant regions. Thus, in another aspect of the invention, the structural features of the anti-LAG-3 antibody of the invention, for example, 25F7, 26H10, 25E3, 8B7, 11F2 or 17E5, are used to create structurally related anti-LAG-3 antibodies that retain at least one functional property antibodies of the invention, such as binding to human LAG-3. For example, one or more CDR regions of antibodies 25F7, 26H10, 25E3, 8B7, 11F2 or 17E5, or mutants thereof, can be combined by recombinant DNA techniques with known framework regions and/or other CDRs to create additional recombinant DNA-derived anti-LAG antibodies -3 according to the invention discussed above. Other modification types include the modification types described in the previous section. The starting material for the engineering method is one or more V _H and/or V _L sequences as described herein, or one or more CDR domains thereof. To create a genetically engineered antibody, it is not necessary to actually produce (ie, express as a protein) an antibody having one or more VH and/or VL sequences herein or one or more CDR domains thereof. Rather, the information contained in the sequence(s) is used as starting material to create second-generation sequence(s) derived from the original sequence(s), and then the second-generation sequence(s) are produced and expressed as squirrel.

Accordingly, another embodiment of the invention includes a method for producing an anti-LAG-3 antibody, comprising:

(a) provision:

(i) an antibody heavy chain variable region sequence comprising a CDR1 sequence selected from the group consisting of SEQ ID NOs: 1-6, a CDR2 sequence selected from the group consisting of SEQ ID NOs: 7-12, and/or a CDR3 sequence , selected from the group consisting of SEQ ID NO: 13-14, GGY and 16-18; and/or (ii) an antibody light chain variable region sequence comprising a CDR1 sequence selected from the group consisting of SEQ ID NOs: 19-24, a CDR2 sequence selected from the group consisting of SEQ ID NOs: 25-30, and/ or a CDR3 sequence selected from the group consisting of SEQ ID NO: 31-36;

(b) altering at least one amino acid residue in the antibody heavy chain variable region sequence and/or the light chain variable region sequence to create at least one altered antibody sequence; and (c) expressing the altered antibody sequence as a protein.

Standard molecular biology techniques can be used to produce and express an antibody with a modified sequence.

Preferably, the antibody encoded by the altered antibody gene sequence(s) is an antibody that retains one, some, or all of the functional properties of the anti-LAG-3 antibodies herein, which functional properties include, but are not limited to:

(i) high affinity binding to human LAG-3;

(ii) binding to monkey LAG-3;

(iii) lack of binding to mouse LAG-3;

(iv) the ability to inhibit the binding of LAG-3 to MHC class II molecules and

- 22 043093 (v) the ability to stimulate an immune response (for example, an antigen-specific T-cell response).

The functional properties of the modified antibodies can be assessed by standard assay methods known in the art and/or presented herein, such as the methods given in the examples.

In some embodiments of the antibody engineering methods of the invention, mutations can be introduced, randomly or selectively, into the sequence encoding the entire antibody or antibody region, and the resulting modified anti-LAG-3 antibodies can be screened for binding activity and/or other functional properties as described herein. Mutation methods are described in the prior art (see, for example, published PCT International Applications WO 02/092780 and WO 03/074679).

Nucleotide molecules encoding antibodies of the invention

Another aspect of the invention relates to nucleotide molecules that encode antibodies of the invention. Nucleic acids may be present in whole cells, cell lysates, or in partially purified or substantially pure form. A nucleic acid is isolated or reduced to a substantially pure state if it is purified from other cellular components or other impurities, such as other cellular nucleic acids or proteins, by standard methods including alkaline SDS treatment, CsCl centrifugation, column chromatography , agarose gel electrophoresis and other methods well known in the art. See Ausubel, et al., ed. (1987) Current Protocols in Molecular Biology, Greene Publishing and Wiley Interscience, New York. The nucleic acid of the invention may be DNA, RNA, and may or may not contain or contain intronic sequences. In a preferred embodiment of the invention, the nucleic acid is DNA.

The nucleic acids of the invention can be obtained by conventional molecular biology methods. In the case of antibodies expressed by hybridomas (for example, hybridomas obtained from transgenic mice carrying human immunoglobulin genes, described in more detail below), cDNAs encoding the light and heavy chains of the antibody produced by the hybridoma can be obtained by standard PCR amplification or cDNA cloning. obtained from a library of immunoglobulin genes (for example, using phage display methods), the nucleic acid encoding such antibodies can be isolated from the gene library.

Preferred nucleotide molecules of the invention are those nucleotide molecules that encode the VH and _VL sequences of monoclonal antibodies 25E3, 25F7, 8B7, 26H10, 11F2 and 17E5. DNA sequences encoding VH sequences 25E3, 25F7, 8B7, 26H10, 11F2 and 17E5 are shown in SEQ ID NOs: 49-54, respectively. DNA sequences encoding the VL sequences of antibodies 25E3, 25F7, 8B7, 26H10, 11F2 and 17E5 are shown in SEQ ID NO: 55-60, respectively.

Once DNA fragments encoding the VH and VL segments are obtained, these DNA fragments can be further processed by conventional rcombinant DNA techniques, for example, to convert the variable region genes into full-length antibody genes, into Fab fragment genes, or into a scFv gene. In this case, a DNA fragment encoding V _L - or V _H - is operably linked to another DNA fragment encoding another protein, such as an antibody constant region or a flexible linker. The expression operably linked in this context is intended to mean that two DNA fragments are linked in such a way that the amino acid sequences encoded by the two DNA fragments remain in reading frame.

The isolated DNA encoding the VH region can be converted into a full-length heavy chain gene by operably linking the DNA encoding the VH to another DNA molecule encoding the heavy chain constant regions ( _CH1 , CH2, and _CH2 ). The gene sequences for the heavy chain constant regions are known in the art (eg, see above, Kabat et al. (1991)), and DNA fragments encompassing these regions can be obtained by conventional PCR amplification. The heavy chain constant region may be an IgG1, IgG2, IgG3, IgG4, IgA, IgE, IgM, or IgD constant region, but is most preferably an IgG1 or IgG4 constant region. For the heavy chain gene of the Fab fragment, DNA encoding _VH can functionally bind to another DNA molecule encoding only the heavy chain _CH1 constant region.

Isolated DNA encoding the _VL region can be converted into a full-length light chain gene (as well as a Fab light chain gene) by functionally linking the DNA encoding the VL to another DNA molecule encoding the light chain constant region, CL. The sequences of human light chain constant region genes are known in the art (eg, Kabat et al., supra), and DNA fragments covering these regions can be obtained by conventional PCR amplification. In preferred embodiments of the invention, the light chain constant region may be a kappa or lambda constant region.

To create the scFv gene, DNA fragments encoding the VH and VL regions are operably linked to another fragment encoding a flexible linker, for example, encoding an amino acid sequence

- 23 043093 activity (Gly ₄ -Ser) ₃ so that the VH and VL sequences can be expressed as a continuous single-chain protein in which the VL and VH regions are linked by a flexible linker (see, for example, Bird et al. (1988) Science 242 : 423-426; Huston et al. (1988) Proc. Natl. Acad. Sci. USA 85: 5879-5883; McCafferty et al. (1990) Nature 348: 552-554).

Preparation of monoclonal antibodies according to the invention

Monoclonal antibodies (mAbs) of the present disclosure can be produced by the well-known somatic cell hybridization method (hybridoma method) Kohler and Milstein (1975) Nature 256: 495. Other options for producing monoclonal antibodies include transformation of B lymphocytes with viruses or oncogenes and phage display methods. Chimeric or humanized antibodies are also well known in the art. See, for example, US Patent Nos. 4,816,567; 5225539; 5530101; 5585089; 5693762 and 6180370, the contents of which are specifically incorporated into this description by reference in their entirety.

In a preferred embodiment, the antibodies of the invention are human monoclonal antibodies. Such human monoclonal antibodies specific for human LAG-3 can be generated using transgenic or transchromosomal mice carrying components of the human immune system rather than components of the murine system. These transgenic and transchromosomal mice include mice referred to herein as HuMAb Mouse® and KM Mouse®, respectively, and generally referred to herein as human Ig-bearing mice.

HuMAb mice (Medarex®, Inc.) contain miniloci of unarranged human immunoglobulin genes that encode human immunoglobulin μ heavy (μ and γ) and light chain sequences, together with targeted mutations that inactivate endogenous μ and γ chain loci (see, e.g. , Lonberg et al (1994) Nature 368 (6474): 856-859). Accordingly, these mice exhibit reduced expression of murine IgM or anti-chain and, in response to immunization, introduced human heavy and light chain transgenes cause antibody class switching and somatic mutation, generating high-affinity human IgGK monoclonal antibodies (Lonberg et al. (1994) , see above; review in

Lonberg (1994) Handbook of Experimental Pharmacology 113: 49-101; Lonberg, N. and

Huszar, D. (1995) Intern. Rev. Immunol. 13: 65-93, and Harding and Lonberg (1995) Ann. N.Y.

Acad. Sci. 764: 536-546).

The generation and use of HuMAb Mouse® mice and the genomic modifications that these mice carry are described in more detail in

Taylor et al.

(1992) Nucleic Acids Research 20: 6287-6295; Chen et al. (1993) International Immunology

5: 647- 656; Tuaillon et al. (1993) Proc. Natl. Acad. Sci. USA 90: 3720- 3724; Choi et al.

(1993) Nature Genetics 4: 117-123; Chen et al. (1993) EMBO J. 12: 821-830; Tuaillon et al.

(1994) J. Immunol. 152: 2912-2920; Taylor et al. (1994) International Immunology 6: 579-

591; and Fishwild et al. (1996) Nature Biotechnology 14: 845-851, the contents of all these materials are specifically incorporated into this description by reference in their entirety. In addition, see US Pat. Nos. 5,545,806; 5569825; 5625126; 5633425; 5789650; 5877397; 5661016; 5814318; 5874299; 5770429; and 5545807; published PCT International Patent Applications No. WO 92/03918; WO 93/12227; WO 94/25585; WO 97/13852; WO 98/24884; WO 99/45962 and WO 01/14424, the contents of which are incorporated herein by reference in their entirety.

In another embodiment, the formation of human antibodies of the invention can be induced using a mouse that carries human immunoglobulin gene sequences on transgenes and trans chromosomes, such as a mouse that carries a human heavy chain transgene and a trans human light chain chromosome. Such a mouse is referred to herein as the KM mouse® and is described in detail in published International Patent Application PCT WO 02/43478. A modified form of this mouse, which further includes disruption of the endogenous FcyRIIB receptor gene in a homozygous state, is also described in published International Patent Application PCT WO 02/43478 and is referred to herein as the KM/FCGR2D mouse®. In addition, mice with either HCo7 or HCo12 heavy chain transgenes can be used.

Other transgenic animal options include XenoMouse (Abgenix, Inc., US Pat. Nos. 5,939,598; 6,075,181; 6,114,598; 6,150,584 and 6,162,963). Other embodiments include mice (Tomizuka et al. (2000) Proc. Natl. Acad. Sci. USA 97: 722-727) and bovine TCs carrying human heavy and light chain trans chromosomes (Kuroiwa et al. (2002) Nature Biotechnology 20: 889-894; published International PCT Application WO 02/092812). The contents of these patents and the published patent application are expressly incorporated herein by reference in their entirety.

- 24 043093

In one embodiment, human monoclonal antibodies of the invention are produced by phage display methods for screening human immunoglobulin gene libraries. See, for example, US Pat. Nos. 5,223,409; 5403484; 5571698; 5427908; 5580717; 5969108; 6172197; 5885793; 6521404; 6544731; 6555313; 6582915 and 6593081, the contents of which are incorporated into this description by reference in their entirety.

Human monoclonal antibodies of the invention can be produced using SCID mice in which human immune cells have been reconstituted such that human antibodies are produced upon immunization (antibody response). See, for example, US Patent Nos. 5,476,996 and 5,698,767, the contents of which are incorporated herein by reference in their entirety.

In another embodiment of the invention, human antibodies against LAG-3 are obtained by the phage display method, in which the phages contain nucleic acids encoding antibodies produced in the body of transgenic animals previously immunized with the LAG-3 protein. In a preferred embodiment of the invention, the transgenic animal is a HuMab, KM or Kirin mouse. See, for example, US Pat. No. 6,794,132, the contents of which are incorporated herein by reference in their entirety.

Immunization of mice carrying human Ig (Human Ig)

In one embodiment, mice with human Ig are immunized with a purified or enriched preparation of LAG-3 antigen, recombinant LAG-3 protein, or cells expressing LAG-3 protein. For example, Lonberg et al. (1994), see above; Fishwild et al. (1996), see above; published PCT International Applications WO 98/24884 or WO 01/14424, the contents of which are incorporated herein by reference in their entirety. In a preferred embodiment, mice 6-16 weeks of age are immunized with 5-50 μg of LAG-3 protein. Alternatively, a LAG-3 region fused to a non-LAG-3 polypeptide is used.

In one embodiment, transgenic mice are immunized intraperitoneally (IP) or intravenously (IV) with LAG-3 antigen in complete Freund's adjuvant followed by IP or IV immunization in incomplete Freund's adjuvant. Other embodiments use adjuvants other than Freund's adjuvant, or whole cells in the absence of an adjuvant. Plasma can be screened by ELISA, and mouse cells with satisfactory titers of human anti-LAG-3 immunoglobulin can be used for fusion.

Preparation of hybridomas producing human monoclonal antibodies according to the invention

To produce hybridomas producing human monoclonal antibodies of the invention, splenocytes and/or lymph node cells from immunized mice can be isolated and fused with an appropriate immortalized cell line, such as a murine myeloma cell line. The resulting hybridomas can be screened for the production of antigen-specific antibodies. The production of hybridomas is well known in the art. See, for example, Harlow and Lane (1988) Antibodies, A Laboratory Manual, Cold Spring Harbor Publications, New York.

Preparation of transfectomes producing human monoclonal antibodies according to the invention

Antibodies of the invention can also be produced in transfectome host cells using, for example, a combination of recombinant DNA and gene transfection techniques well known in the art (e.g., Morrison, S. (1985) Science 229: 1202). In one embodiment, DNA encoding partial or full-length light and heavy chains obtained by conventional molecular biology techniques is inserted into one or more expression vectors such that the genes are operably linked to transcriptional and translational regulatory sequences. In this context, the expression operably linked is intended to mean that the antibody gene is ligated to the vector such that the transcriptional and translational control sequences in the vector serve their intended function of regulating the transcription and translation of the antibody gene.

The term regulatory sequence is intended to include promoters, enhancers, and other expression control elements (eg, polyadenylation signals) that control transcription or translation of antibody chain genes. Such regulatory sequences are described, for example, in Goeddel (Gene Expression Technology. Methods in Enzymology 185, Academic Press, San Diego, CA (1990)). Preferred regulatory sequences for expression in mammalian host cells include viral elements that control high level protein expression in mammalian cells, such as promoters and/or enhancers from cytomegalovirus (CMV), simian virus 40 (SV40), adenovirus (eg, major late adenovirus (AdMLP) and polyoma virus promoter. Or, non-viral regulatory sequences can be used, such as the ubiquitin promoter or the β-globin promoter. In addition, regulatory elements consisting of sequences from different sources, such as the SRa promoter promoter system, which contains sequences from SV40 early promoter and T virus long terminal repeat

- 25 043093 cell leukemia type 1 (Takebe et al. (1988) Mol. Cell. Biol. 8: 466-472). The expression vector and expression control sequences are selected so that they are compatible with the host cell used for expression.

The antibody light chain gene and the antibody heavy chain gene can be inserted into the same expression vector or into different expression vectors. In preferred embodiments of the invention, variable regions are used to create full-length antibody genes of any isotype by inserting them into expression vectors already encoding the heavy and light chain constant regions of a given isotype, so that the VH segment is operably linked to the CH segment(s) in the vector, and V _L the segment was operably linked to the CL segment in the vector. Additionally or alternatively, the recombinant gene expression vector may encode a signal peptide that promotes secretion of the antibody chain by the host cell. The antibody chain gene can be cloned into a vector such that the signal peptide binds in frame to the amino terminus of the antibody chain. The signal peptide may be an immunoglobulin signal peptide or a heterologous signal peptide (eg, a signal peptide from a non-immunoglobulin protein).

In addition to antibody chain genes and regulatory sequences, the recombinant gene expression vector of the invention may carry additional sequences, such as sequences that regulate replication of the vector in host cells (eg, origins of replication) and selectable marker genes. The selectable marker gene assists in the selection of host cells into which the vector is introduced (see, for example, US Pat. Nos. 4,399,216; 4,634,665 and 5,179,017). For example, typically a selectable marker gene confers resistance to drugs such as G418, hygromycin, or methotrexate to the host cell into which the vector is introduced. Preferred selection marker genes include the dihydrofolate reductase (DHFR) gene (for use in selection/amplification in dhfr host cells in the presence of methotrexate) and the peo gene (for G418 selection).

For expression of the light and heavy chains, the expression vector(s) encoding the heavy and light chains are transfected into the host cell by standard methods. The various forms of the term transfection are intended to cover a wide range of techniques commonly used to introduce exogenous DNA into a prokaryotic or eukaryotic host cell, such as electroporation, calcium phosphate precipitation, DEAE-dextran polymer transfection, and the like. Although it is theoretically possible to express antibodies of the invention in both prokaryotic and eukaryotic host cells, expression of antibodies in eukaryotic cells, and most preferably mammalian host cells, is most preferred because such eukaryotic cells, and in particular, mammalian cells, rather than prokaryotic cells, assemble and secrete an appropriately folded (having a suitable spatial structure) and immunologically active antibody.

Preferred mammalian host cells include Chinese hamster ovary cells (CHO cells) (including dhfr ^- CHO cells described in Urlaub and Chasin, (1980) Proc. Natl. Acad. Sci. USA 77: 4216-4220, used with a DHFR selectable marker eg, as described in RJ Kaufman and PA Sharp (1982) J Mol Biol 159: 601-621), NSO myeloma cells, COS cells and SP2 cells. In particular, another preferred expression system for use with NSO myeloma cells is the GS gene expression system disclosed in International Patent Applications WO 87/04462, WO 89/01036 and European Patent EP 338841. If recombinant expression vectors encoding antibody genes , are introduced into mammalian host cells, the antibodies are produced by culturing the host cells for a period of time sufficient for expression of the antibody in the host cells or, more preferably, for secretion of the antibody into the culture medium in which the host cells are grown. Antibodies can be isolated from the culture medium by conventional protein purification methods.

Characteristics of antibody binding to antigen

Antibodies of the invention can be tested for binding to human LAG-3, for example, by standard ELISA methods. Human anti-LAG-3 IgG can be further tested for reactivity to the LAG-3 antigen by Western blotting. The binding specificity of an antibody of the invention can also be determined by monitoring the binding of the antibody to cells expressing the LAG-3 protein, for example, by flow cytometry methods. These methods are known in the art. See, for example, Harlow and Lane (1988), supra.

Immunoconjugates

Antibodies of the present invention can be conjugated with a therapeutic agent to form an immunoconjugate, such as an antibody-drug conjugate (ADC). Suitable therapeutic agents include antimetabolites, alkylating agents, DNA minor groove binders, DNA intercalators, DNA cross-linkers, histone deacetylase inhibitors, nuclear export inhibitors, proteasome inhibitors, topoisomerase I or II inhibitors, heat shock protein inhibitors, tyrosine kinase inhibitors, antibiotics, and antimitotic agents. In an ADC, the antibody and therapeutic agent are preferably conjugated via a cleavable linker, such as a peptidyl, disulfide, or hydrazone linker. More preferably, the linker is a peptidyl

- 26 043093 nal linker, such as

Vai-Cit, Ala-Vai, Val-Ala-Val, Lys-Lys, Pro-Val-GlyVal-Val (SEQ ID NO: 15), Ala-Asn-Val, Val-Leu-Lys, Ala-Ala-Asn , Cit-Cit, Vai-Lys, Lys,

Cit, Ser or Glu.

ADC conjugates can be prepared as described in US Pat. Nos. 7,087,600; 6989452 and 7129261; published International PCT Applications WO 07/038658; WO 07/051081; WO 07/059404; WO 08/083312 and WO 08/103693; Published US Patent Applications 20060024317; 20060004081 and 20060247295; the disclosure of which is incorporated herein by reference.

Bispecific molecules

In another aspect, the present disclosure includes bispecific molecules comprising an anti-LAG-3 antibody linked to at least one different functional molecule, e.g., another peptide or protein (e.g., another antibody or ligand for a receptor), forming a bispecific molecule that binds with at least two different binding sites or target molecules. Thus, a bispecific molecule as used herein includes molecules that have three or more specificities. In a preferred embodiment, the bispecific molecule comprises a first binding specificity for LAG-3 and a second binding specificity for a molecule having trigger properties that recruits cytotoxic effector cells capable of killing a LAG-3 expressing target cell. Examples of suitable trigger molecules are CD64, CD89, CD16 and CD3. See, for example, Kufer et al., TRENDS in Biotechnology, 22 (5), 238-244 (2004).

In one embodiment of the invention, the bispecific molecule contains a third specificity in addition to the anti-Fc binding specificity and the anti-LAG-3 binding specificity. A third specificity may be specificity for an antienhancing factor (EF), for example, a molecule that binds to a surface protein that is involved in cytotoxic activity and thereby enhances the immune response against the target cell. For example, an antienhancement factor can bind a cytotoxic T cell (eg, via CD2, CD3, CD8, CD28, CD4, CD40, or ICAM1) or other immune cell, resulting in an enhanced immune response against the target cell.

Bispecific molecules can come in a variety of formats and sizes. At one extreme are bispecific molecules that retain the normal antibody format except that instead of having two binding arms with identical specificity, they have two binding arms with different specificities. At the other extreme are bispecific molecules consisting of two single-chain antibody fragments (scFv) linked by a peptide chain, the so-called Bs(scFv) ₂ construct. Intermediate-sized bispecific molecules comprise two distinct F(ab) moieties linked by a peptidyl linker. Bispecific molecules of these and other formats can be obtained by genetic engineering (recombinant DNA methods), somatic hybridization or chemical methods. See, for example, Kufer et al., supra; Cao and Suresh, Bioconjugate Chemistry, 9 (6), 635-644 (1998); and van Spriel et al., Immunology Today, 21 (8), 391-397 (2000), and references therein.

Pharmaceutical compositions

In another aspect, the present invention includes a pharmaceutical composition comprising an antibody of the present disclosure formulated in conjunction with a pharmaceutically acceptable carrier. It may contain one or more pharmaceutically active ingredients, such as another antibody or drug. The pharmaceutical compositions of the invention can also be used in combination therapy, for example, with another immunostimulating agent, an anticancer agent and an antiviral agent or vaccine, such that the anti-LAG-3 antibody enhances the immune response against the vaccine.

The pharmaceutical composition may contain a number of excipients. Excipients that may be used include carriers, surfactants, thickeners or emulsifiers, solid binders, dispersing or suspending agents, solubilizers, colorants, flavoring agents, coatings, disintegrants, lubricants, sweeteners, preservatives, isotonic agents, and combinations thereof. . The selection and use of suitable excipients are described in Gennaro, ed., Remington: The Science and Practice of Pharmacy, ^20th Ed. (Lippincott Williams & Wilkins 2003), the disclosure of which is incorporated by reference herein.

Preferably, the pharmaceutical composition is suitable for intravenous, intramuscular, subcutaneous, parenteral, spinal or epidermal administration (eg, by injection or infusion). Depending on the route of administration, the active compound may be coated with a material to protect it from acid and other natural conditions that may inactivate it. The expression parenteral administration as used herein means methods of administration other than enteral (intestinal) and local administration, usually by injection, and includes, but is not limited to, intravenous, intramuscular, intraarterial, intrathecal

- 27 043093 (subsurrent), intracapsulal, intraorbital (intra -head), intraudermal, intraperitoneal (intra -abdominal), trans "intrastricheal, subcuticular, intra -articular, subpapsulus, subarachnoidal, intraspinal and submarine (vone) injection and cobblestones . Alternatively, the antibody of the invention may be administered by a non-parenteral route, such as a topical, epidermal or transmucosal route, such as intranasally, orally, vaginally, rectally, sublingually or topically.

The pharmaceutical compounds of the invention may be in the form of pharmaceutically acceptable salts. The expression pharmaceutically acceptable salt refers to a salt that retains the desired biological activity of the parent compound and does not cause undesirable toxicological effects. Examples of such salts include acid addition salts and base addition salts. Acid addition salts include salts formed from non-toxic inorganic acids such as hydrochloric, nitric, phosphoric, sulfuric, hydrobromic, hydroiodic, phosphorous, etc., as well as from non-toxic organic acids such as aliphatic mono- and dicarboxylic acids, phenyl-substituted alkanoic acids, hydroxyalkanoic acids, aromatic acids, aliphatic and aromatic sulfonic acids, etc. Base addition salts include salts formed from alkali and alkaline earth metals such as sodium, potassium, magnesium, calcium, etc., as well as from non-toxic organic amines such as N,Ndibenzylethylenediamine, N-methylglucamine, chloroprocaine, choline, diethanolamine , ethylenediamine, procaine, etc.

Pharmaceutical compositions may be in the form of sterile aqueous solutions or dispersions. They can also be prepared in the form of microemulsions, liposomes or other ordered structures suitable for high drug concentrations.

The amount of active ingredient that can be mixed with the carrier material to form a unit dosage form varies depending on the subject being treated and the particular route of administration and is generally the amount of the composition that produces a therapeutic effect. Typically, of 100%, the amount is from about 0.01% to about 99% of the active ingredient, preferably from about 0.1-70%, most preferably from about 1-30% of the active ingredient in combination with a pharmaceutically acceptable carrier.

Drug dosage regimens are adjusted to produce the optimal target response (eg, therapeutic response). For example, a single bolus may be administered, several small doses may be administered at specified intervals, or the dose may be proportionally decreased or increased depending on the therapeutic situation. It is particularly preferable to formulate compositions for parenteral administration in unit dosage form for ease of administration and uniformity of dosage. The expression single (single) dosage dose as used herein refers to physically discrete units applicable as a unit dosage form for subjects to be treated; each unit contains a predetermined amount of active compound calculated to achieve a predetermined therapeutic effect, in combination with the desired pharmaceutical carrier. Alternatively, the antibody can be administered as a sustained release formulation, in which case less frequent administration is required.

The administered doses of the antibody are approximately 0.0001-100 mg/kg, usually 0.0-5 mg/kg of host body weight. For example, doses may be 0.3 mg/kg body weight, 1 mg/kg body weight, 3 mg/kg body weight, 5 mg/kg body weight, or 10 mg/kg body weight, or range from 1-10 mg/kg. A typical treatment regimen includes administration once a week, once every two weeks, once every three weeks, once every four weeks, once a month, once every 3 months, or once every 3-6 months. Preferred dosage regimens for the anti-LAG-3 antibody of the invention include 1 mg/kg body weight or 3 mg/kg body weight intravenously, wherein the antibody is given using one of the following dosing regimens:

(i) six doses every four weeks, then every three months;

(ii) every three weeks;

(iii) 3 mg/kg body weight once, then 1 mg/kg body weight every three weeks.

In some methods, the dose is adjusted to achieve a plasma antibody concentration of about 1-1000 μg/ml, and in some methods about 25-300 μg/ml.

A therapeutically effective dose of an anti-LAG-3 antibody of the invention preferably causes a reduction in the severity of disease symptoms, an increase in the frequency and duration of asymptomatic periods of the disease, or prevention of deterioration or disability due to the disease. For example, when treating tumor carriers, a therapeutically effective dose preferably inhibits tumor growth by at least about 20%, more preferably at least about 40%, even more preferably by at least about 60%, and even more preferably, at least about 80% greater than in untreated subjects. A therapeutically effective amount of a therapeutic compound can reduce the size of a tumor or otherwise reduce the intensity of symptoms in a subject, which typically is a human or may be another mammal.

- 28 043093

The pharmaceutical composition may be a controlled release formulation, including implants, transdermal patches, and microencapsulated product delivery systems. Biodegradable, biocompatible polymers such as ethylene vinyl acetate, polyanhydrides, polyglycolic acid, collagen, polyorthoesters and polylactic acid can be used. See, for example, Sustained and Controlled Release Drug Delivery Systems, J.R. Robinson, ed., Marcel Dekker, Inc., New York, 1978.

Therapeutic compositions can be administered using medical devices such as:

(1) needle-free hypodermic injectors (see, for example, US patents 5399163; 5383851; 5312335; 5064413; 4941880; 4790824 and 4596556);

(2) microinfusion pumps (US patent 4487603);

(3) transdermal delivery devices (US Patent 4,486,194);

(4) infusion devices (US patents 4447233 and 4447224) and (5) osmotic devices (US patents 4439196 and 4475196);

the disclosure of these patents is incorporated herein by reference.

In some embodiments, human monoclonal antibodies of the invention can be formulated to ensure desired in vivo distribution. For example, to ensure that the therapeutic compound of the invention penetrates the blood-brain barrier, they can be formulated in liposomes, which may further contain targeting particles to promote selective transport to specific cells or organs. See, for example, US patents 4522811; 5374548; 5416016 and 5399331; V.V. Ranade (1989) J. Clin. Pharmacol. 29: 685; Umezawa et al. (1988) Biochem. Biophys. Res. Commun. 153: 1038; Bloeman et al. (1995) FEBS Lett. 357: 140; M. Owais et al. (1995) Antimicrob. Agents Chemother. 39: 180; Briscoe et al. (1995) Am. J. Physiol. 1233: 134; Schreier et al. (1994) J Biol. Chem. 269: 9090; Keinanen and Laukkanen (1994) FEBS Lett. 346: 123 and Killion and Fidler (1994) Immunomethods 4: 273.

Application and methods of the invention

The antibodies, antibody compositions and methods of the present invention have various in vitro and in vivo usefulness, including, for example, detection of LAG-3 or enhancement of the immune response by blockade of LAG-3. In a preferred embodiment, the antibodies of the present invention are human antibodies. For example, these molecules can be administered to cells in culture in vitro or ex vivo, or to humans, for example, in vivo, to enhance immunity in various situations. Accordingly, in one aspect, the invention includes a method of modifying the immune response in a subject by administering to the subject an antibody, or antigen binding site thereof, of the invention such that the immune response in the subject is modified. Preferably, the response is enhanced, stimulated or upregulated.

Preferred subjects include sick people in need of an enhanced immune response. The methods are particularly useful for treating human patients with a disorder that can be treated by enhancing (increasing) the immune response (eg, T-cell mediated immune response). In a particular embodiment of the invention, the methods are particularly useful for treating cancer in vivo. To achieve antigen-specific immune enhancement, anti-LAG-3 antibodies may be administered concomitantly with the target antigen, or the antigen may already be present in the subject being treated (eg, a tumor carrier or a viral carrier). When anti-LAG-3 antibodies are administered together with another agent, the two agents can be administered in any order or simultaneously.

In addition, the invention includes methods for detecting human LAG-3 antigen in a sample, or quantifying human LAG-3 antigen, which consists of contacting the sample, and a control sample, with a human monoclonal antibody, or antigen-binding site thereof, which specifically binds to human LAG-3 under conditions that promote the formation of a complex between the antibody or its antigen-binding site and human LAG-3. Complex formation is then detected, with the difference in complex formation between the sample compared to the control sample indicating the presence of human LAG-3 antigen in the sample. Moreover, the anti-LAG-3 antibodies of the invention can be used to purify human LAG-3 using immunoaffinity chromatography.

In view of the ability of the anti-LAG-3 antibodies of the invention to inhibit the binding of LAG-3 to MHC class II molecules and stimulate antigen-specific T cell responses, the invention also includes in vitro and in vivo methods of using the antibodies of the invention to stimulate, enhance or upregulation of antigen-specific T-cell reactions (responses). For example, the invention includes a method of stimulating an antigen-specific T-cell response by contacting said T-cell with an antibody of the invention such that an antigen-specific T-cell response is stimulated. To quantify the antigen-specific T-cell response, any suitable indicator of the antigen-specific T-cell response can be used. Non-limiting examples of suitable indicators include increased T cell proliferation in the presence of an antibody and/or increased cytokine production in the presence of an antibody. In a preferred embodiment of the invention, antigen-specific T

- 29 043093 cellular response stimulates the production of interleukin-2.

The invention also includes a method of stimulating an immune response (eg, an antigen-specific T-cell response) in a subject by administering an antibody of the invention to the subject such that an immune response (eg, an antigen-specific T-cell response) is stimulated in the subject. In a preferred embodiment of the invention, the subject is a tumor carrier and an immune response against the tumor is stimulated. In another preferred embodiment of the invention, the subject is a carrier of the virus and an immune response against the virus is stimulated.

In another aspect, the invention includes a method of inhibiting the growth of tumor cells in a subject, comprising administering to the subject an antibody of the invention such that tumor growth in the subject is inhibited. In yet another aspect, the invention includes a method of treating a viral infection in a subject, comprising administering to the subject an antibody of the invention such that the infection in the subject is treated.

These and other methods of the invention are discussed in more detail below.

Cancer

Blockade of LAG-3 with antibodies can enhance the immune response to cancer cells in a patient. In one aspect, the present invention relates to treating a subject in vivo with an anti-LAG-3 antibody so that the growth of cancerous tumors is inhibited. An anti-LAG-3 antibody can be used alone to inhibit the growth of cancerous tumors. Alternatively, the anti-LAG-3 antibody may be used in combination with other immunogenic agents, conventional anticancer drugs, or other antibodies, as described below.

Accordingly, in one embodiment, the invention includes a method of inhibiting the growth of cancer cells in a subject, comprising administering to the subject a therapeutically effective amount of an anti-LAG-3 antibody, or antigen-binding site thereof. Preferably, the antibody is a human anti-LAG-3 antibody (such as any of the human anti-human LAG-3 antibodies described herein). Additionally, or alternatively, the antibody may be a chimeric or humanized anti-LAG-3 antibody.

Preferred cancers whose growth can be inhibited using the antibodies of the invention include cancers typically responsive to immunotherapy. Non-limiting examples of cancers preferred for treatment include melanoma (eg, metastatic malignant melanoma), kidney cancer (eg, clear cell carcinoma), prostate cancer (eg, hormone-resistant prostate adenocarcinoma), breast cancer, colon cancer, and lung cancer (for example, non-small cell lung cancer). In addition, the invention includes persistent (resistant or recurrent malignant tumors, the growth of which can be inhibited by the action of antibodies according to the invention.

Examples of other cancers that can be treated with the methods of the invention include bone cancer, pancreatic cancer, head and neck cancer, cutaneous or intraocular malignant melanoma, uterine cancer, ovarian cancer, rectal cancer, anal cancer, stomach cancer, testicular cancer, fallopian tube carcinoma, endometrial carcinoma, cervical carcinoma, vaginal carcinoma, carcinoma of the external genitalia, Hodgkin's disease, non-Hodgkin's lymphoma, esophageal cancer, small intestinal cancer, endocrine system cancer, thyroid cancer, parathyroid cancer, adrenal cancer, soft tissue sarcoma , urethral cancer, penile cancer, chronic or acute leukemia, including acute myeloid leukemia, chronic myeloid leukemia, acute lymphoblastic leukemia, chronic lymphoblastic leukemia, solid tumors in children, lymphocytic lymphoma, bladder cancer, kidney or ureter cancer, renal pelvic carcinoma, central nervous system (CNS) tumor, primary CNS lymphoma, tumor angiogenesis, spinal tumor, brainstem glioma, pituitary adenoma, Kaposi's sarcoma, epidermoid cancer, squamous cell carcinoma. T-cell lymphoma, cancers caused by environmental causes, including cancers caused by asbestosis, and combinations of these cancers. The present invention is also applicable to the treatment of metastatic cancers, in particular metastatic cancers expressing PD-L1 (Iwai et al. (2005) Int. Immunol. 17: 133-144).

Optionally, antibodies to LAG-3 can be combined with an immunogenic agent such as cancer cells, purified tumor antigens (including recombinant protein, peptide and carbohydrate molecules), cells and cells transfected with genes encoding immunostimulatory cytokines (He et al. (2004) J Immunol 173: 4919-28). Non-limiting examples of tumor vaccines that can be used include peptides of melanoma antigens, such as peptides of gp100, MAGE antigens, Trp-2, MARTI and/or tyrosinase, or tumor cells transfected to express the cytokine GM-CSF (discussed in detail below).

In humans, some tumors have been shown to be immunogenic, such as melanomas. By increasing the threshold for T cell activation using LAG-3 blockade, the host response (response) to the tumor can be activated.

LAG-3 blockade appears to be more effective when combined with a vaccination protocol. Many experimental methods (strategies) of vaccination against tumors have been developed

- 30 043093 (see Rosenberg, S., 2000, Development of

Cancer Vaccines, ASCO Educational Book Spring: 60- 62; Logothetis, C., 2000, ASCO

Educational Book Spring: 300-302; Khayat, D. 2000, ASCO Educational Book Spring: 414428; Foon, K. 2000, ASCO Educational Book Spring: 730- 738; see also Restifo, N. and Sznol, M., Cancer Vaccines, Ch. 61, pp. 3023-3043 in DeVita et al. (eds.), 1997, Cancer: Principles and Practice of Oncology, Fifth Edition).

In one of these methods, a vaccine is prepared using autologous or allogeneic tumor cells. These cell-based vaccines have been shown to be most effective when the tumor cells are transfected to express GM-CSF. GM-CSF has been found to be a potent activator of antigen presentation for tumor vaccination (Dranoff et al. (1993) Proc. Natl. Acad. Sci U.S.A. 90: 3539-43).

The study of gene expression and examples of large-scale gene expression in various tumors has led to the definition of so-called tumor-specific antigens (Rosenberg, SA (1999) Immunity 10: 281-7). In many cases, these tumor-specific antigens are differentiation antigens expressed in tumors and in the cell from which the tumor is formed, for example, melanocyte antigens gp100, MAGE antigens and Tip-2. Moreover, many of these antigens can be shown to be targets of tumor-specific T cells found in the host. LAG-3 blockade can be used in combination with a panel of recombinant proteins and/or peptides expressed in the tumor to generate an immune response to these proteins. These proteins are normally considered by the immune system as self-antigens and, therefore, tolerance is maintained to them. The tumor antigen may include telomerase protein, which is required for the synthesis of chromosomal telomeres and which is expressed in more than 85% of human cancers and only in a limited number of somatic tissues (Kim et al. (1994) Science 266: 2011-2013). (These somatic tissues can be protected from immune attack in a variety of ways.) The tumor antigen may also be a neoantigen expressed in cancer cells due to somatic mutations that alter the protein sequence, or create fusion proteins between two unrelated sequences (ie, bcr-abl in the Philadelphia chromosome), or an idiotype from B - cell tumors.

Other tumor vaccines may include proteins from viruses intrinsically associated with human cancers, such as human papillomaviruses (HPV), hepatitis viruses (HBV and HCV), and Kaposi's sarcoma-associated herpes virus (KASV). KHSV). Another type of tumor-specific antigen that can be used in combination with LAG-3 blockade is purified heat shock proteins (HSPs) isolated from the tumor tissue itself. These heat shock proteins contain protein fragments from tumor cells and these HSPs are highly effective when delivered to antigen presenting cells to elicit immunity to tumors (Suot & Srivastava (1995) Science 269: 1585-1588; Tamura et al. (1997) Science 278: 117 -120).

Dendritic cells (DCs) are potent antigen-presenting cells that can be used to elicit an antigen-specific response. DCs can be generated ex vivo and loaded with various protein and peptide antigens, as well as tumor cell extracts (Nestle et al. (1998) Nature Medicine 4: 328-332). DCs can also be genetically transfected to express these tumor antigens. For immunization, DCs were also directly fused with tumor cells (Kugler et al. (2000) Nature Medicine 6: 332-336). DC immunization in combination with LAG-3 blockade can be used as an effective vaccination method to activate a more powerful antitumor response.

LAG-3 blockade can also be combined with conventional anticancer therapeutic agents. LAG-3 blockade can be effectively combined with chemotherapy. In these cases, the dose of the chemotherapy agent administered may be reduced (Mokyr et al. (1998) Cancer Research 58: 53015304). An example of such a combination is the use of an anti-LAG-3 antibody in combination with dacarbazine for the treatment of melanoma. Another example of such a combination is the combination of an anti-LAG-3 antibody with interleukin-2 (IL-2) for the treatment of melanoma. The scientific rationale for the combined use of LAG-3 blockade and chemotherapy is that cell death, which is a consequence of the cytotoxic effects of most chemotherapeutic compounds, would cause increased levels of tumor antigen in the antigen presentation pathway. Other therapies that, when combined with LAG-3 blockade, may have a synergistic effect via cell death include radiation, surgery, and hormonal deprivation. Each of these protocols creates a source of tumor antigen in the host. Angiogenesis inhibitors can also be combined with LAG-3 blockade. Inhibition of angiogenesis leads to the death of tumor cells, which can ensure the entry (feeding) of tumor antigen into the host antigen presentation pathway.

Antibodies that block LAG-3 can also be used in combination with bispecific anti

- 31 043093 bodies that target effector cells expressing the Fca or Fcy receptor to tumor cells (see, for example, US patent Nos. 5922845 and 5837243). Bispecific antibodies can be used to target two separate antigens. For example, anti-Fc receptor/anti-tumor antigen bispecific antibodies (eg, Her-2/neu) can be used to target macrophages to tumor sites. This targeting may more effectively activate tumor-specific responses. The T cell component (arm) of these responses is enhanced by LAG-3 blockade. Alternatively, the antigen can be delivered directly to DCs using bispecific antibodies that bind to the tumor antigen and a specific dendritic cell cell surface marker.

Tumors evade immune control using a variety of mechanisms. Many of these mechanisms can be overcome by inactivating proteins that are expressed by tumors and that have immunosuppressive activity. These proteins include, among others, TGF-β (Kehrl et al. (1986) J. Exp. Med. 163: 1037-1050), IL-10 (Howard & O'Garra (1992) Immunology Today 13: 198-200 ), and Fas ligand (Hahne et al. (1996) Science 274: 1363-1365). Antibodies to these structures can be used in combination with anti-LAG-3 antibody to counteract the effects of the immunosuppressant and promote the host immune response to the tumor.

Other antibodies that activate the host immune response can be used in combination with the anti-LAG-3 antibody. These antibodies include molecules on the surface of dendritic cells that activate DC function and antigen presentation. Antibodies against CD40 can effectively replace helper T cell activity (Ridge et al. (1998) Nature 393: 474-478) and can be used in combination with antibodies against LAG-3 (Ito et al. (2000) Immunobiology 201 (5) 527-40). Activating antibodies to T cell costimulatory molecules such as CTLA-4 (eg, US Pat. No. 5811097), OX-40 (Weinberg et al. (2000) Immunol 164: 2160-2169), 4-1BB (Melero et al. ( 1997) Nature Medicine 3: 682-685 (1997) and ICOS (Hutloff et al. (1999) Nature 397: 262-266), may also provide increased levels of T cell activation. Bone marrow transplantation is currently used to treat various tumors of hematopoietic origin. Although the consequence of this treatment is homologous disease (graft-versus-host), the graft response to the tumor may provide therapeutic benefit. LAG-3 blockade can be used to improve the effectiveness of transplanted donor tumor-specific T cells.

There are also some experimental treatment protocols that involve ex vivo activation and expansion of antigen-specific T cells and adoptive transfer of these cells to recipients to stimulate antigen-specific T cells against the tumor (Greenberg & Riddell (1999) Science 285: 546-51). These methods can also be used to activate T cell responses to infectious agents such as CMV. Ex vivo activation in the presence of anti-LAG-3 antibodies can increase the frequency and activity of adoptively transferred T cells.

Infectious diseases

Other methods of the invention are used to treat patients exposed to specific toxins or pathogens. Accordingly, another aspect of the invention includes a method of treating an infectious disease in a subject, comprising administering to the subject an anti-LAG-3 antibody, or an antigen-binding site thereof, such that the infectious disease is treated. Preferably, the antibody is a human anti-human LAG-3 antibody (such as any of the human anti-LAG-3 antibodies described herein). Additionally or alternatively, the antibody may be a chimeric or humanized antibody.

Similar to the use discussed above for the treatment of tumors, antibody-mediated blockade of LAG-3 can be used individually (alone), or as an adjuvant in combination with vaccines to stimulate an immune response to pathogens, toxins and autoantigens. Examples of pathogens for which this therapeutic modality may be particularly useful include pathogens for which there is currently no effective vaccine or pathogens for which conventional vaccines are not sufficiently effective. These pathogens include, but are not limited to, HIV, hepatitis (A, B, & C), influenza, herpes, giardiasis, malaria, leishmaniasis, Staphylococcus aureus, Pseudomonas aeruginosa. LAG-3 blockade is particularly useful against resistant infections with agents such as HIV that present changing antigens during infection. These new epitopes are recognized as foreign during administration of anti-human LAG-3 antibody, thereby eliciting a strong T cell response that is not dampened by negative signals from LAG-3.

Some examples of pathogenic viruses that cause infections susceptible to treatment with the methods of the invention include HIV, hepatitis (A, B or C), herpes virus (for example, VZV, HSV-1, HAV-6, HSV-II and CMV, Epstein-Barr), adenovirus, influenza virus, flaviviruses, echovirus, rhinovirus, coxsackievirus, coronavirus, respiratory syncytial virus, mumps virus, rotavirus, measles virus, measles rubella virus, parvovirus, vaccinia virus, HTLV virus, dengue virus , virus

- 32 043093 papillomas, molluscum contagiosum virus, polyoma virus, rabies virus, JC virus and arboviral encephalitis virus.

Some examples of pathogenic bacteria causing infections susceptible to treatment by the methods of the present invention include chlamydia, rickettsial microorganisms, mycobacteria, staphylococci, streptococci, pneumococci, meningococci and gonococci, Klebsiella, Proteus, serratia, pseudomonas, legionella, diphtheria bacteria, salmonella, rod-shaped bacteria, the bacteria that cause cholera (Vibrio cholerae), tetanus, botulism, anthrax, plague, leptospirosis and Lyme disease.

Some examples of pathogenic fungi that cause infections susceptible to treatment with the methods of the invention include Candida (C. albicans, C. krusei, C. glabrata, C. tropicalis, etc.), Cryptococcus neoformans, Aspergillus (A. fumigatus . .

Some examples of pathogenic parasites causing infections susceptible to treatment with the methods of the invention include Entamoeba histolytica, Balantidium coli, Naegleriafowleri, Acanthamoeba sp., Giardia lamblia, Cryptosporidium sp. .), causative agent of pneumocystis Pneumocystis carinri, causative agent of malaria Plasmodium vivax, babesia Babesia microti, trypanosome species Trypanosoma brucei, species Trypanosoma cmzi, leishmania species Leishmania donovani, toxoplasma (Toxoplasma gondii), nematode Nippostrongylus brasiliensis.

In all of the above methods, LAG-3 blockade can be combined with other forms of immunotherapy, such as cytokine treatment (eg, interferons, GM-CSF, G-CSF, IL-2), or bispecific antibody therapy, which provides increased presentation of tumor antigens (see ., for example, Holliger (1993) Proc. Natl. Acad. Sci. USA 90: 6444-6448; Poljak (1994) Structure 2: 1121-1123).

Autoimmune reactions

Antibodies against LAG-3 can cause and enhance autoimmune reactions. Indeed, the induction of antitumor responses using tumor cell and peptide vaccines shows that many antitumor responses involve antiautoreactivity (reaction against self-antigens) (van Elsas et al. (2001) J. Exp. Med. 194: 481-489; Overwijk, et al. (1999) Proc. Natl. Acad. Sci. U.S.A. 96: 2982-2987; Hurwitz, (2000), supra; Rosenberg & White (1996) J. Immunother Emphasis Tumor Immunol 29 (1): 81-4) . Therefore, the use of anti-LAG-3 antibody blockade in combination with various self-proteins can be considered to develop vaccination protocols to effectively generate immune responses against these self-proteins to treat the disease. For example, Alzheimer's disease involves inappropriate accumulation of Av peptide in amyloid deposits in the brain; The antibody response to myeloid is able to clear these amyloid deposits (Schenk et al. (1999) Nature 400: 173-177).

Other self proteins can also be used as targets, such as IgE to treat allergies and asthma, and TNFa to treat rheumatoid arthritis. Finally, antibody responses to various hormones can be induced using anti-LAG-3 antibody. The neutralizing antibody response to reproductive hormones can be used for contraception. Neutralizing antibody responses to hormones and other soluble factors required for the growth of specific tumors can also be considered as possible targets for vaccination.

Methods similar to the anti-LAG-3 antibody methods described above can be used to induce a therapeutic autoimmune response to treat patients with inappropriate accumulation of other autoantigens, such as amyloid deposits, including Alzheimer's disease Ae-peptide, cytokines such as TNFa, and IgE .

Vaccines

Antibodies against LAG-3 can be used to stimulate antigen-specific immune responses using co-administration of an anti-LAG-3 antibody with a target antigen (eg, a vaccine). Accordingly, in another aspect, the invention includes a method of enhancing an immune response to an antigen in a subject, comprising administering to the subject:

(i) an antigen and (ii) an antibody against LAG-3, or an antigen-binding site thereof, such that the immune response to the antigen is enhanced.

Preferably, the antibody is a human anti-human LAG-3 antibody (such as any of the human anti-LAG-3 antibodies described herein). Additionally or alternatively, the antibody may be a chimeric or humanized antibody. The antigen may be, for example, a tumor antigen, a viral antigen, a bacterial antigen, or an antigen from a pathogen. Non-limiting examples of such antigens include antigens discussed

- 33 043093 given in the previous section, such as the tumor antigens (or tumor vaccines) discussed above, or antigens from viruses, bacteria or other pathogens described above.

Appropriate routes of administration of the antibody compositions (eg, human monoclonal antibodies, multispecific (polyspecific) and bispecific molecules and immunoconjugates) of the invention in vivo and in vitro are well known in the art and can be selected by one of ordinary skill in the art. For example, antibody compositions can be administered by injection (eg, intravenous or subcutaneous). The appropriate dosages of molecules used depend on the age and body weight of the subject and the concentration and/or composition of the antibody composition.

As described above, the human anti-LAG-3 antibodies of the invention can be co-administered with one or more other therapeutic agents, for example, a cytotoxic agent, a radiotoxic agent, or an immunosuppressant. The antibody may be associated with the agent (as an immunocomplex) or it may be administered separately from the agent. In the latter case (separate administration), the antibody may be administered before, after, or simultaneously with the administration of the agent, or it may be administered in conjunction with other therapeutic agents, such as anticancer therapy such as radiation. Such therapeutic agents include, but are not limited to, antineoplastic agents such as doxorubicin (Adriamycin), cisplatin, bleomycin sulfate, carmustine, chlorambucil, dacarbazine and cyclophosphamide, hydroxyurea, which, on their own, are effective only at levels that are toxic or subtoxic to the patient. Cisplatin is administered intravenously at a dose of 100 mg/ml once every four weeks, and adriamycin is administered intravenously at a dose of 60-75 mg/ml once every 21 days. Co-administration of human anti-LAG-3 antibodies, or antigen-binding fragments thereof, of the present invention with a chemotherapeutic agent provides two anti-cancer agents that act by different mechanisms that have a cytotoxic effect on human tumor cells. Such co-administration can overcome problems caused by the development of drug resistance or changes in the antigenicity of tumor cells that would render them non-reactive towards the antibody.

Also included within the scope of the present invention are kits containing antibody compositions of the invention (eg, human antibodies, bispecific or multispecific molecules, or immunoconjugates) and instructions for use. In addition, the kit may contain an additional reagent, or one or more additional human antibodies of the invention (eg, a human antibody having additional activity that binds to an epitope on a LAG-3 antigen different from the first human antibody). Kits typically include a label indicating the intended use of the kit's contents. The term label includes any documentation or written material supplied on or with the kit, or otherwise accompanying the kit.

Combination therapy

In another aspect, the invention includes a combination therapy method in which an anti-LAG-3 antibody is co-administered with one or more additional antibodies effective in stimulating an immune response, thereby further enhancing, stimulating or upregulating the immune response in a subject. For example, the invention includes a method of stimulating an immune response in a subject by administering to the subject an anti-LAG-3 antibody and one or more additional immunostimulatory antibodies, such as an anti-PD-1 antibody, an anti-PD-L1 antibody, and/or an anti-CTLA-4 antibody, such that an immune response is stimulated in the subject, for example, tumor growth is inhibited or an antiviral response is stimulated. In one embodiment, the subject is administered an anti-LAG-3 antibody and an anti-PD-1 antibody. In another embodiment, the subject is administered an anti-LAG-3 antibody and an anti-PD-L1 antibody. In yet another embodiment, the subject is administered an anti-LAG-3 antibody and an anti-CTLA-4 antibody. In one embodiment, the anti-LAG-3 antibody is a human antibody, such as the antibody of this disclosure. Or. the anti-LAG-3 antibody may be, for example, a chimeric or humanized antibody (eg, derived from a mouse anti-LAG-3 mAb). In another embodiment, the at least one additional immunostimulatory antibody (eg, anti-PD-1, anti-PD-L1, and/or anti-CTLA-4) is a human antibody. Or, the at least one additional immunostimulatory antibody may be, for example, a chimeric or humanized antibody (eg, derived from a murine anti-PD-1, anti-PD-L1, and/or anti-CTLA-4 antibody).

In one embodiment, the present invention includes a method of treating a hyperproliferative disease (eg, cancer) by administering an anti-LAG-3 antibody and an anti-CTLA-4 antibody. In other embodiments, the anti-LAG-3 antibody is administered at a subtherapeutic dose, the anti-CTLA-4 antibody is administered at a subtherapeutic dose, or both antibodies are administered at a subtherapeutic dose. In another embodiment, the present invention includes a method of modifying side effects associated with treatment of a hyperproliferative disease with an immunostimulating agent, comprising administering an anti-LAG-3 antibody and a subtherapeutic dose of an anti-CTLA-4 antibody to a subject. In some embodiments, the anti-CTLA-4 antibody is a monoclonal antibody with

- 34 043093 human sequence 10D1 (described in the published International Patent Application PCT WO 01/14424), and the anti-LAG-3 antibody is a monoclonal antibody with a human sequence, such as 25F7, 26H10, 25E3, 8B7, 11F2 or 17E5 as described herein . Other anti-CTLA-4 antibodies covered by the methods of the present invention include, for example, those disclosed in: International Patent Applications WO 98/42752; WO 00/37504; US Patent No. 6207156; Hurwitz et al. (1998) Proc. Natl. Acad. Sci. USA 95(17): 10067-10071; Camacho et al. (2004) J. Clin. Oncology 22(145): Abstract No. 2505 (antibody CP-675206) and Mokyr et al. (1998) Cancer Res. 58:5301-5304. In some embodiments, the anti-CTLA-4 antibody binds to human CTLA-4 with a KD of 5x10'8 M or less, binds to human CTLA-4 with a KD of 1x10'8 M or less, binds to human CTLA-4 with a KD of 5x10'9 M or less, or binds to human CTLA-4 with a KD of 1x10' ⁸ M to 1x10' ¹⁰ M or less.

In one embodiment, the present invention includes a method of treating a hyperproliferative disease (eg, cancer) by administering an anti-LAG-3 antibody and an anti-PD-1 antibody. In other embodiments, the anti-LAG-3 antibody is administered at a subtherapeutic dose, the anti-PD-1 antibody is administered at a subtherapeutic dose, or both antibodies are administered at a subtherapeutic dose. In another embodiment, the present invention includes a method of modifying side effects associated with treatment of a hyperproliferative disease with an immunostimulating agent, comprising administering an anti-LAG-3 antibody and a subtherapeutic dose of an anti-PD-1 antibody to a subject. In some embodiments of the invention, the subject is a human. In some embodiments, the anti-PD-1 antibody is a human sequence monoclonal antibody and the anti-LAG-3 antibody is a human sequence monoclonal antibody, such as 25F7, 26H10, 25E3, 8B7, 11F2, or 17E5 as described herein. Examples of anti-PD-1 antibodies with human sequence include antibodies 17D8, 2D3, 4H1, 5C4 and 4A11, described in published International Patent Application PCT WO 06/121168. In some embodiments, the anti-PD-1 antibody binds to human PD-1 with a KD of 5x10'8 M or less, binds to human PD-1 with a KD of 1x10'8 M or less, binds to human PD-1 with a KD of 5x10'9 M or less or binds to human PD-1 with a KD value between 1x10' ⁸ M and 1x10' ¹⁰ M or less.

In one embodiment, the present invention includes a method of treating a hyperproliferative disease (eg, cancer) by administering an anti-LAG-3 antibody and an anti-PD-L1 antibody. In other embodiments, the anti-LAG-3 antibody is administered at a subtherapeutic dose, the anti-PDL1 antibody is administered at a subtherapeutic dose, or both antibodies are administered at a subtherapeutic dose. In another embodiment, the present invention includes a method of modifying side effects associated with treatment of a hyperproliferative disease with an immunostimulating agent, comprising administering an anti-LAG-3 antibody and a subtherapeutic dose of an anti-PD-L1 antibody to a subject. In some embodiments, the subject is a human. In some embodiments, the anti-PD-L1 antibody is a human sequence monoclonal antibody and the anti-LAG-3 antibody is a human sequence monoclonal antibody, such as 25F7, 26H10, 25E3, 8B7, 11F2, or 17E5 as described herein. Examples of anti-PD-L1 antibodies with human sequence include antibodies 3G10, 12A4, 10A5, 5F8, 10H10, 1B12, 7H1, 11E6, 12B7 and 13G4, described in published International Patent Application PCT WO 07/005874. In some embodiments, the anti-PD-L1 antibody binds to human PDL1 with a KD of 5x10'8 M or less, binds to human PD-L1 with a KD of 1x10'8 M or less, binds to human PD-L1 with a KD of 5x10'9 M, or less or binds to human PD-L1 with a KD value between 1x10' ⁸ M and 1x10' ¹⁰ M or less.

Blockade of LAG-3 and one or more second target antigens such as CTLA-4 and/or PD-1 and/or PD-L1 by antibodies can enhance the immune response to cancer cells in a patient. Cancer cells whose growth can be inhibited by the antibodies of the present disclosure include cancer cells typically responsive to immunotherapy. Representative examples of cancers for treatment with combination therapy of the present disclosure include the cancers specifically listed above in the discussion of anti-LAG-3 antibody monotherapy.

In some embodiments, the combination of therapeutic antibodies discussed herein can be administered simultaneously as a single composition in a pharmaceutically acceptable carrier, or simultaneously as separate compositions with each antibody in a pharmaceutically acceptable carrier. In one embodiment of the invention, the combination of therapeutic antibodies can be administered sequentially. For example, an anti-CTLA-4 antibody and an anti-LAG-3 antibody may be administered sequentially such that the anti-CTLA-4 antibody is administered first and the anti-LAG-3 antibody second, or the anti-LAG-3 antibody is administered first and the anti-CTLA antibody -4 second. Additionally or alternatively, the anti-PD-1 antibody and the anti-LAG-3 antibody may be administered sequentially, such that the anti-PD-1 antibody is administered first and the anti-LAG-3 antibody second, or the anti-LAG-3 antibody is administered first and anti-PD-1 antibody second. In addition or alternatively, the anti-PD-L1 antibody and the anti-LAG-3 antibody can be administered sequentially such that the anti

- 35 043093 anti-PD-L1 is administered first and anti-LAG-3 antibody second, or anti-LAG-3 antibody is administered first and anti-PD-L1 antibody second.

Further, if more than one dose of a combination therapeutic drug is administered sequentially, the order of sequential administration can be reversed or the same order of administration maintained at each time point, sequential administration can be combined with simultaneous administration, or any combination thereof. For example, the first administration of a combination of anti-CTLA-4 antibody and anti-LAG-3 antibody may be simultaneous, the second administration may be sequential, with the anti-CTLA-4 antibody administered first and the anti-LAG-3 antibody second, and the third administration may be sequential. with the anti-LAG-3 antibody administered first, and the anti-CTLA-4 antibody second, etc. Additionally or alternatively, the first administration of the combination of anti-PD-1 antibody and anti-LAG-3 antibody may be simultaneous, the second administration may be sequential, with the anti-PD-1 antibody administered first, the anti-LAG-3 antibody second, and the third administration may be sequential, with the anti-LAG-3 antibody administered first and the anti-PD-1 antibody second, etc. Additionally or alternatively, the first administration of the combination of anti-PD-L1 antibody and anti-LAG-3 antibody may be simultaneous, the second administration may be sequential, with the anti-PD-L1 antibody administered first, the anti-LAG-3 antibody second, and the third administration may be sequential, with the anti-LAG-3 antibody administered first and the anti-PD-L1 antibody second, etc. Another typical drug administration regimen may include a first administration being sequential, with anti-LAG-3 antibody administered first and anti-CTLA-4 (and/or anti-PD-1 and/or anti-PD-L1) antibody second, and subsequent administrations can be simultaneous (competitive).

Optionally, a combination of an anti-LAG-3 antibody and one or more additional antibodies (eg, anti-CTLA-4 and/or anti-PD-1 and/or anti-PD-L1 antibodies) can be further combined with an immunogenic agent, such as purified cancer cells tumor antigens (including recombinant protein, peptide and carbohydrate molecules), cells and cells transfected with genes encoding immunostimulatory cytokines (He et al. (2004) J. Immunol. 173: 4919-28). Non-limiting examples of tumor vaccines that can be used include peptides or melanoma antigens, such as gp100 protein peptides, MAGE antigens, Trp-2, MART1 and/or tyrosinase, or tumor cells transfected to express the cytokine GM-CSF (more discussed below). Combined blockade of LAG-3 and CTLA-4 and/or PD-1 and/or PD-L1 can further be combined with a vaccination protocol, such as any of the vaccination protocols discussed in detail above in the description of anti-LAG-3 antibody monotherapy.

Combination blockade of LAG-3 and CTLA-4 and/or PD-1 and/or PD-L1 can be further combined with standard anticancer therapeutic agents. For example, combination blockade of LAG-3 and CTLA-4 and/or PD-1 and/or PD-L1 can be further combined with chemotherapy regimens. In these cases, it is possible to reduce the dose of the other chemotherapeutic reagent administered with the combination of the present disclosure (Mokyr et al. (1998) Cancer Research 58: 5301-5304). An example of such a combination is the combination of anti-LAG-3 antibodies and anti-CTLA-4 antibodies and/or anti-PD-1 antibodies and/or anti-PD-L1 antibodies in combination with dacarbazine for the treatment of melanoma. Another example is the combination of anti-LAG-3 antibodies and anti-CTLA-4 antibodies and/or anti-PD-1 antibodies and/or anti-PD-L1 antibodies in combination with interleukin-2 (IL-2) for the treatment of melanoma. The scientific rationale for the combined use of LAG-3 and CTLA-4 and/or PD-1 and/or PD-L1 blockade with chemotherapy is that cell death, which is a consequence of the cytotoxic effects of most chemotherapeutic compounds, would cause increased levels of tumor antigen in the pathway. antigen presentation. Other combination therapies that, when combined with combined blockade of LAG-3 and CTLA-4 and/or PD-1 and/or PD-L1, may have a synergistic effect through cell death include radiation, surgery, or hormonal deprivation. Each of these protocols creates a source of tumor antigen in the host. Angiogenesis inhibitors can also be combined with blockade of LAG-3 and CTLA-4 and/or PD-1 and/or PD-L1. Inhibition of angiogenesis leads to the death of tumor cells, which can ensure the entry (feeding) of tumor antigen into the host antigen presentation pathway.

A combination of blocking antibodies to LAG-3 and CTLA-4 and/or PD-1 and/or PD-L1 can also be used in combination with bispecific antibodies that target effector cells expressing the Fca or Fcy receptor to tumor cells (see, e.g., US Patent Nos. 5,922,845 and 5,837,243). Bispecific antibodies can be used to target two separate antigens. The T cell component (arm) of these responses is enhanced by combined blockade of LAG-3 and CTLA-4 and/or PD-1 and/or PD-L1.

In another example, a combination of anti-LAG-3 and anti-CTLA-4 antibodies and/or anti-PD-1 antibodies and/or anti-PD-L1 antibodies can be used in combination with anti-tumor antibodies such as Rituxan® (rituximab), Herceptin® (trastuzumab ), Bexxar®, tositumomab, Zevalin® (ibritumomab), Campath® (alemtuzumab), Lymphocide® (epratuzumab), Avastin® (bevaciza

- 36 043093 mab) and Tarceva (Tarceva®, erlotinib), etc. By way of example, and without wishing to be bound by theory, treatment with an anticancer antibody or an anticancer antibody conjugated to a toxin may cause cancer cell death (eg, tumor cells), which could enhance the immune response mediated by CTLA-4, PD-1, PD -L1 or LAG-3. A typical example of treatment for a hyperproliferative disease (eg, cancer) may include an anti-cancer antibody in combination with anti-LAG-3 antibodies and anti-CTLA-4 antibodies and/or anti-PD-1 antibodies and/or anti-PD-L1 antibodies, simultaneously or sequentially, or in a combination of simultaneous and sequential administration, which can enhance the host antitumor immune response.

Tumors evade immune control using a variety of mechanisms. Many of these mechanisms can be overcome by inactivating proteins that are expressed by tumors and that have immunosuppressive activity. These proteins include, among others, TGF-β (Kehrl et al. (1986) J. Exp. Med. 163: 1037-1050), IL-10 (Howard & O'Garra (1992) Immunology Today 13: 198-200 ), and Fas ligand (Hahne et al. (1996) Science 274: 1363-1365). In another example, antibodies to each of these structures can be further combined with a combination of anti-LAG-3 and anti-CTLA-4 and/or anti-PD-1 and/or anti-PD-L1 antibodies to counteract the effects of the immunosuppressant and promote the host immune response to the tumor.

With the combination of anti-LAG-3 and anti-CTLA-4 and/or anti-PD-1 and/or anti-PD-L1 antibodies, other antibodies that activate the host immune response can be additionally used. These antibodies include molecules on the surface of dendritic cells that activate DC function and antigen presentation. Anti-CD40 antibodies (Ridge et al., supra) can be used in combination with a combination of anti-LAG-3 and anti-CTLA-4 and/or anti-PD-1 and/or anti-PD-L1 antibodies (Ito et al., see above (2000) Other activating antibodies to T cell costimulatory molecules (Weinberg et al., supra, Melero et al., supra, Hutloff et al. (1999), supra) may also provide increased levels activation of T cells.

As discussed above, bone marrow transplantation is currently used to treat various tumors of hematopoietic origin. Combined blockade of LAG-3 and CTLA4 and/or PD-1 and/or PD-L1 can be used to improve the efficacy of transplanted donor tumor-specific T cells.

There are also some experimental treatment protocols that involve ex vivo activation and expansion of antigen-specific T cells and adoptive transfer of these cells to recipients to stimulate antigen-specific T cells against the tumor (Greenberg & Riddell, supra). These methods can also be used to activate T cell responses to infectious agents such as CMV. It can be speculated that ex vivo activation in the presence of anti-LAG-3 and anti-CTLA-4 and/or anti-PD-1 and/or anti-PD-L1 antibodies may increase the frequency and activity of adoptively transferred T cells.

Some embodiments of the present invention include a method of modifying a side effect associated with treatment of a hyperproliferative disease (eg, cancer) with an immunostimulating agent, comprising administering to the subject an anti-LAG-3 antibody and a subtherapeutic dose of an anti-CTLA-4 and/or anti-PD-1 and/or PD-L1. For example, the methods of the present invention include a method of reducing the incidence of colitis or diarrhea caused by an immunostimulating therapeutic antibody by administering a non-GI steroid to the patient. Since any patient receiving an immunostimulatory therapeutic antibody has an increased risk of colitis or diarrhea caused by such antibody, therapy with the methods of the present invention is applicable to this entire group of patients. Although steroids are used to treat inflammatory bowel disease (IBD) and to prevent recurrence (exacerbation) of IBD, they can be used to prevent (reduce the incidence) of IBD in patients who have not been diagnosed with IBD. The noticeable side effects associated with steroids, even non-absorbable steroids, preclude their use for prophylactic purposes.

In other embodiments, a combination of blockade of LAG-3 and CTLA-4 and/or PD-1 and/or PD-L1 (i.e., immunostimulatory therapeutic antibodies against LAG-3 and against CTLA-4 and/or against PD-1 and/or or anti-PD-L1) can be further combined with a non-absorbable steroid. The term non-absorbable steroid means a glucocorticoid that has extensive first-pass metabolism such that, after metabolism in the liver, the bioavailability of the steroid is low, namely below about 20%. In one embodiment of the invention, the non-absorbable steroid is budenoside. Budenoside is a topical glucocorticosteroid that, after oral administration, undergoes extensive metabolism in the liver. ENTOCORT EC® (Astra-Zeneca) is a pH- and time-sensitive oral formulation of budenoside designed to optimize drug delivery to the ileum and throughout the colon. ENTOCORT EU® is approved in the United States for the treatment of mild to moderate Crohn's disease affecting the ileum and/or adjacent colon. The usual oral dose of ENTOCORT EC® for the treatment of Crohn's disease is 6-9 mg/day. ENTOCORT EC®

- 37 043093 is released in the intestine before it is absorbed and stored in the intestinal mucosa. Upon passage of the target tissue of the intestinal mucosa, ENTOCORT EC® undergoes extensive metabolism by the cytochrome P450 system in the liver to metabolites with negligible glucocorticoid activity. Therefore, bioavailability is low (about 10%). As a result of its low bioavailability, the therapeutic index of budenoside is low compared to other glucocorticoids with less extensive first-pass metabolism. Budenoside causes fewer side effects, including less hypothalamic-pituitary insufficiency, than systemic steroids. However, chronic use of ENTOCORT EC® may cause systemic glucocorticoid effects such as hypercortisolism and adrenal suppression. See PDR 58 ^th ed. 2004; 608610.

In other embodiments, a combination of LAG-3 and CTLA-4 and/or PD-1 and/or PD-L1 blockade (i.e., anti-LAG-3 and anti-CTLA-4 and/or anti-PD-1 and/or anti-PD-1 immunostimulatory therapeutic antibodies /or anti-PD-L1) in combination with the use of a non-absorbable steroid can be further combined with a salicylate. Salicylates include 5-ASA agents such as: sulfasalazine (AZULFIDINE®, Pharmacia &UpJohn); osalazine (DIPENTUM®, Pharmacia &UpJohn); balsalazide (COLAZAL®, Salix Pharmaceuticals, Inc.); and mesalamine (ASACOL®, Procter & Gamble Pharmaceuticals; PENTASA®, Shire US; CANASA®, Axcan Scandipharm, Inc.; ROWASA®, Solvay).

In accordance with the methods of the present invention, the regimen of salicylate administered in combination with anti-LAG-3 and anti-CTLA-4 and/or anti-PD-1 and/or anti-PD-L1 antibodies and a non-absorbable steroid may include any overlapping (in time ) or sequential administration of a salicylate and a non-absorbable steroid to reduce the incidence of colitis caused by immunostimulating antibodies. For example, methods for reducing the incidence of colitis caused by the immunostimulatory antibodies of the present invention include administering salicylate and a non-absorbable steroid simultaneously or sequentially (eg, salicylate administered 6 hours after the non-absorbable steroid) or any combination thereof. Further, in accordance with the present invention, the salicylate and the non-absorbable steroid may be administered by the same route (eg, both are administered orally) or by different routes (eg, the salicylate is administered orally and the non-absorbable steroid is administered rectally), which may be different from the route(s) of administration of antibodies against LAG-3 and anti-CTLA-4 and/or anti-PD-1 and/or anti-PD-L1.

The present invention is further illustrated by the following examples, which should not be construed as further limiting. The contents of all figures and all references, Genbank sequences, patents and published patent applications are expressly incorporated herein by reference. In particular, the disclosure of published PCT International Applications WO 09/045957, WO 09/073533, WO 09/073546 and Wo 09/054863 is expressly incorporated herein by reference.

Example 1: Preparation of human monoclonal antibodies against LAG-3

Human monoclonal antibodies against LAG-3 are obtained from transgenic mice that express the antibody genes as follows.

Antigens

Recombinant human LAG-3 fusion proteins are used as an immunogen to generate antibodies against human LAG-3. In some immunization regimens, a fusion protein containing the entire extracellular region (domains 1-4) of human LAG3 fused with the Fc domain of human immunoglobulin (R&D Systems, Catalog #2319-L3) (D1-D4 hFc) or the Fc domain of mouse immunoglobulin is used as an immunogen. immunoglobulin (D1-D4 mFc). In other immunization regimens, a fusion protein containing only the first two extracellular domains of human LAG-3 fused to the Fc domain of a mouse immunoglobulin (D1-D2 mFc) is used as an immunogen. LAG-3 fusion proteins are produced by conventional recombinant DNA methods

Transgenic trans-chromosomal mouse lines KM Mouse™ and KM/FCGR2D Mouse™

Fully human monoclonal antibodies to human LAG-3 are obtained from transgenic trans-chromosomal KM Mouse™ and KM/FCGR2D Mouse™ mice, which express human antibody genes.

In KM Mouse™ mice, the endogenous mouse kappa light chain gene is homozygously disrupted as described in Chen et al. (1993) EMBO J. 12: 811-820, and the endogenous mouse heavy chain gene is homozygously disrupted as described in Example 1 of published patent application PCT WO 01/09187. In addition, this mouse strain carries the human kappa light chain transgene, KCo5, described in Fishwild et al., supra. Mice of this strain also carry the trans chromosome SC20, which has the human Ig heavy chain locus, as described in published PCT International Application WO 02/43478. The KM/FCGR2D Mouse™ line is similar to the KM Mouse™ line except that its genome also contains a homozygous disruption of the endogenous FcyRIIB gene. The KM Mouse™ and KM/FCGR2D Mouse™ mouse lines are also described in detail in US Patent Application Published No. 20020199213.

- 38 043093

Immunization of KM Mouse™ and KM/FCGR2D Mouse™ mice

To generate fully monoclonal antibodies to LAG-3, KM Mouse™ and KM/FCGR2D Mouse™ mice are immunized with one of the three different recombinant LAG-3 fusion proteins described above (D1-D4 hFc, D1-D4 mFc, D1-D2, mFc ). General immunization schedules are described in Lonberg et al. (1994), see above; Fishwild et al., supra; and in published PCT International Application WO 98/24884. Mice are 6-16 weeks old before the first antibody infusion. Mice are immunized intraperitoneally (into the peritoneal cavity) (IP) and/or subcutaneously (SC). Mice are immunized biweekly four times with 10 μg of recombinant LAG-3 fusion protein, then immunized twice with 20 μg of the same immunogen using Ribi as an adjuvant. The immune response is monitored by collecting a blood sample from the retroorbital sinus. Plasma is screened by ELISA (as described below), and mice with satisfactory anti-LAG-3 human immunoglobulin titers are used for fusions. Mice are given a booster dose of 20 μg of antigen intravenously and intraperitoneally, followed by a booster dose of 20 μg of antigen intravenously, then sacrificed and the spleen removed.

Selection of KM and KM/FCGR2D Mice Producing Anti-LAG-3 Antibodies

To select mice that produce antibodies that bind the LAG-3 protein, sera from mice immunized with the D1-4 hFc fusion protein are tested by a modified ELISA method as originally described by Fishwild et al. (1996). Briefly, recombinant LAG-3 fusion protein was immobilized on microtiter plates at a concentration of 1 μg/ml in PBS, 50 μl/well, incubated at 4°C overnight, then blocked with 200 μl/well of 5% BSA in PBS. Dilutions of plasma from mice immunized with LAG-3 are added to each well and incubated for 1-2 hours at room temperature. The plates are washed in PBS/Tween and then incubated with goat anti-human immunoglobulin kappa light chain polyclonal antibody conjugated to horseradish peroxidase (HRP) for 1 h at room temperature. The plates are washed, treated with ABTS substrate and analyzed on a spectrophotometer at OD ₄₀₅ .

Serum from mice immunized with D1-D4 mFc or D1-D2 mFc fusion proteins is analyzed by indirect ELISA using goat anti-mouse IgG antibody immobilized on plates for one hour before immobilizing the antigen to exclude nonspecific binding to the mouse Fc region. The same ELISA steps as described above are then performed.

Mice treated with the highest titers of anti-LAG-3 antibodies are used for fusions. Fusions were performed as described below, and hybridoma cell supernatants were tested for anti-LAG-3 activity by ELISA.

Preparation of hybridoma cells producing human monoclonal antibodies against LAG-3 proteins

Murine splenocytes isolated from KM or KM/FCGR2D mice were electrofused using a Cyto Pulse large cell fusion chamber electroporator (Cyto Pulse Sciences, Inc., Glen Burnie, MD) for the murine myeloma cell line. The resulting hybridomas are then screened for the production of antigen-specific antibodies. Single-cell suspensions of spleen lymphocytes from immunized mice are fused with one-fourth the number of P3X63 Ag8.6.53 (ATCC CRL 1580) nonsecreting mouse myeloma cells. Cells are seeded at a density of approximately 1 x 10 ⁵ /well in a flat-bottom microtiter plate and then incubated for approximately two weeks in selective medium containing 10% fetal calf serum supplemented with the ori gene (IGEN) in RPMI, L-glutamine, sodium pyruvate, HEPES , penicillin, streptamycin, gentamicin, 1xHAT and β-mercaptoethanol. After 1-2 weeks, the cells are cultured in a medium in which HAT is replaced by NT. Individual wells are then screened by ELISA (described above) for human anti-LAG-3 monoclonal IgG antibodies. When intensive growth of hybridomas begins, usually after 10-14 days, control the environment. Antibody-secreting hybridomas are subcultured, screened again, and, if still positive for human IgG, anti-LAG-3 monoclonal antibodies are subcloned at least twice by serial dilution. Stable subclones are then cultured in vitro to produce small amounts of antibody in tissue culture medium for further study.

Hybridoma cell clones 25F7, 26H10, 25E3, 8B7, 11F2 and 17E5 were selected for further analysis and sequencing.

Example 2. Structural characteristics of human monoclonal antibodies against LAG-3 25F7, 26H10, 25E3, 8B7, 11F2 and 17E5

The cDNA sequences encoding the variable regions of the heavy and light chains of mAbs expressed by clones 25F7, 26H10, 25E3, 8B7, 11F2 and 17E5, described in example 1, are sequenced in accordance with the following protocol. Total RNA was prepared using 5x106 hybridoma cells using the RNeasy Mini Kit (Qiagen, Valencia, CA). cDNA was obtained according to the 5'-RACE protocol with the SMART RACE cDNA Amplification Kit (Clontech

- 39 043093

Laboratories, Inc., Mountain View, CA) and Superscript II reverse transcriptase (Invitrogen, Carlsbad, CA). The V regions of each antibody are amplified using a 3' primer to the human constant region. A 3' primer specific for the human immunoglobulin constant region is paired with a 5' RACE universal primer mixture. PCR products containing the V region were cloned into the pCR4TOPO vector (Invitrogen, Carlsbad, CA) and transferred into E. coli strain TOP10 (Invitrogen, Carlsbad, CA). Samples were prepared with either DNA miniprep or Templiphi (GE Healthcare Biosciences, Piscataway, NJ, USA) and subjected to DNA sequencing (Sequetech, Mountain View, CA). The resulting DNA sequences are analyzed for in-frame rearrangements and other characteristics. The expressed proteins are examined by conventional protein chemistry methods. Clones 25E3, 25F7 and 26H10 were found to express an IgG1 heavy chain and kappa light chain antibody, while clones 8B7 and 17E5 expressed an IgG4 heavy chain and kappa light chain antibody, and clone 11F2 expressed an IgG4 heavy chain and kappa light chain antibody. IgG2 light chain.

The nucleotide and amino acid sequences of the 25F7 heavy chain variable region are shown in FIG. 1A and SEQ ID NO: 49 and 37, respectively. The nucleotide and amino acid sequences of the kappa 25F7 light chain variable region are shown in FIG. 1B and SEQ ID NO: 55 and 43, respectively. Comparison of the 25F7 immunoglobulin heavy chain sequence with known germline immunoglobulin heavy chain sequences (FIG. 7) shows that the 25F7 heavy chain uses a _VH segment from human germline _VH 4-34 (SEQ ID NO: 61) and a JH segment from human germline JH5b (SEQ ID NO: 62). More detailed analysis of the 25F7 VH sequence using the Kabat CDR region detection system depicts the heavy chain CDR1, CDR2 and CDR3 regions shown in FIG. 1A and SEQ ID NO: 1, 7 and 13, respectively. Comparison of the 25F7 immunoglobulin light chain sequence with known germline immunoglobulin light chain sequences (FIG. 8) shows that the 25F7 kappa light chain uses a V _K segment from human germline _VK L6 (SEQ ID NO: 63) and a _JK segment from human germline JK2 (SEQ ID NO: 64). More detailed analysis of the 25F7 V _K sequence using the Kabat CDR region detection system depicts the light chain CDR1, CDR2 and CDR3 regions shown in FIG. 1B and SEQ ID NO: 19, 25 and 31, respectively.

The nucleotide and amino acid sequences of the 26H10 heavy chain variable region are shown in FIG. 2A and SEQ ID NO: 50 and 38, respectively. The nucleotide and amino acid sequences of the 26H10 light chain variable region are shown in FIG. 2B and SEQ ID NO: 56 and 44, respectively. Comparison of the immunoglobulin 26H10 heavy chain sequence with known germline immunoglobulin heavy chain sequences (Fig. 9) shows that the 26H10 heavy chain uses a VH segment from human germline VH 3-33 (SEQ ID NO: 65) and a JH segment from human germline JH 6B (SEQ ID NO: 66). More detailed analysis of the 26H10 VH sequence using the Kabat CDR region detection system depicts the heavy chain CDR1, CDR2 and CDR3 regions shown in FIG. 2A and SEQ ID NO: 2, 8 and 14, respectively. Comparison of the immunoglobulin 26H10 light chain sequence with known germline immunoglobulin light chain sequences (FIG. 10) shows that the 26H10 kappa light chain uses a VK segment from human germline VK _A27 (SEQ ID NO: 67) and a _JK segment from human germline JK3 lines (SEQ ID NO: 68). More detailed analysis of the 26H10 V _K sequence using the Kabat CDR region determination system reveals the light chain CDR1, CDR2 and CDR3 regions shown in FIG. 2B and SEQ ID NO: 20, 26 and 32, respectively.

The nucleotide and amino acid sequences of the 25E3 heavy chain variable region are shown in FIG. 3A and SEQ ID NO: 51 and 39, respectively. The nucleotide and amino acid sequences of the 25E3 light chain variable region are shown in FIG. 3B and SEQ ID NO: 57 and 45, respectively. Comparison of the immunoglobulin 25E3 heavy chain sequence with known germline immunoglobulin heavy chain sequences (FIG. 11) shows that the 25E3 heavy chain uses a VH segment from human germline VH 3-20 (SEQ ID NO: 69) and a JH segment from human germline JH 4B (SEQ ID NO: 70). More detailed analysis of the 25E3 V _H sequence using the Kabat CDR region determination system depicts the heavy chain CDR1, CDR2 and CDR3 regions shown in FIG. 3A and SEQ ID NO: 3, 9 and GGY, respectively. Comparison of the immunoglobulin 25E3 light chain sequence with known germline immunoglobulin light chain sequences (FIG. 12) shows that the 25E3 kappa light chain uses a VK segment from the human germline VK L18 (SEQ ID NO: 71) and a JK segment from the human germline JK 2 (SEQ ID NO: 64). More detailed analysis of the 25E3 VK sequence using the Kabat CDR region detection system depicts the light chain CDR1, CDR2 and CDR3 regions shown in FIG. 3B and SEQ ID NO: 21, 27 and 33, respectively.

The nucleotide and amino acid sequences of the 8B7 heavy chain variable region are shown in FIG. 4A and SEQ ID NO: 52 and 40, respectively. The nucleotide and amino acid sequences of the 8B7 light chain variable region are shown in FIG. 4B and SEQ ID NO: 58 and 46, respectively. Comparison of the immunoglobulin 8B7 heavy chain sequence with known sequences

- 40 043093 germline immunoglobulin heavy chain activities (FIG. 13) shows that the 8B7 heavy chain uses a VH segment from human germline VH 4-34 (SEQ ID NO: 61) and a JH segment from human germline JH 5B (SEQ ID NO: 62). More detailed analysis of the 8B7 VH sequence using the Kabat CDR region detection system depicts the heavy chain CDR1, CDR2 and CDR3 regions shown in FIG. 4A and SEQ ID NO: 4, 10 and 16, respectively. Comparison of the immunoglobulin 8B7 light chain sequence with known germline immunoglobulin light chain sequences (FIG. 14) shows that the kappa 8B7 light chain uses a V _K segment from the human germline _VK L6 (SEQ ID NO: 63) and a _JK segment from the human germline JK4 (SEQ ID NO: 72). More detailed analysis of the 8B7 V _K sequence using the Kabat CDR region determination system reveals the light chain CDR1, CDR2 and CDR3 regions shown in FIG. 4B and SEQ ID NO: 22, 28 and 34, respectively.

The nucleotide and amino acid sequences of the 11F2 heavy chain variable region are shown in FIG. 5A and SEQ ID NO: 53 and 41, respectively. The nucleotide and amino acid sequences of the 11F2 light chain variable region are shown in FIG. 5B and SEQ ID NO: 59 and 47, respectively. Comparison of the immunoglobulin 11F2 heavy chain sequence with known germline immunoglobulin heavy chain sequences (FIG. 15) shows that the 11F2 heavy chain uses a V _H segment from human germline V _H 1-24 (SEQ ID NO: 73), a D segment from human germline 2-15 and a JH segment from human germline JH 4B (SEQ ID NO: 70). More detailed analysis of the 11F2 V _H sequence using the Kabat CDR region detection system depicts the heavy chain CDR1, CDR2 and CDR3 regions shown in FIG. 13A and SEQ ID NO: 5, 11 and 17, respectively. Comparison of the 11F2 immunoglobulin light chain sequence with known germline immunoglobulin light chain sequences (FIG. 16) shows that the 11F2 kappa light chain uses a VK segment from the human germline VK L6 (SEQ ID NO: 63) and a JK segment from the human germline JK1 (SEQ ID NO: 72). More detailed analysis of the 11F2 VK sequence using the Kabat CDR region detection system depicts the light chain CDR1, CDR2 and CDR3 regions shown in FIG. 5B and SEQ ID NO: 23, 29 and 35, respectively.

The nucleotide and amino acid sequences of the 17E5 heavy chain variable region are shown in FIG. 6A and SEQ ID NO: 54 and 42, respectively. The nucleotide and amino acid sequences of the 17E5 light chain variable region are shown in FIG. 6B and SEQ ID NO: 60 and 48, respectively. Comparison of the immunoglobulin 17E5 heavy chain sequence with known germline immunoglobulin heavy chain sequences (FIG. 17) shows that the 17E5 heavy chain uses a VH segment from human germline VH 3-33 (SEQ ID NO: 65), a D segment from human germline 2-2 and JH segment from human germline JH 4B (SEQ ID NO: 70). More detailed analysis of the 17E5 VH sequence using the Kabat CDR region detection system depicts the heavy chain CDR1, CDR2 and CDR3 regions shown in FIG. 6A and SEQ ID NO: 6, 12 and 19, respectively. Comparison of the 17E5 immunoglobulin light chain sequence with known germline immunoglobulin light chain sequences (FIG. 18) shows that the 17E5 kappa light chain uses a V _K segment from human germline VK _L6 (SEQ ID NO: 63) and a J _K segment from human germline JK5 (SEQ ID NO: 75). More detailed analysis of the 17E5 V _K sequence using the Kabat CDR region determination system depicts the light chain CDR1, CDR2 and CDR3 regions shown in FIG. 6B and SEQ ID NO: 24, 30 and 36, respectively.

The variable regions of antibodies 25F7, 26H10, 25E3, 8B7, 11F2 and 17E5 can be converted into full-length antibodies of any desired isotype using standard recombinant DNA methods. For example, DNA encoding the VH and VL regions can be cloned into an expression vector that carries heavy and light chain constant regions such that the variable regions are operably linked to the constant regions. Alternatively, separate vectors can be used to express the full-length heavy chain and the full-length light chain. Non-limiting examples of the use of expression vectors to create full-length antibodies include the pIE vectors described in US Patent Application Published No. 20050153394.

Example 3. Characteristics of the binding properties of monoclonal antibodies against LAG-3

In this example, the binding of human anti-LAG-3 antibodies to cell surface LAG-3 (human, simian and mouse LAG-3) is studied by flow cytometry. In addition, the kinetics of binding to LAG-3 is studied using the BIACORE optical biosensor. In addition, epitope mapping is carried out using the peptide scanning method.

A. Flow cytometry study 1.

Binding of CHO cells to human LAG-3

To test the ability of the antibodies to bind to the cell surface protein LAG-3, the antibodies are incubated with a CHO cell line transfected to express human LAG-3 on the cell surface. Serial dilutions of monoclonal antibodies 25F7, 26H10, 25E3, 8B7, 11F2 and 17E5 are made in cold 1x PFAE buffer (1x PBS+2% FBS, 0.02% sodium azide, 2mM Na

- 41 043093

EDTA). For the binding reaction, 50 µl of the diluted antibody solution is added to 50 µl of a cell suspension containing 2x105 cells, and the mixture is incubated on ice for 30 minutes. Cells are then washed twice with 1x PFAE buffer. FITC-labeled goat anti-human immunoglobulin kappa light chain antibody (Bethyl Laboratories, Inc., Cat. # A80-115F) was added at a 1:100 dilution and the mixture was incubated for 30 min at 4°C and then washed twice with cold 1x PFAE buffer. . After the last wash, 150 μl of cold 1x PFAE buffer containing 10 μg/ml propidium iodide (Roche Applied Science, Cat #1348639) was added to each solution and antibody binding assay was performed by flow cytometry on a FACScalibur flow pitometer (BD Bioscience).

The results of the flow cytometry analysis are summarized below in Table. 1, which provides EC ₅₀ values for binding to CHO-human LAG-3, showing that antibodies 25F7, 26H10, 25E3, 8B7, 11F2 and 17E5 bind effectively to cell surface human LAG-3, with the EC ₅₀ value of antibody 25F7, approximately 20 times lower than the EC ₅₀ for 25E3, but approximately equivalent to the EC ₅₀ for antibodies 8B7 and 26H10. The EC ₅₀ values for 11F2 and 17E5 are in the same range as for 25E3.

Table 1. Binding of anti-LAG-3 antibodies to CHO cells expressing human LAG-3

Antibody EU so (nM) 25F7 0.45-2.52 8B7 1.93-4.44 26N10 1.81-3.64 11F2 15.12 25EZ 14.9-25.39 17E5 12.3

2. Binding of activated human CD4+ T cells

To test the ability of antibodies to bind to native LAG-3 on the surface of activated human T cells, resting CD4+ T cells were isolated from purified peripheral blood mononuclear cells and stimulated for three days with a combination of anti-CD3 and anti-CD28 antibodies fixed to polystyrene beads. Serial dilutions of monoclonal antibodies 25F7, 8B7 and 26H10 are made in cold 1x PFAE buffer (1x PBS+2% FBS, 0.02% sodium azide, 2 mM Na EDTA). For the binding reaction, 50 μl of the diluted antibody solution is mixed with 50 μl of PE-labeled anti-human CD4 antibody (BD Bioscience, Cat # 555347). Activated T cells are processed according to the above protocol. The antibody binding assay was performed as described above.

The results of the flow cytometry analysis are summarized below in Table. 2, which provides EC ₅₀ values for binding to activated human CD4 ⁺ T cells, showing that all three antibodies similarly bind to human cell surface LAG-3.

Table 2. Binding of anti-LAG-3 antibodies to activated human CD4 ⁺ T cells

Antibody ECso (nM) 25F7 0.27-0.45 26N10 0.41-0.84 8B7 0.69-1.80

3. Binding of mouse LAG-3 antigen

To determine whether anti-LAG-3 antibodies cross-react with mouse LAG-3, the cDNA sequence was cloned by RT-PCR using a cDNA pool preparation obtained by reverse transcription of RNA from a collection of tissue samples from cynomolgus and rhesus macaques. The sequence is first amplified from the cDNA pool using primers (5'-forward primer: 5Mcyn1408; 5'-atgtgggaggctcagttcctg-3' (SEQ ID NO: 91) & 3'-reverse primer: 3Mcyn1408α; 5'-gtcagagctgctccggctc-3' ( SEQ ID NO: 92)) using a GC enriched PCR amplification system (Roche) and cloned into TOPO recipient cloning vector (Invitrogen) for sequence analysis. Clones corresponding to the Macacresus LAG-3 reference sequence in Genbank (Genbank Accession No. XM_001108923) are then reamplified using the TOPO DNA cloning vector using a second set of primers that introduce restriction enzyme sites for directed cloning into the expression vector in mammalian cells.

Clone pa23-5 with simian LAG-3 was isolated and sequenced. The isolated monkey sequence is 99.6% identical to the rhesus monkey LAG-3 reference sequence in Genbank. A comparison of the amino acid sequence of the pa23-3 cDNA clone (SEQ ID NO: 93) with the rhesus monkey LAG-3 sequence (SEQ ID NO: 94) from Genbank (Accession No. XM_001108923) is shown in FIG. 19. The two sequences are identical except for one amino acid difference at position 419 (arginine in clone pa23-5 versus threonine in the rhesus monkey sequence in Genbank), and based on

- 42 043093 this concludes that the cDNA clone pa23-5 represents the sequence of the macacresus LAG-3 gene.

The pa23-5 cDNA is introduced into an expression construct, which is transfected into a CHO-S cell suspension by nucleofection (Amaxa). The expression of LAG-3 by rhesus monkeys from sorted, drug resistance-selected clones was checked by FACS analysis. This clonal CHO cell line overexpressing rhesus LAG-3 was used in FACS assays similar to those described above to quantify cross-reactivity to the simian protein. Briefly, serial dilutions of monoclonal antibodies 25F7, 8B7 and 26H10 are made in cold 1x PFAE buffer (1x PBS + 2% FBS, 0.02% sodium azide, 2 mM Na EDTA). For the binding reaction, 50 μl of a diluted antibody solution is added to 50 μl of a cell suspension containing 2x10 ⁵ cells, and the mixture is incubated on ice for 30 minutes. Cells are processed according to the protocol described above. The antibody binding assay was performed as described above.

In a separate experiment, the antibodies were tested for binding to LAG-3 from cynomolgus monkeys using activated T cells from these monkeys. In vitro activation of these simian T cells is accomplished by treating the T cells with anti-CD3/anti-CD28 antibodies following essentially the same protocol described above for in vitro activation of human T cells, followed by flow cytometry analysis as described above. for in vitro staining of activated human CD4+ T cells.

The results of flow cytometry analysis using CHO-rhesus LAG-3 cells and activated T cells from cynomolgus monkeys are summarized below in table. 3, which presents EC ₅₀ values for binding to two different cell types expressing simian LAG-3. These results indicate that all antibodies bind effectively to both LAG-3 on activated Cynomolgus monkey T cells and to LAG-3 rhesus monkeys (SEQ ID NO: 93) transfected into CHO cells. However, there is a hierarchy, a scale of binding affinities, with clone 26H10 exhibiting the highest affinity, which is approximately 2.5 and 6 times higher than clones 8B7 and 25F7, respectively. Differences in the binding order (hierarchy) between the two cell types may reflect differences between the amino acid sequences of the rhesus and cynomolgus LAG-3 proteins.

Table 3. Binding of anti-LAG-3 antibodies to monkey LAG-3

Antibody EC% (nM) Activated Cyno CD4 ⁺ T cells ECgp (nM) CHO-Rhesus LAG3 26N10 5.19 4.684 25F7 14.18 22.72 8B7 30.45 10.01

4. Binding of mouse LAG-3 antigen

To determine whether antibodies cross-react with murine LAG-3, flow cytometry studies similar to those described above were performed using as target cells a murine T-cell hybridoma line (3A9) transfected to express murine LAG-3. 3 on the cell surface, followed by FACS analysis to detect antibody binding. The results show that, unlike the control anti-mouse LAG-3 antibody, which produces intense staining, none of the human antibodies 25E3, 25F7, 8B7 or 26H10 show binding above background levels to mouse LAG-3, i.e. none of these antibodies cross-react with mouse LAG-3.

B. Analysis using the BIACORE system

The binding of antibodies 25E3, 25F7, 8B7, 26H10 and 17E5 to recombinant LAG-3 protein was studied using the BIAcore™ system using the capture method. Each of the antibodies, 25E3, 25F7, 8B7, 26H10 and 17E5, was immobilized (captured) using anti- _CH1 antibody (anti- _CH1 ), an antibody reagent specific for human heavy chain constant region 1 (Zymed, Clone HP6045, Initial concentration 1.0 mg/ml). Anti-C _H 1 is applied to a CM5 chip (BR-1000-14, for research work) with a high density (9700-11500RU (relative units)). Application is carried out using the usual immobilization method recommended by the manufacturer. Purified antibody 25E3, 25F7, 8B7, 26H10 or 17E5, with a concentration in the range of 0.5-3 μg/ml, is then immobilized (captured) on the anti- _CH 1 coated surface at a flow rate of 10 μl/min for 1 min. A single concentration (20 nM) of recombinant human LAG-3 fusion protein was injected above the antibody (passed through the antibody) for 3 minutes at a flow rate of 25 μg/ml. The antigen is left to dissociate for 7.5 minutes. The surface of the biosensor (chip) is regenerated after each cycle by washing it with 25 μl of a 25 mM NaOH solution, and then 30 μl of HBS-EP. An isotype control is applied to the chip and the results obtained are used to subtract nonspecific binding. All experiments are carried out on a Biacore 3000 surface plasmon resonance instrument using the BIAcore v 3.2 control program. Data analysis using the BiaEvaluation v3.2 program. The results are shown below in table. 4. The results of BIAcore analysis for 25E3, 25F7, 8B7, 26H10 and 17E5 confirm the results obtained by flow cytometry that all five antibodies are able to bind to human

- 43 043093

High affinity LAG-3.

Table 4. Binding kinetics of anti-LAG-3 antibody to recombinant human LAG-3

Antibody K" x ΙΟ ⁹ (M) 25EZ 0.09 8B7 0.09 26N10 0.10 25F7 0.47 17E5 4.53

B. Epitope mapping

In the LAG-3 protein, the immunoglobulin-like first domain of the extracellular region contains an exposed extracellular loop having the amino acid sequence: GPPAAAPGHPLAPGPHPAAPSSWGPRPRRY (SEQ ID NO: 79). Peptide scanning was performed to study the binding of antibodies 25E3, 25F7, 8B7 and 26H10 to this region of LAG-3 and to map the epitope bound by each antibody throughout this region. A group of 10 overlapping peptides are prepared, scanned across the entire outer loop sequence, and conjugated to biotin. For ELISA analysis, microtiter plates with pre-immobilized streptavidin (Sigma-Aldrich, Cat # M5432) are used to capture the biotinylated loop peptide conjugates, which are applied in a volume of 100 μl with a concentration of 2 μg/ml, and incubated for 18 hours at 4°C, after whereupon the plates are washed 3 times and blocked at room temperature for 1 hour with blocking buffer (1 x PBS + 10% FBS). The plates are then loaded with 3-fold serial dilutions of human anti-LAG-3 antibodies in blocking buffer based on an initial concentration of 10 μg/ml, and the plates are incubated at room temperature for 2 hours and then washed three times. To detect bound human antibody, HRP-conjugated goat anti-human immunoglobulin kappa light chain antibody (Bethyl Laboratories, Cat #A80-115P) is diluted to a concentration of 1 μg/ml in blocking buffer and placed in assay wells for 1 h and then washed three times. and TMB substrate (eBioscience, Cat #00-4201-56) was applied. Optical density was determined at a wavelength of 650 nm on a Spectramax 340PC spectrophotometer (Molecular Dynamics, Inc.).

The results of peptide scanning are presented in table. 5.

Table 5. Binding of anti-LAG antibodies to peptide scanning of the extracellular loop ___________________________(Extra Loop) LAG-3 _______,

Peptide scanning of LAG-3 outer loop SEQ GPPAAAPGHPLAPGPHPAAPSSWGPRPRRY 79 25EZ 8B7 25F7 26N10 GPPAAAPGHPLA 80 - - - - PAAAPGHPLAPG 81 ++ - - - AAPGHPLAPGPH 82 ++ - - - PGHPLAPGPHPA 83 + - - - HPLAPGPHPAAP 84 ± - - - LAPGPHPAAPSS 85 - - - - PGPHPAPSSWG 86 - ++ ++ - PHPAAPSWGPR 87 - ++ ++ - PAAPSSWGPRPR 88 - ++ + - APSSWGPRPRRY 89 - - - -

Based on these results, antibody 25E3 is determined to recognize a region in the extracellular loop containing the amino acid sequence PGHPLAPG (SEQ ID NO: 76), while antibody 25F7 recognizes a region in the extracellular loop containing the amino acid sequence HPAAPSSW (SEQ ID NO: 77), and 8B7 appears to recognize a region in the extracellular loop containing the amino acid sequence PAAPSSWG (SEQ ID NO: 78). In contrast, no binding of the full-length loop peptide or any of the shorter scanning peptides by the 26H10 antibody could be detected.

The regions identified in this study are highlighted in the full-length extracellular loop sequence:

IGPPAAAPGHPLAPGPHPAAPSSWGPRPRRY (SEQ ID NO: 79) 125E3 125F7 |δΒ7

In summary, the peptide scan results indicate that antibodies 25E3, 25F7, and 8B7 bind to distinct, albeit closely spaced, epitopes in human LAG-3.

To study in more detail the binding of these antibodies to the outer loop peptide region, additional ELISA assays were performed. An ELISA assay using human full-length Extra Loop peptide (SEQ ID NO: 79) determined EC ₅₀ values for 25E3, 25F7 and 8B7 binding. In addition, a similar peptide ELISA assay is carried out using full

- 44 043093 the extracellular loop peptide sequence (Extra Loop) of the rhesus monkey LAG-3 protein having the sequence GPPAPAPGHPPAPGHRPAAP YSWGPRPRRY (SEQ ID NO: 90), and EC5 ₀ values for binding are determined for antibodies 25F7 and 8B7. The results are presented below in table. 6. These results confirm that antibodies 25E3, 25F7 and 8B7 are capable of recognizing the peptide region of the human LAG-3 protein. In addition, antibodies 25F7 and 8B7 also bind to the extracellular loop peptide sequence of the rhesus monkey LAG-3 protein, although less well than to the human sequence, which may be caused by species-related sequence divergence of this polypeptide. The results also confirm that the 26H10 antibody is unable to recognize the extracellular loop peptide of the LAG-3 protein.

Table 6. Binding of anti-LAG-3 antibodies to the extracellular loop peptide of the human LAG-3 protein _________________________ and rhesus monkeys_________________________

Antibody ECgn (nM) human extracellular loop ECgn (nM) extracellular loop of rhesus macaques 25EZ 0.55 Didn't check 25F7 0.29-0.95 13.09 8B7 0.28-E35 0.60 26N10 No binding No binding

Example 4: Inhibition of LAG-3 Binding to MHC Class II Molecules by Anti-LAG-3 Antibodies

To test the ability of anti-LAG-3 antibodies to inhibit the binding of LAG-3 to MHC Class II molecules, an in vitro binding assay was performed in which a LAG-3 fusion protein containing the extracellular domain of human LAG-3 fused to a mouse Fc (hLAG- 3-mIg), reacts with Daudi cells expressing molecules of human major histocompatibility complex (MHC, MHC) Class II antigens.

To test antibody inhibition of LAG-3 binding to MHC Class II, serial dilutions of antibodies 25E3, 25F7, 8B7 and 26H10 are made starting at a concentration of 20 μg/ml in PFAE buffer, and to these serial dilutions 1 μg/ml hLAG-3 fusion protein is added -mIg. This mixture is incubated for 20 min at room temperature and then added to 2x105 Daudi cells washed in 1x PFAE. The mixture is incubated at 4°C for 30 minutes. Cells are pelleted by centrifugation (3 min, 400xg), washed once with 1x PFAE buffer and pelleted by centrifugation again, and the binding of hLAG-3-mIg to Daudi cells is determined using a PE-labeled secondary antibody (reagent) against mIgG Fcy. LAG-3-mIg binding assay was performed on a FACScalibur flow cytometer (BD Bioscience). The results are presented below in table. 7, which provides IC50 values in nM.

Table 7. Inhibition of LAG-3 binding to MHC Class II by anti-LAG-3 antibodies

Antibody ICgo (nM) 25EZ 0.S-6.7S 25F7 0.12-0.92 8B7 0.19-0.95 26N10 0.10

The results show that all four antibodies effectively inhibit the binding of LAG-3 to Class II MHC antigen molecules, with IC ₅₀ values for 25F7, 8B7 and 26H10 approximately 713 times less than for 25E3.

Example 5 Stimulation of Antigen-Specific T-Cell Response with Anti-LAG-3 mAbs

To test the ability of anti-LAG-3 antibodies to stimulate antigen-specific T cell responses, a 3A9 T cell peptide stimulation assay is used (see, for example, Workman et al. (2003) J. Immunol. 169: 5392-5395; Workman et al. (2002) ) Eur J Immunol 32: 2255-2263).

In this assay, mouse T cell hybridomas, 3A9, specific for the HEL _48-62 peptide, are used as the immunoreactive T cell.

An immunoreactive 3A9 T cell is transduced with a retrovirus so that it expresses either human LAG-3 or murine LAG-3 on its surface. The antigen presenting cell (APC) used to present HEL _48-62 peptide antigen to 3A9 cells is the MHC Class II positive mouse cell line LK35.2. A separate study determined that the human LAG-3 fusion protein was able to bind to mouse MHC Class II molecules, thereby validating the use of LK35.2 murine APCs in this assay. Antigen-specific stimulation of 3A9 cells is indicated by the production of interleukin-2 (IL-2), the secretion of which is quantified using ELISA (mouse IL-2 OptEIA kit, BD Bioscience, Cat #555148 according to the manufacturer's recommendations).

Ectopic expression of human or murine LAG-3 on 3A9 T cells, in the absence of any antibody, has an inhibitory effect on antigen-specific responses when the transfected T cells are incubated with LK35.2 APCs presenting the HEL _48-62 peptide antigen, which indicates an increase in the amount of peptide antigen required to stimulate IL-2 production by 3A9 cells compared to the peptide dose-effect profile of control 3A9 T cells.

- 45 043093

To test the stimulation of an antigen-specific T-cell response, APC (2.5 x 10 ⁴ cells) are first preincubated with antigenic peptide (200 nM) for 30 min at 37°C, and ZA9 T cells (5.0 x ^{10 4} cells expressing mLAG-3, hLAG -3 or control cells) are preincubated with an antibody against hLAG-3 (25E3, 25F7, 8B7, 26H10, 11F2, 17E5), in a three-fold serial dilution, starting from 25 μg/ml) for 15 min at 37°C. ZA9 T cells are then added to antigen-activated APCs and the cultures are incubated for 24 hours at 37°C. The supernatants are then collected and murine IL-2 production is quantified. The results for 3A9 T cells expressing human LAG-3 are presented in table. 8, which gives the values of 1C ₅ o in nM.

Table 8. Stimulation of antigen-specific T-cell response by anti-LAG-3 antibodies

Antibody 3A9-hLAG-3 peptide analysis IQn (nM) 25F7 0.14-E94 26N10 1.45-6.49 8B7 3.25-13.90 25EZ 3.88-70.78 11F2 81.50-240 17E5 No inhibition

The results show that antibodies 25F7, 8B7 and 26H10, and to a lesser extent 25E3, are able to stimulate IL-2 production in the antigen-specific T-cell response assay, whereas antibody 11F2 exhibits minimal inhibitory ability and antibody 17E5 is not active in this assay. Neither antibody quantified IL-2 production by control 3A9 T cells or 3A9 T cells transfected with murine LAG-3 protein, demonstrating the specificity of this effect.

Example 6 Inhibition of Tumor Growth by Anti-LAG-3 Ab, Alone or in Combination

To test the ability of an anti-LAG-3 antibody, alone or in combination with another immunostimulatory antibody, to inhibit tumor cell growth in vivo, two different mouse models of syngeneic tumor grafts were used. The first model uses SalN murine fibrosarcoma cells. The second model uses the mouse colon cancer cell line MSZ 8.

In the first experiment, each mouse (A/J line) is implanted with 2x10 ⁶ SalN fibrosarcoma cells on day 0 and the tumor cells are allowed to grow for seven days. On days 7, 10, and 12 postimplantation, mice were treated with 10 mg/kg of either one anti-LAG-3 mAb (rat anti-mouse LAG-3 antibody, wab C9B7W; eBioscience, Cat. No. 14-2231), one anti-PD- L1 (mouse anti-PDL1 mAb 14D8), anti-LAG-3 and anti-PD-L1 in combination, or an IgG1 isotype control antibody. mAb 14D8 is a rat anti-mouse PD-L1 antibody that is chimeric to contain the mouse IgGl and mouse kappa chain constant regions.

Tumor volumes in mice are measured over 50 days after implantation and mean and median tumor volumes are determined. The average inhibition of tumor growth is calculated (taking into account that the inhibition of the IgGl isotype control antibody is 0%). The results on day 24 after implantation are presented below in table. 9.

Table 9. Average tumor growth inhibition in the SalN tumor model

Day IgGl LAG-3 PD-L1 Combo

I - I 68 74.9 95.8

Thus, the administration of only one antibody against LAG3, or only one antibody against PD-L1 leads to inhibition of tumor growth, while the combination of both antibodies causes significantly more significant inhibition of tumor growth. Regarding the experimental groups, 4 of 10 mice treated with anti-LAG3 antibody alone were tumor-free at the end of the experiment, whereas only 1 of 10 mice treated with the control IgGl antibody were tumor-free at the end of the experiment. Similarly, 4 of 11 mice treated with anti-PD-L1 antibody alone were tumor-free. Treatment of mice with a combination of anti-LAG3 antibody and anti-PD-L1 antibody results in 9 out of 10 mice being tumor-free; the last mouse is left with a slow-healing tumor that remains small throughout the study.

In two additional studies, experimental mice are implanted with murine MC38 colon cancer cells. In the first experiment, on day 0, each C57B1/6 mouse is implanted with 2x10 ⁶ MC38 cells and on days 7, 10 and 12 after implantation, 200 μg/dose of one anti-LAG-3 antibody (C9B7W mAb), one anti-PD-1 antibody ( 4H2 mAb) or a combination of anti-LAG-3 antibody and anti-PD-1 antibody. As a control antibody, I use an antibody related to the IgGl isotype in an amount of 400 mcg/dose. 4H2 mAb is a rat anti-mouse PD-L1 antibody that is formulated into a chimeric antibody so as to contain constant volume

-46043093 lusty mouse IgG1 and mouse kappa chain.

Mean tumor volume, median tumor volume, and % survival were determined 80 days after implantation. The results show that in this tumor model (MC38), anti-LAG-3 monotherapy shows little or no activity in inhibiting tumor growth, and none of the treated mice survived the experiment. In contrast, significant activity was observed with anti-PD-1 antibody monotherapy, such that 4 out of 10 mice were tumor-free by the end of the experiment. In addition, similar to the results obtained in the Sa1N tumor model, combination therapy with anti-LAG-3 antibody plus anti-PD-1 antibody was more effective than therapy with either antibody alone, with 7 of 8 mice being tumor-free by the end of the experiment.

In the second experiment using the MC38 model, on day 0, each C57B1/6 mouse was implanted with 2x106 MC38 cells, and on days 5, 8, and 11 after implantation, 200 μg/dose of test antibody and/or 400 μg/dose of control IgG antibody were administered as follows:

(i) control IgG1 antibody;

(ii) anti-LAG-3 mAb (C9B7W mAb) together with a control IgG1 antibody;

(iii) anti-PD-1 antibody (4H2) together with a control IgG1 antibody;

(iv) anti-CTLA-4 antibody (mouse anti-mouse CTLA-4 mAb 9D9) together with a control IgG1 antibody;

(v) anti-LAG-3 mAb together with anti-PD-1 mAb or (vi) anti-LAG-3 mAb together with anti-CTLA-4 mAb.

9D9 mAb is a mouse anti-mouse CTLA-4 antibody produced in an endogenous mouse CTLA-4 knockout mouse.

Mean tumor volume, median tumor volume, and % survival were determined 100 days after implantation. The results are similar to those of the first experiment, namely, monotherapy directed against LAG-3 has little or no activity in inhibiting MC38 tumor growth, and none of the treated mice survived the experiment. Anti-CTLA-4 monotherapy also has little or no activity in inhibiting tumor growth of MC38, and none of the treated mice survived the experiment. In contrast, significant activity was observed with anti-PD-1 antibody monotherapy, such that 4 out of 10 mice were tumor-free by the end of the experiment. In addition, again, combination therapy is more effective than monotherapy. In the group of mice treated with a combination of anti-LAG-3 antibody and anti-CTLA-4 antibody, 3 out of 10 were tumor-free at the end of the experiment, and in the group of mice treated with a combination of anti-LAG-3 antibody and anti-PD-1 antibody, 8 out of 10 10 mice tumor disappears by the end of the experiment.

Thus, the in vivo tumor cell graft studies described above indicate that, at least in some tumor models, treatment with a single anti-LAG antibody results in marked inhibition of tumor growth in vivo. In addition, in complex tumor models, combination therapy with an anti-LAG-3 antibody together with either an anti-PD-1 antibody, an anti-PD-L1 antibody, or an anti-CTLA-4 antibody has even greater antitumor activity than monotherapy alone.

Example 7. Stimulation of autoimmunity in NOD mice by inhibition of monoclonal antibodies (mAb) against LAG-3

To test the ability of anti-LAG-3 antibodies to stimulate an immune response indicative of the development of autoimmunity, a NOD mouse model of diabetes was used. NOD mice are known to be predisposed to developing autoimmune diabetes. The progression of diabetes in female NOD mice is monitored by quantifying serum glucose. Thus, the effect of treatment with anti-LAG-3 antibody, alone or in combination with any immunostimulatory antibody, on the development of diabetes in female NOD mice is studied.

Female NOD mice are administered 250 μg/dose of reagent on day 0, day 2, and day 5 as follows:

(i) control IgG1 antibody;

(ii) one anti-LAG-3 mAb (C9B7W mAb);

(iii) one anti-PD-1 antibody (4H2 mAb);

(iv) one anti-CTLA-4 mAb (9D9 mAb);

(v) anti-LAG-3 mAb together with anti-PD-1 mAb or (vi) anti-LAG-3 mAb together with anti-CTLA-4 mAb.

The results show that treatment with anti-LAG-3 antibody alone or anti-PD-1 antibody treatment alone (but not treatment with anti-CTLA 4 antibody alone) increases the number of mice converting to the diabetic phenotype. In addition, combination treatment with anti-LAG-3 plus anti-PD-1, or anti-LAG-3 plus anti-CTLA-4 is even more effective at converting mice into a diabetic phenotype.

Thus, the results indicate that blockade of the interaction of LAG-3 with its receptor is superimposed on a negative immunoregulatory signal, which contributes to higher immunological

- 47 043093 gical activity of NOD mice, and this higher immunological activity in mice treated with LAG-3 can be enhanced by combination administration with either an anti-PD-1 antibody or an anti-CTLA-4 antibody.

Example 8 Immunohistochemical study using anti-LAG-3 mAbs

In this experiment, fluorescently labeled human anti-LAG-3 antibodies are used in immunohistochemical studies. The following FITC-labeled human antibodies against LAG-3 are used: 25F7-FITC (F:P= 2.9; IgG1 variant); 25F7-G4-FITC (F:P=2.7; IgG4 variant); 8B7FITC (F:P=2.6) and 26H10-FITC (F:P=3.4). A panel of lymphoid tissues, specifically tonsils (two samples), spleen (two samples) and thymus (two samples), along with pituitary tissue (four samples) are studied. CHO cells transfected with LAG-3 were also used as a control. Cryostat sections fixed in acetone are used. Sections are stained with FITC-labeled anti-LAG-3 antibody (0.2-5 μg/ml), followed by rabbit anti-FITC antibody as a bridging antibody, and then imaged using the EnVision™+System Rabbit Antibody Imaging Kit Kit (Dako USA, Carpinteria, CA). The results are given in table. 10.

Table 10. Immunohistochemical studies. using mAbs against LAG-3

Textile 25F7-FITC 25F7-G4-FITC 8В7-FITC 26Н10-FITC CHO/LAG-3 + + + + Cells (strong) (strong) (strong) (strong) Tonsils + + + + (n=2) (strong; rare in some LCs, 2/2) (strong; rare in some LCs, 2/2) (strong; rare in some LCs, 2/2) (strong; rare in some LCs, 2/2) Spleen + + + + (n=2) (very weak, mainly in the red pulp, 2/2) (very weak, mainly in the red pulp, 2/2) (very weak, mainly in the red pulp, 2/2) (very weak, mainly in the red pulp, 2/2) Thymus + + + + (n=2) Pituitary (strong; very rare in some LCs, 1/2) + (strong; very rare in some LCs, 1/2) + (strong; very rare in selected LCs, 1/2) (strong; very rare in some LCs, 1/2) + (n=4) (strong; episodic in the adenohypophysis, 3/4) (strong; episodic in the adenohypophysis, 3/4) (strong; episodic in the adenohypophysis, 3/4; mild, rare, 1/4)

LC=lymphocyte;

+=positive staining;

-=negative staining

As expected, LAG-3 expression was detected in a panel of lymphoid tissue. In addition, two of the three antibodies studied against LAG-3, 25F7 (IgG1 and IgG4 variants) and 26H10, exhibit retention in pituitary tissue, whereas one of the antibodies studied, 8B7, does not exhibit such retention in pituitary tissue. Thus, immunohistochemical studies allow the identification of two subpopulations of anti-LAG-3 antibodies, with one subpopulation retained in the pituitary tissue and the other subpopulation not retained in the pituitary tissue.

- 48 043093

Summary list of sequences

SEQ ID NO: Subsequence SEQ ID NO: Subsequence 1 V _H CDR1 aa 25F7 49 V _Hnt 25F7 2 V _H CDR1 a. A. 26N10 50 V _Hnt 26H10 3 V _H CDR1 a.a. 25EZ 51 V _Hnt 25E3 4 V _H CDR1 a.a. 8B7 52 V _Hnt 8B7 5 V _H CDR1 a.a. 11F2 53 V _Hnt 11F2 6 V _H CDR1 a.a. 17E5 54 V _Hnt 17E5 7 V _H CDR2 aa 25F7 55 _VKnt 25F7 8 V _H CDR2 a. A. 26N10 56 _VKnt 26H10 9 V _H CDR2 a.a. 25EZ 57 V _Knt 25E3 10 V _H CDR2 a. A. 8B7 58 _VKnt 8B7 AND V _H CDR2 a.a. 11F2 59 _VKnt 11F2 12 V _H CDR2 a.a. 17E5 60 _VKnt 17E5 13 V _H CDR3 aa 25F7 61 V _H 4-34 germline aa 14 V _H CDR3 a. A. 26N10 62 Vh JH5b germline a. a. 15 PVGW 63 Vk L6 germline a. a. 16 V _H CDR3 a.a. 8B7 64 Vk JK2 germline a. a. 17 V _H CDR3 aa 11F2 65 V _H 3-33 germline aa 18 V _H CDR3 a.a. 17E5 66 V _H JH6b germline aa 67 V _k A 27 germline aa 19 V _K CDR1 aa 25F7 68 Vk JK3 germline a. a. 20 V _K CDR1 a. A. 26N10 69 V _H 3-20 germline aa 21 V _K CDR1 a.a. 25EZ 22 V _K CDR1 a. A. 8B7 70 Vh JH4b germline a. a. 23 V _K CDR1 aa 11F2 71 Vk L-18 germline a. a. 24 V _K CDR1 a. A. 17E5 72 Vk JK4 germline a. a.

-

Claims (31)

73 V _H 1-24 germline a. A. 25 V _K CDR2 aa 25F7 74 Vk JK1 germline a. A. 26 V _K CDR2 a. A. 26H10 75 V _k JK5 germline a. A. 27 V _K CDR2 a.a. 25EZ 28 V _K CDR2 a.a. 8В7 76 PGHPLAPG 29 V _K CDR2 a.a. 11F2 77 HPAAPSSW 30 V _K CDR2 a.a. 17E5 78 PAAPSSWG 79 GPPAAAPGHPLAPGPHPAAPSSW GPRPRRY 31 V _K CDR3 aa 25F7 80 GPPAAAPGHPLA 32 V _K CDR3 a. A. 26Н10 81 PAAAPGHPLAPG 33 V _K CDR3 a.a. 25ЕЗ 82 AAPGHPLAPGPH 34 V _K CDR3 a.a. 8В7 83 PGHPLAPGPHPA 35 V _K CDR3 a.a. 11F2 84 HPLAPGPHPAAP 36 V _K CDR3 a.a. 17E5 85 LAPGPHPAAPSS 86 PGPHPAPSSWG 37 V _H aa 25F7 87 PHPAAPSWGPR 38 V _H aa 26Н10 88 PAAPSSWGPRPR 39 V _H a. A. 25ЕЗ 89 APSSWGPRPRRY 40 V _H a.a. 8В7 90 GPPAPAPGHPPAPGHRP AAPYSWGPRPRRY 41 V _H a.a. 11F2 42 V _H a.a. 17E5 91 atgtgggaggctcagttcctg 92 gtcagagctgctccggctc 43 V _K aa 25F7 93 LAG-3 rhesus macaque clone ra23-5 a.a. 44 V _K a. A. 26Н10 94 LAG-3 rhesus monkey (ХМ_001108923) 45 V _K a. A. 25EZ 46 V _K a.a. 8B7 47 V _K a.a. 11F2 48 V _K a.a. 17E5 aa=amino acid; p.t.=nucleotide The term comprising, as used in this specification and claims, means consisting of at least in part. When interpreting sentences in this specification and claims that include the term containing, features other than those implied by that term may also be present in each sentence. Related terms such as contains and consists of should be interpreted in a similar way. In this description, where reference has been made to patent descriptions, other external documents, or other sources of information, generally for the purpose of providing context for a discussion of features of the invention. Unless specifically stated otherwise, reference to such external documents should not be construed as an admission that such documents, or such sources of information, are, in any jurisdiction, prior art, or part of the generally accepted knowledge in the art. CLAIM 1. A method of treating cancer in a subject to be treated, comprising administering to the subject an isolated human monoclonal antibody or an antigen-binding region thereof that binds human lymphocyte activation gene-3 (LAG-3), wherein the antibody comprises heavy and light chain variable regions containing amino acid sequences that are at least 95% homologous to SEQ ID NO: 37 and 43, respectively. 2. The method of claim 1, wherein the antibody or antigen binding region thereof that binds human LAG-3 comprises heavy and light chain variable regions containing amino acid sequences that are at least 98% homologous to the sequences of SEQ ID NOs: 37 and 43 respectively. 3. The method according to claim 1 or 2, wherein the antibody or antigen-binding region thereof that binds human LAG-3 contains heavy and light chain variable regions containing amino acids - 50 043093 lot sequences that are at least 99% homologous to the sequences of SEQ ID NO: 37 and 43, respectively. 4. The method according to any one of claims 1 to 3, wherein the antibody or antigen binding region thereof that binds human LAG-3 comprises the heavy chain variable region sequences CDR1, CDR2 and CDR3 set forth in SEQ ID NO: 1, 7 and 13 respectively, and the light chain variable region sequences CDR1, CDR2, CDR3 shown in SEQ ID NO: 19, 25 and 31, respectively. 5. The method according to any one of claims 1 to 4, wherein the antibody or antigen-binding region thereof that binds human LAG-3 contains a heavy chain variable region containing the amino acid sequence of SEQ ID NO: 37, and a light chain variable region containing the amino acid sequence SEQ ID NO: 43. 6. Method according to any one of claims 1 to 5, wherein the antibody is a full-length antibody. 7. The method according to claim 6, where the full-length antibody is an IgG1, IgG2 or IgG4 isotype. 8. The method according to any one of claims 1 to 7, wherein the antibody or antigen-binding portion thereof that binds human LAG-3 is a chimeric, humanized or human antibody. 9. The method according to any one of claims 1 to 8, characterized in that the antibody or antigen-binding portion thereof that binds human LAG-3 exhibits at least one of the following properties: a) stimulates an antigen-specific T-cell response; b) stimulates the production of interleukin-2 (IL-2) in an antigen-specific T-cell response; c) inhibits the binding of LAG-3 to molecules of the major histocompatibility complex (MHC, MHC) class II; d) stimulates an antitumor response; e) inhibits the growth of tumor cells in vivo; f) binds to human LAG-3 with a K _D of 1x10 ^-7 M or less; g) binds to monkey LAG-3; h) does not bind to mouse LAG-3. 10. The method of claim 9, wherein the antibody or antigen-binding portion thereof that binds human LAG-3 exhibits at least two properties (a)-(h). 11. The method of claim 9, wherein the antibody or antigen-binding portion thereof that binds human LAG-3 exhibits all of the properties (a)-(h). 12. The method according to any one of claims 1 to 11, characterized in that the antibody or antigen-binding portion binds human LAG-3 protein with a K _D of 1x10' ⁹ M or less. 13. The method according to any one of claims 1 to 12, characterized in that the antibody or its antigen-binding portion binds the human LAG-3 protein with a K _D of 5x10 ^-10 M or less. 14. The method according to any one of claims 1 to 13, comprising administering to the subject at least one additional immunostimulating antibody or antigen-binding site thereof. 15. The method of claim 14, wherein the at least one additional immunostimulatory antibody or antigen binding site thereof is an anti-PD-1 antibody, an anti-PD-L1 antibody and/or an anti-CTLA-4 antibody or an antigen binding site thereof. 16. The method of claim 15, wherein the at least one additional immunostimulatory antibody or antigen binding site thereof is an anti-PD-1 antibody or antigen binding site thereof. 17. The method of claim 16, wherein the anti-PD-1 antibody or antigen-binding region comprises the CDR1, CDR2 and CDR3 regions of the heavy and light chain of the 5C4 antibody. 18. The method of claim 16 or 17, wherein the anti-PD-1 antibody or antigen-binding region thereof comprises the heavy and light chain variable regions of the 5C4 antibody. 19. The method according to any one of claims 14 to 18, wherein the at least one additional immunostimulating antibody is a full-length antibody. 20. The method according to any one of claims 14-19, where the full-length antibody is an IgG1, IgG2 or IgG4 isotype. 21. The method according to any one of claims 14 to 20, wherein the at least one additional immunostimulatory antibody or antigen binding site thereof is a chimeric, humanized or human antibody. 22. The method according to any one of claims 14 to 21, wherein the antibody that binds human LAG-3 and the at least one immunostimulatory antibody is a full-length IgG4 isotype human antibody. 23. The method according to any one of claims 14 to 22, wherein an antibody or antigen-binding portion thereof that binds human LAG-3 and at least one immunostimulatory antibody or - 51 043093 antigen-binding site is located in a bispecific type molecule. 24. Method according to any one of paragraphs. 14-22, wherein an antibody or antigen-binding portion thereof that binds human LAG-3 and at least one immunostimulatory antibody or antigen-binding portion thereof are coadministered as separate compositions with each antibody in a pharmaceutically acceptable carrier. 25. Method according to any one of paragraphs. 14-22, wherein an antibody or antigen-binding portion thereof that binds human LAG-3 and at least one immunostimulatory antibody or antigen-binding portion thereof are administered simultaneously as a single composition in a pharmaceutically acceptable carrier. 26. Method according to any one of paragraphs. 14-22, wherein an antibody or antigen-binding portion thereof that binds human LAG-3 and at least one immunostimulatory antibody or antigen-binding portion thereof are administered sequentially. 27. The method according to any one of claims 1 to 26, further comprising administering a chemotherapeutic agent. 28. The method according to any one of claims 1 to 27, further comprising administering an angiogenesis inhibitor. 29. The method according to any one of claims 1 to 28, wherein the cancer is melanoma, metastatic malignant melanoma, kidney cancer, clear cell carcinoma, prostate cancer, hormone-resistant prostate adenocarcinoma, breast cancer, colon cancer, lung cancer or non-small cell cancer lung cancer. 30. The method according to any one of claims 1 to 29, wherein the cancer is a refractory or recurrent malignant tumor or metastatic cancer. 31. The method according to any one of claims 1 to 30, wherein the subject is a human. Eurasian Patent Organization, EAPO Russia, 109012, Moscow, Maly Cherkassky lane, 2