Anti-LAG-3 antibodies

Abstract

Antibodies, or antigen-binding fragments thereof, that bind to Lymphocyte-activation gene-3 (LAG-3) are described, in particular antibodies, or antigen-binding fragments thereof, that are agonists of LAG-3. The antibodies bind to LAG-3 and inhibit antigen-induced CD4.sup.+ and/or CD8.sup.+ T cell proliferation, or antigen-induced CD4.sup.+ and/or CD8.sup.+ T cell activation. The antibodies may be used as medicaments, in particular for the treatment of conditions associated with proliferation and/or activation of CD4.sup.+ and/or CD8.sup.+ T cells, such as inflammatory and autoimmune disorders.

Claims

1. An isolated humanized antibody, or antigen-binding fragment thereof, that binds to lymphocyte activation gene-3 (LAG-3), comprising: a) an antibody VH region comprising a VH CDR1, a VH CDR2, and a VH CDR3, wherein the VH CDR1 has an amino acid sequence of SEQ ID NO:1, the VH CDR2 has an amino acid sequence of SEQ ID NO:2, and the VH CDR3 has an amino acid sequence of SEQ ID NO:3; and an antibody VL region comprising a VL CDR1, a VL CDR2, and a VL CDR3, wherein the VL CDR1 has an amino acid sequence of SEQ ID NO:4, the VL CDR2 has an amino acid sequence of SEQ ID NO:5, and the VL CDR3 has an amino acid sequence of SEQ ID NO:6; or b) an antibody VH region comprising a VH CDR1, a VH CDR2, and a VH CDR3, wherein the VH CDR1 has an amino acid sequence of SEQ ID NO:21, the VH CDR2 has an amino acid sequence of SEQ ID NO:22, and the VH CDR3 has an amino acid sequence of SEQ ID NO:23; and an antibody VL region comprising a VL CDR1, a VL CDR2, and a VL CDR3, wherein the VL CDR1 has an amino acid sequence of SEQ ID NO:24, the VL CDR2 has an amino acid sequence of SEQ ID NO:25, and the VL CDR3 has an amino acid sequence of SEQ ID NO:26; wherein: the antibody or antigen binding fragment thereof comprises a humanized light chain framework region which comprises: a VL framework region 1 (VL FR1) of SEQ ID NO: 68; a VL FR2 of SEQ ID NO: 69; a VL FR3 of SEQ ID NO: 70; and a VL FR4 of SEQ ID NO: 71; a VL framework region 1 (VL FR1) of SEQ ID NO: 72; a VL FR2 of SEQ ID NO: 73; a VL FR3 of SEQ ID NO: 74; and a VL FR4 of SEQ ID NO: 75; a VL framework region 1 (VL FR1) of SEQ ID NO: 76; a VL FR2 of SEQ ID NO: 77; a VL FR3 of SEQ ID NO: 78; and a VL FR4 of SEQ ID NO: 79; or a VL framework region 1 (VL FR1) of SEQ ID NO: 80; a VL FR2 of SEQ ID NO: 81; a VL FR3 of SEQ ID NO: 82; and a VL FR4 of SEQ ID NO: 83; and wherein the antibody or antigen binding fragment thereof comprises a humanized heavy chain framework region which comprises: a VH framework region I (VH FR1) of SEQ ID NO: 52; a VH FR2 of SEQ ID NO: 53; a VH FR3 of SEQ ID NO: 54; and a VH FR4 of SEQ ID NO: 55; a VH framework region 1 (VH FR1) of SEQ ID NO: 56; a VH FR2 of SEQ ID NO: 57; a VH FR3 of SEQ ID NO: 58; and a VH FR4 of SEQ ID NO: 59; a VH framework region 1 (VH FR1) of SEQ ID NO: 60; a VH FR2 of SEQ ID NO: 61; a VH FR3 of SEQ ID NO: 62; and a VH FR4 of SEQ ID NO: 63; or a VH framework region 1 (VH FR1) of SEQ ID NO: 64; a VH FR2 of SEQ ID NO: 65; a VH FR3 of SEQ ID NO: 66; and a VH FR4 of SEQ ID NO: 67.

2. An isolated humanized antibody, or antigen-binding fragment thereof, according to claim 1, which is a humanized monoclonal antibody, or antigen-binding fragment thereof.

3. An isolated humanized antibody, or antigen-binding fragment thereof, according to claim 1, which comprises an antibody VL region comprising: a VL FR1 having an amino acid sequence of SEQ ID NO: 68; a VL CDR1 having an amino acid sequence of SEQ ID NO: 4; a VL FR2 having an amino acid sequence of SEQ ID NO: 69; a VL CDR2 having an amino acid sequence of SEQ ID NO: 5; a VL FR3 having an amino acid sequence of SEQ ID NO: 70; a VL CDR3 having an amino acid sequence of SEQ ID NO: 6; and a VL FR4 having an amino acid sequence of SEQ ID NO: 71; a VL FR1 having an amino acid sequence of SEQ ID NO: 68; a VL CDR1 having an amino acid sequence of SEQ ID NO: 24; a VL FR2 having an amino acid sequence of SEQ ID NO: 69; a VL CDR2 having an amino acid sequence of SEQ ID NO: 25; a VL FR3 having an amino acid sequence of SEQ ID NO: 70; a VL CDR3 having an amino acid sequence of SEQ ID NO: 26; and a VL FR4 having an amino acid sequence of SEQ ID NO: 71; a VL FR1 having an amino acid sequence of SEQ ID NO: 72; a VL CDR1 having an amino acid sequence of SEQ ID NO: 4; a VL FR2 having an amino acid sequence of SEQ ID NO: 73; a VL CDR2 having an amino acid sequence of SEQ ID NO: 5; a VL FR3 having an amino acid sequence of SEQ ID NO: 74; a VL CDR3 having an amino acid sequence of SEQ ID NO: 6; and a VL FR4 having an amino acid sequence of SEQ ID NO: 75; a VL FR1 having an amino acid sequence of SEQ ID NO: 72; a VL CDR1 having an amino acid sequence of SEQ ID NO: 24; a VL FR2 having an amino acid sequence of SEQ ID NO: 73; a VL CDR2 having an amino acid sequence of SEQ ID NO: 25; a VL FR3 having an amino acid sequence of SEQ ID NO: 74; a VL CDR3 having an amino acid sequence of SEQ ID NO: 26; and a VL FR4 having an amino acid sequence of SEQ ID NO: 75; a VL FR1 having an amino acid sequence of SEQ ID NO: 76; a VL CDR1 having an amino acid sequence of SEQ ID NO: 4; a VL FR2 having an amino acid sequence of SEQ ID NO: 77; a VL CDR2 having an amino acid sequence of SEQ ID NO: 5; a VL FR3 having an amino acid sequence of SEQ ID NO: 78; a VL CDR3 having an amino acid sequence of SEQ ID NO: 6; and a VL FR4 having an amino acid sequence of SEQ ID NO: 79; a VL FR1 having an amino acid sequence of SEQ ID NO: 76; a VL CDR1 having an amino acid sequence of SEQ ID NO: 24; a VL FR2 having an amino acid sequence of SEQ ID NO: 77; a VL CDR2 having an amino acid sequence of SEQ ID NO: 25; a VL FR3 having an amino acid sequence of SEQ ID NO: 78; a VL CDR3 having an amino acid sequence of SEQ ID NO: 26; and a VL FR4 having an amino acid sequence of SEQ ID NO: 79; a VL FR1 having an amino acid sequence of SEQ ID NO: 80; a VL CDR1 having an amino acid sequence of SEQ ID NO: 4; a VL FR2 having an amino acid sequence of SEQ ID NO: 81; a VL CDR2 having an amino acid sequence of SEQ ID NO: 5; a VL FR3 having an amino acid sequence of SEQ ID NO: 82; a VL CDR3 having an amino acid sequence of SEQ ID NO: 6; and a VL FR4 having an amino acid sequence of SEQ ID NO: 83; or a VL FR1 having an amino acid sequence of SEQ ID NO: 80; a VL CDR1 having an amino acid sequence of SEQ ID NO: 24; a VL FR2 having an amino acid sequence of SEQ ID NO: 81; a VL CDR2 having an amino acid sequence of SEQ ID NO: 25; a VL FR3 having an amino acid sequence of SEQ ID NO: 82; a VL CDR3 having an amino acid sequence of SEQ ID NO: 26; and a VL FR4 having an amino acid sequence of SEQ ID NO: 83.

4. An isolated humanized antibody, or antigen-binding fragment thereof, according to claim 1, which comprises an antibody VH region comprising: a VH FR1 having an amino acid sequence of SEQ ID NO: 52; a VH CDR1 having an amino acid sequence of SEQ ID NO: 1; a VH FR2 having an amino acid sequence of SEQ ID NO: 53; a VH CDR2 having an amino acid sequence of SEQ ID NO: 2; a VH FR3 having an amino acid sequence of SEQ ID NO: 54; a VH CDR3 having an amino acid sequence of SEQ ID NO: 3; and a VH FR4 having an amino acid sequence of SEQ ID NO: 55; a VH FR1 having an amino acid sequence of SEQ ID NO: 52; a VH CDR1 having an amino acid sequence of SEQ ID NO: 21; a VH FR2 having an amino acid sequence of SEQ ID NO: 53; a VH CDR2 having an amino acid sequence of SEQ ID NO: 22; a VH FR3 having an amino acid sequence of SEQ ID NO: 54; a VH CDR3 having an amino acid sequence of SEQ ID NO: 23; and a VH FR4 having an amino acid sequence of SEQ ID NO: 55; a VH FR1 having an amino acid sequence of SEQ ID NO: 56; a VH CDR1 having an amino acid sequence of SEQ ID NO: 1; a VH FR2 having an amino acid sequence of SEQ ID NO: 57; a VH CDR2 having an amino acid sequence of SEQ ID NO: 2; a VH FR3 having an amino acid sequence of SEQ ID NO: 58; a VH CDR3 having an amino acid sequence of SEQ ID NO: 3; and a VH FR4 having an amino acid sequence of SEQ ID NO: 59; a VH FR1 having an amino acid sequence of SEQ ID NO: 56; a VH CDR1 having an amino acid sequence of SEQ ID NO: 21; a VH FR2 having an amino acid sequence of SEQ ID NO: 57; a VH CDR2 having an amino acid sequence of SEQ ID NO: 22; a VH FR3 having an amino acid sequence of SEQ ID NO: 58; a VH CDR3 having an amino acid sequence of SEQ ID NO: 23; and a VH FR4 having an amino acid sequence of SEQ ID NO: 59; a VH FR1 having an amino acid sequence of SEQ ID NO: 60; a VH CDR1 having an amino acid sequence of SEQ ID NO: 1; a VH FR2 having an amino acid sequence of SEQ ID NO: 61; a VH CDR2 having an amino acid sequence of SEQ ID NO: 2; a VH FR3 having an amino acid sequence of SEQ ID NO: 62; a VH CDR3 having an amino acid sequence of SEQ ID NO: 3; and a VH FR4 having an amino acid sequence of SEQ ID NO: 63; a VH FR1 having an amino acid sequence of SEQ ID NO: 60; a VH CDR1 having an amino acid sequence of SEQ ID NO: 21; a VH FR2 having an amino acid sequence of SEQ ID NO: 61; a VH CDR2 having an amino acid sequence of SEQ ID NO: 22; a VH FR3 having an amino acid sequence of SEQ ID NO: 62; a VH CDR3 having an amino acid sequence of SEQ ID NO: 23; and a VH FR4 having an amino acid sequence of SEQ ID NO: 63; a VH FR1 having an amino acid sequence of SEQ ID NO: 64; a VH CDR1 having an amino acid sequence of SEQ ID NO: 1; a VH FR2 having an amino acid sequence of SEQ ID NO: 65; a VH CDR2 having an amino acid sequence of SEQ ID NO: 2; a VH FR3 having an amino acid sequence of SEQ ID NO: 66; a VH CDR3 having an amino acid sequence of SEQ ID NO: 3; and a VH FR4 having an amino acid sequence of SEQ ID NO: 67; or a VH FR1 having an amino acid sequence of SEQ ID NO: 64; a VH CDR1 having an amino acid sequence of SEQ ID NO: 21; a VH FR2 having an amino acid sequence of SEQ ID NO: 65; a VH CDR2 having an amino acid sequence of SEQ ID NO: 22; a VH FR3 having an amino acid sequence of SEQ ID NO: 66; a VH CDR3 having an amino acid sequence of SEQ ID NO: 23; and a VH FR4 having an amino acid sequence of SEQ ID NO: 67.

5. An isolated antibody, or antigen-binding fragment thereof, that binds to lymphocyte activation gene-3 (LAG-3), which comprises: an antibody VH region comprising: a VH FR1 having an amino acid sequence of SEQ ID NO: 64; a VH CDR1 having an amino acid sequence of SEQ ID NO: 1; a VH FR2 having an amino acid sequence of SEQ ID NO: 65; a VH CDR2 having an amino acid sequence of SEQ ID NO: 2; a VH FR3 having an amino acid sequence of SEQ ID NO: 66; a VH CDR3 having an amino acid sequence of SEQ ID NO: 3; and a VH FR4 having an amino acid sequence of SEQ ID NO: 67; and an antibody VL region comprising: a VL FR1 having an amino acid sequence of SEQ ID NO: 76; a VL CDR1 having an amino acid sequence of SEQ ID NO: 4; a VL FR2 having an amino acid sequence of SEQ ID NO: 77; a VL CDR2 having an amino acid sequence of SEQ ID NO: 5; a VL FR3 having an amino acid sequence of SEQ ID NO: 78; a VL CDR3 having an amino acid sequence of SEQ ID NO: 6; and a VL FR4 having an amino acid sequence of SEQ ID NO: 79.

6. An isolated antibody, or antigen-binding fragment thereof, that binds to lymphocyte activation gene-3 (LAG-3), which comprises: an antibody VH region comprising: a VH FR1 having an amino acid sequence of SEQ ID NO: 64; a VH CDR1 having an amino acid sequence of SEQ ID NO: 21; a VH FR2 having an amino acid sequence of SEQ ID NO: 65; a VH CDR2 having an amino acid sequence of SEQ ID NO: 22; a VH FR3 having an amino acid sequence of SEQ ID NO: 66; a VH CDR3 having an amino acid sequence of SEQ ID NO: 23; and a VH FR4 having an amino acid sequence of SEQ ID NO: 67; and an antibody VL region comprising: a VL FR1 having an amino acid sequence of SEQ ID NO: 76; a VL CDR1 having an amino acid sequence of SEQ ID NO: 24; a VL FR2 having an amino acid sequence of SEQ ID NO: 77; a VL CDR2 having an amino acid sequence of SEQ ID NO: 25; a VL FR3 having an amino acid sequence of SEQ ID NO: 78; a VL CDR3 having an amino acid sequence of SEQ ID NO: 26; and a VL FR4 having an amino acid sequence of SEQ ID NO: 79.

7. An isolated humanized antibody, or antigen-binding fragment thereof, according to claim 1, which lacks complement-dependent cytotoxicity (CDC) and antibody-dependent cell-mediated cytotoxicity (ADCC) activity.

Description

(1) Embodiments of the invention are now described, by way of example only, with reference to the accompanying drawings in which:

(2) FIG. 1 shows amino acid sequence (SEQ ID NO:27) of mature human LAG-3 protein. The four extracellular Ig superfamily domains are at amino acid residues: 1-149 (D1); 150-239 (D2); 240-330 (D3); and 331-412 (D4). The amino acid sequence of the extra-loop structure of the D1 domain of human LAG-3 protein is shown underlined in bold (SEQ ID NO:40);

(3) FIG. 2 shows a graphical representation of the V.sub.H CDR loops of monoclonal antibody 13E2 (Lefranc, M.-P. et al., Dev. Comp. Immunol., 27, 55-77 (2003));

(4) FIG. 3 shows the amino acid sequence of the V.sub.H domain of monoclonal antibody 13E2 aligned with an encoding nucleic acid sequence;

(5) FIG. 4 shows a graphical representation of the V.sub.L CDR loops of monoclonal antibody 13E2 (Lefranc, M.-P. et al., Dev. Comp. Immunol., 27, 55-77 (2003));

(6) FIG. 5 shows the amino acid sequence of the V.sub.L domain of monoclonal antibody 13E2 aligned with an encoding nucleic acid sequence;

(7) FIG. 6 shows a graphical representation of the V.sub.H CDR loops of monoclonal antibody 34F4 (Lefranc, M.-P. et al., Dev. Comp. Immunol., 27, 55-77 (2003));

(8) FIG. 7 shows the amino acid sequence of the V.sub.H domain of monoclonal antibody 34F4 aligned with an encoding nucleic acid sequence;

(9) FIG. 8 shows a graphical representation of the V.sub.L CDR loops of monoclonal antibody 34F4 (Lefranc, M.-P. et al., Dev. Comp. Immunol., 27, 55-77 (2003));

(10) FIG. 9 shows the amino acid sequence of the V.sub.L domain of monoclonal antibody 34F4 aligned with an encoding nucleic acid sequence;

(11) FIG. 10 shows the top V-D-J germline BLAST alignments for nucleotide sequence encoding the V.sub.H region of monoclonal antibody 13E2;

(12) FIG. 11 shows the top V-J germline BLAST alignments for nucleotide sequence encoding the V.sub.L region of monoclonal antibody 13E2;

(13) FIG. 12 shows the top V-D-J germline BLAST alignments for nucleotide sequence encoding the V.sub.H region of monoclonal antibody 34F4;

(14) FIG. 13 shows the top V-J germline BLAST alignments for nucleotide sequence encoding the V.sub.L region of monoclonal antibody 34F4;

(15) FIG. 14 shows the results of binding of different concentrations of agonistic anti-LAG-3 monoclonal antibodies 13E2 and 34F4, and antagonistic anti-LAG-3 monoclonal antibody 1764, to LAG-3-transfected Chinese hamster ovary (CHO) cells, compared to an isotype control antibody (mIgG1);

(16) FIG. 15 shows the results of binding of different concentrations of agonistic anti-LAG-3 monoclonal antibodies 13E2 and 34F4, and antagonistic anti-LAG-3 monoclonal antibody 1764, to CD4.sup.+ and CD8.sup.+ primary cells (SEB-stimulated PBMCs), compared to an isotype control antibody (mIgG1), from a healthy donor (donor 1);

(17) FIG. 16A shows the results of inhibition of binding of IMP321 (LAG-3Ig, 1 μg/ml) to MHC class II-positive B cells by different concentrations of agonistic anti-LAG-3 monoclonal antibodies 13E2 and 34F4, and antagonistic anti-LAG-3 monoclonal antibody 1764, compared to an isotype control antibody (mIgG1). FIG. 16B shows the results of inhibition of activation of THP-1 cells by IMP321 (20 ng/ml) in the presence of different concentrations of agonistic anti-LAG-3 monoclonal antibodies 13E2 and 34F4, and antagonistic anti-LAG-3 monoclonal antibody 1764, compared to an isotype control antibody (mIgG1);

(18) FIG. 17(A) illustrates the CMV-induced proliferation profiles of the CD8.sup.+ T cells of one donor in the presence of mIgG1, 17B4, 13E2 or 34F4, analysed by flow cytometry and the gating strategy used for the assay described in Example 10. FIG. 17(B) shows the results of an assay for the inhibition of CD8.sup.+ T cell proliferation by antibodies 13E2, 34F4, and 1764, compared with the isotype control antibody (mIgG1) for the same donor. The baseline proliferation is also shown (No Stim);

(19) FIG. 18 shows the results of an assay for the inhibition of CD4.sup.+ or CD8.sup.+ T cell proliferation by antibodies 13E2 or 34F4, compared with the isotype control antibody (mIgG1), in several different donors;

(20) FIG. 19 shows the effect of different concentrations of agonistic anti-LAG-3 monoclonal antibodies 13E2 and 34F4 on CD8.sup.+ T cell proliferation;

(21) FIG. 20 shows the heavy chain amino acid sequence of a chimeric 13E2-human IgG4 Fc antibody (the chimeric 13E2-human IgG4 Fc heavy chain amino acid sequence is referred to as 13E2IgG4mut in the figure), and the light chain amino acid sequence of a chimeric 13E2-human IgK antibody (the chimeric 13E2-human IgK light chain amino acid sequence is referred to as 13E2IgK in the figure);

(22) FIG. 21 illustrates the different effects on T cells of depleting, antagonist, and agonist anti-LAG-3 antibodies;

(23) FIG. 22 shows an alignment of the variable regions of humanized VH variants 1-4 (VH.sub.1, VH.sub.2, VH.sub.3, and VH.sub.4), and an alignment of the variable regions of humanized VL variants 1-4 (VL.sub.1, VL.sub.2, VL.sub.3, and VL.sub.4), with the corresponding sequence of the original mouse monoclonal antibody 13E2 (13E2 VH and 13E2 VL, respectively). CDR sequences are highlighted in grey. Changes in the humanized framework sequences of the variants, compared with the original mouse sequence, are shown underlined and in bold;

(24) FIG. 23 shows the heavy chain amino acid sequence of a humanized 13E2-human IgG4 Fc antibody (IMP761) aligned with the chimeric 13E2-human IgG4 Fc heavy chain amino acid sequence (13E2IgG4mut) of antibody Chim13E2IgG4. The V.sub.H region is shown in bold, and the Fc region is shown highlighted. Amino acid residues of the humanized IMP761 sequence that differ from corresponding residues of the chimeric 13E2IgG4mut sequence are single underlined. CDR sequences (based on combined IMGT/Kabat CDR sequence identification) are shown double underlined;

(25) FIG. 24 shows the light chain amino acid sequence of a humanized 13E2-human IgK antibody (IMP761) aligned with the chimeric 13E2-human IgK light chain amino acid sequence (13E2IgK) of antibody Chim13E2IgG4. The V.sub.L region is shown in bold, and the IgK region is shown highlighted. Amino acid residues of the humanized IMP761 sequence that differ from corresponding residues of the chimeric 13E2IgK sequence are single underlined. CDR sequences (based on the combined IMGT/Kabat CDR sequence identification) are shown double underlined;

(26) FIG. 25 shows the results of an assay to test binding of the chimeric 13E2-human IgG4 Fc antibody (Chim13E2IgG4) and IMP761 to CHO-LAG-3+ cells;

(27) FIG. 26 shows the results of an assay to test the effect of IMP761 and Chim13E2IgG4 on: (a) antigen-induced CD8.sup.+ T-cell proliferation; and (b) CD25 expression within CD8.sup.+ T cells;

(28) FIG. 27 shows the effect of IMP761 and Chim13E2IgG4 on antigen-induced CD8.sup.+ T-cell proliferation, and CD25 expression, as a plot of: (a) the percentage inhibition of CD8.sup.+ T-cell proliferation; and (b) the percentage inhibition of CD25 expression within CD8.sup.+ T cells, compared to an isotype-matched control;

(29) FIG. 28 shows the effect of different concentrations of IMP761 and Chim13E2IgG4 on antigen-induced CD8.sup.+ T-cell proliferation;

(30) FIG. 29 shows the results of ADCC assays to determine whether IMP761 has cytotoxic activity against LAG-3-expressing cells, using: an ADCC Reporter Bioassay available from Promega (a); and an ADCC assay using IL-2-stimulated PBMCs as effector cells—the results are plotted as the percentage of: CD4.sup.+ and CD8.sup.+ T cells (b); or LAG-3+CD4.sup.+ and LAG-3+CD8.sup.+ T cells (c) in the PBMC population;

(31) FIG. 30 shows the results of CDC assays using rabbit complement to determine whether IMP761 has cytotoxic activity against LAG-3-expressing cells—the results are plotted as the percentage of: (a) CD4.sup.+ and CD8.sup.+ T cells; or (b) LAG-3+CD4.sup.+ and LAG-3+CD8.sup.+ T cells, in the PBMC population; and

(32) FIG. 31 shows the results of an assay to determine whether IMP761 has cytotoxic activity against LAG-3-expressing cells after culturing antigen-stimulated PBMCs with the antibody for 3 days. The results are plotted as: (a) the percentage of CD4.sup.+ and CD8.sup.+ T cells in the PBMC population; and (b) the percentage of LAG-3+ cells in the CD4.sup.+ and CD8.sup.+ T cell subpopulation.

EXAMPLE 1

(33) Generation of 13E2 anti-LAG-3 Monoclonal Antibody

(34) To generate anti-LAG-3 antibodies, 15 Balb/c mice (referred to as mice numbers 1-15 below) were immunised according to the immunisation protocol described below.

(35) Thirteen mice were each immunised with subcutaneous (s.c.) injections of 100 μg of IMP321 (LAG-3Ig), clinical grade lot S017/LC1/041011 (termed “LC1” below): 3 times (mouse no. 12); 4 times (mouse no. 9); 5 times (mice nos. 5 and 14); or 6 times (mice nos. 3, and 11) at Day 0, Day 15, Day 30, Day 50, Day 67 and Day 108. Mice nos. 1, 2, 4, 8, 10, 13 and 15 were immunised up to 4 additional times. In parallel, two mice (mouse no. 6 and mouse no. 7), used as controls, were immunised with 10 μg of LC1 in Complete Freund's Adjuvant (CFA) once at Day 0, and with the same dose of antigen in Incomplete Freund's Adjuvant (IFA) on Day 15, Day 30, Day 50, Day 67 and Day 108.

(36) Twelve days after the 6th immunisation, serum of mice nos. 1, 2, 3, 4, 8, 10, 11, 13 and 15 was drawn (mice nos. 5, 9, 12 and 14 had already been sacrificed) and analyzed in an Enzyme-Linked Immunosorbent Assay (ELISA) using LAG-3 D1-D4 as coated antigen, purified anti-LAG-3 17B4 murine monoclonal antibody as reference, and a goat anti-mouse Ig-HRP as labelled secondary antibody, to determine the concentration of anti-LAG-3 antibodies present in the serum. The results are shown in Table 2 below:

(37) TABLE-US-00004 TABLE 2 [Anti-LAG-3 Mouse No. Ab] μg/ml 1 1049 2 281 3 423 4 193 6 131 7 394 8 242 10 275 11 506 13 188 15 848

(38) After six immunisations, serum from several mice (including mouse no. 3) surprisingly gave better results than serum from the two positive control mice (nos. 6 and 7), even though these mice were immunised with IMP321 in the absence of CFA or IFA as adjuvant. This unconventional technique (i.e. immunisation with IMP321 in PBS, without using CFA or IFA as adjuvant) also gave good results compared with other experiments using LAG-3-expressing CHO cells in IFA (data not shown). Without being bound by theory, it is believed that this may be because IMP321 itself is an adjuvant, i.e. it directly triggers activation and maturation of dendritic cells.

(39) The same serum samples were assessed for their ability to inhibit the binding of IMP321 to its ligand, MHC class II, expressed on Raji B cells. Ten microliters of a solution of 10 μg/ml of Alexa-Fluor488-conjugated IMP321 was preincubated with and without 5 μl of the serum collected from each mouse or naïve serum for 30 min at 4° C. Then Raji cells were added to a final volume of 50 μl and incubated for 30 minutes at 4° C. The cell-bound fluorescence was analyzed by flow cytometry.

(40) Mouse no. 3 was selected because the serum from this mouse showed a high titre (423 μg/ml) in the D1-D4 serum ELISA assay (Table 1), and a strong capacity to inhibit the binding of IMP321 to MHC class II.sup.+ Raji B cells.

(41) Twelve days after the sixth immunisation, mouse no. 3 received an intravenous (i.v.) boost with 10 μg D1-D4 LAG-3 (with no Fc tail) recombinant protein (produced in CHO cells and purified). Three days after the i.v. boost injection, mouse no. 3 was sacrificed and the spleen was removed. The splenocytes were extracted by squeezing the pieces of spleen with a 5 ml syringe rubber plunger in a Petri dish containing complete serum-free DMEM. The splenocytes and Sp2/0 myeloma cells (cultivated in RPMI 1640+10% FCS+2 mM Glutamine+0.5% P/S) were washed in serum-free medium. The two cell types were mixed together in a ratio of 5:1 of splenocytes:myeloma, and then pelleted by centrifugation. The fusion agent (PEG-1500, Polyethylene Glycol solution 50% w/v in PBS, Roche 10783641001, 1 ml per 10.sup.8 cells) was pre-warmed at 37° C. and added to the cell pellet drop by drop. The cells were very gently resuspended, and diluted by doubling the volume after 90 seconds. The cells were further diluted at regular intervals to a final dilution of 1 to 15 over approximately 5 minutes. The cells were then centrifuged and resuspended in medium containing 10% FBS and incubated for approximately one hour at 37° C. The cells (10,000 cells/well) were then plated in 46 96-well plates in complete RPMI containing 10% of Ultralow Ig FBS (Gibco 16250), 2% HAT (Gibco 21060) and supplemented with 10% BM Condimed H1 (a supplement to the culture medium to support the growth of B-cell hybridomas after fusion and during cloning, Roche 11088947001), and cultured until screening.

(42) The screening was performed by cytometry using CHO cells, with membrane-expressed LAG-3, to analyse the binding ability of antibodies present in the supernatants of growing hydridomas, revealed by a FITC-conjugated goat-anti mouse Ig. The positive hybridomas were expanded and rescreened on CHO LAG-3.sup.+ cells, and wild type CHO cells. From this fusion, a total of 632 wells were screened by FACS analysis on LAG-3 expressing CHO cells (a yield of 14%). 4 hybridoma clones were found to stably express anti-LAG-3 antibody, including 13E2. The hybridoma was then subcloned by limiting dilution.

(43) TABLE-US-00005 TABLE 3 Fusion N °5 Splenocytes (millions) 33 Fusion characteristic BM condimed H1, 10,000 cells/well Number of wells to start with 4416 Number of wells with dividing 632 cells that were screened Recovery 14% Number of stable anti LAG-3 4 hybridomas

EXAMPLE 2

(44) Amino Acid Sequence of Variable Region of Antibody 13E2

(45) TABLE-US-00006 V.sub.H Amino Acid Sequence: (SEQ ID NO: 7) QVTLKESGPGILQPSQTLSLTCSFSGFSLSTSGMGLGWIRQPSGKGLEWL THIWWDDIKRYNPDLRSRLTISKDTSSSQIFLKIASVDTADTATYYCARI VEGSYSSSYFDVWGAGTTVTVSS.

(46) SEQ ID NO:7 is the amino acid sequence of the heavy chain variable (V.sub.H) domain of antibody 13E2. The Complementarity Determining Regions (CDRs), as determined by the IMGT numbering system (Lefranc, M.-P. et al., Nucleic Acids Research, 27, 209-212 (1999)), are underlined. The CDRs, as determined by the Kabat numbering system, are shown in bold.

(47) FIG. 2 shows a graphical representation of the V.sub.H CDR loops of monoclonal antibody 13E2 (Lefranc, M.-P. et al., Dev. Comp. Immunol., 27, 55-77 (2003)). Shaded circles (residue nos. 4, 12, 13, 19, 21, 23, 25, 41, 50, 52, 53, 71, 76, 78, 87, 89, 91, 94, 100) are hydrophobic (non-polar) residues in frameworks 1-3 at sites that are hydrophobic in the majority of antibodies. Squares are key residues at the start and end of each CDR. Amino acid residue nos. 23, 41, 89, 104, 118 in the framework are structurally conserved amino acids;

(48) TABLE-US-00007 V.sub.L Amino Acid Sequence: (SEQ ID NO: 8) DIVMTQPHKFMSTSVEDRVTITCKASQDVIFDVAWYQQKPGQSPKLLIYS ASSRVSGVPDRFTGSGSGTDFTFTISSVQAEDLAVYYCQQHYSTPYTFGG GTTLEIK.

(49) SEQ ID NO:8 is the amino acid sequence of the light chain variable (V.sub.L) domain of antibody 13E2. The Complementarity Determining Regions (CDRs), as determined by the IMGT numbering system (Lefranc, M.-P. et al., Nucleic Acids Research, 27, 209-212 (1999)), are underlined. The CDRs, as determined by the Kabat numbering system, are shown in bold.

(50) TABLE-US-00008 TABLE 4 13E2 VH and VL CDR Sequences Antibody CDR-1 CDR-2 CDR-3 13E2 VH GFSLSTSGMG IWWDDIK ARIVEGSYSSSYFDV (IMGT numbering) (SEQ ID NO: 1) (SEQ ID NO: 2) (SEQ ID NO: 3) 13E2 VL QDVIFD SAS QQHYSTPYT (IMGT numbering) (SEQ ID NO: 4) (SEQ ID NO: 5) (SEQ ID NO: 6) 13E2 VH TSGMGLG HIWWDDIKRYNPDLRS IVEGSYSSSYFDV (Kabat numbering) (SEQ ID NO: 21) (SEQ ID NO: 22) (SEQ ID NO: 23) 13E2 VL KASQDVIFDVA SASSRVS QQHYSTPYT (Kabat numbering) (SEQ ID NO: 24) (SEQ ID NO: 25) (SEQ ID NO: 26)

(51) FIG. 4 shows a graphical representation of the V.sub.L CDR loops of monoclonal antibody 13E2 (Lefranc, M.-P. et al., Dev. Comp. Immunol., 27, 55-77 (2003)). Shaded circles (residue nos. 4, 11, 19, 21, 23, 25, 40, 41, 52, 53, 54, 71, 76, 87, 89, 91, 94, 96, 100, 101) are hydrophobic (non-polar) residues in frameworks 1-3 at sites that are hydrophobic in the majority of antibodies. Squares are key residues at the start and end of each CDR. Amino acids residue nos. 23, 41, 89, 104, 118 in the framework are structurally conserved amino acids.

EXAMPLE 3

(52) Nucleic Acid Sequence Encoding the Variable Domains of Antibody 13E2

(53) Nucleic acid sequence encoding the VH domain of monoclonal antibody 13E2 is shown in FIG. 3, and below:

(54) TABLE-US-00009 (SEQ ID NO: 9) CAGGTTACTCTGAAAGAGTCTGGCCCTGGGATATTGCAGCCCTCCCAGAC CCTCAGTCTGACTTGTTCTTTCTCTGGGTTTTCACTGAGCACTTCTGGTA TGGGTCTAGGCTGGATTCGTCAGCCATCAGGGAAGGGTCTGGAGTGGCTG ACACACATTTGGTGGGATGATATCAAGCGCTATAACCCAGACCTGAGGAG CCGACTGACTATCTCCAAGGATACCTCCAGCAGCCAGATTTTCCTCAAGA TCGCCAGTGTGGACACTGCAGATACTGCCACATATTACTGTGCTCGAATA GTGGAGGGTTCATACAGTAGTAGTTACTTCGATGTCTGGGGCGCAGGGAC CACGGTCACCGTCTCCTCAG.

(55) Nucleic acid BLAST alignment shows that nucleic acid sequence encoding the VH domain of monoclonal antibody 13E2 has significant identity to the sequence of the following germline genes: IGHV8-8*01, IGHV8-11*01, IGHV8-12*01, IGHD2-12*01, IGHD1-1*01, IGHJ1*01, IGHJ1*02, IGHJ1*03.

(56) FIG. 10 shows an alignment of nucleic acid sequence encoding the VH domain of the 13E2 antibody with its top germline gene match. FIG. 10 shows that a portion comprising nucleotides 1-301 of the 13E2 VH region (which encompasses heavy chain framework regions FR1, FR2, and FR3) has 92.7% nucleic acid sequence identity with nucleic acid sequence of V gene IGHV8-8*01.

(57) TABLE-US-00010 TABLE 5 Alignment summary between 13E2 VH and IGHV8-8*01 length matches mismatches gaps identity(%) FR1-IMGT 75 75 0 0 100 CDR1-IMGT 30 29 1 0 96.7 FR2-IMGT 51 48 3 0 94.1 CDR2-IMGT 21 18 3 0 85.7 FR3-IMGT 114 99 15 0 86.8 CDR3-IMGT 10 10 0 0 100 (germline) Total 301 279 22 0 92.7

(58) Nucleic acid sequence encoding the VL domain of monoclonal antibody 13E2 is shown in FIG. 5, and below:

(59) TABLE-US-00011 (SEQ ID NO: 10) GACATTGTGATGACCCAGCCTCACAAATTCATGTCCACATCAGTGGAAGA CAGGGTCACCATCACCTGCAAGGCCAGTCAGGATGTGATTTTTGATGTAG CCTGGTATCAACAGAAACCAGGACAATCTCCTAAATTACTGATTTACTCG GCATCCTCCCGGGTCAGTGGAGTCCCTGATCGCTTCACTGGCAGTGGATC TGGGACGGATTTCACTTTCACCATCAGTAGTGTGCAGGCTGAAGACCTGG CAGTTTATTACTGTCAGCAACACTATAGTACTCCGTACACGTTCGGAGGG GGGACCACGCTGGAAATAAAAC.

(60) Nucleic acid BLAST alignment shows that nucleic acid sequence encoding the VL domain of monoclonal antibody 13E2 has significant identity to the sequence of the following germline genes: IGKV6-17*01, IGKV6-25*01, IGKV6-23*01, IGKJ2*01, IGKJ2*03, IGKJ2*02.

(61) FIG. 11 shows an alignment of nucleic acid sequence encoding the VL domain of the 13E2 antibody with its top germline gene match. FIG. 11 shows that a portion comprising nucleotides 1-284 of the 13E2 VL region (which encompasses light chain framework regions FR1, FR2, and FR3) has 94.7% nucleic acid sequence identity with nucleic acid sequence of V gene IGKV6-17*01.

(62) TABLE-US-00012 TABLE 6 Alignment summary between 13E2 VL and IGKV6-17*01 length matches mismatches gaps identity(%) FR1-IMGT 78 74 4 0 94.9 CDR1-IMGT 18 14 4 0 77.8 FR2-IMGT 51 50 1 0 98 CDR2-IMGT 9 9 0 0 100 FR3-IMGT 108 103 5 0 95.4 CDR3-IMGT 20 19 1 0 95 (germline) Total 284 269 15 0 94.7

EXAMPLE 4

(63) Generation of 34F4 Anti-LAG-3 Monoclonal Antibody

(64) 4 Balb/c mice (referred to as mice nos. 1-4 below) were each immunised with four (mice nos. 2 and 4), or five (mice nos. 1 and 3) s.c. injections of 100 μg IMP321 (LAG-3Ig), clinical grade lot S017/LC1/041011 (termed “LC1” below), at Day 0, Day 14, Day 28, Day 43 and Day 70. An additional Balb/c mouse (referred to as mouse no. 5 below) was immunized with three s.c. injections of D1-D4 LAG-3.

(65) Two weeks after the third injection, serum from each mouse was tested for anti-LAG-3 antibody content in an ELISA assay (as described in Example 1). The results are presented in Table 7 below:

(66) TABLE-US-00013 TABLE 7 Mouse No. [Anti-LAG-3 Ab] μg/ml 1 31 2 10 3 13 4 28 5 10

(67) The ability of the serum from each mouse to inhibit binding of IMP321 to the Raji B cells was determined by FACS analysis (as described in Example 1).

(68) As can be seen from Table 7, the antibody titres were low in all mice. None of the sera inhibited the binding of IMP321 to MHC class II.sup.+ Raji B cells after three immunisations. No other bleeding was performed to test serum titres and inhibition capacity. The immunisation process was continued and mouse no. 3, which received five s.c. injections of LC1, was boosted with 10 μg of D1-D4 LAG-3 i.v. at Day 92.

(69) Three days after the i.v. boost injection, 73 million splenocytes from mouse no. 3 were fused with 15 million Sp2/0 myeloma cells following the same procedure described in Example 1. The wells of 40 96-well plates were seeded with approximately 25,500 cells per well, and then cultured with supplementation of the culture medium with 10% BM Condimed Ht 2256 wells were screened by FACS analysis on LAG-3 expressing CHO cells (a yield of 59%). Two stable anti-LAG-3 hybridomas were selected, including 34F4 (see Table 8 below).

(70) TABLE-US-00014 TABLE 8 Fusion N °17 Splenocytes (millions) 73 Fusion characteristic BM condimed H1, 25,000 cells/well Number of wells to start with 3840 Number of wells with dividing 2256 cells that were screened Recovery 59% Number of stable anti LAG-3 2 hybridomas

EXAMPLE 5

(71) Amino Acid Sequence of Variable Region of Antibody 34F4

(72) TABLE-US-00015 V.sub.H Amino Acid Sequence: (SEQ ID NO: 17) QVTLKESGPGILQPSQTLSLTCSFSGFSLNTSGMGVGWIRQPSGKGLEWL THIWWDDVKRYNPALKSRLTISKDTSSSQVFLKIASVDTADTATYYCARI EGDTYYDYYFDYWGQGVTLTVSS.

(73) SEQ ID NO: 17 is the amino acid sequence of the heavy chain variable (V.sub.H) domain of antibody 34F4. The Complementarity Determining Regions (CDRs), as determined by the IMGT numbering system (Lefranc, M.-P. et al., Nucleic Acids Research, 27, 209-212 (1999)), are underlined. The CDRs, as determined by the Kabat numbering system, are shown in bold.

(74) FIG. 6 shows a graphical representation of the V.sub.H CDR loops of monoclonal antibody 34F4 (Lefranc, M.-P. et al., Dev. Comp. Immunol., 27, 55-77 (2003)). Shaded circles (residue nos. 4, 12, 13, 19, 21, 23, 25, 41, 50, 52, 53, 71, 76, 78, 87, 89, 91, 94, 100) are hydrophobic (non-polar) residues in frameworks 1-3 at sites that are hydrophobic in the majority of antibodies. Squares are key residues at the start and end of each CDR. Amino acid residue nos. 23, 41, 89, 104, 118 in the framework are structurally conserved amino acids.

(75) TABLE-US-00016 V.sub.L Amino Acid Sequence: (SEQ ID NO: 18) DIVMTQSHKLMSTSVGDGLSITCKASQDVSIAVVWYQQKPGQSPKLLIYS ASFRYTGVPDRFTGSGSGTDFTFTISSVQAEDLAVYYCQQHYSIPWTFGG GTKLEIK.

(76) SEQ ID NO: 18 is the amino acid sequence of the light chain variable (V.sub.L) domain of antibody 34F4. The Complementarity Determining Regions (CDRs), as determined by the IMGT numbering system (Lefranc, M.-P. et al., Nucleic Acids Research, 27, 209-212 (1999)), are underlined. The CDRs, as determined by the Kabat numbering system, are shown in bold.

(77) TABLE-US-00017 TABLE 9 34F4 VH and VL CDR Sequences Antibody CDR-1 CDR-2 CDR-3 34F4 VH GFSLNTSGMG IWWDDVK ARIEGDTYYDYYFDY (IMGT numbering) (SEQ ID NO: 11) (SEQ ID NO: 12) (SEQ ID NO: 13) 34F4 VL QDVSIA SAS QQHYSIPWT (IMGT numbering) (SEQ ID NO: 14) (SEQ ID NO: 15) (SEQ ID NO: 16) 34F4 VH TSGMGVG HIWWDDVKRYNPALKS IEGDTYYDYYFDY (Kabat numbering) (SEQ ID NO: 31) (SEQ ID NO: 32) (SEQ ID NO: 33) 34F4 VL KASQDVSIAVV SASFRYT QQHYSIPWT (Kabat numbering) (SEQ ID NO: 34) (SEQ ID NO: 35) (SEQ ID NO: 36)

(78) FIG. 8 shows a graphical representation of the V.sub.L CDR loops of monoclonal antibody 34F4 (Lefranc, M.-P. et al., Dev. Comp. Immunol., 27, 55-77 (2003)). Shaded circles (residue nos. 4, 11, 19, 21, 23, 25, 40, 41, 52, 53, 54, 71, 76, 87, 89, 91, 94, 96, 100, 101) are hydrophobic (non-polar) residues in frameworks 1-3 at sites that are hydrophobic in the majority of antibodies. Squares are key residues at the start and end of each CDR. Amino acids residue nos. 23, 41, 89, 104, 118 in the framework are structurally conserved amino acids.

EXAMPLE 6

(79) Nucleic Acid Sequence Encoding the Variable Domains of Antibody 34F4

(80) Nucleic acid sequence encoding the VH domain of monoclonal antibody 34F4 is shown in FIG. 7, and below:

(81) TABLE-US-00018 (SEQ ID NO: 19) CAGGTTACTCTGAAAGAGTCTGGCCCTGGGATATTGCAGCCCTCCCAGAC CCTCAGTCTGACTTGTTCTTTCTCTGGGTTTTCACTGAACACTTCTGGTA TGGGTGTAGGCTGGATTCGTCAGCCATCAGGGAAGGGTCTGGAGTGGCTG ACACACATTTGGTGGGATGATGTCAAGCGCTATAATCCAGCCCTGAAGAG CCGACTGACTATCTCCAAGGATACCTCCAGCAGCCAGGTATTCCTCAAGA TCGCCAGTGTGGACACTGCAGATACTGCCACATACTACTGTGCTCGAATA GAGGGGGATACTTACTACGACTATTACTTTGACTACTGGGGCCAAGGCGT CACTCTCACAGTCTCCTCAG.

(82) Nucleic acid BLAST alignment shows that nucleic acid sequence encoding the VH domain of monoclonal antibody 34F4 has significant identity to the sequence of the following germline genes: IGHV8-8*01, IGHV8-12*01, IGHV8-11*01, IGHD1-1*01, IGHD1-2*01, IGHD2-3*01, IGHJ2*01, IGHJ2*02, IGHJ2*03.

(83) FIG. 12 shows an alignment of nucleic acid sequence encoding the VH domain of the 34F4 antibody with its top germline gene match. FIG. 12 shows that a portion comprising nucleotides 1-301 of the 34F4 VH region (which encompasses heavy chain framework regions FR1, FR2, and FR3) has 94.4% nucleic acid sequence identity with nucleic acid sequence of V gene IGHV8-8*01.

(84) TABLE-US-00019 TABLE 10 Alignment summary between 34F4 VH and IGHV8-8*01 length matches mismatches gaps identity(%) FR1-IMGT 75 75 0 0 100 CDR1-IMGT 30 28 2 0 93.3 FR2-IMGT 51 49 2 0 96.1 CDR2-IMGT 21 19 2 0 90.5 FR3-IMGT 114 103 11 0 90.4 CDR3-IMGT 10 10 0 0 100 (germline) Total 301 284 17 0 94.4

(85) Nucleic acid sequence encoding the VL domain of monoclonal antibody 34F4 is shown in FIG. 9, and below:

(86) TABLE-US-00020 (SEQ ID NO: 20) GACATTGTGATGACCCAGTCTCACAAACTCATGTCCACATCAGTTGGAGA CGGGCTCAGCATCACCTGCAAGGCCAGTCAGGATGTGAGCATTGCTGTAG TCTGGTATCAACAGAAACCAGGACAATCTCCTAAACTGCTGATTTACTCG GCATCCTTCCGGTACACTGGAGTCCCTGATCGCTTCACTGGCAGTGGATC TGGGACGGATTTCACTTTCACCATCAGCAGTGTGCAGGCTGAAGACCTGG CAGTTTATTACTGTCAGCAACATTATAGTATTCCGTGGACGTTCGGTGGA GGCACCAAGCTGGAAATCAAAC.

(87) Nucleic acid BLAST alignment shows that nucleic acid sequence encoding the VL domain of monoclonal antibody 34F4 has significant identity to the sequence of the following germline genes: IGKV6-17*01, IGKV6-25*01, IGKV6-23*01, IGKJ1*01, IGKJ1*02, IGKJ2*01.

(88) FIG. 13 shows an alignment of nucleic acid sequence encoding the VL domain of the 34F4 antibody with its top germline gene match. FIG. 13 shows that a portion comprising nucleotides 1-284 of the 34F4 VL region (which encompasses light chain framework regions FR1, FR2, and FR3) has 94.7% nucleic acid sequence identity with nucleic acid sequence of V gene IGKV6-17*01.

(89) TABLE-US-00021 TABLE 11 Alignment summary between 34F4 VL and IGKV6-17*01 length matches mismatches gaps identity(%) FR1-IMGT 78 74 4 0 94.9 CDR1-IMGT 18 16 2 0 88.9 FR2-IMGT 51 49 2 0 96.1 CDR2-IMGT 9 9 0 0 100 FR3-IMGT 108 107 1 0 99.1 CDR3-IMGT 20 19 1 0 95 (germline) Total 284 274 10 0 96.5

EXAMPLE 7

(90) Binding of the Agonist 13E2 and 34F4 Monoclonal Antibodies to LAG-3.sup.+ Transfected, and Primary Cells, Compared to the 1764 Antagonist Monoclonal Antibody

(91) LAG-3.sup.+-transfected CHO cells, or SEB-stimulated PBMCs from a healthy donor were incubated with anti-LAG-3 monoclonal antibody, or an isotype control (mIgG1) for 30 minutes in PBS, BSA 0.5%, Azide 0.1% at 4° C. Cells were washed, and cell-bound antibody was revealed by a FITC-conjugated goat F(ab′)2-anti-mouse Ig (H+L) (Coulter). The secondary antibody was washed away, and the CHO cells were directly analyzed by flow cytometry. The PBMCs were phenotyped using CD4-PE-Cy7 and CD8-APC-Cy7.

(92) Binding to CHO LAG-3.sup.+ Cells

(93) The results are presented as the mean of fluorescence intensity (MFI) of CHO cells transfected with human LAG-3-encoding plasmid as a function of antibody concentration. The results are shown in Table 12 below, and in FIG. 14.

(94) TABLE-US-00022 TABLE 12 Binding to CHO LAG3+ ng/ml mIgG1 17B4 13E2 34F4 10000 440 6087 6578 6618 2000 251 5477 6904 6411 500 168 4759 6069 4988 100 nd 3170 4583 3653 20 nd 1378 2475 1763 6 nd 549 925 693 1 nd 247 379 329 0.4 nd 179 222 222 0 156 156 156 156 nd = not determined

(95) Based on the results in Table 12, the EC.sub.50 value for binding of each antibody to LAG-3-expressing CHO cells is: 1764: 0.7 nM; 13E2: 0.3 nM; 34F4: 0.5 nM.

(96) The mean EC.sub.50 values from four independent experiments (data not shown) for binding of each antibody to LAG-3-expressing CHO cells is: 1764: 0.7 nM; 13E2: 0.4 nM; 34F4: 0.5 nM.

(97) The mean, from the four independent experiments, of the 13E2 EC.sub.50 value is 2.7 times the mean of the 1764 EC.sub.50 value.

(98) The mean, from the four independent experiments, of the 34F4 EC.sub.50 value is 1.6 times the mean of the 1764 EC.sub.50 value.

(99) Binding to SEB-Stimulated PBMCs

(100) The results are presented as the mean of fluorescence intensity on the CD4+ or CD8+ cells from PBMCs of a donor (Donor 1) stimulated for three days with 0.5 μg/ml SEB as a function of antibody concentration. The results for binding to CD4.sup.+ and CD8.sup.+ cells, for Donor 1, are shown in Table 13 below, and in FIG. 15.

(101) TABLE-US-00023 TABLE 13 Binding on CD4 and CD8 (Donor 1) CD4 CD8 Antibody ng/ml mIgG1 17B4 13E2 34F4 ng/ml mIgG1 17B4 13E2 34F4 10000 116 1598 1947 1787 10000 140 5142 5107 4719 2326 113 1944 1869 1662 2326 137 4826 4850 4455 541 116 1746 1601 1590 541 120 4092 4359 3867 126 109 1239 1491 1384 126 115 2837 3677 3302 29 108 675 1160 1139 29 110 1397 2432 2332 7 110 324 600 621 7 114 606 1131 1183 1.6 127 192 272 282 1.6 116 309 458 446 0.4 106 143 165 169 0.4 112 195 232 288 0 106 106 106 106 0 112 112 112 112

(102) Based on the results given in Table 13, the EC.sub.50 value for binding of each antibody to CD4.sup.+ cells is: 1764: 0.5 nM; 13E2: 0.1 nM; 34F4: 0.1 nM.

(103) The mean EC.sub.50 value from three donors (data not shown) for binding of each antibody to CD4.sup.+ cells is: 1764: 0.8 nM; 13E2: 0.2 nM; 34F4: 0.2 nM.

(104) The mean of the 13E2 and 34F4 CD4.sup.+ EC.sub.50 values from the three donors is 3.8 times the mean of the 1764 EC.sub.50 value.

(105) Based on the results given in Table 13, the EC50 value for binding of each antibody to CD8.sup.+ cells is: 1764: 0.7 nM; 13E2: 0.3 nM; 34F4: 0.2 nM.

(106) The mean EC.sub.50 value from three donors (data not shown) for binding of each antibody to CD8.sup.+ cells is: 1764: 1 nM; 13E2: 0.4 nM; 34F4: 0.5 nM.

(107) The mean of the 13E2 and 34F4 CD8+EC.sub.50 values from the three donors is 2.5 times the mean of the 1764 EC.sub.50 value.

(108) The results show that the 13E2 and 34F4 monoclonal antibodies each bind to CHO cells expressing LAG-3.sup.+, and to CD4.sup.+-T cells and CD8.sup.+-T cells, with a higher affinity than 1764.

(109) Biacore analysis with LAG-3Ig on the chip and 1764 antibody in the running buffer gave the following results:

(110) TABLE-US-00024 ka (1/Ms) 1.09 × 10.sup.6 kd (1/s) 1.32 × 10.sup.−4 KD (M) 1.21 × 10.sup.−10

(111) Biacore analysis with 1764 antibody on the chip and LAG-3Ig in the running buffer gave the following results:

(112) TABLE-US-00025 ka (1/Ms) 2.22 × 10.sup.5 kd (1/s) 8.18 × 10.sup.−4 KA (1/M) 2.71 × 10.sup.8 KD (M) 3.69 × 10.sup.−9

EXAMPLE 8

(113) Inhibition of Binding of IMP321 (LAG-3IG) to MHC Class II-Positive Cells by 13E2 and 34F4

(114) Binding of an IMP321 conjugate (LAG-3Ig-Alexa 488) to MHC class II-positive B cells (Raji cells) was determined following pre-incubation of the conjugate (1 pg/ml at 4° C.) with an anti-LAG-3 monoclonal antibody (13E2, 34F4, or 1764), or an isotype control (mIgG1). Analysis of cell-bound fluorescence was carried out using fluorescence-activated cell sorting (FACS).

(115) The mean fluorescence intensity (MFI) corresponding to the cell-bound LAG-3Ig as a function of antibody concentration is shown in Table 14 below, and in FIG. 16A.

(116) TABLE-US-00026 TABLE 14 MFI of sLAG-3Ig-Alexa 488 Antibody Ratio Ab:IMP321- (ng/ml) mIgG1 17B4 13E2 34F4 Alexa (1 μg/ml) 11000 328 −11 −2 −7   10:1 3300 359 13 20 18   3:1 990 361 35 37 35   1:1 297 373 73 51 46  0.3:1 89 324 214 179 185  0.1:1 27 254 260 207 258 0.03:1 0 284 284 284 284

(117) The results show that binding of IMP321 to Raji cells was inhibited by pre-incubation with each of the LAG-3-specific monoclonal antibodies.

EXAMPLE 9

(118) Inhibition of IMP321 (LAG-3Ig)-Induced Monocyte Activation by 13E2 and 34F4

(119) IMP321 (20 ng/ml) was preincubated with anti-LAG-3 monoclonal antibody 13E2, 34F4, or 17B4, or an isotype control (mIgG1), for 5 minutes at 37° C., before incubation of the mixture with THP-1 cells for 4 hours at 37° C. The amount of CCL4 secretion by the THP-1 cells was used to determine the level of monocyte activation.

(120) The CCL4 concentration (expressed in pg/ml) as a function of Ab concentration is shown in Table 15 below, and in FIG. 16B.

(121) TABLE-US-00027 TABLE 15 Read-out [CCL4] in THP-1 monocytic cells supernatant (pg/ml) Antibody Ratio Ab:IMP321 (ng/ml) mIgG1 17B4 13E2 34F4 (20 ng/ml) 20000 2901 860 97 130 1000:1 2000 2261 947 77 111  100:1 200 2222 860 94 135  10:1 20 2145 1548 224 338   1:1 2 2052 1963 1661 1798  0.10:1 0 2112 2112 2112 2112

(122) The results show that IMP321-induced monocyte activation is inhibited by pre-incubation of IMP321 with the antagonist anti-LAG-3 monoclonal antibody 17B4, and also by pre-incubation with the agonist monoclonal antibodies 13E2 and 34F4.

(123) It was concluded from these results, and from the results in Example 8, that the agonist monoclonal antibodies 13E2 and 34F4, like the antagonist 17B4 monoclonal antibody, interact with or close to the MHC class II binding site of LAG-3, as shown by their ability to block the binding and the activity of LAG-3Ig (IMP321).

EXAMPLE 10

(124) Inhibition of T Cell Proliferation by 13E2 and 34F4 Compared to 17E34

(125) PBMCs from 3 healthy donors (0.2×10.sup.6 cells/well, at 1×10.sup.6/ml in complete RPMI+10% FBS) were labelled with Carboxyfluorescein succinimidyl ester (CFSE) and incubated with a pool of peptides covering the sequence of CMV pp35 in the presence of monoclonal anti-LAG-3 antibody 13E2, 34F4, 1764, or an isotype control (mIgG1) (supra-optimal dose, 300 ng/ml for donor #1 and #2, 100 ng/ml for donor #3).

(126) The T cell response was investigated by measuring the CFSE-based proliferation of CD4.sup.+ or CD8.sup.+ T cells on day 5. The FACS profiles of the CD8.sup.+ T cells of donor #1 in the presence of each different antibody, as well as the gating strategy, are shown in FIG. 17(A). FIG. 17(B) shows the percentage of CD8.sup.+ T cells under each division peak, as a function of cell division, for the same donor. The results for the 3 donors are shown in Table 16 below. The baseline proliferation without antigenic peptides (No stim) was also measured (see FIG. 17(A) lowest panel and Table 16). The CD4.sup.+ T cells of donor #1 did not display any CMV-specific proliferation, so the results for this population are not included.

(127) TABLE-US-00028 TABLE 16 Percentage of CD4.sup.+ and CD8.sup.+ T cells under each division peak Division No. 1 2 3 4 5 6 7 Donor #1 CD8.sup.+ mIgG1 0.68 0.54 1.28 2.20 1.50 1.22 0.78 17B4 0.84 1.02 1.81 2.82 2.22 2.10 1.51 13E2 1.02 0.41 0.61 0.67 0.39 0.36 0.31 34F4 1.16 0.68 1.32 1.22 0.66 0.52 0.27 No Stim 1.77 0.36 0.33 0.52 0.38 0.13 0.04 Donor #2 CD4.sup.+ mIgG1 1.90 2.26 2.67 1.84 0.78 0.20 0.18 17B4 1.87 2.44 2.47 1.72 0.69 0.17 0.14 13E2 1.72 1.89 2.04 1.35 0.64 0.21 0.16 34F4 1.61 1.88 2.01 1.34 0.60 0.16 0.15 No Stim 1.34 1.14 1.48 0.81 0.41 0.17 0.12 Donor #2 CD8.sup.+ mIgG1 1.24 1.72 2.54 3.37 1.28 0.37 0.24 17B4 1.24 1.79 2.57 3.30 1.13 0.37 0.17 13E2 1.27 1.35 1.83 2.14 0.75 0.22 0.14 34F4 1.25 1.47 1.95 2.17 0.86 0.25 0.23 No Stim 0.77 0.92 1.03 1.22 0.33 0.22 0.10 Donor #3 CD4.sup.+ mIgG1 2.93 2.85 4.32 8.18 5.03 2.1 17B4 2.95 2.8 4.51 9.9 6.71 2.81 13E2 3.42 2.62 3.01 4.66 2.53 1.07 No Stim 2.77 2.07 2.79 3.71 1.71 0.5 Donor #3 CD8.sup.+ mIgG1 4.62 1.3 1.41 4.26 4.64 2.6 17B4 4.74 1.63 1.56 4.24 4.56 2.67 13E2 5.1 1.56 1.24 2.81 1.9 0.97 No Stim 4.04 0.44 0.37 0.38 0.11 0.07

(128) The Proliferation Index (PI) (calculated as the sum of: the percentage of CD4.sup.+ or CD8.sup.+ T cells under each division peak, multiplied by the division number) is provided in Table 17. This index emphasises the percentages for cells which have experienced several rounds of division. Table 17 also records the percentage inhibition for each antibody compared with the isotype control (mIgG1) based on the PI values.

(129) TABLE-US-00029 TABLE 17 Effect of antibodies 13E2 and 34F4 on CD4+ and CD8+ T cell proliferation compared to 17B4 % % Donor PI inhibition PI inhibition No. Antibody (CD4) (CD4) (CD8) (CD8) 1 mIgG1 34.64 17B4 53.80 −55 13E2 12.60 64 34F4 19.65 43 No Stim 8.54 2 mIgG1 28.16 36.09 17B4 26.52 6 34.77 4 13E2 22.59 20 24.09 33 34F4 21.76 23 26.13 28 No Stim 15.20 14.25 3 mIgG1 92.06 67.29 17B4 112.09 −22 68.46 −2 13E2 55.40 40 38.50 43 No Stim 41.67 8.52

(130) The results show that monoclonal antibodies 13E2 and 34F4 consistently inhibit the proliferation of CD4.sup.+ and CD8.sup.+ T cells induced by antigenic peptides, while monoclonal antibody 17B4 tends to have a minor positive effect at the tested concentration.

EXAMPLE 11

(131) Inhibition of T Cell Proliferation by 13E2 and 34F4

(132) PBMCs from 12 healthy donors (0.2×10.sup.6 cells/well, at 1×10.sup.6/ml in complete RPMI+10% FBS) were labelled with CFSE and incubated with a pool of peptides covering the sequence of CMV pp35 in the presence of monoclonal anti-LAG-3 antibody 13E2, 34F4, or an isotype control (mIgG1).

(133) The T cell response was investigated by measuring the CFSE-based proliferation of CD4.sup.+ or CD8.sup.+ T cells on day 5. The percentage of CD4.sup.+ or CD8.sup.+ T cells under each division peak was calculated as a function of cell division, using the gating strategy illustrated in FIG. 17(A). The baseline proliferation without antigenic peptides (No stim) was also measured. The CD4.sup.+ T cells of donors #1, #5, and #12 did not display any CMV-specific proliferation, so the results for these samples are not included.

(134) The Proliferation Index (PI) (calculated as the sum of: the percentage of CD4.sup.+ or CD8.sup.+ T cells under each division peak, multiplied by the division number) for each donor is provided in Table 18, and the results are plotted in FIG. 18. Table 18 also records the percentage inhibition for each antibody compared with the isotype control (mIgG1) based on the PI values.

(135) TABLE-US-00030 TABLE 18 Effect of antibodies 13E2 and 34F4 on CD4.sup.+ and CD8.sup.+ T cell proliferation Donor PI % Inhibition PI % Inhibition No. Antibody (CD4) (CD4) (CD8) (CD8) 1 mIgG1 34.64 13E2 12.60 64 34F4 19.65 43 No Stim 8.54 2 mIgG1 28.16 36.09 13E2 22.59 20 24.09 33 34F4 21.76 23 26.13 28 No Stim 15.20 14.25 3 mIgG1 92.06 67.29 13E2 55.40 40 38.50 43 No Stim 41.67 8.52 4 mIgG1 10.10 43.05 13E2 8.83 13 19.27 55 34F4 9.39 7 19.32 55 No Stim 6.14 3.26 5 mIgG1 11.01 13E2 7.00 36 34F4 6.71 39 No Stim 3.08 6 mIgG1 14.56 78.19 13E2 6.30 57 15.56 80 34F4 6.14 58 17.01 78 7 mIgG1 36.25 31.62 13E2 18.90 48 12.95 59 34F4 32.00 12 21.99 30 No Stim 29.59 15.53 8 mIgG1 6.20 10.96 13E2 4.32 30 6.77 38 No Stim 2.60 4.48 9 mIgG1 7.60 10.44 13E2 4.63 39 6.57 37 No Stim 1.28 1.57 10 mIgG1 15.80 75.10 34F4 11.70 26 25.50 66 No Stim 17.00 6.60 11 mIgG1 15.19 19.14 13E2 5.51 64 12.77 33 34F4 4.96 67 1.73 91 12 mIgG1 20.69 13E2 10.71 48 No Stim 10.71 Donor # 1, 2, 4, 5, 6, 7, 8, 9: 300 ng/ml; Donor # 3: 100 ng/ml; Donor # 10, 11, 12: 1000 ng/ml

(136) The mean values from these results are set out in Table 19.

(137) TABLE-US-00031 TABLE 19 Effect of antibodies 13E2 and 34F4 on CD4.sup.+ and CD8.sup.+ T cell proliferation (donor averages) Mean of 7 donors Mean of Mean of 11 Mean of 8 tested with both percentage donors tested donors tested antibodies inhibition with 13E2 with 34F4 13E2 34F4 CD4 39% 36% 40% 33% CD8 48% 55% 52% 52%

(138) The results show that monoclonal anti-LAG-3 antibodies 13E2 and 34F4 inhibit the proliferation of CD4.sup.+ and CD8.sup.+ T cells induced by antigenic peptides. The results suggest that the inhibitory effect of each antibody may be more pronounced for CD8.sup.+ T cells than for CD4.sup.+ T cells. In most of the donors tested, the effects of the 13E2 and 34F4 antibodies were very similar, so the antibodies appear to have comparable activity.

EXAMPLE 12

(139) Dose-Response of Agonist Antibody on CD8.sup.+ T Cell Proliferation

(140) CFSE-labelled PBMCs were stimulated by CMV peptide, as described above, in the presence of various concentrations of agonist anti-LAG-3 monoclonal antibody 13E2, 34F4, or an isotype control (mIgG1).

(141) The T cell response was investigated by measuring CFSE-based proliferation of CD8.sup.+ T cells on day 5. The Proliferation Index (calculated as the sum of the percentage of CD8.sup.+ T cells under each division peak, multiplied by the division number) is provided in Table 20.

(142) TABLE-US-00032 TABLE 20 Effect of 13E2 and 34F4 antibodies on CD8.sup.+ T cell proliferation Antibody Proliferation (30 ng/ml) Index mIgG1 69.9 13E2 26.2 34F4 16.4 None 95.3

(143) Table 21 below records the CD8.sup.+ T cell Proliferation Index as a function of antibody concentration. The results in Table 21 are plotted in FIG. 19.

(144) TABLE-US-00033 TABLE 21 Effect of different concentrations of antibody on CD8+ T cell proliferation Antibody CD8+ T cell Proliferation concentration Index (ng/ml) mIgG1 13E2 34F4 3000 82.9 17.6 14.9 952 72.9 13.7 10.5 302 69.9 16.4 26.2 96 65.3 19.6 13.7 30 87.4 14.4 22.8 10 78.2 50.4 46.2 3 77.8 59.2 66.6 0 95.3 95.3 95.3

(145) The results show that a dose as low as 30 ng/ml of monoclonal anti-LAG-3 antibody 13E2, or 34F4, causes maximal inhibition of CD8.sup.+ T cell proliferation. The results also show that the effects of the antibodies were very similar.

EXAMPLE 13

(146) Inhibition of CD8.sup.+ T Cell Proliferation by 34F4 is Reversed by Preincubation with IMP321

(147) CFSE-labelled PBMCs from 2 donors were stimulated by CMV peptide, as described above, in the presence of various concentrations of 34F4 antibody. The 1 μg/ml dose of 34F4 was also assessed after neutralisation with a 10-fold excess of IMP321.

(148) The T cell response was investigated by measuring the CFSE-based proliferation of CD8.sup.+ T cells on day 5. The percentage of inhibition of the proliferation of CD8.sup.+ T cells was calculated based on the percentage of dividing cells observed in the presence of 34F4 antibody, or 34F4 antibody and IMP321 (LAG-3Ig), compared to a control with or without IMP321.

(149) The results are shown in Table 22 below.

(150) TABLE-US-00034 TABLE 22 Effect of IMP321 on inhibition of CD8.sup.+ T cell proliferation by 34F4 antibody Donor #1 Donor #2 Incubation Antibody concentration Antibody concentration condition 1000 ng/ml 100 ng/ml 10 ng/ml 1000 ng/ml 100 ng/ml 10 ng/ml 34F4 64% 68% 44% 65% 35% 41% 34F4 + IMP321  0% nd nd 41% nd nd

(151) The results show that pre-incubation of 34F4 antibody with IMP321 reverses the inhibitory effect of 34F4 antibody on proliferation of CD8.sup.+ T cells. This shows that the inhibition of CD8.sup.+ T cell proliferation mediated by 34F4 is dependent on binding of 34F4 antibody to LAG-3.

EXAMPLE 14

(152) Inhibition of CD8.sup.+ T Cell Proliferation by 13E2 and 34F4 is not Reversed by IL-2

(153) CFSE-labelled PBMCs were stimulated by CMV peptides, as described above, in the presence of 13E2 or 34F4 antibody, with or without IL-2.

(154) The T cell response was investigated by measuring the CFSE-based proliferation of CD8.sup.+ T cells on day 5. The percentage of inhibition of the proliferation of CD8.sup.+ T cells was calculated based of the percentage of dividing cells observed in the presence of 13E2 or 34F4 antibody, with or without or IL-2, compared to an isotype control with or without IL-2.

(155) The results are shown in Table 23 below.

(156) TABLE-US-00035 TABLE 23 Effect of IL-2 on inhibition of CD8.sup.+ T cell proliferation by 13E2 and 34F4 antibody Percentage of inhibition of CD8.sup.+ T cell proliferation Antibody Without IL-2 With IL-2 13E2 38% 44% 34F4 59% 64%

(157) The results show that the addition of exogenous IL-2 was not able to overcome the inhibitory effect of 13E2 or 34F4 antibody on proliferation of CD8.sup.+ T cells. It is concluded from these results that antibodies 13E2 and 34F4 each directly inhibit signal 1 (response to the CMV antigen, a T-cell receptor dependent pathway) but not signal 2 (response to IL-2, that is help from CD4 cells) in CD8.sup.+ T cells.

EXAMPLE 15

(158) Effect of 13E2 on T Cell Activation Marker Secretion

(159) Peripheral blood mononuclear cells (PBMCs) include lymphocytes (T cells, B cells, and NK cells), monocytes, and dendritic cells. IFN-γ is predominantly secreted by activated CD4.sup.+ and CD8.sup.+ memory and effector T cells and by NK cells upon activation. After re-stimulation with specific antigen in vitro, secretion of IFN-γ is induced.

(160) PBMCs from four healthy donors (0.2×10.sup.6 cells/well, at 1×10.sup.6/ml in complete

(161) RPMI+10% FBS) were labelled with CFSE, and incubated with a pool of peptides covering the sequence of CMV pp35 in the presence of monoclonal anti-LAG-3 antibody 13E2 or an isotype control (mIgG1). The T cell response was investigated by measuring the release of IFN-γ in cell supernatant on day 2. The concentration of IFN-γ, and the percentage inhibition of IFN-γ secretion by 13E2, is presented in Table 24 below.

(162) TABLE-US-00036 TABLE 24 Agonist/ [IFN] Donor Control (pg/ml) Donor 1 mIgG1 297 13E2 226 % Inhibition 24% Donor 2 mIgG1 1043 13E2 255 % Inhibition 76% Donor 3 mIgG1 499 13E2 91 % Inhibition 82% Donor 4 mIgG1 1151 13E2 75 % Inhibition 93%

(163) The results show that monoclonal antibody 13E2 inhibited the secretion of IFN-γ in each of the donors tested. This provides evidence that monoclonal antibody 13E2 inhibits activation of T cells.

EXAMPLE 16

(164) Effect of 13E2 and 34F4 on T Cell Activation Marker Expression

(165) PBMCs from four healthy donors (0.2×10.sup.6 cells/well, at 1×10.sup.6/ml in complete RPMI+10% FBS) were labelled with CFSE, and incubated with a pool of peptides covering the sequence of CMV pp35 in the presence of monoclonal anti-LAG-3 antibody 13E2 or 34F4, or an isotype control (mIgG1).

(166) The T cell response was investigated by measuring the expression of CD25, as an activation marker, on CD8.sup.+ T cells on day 5. The percentage of CD8.sup.+ T cells expressing CD25, as well as the percentage inhibition of CD25 expression by 13E2 or 34F4, is presented in Table 25 below.

(167) TABLE-US-00037 TABLE 25 Agonist/ % CD25+ in Donor Control CD8+ T cells Donor 1 mIgG1 3.5 13E2 1.2 % Inhibition 66% Donor 2 mIgG1 6.6 13E2 5 % Inhibition 24% Donor 3 mIgG1 27.5 34F4 8.8 % Inhibition 68% Donor 4 mIgG1 9.2 34F4 7.0 % Inhibition 24%

(168) The results show that each monoclonal antibody, 13E2 and 34F4, significantly inhibited the expression of CD25 on CD8.sup.+ T cells. This provides evidence that each antibody inhibits activation of CD8.sup.+ T cells.

EXAMPLE 17

(169) Chimeric 13E2-Human Antibody Sequences

(170) Nucleotide sequences encoding the murine variable region of the 13E2 heavy and light chains were fused to sequences encoding the constant region of the human IgG4 heavy and Kappa light chains, respectively. These synthetic chimeric sequences were subcloned into an expression vector, and expressed in CHO cells grown in suspension.

(171) The heavy chain amino acid sequence of a chimeric 13E2-human IgG4 Fc antibody is shown below, and in FIG. 20(A). The antibody comprises the V.sub.H domain of mouse monoclonal antibody 13E2, and a human IgG4 Fc portion with an S228P mutation (to abolish Fab arm exchange) (13E2IgG4mut). In the figure, the V.sub.H region is shown in bold, and the Fc region is shown highlighted.

(172) TABLE-US-00038 13E2IgG4mut (SEQ ID NO: 30) MGWTLVFLFLLSVTAGVHSQVTLKESGPGILQPSQTLSLTCSFSGFSLST SGMGLGWIRQPSGKGLEWLTHIWWDDIKRYNPDLRSRLTISKDTSSSQIF LKIASVDTADTATYYCARIVEGSYSSSYFDVWGAGTTVTVSSASTKGPSV FPLAPCSRSTSESTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQ SSGLYSLSSVVTVPSSSLGTKTYTCNVDHKPSNTKVDKRVESKYGPPCPP CPAPEFLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSQEDPEVQFNWY VDGVEVHNAKTKPREEQFNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKGL PSSIEKTISKAKGQPREPQVYTLPPSQEEMTKNQVSLTCLVKGFYPSDIA VEWESNGQPENNYKTTPPVLDSDGSFFLYSRLTVDKSRWQEGNVFSCSVM HEALHNHYTQKSLSLSLGK

(173) The light chain amino acid sequence of a chimeric 13E2-human IgK antibody is shown below, and in FIG. 20(B). The antibody comprises the V.sub.L domain of monoclonal antibody 13E2, and a wild-type human Ig kappa (IgK) chain C portion (13E2IgK). In the figure, the V.sub.L region is shown in bold, and the IgK region is shown highlighted.

(174) TABLE-US-00039 13E2IgK (SEQ ID NO: 37) MVSSAQFLGLLLLCFQGTRCDIVMTQPHKFMSTSVEDRVTITCKASQDVI FDVAWYQQKPGQSPKLLIYSASSRVSGVPDRFTGSGSGTDFTFTISSVQA EDLAVYYCQQHYSTPYTFGGGTTLEIKRTVAAPSVFIFPPSDEQLKSGTA SVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLSSTLT LSKADYEKHKVYACEVTHQGLSSPVTKSFNRGEC

(175) A chimeric 13E2-human antibody (referred to as Chim13E2IgG4) comprises the chimeric heavy and light chains: 13E2IgG4mut; and 13E2IgK.

EXAMPLE 18

(176) Humanized 13E2 Monoclonal Antibody (IMP761) Sequences

(177) For optimal retention of CDR-loop conformation, combined IMGT/Kabat CDR sequence identification was used to graft CDRs from murine 13E2 to human frameworks in order to obtain a humanized version of 13E2. These synthetic chimeric sequences were subcloned into an expression vector, and expressed in CHO cells grown in suspension.

(178) Heavy and light chain amino acid sequences of the humanized 13E2 monoclonal antibody (referred to as IMP761) are shown below. Variable domains are shown in bold, CDR sequences are shown underlined.

(179) TABLE-US-00040 Amino acid sequence of IMP761 heavy chain (SEQ ID NO: 84) MGWTLVFLFLLSVTAGVHSQITLKESGPTLVKPTQTLTLTCTFSGFSLST SGMGLGWIRQPPGKTLEWLTHIWWDDIKRYNPDLRSRLSITKDTSKNQVV LTMTNMDPLDTGTYYCARIVEGSYSSSYFDVWGQGTLVTVSSASTKGPSV FPLAPCSRSTSESTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQ SSGLYSLSSVVTVPSSSLGTKTYTCNVDHKPSNTKVDKRVESKYGPPCPP CPAPEFLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSQEDPEVQFNWY VDGVEVHNAKTKPREEQFNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKGL PSSIEKTISKAKGQPREPQVYTLPPSQEEMTKNQVSLTCLVKGFYPSDIA VEWESNGQPENNYKTTPPVLDSDGSFFLYSRLTVDKSRWQEGNVFSCSVM HEALHNHYTQKSLSLSLGK

(180) Alignment of this sequence (IMP761 heavy chain) with the chimeric heavy chain sequence, 13E2IgG4mut, of Example 17 is shown in FIG. 23. In the figure, the V.sub.H region is shown in bold, and the Fc region is shown highlighted. Amino acid residues of the humanized IMP761 sequence that differ from corresponding residues of the chimeric 13E2IgG4mut sequence are single underlined. CDR sequences (based on the combined IMGT/Kabat CDR sequence identification) are shown double underlined. The changed residues in the humanized sequence are also set out in Table 26 below (as VH variant 4, VH.sub.4, as well as the changed residues in three other humanized variants of the original 13E2 heavy chain sequence: VH variants 1, 2, and 3, VH.sub.1, VH.sub.2, and VH.sub.3).

(181) TABLE-US-00041 TABLE 26 13E2 mouse Humanized residue Heavy chain heavy chain VH.sub.4 residue no. residue VH.sub.1 VH.sub.2 VH.sub.3 (IMP761) 21 V I I I 26 S T 29 G A A T T 30 I L L L L 31 L V V V V 32 Q K K K K 34 S T T T T 38 S T T T T 42 S T T T 62 S P P P P 65 G A A A T 69 L V 70 T A A 88 L V 89 T S 91 S R T 96 S K K K K 97 S N N N 99 I V V V V 100 F I V A V 102 K N T T T 103 I M M M M 104 A T T T 105 S N N N N 106 V M M M M 108 T P P P P 109 A V V L L 112 A G G 134 A Q Q Q Q 137 T L L

(182) The heavy chain framework sequences of the humanized antibody (antibody IMP761) are:

(183) TABLE-US-00042 (SEQ ID NO: 64) VH FR1: QITLKESGPTLVKPTQTLTLTCTFS; (SEQ ID NO: 65) VH FR2: WIRQPPGKTLEWLT; (SEQ ID NO: 66) VH FR3: RLSITKDTSKNQVVLTMTNMDPLDTGTYYC; and (SEQ ID NO: 67) VH FR4: WGQGTLVTVSS.

(184) TABLE-US-00043 Amino acid sequence of IMP761 light chain (SEQ ID NO: 85) MVSSAQFLGLLLLCFQGTRCDIVMTQTPSSLSASVGDRVTITCKASQDVI FDVAWYQQRPGQAPKLLIYSASSRVSGVPSRFSGSGSGTDFTLTISSLQP EDFATYYCQQHYSTPYTFGQGTRLDIKRTVAAPSVFIFPPSDEQLKSGTA SVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLSSTLT LSKADYEKHKVYACEVTHQGLSSPVTKSFNRGEC

(185) Alignment of this sequence (IMP761 light chain) with the chimeric light chain sequence, 13E2IgK, of Example 17 is shown in FIG. 24. In the figure, the V.sub.L region is shown in bold, and the IgK region is shown highlighted. Amino acid residues of the humanized IMP761 sequence that differ from corresponding residues of the chimeric 13E2IgK sequence are single underlined. CDR sequences (based on the combined IMGT/Kabat CDR sequence identification) are shown double underlined. The changed residues in the humanized sequence are set out in Table 27 below (as VL variant 3, VH.sub.3, as well as the changed residues in three other humanized variants of the original 13E2 light chain sequence: VL variants 1, 2, and 4, VL.sub.1, VL.sub.2, and VL.sub.4).

(186) TABLE-US-00044 TABLE 27 13E2 Light mouse chain light Humanized residue residue chain VL.sub.3 no. residue VL.sub.1 VL.sub.2 (IMP761) VL.sub.4 21 D E 23 V Q 24 M L 27 P S S T S 28 H P P P P 29 K D S S D 30 F S S S S 31 M L L L L 32 S A A 33 T V A A V 35 V L L 36 E G G G G 37 D E E 39 V A A 59 K R 60 P A 63 S P A A 69 Y F 80 D S S 83 T S S S S 93 F L L L L 96 S D 98 V L L L L 100 A P P 103 L V F F V 105 V T T 120 G Q Q Q 123 T K K R K 124 L V V 125 E D

(187) The light chain framework sequences of the humanized antibody (antibody IMP761) are:

(188) TABLE-US-00045 (SEQ ID NO: 64) VL FR1: DIVMTQTPSSLSASVGDRVTITC; (SEQ ID NO: 65) VL FR2: WYQQRPGQAPKLLIY; (SEQ ID NO: 66) VL FR3: GVPSRFSGSGSGTDFTLTISSLQPEDFATYYC; and (SEQ ID NO: 67) VL FR4: FGQGTRLDIK

EXAMPLE 19

(189) Binding of Chimeric 13E2-Human Antibody (Chim13E2IgG4) and Humanized 13E2 Antibody (IMP761) to CHO-LAG-3+ Cells

(190) CHO cells expressing LAG-3 on their surface (0.05×10.sup.6 cells/well in PBS, 0.5% BSA, 0.1% azide) were incubated with different concentrations of a chimeric 13E2-human antibody (referred to as Chim13E2IgG4) comprising the chimeric heavy and light chains described in Example 17 (heavy chain: 13E2IgG4mut; light chain: 13E2IgK), IMP761, or human IgG4 (as an isotype-matched negative control). Secondary antibody goat anti-human IgG-FITC was used to detect the presence of antibodies on the surface of the CHO-LAG-3.sup.+ cells. FITC mean fluorescence of intensity (MFI) was determined after analysis by flow cytometry.

(191) The results are shown in Table 28 below, and in FIG. 25.

(192) TABLE-US-00046 TABLE 28 MFI FITC CHO cells ng/ml Chim13E2IgG4 IMP761 huigG4 6000 2039 2086 71 1905 2024 1989 77 605 1874 1821 46 192 1580 1429 59 61 1058 898 69 20 593 460 49 6 279 234 37 0 35 35 35

(193) The results show that the humanized monoclonal antibody IMP761 binds to CHO cells expressing LAG-3 on their surface in a very similar manner to the chimeric antibody.

EXAMPLE 20

(194) Binding Affinity of Chimeric 13E2-Human Antibody (Chim13E2IgG4) and Humanized 13E2 Antibody (IMP761) to Human LAG-3Ig Protein

(195) A Biacore™ surface plasmon resonance analysis was performed using the chimeric antibody Chim13E2IgG4 (comprising the chimeric heavy chain 13E2IgG4mut, and the chimeric light chain 13E2IgK, described in Example 17), or the humanized 13E2 antibody (IMP761) described in Example 18, covalently immobilized to a C1 sensor chip. The coating was performed in 10 mM sodium acetate, pH 5.0, to reach 13±1 RU. The recombinant human LAG-3Ig protein (IMP321) was then passed over the captured antibodies at 6 different concentrations, ranging from 0.078-2.5 nM, in analysis buffer (10 mM HEPES pH 7.4, 150 mM NaCl, 3 mM EDTA, 0.05% Tween 20) at 25° C., with regeneration in every cycle. The analysis was carried out on Biacore™ T200 and the data was fitted using the kinetic global fit (Langmuir 1:1) model. The kinetics parameters are recorded in Table 29, and represent the average of three runs.

(196) TABLE-US-00047 TABLE 29 Antibody ka (M.sup.−1s.sup.−1) kd (s.sup.−1) KD (pM) Chim13E2IgG4 2.9 ± 0.2 × 10.sup.7 6.4 ± 0.6 × 10.sup.−4 21.9 ± 0.1 IMP761 2.9 ± 0.2 × 10.sup.7 6.6 ± 0.6 × 10.sup.−4 22.8 ± 0.9

(197) The results show that the humanized monoclonal antibody IMP761 has the same affinity to human LAG-3Ig protein as the chimeric antibody. Both antibodies display a very rapid association rate, which explains the high affinity of the 13E2-derived antibodies for LAG-3.

EXAMPLE 21

(198) Effect of Humanized 13E2 Antibody (IMP761) on CD8.sup.+ T Cell Proliferation and CD25 Expression Induced by Antigenic Stimulation

(199) CFSE-labelled PBMCs from healthy donors (0.2×10.sup.6 cells/well in complete RPMI+10% FBS) were incubated with a pool of peptides covering the sequence of CMV pp35 in triplicate, with 300 ng/ml human IgG4 (isotype control), Chim13E2IgG4 or IMP761. The T cell response was evaluated by measuring the proliferation, evaluated using a proliferation index (calculated as the sum of the percentage of CD8.sup.+ T cells under each division peak, multiplied by the division number), and the expression of CD25 at day 5 by flow cytometry. The percentage inhibition for each antibody, compared with the isotype-matched negative control (huIgG4), was calculated based on the proliferation index values, or the percentage of CD25.sup.+ T cells within the CD8.sup.+ T cell population.

(200) The results are shown in Tables 30 and 31 below, and in FIGS. 26 and 27.

(201) TABLE-US-00048 TABLE 30 Proliferation index PI (CD8) % inhibition donor 1 huIgG4 8.3 Chim13E2IgG4 3.6 56.6 IMP761 3.9 53.0 donor 2 huIgG4 29.4 Chim13E2IgG4 17.7 39.8 IMP761 10.9 62.9 donor 3 huIgG4 3.7 Chim13E2IgG4 0.8 78.4 IMP761 1.4 62.2 donor 4 huIgG4 9.7 Chim13E2IgG4 4.5 53.6 IMP761 3.6 62.9 Average huIgG4 12.8 Chim13E2IgG4 6.7 57.1 IMP761 5.0 60.2

(202) TABLE-US-00049 TABLE 31 CD25 expression in CD8+ T cells % CD25 % inhibition donor 1 huIgG4 3.3 Chim13E2IgG4 1.7 48.5 IMP761 2.1 36.4 donor 2 huIgG4 15.1 Chim13E2IgG4 9.6 36.4 IMP761 6.9 54.3 donor 3 huIgG4 3.3 Chim13E2IgG4 1.3 60.6 IMP761 2.2 33.3 donor 4 huIgG4 5.1 Chim13E2IgG4 2.8 45.1 IMP761 2.3 54.9 Average huIgG4 6.7 Chim13E2IgG4 3.9 47.7 IMP761 3.4 44.7

(203) The results show that the humanized monoclonal antibody IMP761 had a similar effect on inhibition of antigen-induced CD8.sup.+ T-cell proliferation, and CD25.sup.+ expression, as the chimeric antibody Chim13E2IgG4. Both antibodies caused, on average, approximately 60% inhibition of antigen-induced CD8.sup.+ T-cell proliferation, and approximately 45% inhibition of CD25.sup.+ T cells within the CD8.sup.+ T cell population.

EXAMPLE 22

(204) Effect of Different Doses of Chimeric 13E2-Human Antibody (Chim13E2IgG4) and Humanized 13E2 Antibody (IMP761) on CD8.sup.+ T Cell Response

(205) CFSE-labelled PBMCs from healthy donors (0.2×10.sup.6 cells/well in complete RPMI+10% FBS) were incubated with a pool of peptides covering the sequence of CMV pp35 in triplicate, with different doses of Chim13E2IgG4, IMP761, or human IgG4 (an isotype-matched negative control). The T cell response was evaluated by measuring the proliferation (CFSE dilution) at day 5 by flow cytometry. The percentage of CD8.sup.+ T cells for each division number was calculated for the different antibody doses used.

(206) The results are shown in Table 32 below, and in FIG. 28.

(207) TABLE-US-00050 TABLE 32 Antibody concentration cell division (ng/ml) Antibody 1 2 3 4 5 6 7 and more 10 huIgG4 2.3 5.1 5.9 3.7 1.6 0.5 0.3 Chim13E2IgG4 1.9 4 4.4 2.6 1.1 0.4 0.1 IMP761 2.2 2.6 3.5 2.1 0.9 0.3 0 30 huIgG4 2.1 4.9 5.4 2.4 1.2 0.3 0.2 Chim13E2IgG4 2.1 3.4 3.4 2 0.9 0.2 0.1 IMP761 1.5 2.3 2.5 1.5 0.7 0.1 0 100 huIgG4 1.9 4.3 4.4 2 0.6 0.3 0 Chim13E2IgG4 1.2 1.4 1.9 0.9 0.6 0.2 0 IMP761 1.4 1.6 1.7 0.9 0.3 0.1 0 300 huIgG4 2.9 5.5 5.8 2.4 1.1 0.2 0.1 Chim13E2IgG4 1.5 2.1 2.4 1.5 0.7 0.1 0.1 IMP761 1.4 1.9 2.1 1.2 0.6 0.2 0.1 Unstim 1.2 0.1 0.1 0.1 0.1 0 0

(208) The results show that the inhibitory effect of IMP761 and Chim13E2IgG4 on antigen-induced CD8.sup.+ T-cell proliferation was dose-dependent. In particular, the inhibitory effect of each antibody increased as the dose increased from 10 ng/ml to 100 ng/ml antibody. At 300 ng/ml antibody, the inhibitory effect was similar to that of 100 ng/ml antibody. The inhibitory effect of IMP761 was similar to that of Chim13E2IgG4 at all doses tested.

EXAMPLE 23

(209) Humanized 13E2 Antibody (IMP761) does not Possess Cytotoxic Activity Against LAG-3-Expressing Cells

(210) Several types of assay were used to confirm that the humanized 13E2 antibody (IMP761) does not possess cytotoxic activity against LAG-3-expressing cells.

(211) 1) ADCC Reporter Bioassay (Promega, G7015)

(212) In this assay, primary donor PBMC or NK cells are replaced with Jurkat cells stably expressing human FcγRIIIa (the high-affinity V158 receptor) and NFAT-responsive element driving expression of a luciferase reporter gene. If a test antibody has ADCC activity, it will bind together a target cell and the FcγRIIIa receptor of a Jurkat cell. Resulting activation of the signalling pathway downstream of the FcγRIIIa receptor results in NFAT pathway activation thereby inducing luciferase reporter gene expression. Luciferase activity is quantified by luminescence read-out.

(213) LAG-3-transfected CHO and Jurkat cells, and PBMCs stimulated by SEB for 2 days to cause expression of LAG-3 (55% of the PBMCs were LAG-3.sup.+), were used as targets cells to assay the ADCC activity of IMP761, compared with an isotype-matched negative control antibody, hIgG4 (recombinant mAb from BioRad). Anti-CD20 antibody and Raji cells, provided with the assay kit, were used as a positive control. The anti-CD20 antibody was also tested on SEB-stimulated PBMCs. The assays were carried out, following the manufacturer's instructions, using 75,000 effector cells with 12,500 target cells. After incubation for 6 hours at 37° C., Bio-Glo luciferase assay system was used, according to the manufacturer's instructions, to measure luminescence using a PerkinElmer EnVision 2103 luminometer (integration time of 0.5 sec/well).

(214) The results are shown in Table 33 below, and in FIG. 29(a). The results are presented as the fold change in the relative luminescence unit (RLU), calculated by dividing the RLU obtained in the presence of the test antibody (at the maximal concentration recommended by the manufacturer, 3 μg/ml), by the RLU obtained without antibody.

(215) TABLE-US-00051 TABLE 33 Fold change in RLU Target cells anti-CD20 hIgG4 IMP761 Raji 36.8 CHO-LAG3.sup.+ 1.4 1.1 Jurkat WT 2.3 0.9 Jurkat-LAG3.sup.+ 2.5 1.2 LAG3.sup.+ PBMCs 5.0 1.9 1.3

(216) The results show that the fold change in RLU for IMP761 antibody was approximately 1-fold for each of the different target cells tested, irrespective of whether the target cell expresses LAG-3 or not. The fold change in RLU obtained for the isotype-matched negative control antibody, hIgG4, was slightly higher and ranged from a 1.4- to 2.5-fold for the different target cells. The positive control anti-CD20 antibody showed significant ADCC activity against the Raji cells (a B cell line), and against SEB-stimulated PBMCs, which contain a small percentage of B cells.

(217) It was concluded from these results that the IMP761 antibody does not have any ADCC activity against LAG-3-expressing cells.

(218) 2) Conventional ADCC Assay

(219) This assay uses PBMCs stimulated for one day in X-Vivo 10 medium (Lonza) with 100 IU/ml of IL-2 (Roche) and CFSE-labelled PBMCs, stimulated with SEB for two days to allow the expression of LAG-3 on T cells. The assay was carried out in X-Vivo 10 medium at a 50:1 effector:target ratio, with high dose (3 μg/ml) IMP761 or an isotype-matched negative control antibody, hIgG4. After 4 hours, the cell mixtures were harvested and stained for CD4, CD8, CD25 and LAG-3 using fluorochrome-conjugated antibodies. Cell viability in each blood cell population was then assessed by flow cytometry, after exclusion of cells which appeared positive for 7-Amino-Actinomycin D (7-AAD) staining, a fluorescent dye which labels cells which have lost their membranous integrity, a phenomenon which appears rapidly after cell death.

(220) The results are shown in Table 34, and in FIGS. 29(b) and (c). The results are presented as the percentage of living CD4.sup.+ or CD8.sup.+ cells in the PBMC population (b), and the percentage of living LAG-3.sup.+CD4.sup.+ or LAG-3+CD8.sup.+ cells in the PBMC population (c).

(221) TABLE-US-00052 TABLE 34 % living % living LAG-3.sup.+CD8.sup.+ CD8.sup.+ or CD4.sup.+ cells or LAG-3.sup.+CD4.sup.+ cells in PBMC population in PBMC population No Ab hIgG4 IMP761 No Ab hIgG4 IMP761 CD8.sup.+ cells 24.9 23.9 22.9 10.5 8.4 13.9 CD4.sup.+ cells 43.3 41.0 44.1 14.0 9.0 17.1

(222) The results show that IMP761 antibody does not reduce the percentage of CD8.sup.+ or CD4.sup.+ T cells in the PBMC population, or the percentage of LAG-3.sup.+CD8.sup.+ cytotoxic T cells, or LAG-3.sup.+CD4.sup.+ helper T cells in the PBMC population. The isotype-matched negative control antibody, hIgG4, caused a slight reduction in the viability of T cells in the PBMC population, especially of activated T cells expressing LAG-3.

(223) It was concluded from these results that IMP761 antibody does not have any ADCC activity against LAG-3-expressing T cells.

(224) 3) CDC Assay

(225) For CDC testing, SEB-stimulated cells, used as target cells, were incubated with 3 μg/ml of IMP761, an isotype-matched negative control antibody, hIgG4, a CDC-positive anti-CD3 control antibody (clone MEM-57, Cerdalane), or an isotype-matched negative control mouse antibody, mIgG2a, for 45 minutes in PBS, 0.5% BSA. Unbound antibodies were then washed away, and the cells were incubated with rabbit complement diluted by 3 volumes in RPMI medium for 1 hour at 37° C. The cells were stained for CD4, CD8, CD25 and LAG-3 using fluorochrome-conjugated antibodies. Cell viability in each blood cell population was then assessed by flow cytometry, after exclusion of cells labelled by 7-AAD.

(226) The results are shown in Table 35, and in FIG. 30. The results are presented as the percentage of living CD4.sup.+ or CD8.sup.+ in the PBMC population (a), and the percentage of living LAG-3.sup.+CD4.sup.+ or LAG-3.sup.+CD8.sup.+ cells in the PBMC population (b).

(227) TABLE-US-00053 TABLE 35 % living CD8.sup.+ or CD4.sup.+ cells % living LAG3.sup.+CD8.sup.+ in PBMC population or LAG-3.sup.+CD4.sup.+ cells in PBMC population hIgG4 IMP761 mIgG2a CD3 hIgG4 IMP761 mIgG2a CD3 CD8.sup.+ cells 13.0 13.4 13.7 9.3 10.0 10.5 8.6 5.9 CD4.sup.+ cells 35.8 36.6 32.9 26.9 13.3 13.9 12.1 8.9

(228) The results show that IMP761 antibody does not reduce the percentage of CD8.sup.+ or CD4.sup.+ T cells in the PBMC population, or the percentage of LAG-3.sup.+CD8.sup.+ cytotoxic T cells, or LAG-3.sup.+CD4.sup.+ helper T cells in the PBMC population. As expected, the anti-CD3 positive control antibody did cause a decrease in the percentage of T cells in the PBMC population, and of activated T cells expressing LAG-3.

(229) It was concluded from these results that IMP761 antibody does not have any CDC activity against LAG-3-expressing T cells.

(230) 4) Assessment of Cytotoxicity in T Cell Proliferation Assay

(231) IMP761 antibody showed no cytotoxic activity in any of the short-term cytotoxicity assays described in (1)-(3) above. The cytotoxicity of IMP761 against LAG-3-expressing T cells was also evaluated after culturing antigen-stimulated PBMCs for several days. In a similar manner to the proliferation assays described in the previous examples, PBMCs from healthy donors (0.2×10.sup.6 cells/well in complete RPMI+10% FBS) were incubated in triplicate with a pool of peptides covering the sequence of CMV pp35, in the presence of 300 ng/ml IMP761, or human IgG4 (as an isotype-matched negative control). After three days, the percentage of CD8.sup.+ and CD4.sup.+ T cells gated in living lymphocytes, as well as the percentage of LAG-3.sup.+ cells in these T cell subsets, was measured by flow cytometry.

(232) The results are shown in Table 36, and in FIG. 31.

(233) TABLE-US-00054 TABLE 36 % living CD8+ or CD4+ among lymphocytes % living LAG3+ among CD8+ or CD4+ lymphocytes Donor 1 Donor 2 Donor 1 Donor 2 hIgG4 IMP761 hIgG4 IMP761 hIgG4 IMP761 hIgG4 IMP761 CD8+ cells 13.9 13.9 36.3 36.2 3.2 3.8 0.9 1.2 CD4+ cells 35.8 36.6 35.3 35.3 2.6 3.4 2.5 2.8

(234) The results show that IMP761 antibody does not reduce the percentage of CD8.sup.+ or CD4.sup.+ T cells in the lymphocyte population, or the percentage of LAG-3.sup.+CD8.sup.+ cytotoxic T cells, or LAG-3.sup.+CD4.sup.+ helper T cells in the lymphocyte population.

(235) It was concluded from these results that IMP761 antibody does not show any cytotoxic activity against LAG-3-expressing T cells in this proliferation assay.

(236) It was concluded from the results presented in this example that IMP761 antibody does not possess cytotoxic activity, so inhibition of antigen-induced T cell proliferation and activation by this antibody is not due to any cytotoxic activity against activated T cells.