Multispecific antigen-binding protein

Abstract

The invention relates to a multispecific antigen-binding protein that specifically binds PD-L1 and PVRIG. The invention further relates to antigen-binding proteins specific for PD-L1 and to antigen-binding proteins specific for PVRIG. The invention further relates to nucleic acids encoding the above antigen-binding proteins, and to pharmaceutical compositions, medical uses and methods of treatment comprising the antigen-binding proteins or their encoding nucleic acids.

Description

[0001] MULTISPECIFIC ANTIGEN-BINDING PROTEIN

[0002] Field of the Invention

[0003] The invention relates to a multispecific antigen-binding protein that specifically binds PD-L1 and PVRIG. The invention further relates to anti gen -binding proteins specific for PD-L1 and to anti gen -binding proteins specific for PVRIG. The invention further relates to nucleic acids encoding the above antigen-binding proteins, and to pharmaceutical compositions, medical uses and methods of treatment comprising the antigen-binding proteins or their encoding nucleic acids.

[0004] Background to the Invention

[0005] Programmed death-ligand 1 (PD-L1 or CD274 or B7-H1) is a transmembrane protein with a size of 40kDa. In healthy tissues, cells express PD-L1 in response to inflammatory situations and it plays a role in dampening cytotoxic CD8+ T cells or CD4+ helper cells. PD-L1 is mostly expressed on macrophages, on certain activated T and B cell populations, on DCs and some epithelial cells. PD-1 is a 55kDa co-inhibitory transmembrane molecule with an ITIM intracellular domain that is expressed on activated T, NK, B cells, macrophages, and dendritic cells (DC). Binding of PD-L1 to PD-1 inhibits the functional activities of these effector cells.

[0006] In cancer, the PD-1 / PD-L1 pathway is one of the pathways that regulates immune tolerance within the tumour microenvironment. PD-1 is especially highly expressed on T cells within the tumor microenvironment. PD-1 expression can be triggered by transcription factors such as NF AT, NOTCH, FOXO1, and IRF9. Besides cancer tissues, PD-1 expression is upregulated highly in chronic infections. Tumor cells also express PD- L1 to evade the immune response elicited by CD8+ T cells. In particular, tumor cells upregulate their PD-L1 expression in response to IFN-g release by the effector lymphocytes in the tumor microenvironment. The engagement of PD-1 and PD-L1 can occur in the tumor microenvironment, tertiary lymphoid structures and draining lymph nodes. Disruption of the PD-1 / PD-L1 pathway may result in reactivation of tumor-specific cytotoxic lymphocytes. In particular, since PD-1 is also associated with the exhausted cytotoxic lymphocyte phenotype, where tumor specific T cells become dysfunctional, blocking PD-1 engagement to PD-L1 can rescue these cells to be functional again. In addition, it may enable functional tumor antigen presentation by macrophages and dendritic cells.

[0007] Anti-PDl and anti-PD-Ll blocking antibodies have been used in cancer immunotherapies for multiple tumor indications. These therapies have shown a significant success with the tumors in patients completely regressed.

[0008] PVRIG (Poliovirus receptor-related immunoglobulin domain-containing protein), also known as CD112 receptor (CD112R), is a transmembrane receptor expressed on the surface of T cells or NK cells. PVRIG is part of the DNAM-1 / TIGIT / CD96 axis for coregulating lymphocyte activation, and contains in its cytoplasmic tail an immunoreceptor tyrosine-based inhibitory motif (ITIM) domain.

[0009] Upon binding of PVRIG to its ligand PVRL2 (CD112 or Nectin2) expressed on the antigen presenting cells (APCs) or tumor cells, T cell or NK cells are negatively regulated (Zhu et al., 2016). PVRL2 can additionally bind DNAX accessory molecule 1 (DNAM-1) with a lower affinity, which is a co-activating receptor found also on T cells or NK cells. PVRL2 can also bind to another co-inhibitory molecule expressed on the cytotoxic T cells, called T-cell immunoreceptor with Ig and ITIM domains (TIGIT). Thus, the same ligand can be either co-inhibitory or co-activatory depending on the receptor expression.

[0010] PVRL2 is a surface adhesion protein playing a role in the formation of cell / cell junctions and overexpressed in various malignant solid tumors. Furthermore, PVRIG expression has been shown to be upregulated on CD8+ T cells found in colorectal tumor patient tissue samples (Yang et al., 2023). Blocking PVRIG by antibodies has been shown to increase NK cell cytotoxicity against breast cancer cell lines in vitro (Xu et al., 2017). Similarly, anti-PVRIG blocking antibodies reversed CD8+ dysfunction within the tumor microenvironment through disruption of PVRIG binding to PVRL2 in the tumors (Whelan et al., 2019).

[0011] There is a need for further therapeutic agents for treatment of diseases such as cancer. Summary of the Invention

[0012] The inventors have generated multispecific antigen-binding proteins that bind to both PD-L1 and PVRIG. These multispecific antigen-binding proteins provide for dual targeting of PD-L1 and PVRIG, allowing for engagement of PD-L1 on cancer cells and / or antigen-presenting cells, and engagement of PVRIG on immune cells. It is believed that the multispecific antigen-binding proteins may thus be able to block immunosuppressive pathways in cancer and also promote immune cell activation. The multispecific antigenbinding proteins thus provide a particularly effective means of treating cancer by reducing immunosuppression and activating immune cells to specifically target cancer cells. As shown in the examples, the multispecific antigen-binding proteins are able to activate T cells, NK cells and induce ADCC against cancer cells. They are further able to crosslink PD-L1 on cancer cells and PVRIG on immune cells, which may enhance anti-cancer activity. The inventors further screened a large number of PD-L1 and PVRIG binders to identify PD-L1 and PVRIG-specific antigen-binding proteins which have optimal characteristics. They then screened a large number of multispecific antigen-binding proteins representing combinations of the optimal binders and showed that particular designs of multi-specific binding proteins have optimal properties.

[0013] The invention accordingly provides a multispecific antigen-binding protein comprising a first antigen-binding domain specific for PD-L1 and a second antigenbinding domain specific for PVRIG. The invention further provides an antigen-binding protein specific for PD-L1, comprising the three heavy chain CDRs and three light chain CDRs contained within a HCVR / LCVR pair selected from SEQ ID NOs 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156, 147 / 157, 148 / 158, 149 / 159, and 150 / 160. The invention additionally provides an antigen-binding protein specific for PVRIG, comprising the three heavy chain CDRs and three light chain CDRs contained within a HCVR / LCVR pair selected from SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140.

[0014] The invention also provides a combination of an antigen-binding protein specific for PD-L1 according to the invention and an antigen-binding protein specific for PVRIG according to the invention. The invention further provides one or more nucleic acids encoding a multispecific antigen-binding protein according to the invention, an antigen- binding protein specific for PD-L1 according to the invention or an antigen-binding protein specific for PVRIG according to the invention. The invention additionally provides a pharmaceutical composition comprising a multispecific antigen-binding protein according to the invention, an antigen-binding protein specific for PD-L1 according to the invention, an antigen-binding protein specific for PVRIG according to the invention, a combination of antigen-binding proteins according to the invention, or one or more encoding nucleic acids according to the invention, and a pharmaceutically acceptable excipient.

[0015] The invention also provides a multispecific antigen-binding protein according to the invention, an antigen-binding protein specific for PD-L1 according to the invention, an antigen-binding protein specific for PVRIG according to the invention, a combination of antigen-binding proteins according to the invention, one or more encoding nucleic acids according to the invention, or a pharmaceutical composition according to the invention, for use in a method of treatment of the human or animal body by therapy. The invention further a multispecific antigen-binding protein according to the invention, an antigenbinding protein specific for PD-L1 according to the invention, an antigen-binding protein specific for PVRIG according to the invention, a combination of antigen-binding proteins according to the invention, one or more encoding nucleic acids according to the invention, or a pharmaceutical composition according to the invention, for use in a method of treatment of a cancer or tumour.

[0016] Description of the Figures

[0017] Figure 1 PVRIG expression was evaluated on the CD8+ T cells (A), CD56+ NK cells (B) and Foxp3+ regulatory T (Treg) cells (C) within the tumor tissues and stromal samples of the triple negative breast cancer (breast tumor and breast stroma), gastric cancer (gastric tumor and gastric stroma), non-small cell lung cancer (NSCLC tumor and NSCLC stroma), renal cell carcinoma (RCC tumor and RCC stroma) and melanoma (melanoma tumor and melanoma stroma). Abbreviations of the sample names shown in the figure 1 A, IB and 1C are in brackets. Figures show the percentages of PVRIG expressing CD8+ T cells (A), CD56+ NK cells (B) and Foxp3+ Treg cells per mm2in the tumor tissue samples and stromal samples of the different cancer types. Figure 2 PD-L1 expression was evaluated on the CD68+ tumor associated macrophages (A) and Epcam expressing tumor cells (B) within the tumor tissues and stromal samples of the triple negative breast cancer (breast tumor and breast stroma), gastric cancer (gastric tumor and gastric stroma), non-small cell lung cancer (NSCLC tumor and NSCLC stroma), renal cell carcinoma (RCC tumor and RCC stroma) and melanoma (melanoma tumor and melanoma stroma). Abbreviations of the sample names shown in the figure 2A and 2B are in brackets. Figures show the percentages of PD-L1 expressing CD68+ tumor associated macrophages (A) and Epcam expressing tumor cells (B) per mm2in the tumor tissue samples and stromal samples of the different cancer types. Melanoma samples were not analysed from Epcam expressing tumor cells.

[0018] Figure 3 The percentage of CD3+ T cells, CD4+ T cells, CD8+ T cells, CD3- CD56+ NK cells and CD3+CD56+ NKT cells in CD45+ cells of peripheral blood mononuclear cells (PBMCs) from healthy controls (ctrl PBMCs), PBMCs from lung cancer patients (patient PBMCs) and lung cancer dissociated tumor cells (DTC) (lung DTC).

[0019] Figure 4 CD4+ T cells, CD8+ T cells, CD3-CD56+ NK cells and CD3+CD56+ NKT cells expressing PVRIG, TIGIT, PD-1 and DNAM-1. Cells were analysed from peripheral blood mononuclear cells (PBMCs) from healthy controls (ctrl PBMCs), PBMCs from lung cancer patients (patient PBMCs) and lung cancer dissociated tumor cells (DTC) (lung DTC).

[0020] Figure 5 The expression of PD-L1, PVR and PVRL2 on CD1 Ib+CDl lc+ cell population including monocytes, dendritic cells and macrophages. The cell population was isolated from healthy controls PBMCs (healthy PBMCs), lung cancer patients PBMCs (lung cancer PBMCs) and lung cancer dissociated tumor cells (lung cancer DTCs).

[0021] Figure 6 PD-L1, PVR and PVRL2 expression was studied in Epcam+ lung cancer cells (CD45-).

[0022] Figure 7 PVRIG expression on the CD3+ T cells, CD3-CD56+ NK cells, CD3+CD56+ NKT cells and gamma delta TCR expressing T cells (gamma delta T cells) was studied by flow cytometry. Figure shows the percentages of the cell populations present in PBMC, the percentage of PVRIG expressing cells in each cell population and PVRIG median values. Figure 8 PD-L1 expression on the CD3+ T cells, NK cells, CD3+CD56+ NKT cells and the gamma delta T cells was studied by flow cytometry. Figure shows the percentages of the cell populations present in PBMC, the percentage of PD-L1 expressing cells in each cell population and PD-L1 median values.

[0023] Figure 9 (A) The PD-L1 binding domains of Multi-246 and Multi-256 were epitope mapped. (B) The PVRIG binding domains of Multi -246, Multi -256 and anti-PVRIG antibody 743 were epitope mapped.

[0024] Figure 10 PVRIG binding and blocking towards its ligand PVRL2 by PVRIG x PD-L1 binding proteins measured with a reporter assay. The figure shows the fold change with different concentrations of each PVRIG x PD-L1 binding protein and an anti-PVRIG antibody. An anti-PVRIG antibody has the highest capacity to block PVRIG / PVRL2 interaction. All the selected PVRIG x PD-L1 binding proteins multi-216, multi-246, multi- 256 and multi-61 showed similar blocking capacities. The EC50 values for the PVRIG x PD-L1 binding proteins were 1.2-2.7 nM.

[0025] Figure 11 Efficacy of the PVRIG x PD-L1 binding proteins multi-216, multi -246, multi-256 and multi-61 and anti-PD-Ll antibody Tecentriq (atezolizumab) show binding to PD-L1 and blocking its binding to its receptor PD-1. The figure shows the fold change with different concentrations of each PVRIG x PD-L1 binding protein and an anti-PD-Ll antibody. All the PVRIG x PD-L1 binding proteins blocked the PD-L1 binding to PD-1 with a comparable EC50 values between 3.0-5.7 nM.

[0026] Figure 12 Blocking capacity of PVRIG x PD-L1 binding proteins to block PVRL2 binding to PVRIG was measured with CHO cells overexpressing PVRIG on their surface in the presence of different concentrations of PVRIG x PD-L1 binding proteins multi-216, multi-246, multi -256. The figure shows percentage of PVRIG binding PVRL2 conjugated to human Fc protein and biotin was added with streptavidin conjugated to APC and measured by flow cytometry. PVRIG x PD-L1 binding proteins multi-216, multi-246, and multi-256 blocked the PVRIG / PVRL2 interaction in similar range to each other which were observed at subnanomolar level. An anti-PVRIG antibody was included as positive control.

[0027] Figure 13 Blocking capacity of PVRIG x PD-L1 binding proteins to block PD-L1 binding to PD-1 was measured with CHO cells overexpressing PD-L1 on their surface in the presence of different concentrations of PVRIG x PD-L1 binding proteins multi-216, multi-246 and multi-256. The figure shows percentage of PD-L1 binding PD-1 conjugated to human Fc protein and biotin was added with streptavidin conjugated to APC and measured by flow cytometry. PVRIG x PD-L1 binding proteins multi-216, multi-246 and multi-256 blocked the PD-1 / PD-L1 interaction in similar range to each other which were observed at subnanomolar level. An anti-PD-Ll antibody was included as positive control and streptavidin conjugated to APC as a negative control in the assay.

[0028] Figure 14 Blocking capacity of PVRIG x PD-L1 binding proteins to block CD80 binding to PD-L1 was measured with CHO cells overexpressing PD-L1 on their surface in the presence of different concentrations of PVRIG x PD-L1 binding proteins multi-216, multi-246, multi-256 and multi-61. The figure shows percentage of PD-L1 binding CD80 conjugated to human Fc protein and biotin was added with streptavidin conjugated to APC and measured by flow cytometry. PVRIG x PD-L1 binding proteins multi-216, multi-246, multi-256 and multi-61 blocked the CD80 / PD-L1 interaction in similar range to each other which were observed at subnanomolar level. An anti-PD-Ll antibody was included as positive control and streptavidin conjugated to APC as a negative control in the assay.

[0029] Figure 15 Functionality of the PVRIG x PD-L1 binding proteins was evaluated by co-culturing primary NK cells and K562 tumor cells. The cells were co-cultured in the presence of isotype control bispecific antibody multi-287, anti-PVRIG antibody, multi-72 (a human IgGl isotype bispecific protein targeting GFP on Fab and PD-L1 on MAb domains), or PVRIG x PD-L1 binding proteins multi-61, multi-216, multi-246 and multi- 256. The secreted IFN-gamma was measured by ELISA and the fold change of IFN- gamma in each sample was calculated in comparison to multi-72. The anti-PVRIG antibody and the PVRIG x PD-L1 binding proteins induce a PVRIG dependent enhancement of the NK cell activation towards K562 tumor cells.

[0030] Figure 16 Functionality of the PVRIG x PD-L1 binding proteins in different concentrations was evaluated by co-culturing NK-92 cells and K562 tumor cells. The cells were co-cultured in the presence of an anti-PVRIG antibody (SOC35), an anti-PD-Ll antibody (Atezolizumab / Tecentriq), a combination of the anti-PVRIG and anti-PD-Ll antibodies and PVRIG x PD-L1 binding proteins multi-216, multi-246, multi-256 and multi-61 in different concentrations. The secreted IFN-gamma was quantified, IFN-gamma concentration in human IgGl sample was normalized to 1 and fold change of IFN-gamma was calculated for each sample. Both PVRIG x PD-L1 binding proteins and the anti- PVRIG antibody induced IFN-gamma production in the cells. The anti-PD-Ll antibody (human IgGl type) alone does not increase IFN-gamma secretion in the NK-92 cells. The combination of both anti-PVRIG and anti-PD-Ll antibodies also shows a similar increase in IFN-gamma secretion as the anti-PVRIG monospecific antibody and the PVRIG x PD- L1 binding proteins multi-216, multi-246 and multi -256. EC50 values were calculated for the PVRIG x PD-L1 binding proteins multi-216, multi-246 and multi-256 being 0,5022 nM, 0,5295 nM and 0,5317 nM, respectively.

[0031] Figure 17 The OKT3 dual targeting assay was carried out to evaluate the synergistic effect of PVRIG x PD-L1 binding proteins. Primary CD4+ T cells isolated from healthy donor PBMCs (CD4+ T cell Isolation Kit (human) (Miltenyi Biotec) that express PVRIG were used as effector cells. MDA-MB-231 cell line that express PD-L1 and PVRL2 on their surface was used as target cells. MDA-MB-231 cells were transfected to express OKT3 clone anti-CD3 antibody on their surface. CD4+ T cells and MDA-MB- 231 cells expressing anti-CD3 antibody on their surface were co-cultured for 48h in the presence of an anti-PVRIG antibody, an anti-PD-Ll antibody Tecentriq (atezolizumab), their combination or PVRIG x PD-L1 binding proteins. After the co-culture IL-2 cytokine levels in cell culture mediums were measured by ELISA (ELISA Pro: Human IL-2 kit, Mabtech). IL-2 secretion indicates activation of the T cells.

[0032] Figure 18 PVRIG and PD-L1 expression on anti-CD3 stimulated and nonstimulated T cells, NK cells, B cells, and monocytes. Percentage of PVRIG and PD-L1 expressing cells was measured with flow cytometry.

[0033] Figure 19 The binding of an anti-PVRIG antibody, anti-PD-Ll antibody Tecentriq (atezolizumab) and PVRIG x PD-L1 binding proteins multi-216, multi-246 and multi-256 to specific immune cell populations was evaluated in anti-CD3 stimulated CD3+ T cells, CD14+ monocytes, CD19+ B cells and CD56+ NK cells with increasing concentrations. The anti-PVRIG and anti-PD-Ll antibodies and PVRIG x PD-L1 binding proteins contained biotin in their Fc domain. Binding was detected by adding streptavidin conjugated fluorochromes and flow cytometry. The figures represent percentages of positive cells in cell population.

[0034] Figure 20 Binding of PVRIG x PD-L1 binding proteins with different concentrations to PVRIG and PD-L1 on cell surface was studied with CHO cells overexpressing PVRIG or PD-L1. Figure 20A shows binding of PVRIG x PD-L1 binding proteins multi-216, multi-224, multi-226, multi-236, multi-244, multi-246 and multi-256 to PVRIG on CHO cells. Binding of anti-PVRIG antibody and anti-PD-Ll antibody was also studied. PVRIG x PD-Llbinding proteins multi-216, multi-226, multi-236, multi-246 and multi-256 showed higher binding to PVRIG than anti-PVRIG antibody, while multi-224 and multi-244 had lower binding. Figure 20B shows the binding of PVRIG x PD-L1 binding proteins multi-216, multi-224, multi -244, multi -226, multi -236, multi-246, multi- 256 and multi-61 to PD-L1 on CHO cells. Binding of anti-PD-Ll antibody was also studied. All the PVRIG x PD-L1 binding proteins bound similarly to PD-L1 as the anti- PD-Ll antibody. Figure 20C shows binding of PVRIG x PD-L1 binding proteins multi- 216, multi-224, multi-226, multi-236, multi-244, multi-246 and multi-256 to breast cancer cell line MDA-MB-231 that expresses PD-L1 but not PVRIG.

[0035] Figure 21 Tumor cell lines MDA-MB-231 (breast cancer), Pane 05.04 (pancreas adenocarcinoma) HTC116 (colon cancer) that express PD-L1 but not PVRIG were stained with anti -PVRIG- APC and anti-PD-Ll -APC antibodies (Biolegend) for flow cytometry. Figure 21 A shows that MDA-MD-231 cells have high PD-L1 expression, whereas Pane 05.04 cells have intermedial PD-L1 expression and HCT116 cells have low PD-L1 expression. Figures 21B to 21D show binding of PVRIG x PD-L1 binding proteins multi- 216, multi-246, multi-256 and multi-61 with different concentrations to PD-L1 expressed on MDA-MB-231, Pane 05.04 and HTC116 cells, as studied by flow cytometry. All the tested PVRIG x PD-L1 binding proteins multi-216, multi-246, multi-256 and multi-16 as well as anti-PD-Ll antibody Tecentriq (atezolizumab) bound to the cells in a dosedependent fashion and the binding correlated with the PD-L1 expression levels on the tumor cells. Anti-PVRIG antibody and isotype control bispecific antibody multi-287 did not bind to the cells.

[0036] Figure 22 Anti-PVRIG antibody, anti-PD-Ll antibody Tecentriq and PVRIG x PD- L1 binding proteins multi-61, multi-246 and multi-256 binding to Fc-gamma receptor family members was studied on CHO cells expressing Fc gamma RI (A), Fc gamma Rlla (Hl 31) (C), Fc gamma Rlla (R131) (D), Fc gamma Rllb (E) and Fc gamma Rllla (F), and in control cells (B). Concentration series was 0.006 nM, 0.024 nM, 0.098 nM, 0.39 nM, 1.56 nM, 6.3 nM, 25 nM, 100 nM, and 400 nM.

[0037] Figure 23 Enhanced functional activity effect of the PVRIG x PD-L1 binding proteins was studied in co-culture of autologous memory T cells and monocytes. The two immune cell populations were co-cultured in the presence of a hlgGl antibody (isotype control), an anti-PVRIG antibody (SOC35), an anti-PD-Ll antibody (Tecentriq, atezolizumab), a combination of the anti-PVRIG and anti-PD-Ll antibodies, and PVRIG x PD-L1 binding proteins multi-216, multi-246, multi-256 and multi-61. Each molecule was studied in six different concentrations 69 nM, 13.8 nM, 2.8 nM, 0.55 nM, 0.11 nM, and 0.02 nM (In the figure from left to right) At the end of the co-culture the IFN-gamma secretion to the cell culture medium was measured with ELISA. When the anti-PD-Ll antibody, combination of the anti-PD-Ll and anti-PVRIG antibodies and the PVRIG x PD- L1 binding proteins are applied to the mixed lymphocyte reaction a functional increase on memory T cell activation in a dose-dependent fashion is observed. However, the anti- PVRIG antibody alone does not have effect on memory T cell activation. Memory T cell activation in this assay is dependent on PD-L1 blocking. The increase of IFN-gamma fold change is calculated and shown normalized to hlgGl antibody. Based on the concentration series EC50 values were calculated for all the PVRIG x PD-L1 binding proteins multi-216, multi-246, and multi-256 with EC50 values 0,5965 nM, 0,4389 nM, and 0,7442 nM, respectively.

[0038] Figure 24 Mixed lymphocyte reaction was used to assess the immune response augmentation to tissue mismatched reaction. PBMCs and monocyte-derived dendritic cells from different donors were co-cultured in the presence of different concentrations of an anti-PD-Ll antibody (Tecentriq, atezolizumab) and PVRIG x PD-L1 binding proteins multi-61, multi-216, multi-246, and multi-256. In the upper figures PBMCs isolated from a sample of one donor were cultured with monocyte-derived dendritic cells isolated from samples of two different donors. In lower figures monocyte-derived dendritic cells isolated from a sample of one donor were cultured with PBMCs isolated from samples of two different donors. After the co-culture 4- IBB (CD 137) protein on cell surface was measured with flow cytometry to assess the activation of the cells. Addition of PVRIG x PD-L1 binding proteins increase 4- IBB expression on the NK cell surface dose-dependently, whereas anti-PD-Ll antibody Tecentriq (atezolizumab) does not affect to the expression of 4-1BB on the NK cells. This effect is likely to be attributed to active Fc domain present in the antibody and PVRIG x PD-L1 binding proteins that is binding to the FcgRIIIa on NK cells. All the selected PVRIG x PD-L1 binding proteins multi-216, multi-246, multi-256 and multi-61 increase the 4-1BB expression, showing the effect on NK cells in this assay. Multi-246 showing the highest effect.

[0039] Figure 25 Antibody-dependent cellular cytotoxicity (ADCC) was studied in CHO cells expressing PVRIG or PD-L1 in the presence of anti-PVRIG antibody, anti-PD-Ll antibody, combination of anti-PVRIG and anti-PD-Ll antibodies, PVRIG x PD-L1 binding proteins multi-61, multi- 108, multi-216, multi -224, multi -226, multi-244, multi -246, and multi-256 and multi-72 (a human IgGl isotype bispecific protein targeting GFP on Fab and PD-L1 on MAb domains) and isotype control bispecific antibody multi-287. The ADCC was detected with a PD-L1 antibody and PVRIG x PD-L1 binding proteins multi-61, multi-216, multi-224, multi-226, multi-244, multi-246 and multi-256 (PD-L1 binding domain in Mab position and PVRIG binding domain in fab position) in CHO cells expressing PD-L1 (figures 25A and 25C). The ADCC also was detected in CHO cells expressing PVRIG, when anti-PVRIG antibody or multi- 108 (PVRIG binding domain in Mab position and PD-L1 binding domain in fab position) was present. However, ADCC was not detected in PVRIG expressing CHO cells in the presence of PVRIG x PD-L1 binding proteins in which PD-L1 binding domain in Mab position and PVRIG binding domain in fab position) (figures 25B and 25D). The symbols for each sample in figure 25A are the same as in figure 25B.

[0040] Figure 26 The PVRIG x PD-L1 binding proteins multi-61, multi-216, multi-246 and multi-256, a combination of anti-PVRIG and anti-PD-Ll antibodies (Combo (Tec + SOC35)), and anti-CD52 antibody Campath (alemtuzumab) were incubated in the PBMCs. Then, the number of CD3+, CD4+, CD8+ and dead cells cells (in pl volume per cell populations) were analysed by flow cytometry to detect cell depletion compared the IgGl control sample and sample without treatment. Figure 26A shows that Campath caused a pronounced cell death effect, which was almost 6-fold as median compared to all tested PVRIG x PD-L1 binding proteins and control IgG. In CD3+, CD4+, and CD8+ cell populations Campath clearly depleted cells more than any of the studied PVRIG x PD-L1 binding proteins. There was minimal decrease in CD3+, CD4+ and CD8+ cells with PVRIG x PD-L1 binding proteins multi-61, multi-216 and multi-246. PVRIG x PD-L1 binding protein multi-256 did not reduce CD3+, CD4+ and CD8+ cells compared to sample without treatment or with IgGl isotype. CD3+ cells in figure 26B, CD4+ cells in figure 26C, and CD8+ cells in figure 26D. Figure 27 Schematic presentation of PVRIG x PD-L1 binding proteins bringing PVRIG expressing T or NK cells together with PD-L1 expressing antigen presenting cells (APCs) or tumor cells (figure 27A). The combination of anti-PVRIG and anti-PD-Ll antibodies does not cause this effect (figure 27B).

[0041] Figure 28 Cluster formation induced by PVRIG x PD-L1 binding proteins was studied with CH0-K1 cells engineered to overexpress either PVRIG or PD-L1. CH0-K1 cells expressing PVRIG were intracellularly stained with CFSE and CH0-K1 cells expressing PD-L1 with CellTrace Violet. The cells were combined and incubated together with by PVRIG x PD-L1 binding proteins multi-61, multi-246 and multi-256 in a concentration series ranging from 0.02 nM to 930 nM. Cells were analysed by flow cytometry by analyzing all live cells and gated doublets (the clustered cells). The cell gates and their flow cytometry analysis results are shown in figure 28A. The flow cytometry events that appear in concomitant colors represent the two CHO cell lines that expresses PD-L1 and PVRIG that have been clustered. The clustering assay titrates the concentration when PVRIG x PD-L1 binding proteins start to cluster the PVRIG expressing and PD-L1 expressing cells (figure 28B). The EC50 values were calculated for multi-246, multi-256 and multi-61. Clustering peak for multi-246 is at 0.36 nM, for multi-256 at 0.49 nM and for multi-61 at 0.82 nM.

[0042] Figure 29 Individual tumor growth curves for the mice treated with isotype control antibody, anti-PD-Ll antibody Tecentriq, combination of anti-PD-Ll hlgGl and anti- PVRIG hlgGl, PVRIG x PD-L1 binding proteins multi-246 and multi -256 and anti- PVRIG hlgGl antibody. Tumor volumes, expressed in mm3, are shown as a function of time (days post randomization). Individual values are shown. n=l 1 mice per group. Dotted vertical line indicates days of dosing.

[0043] Figure 30 Average tumor growth in mice groups treated with isotype control antibody, anti-PD-Ll antibody Tecentriq, combination of anti-PD-Ll hlgGl and anti- PVRIG hlgGl, PVRIG x PD-L1 binding proteins multi-246 and multi -256 and anti- PVRIG hlgGl antibody. Average tumor volumes, expressed in mm3, are shown as a function of time (days post randomization). Mean values and standard error means (SEM) are shown. n=l 1 mice per group. A two-way ANOVA followed by a Dunnett post-test was used to compare the treated groups to the isotype control treated group. * p < 0.05. Figure 31 Gene expression in tumors from the mice treated with isotype control antibody, anti-PD-Ll antibody Tecentriq, combination of anti-PD-Ll hlgGl and anti- PVRIG hlgGl (Combo), PVRIG x PD-L1 binding proteins multi-246, multi-256, and multi-61. Expression of selected genes CXCL9, CXCL10, CXCL12, TIGIT, LAG3, CTLA4, PRF1, and GZMB in different study groups is represented in the figure.

[0044] Description of the Sequences

[0045] SEQ ID NOs 1-60 are the amino acid sequences of CDRs of PVRIG binders.

[0046] SEQ ID NOs 61-120 are the amino acid sequences of CDRs of PD-L1 binders.

[0047] SEQ ID NOs 121-140 are the amino acid sequences of heavy and light chain variable domains of PVRIG binders.

[0048] SEQ ID NOs 141-160 are the amino acid sequences of heavy and light chain variable domains of PD-L1 binders.

[0049] SEQ ID NOs 161-169 are the amino acid sequences of heavy and light chains of particular multispecific antigen-binding proteins.

[0050] SEQ ID NOs 170-171 are the amino acid sequences of heavy and light chain variable domains of a reference anti-PVRIG antibody.

[0051] SEQ ID NOs 172-173 are the amino acid sequences of heavy and light chain variable domains of a reference anti-PD-Ll antibody.

[0052] SEQ ID NOs 174-176 are the amino acid sequences of heavy and light chains of a multispecific antigen-binding protein based on reference anti-PVRIG and anti-PD-Ll antibodies.

[0053] SEQ ID NOs 177-178 are the amino acid sequences of heavy and light chains of a reference anti-PVRIG antibody.

[0054] SEQ ID NOs 179-180 are the amino acid sequence of heavy and light chains of a reference anti-PVRIG antibody.

[0055] SEQ ID NOs 181-219 are the amino acid sequences of heavy and light chains of particular multispecific antigen-binding proteins.

[0056] SEQ ID NOs 220-228 are the nucleotide sequences of heavy and light chains of particular multispecific antigen-binding proteins.

[0057] SEQ ID NO: 229 is the amino acid sequence of a linker. SEQ ID NO: 230 is a human PD-L1 amino acid sequence.

[0058] SEQ ID NO: 231 is a human PVRIG amino acid sequence.

[0059] Detailed Description

[0060] General definitions

[0061] It is to be understood that different applications of the disclosed antigen-binding proteins of the invention may be tailored to the specific needs in the art. It is also to be understood that the terminology used herein is for the purpose of describing particular embodiments of the invention only, and is not intended to be limiting.

[0062] In addition, as used in this specification and the appended claims, the singular forms “a”, “an”, and “the” include plural references unless the content clearly dictates otherwise. Thus, for example, reference to “an antigen-binding protein” includes two or more “anti gen -binding proteins”.

[0063] It is to be understood that the term “nucleic acid” and the term “polynucleotide” are used interchangeably herein.

[0064] For the purpose of this invention, in order to determine the percent identity of two sequences (such as two nucleic acids or two nucleic acids sequences), the sequences are aligned for optimal comparison purposes (e.g. gaps can be introduced in a first sequence for optimal alignment with a second sequence). The nucleotide or amino acid residues at each position are then compared. When a position in the first sequence is occupied by the same nucleotide or amino acid as the corresponding position in the second sequence, then the nucleotides or amino acids are identical at that position. The percent identity between the two sequences is a function of the number of identical positions shared by the sequences (i.e., % identity = number of identical positions / total number of positions in the reference sequence x 100).

[0065] Typically, the sequence comparison is carried out over the length of the reference sequence. For example, if the user wished to determine whether a given (“test”) sequence is 95% identical to SEQ ID NO: 1, SEQ ID NO: 1 would be the reference sequence. To assess whether a sequence is at least 95% identical to SEQ ID NO: 1 (an example of a reference sequence), the skilled person would carry out an alignment over the length of SEQ ID NO: 1, and identify how many positions in the test sequence were identical to those of SEQ ID NO: 1. If at least 95% of the positions are identical, the test sequence is at least 95% identical to SEQ ID NO: 1. If the sequence is shorter than SEQ ID NO: 1, the gaps or missing positions should be considered to be non-identical positions.

[0066] The skilled person is aware of different computer programs that are available to determine the homology or identity between two sequences. For instance, a comparison of sequences and determination of percent identity between two sequences can be accomplished using a mathematical algorithm. In an embodiment, the percent identity between two amino acid or nucleic acid sequences is determined using the Needleman and Wunsch (1970) algorithm which has been incorporated into the GAP program in the Accelrys GCG software package (available at http: / / www.accelrys.com / products / gcg / ), using either a Blosum 62 matrix or a PAM250 matrix, and a gap weight of 16, 14, 12, 10, 8, 6, or 4 and a length weight of 1, 2, 3, 4, 5, or 6.

[0067] By “specific for” or "specifically binds", it is typically meant that the relevant antigen-binding domain of an antigen-binding protein binds to one or more antigenic determinants of the desired antigen and does not bind or does not bind significantly to other polypeptides. For example, an antigen-binding domain that specifically binds PD-L1 may bind to an antigen of PD-L1 but not bind to an antigen of a different polypeptide such as bovine serum albumin. An antigen-binding domain may specifically bind if it binds to the desired antigen with a stronger affinity when compared to binding an antigen of a different polypeptide such as bovine serum albumin. Binding affinity may be quantified by determining the dissociation equilibrium constant (KD), dissociation rate constant (Kd) and / or association rate constant (Ka) for an antibody and its target. Similarly, the specificity of binding of an antibody to its target may be defined in terms of the comparative binding affinity of the antibody for its target as compared to binding affinity with respect to the antibody and another, non-target molecule. The dissociation equilibrium constant (KD) (M) refers to the dissociation equilibrium constant of a particular antigen-binding protein: antigen interaction. In particular, the KD represents the ratio of the free concentrations of antigen-binding protein and antigen to the concentration of the antigen-binding protein: antigen complex (at equilibrium). There is an inverse relationship between KD and binding affinity; thus the smaller the KD, the higher, i.e., stronger, the affinity. The dissociation rate constant (Kd) (sec -1 or 1 / s) refers to the dissociation rate constant of a particular antigen-binding protein: antigen interaction, and may also be referred to as koff. The association rate constant (Ka) (M-l x sec -1 or 1 / M) refers to the association rate constant of a particular antigen-binding protein: antigen interaction, and may also be referred to as kon. Methods for measuring the affinity of binding are well known in the art. The affinity of binding may be measured for example by surface plasmon resonance or bio-layer interferometry. For instance, binding affinity may correspond to a KD value of about 10'7M or less, such as about 10'8M or less, such as about 10'9M or less when determined by, for instance, surface plasmon resonance using the antigen as ligand and the antigen-binding protein as analyte.

[0068] As used herein, the term “about” may be interpreted to mean a value within + / - 10% of the recited value.

[0069] All publications, patents and patent applications cited herein, whether supra or infra, are hereby incorporated by reference in their entirety.

[0070] Multispecific antigen-binding protein

[0071] The invention relates to an antigen-binding protein that is multispecific in that it is specific for multiple antigens. The multispecific antigen-binding protein may specifically bind multiple antigens. The multispecific antigen-binding protein typically comprises multiple antigen-binding domains. The multiple antigen-binding domains typically comprise different antigen-binding domains specifically binding different target antigens. It should be understood that the multispecific antigen-binding protein may comprise more than one antigen-binding domain for each target antigen, such as at least two copies of an antigen-binding domain for each target antigen. The multispecific antigen-binding protein is specific for or specifically binds at least PD-L1 and PVRIG. The multispecific antigenbinding protein thus typically comprises an antigen-binding domain that specifically binds PD-L1 (also described herein as a “first” antigen-binding domain) and an antigen-binding domain that specifically binds PVRIG (also described herein as a “second” antigen-binding domain). The multispecific antigen-binding protein may also be specific for one or more additional antigens. The multispecific antigen-binding protein may be a bispecific antigenbinding protein specific for PD-L1 and PVRIG. The multispecific antigen-binding protein may be a trispecific antigen-binding protein specific for PD-L1, PVRIG and an additional antigen.

[0072] Antigen-binding domains and configurations

[0073] The multispecific antigen-binding protein may be selected from any class of antigen-binding protein. Typically, an antigen-binding protein is any protein that comprises an antigen-binding site or domain. The antigen-binding protein may comprise an antibody or antibody derivative, or an antigen-binding fragment of either thereof. Alternatively, the antigen-binding protein may comprise a synthetic antigen-binding scaffold or synthetic antibody, for example an Alphabody, Affibody, Affitin, Anticalin, Monobody or Adnectin.

[0074] An antigen-binding domain may be selected from any domain able to specifically bind a target antigen. Preferably, an antigen-binding domain comprises one or more immunoglobulin variable domains. An anti gen -binding domain may comprise a heavy chain variable domain and / or a light chain variable domain. Each immunoglobulin variable domain typically comprises three complementarity determining regions (CDRs). The CDRs typically are responsible for antigen specificity (e.g. by making direct interactions with the antigen). Exemplary conventions that can be used to identify the boundaries of CDRs include the Kabat definition, the Chothia definition and the IMGT definition (see, for example, Kabat, Elvin Abraham. Sequences of proteins of immunological interest. No. 91. US Department of Health and Human Services, Public Health Service, National Institutes of Health, 1991; Lefranc, Marie-Paule, et al. "IMGT unique numbering for immunoglobulin and T cell receptor variable domains and Ig superfamily V-like domains." Developmental & Comparative Immunology 27.1 (2003): 55-77). The immunoglobulin variable domains may also comprise framework regions, which provide the immunoglobulin-like structure of the domain and typically do not make direct interactions with an antigen.

[0075] An antibody is typically an immunoglobulin molecule and may be selected from a monoclonal antibody or a polyclonal antibody. A multispecific antibody may be a bispecific or trispecific antibody, such as bispecific antibody specific for PD-L1 and PVRIG. An antigen-binding domain or antibody fragment may be selected from an scFv (typically arranged VH-VL in an N- to C-terminal orientation), Fab, a Fab’, a F(ab’)2 fragment, a heavy chain variable domain (VH), a light chain variable domain (VL), a single domain antibody (sdAb), a nanobody (VHH), or a VNAR domain. An antigen-binding protein is typically human or humanized.

[0076] An antigen-binding protein may comprise a complex of multiple antigen-binding domains or antibody fragments. For example, an antigen-binding protein may comprise a monoclonal antibody specific for a first antigen and a Fab fragment specific for a second antigen. An antigen-binding protein may comprise a fusion polypeptide comprising multiple antigen-binding domains in whole or in part. For example, a multispecific antigen-binding protein may comprise a fusion polypeptide comprising a heavy chain variable region of a first antigen-binding domain and a light or heavy chain variable region of a second antigen-binding domain.

[0077] The antigen-binding proteins described herein may comprise one or more constant regions. The constant regions may be heavy or light chain constant regions. Preferably, the constant regions are human. A heavy chain constant region may comprise one or more constant domains selected from CHi, CH2 and CH3. Preferably, a heavy constant region comprises an Fc region. A heavy chain constant region may be of any isotype and any subclass. The isotype may be IgM, IgD, IgA, IgE or IgG. A preferred isotype is IgG. The IgG may be an IgGl, IgG2, IgG3 or IgG4. A particularly preferred subclass is IgGl. The light chain constant region may be a kappa or a lambda constant region. Constant domains may be directly linked to one another or may be linked by a full or partial hinge or linker region.

[0078] The antigen-binding proteins described herein may comprise one or more mutations at interfaces between polypeptides (typically heavy and light chains) to promote heterodimerisation of heavy and light chains. The mutations may be present in a heavy chain variable region, a light chain variable region, a CHi domain and / or a light chain constant region. The skilled person is aware of a range of suitable such mutations. Suitable mutations may for example be any mutations described in WO2020127354 and Nyesiga et al, incorporated herein by reference, and the specific mutations discussed below.

[0079] A preferred multispecific antigen-binding protein according to the invention comprises a polypeptide comprising a first heavy chain polypeptide specific for a first antigen fused to a second light chain polypeptide specific for a second antigen. The second light chain polypeptide also forms part of a Fab fragment comprising a second heavy chain polypeptide specific for the second antigen. The multispecific antigen-binding protein further comprises a light chain polypeptide pairing with the first heavy chain polypeptide. The multi-specific antigen-binding protein thus typically comprises an antibody specific for a first antigen (comprising the first heavy and first light chain polypeptides) and a Fab fragment specific for a second antigen (comprising the second heavy chain polypeptide and second light chain polypeptide).

[0080] In all of the above aspects, the first antigen is preferably PD-L1 and the second antigen PVRIG. It has been surprisingly identified by the inventors that multi-specific binding proteins of the above configuration (which is described in more detail below) have optimal properties. In particular, the use of an antibody specific for PD-L1 (comprising an effector region) in the above multi-specific antigen-binding protein configuration optimally allows for PD-L1 -dependent ADCC, whereas an antibody specific for PVRIG in the same multi-specific antigen-binding protein configuration is less effective at inducing ADCC. The affinity of the multi-specific antigen-binding protein for PD-L1 may advantageously be higher when the first antigen-binding domain (typically an antibody specific for a first antigen) is specific for PD-L1. In contrast, the affinity for PD-L1 may be decreased, with corresponding impact on PD-L1 associated functions, when the second antigen-binding domain (typically a Fab fragment specific for a second antigen) is specific for PD-L1. It is thus preferred that the first antigen-binding domain be specific for PD-L1 and the second antigen-binding domain specific for PVRIG.

[0081] In more detail, a preferred multi-specific antigen-binding protein of the invention may comprise a first antigen-binding domain specific for PD-L1 fused to a second antigenbinding domain specific for PVRIG. In this aspect, the first antigen-binding domain may be an antibody, comprising an effector region, typically an Fc region (preferably a monoclonal antibody (mAb)) and the second antigen-binding domain a Fab or an scFv, preferably a Fab. The first antigen-binding domain of the preferred multi-specific antigenbinding protein may be described herein as being in the mAb position and the second antigen-binding domain as being in the Fab position. The first antigen-binding domain may comprise a first heavy chain polypeptide and a first light chain polypeptide. The second antigen-binding domain may comprise a second heavy chain polypeptide and a second light chain polypeptide. The second antigen-binding domain may be fused to the first antigen binding domain via the second heavy or light chain polypeptide. Thus, the first heavy chain polypeptide may be fused to the second heavy or light chain polypeptide. Preferably, the second light chain polypeptide is C-terminally fused to the first heavy chain polypeptide.

[0082] In a particularly preferred embodiment, the first antigen-binding domain is an antibody, the second antigen-binding domain is a Fab, and the second light chain polypeptide is C-terminally fused to the first heavy chain polypeptide. The heavy chain polypeptide preferably comprises a constant region, typically comprising an Fc region. The constant region is preferably an IgG, particularly preferably IgGl. The first heavy chain polypeptide and the second light chain polypeptide may be directly fused or may be fused via a linker sequence. Any suitable linker may be used to link the first and second antigenbinding domains. The linker is preferably non-immunogenic. The linker may comprise glycine and / or serine residues. The linker may comprise multiple repeats of an amino acid sequence comprising glycine and / or serine residues. A preferred linker comprises GGGGS (SEQ ID NO: 229). A linker may comprise multiple repeats of SEQ ID NO: 229, such as two, three, four or five copies of SEQ ID NO: 229. The multiple repeats are typically directly fused to one another. A preferred linker comprises three repeats of SEQ ID NO: 229 directly fused to one another. Thus, the heavy chain polypeptide may comprise, in the direction N-C, a heavy chain variable region specific for PD-L1, a heavy chain constant region (preferably IgGl), optionally a linker, a light chain variable region specific for PVRIG, and optionally a light chain constant region.

[0083] The above preferred multi-specific antigen-binding protein is typically tetravalent. The above preferred multi-specific antigen-binding protein typically comprises an antibody binding a first target and two Fabs binding a second target. The above preferred multispecific antigen-binding protein typically comprises three different polypeptide chains. The first polypeptide chain comprises a heavy chain polypeptide of the first antigen-binding domain fused to a light chain polypeptide of the second antigen-binding domain. The second polypeptide chain comprises a light chain that pairs with the heavy chain polypeptide of the first antigen-binding domain. The third polypeptide chain comprises a heavy chain polypeptide that pairs with the light chain polypeptide of the second antigenbinding domain. The multi-specific antigen-binding protein typically comprises two copies of each of the three different polypeptide chains. The above preferred multi-specific antigen-binding protein may comprise one or more mutations at interfaces between heavy and light chains to promote heterodimerisation or to decrease immunogenicity. The first polypeptide chain may comprise one or more of the mutations Q44R in the heavy chain variable region or a mutation of any residue at position 44 in the heavy chain variable region to R, Q44R in the light chain variable region or a mutation of any residue in the light chain variable region at position 44 to R, Hl 68 A and / or F170G in a CHi domain, and / or T109P, S114A / T114A, and / or N137K / S137K in a kappa light chain constant region. The first polypeptide chain may comprise all of the above mutations. Preferably, the first polypeptide chain comprises the mutation Q44R in the heavy chain variable region, Q44R in the light chain variable region, the mutations H168A and / or F170G in a CHi domain, and the mutations SI 14A and N137K in a kappa light chain constant region. The second polypeptide chain may comprise one or more of the mutations Q44E in the light chain variable region or a mutation of any residue at position 44 in the light chain variable region to E, and / or L135Y and / or S176W in a kappa light chain constant region. The second polypeptide chain may comprise all of the above mutations. Preferably, the second polypeptide chain comprises the mutation Q44E in the light chain variable region and the mutations L135Y and S176W in a kappa light chain constant region. The third polypeptide chain may comprise one or more of the mutations Q44E in the heavy chain variable region or a mutation of any residue at position 44 in the heavy chain variable region to E, and / or S183T and / or T187E in a CHi domain. The third polypeptide chain may comprise all of the above mutations. Preferably, the third polypeptide chain comprises the mutation Q44E in the heavy chain variable region and the mutation T187E in a CHi domain.

[0084] The residue numbering for mutations in heavy chain and light chain variable regions described above is typically IMGT numbering. The residue numbering for mutations in constant domains described above is typically EU numbering.

[0085] Other preferred configurations for multispecific antigen-binding proteins are described in WO2020127354 and Nyesiga et al., the configurations of each of which are incorporated herein by reference.

[0086] Functional properties The multispecific antigen-binding protein of the invention may target PD-L1 on a first cell and PVRIG on a second cell. The cells are typically mammalian, preferably human cells. Typically, the first cell is a cancer cell or antigen-presenting cell. Typically, the first cell has increased PD-L1 expression, typically increased PD-L1 surface expression. The first cell may additionally or alternatively have increased PVRL2 expression. A cancer cell may be a cell of a solid cancer or tumour or of a liquid cancer, such as a haematological cancer. The solid cancer or tumour may be prostate cancer, breast cancer, lung cancer, colorectal cancer, melanoma, bladder cancer, brain / CNS cancer, cervical cancer, oesophageal cancer, gastric cancer, head / neck cancer, kidney cancer, liver cancer, lymphoma, ovarian cancer, pancreatic cancer or sarcoma. The solid cancer or tumour may preferably be breast cancer, gastric cancer, non-small cell lung cancer, renal cell carcinoma or melanoma. The breast cancer may be triple negative breast cancer. The haematological cancer may be myeloid leukemia, such as acute myeloid leukemia. An antigen-presenting cell is typically a professional antigen-presenting cell and may be a monocyte, a macrophage, a B cell or a dendritic cell. The antigen-presenting cell is preferably a dendritic cell. A dendritic cell may be a monocyte-derived dendritic cell, for example a CD14+ CD1 lc+ monocyte-derived dendritic cell. Typically, the second cell is an immune cell, typically any immune cell expressing PVRIG, preferably a T cell or an NK cell. The first cell may be a cancer cell and the second cell a T cell or NK cell. The first cell may be a dendritic cell and the second cell a T cell or NK cell.

[0087] As discussed above, the antigen binding profile of the multi-specific antigen binding protein may allow for reduced immunosuppression by a cancer cell through targeting immunosuppressive pathways involving PD-L1 and PVRIG. Targeting of immune cells such as T and NK cells may also allow for activation of an anti-cancer immune response. Engagement of a first antigen-presenting cell expressing PD-L1 by the multi-specific antigen binding protein may assist in presentation of tumor associated antigens to a second immune cell such as a T cell. Additionally or alternatively, engagement of PD-L1 on APCs may alleviate or prevent APC-mediated immune suppression, such as APC-mediated inhibition of T cells. Engagement of PD-L1 on APCs may thus enhance T cell responses. The inventors consider that this effect may contribute to the enhanced memory T cell responses seen with PD-L1 blocking in the autologous MLR assay with co-culture of memory T cells and monocytes in the examples (Example 9).

[0088] Accordingly, the multispecific antigen-binding protein may bind to PD-L1 on a cancer or antigen-presenting cell. The multispecific antigen-binding protein may bind to PVRIG on an immune cell. The immune cell may be a T cell or an NK cell. The multispecific antigen-binding protein may bind to PD-L1 on a cancer cell and PVRIG on a T cell. The multispecific antigen-binding protein may bind to PD-L1 on a cancer cell and PVRIG on an NK cell. The multispecific antigen-binding protein may bind to PD-L1 on an antigen-presenting cell and PVRIG on a T cell. The multispecific antigen-binding protein may bind to PD-L1 on an antigen-presenting cell and PVRIG on an NK cell.

[0089] The multispecific antigen-binding protein may activate T cells. The T cells may be CD4+ or CD8+ T cells. The T cells may be gamma delta T cells or NKT cells, such as CD3+CD56+ NKT cells. The T cells may be memory cells, such as CD45RO+ T cells. The T cells may be primary T cells or may be present in a PBMC sample. CD4+ T cells may be activated by co-culture with a target cell, typically a cancer cell. The target cell may be a PD-L1 expressing cell. The target cell may further express PVRL2. Memory T cells may be activated by co-culture with monocytes or antigen-presenting cells, such as dendritic cells. The dendritic cells may be monocyte-derived dendritic cells.

[0090] In a preferred embodiment, the T cell activation is PD-L1 dependent. The T cell activation may not be PVRIG-dependent. The dependence of a given function of a multispecific binding protein on a particular target (i.e. PD-L1 or PVRIG) may be determined by the skilled person in any suitable way. For example, a control multi-specific antigen-binding protein may be provided comprising a first antigen-binding domain specific for one target of interest and a second antigen-binding domain specific for a control protein. The control protein is typically a protein not present in the functional assay, for example GFP. An example of such a protein is multi-72, as described in the examples. The functional effect of the control protein is then compared to that of a test multi-specific antigen-binding protein comprising the same first antigen-binding domain and a second antigen-binding domain for the second target of interest. If the test multispecific antigen-binding protein has enhanced function compared to that of the control multi-specific antigen-binding protein, it can be inferred that the relevant function is dependent on the second target of interest. The function may be dependent on the relevant target if it is increased by at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 100%, at least 200%, at least 300%, at least 400% or more compared to a control multispecific binding protein not binding the relevant target.

[0091] The activation of T cells may be determined by any suitable assay known to the person skilled in the art, such as a cytokine release or proliferation assay. Particular assays for T cell activation by a multispecific antigen-binding protein of the invention are described in the examples and include IL-2 release for activation of CD4+ T cells and / or IFN-gamma release for activation of CD45RO+ T cells. The multispecific antigen-binding protein may increase T cell activation, such as memory T cell activation by at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 100%, at least 200%, at least 300%, at least 400% or more. The multispecific antigen-binding protein may increase cytokine release (such as release of IL- 2 and / or IFN-gamma) by at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 100%, at least 200%, at least 300%, at least 400% or more. An increase discussed above is typically with respect to the level of T cell activation in a control sample not contacted with the multispecific antigen-binding protein.

[0092] The multispecific antigen-binding protein may activate NK cells. NK cells are typically CD56+. The NK cells may be primary NK cells or may be present in a PBMC sample. The NK cells may be activated by co-culture with a target cell, typically a cancer cell. The target cell may be a PD-L1 expressing cell. The target cell may further express PVRL2. The activation of NK cells may be determined by any suitable assay known to the person skilled in the art, such as a cytokine release or proliferation assay.

[0093] In a preferred embodiment, the NK cell activation is PVRIG dependent. The NK cell activation may not be PD-L1 dependent.

[0094] Particular assays for NK cell activation by a multispecific antigen-binding protein of the invention are described in the examples and include IFN-gamma release on coculture with target cancer cells. The multispecific antigen-binding protein may increase NK cell activation, such as NK cell activation by at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 100%, at least 150%, at least 200%, or more. The multispecific antigen-binding protein may increase cytokine release (such as release of IL-2 and / or IFN-gamma) by at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 100%, at least 150%, at least 200%, or more. An increase discussed above is typically with respect to the level of NK cell activation in a control sample not contacted with the multispecific antigen-binding protein.

[0095] The multispecific antigen-binding protein may be capable of crosslinking a first cell expressing PD-L1 and a second cell expressing PVRIG. The first cell may be a cancer cell or antigen-presenting cell, such as a dendritic cell, and the second cell may be an immune cell such as a T cell or an NK cell. The multispecific antigen-binding protein may cluster cells expressing PD-L1 and cells expressing PVRIG. The multispecific antigenbinding protein may thus cluster cancer cells and immune cells, such as cancer cells and T cells and / or NK cells. The multispecific antigen-binding protein may cluster antigen- presenting cells and immune cells, such as antigen presenting cells and T cells and / or NK cells. The antigen-presenting cells may be dendritic cells. It is considered that such crosslinking or clustering may advantageously enhance anti-cancer activity. The multispecific antigen-binding protein may form an immunological synapse between cancer cells or antigen-presenting cells and immune cells, such as T cells and / or NK cells. The immunological synapse typically comprises a colocalization of the relevant cells.

[0096] Crosslinking or clustering of the above cells by the multispecific antigen-binding protein may be determined by any suitable assay known to the person skilled in the art. The cells may be co-cultured or otherwise combined or incubated together. The formation of crosslinked or clustered cells may be determined by any technique that allows observation of crosslinking or clustering. A population of cells comprising both types of cells may be identified. The different types of cells may comprise markers or labels allowing determination of their respective localisation or their co-localisation. One suitable technique is flow cytometry. A particular flow cytometry-based assay is described in the examples.

[0097] The multispecific antigen-binding protein may increase crosslinking or clustering of cancer cells and immune cells, for example cancer cells and T cells and / or NK cells, by at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 100%, or more. The multispecific antigen-binding protein may increase crosslinking or clustering of antigen-presenting cells and immune cells, such as antigen presenting cells and T cells and / or NK cells, by at least 5%, 10%, at least 15%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 100%, or more. The antigen-presenting cells may be dendritic cells. An increase discussed above is typically with respect to the level of crosslinking or clustering in a control sample not contacted with the multispecific antigenbinding protein.

[0098] The multispecific antigen-binding protein may induce any Fc-mediated antibody functionality against a target cell, such as any biological activity mediated by binding of an Fc region to an Fc receptor. The Fc receptor may be any Fc receptor, such as FcyR or FcaR. The Fc-mediated antibody functionality may be antibody-dependent cytotoxicity (ADCC) and / or antigen-dependent phagocytosis (ADCP) against a target cell. The Fc- mediated antibody functionality may be any FcR cross-linkage dependent activation of an immune cell, such as activation of B cells, macrophages, dendritic cells, cytotoxic T cells, helper T cells, NK cells and / or other myeloid cells. A preferred Fc-mediated antibody functionality is ADCC. In these aspects, the multispecific antigen-binding protein typically comprises an effector region capable of inducing the Fc-mediated antibody functionality against a target cell. The effector region is typically an Fc region. The effector region may be of any isotype discussed above. A preferred isotype is IgG. A particularly preferred subclass is IgGl. The target cell is typically a cancer cell. The target cell may be a PD-L1 expressing cell. The target cell may further express PVRL2. The multispecific antigenbinding protein is typically bound to the target cell such that its effector region may mediate an Fc-mediated antibody functionality such as ADCC.

[0099] Preferably, the multispecific antigen-binding protein induces ADCC. In a preferred embodiment, the ADCC or other Fc-mediated antibody functionality is PD-L1 dependent. The ADCC or other Fc-mediated antibody functionality may not be PVRIG-dependent. The inventors have shown that in a multi-specific antigen-binding protein where the first antigen-binding domain comprises an effector region (for example is an antibody comprising an Fc region) and the second antigen-binding domain is a Fab or an scFv, it is preferable for ADCC if the first antigen-binding domain is specific for PD-L1. It is considered that the respective locations of the antigen-binding domains for PD-L1 and PVRIG and the Fc region may determine the efficacy of ADCC or any other Fc-mediated antibody functionality. Such a multi-specific antigen-binding protein may comprise the first antigen-binding domain specific for PD-L1 fused to a second antigen-binding domain specific for PVRIG. The second antigen-binding domain may be fused to a heavy chain polypeptide of the first antigen binding domain via a heavy or light chain polypeptide of the second antigen-binding domain, preferably a light chain polypeptide. The second light chain polypeptide may be C-terminally fused to the first heavy chain polypeptide. It is considered that in the above configuration, with a PD-L1 antigen-binding domain in the specified orientation, the availability of Fc regions for binding Fc receptors (for example on dendritic cells) and inducing ADCC or any other Fc-mediated antibody functionality may be enhanced.

[0100] It is further considered that provision of a PD-L1 dependent Fc-mediated antibody functionality (such as ADCC) may advantageously allow for conditional Fc-mediated antibody functionality towards PD-L1 expressing cancer cells, without Fc-mediated antibody functionality towards immune cells expressing PVRIG. The multispecific antigen-binding protein may only induce Fc-mediated antibody functionality when bound to PD-L1. The multispecific antigen-binding protein may not induce Fc-mediated antibody functionality or may exhibit reduced Fc-mediated antibody functionality when bound only to PVRIG. The multispecific antigen-binding protein may thus conditionally or selectively target Fc-mediated antibody functionality (such as ADCC) towards cancer cells. The multispecific antigen-binding protein may selectively deplete cancer cells but not immune cells. The multispecific antigen-binding protein may conditionally or selectively target Fc- mediated antibody functionality towards cancers or tumours comprising both cancer cells expressing PD-L1 and immune cells expressing PVRIG.

[0101] The ADCC may be mediated by any immune cell which expresses a receptor for the effector region, typically an Fey receptor, and which is capable of mediating ADCC. The Fey receptor may be FcyRIIIA (also described as CD16a). The immune cell may be an NK cell or an eosinophil, preferably an NK cell.

[0102] The induction of ADCC may be determined by any suitable assay known to the person skilled in the art, such as a cell killing or cell lysis assay. The multispecific antigenbinding protein may thus lyse or kill target cells. Typically, target cells are incubated with the multi-specific antigen-binding protein. Target cells bound to or labelled with the multispecific antigen-binding protein may then be incubated or co-cultured with the immune cell capable of mediating ADCC. The ability of the immune cell to mediate ADCC may then be determined by the level of killing or lysis of the target cells. A particular ADCC assay is described in the examples and involves co-culture of NK cells and PD-L1 expressing cancer cells.

[0103] The multispecific antigen-binding protein may increase ADCC against target cells, such as increase killing or lysis of target cells, by at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 100%, at least 200%, at least 300%, at least 500%, at least 1000% or more. The increase is typically with respect to the level of ADCC against target cells not contacted with the multispecific antigen-binding protein.

[0104] The multispecific antigen-binding protein may also have one or more anti-cancer or anti-tumour effects. The multispecific antigen-binding protein may inhibit tumour cell proliferation or growth. The multispecific antigen-binding protein may increase tumour cell killing. The multispecific antigen-binding protein may increase tumour clearance or regression. The multi-specific antigen-binding may increase survival from a cancer or tumour, such as tumour-free survival. The multispecific antigen-binding protein may enhance an anti-tumour immune response. The anti-tumour immune response may comprise T cell and / or NK cell activation. The anti-tumour immune response may additionally or alternatively comprise ADCC of tumour cells. The anti-tumour response may comprise an increase in expression of one or more genes associated with an antitumour response, optionally selected from one or more, such as all, of CXCL9, CXCL10, CXCL12, PRF1 and GZMB. The increase in expression may be at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 100%, or more. The anti -turn our response may occur without T cell exhaustion. Thus, the multispecific antigen binding protein may not induce T cell exhaustion. The multispecific antigen binding protein may not induce or may not significantly induce expression of one or more genes associated with T cell exhaustion, such as TIGIT, LAG3 and / or CTLA4. The multispecific antigen binding protein may not induce or may not significantly induce expression of TIGIT and CTLA4. The above expression changes may be determined in cancer cells or in suitable in vivo models of cancer treatment, for example as illustrated in the examples. The above anti-cancer or tumour effects may be increased by at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, at least 100%, or more by the multispecific antigen-binding protein.

[0105] The multispecific antigen-binding protein may display any combination of the above functional properties. For example, the multispecific antigen-binding protein may activate T cells and may activate NK cells. The multispecific antigen-binding protein may activate T cells and induce ADCC against target cells. The multispecific antigen-binding protein may activate NK cells and induce ADCC against target cells. The multispecific antigen-binding protein may crosslink or cluster cancer cells and T cells and / or NK cells, and induce ADCC against target cells. The multispecific antigen-binding protein may crosslink or cluster cancer cells and T cells and / or NK cells and activate T cells and / or NK cells. The multispecific antigen-binding protein may crosslink or cluster antigen-presenting cells and T cells and / or NK cells and activate T cells and / or NK cells. The multispecific antigen-binding protein may inhibit tumour cell proliferation or growth and activate T cells and / or NK cells.

[0106] The multispecific antigen-binding protein may bind PD-L1 and / or PVRIG with a high binding affinity. The KD of the multispecific antigen-binding protein for PD-L1 may be less than 500nM, less than 200 nM, less than 100 nM, less than 50 nM, less than 20nM, less than 10 nM, less than 5nM, preferably less than 2nM or more preferably less than 1 nM. The KD of the multispecific antigen-binding protein for PVRIG may be less than 500nM, less than 200 nM, less than 100 nM, less than 50 nM, preferably less than 20 nM, more preferably less than 10 nM. In some embodiments, the KD of the multispecific antigen-binding protein for PVRIG less than 5nM or less than 1 nM. The KD of the multispecific antigen-binding protein for PD-L1 may be less than 5nM (preferably less than 2nM) and the KD of the multispecific antigen-binding protein for PVRIG less than 20nM.

[0107] The multispecific antigen-binding protein may have a high association rate constant (Ka) for PD-L1 and / or PVRIG. The Ka for PD-L1 may be greater than lx 105M-l x sec - 1. The Ka for PD-L1 may be in the range of lx 105- 5 x 106M-l x sec -1, such as 5 x 105- 3 x 106M-l x sec -1. The Ka for PVRIG may be greater than lx 105M-l x sec -1. The Ka for PVRIG may be in the range of lx 105- 10 x 106M-l x sec -1, such as 5 x 105to 1 x 106M-l x sec -1. The Ka for PD-L1 may be greater than lx 105M-l x sec -1 and the Ka for PVRIG greater than lx 105M-l x sec -1. The multispecific antigen-binding protein may have a low dissociation rate constant (Kd) for PD-L1 and / or PVRIG. The Kd for PD- L1 may be less than 1 x 10'21 / s, preferably less than 5 x 10'31 / s. The Kd for PD-L1 may be in the range of 1 x 10'4to 1 x 10'21 / s, such as 1 x 10'4to 5 x 10'31 / s. The Kd for PVRIG may be less than 1 x 10'21 / s, preferably of the order of 10'31 / s. The Kd for PVRIG may be in the range of 1 x 10'31 / s to 1 x 10'21 / s. The Kd for PD-L1 may be less than 5 x 10'31 / s and the Kd for PVRIG less than 1 x 10'21 / s.

[0108] The above binding affinities may be as determined by surface plasmon resonance e.g. by using a simple 1 : 1 interaction model for captured ligand (antibody) and soluble analyte (target protein). The binding affinity may be determined in a Biacore 8K+ system at a temperature of 25°C. A range of suitable binding assays for determining the KD, Ka and / or Kd of an antigen-binding protein are known to the skilled person and also described in the Examples. The above binding affinities are typically for human PD-L1 and human PVRIG. The above binding affinities may be for human PD-L1 and human PVRIG proteins comprising respectively the amino acid sequences of SEQ ID NOs 230 and 231.

[0109] The multispecific antigen-binding protein may block binding of PD-L1 to PD1 and / or CD80. The multispecific antigen-binding protein may block binding of PVRIG to PVRL2. The multispecific antigen-binding protein may block binding of PD-L1 to PD1 and binding of PVRIG to PVRL2. The multispecific antigen-binding protein may block binding of PD-L1 to CD80 and binding of PVRIG to PVRL2. The multispecific antigenbinding protein may block binding of PD-L1 to CD80 and PD-1 and binding of PVRIG to PVRL2.

[0110] Binding may be blocked by at least 10%, at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90%, substantially completely or completely by the multispecific antigen-binding protein. Blocking of binding of a given ligand to its receptor by a multispecific antigen-binding protein or antigen-binding protein specific for the receptor as described herein can be determined by any suitable in vitro assay, for example a functional report assay, an ELISA-based immunoassay or flow cytometry. The receptor may be attached to a solid support or expressed on a cell. Typically, the receptor is pre-incubated with a multispecific antigen-binding protein or antigen-binding protein described herein, and the ability of this pre-incubation to block binding of free ligand to the receptor is then determined. Suitable assays for determining blocking of PD-1 and CD80 to PD-L1, and blocking of PVRIG to PVRL2 are described for instance in Example 6.

[0111] The multispecific antigen-binding protein may specifically bind to human PD-L1 and cynomolgus PD-L1 (i.e. be cross-reactive for both species), typically with a binding affinity for each protein in the above range. The multispecific antigen-binding protein may specifically bind to human PD-L1 and cynomolgus PD-L1 with a KD of less than 50 nM, preferably less than 20nM. The multispecific antigen-binding protein may specifically bind to human PVRIG and cynomolgus PVRIG, typically with a binding affinity for each protein in the above range.

[0112] The inventors have further identified epitopes of human PD-L1 and human PVRIG bound by preferred multi-specific antigen-binding proteins. The multispecific antigenbinding protein may thus specifically bind to an epitope of human PD-L1 comprising at least one amino acid residue selected from V23, D26, E58, Ml 15, D122, Y123, K124, and R125. This residue numbering may refer to a human PD-L1 amino acid sequence as shown in SEQ ID NO: 230. The multispecific antigen-binding protein may bind to an epitope comprising corresponding amino acid residues to those specified above in any human PD- L1 amino acid sequence. The multispecific antigen-binding protein may specifically bind to an epitope which comprises two, three, four, five, six, seven or all eight of these residues. The multi-specific antigen-binding protein may bind at least two non-contiguous amino acid sequences comprising amino acid residues described above, and thus may for example bind at least one amino acid residue selected from V23 and D26 and at least one amino acid residue selected from D122, Y123, K124, and R125. Preferably, the multispecific antigen-binding protein binds to an epitope of human PD-L1 comprising at least one amino acid residue selected from D26, D122, Y123, K124, and R125. The multispecific antigen-binding protein may preferably bind to an epitope which comprise two, three, four or all five of these residues.

[0113] Alternatively, the multispecific antigen-binding protein may specifically bind to an epitope of human PD-L1 comprising at least one amino acid residue selected from Y56, E58, R113, Ml 15, SI 17, A121, and D122. The multispecific antigen-binding protein may specifically bind to an epitope which comprises two, three, four, five, six, or all seven of these residues. The multi-specific antigen-binding protein may bind at least two noncontiguous amino acid sequences comprising amino acid residues described above, and thus may for example bind at least one amino acid residue selected from V56 and E58 and at least one amino acid residue selected from R113, Ml 15, SI 17, A121, and D122. Preferably, the multispecific antigen-binding protein binds to an epitope of human PD-L1 comprising at least one amino acid residue selected from E58, Ml 15, and A121. The multispecific antigen-binding protein may preferably bind to an epitope which comprises two or all three of these residues.

[0114] Additionally or alternatively, the multispecific antigen-binding protein may specifically bind to an epitope of human PVRIG comprising at least one amino acid residue selected from S71, L72, V90, H92, P93, R95, G96, R98, W100, A137, F139, P140, G142, and S143. This residue numbering may refer to a human PVRIG amino acid sequence as shown in SEQ ID NO: 231. Various of the above residues have been identified as being involved in binding interactions of PVRIG and PVRL2, as described in Hu et al, incorporated herein by reference. Thus, binding of the multispecific antigen-binding protein to PVRIG may block binding of PVRL2 to PVRIG, as described above. The multispecific antigen-binding protein may bind to an epitope comprising corresponding amino acid residues to those specified above in any human PVRIG amino acid sequence. The multispecific antigen-binding protein may specifically bind to an epitope which comprises two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen or all fourteen of these residues. Preferably, the multispecific antigen-binding protein binds to an epitope of human PVRIG comprising at least one amino acid residue selected from L72, V90, R95, G96, R98, W100, A137, F139, P140, and G142. The multispecific antigenbinding protein may preferably bind to an epitope which comprises two, three, four, five, six, seven, eight, nine or all ten of these residues.

[0115] Various techniques known in the art may be used to determine specific binding of the multispecific antigen-binding protein to particular amino acid residues, such as crossblocking assays, alanine scanning mutational analysis or hydrogen / deuterium exchange detected by mass spectrometry.

[0116] Sequences of multispecific antigen-binding domains

[0117] The inventors have identified a number of preferred antigen-binding domains for PD-L1 and PVRIG and characterised their sequences. They have also characterised preferred combinations of the preferred antigen-binding domains for use in a multispecific antigen-binding protein.

[0118] Accordingly, the invention provides a multispecific antigen-binding protein comprising a first antigen-binding domain (specific for PD-L1) comprising the three heavy chain CDRs contained within a HCVR selected from any of SEQ ID NOs 141-150 and the three light chain CDRs contained within a LCVR selected from any of 151-160. Typically, the heavy chain CDRs and light chain CDRs are selected from those contained within the HCVR / LCVR pairs represented by SEQ ID NOs: 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156, 147 / 157, 148 / 158, 149 / 159, and 150 / 160. The invention also provides a multispecific antigen-binding protein comprising a second antigen-binding domain (specific for PVRIG) comprising the three heavy chain CDRs contained within a HCVR selected from any of SEQ ID NOs 121-130 and the three light chain CDRs contained within a LCVR selected from any of 131-140. Typically, the heavy chain CDRs and light chain CDRs are selected from those contained within the HCVR / LCVR pairs represented by SEQ ID NOs: 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140. The invention further provides a multispecific antigenbinding protein comprising a first antigen-binding domain with heavy and light chain CDRs contained in an HCVR / LCVR as defined above, and a second antigen-binding domain with heavy and light chain CDRs contained in an HCVR / LCVR as defined above.

[0119] The CDRs contained within a HCVR / LCVR may be identified according to the Kabat definition, the Chothia definition or the IMGT definition (see, for example, Kabat, Elvin Abraham. Sequences of proteins of immunological interest. No. 91. US Department of Health and Human Services, Public Health Service, National Institutes of Health, 1991; Lefranc, Marie-Paule, et al. "IMGT unique numbering for immunoglobulin and T cell receptor variable domains and Ig superfamily V-like domains." Developmental & Comparative Immunology 27.1 (2003): 55-77).

[0120] The invention also relates to a multispecific antigen-binding protein comprising a first antigen-binding domain (specific for PD-L1) comprising an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from the following combinations of CDR sequences: SEQ ID NOs 61-66, SEQ ID NOs 67-72, SEQ ID NOs 73-78, SEQ ID NOs 79-84, SEQ ID NOs 85-90, SEQ ID NOs 91-96, SEQ ID NOs 97-102, SEQ ID NOs 103-108, SEQ ID NOs 109-114 and SEQ ID NOs 115-120. Particularly preferred HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combinations for first antigen-binding domains are SEQ ID NOs 79-84, SEQ ID NOs 97-102 and SEQ ID NOs 103-108. The invention further relates to a multispecific antigen-binding protein comprising a second antigen-binding domain (specific for PVRIG) comprising an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from the following combinations of CDR sequences: SEQ ID NOs 1-6, SEQ ID NOs 7-12, SEQ ID NOs 13-18, SEQ ID NOs 19-24, SEQ ID NOs 25-30, SEQ ID NOs 31-36, SEQ ID NOs 37-42, SEQ ID NOs 43-48, SEQ ID NOs 49-54 and SEQ ID NOs 55-60. A particularly preferred HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination for a second antigen-binding domain is SEQ ID NOs 13-18.

[0121] The invention further relates to a multispecific antigen-binding protein comprising a first antigen-binding domain with an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination as defined above, and a second antigen-binding domain with an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination as defined above. Thus, the first antigen-binding domain may comprise a HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from one of the following combinations of CDR sequences: SEQ ID NOs 61-66, SEQ ID NOs 67-72, SEQ ID NOs 73-78, SEQ ID NOs 79-84, SEQ ID NOs 85-90, SEQ ID NOs 91-96, SEQ ID NOs 97-102, SEQ ID NOs 103-108, SEQ ID NOs 109-114 and SEQ ID NOs 115-120, and the second antigen-binding domain may comprise a HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from the following combinations of CDR sequences: SEQ ID NOs 1-6, SEQ ID NOs 7-12, SEQ ID NOs 13-18, SEQ ID NOs 19-24, SEQ ID NOs 25-30, SEQ ID NOs 31-36, SEQ ID NOs 37-42, SEQ ID NOs 43-48, SEQ ID NOs 49-54 and SEQ ID NOs 55-60.

[0122] Preferred combinations of first and second antigen-binding domains include:

[0123] • a first antigen-binding domain comprising the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 79-84, and a second antigen binding domain comprising the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 13-18;

[0124] • a first antigen-binding domain comprising the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 97-102, and a second antigen binding domain comprising the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 13-18; and

[0125] • a first antigen-binding domain comprising the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 103-108, and a second antigen binding domain comprising the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 13-18.

[0126] The first and second antigen-binding domains may comprise a HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination as disclosed above and comprise an HCVR and LCVR having at least 95% sequence identity to the HCVR and LCVR from which the CDRs are derived. Thus for example, the first antigen binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 79-84 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 144 and an LCVR having at least 95% sequence identity to SEQ ID NO: 154. The first antigen binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 97-102 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 147 and an LCVR having at least 95% sequence identity to SEQ ID NO: 157. The first antigen binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 103-108 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 148 and an LCVR having at least 95% sequence identity to SEQ ID NO: 158. The second antigen binding domain (which may be provided in combination with any of the above three first antigen binding domains) may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 13-18 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 123 and an LCVR having at least 95% sequence identity to SEQ ID NO: 133.

[0127] The invention further relates to a multispecific antigen-binding protein comprising a first antigen-binding domain comprising an HCVR / LCVR pair selected from the following HCVR / LCVR pairs: SEQ ID NOs 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156, 147 / 157, 148 / 158, 149 / 159, and 150 / 160. Preferred HCVR / LCVR pairs for a first antigen-binding domain include SEQ ID NOs 144 / 154, 147 / 157 and 148 / 158. The invention further relates to a multispecific antigen-binding protein comprising a second antigen-binding domain comprising an HCVR / LCVR pair selected from the following HCVR / LCVR pairs: SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140. A preferred HCVR / LCVR pair for a second antigen-binding domain is SEQ ID NOs 123 / 133.

[0128] The invention also relates to a multispecific antigen-binding protein comprising a first antigen-binding domain comprising an HCVR / LCVR pair selected from the following HCVR / LCVR pairs: SEQ ID NOs 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156, 147 / 157, 148 / 158, 149 / 159, and 150 / 160, and a second antigen-binding domain comprising an HCVR / LCVR pair selected from the following HCVR / LCVR pairs: SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140.

[0129] Preferred combinations of first and second antigen-binding domains include:

[0130] • a first antigen-binding domain comprising the HCVR / LCVR pair of SEQ ID NOs 144 / 154 and a second antigen-binding domain comprising the HCVR / LCVR pair of SEQ ID NOs 123 / 133;

[0131] • a first antigen-binding domain comprising the HCVR / LCVR pair of SEQ ID NOs 147 / 157 and a second antigen-binding domain comprising the HCVR / LCVR pair of SEQ ID NOs 123 / 133; and

[0132] • a first antigen-binding domain comprising the HCVR / LCVR pair of SEQ ID NOs 148 / 158 and a second antigen-binding domain comprising the HCVR / LCVR pair of SEQ ID NOs 123 / 133.

[0133] The first and second antigen-binding domains of the above preferred combinations may preferably be present in a preferred multispecific antigen-binding format as described above. Thus, the first and second antigen-binding domains are preferably fused. The fusion may particularly preferably be between a heavy chain polypeptide of the first antigenbinding domain and a light chain polypeptide of the second antigen-binding domain. The light chain polypeptide of the second antigen-binding domain may preferably be C- terminally fused to the heavy chain polypeptide of the first antigen-binding domain. The first antigen-binding domain may preferably be an antibody and the second antigenbinding domain preferably a Fab. The N-terminus of the light chain variable region of the second antigen-binding domain, which may preferably be a Fab, may be fused to the C- terminus of the Fc region of the heavy chain polypeptide of the first-antigen -binding domain, which may preferably be an antibody (mAb). The above fusion may be by a linker. The antibody may preferably comprise an IgGl constant region. The first antigenbinding domain (binding PD-L1) is thus preferably in the mAb position and the second antigen-binding domain (binding PVRIG) preferably in the Fab position as described herein. However, multi-specific antigen binding proteins based on the preferred antigenbinding domain sequences described herein may alternatively comprise an antigen-binding domain specific for PVRIG in the mAb position and an antigen-binding domain specific for PD-L1 in the Fab position. The multispecific antigen-binding protein may comprise three different polypeptide chains as described above.

[0134] A multispecific antigen-binding protein comprising first and second antigenbinding domains of the above preferred combinations may thus comprise:

[0135] • a first polypeptide chain comprising the amino acid sequence of SEQ ID NO: 161, a second polypeptide chain comprising the amino acid sequence of SEQ ID NO: 162, and a third polypeptide chain comprising the amino acid sequence of SEQ ID NO: 163;

[0136] • a first polypeptide chain comprising the amino acid sequence of SEQ ID NO: 164, a second polypeptide chain comprising the amino acid sequence of SEQ ID NO: 165, and a third polypeptide chain comprising the amino acid sequence of SEQ ID NO: 166; or

[0137] • a first polypeptide chain comprising the amino acid sequence of SEQ ID NO: 167, a second polypeptide chain comprising the amino acid sequence of SEQ ID NO: 168, and a third polypeptide chain comprising the amino acid sequence of SEQ ID NO: 169.

[0138] A further multispecific antigen-binding protein described herein comprises a first antigenbinding domain (specific for PD-L1) comprising the HCVR / LCVR pair of SEQ ID NOs 172 / 173 and a second antigen-binding domain (specific for PVRIG) comprising the HCVR / LCVR pair of SEQ ID NOs 170 / 171. The above multispecific antigen-binding protein may comprise a first polypeptide chain comprising the amino acid sequence of SEQ ID NO: 174, a second polypeptide chain comprising the amino acid sequence of SEQ ID NO: 175, and a third polypeptide chain comprising the amino acid sequence of SEQ ID NO: 176. Any of the above described multispecific antigen-binding proteins comprising a first antigen-binding domain comprising a specified combination of CDRs or specified HCVR / LCVR and / or a second antigen-binding domain comprising a specified combination of CDRs or specified HCVR / LCVR may have one or more of the functional properties described above. Thus, a multispecific antigen-binding protein as described above may have a KD for PD-L1 of less than 5nM (preferably less than 2nM) and a KD of for PVRIG of less than 20nM. It has been surprisingly identified by the inventors that the binding affinities of the first and second antigen-binding domains in preferred multispecific antigen-binding domains may be comparable to the binding affinities of corresponding individual antigen-binding proteins representing the relevant antigenbinding domains. A multispecific antigen-binding protein as described above may bind to an epitope of human PD-L1 comprising at least one amino acid residue selected from D26, D122, Y123, K124, and R125 (preferably all five residues) and / or an epitope of human PVRIG comprising at least one amino acid residue selected from L72, V90, R95, G96, R98, W100, A137, F139, P140, and G142 (preferably all ten residues).

[0139] Additionally, or alternatively, a multispecific antigen-binding protein as described above may activate T cells and may activate NK cells. The multispecific antigen-binding protein may activate T cells and induce ADCC against target cells. The multispecific antigen-binding protein may activate NK cells and induce ADCC against target cells. The multispecific antigen-binding protein may crosslink or cluster cancer cells and T cells and / or NK cells, and induce ADCC against target cells. The multispecific antigen-binding protein may crosslink or cluster cancer cells and T cells and / or NK cells and activate T cells and / or NK cells. The multispecific antigen-binding protein may block binding of PD- L1 to PD1 and binding of PVRIG to PVRL2. The multispecific antigen-binding protein may block binding of PD-L1 to CD80 and binding of PVRIG to PVRL2. The multispecific antigen-binding protein may block binding of PD-L1 to CD80 and PD-1 and binding of PVRIG to PVRL2.

[0140] In more detail, a multispecific antigen-binding protein as described above may comprise a first antigen-binding domain specific for PD-L1 comprising an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from SEQ ID NOs 79-84, SEQ ID NOs 97-102 and SEQ ID NOs 103-108 and a second antigen-binding domain specific for PVRIG, and may activate NK cells. Alternatively or additionally, this multispecific antigen-binding protein may crosslink or cluster cancer cells and T cells and / or NK cells. Alternatively or additionally, this multispecific antigen-binding protein may have a KD for PVRIG of less than 20 nM, preferably less than 10 nM. Alternatively or additionally, this multispecific antigen-binding protein may bind an epitope of human PVRIG comprising at least one amino acid residue selected from L72, V90, R95, G96, R98, W100, A137, F139, P140, and G142 (preferably all ten residues). The first antigenbinding domain in the above examples may comprise an HCVR / LCVR pair selected from SEQ ID NOs 144 / 154, 147 / 157 and 148 / 158. The multispecific antigen-binding protein may block binding of PD-L1 to PD1 and binding of PVRIG to PVRL2. The multispecific antigen-binding protein may block binding of PD-L1 to CD80 and binding of PVRIG to PVRL2. The multispecific antigen-binding protein may block binding of PD-L1 to CD80 and PD-1 and binding of PVRIG to PVRL2.

[0141] Additionally, a multispecific antigen-binding protein as described above may comprise a first antigen-binding domain specific for PD-L1 and a second antigen-binding domain specific for PVRIG comprising the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 13-18, and may activate T cells. Alternatively or additionally, this multispecific antigen-binding protein may crosslink or cluster cancer cells and T cells and / or NK cells. Alternatively or additionally, this multispecific antigen-binding protein may have a KD for PD-L1 of less than 5 nM, preferably less than 2 nM. The multispecific antigen-binding protein may block binding of PD-L1 to PD1 and binding of PVRIG to PVRL2. The multispecific antigenbinding protein may block binding of PD-L1 to CD80 and binding of PVRIG to PVRL2. The multispecific antigen-binding protein may block binding of PD-L1 to CD80 and PD-1 and binding of PVRIG to PVRL2. Alternatively or additionally, this multispecific antigenbinding protein as described above may bind to an epitope of human PD-L1 comprising at least one amino acid residue selected from D26, D122, Y123, K124, and R125 (preferably all five residues) The second antigen-binding domain in either of the above examples may comprise the HCVR / LCVR pair of SEQ ID NOs 123 / 133.

[0142] Further described herein are multi-specific antigen-binding proteins which bind to the same epitopes on PD-L1 and PVRIG as any of the preferred multi-specific antigenbinding proteins described herein. Thus, described herein is a multi-specific antigen- binding protein which binds to the same epitopes on PD-L1 and PVRIG as a reference multi-specific antigen-binding protein comprising

[0143] • a first antigen-binding domain comprising the HCVR / LCVR pair of SEQ ID NOs 144 / 154 and a second antigen-binding domain comprising the HCVR / LCVR pair of SEQ ID NOs 123 / 133;

[0144] • a first antigen-binding domain comprising the HCVR / LCVR pair of SEQ ID NOs 147 / 157 and a second antigen-binding domain comprising the HCVR / LCVR pair of SEQ ID NOs 123 / 133; or a first antigen-binding domain comprising the HCVR / LCVR pair of SEQ ID NOs 148 / 158 and a second antigen-binding domain comprising the HCVR / LCVR pair of SEQ ID NOs 123 / 133.

[0145] Also described is a multi-specific antigen-binding protein which competes for binding or cross-competes for binding to PD-L1 and PVRIG with a reference multispecific antigen-binding protein as described above.

[0146] Binding to the same epitopes as a reference multi-specific antigen-binding protein can be readily determined by using routine methods known in the art and exemplified herein. Thus, the reference multi-specific antigen-binding protein may be allowed to bind PD-L1 and PVRIG. The ability of the test multi-specific antigen-binding protein to bind PD-L1 and PVRIG may then be assessed. If the test multi-specific antigen-binding protein is able to bind to PD-L1 and PVRIG following saturation binding with the reference multispecific antigen-binding protein, the test multi-specific antigen-binding protein may be determined not to bind the same epitopes as those bound by the reference multi-specific antigen-binding protein. In contrast, if the test multi-specific antigen-binding protein is not able to bind to PD-L1 and PVRIG following saturation binding with the reference multispecific antigen-binding protein, the test multi-specific antigen-binding protein may be determined to bind the same epitopes. Additional analyses may be carried out to determine that an observed lack of binding is not due to steric blocking or another phenomenon, and is instead due to binding of the same epitopes.

[0147] PD-L1 antigen-binding protein The invention also provides an antigen-binding protein specific for PD-L1. The antigen-binding protein may also be specific for additional antigens (typically additional antigens other than PVRIG), or may be monospecific.

[0148] The antigen-binding protein specific for PD-L1 may be selected from any class of antigen-binding protein that comprises an antigen-binding domain comprising immunoglobulin variable domains. The antigen-binding protein may comprise an antibody or antibody derivative, or an antigen-binding fragment of either thereof. The antigenbinding domain or antibody fragment may be selected from an scFv (typically arranged VH-VL in an N- to C-terminal orientation), Fab, a Fab’, or a F(ab’)2 fragment. The antigen-binding protein may comprise one or more constant regions as described above. A preferred constant region is IgG. The IgG may be IgGl.

[0149] The antigen-binding protein specific for PD-L1 comprises an antigen-binding domain specific for PD-L1 comprising the three heavy chain CDRs contained within a HCVR selected from any of SEQ ID NOs 141-150 and the three light chain CDRs contained within a LCVR selected from any of 151-160. Typically, the heavy chain CDRs and light chain CDRs are selected from those contained within the HCVR / LCVR pairs represented by SEQ ID NOs: 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156, 147 / 157, 148 / 158, 149 / 159, and 150 / 160. The CDRs may be identified as described above.

[0150] The antigen-binding protein specific for PD-L1 may comprise an antigen-binding domain specific for PD-L1 comprising an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from the following combinations of CDR sequences: SEQ ID NOs 61-66, SEQ ID NOs 67-72, SEQ ID NOs 73-78, SEQ ID NOs 79-84, SEQ ID NOs 85-90, SEQ ID NOs 91-96, SEQ ID NOs 97-102, SEQ ID NOs 103-108, SEQ ID NOs 109-114 and SEQ ID NOs 115-120. Particularly preferred HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combinations for antigen-binding domains are SEQ ID NOs 79-84, SEQ ID NOs 97-102 and SEQ ID NOs 103-108.

[0151] The antigen-binding protein specific for PD-L1 may comprise an antigen-binding domain comprising an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination as disclosed above and comprise an HCVR and LCVR having at least 95% sequence identity to the HCVR and LCVR from which the CDRs are derived. Thus, the antigen binding domain may preferably comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 79-84 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 144 and an LCVR having at least 95% sequence identity to SEQ ID NO: 154.

[0152] The antigen binding domain may preferably comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 97-102 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 147 and an LCVR having at least 95% sequence identity to SEQ ID NO: 157. The antigen binding domain may preferably comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 103-108 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 148 and an LCVR having at least 95% sequence identity to SEQ ID NO: 158.

[0153] The antigen-binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 61-66 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 141 and an LCVR having at least 95% sequence identity to SEQ ID NO: 151. The antigen-binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 67-72 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 142 and an LCVR having at least 95% sequence identity to SEQ ID NO: 152. The antigen-binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 73-78 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 143 and an LCVR having at least 95% sequence identity to SEQ ID NO: 153. The antigen-binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 85-90 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 145 and an LCVR having at least 95% sequence identity to SEQ ID NO: 155. The antigen-binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 91-96 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 146 and an LCVR having at least 95% sequence identity to SEQ ID NO: 156. The antigen-binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 103-108 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 148 and an LCVR having at least 95% sequence identity to SEQ ID NO: 158. The antigen-binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 109-114 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 149 and an LCVR having at least 95% sequence identity to SEQ ID NO: 159. The antigen-binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 115-120 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 150 and an LCVR having at least 95% sequence identity to SEQ ID NO: 160.

[0154] The antigen-binding domain may comprise an HCVR / LCVR pair selected from the following HCVR / LCVR pairs: SEQ ID NOs 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156, 147 / 157, 148 / 158, 149 / 159, and 150 / 160. Preferred HCVR / LCVR pairs include SEQ ID NOs 144 / 154, 147 / 157 and 148 / 158.

[0155] An antigen-binding protein specific for PD-L1 as described above typically binds to PD-L1 on a cancer or antigen-presenting cell. An antigen-binding protein specific for PD-L1 as described above may activate T cells. The T cell activation may be as described above. An antigen-binding protein specific for PD-L1 as described above may activate NK cells. The NK cell activation may be as described above. An antigen-binding protein specific for PD-L1 as described above may bind to PD-L1 on a cancer cell. An antigenbinding protein specific for PD-L1 as described above may induce ADCC or and / or ADCP against a target cell. The target cell is typically a cancer cell. The target cell may be a PD- L1 expressing cell. The ADCC may be as described above. The antigen-binding protein may block binding of PD-L1 to PD1 and / or CD80. An antigen-binding protein specific for PD-L1 as described above may have a KD for PD-L1 of less than 20nM, less than 10 nM, less than 5nM, preferably less than 2nM or more preferably less than 1 nM. An antigen-binding protein specific for PD-L1 as described above may have a Ka for PD-L1 of greater than lx 105M-l x sec -1. The Ka for PD-L1 may be in the range of lx 105- 5 x 106M-l x sec -1, such as 5 x 105- 3 x 106M-l x sec -1. An antigen-binding protein specific for PD-L1 as described above may have a Kd for PD-L1 of less than 1 x 10'21 / s, preferably less than 5 x 10'31 / s. The above binding affinities are typically for human PD- Ll. The above binding affinities may be determined as discussed above in relation to the multispecific antigen-binding proteins described herein.

[0156] PVRIG antigen-binding protein The invention also provides an antigen-binding protein specific for PVRIG. The antigen-binding protein may also be specific for additional antigens (typically additional antigens other than PDL-1), or may be monospecific.

[0157] The antigen-binding protein specific for PVRIG may be selected from any class of anti gen -binding protein that comprises an antigen-binding domain comprising immunoglobulin variable domains. The antigen-binding protein may comprise an antibody or antibody derivative, or an antigen-binding fragment of either thereof. The antigenbinding domain or antibody fragment may be selected from an scFv (typically arranged VH-VL in an N- to C-terminal orientation), Fab, a Fab’, or a F(ab’)2 fragment. The antigen-binding protein may comprise one or more constant regions as described above. A preferred constant region is IgG. The IgG may be IgGl.

[0158] The antigen-binding protein specific for PVRIG comprises an antigen-binding domain specific for PVRIG comprising the three heavy chain CDRs contained within a HCVR selected from any of SEQ ID NOs 121-130 and the three light chain CDRs contained within a LCVR selected from any of 131-140. Typically, the heavy chain CDRs and light chain CDRs are selected from those contained within the HCVR / LCVR pairs represented by SEQ ID NOs: 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140. The CDRs may be identified as described above.

[0159] The antigen-binding protein specific for PVRIG may comprise an antigen-binding domain specific for PVRIG comprising an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from the following combinations of CDR sequences: SEQ ID NOs 1-6, SEQ ID NOs 7-12, SEQ ID NOs 13-18, SEQ ID NOs 19-24, SEQ ID NOs 25-30, SEQ ID NOs 31-36, SEQ ID NOs 37-42, SEQ ID NOs 43-48, SEQ ID NOs 49-54 and SEQ ID NOs 55-60. A particularly preferred HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination is SEQ ID NOs 13-18.

[0160] The antigen-binding protein specific for PVRIG may comprise an antigen-binding domain comprising an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination as disclosed above and comprise an HCVR and LCVR having at least 95% sequence identity to the HCVR and LCVR from which the CDRs are derived. The antigen binding domain may preferably comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs SEQ ID NOs 13-18 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 123 and an LCVR having at least 95% sequence identity to SEQ ID NO: 133.

[0161] The antigen-binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs SEQ ID NOs 1-6 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 121 and an LCVR having at least 95% sequence identity to SEQ ID NO: 131. The antigenbinding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs SEQ ID NOs 7-12 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 122 and an LCVR having at least 95% sequence identity to SEQ ID NO: 132. The antigen-binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 19-24 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 124 and an LCVR having at least 95% sequence identity to SEQ ID NO: 134. The antigen-binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 25-30 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 125 and an LCVR having at least 95% sequence identity to SEQ ID NO: 135. The antigen-binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 31-36 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 126 and an LCVR having at least 95% sequence identity to SEQ ID NO: 136. The antigen-binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 37-42 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 127 and an LCVR having at least 95% sequence identity to SEQ ID NO: 137. The antigen-binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 43-48 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 128 and an LCVR having at least 95% sequence identity to SEQ ID NO: 138. The antigen-binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 49-54 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 129 and an LCVR having at least 95% sequence identity to SEQ ID NO: 139.

[0162] The antigen-binding domain may comprise the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 55-60 and comprise an HCVR having at least 95% sequence identity to SEQ ID NO: 130 and an LCVR having at least 95% sequence identity to SEQ ID NO: 140.

[0163] The antigen-binding domain may comprise an HCVR / LCVR pair selected from the following HCVR / LCVR pairs: SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140. A preferred HCVR / LCVR pair is SEQ ID NOs 123 / 133.

[0164] An antigen-binding protein specific for PVRIG as described above typically binds to PVRIG on an immune cell. The immune cell may be a T cell or an NK cell. An antigenbinding protein specific for PVRIG as described above may activate NK cells. The NK cell activation may be as described above. An antigen-binding protein specific for PVRIG may bind to PVRIG on a PVRIG-expressing cancer cell. The PVRIG-expressing cancer cell may be of a liquid cancer or tumour, for example myeloid leukemia. The antigen-binding protein specific for PVRIG may induce ADCC of the cancer cell. The antigen-binding protein may block binding of PVRIG to PVRL2.

[0165] An antigen-binding protein specific for PVRIG as described above may have a KD for PVRIG of less than lOOnM, less than 50nM, or less than 20 nM, preferably less than 10 nM, more preferably less than 5nM or less than 1 nM. An antigen-binding protein specific for PVRIG as described above may have a Ka for PVRIG of greater than lx 105M-l x sec -1 or greater than lx 106M-l x sec -1. An antigen-binding protein specific for PVRIG as described above may have a Kd for PVRIG of less than 1 x 10'21 / s, preferably of the order of 10'31 / s. The above binding affinities are typically for human PVRIG. The above binding affinities may be determined as discussed above in relation to the multispecific antigen-binding proteins described herein.

[0166] Combination

[0167] The invention further relates to a combination of an antigen-binding protein specific for PD-L1 and an antigen-binding protein specific for PVRIG as discussed above. The combination may be any means of providing the two antigen-binding proteins together. The antigen-binding proteins may both be present in a composition or may be administered separately, simultaneously or sequentially. The combination may be of an antigen-binding protein specific for PD-L1 comprising an antigen-binding domain comprising an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from the following combinations of CDR sequences: SEQ ID NOs 61-66, SEQ ID NOs 67-72, SEQ ID NOs 73-78, SEQ ID NOs 79-84, SEQ ID NOs 85-90, SEQ ID NOs 91-96, SEQ ID NOs 97-102, SEQ ID NOs 103-108, SEQ ID NOs 109-114 and SEQ ID NOs 115- 120, and an antigen-binding protein specific for PVRIG comprising an antigen binding domain comprising HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from the following combinations of CDR sequences: SEQ ID NOs 1-6, SEQ ID NOs 7-12, SEQ ID NOs 13-18, SEQ ID NOs 19-24, SEQ ID NOs 25-30, SEQ ID NOs 31- 36, SEQ ID NOs 37-42, SEQ ID NOs 43-48, SEQ ID NOs 49-54 and SEQ ID NOs 55-60. Preferred combinations include an antigen-binding protein specific for PD-L1 comprising an antigen binding domain comprising HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from SEQ ID NOs 79-84, SEQ ID NOs 97-102 and SEQ ID NOs 103-108 and an antigen-binding protein specific for PVRIG comprising an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 13-18. The antigen-binding protein specific for PD-L1 may comprise an antigen binding domain comprising an HCVR / LCVR pair selected from the following HCVR / LCVR pairs: SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140, and the antigen-binding protein specific for PVRIG may comprise an antigen binding domain comprising HCVR / LCVR pair selected from the following HCVR / LCVR pairs: : SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140 . Preferred HCVR / LCVR pairs for the antigen-binding domain of the antigen-binding protein specific for PD-L1 are SEQ ID NOs 144 / 154, 147 / 157 and 148 / 158. A preferred HCVR / LCVR pair for the antigen-binding domain of the antigen-binding protein specific for PVRIG is SEQ ID NOs 123 / 133.

[0168] Nucleic acids

[0169] Also provided are one or more isolated nucleic acids or polynucleotides encoding a multispecific antigen-binding protein or antigen-binding protein of the invention. In some cases, the encoding nucleic acid sequence may be provided by more than one nucleic acid sequence, optionally present on more than one nucleic acid molecule. Collectively, the nucleic acid sequences are able to encode the multispecific antigen-binding protein or antigen-binding protein of the invention.

[0170] Nucleic acids which encode a multispecific antigen-binding protein or antigenbinding protein of the invention can be obtained by methods well known to those skilled in the art. The nucleic acids may encode for one or more particular amino acid sequences described herein. The nucleic acids may comprise any degenerate nucleic acid sequences capable of encoding for one or more particular amino acid sequences described herein.

[0171] For example, the nucleic acids may encode heavy and / or light chain variable region sequences comprising any heavy and / or light chain CDR sequences described herein. The nucleic acids may encode any heavy and / or light chain variable region sequences described herein, including any heavy and / or light chain variable region sequence pair described herien. The nucleic acids may encode any heavy and / or light chain sequence described herein.

[0172] In aspects relating to anti gen -binding molecules specific for PD-L1, the nucleic acids may comprise a first nucleic acid encoding a heavy chain variable region sequence comprising the CDR sequences of an antigen-binding domain specific for PD-L1 described herein and a second nucleic acid encoding a light chain variable region sequence comprising the CDR sequences of of an antigen-binding domain specific for PD-L1 described herein. The first and second nucleic acids may encode a heavy chain variable region sequence and light chain variable region sequence of any antigen-binding domain specific for PD-L1 described herein, typically of a heavy chain variable region sequence / light chain variable region sequence pair described herein. For example, the first and second nucleic acids may encode heavy and light chain variable region sequences of an antigen-binding domain specific for PD-L1 comprising an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from the following combinations of CDR sequences: SEQ ID NOs 61-66, SEQ ID NOs 67-72, SEQ ID NOs 73-78, SEQ ID NOs 79-84, SEQ ID NOs 85-90, SEQ ID NOs 91-96, SEQ ID NOs 97-102, SEQ ID NOs 103-108, SEQ ID NOs 109-114 and SEQ ID NOs 115-120. Particularly preferred HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combinations for antigen-binding domains are SEQ ID NOs 79-84, SEQ ID NOs 97-102 and SEQ ID NOs 103-108. The antigen-binding domain may comprise an HCVR / LCVR pair selected from the following HCVR / LCVR pairs: SEQ ID NOs 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156, 147 / 157, 148 / 158, 149 / 159, and 150 / 160. Preferred HCVR / LCVR pairs include SEQ ID NOs 144 / 154, 147 / 157 and 148 / 158.

[0173] In aspects relating to antigen-binding molecules specific for PVRIG, the nucleic acids may comprise a first nucleic acid encoding a heavy chain variable region sequence comprising the CDR sequences of an antigen-binding domain specific for PVRIG described herein and a second nucleic acid encoding a light chain variable region sequence comprising the CDR sequences of of an antigen-binding domain specific for PVRIG described herein. The first and second nucleic acids may encode a heavy chain variable region sequence and light chain variable region sequence of any antigen-binding domain specific for PVRIG described herein, typically of a heavy chain variable region sequence / light chain variable region sequence pair described herein. For example, the first and second nucleic acids may encode heavy and light chain variable region sequences of an antigen-binding domain specific for PVRIG comprising an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from the following combinations of CDR sequences: SEQ ID NOs 1-6, SEQ ID NOs 7-12, SEQ ID NOs 13-18, SEQ ID NOs 19-24, SEQ ID NOs 25-30, SEQ ID NOs 31-36, SEQ ID NOs 37-42, SEQ ID NOs 43-48, SEQ ID NOs 49-54 and SEQ ID NOs 55-60. A particularly preferred HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination is SEQ ID NOs 13-18. The antigen-binding domain may comprise an HCVR / LCVR pair selected from the following HCVR / LCVR pairs: SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140. A preferred HCVR / LCVR pair is SEQ ID NOs 123 / 133.

[0174] Alternatively, the above described nucleic acid sequences encoding heavy and light chain variable region sequences may be present on a single nucleic acid molecule.

[0175] In aspects relating to multispecific antigen-binding molecules described herein, the nucleic acids may comprise one or more nucleic acids encoding heavy and / or light chain variable sequences of any antigen-binding domain specific for PD-L1 and antigen-binding domain specific for PVRIG described herein. The one or more nucleic acids may be as described above in relation to nucleic acids encoding antigen-binding domains specific for PD-L1 and PVRIG. The nucleic acids may comprise separate nucleic acids encoding each heavy and light chain variable region sequence of each antigen-binding domain of the multispecific antigen-binding molecule, or each polypeptide chain sequence of the multispecific antigen-binding molecule. Thus, three separate nucleic acids may encode first, second and third polypeptide chains of the multispecific antigen-binding molecule. The nucleic acids may comprise separate vectors encoding each heavy and light chain variable region sequence of each antigen-binding domain of the multispecific antigenbinding molecule, or each polypeptide chain sequence of the multispecific antigen-binding molecule. Thus, three separate vectors may encode the first, second and third polypeptide chains of the multispecific antigen-binding molecule.

[0176] In aspects relating to preferred multispecific antigen-binding molecules comprising first, second and third polypeptides, separate nucleic acids or vectors may encode each of the first, second and third polypeptides. Thus, the nucleic acids or vectors may for example encode:

[0177] • a first polypeptide chain comprising the amino acid sequence of SEQ ID NO: 161 (a preferred encoding nucleic acid sequence comprises the nucleic acid sequence of SEQ ID NO: 220), a second polypeptide chain comprising the amino acid sequence of SEQ ID NO: 162 (a preferred encoding nucleic acid sequence comprises the nucleic acid sequence of SEQ ID NO: 221), and a third polypeptide chain comprising the amino acid sequence of SEQ ID NO: 163 (a preferred encoding nucleic acid sequence comprises the nucleic acid sequence of SEQ ID NO: 222);

[0178] • a first polypeptide chain comprising the amino acid sequence of SEQ ID NO: 164 (a preferred encoding nucleic acid sequence comprises the nucleic acid sequence of SEQ ID NO: 223), a second polypeptide chain comprising the amino acid sequence of SEQ ID NO: 165 (a preferred encoding nucleic acid sequence comprises the nucleic acid sequence of SEQ ID NO: 224), and a third polypeptide chain comprising the amino acid sequence of SEQ ID NO: 166 (a preferred encoding nucleic acid sequence comprises the nucleic acid sequence of SEQ ID NO: 225); or

[0179] • a first polypeptide chain comprising the amino acid sequence of SEQ ID NO: 167 (a preferred encoding nucleic acid sequence comprises the nucleic acid sequence of SEQ ID NO: 226), a second polypeptide chain comprising the amino acid sequence of SEQ ID NO: 168 (a preferred encoding nucleic acid sequence comprises the nucleic acid sequence of SEQ ID NO: 227), and a third polypeptide chain comprising the amino acid sequence of SEQ ID NO: 169 (a preferred encoding nucleic acid sequence comprises the nucleic acid sequence of SEQ ID NO: 228).

[0180] In other aspects relating to preferred multispecific antigen-binding molecules, the first, second and third polypeptides may be encoded on one or two nucleic acids or vectors.

[0181] The nucleic acids described herein may comprise a DNA sequence, or an RNA, such as mRNA sequence. The nucleic acids may be comprised in one or more vectors. A vector may be a viral vector. Conventional viral based expression systems could include retroviral, alpha-retroviral, lentivirus, adenoviral, adeno-associated (AAV) and herpes simplex virus (HSV) vectors for gene transfer. Non-viral transduction vectors include transposon-based systems including PiggyBac and Sleeping Beauty systems. Methods for producing and purifying such vectors are known in the art.

[0182] The vectors may be cloning vectors or expression vectors. A suitable vector may be any vector which is capable of carrying a sufficient amount of genetic information, and allowing expression of a polypeptide of the invention. General methods by which the vectors may be constructed, transfection methods and culture methods are well known to those skilled in the art. In this respect, reference is made to “Current Protocols in Molecular Biology”, 1999, F. M. Ausubel (ed), Wiley Interscience, New York and the Maniatis Manual produced by Cold Spring Harbor Publishing.

[0183] A nucleic acid may be provided in the form of an expression cassette, which includes control sequences operably linked to the inserted sequence, thus allowing for expression of the encoded sequence in vivo. Hence, also provided is one or more expression cassettes encoding the one or more nucleic acids that encode an antigen-binding protein or multispecific antigen-binding protein described herein. These expression cassettes, in turn, are typically provided within vectors (e.g. plasmids or recombinant viral vectors). Hence, also provided is a vector encoding an antigen-binding protein or multispecific antigen-binding protein described herein. Further provided is a set of nucleic acids or vectors which collectively encode an antigen-binding protein or multispecific antigen-binding protein described herein.

[0184] The vector may be a human artificial chromosome. Human artificial chromosomes are described in e.g. Kazuki et al., Mol. Ther. 19(9): 1591-1601 (2011), and Kouprina et al., Expert Opinion on Drug Delivery 11(4): 517-535 (2014). The vector may be a non-viral delivery system, such as DNA plasmids, naked nucleic acid (e.g. naked RNA), and nucleic acid complexed with a delivery vehicle, such as a liposome or a nanoparticle.

[0185] The nucleic acids, expression cassettes or vectors described herein may be introduced into a host cell, e.g. by transfection. Hence, also provided is a host cell comprising the one or more nucleic acids, expression cassettes or vectors of the invention. The nucleic acids, expression cassettes or vectors described herein may be introduced transiently or permanently into the host cell, allowing expression of an antibody from the one or more nucleic acids, expression cassettes or vectors. Such host cells include transient, or preferably stable higher eukaryotic cell lines, such as mammalian cells or insect cells, lower eukaryotic cells, such as yeast, or prokaryotic cells, such as bacteria cells. Particular examples of cells include mammalian HEK293, such as HEK293F, HEK293T, HEK293S or HEK Expi293F, CHO, HeLa, NSO and COS cells, or any other cell line used herein. Preferred host cells are the immune effector cells described herein. Preferably, the nucleic acids, expression cassettes or vectors described herein are introduced transiently into the host cell.

[0186] Also provided is a method of expressing an antigen-binding protein or multispecific antigen-binding protein described herein, comprising expressing said protein in a host cell described herein.

[0187] Pharmaceutical compositions

[0188] Also provided is a composition comprising an antigen-binding protein or multispecific antigen-binding protein described herein, or one or more encoding nucleic acids described herein.

[0189] The composition is typically a pharmaceutical composition. Typically, such compositions are prepared as liquid suspensions. The protein or nucleic acid(s) may be mixed with an excipient which is pharmaceutically acceptable and compatible with the active ingredient. Suitable excipients are, for example, water, saline, dextrose, glycerol, of the like and combinations thereof. In addition, if desired, the pharmaceutical compositions may contain minor amounts of auxiliary substances such as wetting or emulsifying agents, and / or pH buffering agents. The composition may comprise one or more additional therapeutic agent, such as an anti-cancer or tumour agent such as a chemotherapeutic agent or immunotherapeutic agent. The composition may comprise one or more preservatives, such as an anti-fungal and / or anti-viral agent.

[0190] Also provided is a method of formulating a composition described herein, comprising mixing an antigen-binding protein or multispecific antigen-binding protein described herein, or one or more encoding nucleic acids described herein, with a pharmaceutically acceptable excipient.

[0191] A composition, antigen-binding protein, multispecific antigen-binding protein or encoding nucleic acid(s) described herein may be suitable for administration or administered by any route. Suitable routes include, but are not limited to, the intravenous, intrathecal, intracerebral ventricular, intramuscular, intraperitoneal, subcutaneous, intradermal, transdermal, nasal and oral / buccal routes.

[0192] The composition, antigen-binding protein, multispecific antigen-binding protein or encoding nucleic acid(s) is administered in a manner compatible with the dosage formulation and in such amount will be therapeutically effective. The quantity to be administered depends, for example, on the subject to be treated, the nature of the disorder, and so on. Precise amounts of the agent required to be administered may depend on the judgement of the practitioner and may be peculiar to each subject. The composition, antigen-binding protein, multispecific antigen-binding protein or encoding nucleic acid(s) may be administered as a single dose or in a multiple dose regimen. For example, the initial dose may be followed by administration of a second or plurality of subsequent doses. The second and subsequent doses may be separated by an appropriate time. For example, the doses may be administered once about every week, once about every 2 weeks, once about every 3 weeks, once about every four weeks, or once about every month.

[0193] Medical uses and methods of treatment

[0194] Also provided herein is the use of a composition, antigen-binding protein, multispecific antigen-binding protein, or encoding nucleic acid(s) described herein in a method of treatment of the human or animal body by therapy. Further provided is a method of treating a cancer or tumour in a subject, the method comprising administering to the subject an effective amount of a composition, antigenbinding protein, multispecific antigen-binding protein, or encoding nucleic acid(s) described herein. Also provided is a use of a composition, antigen-binding protein, multispecific antigen-binding protein, or encoding nucleic acid(s) described herein for the manufacture of a medicament for the treatment of a cancer or tumour.

[0195] The subject is typically a mammal. Preferably, the mammal is a human.

[0196] The subject may be of any age. For example, the subject may be a juvenile. The subject may, for example, be an adult.

[0197] The cancer or tumour may be a solid cancer or tumour. The solid cancer or tumour may be breast cancer, gastric cancer, non-small cell lung cancer, renal cell carcinoma or melanoma. The breast cancer may be triple negative breast cancer. The cancer may be a liquid cancer, such as a haematological cancer. The haematological cancer may be myeloid leukemia, such as acute myeloid leukemia. The cancer or tumour may comprise an overexpression of PD-L1 and / or PVRL2. The cancer or tumour may be refractory to treatment with an antagonist of PD-L1 (such as an antigen-binding protein specific for PD- Ll) or an antagonist of PVRIG (such as an antigen-binding protein specific for PVRIG).

[0198] The therapeutic methods and uses may comprise, prior to treatment with the composition, antigen-binding protein, multispecific antigen-binding protein, or encoding nucleic acid(s) described herein, determining whether the cancer or tumour comprises an overexpression of PD-L1.

[0199] The therapeutic methods and uses described herein may comprise inhibiting the disease state (i.e. the cancer or tumour), for example by arresting its development and / / or causing regression of the disease state until a desired end point is reached. The therapeutic methods and uses of the invention may comprise achieving a partial response, a full response by the cancer or tumour. The therapeutic methods and uses of the invention may achieve remission of the cancer or tumour.

[0200] The therapeutic methods and uses described herein may delay the growth of the cancer, arrest the growth of the cancer or tumour and / or reverse the growth of the cancer or tumour. The therapeutic methods and uses of the invention may reduce the size of the cancer or tumour by at least 10%, such as at least 20%, at least 30%, at least 40%, at least 50%, at least 60%, at least 70%, at least 80%, at least 90% or by 100%. The composition, antigen-binding protein, multispecific antigen-binding protein, or encoding nucleic acid(s) described herein may be administered with one or more additional therapy, such as one or more additional therapeutic agents. The additional therapeutic agent may be an anti-cancer or tumour agent such as a chemotherapeutic agent or immunotherapeutic agent. Combined administration of the composition, antigen-binding protein, multispecific antigen-binding protein, or encoding nucleic acid(s) with the additional therapeutic agent may be achieved in a number of different ways. All the components may be administered together in a single composition. Each component may be administered separately as part of a combined therapy.

[0201] For example, the composition, antigen-binding protein, multispecific antigenbinding protein, or encoding nucleic acid(s) of the invention may be administered before, after or concurrently with the additional therapeutic agent.

[0202] The additional therapy may be chemotherapy, immunotherapy, radiotherapy and / or surgery.

[0203] Further embodiments of the invention

[0204] The invention further provides:

[0205] 1. A multispecific antigen-binding protein comprising a first antigen-binding domain specific for PD-L1 and a second antigen-binding domain specific for PVRIG.

[0206] 2. The multispecific antigen-binding protein according to embodiment 1, which is a bispecific antibody comprising a first antigen-binding domain specific for PD-L1 and a second antigen-binding domain specific for PVRIG.

[0207] 3. The multispecific antigen-binding protein according to embodiment 1, which binds to PD-L1 on a cancer cell or an antigen presenting cell.

[0208] 4. The multispecific antigen-binding protein according to any one of embodiments 1 to 3, which binds to PVRIG on an immune cell.

[0209] 5. The multispecific antigen-binding protein according to embodiment 4, wherein the immune cell is a T cell or an NK cell. 6. The multispecific antigen-binding protein according to any one of the preceding embodiments, which activates T cells, optionally wherein T cell activation is PD-L1 dependent.

[0210] 7. The multispecific antigen-binding protein according to any one of the preceding embodiments, which activates NK cells, optionally wherein NK cell activation is PVRG-dependent.

[0211] 8. The multispecific antigen-binding protein according to any one of the preceding embodiments, which induces ADCC against cancer cells.

[0212] 9. _The multispecific antigen-binding protein according to any one of the preceding embodiments, which binds to PD-L1 on a cancer cell and PVRIG on an immune cell, such as a T cell or an NK cell and which induces clustering of said cancer cells and immune cells.

[0213] 10. The multispecific antigen binding protein according to any one of the preceding embodiments, which blocks binding of CD80 and / or PD-1 to PD-L1.

[0214] 11. The multispecific antigen binding protein according to any one of the preceding embodiments, which blocks binding of PVRL2 to PVRIG.

[0215] 12. The multispecific antigen binding protein according to any one of the preceding embodiments, which induces an anti-tumour response without inducing T cell exhaustion, optionally which does not increase expression of one or more genes associated with T cell exhaustion, such as TIGIT, LAG3 and / or CTLA4.

[0216] 13. The multispecific antigen-binding protein according to any one of the preceding embodiments, which binds to an epitope of human PD-L1 comprising at least one amino acid residue selected from V23, D26, E58, Ml 15, D122, Y123, K124, and R125 or to an epitope of human PD-L1 comprising at least one amino acid residue selected from Y56, E58, R113, Ml 15, SI 17, A121, and D122.

[0217] 14. The multispecific antigen-binding protein according to embodiment 13, which binds to an epitope of human PD-L1 comprising the amino acid residues D26, D122, Y123, K124, and R125 or to an epitope of human PD-L1 comprising the amino acid residues E58, Ml 15, and A121.

[0218] 15. The multispecific antigen-binding protein according to any one of the preceding embodiments, which binds to an epitope of human PVRIG comprising at least one amino acid residue selected from S71, L72, V90, H92, P93, R95, G96, R98, W100, A137, F139, P140, G142, and S143.

[0219] 16. The multispecific antigen-binding protein according to embodiment 15, which binds to an epitope of human PVRIG comprising the amino acid residues L72, V90, R95, G96, R98, W100, A137, F139, P140, and G142.

[0220] 17. The multispecific antigen-binding protein according to any one of the preceding embodiments, wherein the first and second antigen-binding domains are selected from antibodies or antigen-binding fragments thereof.

[0221] 18. The multispecific antigen-binding protein according to any one of the preceding embodiments, wherein the first and second antigen-binding domains are fused, optionally wherein a heavy chain polypeptide of the first antigen-binding domain is fused to a light chain polypeptide of the second antigen-binding domain, optionally wherein the fusion is via a linker.

[0222] 19. The multispecific antigen-binding protein according to any one of the preceding embodiments, wherein the first antigen-binding domain is an antibody comprising an Fc region.

[0223] 20. The multispecific antigen-binding protein according to any one of embodiments 17-

[0224] 19, wherein the second antigen-binding domain is a Fab fragment or an scFv fragment.

[0225] 21. The multispecific antigen-binding protein according to any one of embodiments 18-

[0226] 20, comprising a first polypeptide chain comprising a heavy chain polypeptide of the first antigen-binding domain fused to a light chain polypeptide of the second antigenbinding domain, optionally via linker, a second polypeptide chain comprising a light chain that pairs with the heavy chain polypeptide of the first antigen-binding domain, and a third polypeptide chain that comprises a heavy chain polypeptide that pairs with the light chain polypeptide of the second antigen-binding domain, optionally wherein the multispecific antigen-binding protein comprises two copies of each of the three different polypeptide chains.

[0227] 22. The multispecific antigen-binding protein according to any one of the preceding embodiments, comprising an IgGl Fc region.

[0228] 23. The multispecific antigen-binding protein according to any one of the preceding embodiments, wherein the first antigen-binding domain comprises the three heavy chain CDRs and three light chain CDRs contained within a HCVR / LCVR pair selected from SEQ ID NOs 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156, 147 / 157, 148 / 158, 149 / 159, and 150 / 160. The multispecific antigen-binding protein according to embodiment 23, wherein the first antigen-binding domain comprises an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from SEQ ID NOs 61-66, SEQ ID NOs 67-72, SEQ ID NOs 73-78, SEQ ID NOs 79-84, SEQ ID NOs 85-90, SEQ ID NOs 91-96, SEQ ID NOs 97-102, SEQ ID NOs 103-108, SEQ ID NOs 109-114 and SEQ ID NOs 115-120, optionally wherein the first antigen-binding domain comprises a HCVR / LCVR pair having at least 95% identity to a HCVR / LCVR pair selected from SEQ ID NOs 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156, 147 / 157, 148 / 158, 149 / 159, and 150 / 160. The multispecific antigen-binding protein according to embodiment 24, wherein the first antigen-binding domain comprises an HCVR / LCVR pair selected from SEQ ID NOs 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156, 147 / 157, 148 / 158, 149 / 159, and 150 / 160. The multispecific antigen-binding protein according to embodiment 24, wherein the first antigen-binding domain comprises an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from SEQ ID NOs 79-84, SEQ ID NOs 97-102 and SEQ ID NOs 103-108, optionally wherein the first antigen-binding domain comprises a HCVR / LCVR pair having at least 95% identity to a HCVR / LCVR pair selected from SEQ ID NOs 144 / 154, 147 / 157 and 148 / 158. The multispecific antigen-binding protein according to embodiment 26, wherein the first antigen-binding domain comprises an HCVR / LCVR pair selected from SEQ ID NOs 144 / 154, 147 / 157 and 148 / 158. The multispecific antigen-binding protein according to any one of the preceding embodiments, wherein the second antigen-binding domain comprises the three heavy chain CDRs and three light chain CDRs contained within a HCVR / LCVR pair selected from SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140. The multispecific antigen-binding protein according to embodiment 28, wherein the second antigen-binding domain comprises an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from SEQ ID NOs 1-6, SEQ ID NOs 7-12, SEQ ID NOs 13-18, SEQ ID NOs 19-24, SEQ ID NOs 25-30, SEQ ID NOs 31-36, SEQ ID NOs 37-42, SEQ ID NOs 43-48, SEQ ID NOs 49-54 and SEQ ID NOs 55-60, optionally wherein the second antigen-binding domain comprises a HCVR / LCVR pair having at least 95% identity to a HCVR / LCVR pair selected from SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140.

[0229] 30. The multispecific antigen-binding protein according to embodiment 29, wherein the second antigen-binding domain comprises an HCVR / LCVR pair selected from SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140.

[0230] 31. The multispecific antigen-binding protein according to embodiment 29 wherein the second antigen-binding domain comprises the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 13- 18, optionally wherein the second antigen-binding domain comprises a HCVR / LCVR pair having at least 95% identity to SEQ ID NOs 123 / 133.

[0231] 32. The multispecific antigen-binding protein according to embodiment 31, wherein the second antigen-binding domain comprises the HCVR / LCVR pair of SEQ ID NOs 123 / 133.

[0232] 33. An antigen-binding protein specific for PD-L1, comprising the three heavy chain CDRs and three light chain CDRs contained within a HCVR / LCVR pair selected from SEQ ID NOs 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156, 147 / 157, 148 / 158, 149 / 159, and 150 / 160.

[0233] 34. The antigen-binding protein specific for PD-L1 according to embodiment 33, comprising an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from the following combinations of CDR sequences: SEQ ID NOs 61-66, SEQ ID NOs 67-72, SEQ ID NOs 73-78, SEQ ID NOs 79-84, SEQ ID NOs 85-90, SEQ ID NOs 91-96, SEQ ID NOs 97-102, SEQ ID NOs 103-108, SEQ ID NOs 109- 114 and SEQ ID NOs 115-120, optionally wherein the antigen-binding protein comprises a HCVR / LCVR pair having at least 95% identity to a HCVR / LCVR pair selected from SEQ ID NOs 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156,

[0234] 147 / 157, 148 / 158, 149 / 159, and 150 / 160. 35. The antigen-binding protein specific for PD-L1 according to embodiment 33 or 34, comprising an HCVR / LCVR pair selected from SEQ ID NOs 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156, 147 / 157, 148 / 158, 149 / 159, and 150 / 160.

[0235] 36. An antigen-binding protein specific for PVRIG, comprising the three heavy chain CDRs and three light chain CDRs contained within a HCVR / LCVR pair selected from SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140.

[0236] 37. The antigen-binding protein specific for PVRIG according to embodiment 36, comprising an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from the following combinations of CDR sequences: SEQ ID NOs 1-6, SEQ ID NOs 7-12, SEQ ID NOs 13-18, SEQ ID NOs 19-24, SEQ ID NOs 25-30, SEQ ID NOs 31-36, SEQ ID NOs 37-42, SEQ ID NOs 43-48, SEQ ID NOs 49-54 and SEQ ID NOs 55-60, optionally wherein the antigen-binding protein comprises a HCVR / LCVR pair having at least 95% identity to a HCVR / LCVR pair selected from SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140.

[0237] 38. The antigen-binding protein specific for PVRIG according to embodiment 36 or 37, comprising an HCVR / LCVR pair selected from SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140.

[0238] 39. A combination of an antigen-binding protein specific for PD-L1 as defined in any one of embodiments 33-35 and an antigen-binding protein specific for PVRIG as defined in any one of embodiments 36-38.

[0239] 40. One or more nucleic acids encoding a multispecific antigen-binding protein as defined in any one of embodiments 1-32, an antigen-binding protein specific for PD-L1 as defined in any one of embodiments 33-35 or an antigen-binding protein specific for PVRIG as defined in any one of embodiments 36-38.

[0240] 41. A pharmaceutical composition comprising a multispecific antigen-binding protein as defined in any one of embodiments 1-32, an antigen-binding protein specific for PD- L1 as defined in any one of embodiments 33-35, an antigen-binding protein specific for PVRIG as defined in any one of embodiments 36-38, a combination of antigenbinding proteins as defined in embodiment 39, or one or more encoding nucleic acids as defined in embodiment 40, and a pharmaceutically acceptable excipient. 42. A multispecific antigen-binding protein as defined in any one of embodiments 1-32, an antigen-binding protein specific for PD-L1 as defined in any one of embodiments 33-35, an antigen-binding protein specific for PVRIG as defined in any one of embodiments 36-38, a combination of antigen-binding proteins as defined in embodiment 39, one or more encoding nucleic acids as defined in embodiment 40, or a pharmaceutical composition according to embodiment 41, for use in a method of treatment of the human or animal body by therapy.

[0241] 43. A multispecific antigen-binding protein as defined in any one of embodiments 1-32, an antigen-binding protein specific for PD-L1 as defined in any one of embodiments 33-35, an antigen-binding protein specific for PVRIG as defined in any one of embodiments 36-38, a combination of antigen-binding proteins as defined in embodiment 39, one or more encoding nucleic acids as defined in embodiment 40, or a pharmaceutical composition according to embodiment 41, for use in a method of treatment of a cancer or tumour.

[0242] 44. The multispecific antigen-binding protein, one or more encoding nucleic acids for the multispecific antigen-binding protein, or combination of antigen-binding proteins for use according to embodiment 43, wherein the cancer or tumour is a solid cancer or tumour or a haematological cancer.

[0243] 45. The multispecific antigen-binding protein, one or more encoding nucleic acids for the multispecific antigen-binding protein , or combination of antigen-binding proteins for use according to embodiment 43 or 44, wherein the cancer or tumour is refractory to treatment with an antagonist of PD-L1 or an antagonist of PVRIG.

[0244] Examples

[0245] EXAMPLE 1: Expression of PVRIG and PD-L1 in cancer tissue and in immune populations associated with cancer.

[0246] Tissue microarray (TMA) analysis of PVRIG andPD-Ll expression ofT cells, NK cells and regulatory T cells in tumor tissue and stromal samples in different cancers Multiplex immunohistochemistry of five different cancer types using 25 patient sample tissue microarray cores (CRO - OracleBio) was performed to evaluate PVRIG expression in immune cells within the tumor tissue compartments and PD-L1 expression in the tumor cells and tumor associated macrophages.

[0247] PVRIG expression was evaluated on the CD8+ T cells, CD56+ NK cells and Foxp3+ regulatory T (Treg) cells within the tumor tissues and stromal samples of the triple negative breast cancer, gastric cancer, non-small cell lung cancer (NSCLC), renal cell carcinoma (RCC) and melanoma (Figure 1). Table 1.1 summarizes PVRIG expression on the CD8+ T cells, CD56+ NK cells and Foxp3+ regulatory T (Treg) cells within the tumor tissues and stromal samples.

[0248] Figure 1 A presents tissue microarray stainings showing that the percentage (median) of PVRIG expressing CD8+ T cells of triple negative breast cancer was 77%, RCC 65%, gastric cancer 49%, NSCLC 42% and melanoma 31%. In the stromal compartment the percentage (median) of PVRIG expressing CD8+ T cells was in triple negative breast cancer 72%, RCC 57%, gastric cancer 51%, NSCLC 34% and melanoma 23%.

[0249] Figure IB shows CD56+ NK cells expressing PVRIG. The percentage (median) of PVRIG expressing CD56+ NK cells of triple negative breast cancer was 81%, RCC 70%, gastric cancer 65%, NSCLC 64% and melanoma 34%. In the stromal compartment the percentage (median) of PVRIG expressing CD56+ NK cells was in triple negative breast cancer 88%, RCC 74%, gastric cancer 72%, NSCLC 65% and melanoma 46%.

[0250] Figure 1C shows that the percentage of PVRIG expressing Foxp3+ regulatory T (Treg) cells. The percentage (median) of PVRIG expressing Foxp3+ regulatory T (Treg) cells of triple negative breast cancer was 46%, RCC 37%, gastric cancer 18%, NSCLC 54% and melanoma 1%. In the stromal compartment the percentage (median) of PVRIG expressing Foxp3+ regulatory T (Treg) cells was in triple negative breast cancer 32%, RCC 46%, gastric cancer 20%, NSCLC 54% and melanoma 2%.

[0251] Table 1 PVRIG expression on the CD8+ T cells, CD56+ NK cells andFoxp3+ regulatory T (T reg) cells within the tumor tissues and stromal samples of the triple negative breast cancer (breast), gastric cancer, non-small cell lung cancer (NSCLC), renal cell carcinoma (RCC) and melanoma.

[0252] PD-L1 expression was evaluated on the CD68+ tumor associated macrophages and Epcam+ tumor cells within the tumor tissues and stromal samples of the triple negative breast cancer, gastric cancer, non-small cell lung cancer, renal cell carcinoma (RCC) and melanoma (Figure 2). Table 2 summarizes PD-L1 expression on the CD68+ tumor associated macrophages and tumor cells within the tumor tissues and stromal samples.

[0253] Figure 2A presents tissue microarray stainings showing that the percentage (median) of PD-L1 expressing CD68+ tumor associated macrophages of triple negative breast cancer was 80%, RCC 9%, gastric cancer 71%, NSCLC 92% and melanoma 57%. In the stromal compartment the percentage (median) of PD-L1 expressing CD68+ tumor associated macrophages was in triple negative breast cancer 54%, RCC 2%, gastric cancer 36%, NSCLC 74% and melanoma 41%.

[0254] Figure 2B shows the percentage (median) of Epcam+ cancer cells expressing PD- Ll. The percentage (median) of PD-L1 expressing Epcam+ cancer cells of triple negative breast cancer was 83%, RCC 10%, gastric cancer 59%, and NSCLC 95%. In the stromal compartment the percentage (median) of PD-L1 expressing Epcam+ cancer cells was in triple negative breast cancer 56%, RCC 3%, gastric cancer 33%, and NSCLC 69%. Melanoma samples were not analysed.

[0255] Table 2 PD-L1 expression on the CD68+ tumor associated macrophages and Epcam+ cancer cells within the tumor tissues and stromal samples of the triple negative breast cancer (breast), gastric cancer, non-small cell lung cancer (NSCLC), renal cell carcinoma (RCC) and melanoma.

[0256] Flow cytometry analysis ofPVRIG, TIGIT, PD-1 and DNAM-1 expression in immune cell populations and PD-L1, PVR and PVRL2 expression in cancer cells

[0257] Immune cells of peripheral blood mononuclear cells (PBMCs) from healthy controls, PBMCs from lung cancer patients (patient PBMCs) and dissociated tumor cells (DTC) resected and homogenized from cancerous tissues of lung cancer patients (Discovery Life Sciences) were stained for flow cytometry analysis. First, immunophenotyping of the immune cell populations was carried out by staining the cells with antibodies specific for the surface markers. Table 3 summarizes the antibodies used for staining.

[0258] Figure 3 shows CD3+ T cells, CD4+ T helper cells, CD8+ cytotoxic T cells, CD3- CD56+ NK cells and CD3+CD56+ NKT cells present in each sample type. The percentage of CD3+ T cells were at similar levels in control healthy and patient PBMCs and also in lung DTCs (40-60% mean out of CD45+ immune cells), of which CD4+ T cells were also in similar levels throughout the healthy, patient PBMCs and lung DTCs, whereas CD8+ T cells were higher in the lung DTCs compared to healthy and patient PBMCs (10% mean out of CD3+ T cells). NK cells were varying between 1-20% within the tumor tissue, whereas in healthy and patient PBMCs this was 10% out of CD45+ immune cell populations in average.

[0259] Second, the expression ofPVRIG, TIGIT, PD-1 and DNAM-1 was studied in CD4+ T cells, CD8+ T cells, NK cells and NKT cells. Cells from healthy controls PBMCs, patient PBMCs and lung cancer DTCs were double-stained with antibodies immune cell markers and the selected proteins PVRIG, TIGIT, PD-1 and DNAM-1. Staining antibodies are summarized in table 3. Cells were then analysed by flow cytometry. Figure 4 shows the flow cytometry results calculated based on the percentages of the cells that are expressing the stained proteins stained within the immune subgroups CD4+ and CD+ T cells, NK cells and NKT cells.

[0260] Out of healthy and patient PBMC CD8+ T cells between 10% and 30% were PVRIG positive, while this was 2-15% in lung DTCs. Out of healthy and patient PBMC CD4+ T cells between 2% and 8% were PVRIG positive, while this was 1-3% in lung DTCs. Out of healthy PBMC CD3-CD56+ NK cells between 15% and 50% were PVRIG positive with the mean at 32%, while PVRIG expression in lung DTC was 1-10% with mean at 5%. Out of healthy PBMCs CD3+CD56+ NKT cells between 18% and 30% were PVRIG positive with the mean 25%, out of patient PBMCs NKT cells PVRIG expression was between 15% and 30%, while this was 1-3% in lung DTCs.

[0261] Out of healthy PBMC CD8+ T cells with a mean value 19% were TIGIT positive and out of patient PBMCs the CD8+ T cells between 18% and 50% with a mean value at 20% TIGIT positive, while TIGIT expressing cells were 25-70% with a mean value at 40% in lung DTCs. Out of healthy PBMC CD4+ T cells 5% on average were TIGIT positive, while out of patient PBMC CD4+ T cells this was 10%, with values between 5-22%. In lung DTCs 30% to 70% of the CD4+ T cells with a mean value at 40% were expressing TIGIT. Out of healthy PBMC CD3-CD56+ NK cells between 17% and 28% were TIGIT positive with the mean at 20%, while TIGIT expression in lung DTC was more varying 5- 35% with mean at 18%. The TIGIT expressing NK cells in lung DTCs were also varying as low as 0% and as high as 22% with a mean value of 7%. Out of healthy PBMCs CD3+CD56+ NKT cells between 2% and 30% were TIGIT positive with the mean 18%, out of patient PBMCs NKT cells TIGIT expression was between 10% and 40% with the mean 22%, while TIGIT expression among NKT cells in lung DTCs this was 5-60% with a mean value at 40%.

[0262] Out of healthy PBMC CD8+ T cells on average 1% were PD-1 positive and out of patient PBMCs the CD8+ T cells less than 10% were PD-1 positive, while out of CD8+ T cells in lung DTCs 20%-50% with a mean value of 35% were expressing PD-1. Out of healthy and patient PBMC CD4+ T cells less than 5% were PD-1 positive, while PD-1 expressing CD4+ T cells in lung DTCs comprised of 18% to 58% of the CD4+ T cells with a mean value at 40%. Out of healthy and patient PBMCs including within lung DTCs CD3-CD56+ NK cells less than 2% were expressing PD-1. Out of healthy PBMCs CD3+CD56+ NKT cells were less than 2% PD-1 positive, while PD-1 expression among NKT cells in lung DTCs was 5-35% with a mean value at 20%.

[0263] Out of healthy and patient PBMC CD8+ T cells between 39% and 37%, respectively, in average were DNAM-1 positive, while this was 15% in lung DTCs. Out of healthy PBMC CD4+ T cells 5% in average were DNAM-1 positive, while out of patient PBMCs CD4+ T cells DNAM-1 expressing cell range was 0-17% with an average at 7%. In lung DTCs out of CD4+ T cells 5% in average were expressing DNAM-1. Out of healthy PBMC CD3-CD56+ NK cells in average 40% were DNAM-1 positive, while DNAM-1 expression in patient PBMCs NK cells was between 18% and 65% with a mean value 32%. Out of healthy PBMCs CD3+CD56+ NKT cells in average 50% were DNAM- 1 positive, out of patient PBMCs NKT cell DNAM-1 expression was between 20% and 58% with a mean value at 37%, while the DNAM-1 expression on NKT cells in lung DTCs was in average 10%.

[0264] Then the expression of PD-L1, PVR and PVRL2 was studied on a CD1 Ib+CDl lc+ cell population including monocytes, dendritic cells and macrophages. Cell populations isolated from healthy controls PBMCs, patient PBMCs and lung cancer DTCs were stained with antibodies binding to PD-L1, PVR and PVRL2. Antibodies used are summarized in table 4. Figure 5 shows flow cytometry results that out of healthy PBMC CD1 Ib+CDl lc+ cells between 18% and 52% with mean value 35% were PD-L1 positive, 6% on average was PVR positive and 75% on average was PVRL2 positive. Out of patient PBMCs the CD1 Ib+CDl lc+ cells between 10% and 40% with mean value 30% were PD-L1 positive, 3% to 12% with an average value of 7% were PVR positive and 70% on average were PVRL2 expressing. Out of CD1 Ib+CDl lc+ cells in lung DTCs between 10% and 40% with mean value 30% were PD-L1 expressing, 6% in average were PVR positive and 60% to 90% with an average of 75% were PVRL2 positive.

[0265] PD-L1, PVR and PVRL2 expression was also studied in Epcam+ lung cancer cells that were DTCs. Figure 6 shows the flow cytometry results where the PD-L1 expression on the CD45-EpCAM+ tumor cells have donor dependent varying between 5% to 80% where the median falls to less than 10%, whereas PVR expression also varying between no expression to 20% of the tumor cells expressing with a median at 10%. PVRL2 expression was found to be more consistently expressed on the EpCAM+ tumor cells ranging from 20% to 45% with a median value of 35%. Table 3 Antibodies used in flow cytometry analysis.

[0266] Table 4 Antibodies used in flow cytometry analysis.

[0267] PVRIG expression on NKT cells and gamma delta T cells of peripheral blood mononuclear cells (PBMCs) Frozen PBMCs were thawed and stained immediately for flow cytometry analysis, to show PVRIG expression on the CD3+CD56+ NKT cells and gamma delta TCR expressing T cells (gamma delta T cells). The staining was done using the antibodies anti- CD3 APC, anti-CD56 BV650, anti-TCR gamma delta BV421, anti-PD-Ll PerCP-Cy5.5, and anti-PVRIG-PE. The antibodies were conjugated to fluorescent dyes for staining (all antibodies from Biolegend). In addition, viability dye BV570 was used to measure the viability of the cells.

[0268] Figure 7 shows the PVRIG expression on the CD3+ T cells, NK cells, NKT cells and the gamma delta T cells. CD3+ T cells represent 50 %, NK cells 10 %, NKT cells 5 % and gamma delta T cells 1% (mean percentage values) from the alive PBMCs. All the cell populations have high percentage of PVRIG expression (the mean percentage > 75 %). At least 90 % of the NKT cells were PVRIG+ with highest geometric mean fluorescent intensity (MFI). At least 80 % of the gamma delta T cells were expressing PVRIG. Figure 8 shows the PD-L1 expression on the CD3+ T cells, NK cells, NKT and the gamma delta T cells. The percentage of PD-L1 expressing cells in these cell populations were rather low, around 20 % of the gamma delta T cells and around 15 % of the CD3+ T cells, NK cells and NKT cell (< 20 %).

[0269] EXAMPLE 2 Screening of Fab and scFv fragments binding to PVRIG

[0270] Table 5 summarizes the steps performed to screen PVRIG binding Fab fragments and scFv fragments. The final ranking of the remaining clones after step 5 and confirmatory affinity screening was based on affinity (KD), off-rate (kd), blocking ELISA, cynomolgus cross-reactivity, and in silico predictions of immunogenicity and sequence liabilities. A final eight Fab and two scFv fragments were selected that bind to the PVRIG target. Screening for PD-L1 binders is described further below. Selected fragments in Fab format for both PVRIG and PD-L1 were then used to produce multi-specific PVRIG and PD-L1 antigen-binding proteins in a RUBY™ bispecific antibody format with each binder configured in two different possible orientations, resulting in lO x 10 x 2 = 200 different proteins. The RUBY™ format is described in Nyesiga et al.

[0271] For production of the multi-specific antigen-binding proteins, Expi293 cells or Freestyle-HEK cells were transfected according to the manufacturer’s instructions with three different vectors encoding each of the three polypeptide chains. Cell culture supernatants were protein A purified on an NGC10 chromatography system (Bio-Rad) using Mab Select SuRe columns (GE Healthcare). If necessary, further purification was performed by SEC. Some batches were filtrated using Amicon Ultra-2mL, 100K filters (Sigma Aldrich) by washing and filtering the sample 5 times. Buffer exchange to 20 mM histidine, 150 mM arginine (pH 6.5) was performed with HiTrap desalting columns (GE Healthcare). Concentration was measured using BigLunatic UV / Vis spectrophotometer (UnchainedLabs). If necessary, antibodies were concentrated using Vivaspin 6 centrifugal concentrators (GE Healthcare) according to manufacturer’s instructions.

[0272] Table 5 Screening cascade of PVRIG and PD-L1 binding Fab fragments and scFv fragments.

[0273] PVRIG Phage display selections

[0274] AlligatorGold (scFv) and AlligatorFABl-4 (Fab) libraries were used in phage display selections of PVRIG binding fragments. PVRIG antigens used in the phage display selection are summarized in table 6.

[0275] Selection strategy included decreasing concentrations of antigen from 100 nM in round 1 to 0.5 nM in round 5. Negative selections were performed overnight at 2-8 °C prior to all selection rounds to remove nonspecific binding fragments. Antigens used were CTLA-4-Fc (Orencia) bound to streptavidin beads (Invitrogen). Either CD28-Avi

[0276] (AcroBiosystems) and ubiquitin-His (R&D Systems) or CD47-Avi-His (AcroBiosystems) were also used as counter selection antigens. In addition to decreasing concentration of antigen, wash conditions were made more stringent over time. Round 1 included a thermal challenge to aid the isolation of the fragments with good stability.

[0277] Five independent selection tracks were performed including one that alternated with cynomolgus monkey PVRIG (cPVRIG) in order to discover cross-reactive antibodies and another that used cPVRIG for all rounds. One track also included two rounds of selections using cells overexpressing PVRIG and used HEK293 cells for counterselections during these rounds.

[0278] Table 6 PVRIG and PD-L1 antigens used in the phage display selection.

[0279] PVRIG Fab screening

[0280] After phage display selections, 1080 individual Fab clones were picked for clonal analysis as soluble fragments using a high-throughput surface plasmon resonance (SPR) system (LSA, Carterra). Fab fragments were captured using biotinylated anti-Fab Kappa antibody (ThermoFisher) coupled to streptavidin immobilized on an HC200 chip (Carterra). 874 Fab clones reached the threshold KD value of <10 nM and kd <7xl0'4s_1. A total of 646 Fab clones were picked for Sanger sequencing (Eurofins Genomics GmbH) to reveal 104 unique clones.

[0281] PVRIG scFv screening

[0282] 1080 scFv clones were picked for clonal analysis using high-throughput SPR via fragment capture on HC200 chips (Carterra) immobilized with anti-FLAG antibody (Sigma Aldrich) to bind to the FLAG-tag on scFv fragments. 490 scFv clones reached the threshold KD value of <10 nM and kd <3xl0'3s_1. A total of 480 Fab clones were picked for Sanger sequencing (Eurofins Genomics GmbH) to reveal 63 unique clones.

[0283] Secondary screening

[0284] All unique Fab and scFv clones were subjected to a secondary screening that included a specificity ELISA, cell binding using flow cytometry, and ligand block assay.

[0285] Specificity ELISA

[0286] The specificity ELISA evaluated nonspecific binding to tags present on recombinant antigens used in selections and screening, such as His-tag, Avi-tag, Fc-tag, and streptavidin. Streptavidin coated plates were used to couple biotinylated PVRIG-Fc at 0.5 ug / ml. Non-target proteins included CD47-His-Avi, streptavidin, and CTLA4-Fc. Cutoff values were set to target signal >300 000, ratio target / non-target >50 and signal non-target <30 000 for Fabs and target signal >100 000, ratio target / non-target >50 and signal non-target <10 000 for scFvs. 75 Fab clones and 59 scFv clones passed the specificity ELISA.

[0287] Cell Binding

[0288] All unique Fab and scFv clones were stained and analysed by flow cytometry for binding to HEK cells transiently transfected with hPVRIG. Cutoffs were based on positive and negative control antibody median signal to target cell, percent positive cells, and median signal to wild-type (wt) HEK cells (wt HEK). Fab fragment median HEK hPVRIG >14.5, percent positive HEK hPVRIG >4.5, and median wt HEK <14.5 were determined as cut-off values. scFv fragment median HEK hPVRIG >20, percent positive HEK hPVRIG >2, and median wt HEK <10 were determined as cut-off values. 62 Fab fragments passed cell binding assay cutoffs while 38 scFv fragments passed.

[0289] Ligand blocking ELISA All clones were subjected to a single point ligand block assay performed twice. Fab fragments were ranked as either good (<50% max signal), average (50 - 70% max signal), or bad (>70% max signal) while scFv fragments were deemed good (<15%), average (15- 40%), or bad (>40%). 59 Fab fragments and 17 scFv fragments were deemed good, 38 Fab fragments and 34 scFv fragments were deemed average, and 8 Fab fragments and 12 scFv fragments were ranked as bad. Only clones that had good or average rankings were moved forward.

[0290] In total after the specificity ELISA, cell binding, and ligand blocking assay, 58 Fab and 32 scFv fragments were analysed further.

[0291] Secondary affinity measurements

[0292] All Fab fragments and scFv fragments that passed the secondary screening were subjected to more detailed affinity measurements by SPR using the LSA assays against monomeric human (hPVRIG), cynomolgus (cPVRIG) and mouse (mPVRIG). 1 of 58 Fab fragments did not show binding to hPVRIG. 12 of the 58 Fab fragments and 7 of 32 scFv fragments showed binding to cPVRIG. No Fab fragments or scFv fragments showed binding to mPD-Ll, likely due to low homology. Affinity measurements for scFv fragments were repeated using hPVRIG-Fc due to low signal from monomeric hPVRIG.

[0293] A list of top 10 KD and top 10 kd (off-rate) against hPVRIG was created from these measurements. 8 clones had off rates < 1x1 O'3and were included in the set of top kd clones.

[0294] In silico sequence analysis

[0295] 1) PTM analysis

[0296] All clones were analysed for potential for post translational modifications (PTM). 10 Fab clones and 25 scFv clones showed no potential for PTMs and were moved forward. 86 Fab fragments and 33 scFv fragments were found to show minor potential for PTMs and therefore moved forward with further monitoring. 2) In silico immunogenicity analysis

[0297] Only top clones that had passed secondary affinity measurements were subjected to in silico immunogenicity analysis. Immunogenicity predictions were performed using the AbEpiAnalyser tool (EIR Sciences). VH and VL sequences, including mutations inherent to the multispecific antigen-binding protein format were submitted to the tool, and position-specific risk scores were calculated for the top-ranking 15-mer peptides, based on their predicted binding to a large panel (the world average population) of human MHC class II alleles. For each antibody, the sum of the position-specific risk scores is given as the immunogenicity score. The scores for all 15-mers that are identical with the closest germline (self-peptides) were subtracted from the scores for the VH and VL sequences. 20 of the 22 Fab fragments and 6 of the 19 scFv fragments analysed were found to be free from immunogenicity risk.

[0298] Final selection of 10 PVRIG clones

[0299] Based on all of the data above in relation to predicted post-translational modifications, immunogenicity, ligand blocking, KD and kd values, 8 Fab fragments and 2 scFv fragments were finally selected as preferred PVRIG binders. These are summarized with their characteristics in Tables 7 and 8.

[0300] Table 7 Characterization of the Fab and scFv fragments binding to PVRIG that were selected for further studies. Affinities of the PVRIG binding fragments were single- or double-digit nM, but cross-reactivity with cynomolgus PVRIG was only seen for half of the binders. Two of them appeared to be within a ten-fold range from the human affinity. PVRIG binders 1 and 2 are scFv fragments and PVRIG binders 3-10 are Fab fragments. Abbreviations: ka, association rate constant; kd, dissociation rate constant; KD, equilibrium constant (affinity); hu, human; cy, cynomolgus.

[0301] Table 8 Characterization of the Fab and scFv fragments binding to PVRIG that were selected for further studies. Percentage of ligand blocking was determined by ligand blocking assay and immunogenicity scores were calculated. PVRIG binders 1 and 2 are scFv fragments and PVRIG binders 3-10 are Fab fragments. EXAMPLE 3 Screening of Fab and scFv fragment binding to PD-L1

[0302] Table 5 summarizes the steps performed to screen PD-L1 binding Fab fragments and scFv fragments. The final ranking of the remaining clones after step 5 and confirmatory affinity screening was based on affinity (KD), off-rate (kd), blocking ELISA, cynomolgus cross-reactivity, and in silico predictions of immunogenicity and sequence liabilities. A final eight Fab and two scFv fragments were selected that bind to the PD-L1 target. As discussed above, selected fragments in Fab format for both PD-L1 and PVRIG were then used to produce multi-specific PVRIG and PD-L1 antigen-binding proteins in a RUBY™ bispecific antibody format with each binder configured in two different possible orientations, resulting in lO x 10 x 2 = 200 different proteins. The RUBY™ format is described in Nyesiga et al.

[0303] PD-L1 Phage display selections

[0304] AlligatorGold (scFv) and AlligatorFABl-4 (Fab) libraries were used in phage display selections of PD-L1 binding fragments. PD-L1 antigens used in the phage display selection are summarized in table 6.

[0305] Selection strategy included decreasing concentrations of antigen from 100 nM in round 1 to 0.5 nM in round 5. Negative selections were performed overnight at 2-8 °C prior to all selection rounds to remove nonspecific binding fragments. Antigens used were CTLA-4-Fc (Orencia) bound to streptavidin beads (Invitrogen). Either CD28-Avi (AcroBiosystems) and ubiquitin-His (R&D Systems) or CD47-Avi-His (AcroBiosystems) were also used as counter selection antigens. In addition to decreasing concentration of antigen, wash conditions were made more stringent over time. Round 1 included a thermal challenge to aid the isolation of binders with good stability.

[0306] Five independent selection tracks were performed including one that alternated with cynomolgus monkey PD-L1 in order to discover cross-reactive antibodies. One track also included two rounds of selections using cells overexpressing PD-L1 and used CHO cells for counterselections during these rounds.

[0307] PD-L1 Fab screening

[0308] After phage display selections, 2160 individual Fab clones were picked for clonal analysis as soluble fragments using a high-throughput surface plasmon resonance (SPR) system (LSA, Carterra). Fab fragments were captured using biotinylated anti-Fab Kappa antibody (ThermoFisher) coupled to streptavidin immobilized on an HC200 chip. 1185 Fab clones reached the threshold KD value of <100 nM and kd <lxl0'2s_1. A total of 1274 Fab clones were picked for Sanger sequencing (Eurofins Genomics GmbH) to reveal 65 unique clones. PD-L1 scFv screening

[0309] 2160 scFv clones were picked for clonal analysis using high-throughput SPR via fragment capture on HC200 chips immobilized with anti-FLAG antibody (Sigma Aldrich) to bind to the FLAG-tag on scFv fragments. ELISA assays were used to identify positive clones. A threshold signal to human PD-Ll-Fc > 200 000 and CTLA-4 < 8000 was used as a cutoff. 189 clones were Sanger sequenced to reveal 55 unique scFv clones.

[0310] Secondary screening

[0311] All unique Fab and scFv clones were subjected to secondary screening that included a specificity ELISA, cell binding using flow cytometry, and ligand block assay.

[0312] Specificity ELISA

[0313] The specificity ELISA evaluated nonspecific binding to tags present on recombinant antigens used in selections and screening, such as His-tag, Avi-tag, Fc-tag, and streptavidin. Streptavidin coated plates were used to couple biotinylated PD-Ll-His at 0.5 ug / ml. Non-target proteins included CD47-His-Avi, streptavidin, and CTLA4-Fc. Cutoff values were set to target signal > 100 000, ratio target / non-target > 50 and signal non-target < 10 000. 52 Fab clones and 53 scFv clones passed the specificity ELISA.

[0314] Cell Binding

[0315] All unique Fab and scFv clones were stained and analysed by flow cytometry for binding to CHO cells expressing human PD-L1. Cutoffs were based on positive and negative control antibody median signal to target cell, percent positive cells, and median signal to non-expressing CHO cells (wt CHO). Fab fragment median CHO PD-L1 > 70, percent positive CHO hPD-Ll > 50, and median wt CHO < 30 were determined as cut-off values. scFv fragment median CHO PD-L1 > 14, percent positive CHO hPD-Ll > 5, and median wt CHO < 14 were determined as cut-off values. Cell binding assay cutoffs were passed by 59 Fab fragments and 28 scFv fragments.

[0316] Ligand blocking ELISA

[0317] All clones were subjected to a single point ligand block assay performed twice. Fragments were ranked as either good (< 50% max signal), average (50 - 70% max signal), or bad (> 70% max signal). 10 Fab and 35 scFv fragments were deemed good, 33 Fab and 10 scFv fragments were deemed average, and 22 Fab and 10 scFv fragments were ranked as bad. Only clones that had good or average rankings were moved forward.

[0318] After the specificity ELISA, cell binding, and ligand blocking assay, a total of 40 Fab fragments and 26 scFv fragments were selected for further analysis.

[0319] Secondary affinity measurements

[0320] All Fab fragments and scFv fragments that had passed secondary screening were subjected to more detailed affinity measurements by biolayer interferometry (BLI, using an Octet) against monomeric human (hPD-Ll), cynomolgus (cPD-Ll) and mouse PD-L1 (mPD-Ll). 5 of 40 Fabs did not show binding to hPD-Ll and were removed from the Fab set. All clones showed binding to cPD-Ll while none showed binding to mPD-Ll, likely due to low homology.

[0321] A list of top 10 KD and top 10 kd (off-rate) against hPD-Ll was created from these measurements. 8 clones had off rates < IxlO'3and were included in the set of top kd clones.

[0322] In silico sequence analysis

[0323] 1) PTM analysis

[0324] All clones were analysed for potential for post translational modifications (PTM). 9 Fab clones and 26 scFv clones showed no potential for PTMs and were moved forward. 50 Fabs and 28 scFvs were found to show minor potential for PTMs and therefore moved forward with further monitoring.

[0325] 2) In silico immunogenicity analysis

[0326] Only top clones that had passed secondary affinity measurements were subjected to in silico immunogenicity analysis. Immunogenicity analysis and immunogenicity score calculation was done similarly as described for PVRIG binding Fabs and scFvs. 10 of the 12 Fab fragments and 7 of the 8 scFv fragments analysed were found to be free from immunogenicity risk.

[0327] Final selection of 10 PD-L1 clones

[0328] Based on performance in specificity ELISA assays, cell binding assays, ligand blocking assays a total of 40 Fab fragments and 26 scFv fragments were selected for further analyses. Of these, a total of 12 Fab fragments were selected based on KD and kd as well as ligand blocking assays. Of the 26 scFv fragments, 8 clones were selected based on a measured kd < 10'3. Two of these scFv clones were chosen based on affinity to cPD- Ll-Fc rather than hPD-Ll. Based on all of the data above in relation to predicted post- translational modifications, immunogenicity, ligand blocking, KD and kd values, 8 Fab fragments and 2 scFv fragments were finally selected as preferred PD-L1 binders that are summarized with their characteristics in tables 9 to 11.

[0329] Table 9 Characterization of the Fab fragments binding to PD-L1 that were selected for further studies. Most PD-L1 binding fragments had single -digit nanomolar affinities for human PD-L1, and binding to cynomolgus PD-L1 was at comparable levels.

[0330] Abbreviations: ka, association rate constant; kd, dissociation rate constant; KD, equilibrium constant (affinity); hu, human; cy, cynomolgus.

[0331] Table 10 Characterization of the scFv fragments binding to PD-L1 that were selected for further studies. Dissociation rates of the PD-L1 binding scFv fragments was determined with Octet (BLI). Abbreviations: kd, dissociation rate constant; hu, human; cy, cynomolgus. Table 11 Characterization of the Fab and scFv fragments binding to PD-L1 that were selected for further studies. Percentage of ligand blocking was determined by ligand blocking assay and immunogenicity scores were calculated. PD-L1 binders 1 and 3 are scFv fragments and PVRIG binders 2, 4-10 are Fab fragments. EXAMPLE 4 Screening of multi-specific PVRIG and PD-L1 antigen-binding proteins (PVRIG x PD-L1 binding proteins or bispecific proteins)

[0332] The two hundred different PVRIG x PD-L1 binding proteins generated from the preferred PVRIG and PD-L1 binders (as discussed above) were screened in various assays. PVRIG x PD-L1 binding proteins were tested in dual ELISA, followed by separate reporter assays (blocking of PD-1 / PD-L1 or CD112 / PVRIG) with three concentrations. In addition, a high-throughput SPR instrument (Carterra LSA) was used to determine binding affinities. In some cases, binding studies were also performed using BLI (Octet). The above characterizations represent steps 1-3 in table 12. After affinity ranking, additional reporter blocking assays were performed (9-10 concentrations), and EC50 values were calculated (step 4). PVRIG x PD-L1 binding proteins were also tested in two primary cell functional assays, the 0KT3-MDA and the SEE assay. After ranking based on IL-2 release, the best PVRIG x PD-L1 binding proteins were evaluated in cell binding studies (steps 6-7). First, binding to target and nontarget on CHO cells untransfected or stably expressing PD-L1 or PVRIG was measured using flow cytometry (step 6). Second, binding to endogenously expressed targets on MDA-MB-231 cells and PBMC (step 7) was measured. Step 8 involved flow cytometry-based ligand blocking (competition) assays, where the ability of the PVRIG x PD-L1 binding proteins to block the binding of biotinylated ligands was measured. PVRIG x PD-L1 binding proteins were further analysed by functional studies, effects on lymphocytes, antibody-dependent cellular cytotoxicity, and cell clustering (steps 9-12). Finally, PVRIG x PD-L1 binding proteins were studied in humanized mouse tumor model and gene expression was analysed (steps 13 and 14).

[0333] Sequence information for the CDRs, heavy and light chain variable regions and heavy and light chains of preferred individual binders and for various multi-specific antigen-binding proteins, including the preferred Multi-216, Multi-246 and Multi -256 bispecific molecules is provided in the sequence listing and tables correlating binder IDs and sequences included herein. The Multi-61 bispecific molecule was constructed based on sequences of previously described anti-PD-Ll and anti-PVRIG antibodies (SEQ ID NOs: 179 / 180 and 177 / 178 respectively), for comparison with multi-specific antigen-binding proteins based on the binder sequences identified by the inventors. For reference, Multi-216 comprises a first antigen-binding domain specific for PD- L1 (in the mAb position, PD-L1 4) comprising the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 79-84. This antigen-binding domain comprises the HCVR / LCVR pair of SEQ ID NOs 144 / 154. Multi-216 further comprises a second antigen-binding domain specific for PVRIG (in the Fab position, PVRIG 3) comprising the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 13-18. This antigen-binding domain comprises the HCVR / LCVR pair of SEQ ID NOs 123 / 133. Multi-216 comprises first, second and third polypeptide chains of SEQ ID NOs 161-163 respectively.

[0334] Multi-246 comprises a first antigen-binding domain specific for PD-L1 (in the mAb position, PD-L1 7) comprising the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 97-102. This antigen-binding domain comprises the HCVR / LCVR pair of SEQ ID NOs 147 / 157. Multi-246 further comprises a second antigen-binding domain specific for PVRIG (in the Fab position, PVRIG 3), for which sequences are provided above. Multi-246 comprises first, second and third polypeptide chains of SEQ ID NOs 164-166 respectively.

[0335] Multi-256 comprises a first antigen-binding domain specific for PD-L1 (in the mAb position, PD-L1 8) comprising the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 103-107. This antigen-binding domain comprises the HCVR / LCVR pair of SEQ ID NOs 148 / 158. Multi-256 further comprises a second antigen-binding domain specific for PVRIG (in the Fab position, PVRIG 3), for which sequences are provided above. Multi-246 comprises first, second and third polypeptide chains of SEQ ID NOs 167-169 respectively.

[0336] Table 12 Summary of characterization of PVRIG x PD-L1 binding proteins.

[0337] Abbreviations: E, example number; PBMC, peripheral blood mononuclear cells; ADCC, antibody-dependent cellular cytotoxicity; SPR, surface plasmon resonance; KD, equilibrium constant (affinity); DC, dendritic cell; IFN, interferon EXAMPLE 5 Binding of the bispecific proteins to PVRIG and PD-L1 and effect of position of the binders

[0338] Characterization of selected PVRIG x PD-L1 binding proteins

[0339] Strength of binding and consequent functionality of PVRIG or PD-L1 binders was found to differ depending on positioning within the bispecific format (the RUBY™ format described in Nyesiga et al.) In particular, effects varied depending on whether a PVRIG or PD-L1 binder were provided in the mAb or Fab position (i.e. as the monoclonal antibody of the bispecific molecule or the Fab fragment of the bispecific molecule). Chosen bispecific molecules were purified using protein A chromatography and tested for binding using SPR.

[0340] In particular, PVRIG 3 affinity was found to be 1.3 nM in the mAb position (multi- 108) but 7-15 nM in the Fab position. PVRIG reporter assays performed using these bispecific molecules showed that the drop in affinity led to reduced reporter activity when PVRIG 3 was in the Fab position (EC50 = 1.1 - 2.7 nM) instead of the mAb position (EC50 = 0.3 nM). Similar results could be seen for the PD-L1 binders such as PD-L1 5 which had an affinity to PD-L1 of around 1 nM and a corresponding reporter assay EC50 between 2-3 nM in the mAb position. The affinity for PD-L1 5 dropped to 17 nM in the Fab position while PD-L1 reporter activity dropped to 13 nM.

[0341] Table 13, table 14, table 15 and table 16 present characteristics of selected preferred PVRIG x PD-L1 binding proteins. The EC50 values in reporter assays were mostly in the single-digit nM range. In contrast, in the 0KT3-MDA dual functional assay, EC50 values were in single- or double-digit pM range. The SPR data shown was obtained at the screening stage, using antibodies purified only by affinity chromatography (protein A). Low nM affinities were seen for human PD-L1. For PVRIG binding kinetics, a PVRIG protein fused with albumin was used, which weakens binding compared to a PVRIG protein without any bulky additions. This is why the PVRIG binding data shown here is slightly weaker than in later experiments where PVRIG without albumin was used as the analyte. Affinity for PVRIG was 1.3 nM, when the PVRIG 3 binder was in the mAb position (see multi-108), but 6-13 nM when the same binder was in the Fab position. This indicates that the albumin fusion does not impair PVRIG binding to the normal (mAb) binding domains but interferes slightly with binding to the appended Fab arms.

[0342] It should be noted that EC50 values of PVRIG reporter assays from PVRIG x PD- L1 binding proteins when PVRIG 3 was in the mAb position (multi- 108) was in the pM range as opposed to single digit nM range when in the Fab position (multi-246 and multi- 256) reflecting the lowered affinity of PVRIG 3 in the Fab position.

[0343] Table 13 Characterization of PVRIG x PD-L1 binding proteins, PD-L1 equilibrium constant (affinity) (KD), association rate constant (ka), dissociation rate constant (kd).

[0344] Table 14 Charaterization of PVRIG x PD-L1 binding proteins, PVRIG equilibrium constant (affinity) (KD), association rate constant (ka), dissociation rate constant (kd).

[0345]

[0346] Table 15 Charaterization of PVRIG x PD-L1 binding proteins, PD-L1 and PVRIG reporter assays.

[0347]

[0348] Table 16 Characterization of PVRIG x PD-L1 binding proteins, 0KT3-MDA assays.

[0349] The effect of binder position on the efficacy in the reporter assays Table 17 and table 18 summarize the effect of binder position in PVRIG x PD-L1 binder proteins on their efficacy in a reporter assay. Fold induction of reporter signal using 50 nM of each PVRIG x PD-L1 binding protein was compared with that observed with the PD-L1 and PVRIG binding domains in opposite orientations. For most binding domains, the ’mAb position’ was found to work better. Induction of the reporter signal was stronger when the anti-PD-Ll arms were in the ‘mAb position’. The PVRIG blocking effect was also impaired when the anti -PVRIG arm was in the ‘Fab position’, but not as much as observed with PD-L1. These differences may depend on both the targets and the individual PVRIG x PD-L1 binding proteins (the target epitope and binding mode), as well as the compatibility of the binding domains in a specific orientation in the PVRIG x PD-L1 binding protein.

[0350] Table 17 The effect of binder position in PVRIG x PD-L1 binding proteins, with PD-L1 binder in mAb position and PVRIG binder in Fab position.

[0351] Table 18 The effect of binder position in PVRIG x PD-L1 binding proteins, with PVRIG binder in mAb position and PD-L1 binder in Fab position.

[0352] Binding kinetics (SPR) for PD-L1 Table 19 presents binding kinetics (SPR) for PD-L1. For the particularly preferred bispecific molecules Multi-216, Multi -246 and Multi -256, binding kinetics were very similar for human and cynomolgus PD-L1.

[0353] Antibodies were captured to SPR chips and soluble target proteins were used as analytes. The analytes were human and cynomolgus PD-L1, where amino acids 1-239 of the canonical Uniprot PD-L1 sequences were included. This part represents the extracellular domain of PD-L1.

[0354] Table 19 Binding kinetics (SPR) for human and cynomolgus PD-L1

[0355] Binding kinetics (SPR) for PVRIG

[0356] Table 20 presents binding kinetics (SPR) for PVRIG. For Multi-216, Multi-246 and Multi-256, binding to human PVRIG was stronger than binding to cynomolgus PVRIG. Antibodies were captured to SPR chips and soluble target proteins were used as analytes. The analytes were human and cynomolgus PVRIG, where amino acids 41-171 (human) or 41-172 (cynomolgus) of the canonical Uniprot PVRIG sequences were included.

[0357] Table 20 Binding kinetics (SPR) for human and cynomolgus PVRIG

[0358] Epitope mapping

[0359] Each amino acid of the relevant antigen sequences was replaced with the 20 other amino acids, and the effect of each mutation on antibody-antigen interaction was determined. For this, a combination of yeast surface display, flow cytometry and high- throughput sequencing was used (Van Blarcom et al. 2015, Medina-Cucurella and Whitehead 2018, Plaisance et al. 2000, Pruvost et al. 2023). Yeast cells were induced to express either human PD-L1 or human PVRIG, both with hemagglutinin (HA) tag, and expression was verified using biotinylated antibodies that recognize the antigen, streptavidin-PE (phycoerythrin), and by flow cytometry. Both antigens were well expressed and displayed on the yeast cell surface. Binding strength was assessed by flow cytometry at several concentrations ranging from 12.8 pM to 134 nM (PD-L1) or 2.6 pM to 25 nM (PVRIG) using either biotinylated antibodies (and subsequent detection using streptavidin-PE) or antibodies coupled to allophycocyanin (APC).

[0360] In separate experiments, yeast cells were incubated with antibodies (0.5-2.5 nM depending on the antibody) and cell sorting was performed using an SRT cell sorter (Beckman Coulter). For epitope mapping, DMS (Deep Mutational Scanning) libraries of the antigens were generated (for PD-L1, from position 18 to 131 and from 132 to 238; for PVRIG, from position 41 to 172) by Deeptope. These libraries contained all the possible single mutants of the antigens. The different pools of mutants (loss of binding) were identified and sorted using cytometry, and the mutations causing loss of binding to the antibody of interest but not to the control IgG identified by high-throughput sequencing.

[0361] The sorting for both antigens included a conformational control to exclude mutants with loss of conformation. For each alternative amino acid, the enrichment in the pool with lost binding to the IgG of interest was determined, and the number of such amino acid substitutions at each position of the antigen was counted. Structural positions were those where mutations caused loss of binding to two IgGs with non-overlapping epitopes, or positions not exposed on the antigen surface. The results were summarized into heatmaps, where key antigen residues for the binding of the tested antibodies were identified (Figure 9). The different epitopes were visualized on antigen structures modeled with Alphafold2.

[0362] Epitope mapping for PD-L1

[0363] The PD-L1 binding domains PD-L1 7 and PD-L1 8 of Multi-246 and Multi-256, respectively, were epitope mapped (Figure 9A). For Multi-246 (PD-L1 7) the key binding amino acids of PD-L1 in contact with the PD-L1 binding domain were determined to be D26, D122, Y123, K124, and R125. A more extensive list of all potential epitope residues is as follows: V23, D26, E58, Ml 15, D122, Y123, K124, R125. For Multi-256 (PD-L1 8) the key binding amino acids of PD-L1 in contact with the PD-L1 binding domain were determined to be E58, Ml 15, and A121. A more extensive list of all potential epitope residues is as follows: Y56, E58, R113, Ml 15, SI 17, A121, D122. Epitope mapping for PVRIG

[0364] The epitopes for Multi-246, Multi-256 and anti-PVRIG antibody 743 are identical, as they all contain the same anti-PVRIG binding domain, namely PVRIG 03 (Figure 9B). For Multi-246, Multi-256 and anti-PVRIG antibody 743 the key binding amino acids of PVRIG in contact with PVRIG binding domain were determined to be L72, V90, R95, G96, R98, W100, A137, F139, P140, and G142. A more extensive list of all potential epitope residues is as follows: S71, L72, V90, H92, P93, R95, G96, R98, W100, A137, F139, P140, G142, S143.

[0365] EXAMPLE 6 Reporter assays for PVRIG and PD-L1 blocking

[0366] PVRIG / PVRL2 blocking functional reporter assay

[0367] PVRIG binding and blocking towards its ligand PVRL2 and functional blocking efficacy by the PVRIG x PD-L1 binding proteins and an anti-PVRIG antibody were evaluated by using CD112R / CD112 blocking functional bioassay produced by Promega.

[0368] The bioassay involves two engineered cell lines, namely Jurkat T cells engineered to have a NFAT-based luciferase activity in the presence of aAPC CHO-K1 cells that express PVRL2 and a TCR activating molecule on their surface. Blocking the PVRIG on Jurkat cells leads to disruption of interaction between PVRIG and PVRL2 that results in engagement of these two cell lines and thereby TCRs activation on Jurkat cells followed by NFAT-driven luciferase production, which then cleaves luciferin substrate and results in luminescence signal. The luciferase activity shows how well the Jurkat TCRs are activated, and thereby how efficaciously the PVRIG / PVRL2 interaction has been blocked.

[0369] Figure 10 summarizes the PVRIG reporter assay results for selected PVRIG x PD- L1 binding proteins. The fold increase is calculated based on the division of the relative luminescence units detected within each group where the highest concentration of the PVRIG x PD-L1 binding proteins are versus where there is no binding proteins or antibodies. Besides, an anti-PVRIG antibody has the highest capacity to block PVRIG / PVRL2 interaction, all the selected PVRIG x PD-L1 binding proteins multi-216, multi-246, multi-256 and multi-61 had similar blocking capacities. The EC50 values for the PVRIG x PD-L1 binding proteins were 1.2-2.7 nM.

[0370] PD-1 / PD-L1 blocking functional reporter assay

[0371] Efficacy of the PVRIG x PD-L1 binding protein in binding to PD-L1 and blocking of its binding to its receptor PD-1 was evaluated by using PD-1 / PD-L1 Bio-IC assay produced by Invivogen.

[0372] The bioassay involves a Jurkat-Lucia TCR-PD-1 T cell line, that is engineered to have a NF AT -based Lucia activity upon the interaction with another engineered target cell line Raji-APC-hPD-Ll that expresses PD-L1 and a TCR activator on their surface. Blocking the PD-L1 on Jurkat cells leads to disruption of the interaction between PD-1 and PD-L1 that results in luciferase transcription that cleaves QUANTI-Luc substrate resulting in luminescence signal. The signal is higher compared to PVRIG reporter assay because this reporter kit has six NF AT regions that amplify the signal.

[0373] Figure 11 summarizes the PD-L1 reporter assay results for selected PVRIG x PD- L1 binding proteins multi-216, multi-246, multi-256 and multi-61 and a PD-Ll binding antibody Tecentriq (atezolizumab). The fold increase is calculated based on the division of the relative luminescence units detected within each group where the highest concentration of the PVRIG x PD-L1 binding proteins are versus where there is no binding proteins or antibodies. All the PVRIG x PD-L1 binding proteins blocked the PD-L1 binding to PD-1 with a comparable EC50 values between 3.0-5.7 nM.

[0374] PVRIG PVRL2 ligand blocking assay

[0375] Blocking capacity of PVRIG x PD-L1 binding proteins to block PVRL2 binding to PVRIG was measured by a bioassay utilizing CHO cells overexpressing PVRIG on their surface. Briefly, in a 96-well round bottom plate, after addition of the PVRIG x PD-L1 binding proteins the PVRL2 conjugated with hFc protein containing biotin (CD112 hFc- biotin, Aero Biosystems) were added into the reaction. Thereafter, Streptavidin molecules that are conjugated with APC (SA-APC, BD Biosciences) as a color indicator. The detection of PVRL2-hFc-Biotin (15 pg / ml) and SA-APC binding to PVRIG expressing CHO cells was measured by flow cytometry. The APC signal decrease means the biotinylated ligand PVRL2 binds less to PVRIG.

[0376] PVRIG x PD-L1 binding proteins multi-216, multi-246, multi-256 and multi-61 blocked the PVRIG / PVRL2 interaction in similar range to each other which were observed at subnanomolar level (Figure 12). An anti-PVRIG antibody was included as a positive control in the assay.

[0377] PD-1 receptor - PD-L1 ligand blocking assay

[0378] Blocking capacity of PVRIG x PD-L1 binding proteins to block PD-L1 binding to PD-1 was measured by a bioassay utilizing CHO cells overexpressing PD-L1 on their surface.

[0379] Briefly, in a 96-well round bottom plate, after addition of the PVRIG x PD-L1 binding proteins the PD-1 conjugated with hFc protein containing biotin (hPD-1 hFc- biotin, RnD Systems) were added into the reaction. Thereafter, Streptavidin-APC (SA- APC, BD Biosciences) was added as a color indicator. The detection of PD-l-hFc-Biotin (15 pg / ml) and SA- APC binding to PD-L1 expressing CHO cells was measured by flow cytometry. The APC signal decrease means the biotinylated PD-1 binds less to PD-L1.

[0380] PVRIG x PD-L1 binding proteins multi-216, multi-246 and multi-256 blocked PD- 1 binding to PD-L1 in comparable levels with anti-PD-Ll antibody Tecentriq (atezolizumab) (Figure 13).

[0381] PD-L1 - CD80 ligand blocking assay

[0382] Although expressed on the same Antigen Presenting Cells, CD80 molecules can interact with PD-L1 in cis-mode as well as trans-mode. Blocking PD-L1 cis-or transbinding to CD80 was shown to augment and be essential for a durable response in antitumor immune checkpoint therapies (Sugiura et al., 2019).

[0383] In this assay the blocking capacity of PVRIG x PD-L1 binding proteins to block CD80 binding to PD-L1 was evaluated by a bioassay utilizing CHO cells overexpressing PD-L1 on their surface. Briefly, in a 96-well round bottom plate, after addition of the PVRIG x PD-L1 binding proteins the CD80 conjugated with hFc protein that contains Biotin (hCD80 hFc- biotin, Aero Biosystems) were added into the reaction followed by addition of Streptavidin molecules that are conjugated with APC as a color indicator (SA-APC, BD Biosciences). CD80-hFc-Biotin (15 pg / ml) and SA-APC binding to CHO cells expressing PD-L1 was measured with flow cytometry. The APC signal decrease means the biotinylated CD80 binds less to PD-L1.

[0384] All the PVRIG x PD-L1 binding proteins multi-216, multi-246, multi-256 and multi-61 blocked the interaction between PD-L1 and CD80 at very similar concentrations and efficacies (Figure 14).

[0385] EXAMPLE 7 Functionality in NK cells and dual-targeting OKT3-MDA assay

[0386] Functionality of the PVRIG x PD-L1 binding protein in co-culture of primary NK cells and K562 tumor cells

[0387] Functionality of the PVRIG x PD-L1 binding proteins was evaluated by coculturing primary NK cells and K562 tumor cells. Primary NK cells were isolated from peripheral blood mononuclear cells (PBMCs) using EasySep Human NK cell isolation kit (StemCell Technologies). The NK cells were co-cultured together with K562 tumor cells in the presence of IL-12 (R&D Systems) and the isotype control bispecific antibody multi- 287 (a human IgGl isotype antibody that targets GFP (Mab position) and FITC (Fab position)), anti-PVRIG antibody, multi-72 (a human IgGl isotype bispecific protein targeting GFP on Fab and PD-L1 on MAb domains), or PVRIG x PD-L1 binding proteins multi-61, multi-216, multi-246 and multi -256. After 48h incubation, the secreted IFN- gamma_was measured by ELISA (ELISA Pro: Monkey IFN-gamma (crossreacts with human IFN-gamma, Mabtech)).

[0388] To eliminate the donor dependent variation within the assay, IFN-gamma quantity data was normalized to the functional effect of multi-72 (a human IgGl isotype bispecific protein targeting GFP on Fab and PD-L1 on MAb domains) as baseline and the fold change was calculated (Figure 15). The results show that there is an added effect of anti-PVRIG antibody compared to multi-72 meaning PVRIG blockade enhances NK cell functions towards K562 tumor cells. All the tested PVRIG x PD-L1 binding proteins multi-61, multi-216, multi-246 and multi- 256 augmented NK cell functions towards K562 tumor cells similarly as an anti-PVRIG antibody.

[0389] Functionality of the PVRIG x PD-L1 binding proteins with different concentrations in coculture of NK-92 and K562 tumor cells

[0390] Functionality of the PVRIG x PD-L1 binding protein in different concentrations was evaluated by co-culturing NK-92 cells and K562 tumor cells. The NK-92 cells were co-cultured with K562 cells in IL-2 (Human IL-2 IS, Miltenyi Biotec) containing cell culture medium for 48h in the presence of an anti-PVRIG antibody, an anti-PD-Ll antibody, combination of the anti-PVRIG and anti-PD-Ll antibodies and PVRIG x PD-L1 binding proteins in different concentrations. After the co-culture the cytokine IFN-gamma secreted to the supernatant was quantified with ELISA (Human IFN-gamma High Sensitivity ELISA Kit, Abeam).

[0391] Figure 16 presents the results from co-culture of NK-92 and K562 cells with an anti-PVRIG antibody, an anti-PD-Ll antibody and their combination, and different PVRIG x PD-L1 binding proteins multi-216, multi-246, multi-256 and multi-61. Both PVRIG x PD-L1 binding proteins and the anti-PVRIG antibody induced IFN-gamma production in the cells. The anti-PD-Ll antibody (human IgGl type) alone does not increase IFN-gamma secretion in the NK-92 cells. The combination of both anti-PVRIG and anti-PD-Ll antibodies also shows a similar increase in IFN-gamma secretion as the anti-PVRIG monospecific antibody and the PVRIG x PD-L1 binding proteins multi-216, multi-246 and multi-256. EC50 values were calculated for the PVRIG x PD-L1 binding proteins multi- 216, multi-246 and multi-256 being 0,5022 nM, 0,5295 nM and 0,5317 nM, respectively.

[0392] OKT3 dual target assay

[0393] The OKT3 dual targeting assay was carried out to evaluate the synergistic effect of PVRIG x PD-L1 binding proteins. In this assay primary CD4+ T cells isolated from healthy donor PBMCs (CD4+ T cell Isolation Kit (human) (Miltenyi Biotec) that express PVRIG were used as effector cells. MDA-MB-231 cell line that express PD-L1 and PVRL2 on their surface was used as target cells. MDA-MB-231 cells were transfected to express 0KT3 clone anti-CD3 antibody on their surface. In the assay, the functional activation of the CD4+ T cells is dependent on the blockade of both PVRIG and PD-1 receptors, since the engagement of these receptors with their corresponding ligands regulate the activation signaling negatively despite on the activating TCR-engaging anti- CD3 antibody on the MDA-MB-231 tumor cells. The aim was to assess the dual synergistic effect of binding of the PVRIG x PD-L1 binding proteins to the targets and evaluate their effectiveness on the activation of the CD4+ T cells.

[0394] Briefly, the CD4+ T cells and MDA-MB-231 cells expressing anti-CD3 antibody on their surface were co-cultured for 48h in the presence of an anti-PVRIG antibody, an anti-PD-Ll antibody Tecentriq (atezolizumab), their combination or PVRIG x PD-L1 binding proteins. After the co-culture IL-2 cytokine levels in cell culture mediums were measured by ELISA. (ELISA Pro: Human IL-2 kit, Mabtech) IL-2 secretion indicates activation of the T cells. For efficacy determination, the IL-2 concentrations were normalized to Max response for anti-PD-Ll (Tecentriq) in each plate and plotted as mean dose-response for 6 to 18 donors (Figure 17). Equation Normalization for efficacy Normalized response = 100*(Response test article - Response background) / (Max Response Anti-PD-Ll (Tecentriq - Response background). Background response may be response from blank sample or from isotype control showed that all anti-PVRIG x PD-L1 binding proteins are comparable to or more efficacious than the combination of the anti- PD-Ll antibody and anti-PVRIG antibody.

[0395] EXAMPLE 8 Binding of bispecific protein to PD-L1 and PVRIG overexpressing cells, PBMCs and tumor cell lines

[0396] Specific binding of PVRIG x PD-L1 binding proteins on stimulated and non-stimulated PBMCs

[0397] To understand the specific binding of PVRIG x PD-L1 binding proteins to different cell populations in unstimulated and anti-CD3 stimulated PBMCs, the expression of PVRIG and PD-L1 was studied by flow cytometry in T cells, NK cells, B cells, and monocytes.

[0398] PBMCs were stimulated with 3 pg / ml anti-CD3 (OKT3) antibody for 24h. Then PVRIG and PD-L1 expression on different immune cell populations was determined with flow cytometry.

[0399] Figure 18 shows flow cytometry results of the PD-L1 and PVRIG expression. PD- L1 was expressing on NK cells, T cells, B cells and monocytes, while PVRIG on NK cells, T cells, B cells and to a lesser degree on monocytes.

[0400] The binding of an anti -PVRIG antibody, anti-PD-Ll antibody Tecentriq (atezolizumab) and of PVRIG x PD-L1 binding proteins multi-216, multi-246 and multi- 256 to specific immune cell populations was evaluated in stimulated PBMCs with increasing concentrations (Figure 19). The anti -PVRIG and anti-PD-Ll antibodies and PVRIG x PD-L1 binding proteins contained biotin in their Fc domain. Binding was detected by adding streptavidin conjugated fluorochromes and flow cytometry. The percentage of the stained samples was determined by the mean fluorescence intensity (MFI).

[0401] Binding of PVRIG x PD-L1 binding proteins on CHO cells expressing PVRIG and PD-L1

[0402] Binding of PVRIG x PD-L1 binding proteins with different concentrations to PVRIG and PD-L1 on cell surface was studied with CHO cells overexpressing PVRIG or PD-L1. Figure 20A shows the binding of PVRIG x PD-L1 binding proteins multi-216, multi-224, multi-226, multi-236, multi-244, multi-246 and multi-256 to PVRIG on CHO cells. Binding of anti-PVRIG antibody and anti-PD-Ll antibody was also studied. PVRIG x PD-Llbinding proteins multi-216, multi-226, multi-236, multi-246 and multi-256 showed higher binding to PVRIG than anti-PVRIG antibody, while multi-224 and multi- 244 had lower binding. Figure 20B shows the binding of PVRIG x PD-L1 binding proteins multi-216, multi-224, multi-244, multi-226, multi-236, multi-246, multi-256 and multi-61 to PD-L1 on CHO cells. Binding of anti-PD-Ll antibody was also studied. All the PVRIG x PD-L1 binding proteins bound similarly to PD-L1 as the anti-PD-Ll antibody. Figure 20C shows binding of PVRIG x PD-L1 binding proteins multi-216, multi -224, multi-226, multi-236, multi-244, multi-246 and multi-256 to breast cancer cell line MDA-MB-231 that expresses PD-L1 but not PVRIG.

[0403] Binding of PVRIG x PD-L1 binding proteins on tumor cell lines

[0404] To identify tumor cell lines that express PD-L1 but not PVRIG, tumor cell lines MDA-MB-231 (breast cancer), Pane 05.04 (pancreas adenocarcinoma) HTC116 (colon cancer) were stained with anti -PVRIG- APC and anti-PD-Ll-APC antibodies (Biolegend) for flow cytometry. MDA-MD-231 cells were found to have high PD-L1 expression, whereas Pane 05.04 cells had intermedial PD-L1 expression and HCT116 cells low PD-L1 expression (Figure 21A).

[0405] Binding of PVRIG x PD-L1 binding proteins multi-216, multi-246, multi-256 and multi-61 with different concentrations to PD-L1 expressed on MDA-MB-231, Pane 05.04 and HTC116 cells was studied by flow cytometry. Secondary anti-human Fc gamma binding PE labeled antibody (Jackson ImmunoResearch) was used to detection. Figures 21B to 21D show that all the tested PVRIG x PD-L1 binding proteins multi-216, multi- 246, multi-256 and multi-16 as well as anti-PD-Ll antibody Tecentriq (atezolizumab) bound to the cells in a dose-dependent fashion and the binding correlated with the PD-L1 expression levels on the tumor cells. Anti-PVRIG antibody and isotype control bispecific antibody multi-287 (a human IgGl isotype antibody that targets GFP (Mab) and FITC (Fab)) did not bind to the cells. Binding of anti-PVRIG antibody, anti-PD-Ll antibody Tecentriq and PVRIG x PD-L1 binding proteins multi-61, multi-246 and multi-256 to Fc- gamma receptor (FcgR) family members was studied in CHO cells expressing Fc gamma RI (Figure 22 (A), Fc gamma Rlla (H131) (C), Fc gamma Rlla (R131) (D), Fc gamma Rllb (E) and Fc gamma Rllla (F)). Concentration series was 0.006 nM, 0.024 nM, 0.098 nM, 0.39 nM, 1.56 nM, 6.3 nM, 25 nM, 100 nM, and 400 nM. Figure 22A show the binding based on the PE binding mean fluorescence intensity (MFI) as a comparison to Figure 22B, which shows the control non-specific PE binding MFI in the same set of studies.

[0406] All Fc active wild-type IgG containing anti-PVRIG antibody and PVRIG x PD-L1 binding proteins multi-61, multi-246 and multi-256 bind to FcgRI (CD64) similarly (Figure 22A). Anti-PD-Ll antibody Tecentriq did not bind to FcgRI due to its mutated Fc domain. Figure 22B shows that all the tested molecules act on FcgRI specifically except multi-216 that showed a slight tendency of causing a non-specific binding to wild type CH0-K1 cells.

[0407] Figures 22C and 22D show binding to FcgRIIa high affinity variant H131 and low affinity variant R131. Multi-216 showed the highest binding affinity (this may be due to the non-specific binding to WT-CH0-K1 cell binding). Multi-246 and multi-256 showed similar but 2 to 10-fold lower binding compared to multi-216, multi-61 and anti-PVRIG antibody in FcgRIIA binding of low and high affinity variants. In Figure 22E FcgRIIb binding is evaluated with the same concentrations. The data in Figure 22E show that anti- PVRIG antibody had the highest binding to FgRIIb, followed by multi-61 and multi-216, whereas multi-246 and multi-256 had several-fold lower binding affinity compared to high affinity binding multi-61. Anti-PD-Ll antibody Tecentriq having mutated Fc domain did not show any binding to FcgRIIb. Figure 22F show the binding to FgRIIIA (V158). Anti- PD-Ll antibody Tecentriq with mutated Fc domain (used as a negative control), at 6.3 nM all the monospecific and bispecific antibodies bind to FcgRIIIA in 100% efficiency, whereas at 1.56 nM multi-256 had the lowest binding at 32%, followed by 52% binding of multi-61, 76% anti-PVRIG antibody, 78% multi -246 and 96% multi-216.

[0408] EXAMPLE 9 Mixed lymphocyte reaction (MLR) assays: Functionality in memory T cells and monocytic dendritic cells

[0409] Autologous MLR assay in co-culture of memory T cells and monocytes

[0410] Enhanced functional activity effect of the PVRIG x PD-L1 binding proteins was studied in co-culture of autologous memory T cells and monocytes. Primary memory T cells that express CD45RO+ and primary monocytes were isolated from peripheral blood mononuclear cells (PBMCs) separated from buffy coats. The two immune cell populations were co-cultured for 96h in the presence of a hlgGl antibody (isotype control) (Bio X Cell), an anti-PVRIG antibody, an anti-PD-Ll antibody (Tecentriq, atezolizumab), a combination of the anti-PVRIG and anti-PD-Ll antibodies, and PVRIG x PD-L1 binding proteins multi-216, multi-246, multi-256 and multi-61. Each molecule was studied in six different concentrations 69 nM, 13.8 nM, 2.8 nM, 0.55 nM, 0.11 nM, and 0.02 nM. At the end of the co-culture the IFN-gamma secretion to the cell culture medium was measured with ELISA (Promega Lumit).

[0411] When the anti-PD-Ll antibody, combination of the anti-PD-Ll and anti-PVRIG antibodies and the PVRIG x PD-L1 binding proteins were applied to the mixed lymphocyte reaction a functional increase on memory T cell activation in a dose-dependent fashion was observed (Figure 23). However, the anti-PVRIG antibody alone did not have an effect on memory T cell activation (Figure 23). Memory T cell activation in the assay was dependent on PD-L1 blocking. The increase of IFN-gamma fold change was calculated and shown normalized to hlgGl antibody (Bio X Cell).

[0412] Based on the concentration series EC50 values were calculated for all the PVRIG x PD-L1 binding proteins multi-216, multi-246, and multi-256 with EC50 values 0,5965 nM, 0,4389 nM, and 0,7442 nM, respectively. Thus, all the selected PVRIG x PD-L1 binding proteins showed efficacy in memory T cell activation.

[0413] Allogeneic MLR assay in co-culture of peripheral blood mononuclear cells (PBMC) and monocytic dendritic cells - Tissue mismatch immune cell activation

[0414] To assess whether PVRIG x PD-L1 binding proteins augmented the immune reactions caused by a haplotype mismatch, PBMCs separated from buffy coats were mixed and co-cultured with another donor’s monocyte-derived dendritic cells (CD14, CD11c). For this mixed lymphocyte reaction, the PBMCs and the dendritic cells were co-cultured for 96h in the presence of different concentrations of an anti-PD-Ll antibody (Tecentriq, atezolizumab) and PVRIG x PD-L1 binding proteins multi-61, multi-216, multi-246, and multi-256 with different concentrations.

[0415] At the end of the co-culture, 4-1BB (CD137) protein on cell surface was measured with flow cytometry (anti-CD137 PE, Biolegend) to assess the activation of the cells. Figure 24 presents the assay results showing that addition of PVRIG x PD-L1 binding proteins increase 4- IBB expression on the NK cell surface dose-dependently, whereas anti- PD-Ll antibody Tecentriq (atezolizumab) did not affect the expression of 4-1BB on the NK cells. This effect is likely to be attributed to active Fc domain present in the antibody and PVRIG x PD-L1 binding proteins that is binding to the FcgRIIIa on NK cells. All the selected PVRIG x PD-L1 binding proteins multi-216, multi -246, multi -256 and multi-61 increased 4-1BB expression, showing the effect on NK cells in this assay. Multi-246 showed the highest effect. Table 21 Antibodies used in flow cytometry analysis for Allogeneic MLR assay EXAMPLE 10 Antibody-dependent cellular cytotoxicity (ADCC)

[0416] ADCC Reporter Assay

[0417] One further mechanism of action observed for PVRIG x PD-L1 binding proteins was antibody-dependent cellular cytotoxicity (ADCC) exerted by the active Fc domain present in the proteins. The ADCC occurs when an active Fc domain binds to a Fc(gamma)RIIIA receptor (CD 16a) mainly expressed on the NK cells. CD 16a can be expressed also on other immune cells such as monocytes, macrophages, neutrophils, and eosinophils for various functions. For NK cells to be activated for killing the antibody opsonized target cells there needs to be a certain number of antibodies to create an avidity (bundle) effect.

[0418] ADCC reporter assay V-variant (Promega) was used to study ADCC effect caused by PVRIG x PD-L1 binding proteins. The ADCC was studied in CHO cells expressing PVRIG or PD-L1 in the presence of anti -PVRIG antibody, anti-PD-Ll antibody, combination of anti-PVRIG and anti-PD-Ll antibodies, PVRIG x PD-L1 binding proteins multi-61, multi-108, multi-216, multi-224, multi-226, multi-244, multi-246 and multi-256 and multi-72 (a control hlgGl isotype bispecific protein targeting GFP on Fab and PD-L1 on Mab domains) and isotype control bispecific antibody multi-287 (a hlgGl isotype antibody that targets GFP (Mab) and FITC (Fab)) (Figure 25). The ADCC was detected with a PD-L1 antibody and PVRIG x PD-L1 binding proteins multi-61, multi-216, multi- 224, multi-226, multi-244, multi-246 and multi-256 (PD-L1 binding domain in Mab position and PVRIG binding domain in Fab position) in CHO cells expressing PD-L1 (Figures 25A and 25C). The ADCC also was detected in CHO cells expressing PVRIG, when anti-PVRIG antibody or multi- 108 (PVRIG binding domain in Mab position and PD- L1 binding domain in fab position) was present. However, ADCC was not detected in PVRIG expressing CHO cells in the presence of PVRIG x PD-L1 binding proteins in which PD-L1 binding domain in Mab position and PVRIG binding domain in Fab position) (Figures 25B and 25D). This is considered to be due to structural properties of the PVRIG x PD-L1 binding proteins and the location of the PVRIG and PD-L1 binders in comparison with the Fc domain of the proteins. PBMCADCC

[0419] Since PVRIG and PD-L1 are also expressed in functional immune subsets of healthy humans, it was important to evaluate whether these PVRIG x PD-L1 binding proteins could cause a reduction in any immune cell populations. The PVRIG x PD-L1 binding proteins and a combination of anti-PVRIG and anti-PD-Ll antibodies were incubated in the PBMCs for 4h. Then, the PBMCs were stained and CD3+, CD4+, CD8+ and dead cell populations were evaluated for their depletion compared to the IgG control and no treatment groups by flow cytometry. During the analysis the absolute counts of the detected cells in pl volume per cell populations were taken into account. An anti-CD52 antibody Campath (alemtuzumab) has been previously identified as an ADCC-driven antibody (Rao, S. P. et al. 2012) and is known to deplete the effector CD4+ and CD8+ T cells. For this reason, it was included as positive control to observe and compare whether any of the PVRIG x PD-L1 binding proteins cause any the reduction of the immune cell populations similar to Campath. Figure 26 shows the results that Campath caused a pronounced cell death effect, which was almost 6-fold as median compared to all tested PVRIG x PD-L1 binding proteins and control IgG.

[0420] Among the analyzed cell populations Campath clearly depleted the CD3+, CD4+ and CD8+ T cells more than any of the studied PVRIG x PD-L1 binding proteins. There was minimal decrease in CD3+, CD4+ and CD8+ cells with PVRIG x PD-L1 binding proteins multi-61, multi-216 and multi-246. With PVRIG x PD-L1 binding protein multi- 256 there was no reduction in CD3+, CD4+ and CD8+ cells compared to sample without treatment or with IgG isotype. (Figure 26)

[0421] EXAMPLE 11 PVRIG x PD-L1 binding proteins induce clustering of PD-L1 expressing cells and PVRIG expressing cells

[0422] Clustering assay using CH0-K1 cells expressing either PVRIG or PD-L1

[0423] PVRIG x PD-L1 binding proteins allow to bring PVRIG expressing T or NK cells together with PD-L1 expressing antigen presenting cells (APCs) or tumor cells. This may be important to trigger the immune synapse formation that involves T cell receptors (TCRs) on T cells and APCs and / or tumor cells. In addition, NK cells and APCs or tumor cells can form a similar synapse using for example KIRs (Killer Cell Immunoglobulin-like receptors) and certain HLA molecules. PVRIG x PD-L1 binding proteins are considered to enforce these synapses by augmenting through cluster formation. A combination of individual anti-PVRIG and anti-PD-Ll antibodies however was not able to cause this effect. The clustering effect is schematically presented in Figure 27.

[0424] To evaluate the cluster formation induced by PVRIG x PD-L1 binding proteins, CH0-K1 cells were engineered to overexpress either PVRIG or PD-L1. Both cell lines were intracellularly stained, namely CH0-K1 cells expressing PVRIG with CFSE and CH0-K1 cells expressing PD-L1 with CellTrace Violet. The cells were combined and incubated together with by PVRIG x PD-L1 binding proteins in a concentration series ranging from 0.02 nM to 930 nM for 30 minutes (on ice, every 10 minutes gently mixed by pipetting). Flow cytometry analysis was done in two different ways, by analyzing all live cells and gated doublets (the clustered cells). The gated cells in flow cytometry analysis are shown in Figure 28. The flow cytometry events that appear in concomitant colors represent the two CHO cell lines that expresses PD-L1 and PVRIG that have been clustered (Figure 28A).

[0425] The clustering assay titrated the concentration when PVRIG x PD-L1 binding proteins start to cluster the PVRIG expressing and PD-L1 expressing cells. The EC 50 values were calculated for multi -246, multi-256 and multi-61 (Figure 28B). Clustering peak for multi-246 was at 0.36 nM, for multi-256 at 0.49 nM and for multi-61 at 0.82 nM. The clustering of the cells was induced and also decreased with multi-61 in lower concentrations than with multi-246 and multi-256.

[0426] In addition, the assay showed at which concentration the clustering started to decrease. This is due to the saturation of the binding sites on both cells.

[0427] EXAMPLE 12: PVRIG x PD-L1 binding proteins show improved anti-tumor efficacy in humanized mice in vivo.

[0428] An efficacy study was performed using CD34+ humanized-mice, post humanization mice were implanted subcutaneously into the right flank with 5xlO6MDA-MB-231 tumor cells. Starting from Day 0, animals were observed daily and weighed thrice weekly; data including individual and mean gram weights, mean percent weight change versus Day 0 (%vDO) were recorded for each group and %vD0 plotted at study completion. Animals exhibiting > 10% weight loss when compared to Day 0, if any, were provided with DietGel™ (ClearH2O®, Westbrook, ME) ad libitum. Groups reporting a mean %vD0 > 20% and / or >10% mortality were considered above the maximum tolerated dose (MTD) for that treatment on the evaluated regimen. Additional study toxicity endpoints were mice found moribund or displayed >20% net weight loss for a period lasting 7 days or if the mice displayed >30% net weight loss. Tumor growth was monitored twice a week using digital calipers and the tumor volume (TV) was calculated using the formula (0.52 x [length x width2]). When the average tumor reached 72-73 mm3, animals were matched by tumor size, cord blood donor, and humanization rate and assigned into control or treatment groups (n=l l / group for the efficacy study). Tumor size and body weight were measured twice / thrice weekly. In the efficacy study the mice were treated with isotype control human IgGl, anti-PVRIG antibody, combination of anti-PD-Ll and anti-PVRIG antibodies with hlgGl isotype, anti-PD-Ll antibody Tecentriq and PVRIG x PD-L1 binding proteins multi- 246 and multi-256. Table 22 summarizes the used agents, their doses and study groups. Tolerability was assessed by body weight loss, lethality, and clinical signs of adverse treatment-related side effects. None of the agents tested in this study induced treatment- related side effect.

[0429] Table 22 Summary of used agents and study groups used in the efficacy study Figure 29 shows individual tumor growth curves for the mice treated with the agents (table 22). The group that received a combination of anti-PVRIG and anti-PDLl antibodies, both in human IgGl isotype format, displayed significant tumour growth delay with 3 out of 11 having almost cleared their tumors at the end of the study.

[0430] Mice dosed with PVRIG x PD-L1 binding proteins displayed tumour growth delay. For mice dosed with multi-246, 4 out of 11 mice showed strong anti-tumour response, with 2 mice showing long tumour growth delay. Mice dosed with multi-256 also displayed comparable anti-tumour efficacy to multi-246, with 4 out of 11 mice showing strong antitumour response. As expected, treatment with anti-PVRIG antibody had marginal tumour growth delay when compared to mice dosed with PVRIG x PD-L1 binding proteins and combinations of anti-PVRIG and anti-PD-Ll. Figure 30 and Table 23 summarizes the average tumor growth is the different study groups dosed with the different agents.

[0431] This study highlighted that the PVRIG x PD-L1 binding proteins were able to induce better anti-tumour response when compared to mice dosed with anti-PVRIG or anti- PD-Ll antibodies alone or in combination. The data suggests that simultaneous targeting of anti-PVRIG and anti-PDLl promote a strong antitumour response, that is further improved when simultaneous targeting is done by a single molecule, i.e. a PVRIG x PD-L1 binding protein, compared to combination treatment by separate antibodies.

[0432] Table 23 Summary of Anti-Tumor Activities in humanized mice engrafted with MDA-MB- 231.

[0433] PVRIG x PD-L1 binding proteins promote immune activation without promoting strong T cell exhaustion in in vivo mechanistic study.

[0434] Similarly to the efficacy study, humanized-mice were implanted with 5 x 106MDA- MB-231 cells and few days later were matched by tumor size, cord blood donor and H-rate when average tumor volume reached 191-193 mm3. They were then treated thrice and euthanized 24 hours after the third dose. Table 24 summarizes the used agents, their doses and study groups. Blood, tumors, and spleens were collected at study completion. The resulting serum, and tumors and spleen samples (1 / 2 snap frozen in RNAlater and / i FFPE) were processed for ex vivo analysis. RNA was extracted from tumor and spleen tissues for Nanostring gene expression analysis. Table 24 Summary of used agents and study groups used in the mechanistic study.

[0435] Nanostring gene expression was performed from the tumors of mice dosed with the agents as shown in the table 24. Figure 31 summarizes gene expression of selected genes. The data suggest that PVRIG x PD-L1 binding proteins multi-246, multi-256 and multi-61 were able to induce genes associated with anti-tumor response, CXCL9, CXCL10, CXCL12, PRF1, and GZMB when compared to isotype control or anti-PD-Ll Tecentriq dosed mice. Although, compared to mice dosed with PVRIG x PD-L1 binding proteins, a combination of anti-PVRIG and anti-PDLl antibodies displayed increased expression of CXCL9, CXCL10, CXCL12, PRF1, and GZMB, the combination promoted T cell exhaustion more than the PVRIG x PD-L1 binding proteins as evidenced by an increase in TIGIT, LAG3, and CTLA4 genes in the combination group. This may partially explain the inferior anti-tumor efficacy in the combination group compared to the group treated with PVRIG x PD-L1 binding proteins. This data suggests that compared to the anti-PD-Ll and anti-PVRIG combination group, the PVRIG x PD-L1 binding proteins were highly effective in promoting genes associated with anti-tumour response while not inducing the genes associated with T cells exhaustion especially, CTLA4 and TIGIT.

[0436] The PVRIG x PD-L1 binding proteins provided for sufficient immune cell activation, and reduced T cell exhaustion when compared to mice dosed with anti-PD-Ll Tecentriq or a combination of anti-PVRIG and anti-PD-Ll antibodies. The lack of T cell exhaustion and sufficient immune activation may further explain for the improved antitumor response with PVRIG x PD-L1 binding proteins multi-246, multi -256 and multi-61 when compared to a combination of anti-PVRIG and anti-PD-Ll antibodies.

[0437] equence listing VRIG binder CDR sequences (1 / 2)

[0438] 110

[0439] VRIG binder CDR sequences (2 / 2)

[0440] Ill

[0441] D-L1 binder CDR sequences 1 / 2

[0442] 112

[0443] D-L1 binder CDR sequences (2 / 2)

[0444] 113

[0445] Variable region sequences ofPVRIG binders

[0446] PVRIG 1 variable heavy chain SEQ ID NO: 121

[0447] EVQLLESGGGLVQPGGSLRLSC AASGFTF S S YAMSWVRQAPGKGLEW VS YIS YS Y

[0448] GGTYYADSVKGRFTISRDNSKNTLYLQMNSLRAEDTAVYYCARHVYGPGDGALD YWGQGTLVTVSS

[0449] PVRIG 1 variable light chain SEQ ID NO: 131

[0450] DIQMTQSPSSLSASVGDRVTITCRASQGIRDYLNWYQQKPGKAPKLLIYDASSLQS

[0451] GVPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQHHYSSPYTFGQGTKLEIK

[0452] PVRIG 2 variable heavy chain SEQ ID NO: 122

[0453] EVQLLESGGGLVQPGGSLRLSC AASGFTF S S YAMSWVRQAPGKGLEW VSGISYGY

[0454] SSTYYADSVKGRFTISRDNSKNTLYLQMNSLRAEDTAVYYCARSANSGPGYSSYI D YWGQGTLVTVSS

[0455] PVRIG 2 variable light chain SEQ ID NO: 132

[0456] DIQMTQSPSSLSASVGDRVTITCRASQSISSYLNWYQQKPGKAPKLLIYAASSLQSG

[0457] VPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQGYYFYPYTFGQGTKLEIK

[0458] PVRIG 3 variable heavy chain SEQ ID NO: 123

[0459] Q VQLVQSGAEVKKPGS S VKVSCKASGGTF S SDSMWVRQAPGQGLEWMGGIIP YF

[0460] DTADYAQKFQGRVTITADESTSTAYMELSSLRSEDTAVYYCARYPSYGYIDYWGQ GTLVTVSS

[0461] PVRIG 3 variable light chain SEQ ID NO: 133

[0462] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQQKPGQAPRLLIYGASSRAT

[0463] GIPDRFSGSGSGTDFTLTISRLEPEDFAVYYCQQNTWTLLTFGQGTKLEIK

[0464] PVRIG 4 variable heavy chain SEQ ID NO: 124

[0465] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSDYDIHWVRQAPGQGLEWMGGIDPY

[0466] FGGANYAQKFQGRVTITADESTSTAYMELSSLRSEDTAVYYCARSVWWASSIDY WGQGTLVTVSS

[0467] PVRIG 4 variable light chain SEQ ID NO: 134

[0468] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQQKPGQAPRLLIYGASSRAT

[0469] GIPDRFSGSGSGTDFTLTISRLEPEDFAVYYCQQGGYSLFTFGQGTKLEIK

[0470] PVRIG 5 variable heavy chain SEQ ID NO: 125

[0471] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSDGDISWVRQAPGQGLEWMGGIDPY

[0472] FGYANYAQKFQGRVTITADESTSTAYMELSSLRSEDTAVYYCARGYWHSPSLDY WGQGTLVTVSS

[0473] PVRIG 5 variable light chain SEQ ID NO: 135

[0474] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQQKPGQAPRLLIYGASSRAT

[0475] GIPDRFSGSGSGTDFTLTISRLEPEDFAVYYCQQYFTFLPTFGQGTKLEIK PVRIG 6 variable heavy chain SEQ ID NO: 126

[0476] QVQLVQSGAEVKKPGSSVKVSCKASGGTFYHSAISWVRQAPGQGLEWMGGIIPIF DDAHYAQKFQGRVTITADESTSTAYMELSSLRSEDTAVYYCARNDVSGSVFDYW GQGTLVTVSS

[0477] PVRIG 6 variable light chain SEQ ID NO: 136

[0478] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQQKPGQAPRLLIYGASSRAT

[0479] GIPDRFSGSGSGTDFTLTISRLEPEDFAVYYCQQDGYWNPYTFGQGTKLEIK

[0480] PVRIG 7 variable heavy chain SEQ ID NO: 127

[0481] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSHYDIDWVRQAPGQGLEWMGSIIPGF

[0482] HTANYAQKFQGRVTITADESTSTAYMELSSLRSEDTAVYYCARYSDYWSAHGLD YWGQGTLVTVSS

[0483] PVRIG 7 variable light chain SEQ ID NO: 137

[0484] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQQKPGQAPRLLIYGASSRAT

[0485] GIPDRFSGSGSGTDFTLTISRLEPEDFAVYYCQQGYTSLYTFGQGTKLEIK

[0486] PVRIG 8 variable heavy chain SEQ ID NO: 128

[0487] QVQLVQSGAEVKKPGSSVKVSCKASGGTFDHYDISWVRQAPGQGLEWMGSIDPY

[0488] FGYANYAQKFQGRVTITADESTSTAYMELSSLRSEDTAVYYCARGSWWAPVMDY WGQGTLVTVSS

[0489] PVRIG 8 variable light chain SEQ ID NO: 138

[0490] DIQMTQSPSSLSASVGDRVTITCRASQSISSYLNWYQQKPGKAPKLLIYAASSLQSG

[0491] VPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQGDHPYTFGQGTKLEIK

[0492] PVRIG 9 variable heavy chain SEQ ID NO: 129

[0493] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSYYDIHWVRQAPGQGLEWMGGIDPY

[0494] FGTADYAQKFQGRVTITADESTSTAYMELSSLRSEDTAVYYCARNYWWASSIDY WGQGTLVTVSS

[0495] PVRIG 9 variable light chain SEQ ID NO: 139

[0496] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQQKPGQAPRLLIYGASSRAT

[0497] GIPDRFSGSGSGTDFTLTISRLEPEDFAVYYCQQHYRPLTFGQGTKLEIK

[0498] PVRIG 10 variable heavy chain SEQ ID NO: 130

[0499] EVQLLESGGGLVQPGGSLRLSCAASGFTFGSHAMHWVRQAPGKGLEWVSAISGD

[0500] YGSTGYADSVKGRFTISRDNSKNTLYLQMNSLRAEDTAVYYCARYSHAWASPTL D YWGQGTLVTVSS

[0501] PVRIG 10 variable light chain SEQ ID NO: 140

[0502] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQQKPGQAPRLLIYGASSRAT

[0503] GIPDRFSGSGSGTDFTLTISRLEPEDFAVYYCQQAYRVPLFTFGQGTKLEIK

[0504] Variable region sequences ofPD-Ll binders PD-Ll l variable heavy chain SEQ ID NO: 141

[0505] EVQLLESGGGLVQPGGSLRLSCAASGFTFYSYSMYWVRQAPGKGLEWVSSISGYG

[0506] GYTYYADSVKGRFTISRDNSKNTLYLQMNSLRAEDTAVYYCARYSNHYFSFDYW GQGTLVTVSS

[0507] PD-Ll l variable light chain SEQ ID NO: 151

[0508] DIQMTQSPSSLSASVGDRVTITCRASQSISSYLNWYQQKPGKAPKLLIYAASSLQSG

[0509] VPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQPFSPPTFGQGTKLEIK

[0510] PD-L1 2 variable heavy chain SEQ ID NO: 142

[0511] QVQLVQSGAEVKKPGSSVKVSCKASGGTFHYYDISWVRQAPGQGLEWMGSIIGG

[0512] FGTAGYAQKFQGRVTITADESTSTAYMELSSLRSEDTAVYYCARSTHFRHWSRYT FD YWGQGTL VT VS S

[0513] PD-L1 2 variable light chain SEQ ID NO: 152

[0514] DIQMTQSPSSLSASVGDRVTITCRASQSISSYLNWYQQKPGKAPKLLIYAASSLQSG

[0515] VPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQSHSLHTFGQGTKLEIK

[0516] PD-L1 3 variable heavy chain SEQ ID NO: 143

[0517] EVQLLESGGGLVQPGGSLRLSC AASGFTF S S YYMYWVRQAPGKGLEWVSGIS S SY

[0518] SYTGYADSVKGRFTISRDNSKNTLYLQMNSLRAEDTAVYYCARGGAHYIDYWGQ GTLVTVSS

[0519] PD-L1 3 variable light chain SEQ ID NO: 153

[0520] DIQMTQSPSSLSASVGDRVTITCRASQDIRNYLNWYQQKPGKAPKLLIYAASSLQS

[0521] GVPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQHYWFYLPTFGQGTKLEIK

[0522] PD-L1 4 variable heavy chain SEQ ID NO: 144

[0523] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSHGAIHWVRQAPGQGLEWMGGIIHIY

[0524] GTAHYAQKFQGRVTITADESTSTAYMELSSLRSEDTAVYYCARGSAWFTDLDYW GQGTLVTVSS

[0525] PD-L1 4 variable light chain SEQ ID NO: 154

[0526] DIQMTQSPSSLSASVGDRVTITCRASQSISSYLNWYQQKPGKAPKLLIYAASSLQSG

[0527] VPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQGWYLHTFGQGTKLEIK

[0528] PD-L1 5 variable heavy chain SEQ ID NO: 145

[0529] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSGSAIHWVRQAPGQGLEWMGGIIHY

[0530] FGTAYYAQKFQGRVTITADESTSTAYMELSSLRSEDTAVYYCARSGAWAAWIDY WGQGTLVTVSS

[0531] PD-L1 5 variable light chain SEQ ID NO: 155

[0532] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQQKPGQAPRLLIYGASSRAT

[0533] GIPDRFSGSGSGTDFTLTISRLEPEDFAVYYCQQDWAGPPTFGQGTKLEIK PD-L1 6 variable heavy chain SEQ ID NO: 146

[0534] EVQLLESGGGLVQPGGSLRLSCAASGFTFSYDSMSWVRQAPGKGLEWVSAISGDG GSTSYADSVKGRFTISRDNSKNTLYLQMNSLRAEDTAVYYCARSYRFYMDYWGQ GTLVTVSS

[0535] PD-L1 6 variable light chain SEQ ID NO: 156

[0536] DIQMTQSPSSLSASVGDRVTITCRASQSISSYLNWYQQKPGKAPKLLIYAASSLQSG

[0537] VPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQPFHPLTFGQGTKLEIK

[0538] PD-L1 7 variable heavy chain SEQ ID NO: 147

[0539] EVQLLESGGGLVQPGGSLRLSCAASGFTFDDYAMGWVRQAPGKGLEWVSAISDY

[0540] GGGTYYADSVKGRFTISRDNSKNTLYLQMNSLRAEDTAVYYCARRFWLDYWGQ GTLVTVSS

[0541] PD-L1 7 variable light chain SEQ ID NO: 157

[0542] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQQKPGQAPRLLIYGASSRAT

[0543] GIPDRFSGSGSGTDFTLTISRLEPEDFAVYYCQQDWHHPLTFGQGTKLEIK

[0544] PD-L1 8 variable heavy chain SEQ ID NO: 148

[0545] EVQLLESGGGLVQPGGSLRLSC AASGFTF S SHDMGWVRQAPGKGLEWVSHIS SGG

[0546] GSTYYADSVKGRFTISRDNSKNTLYLQMNSLRAEDTAVYYCARPFWFFDAHFDY WGQGTLVTVSS

[0547] PD-L1 8 variable light chain SEQ ID NO: 158

[0548] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQQKPGQAPRLLIYGASSRAT

[0549] GIPDRFSGSGSGTDFTLTISRLEPEDFAVYYCQQTWSPFTFGQGTKLEIK

[0550] PD-L1 9 variable heavy chain SEQ ID NO: 149

[0551] EVQLLESGGGLVQPGGSLRLSCAASGFTFGSGYMSWVRQAPGKGLEWVSGISGSY

[0552] GSTDYADSVKGRFTISRDNSKNTLYLQMNSLRAEDTAVYYCARSGWRGGWLDY WGQGTLVTVSS

[0553] PD-L1 9 variable light chain SEQ ID NO: 159

[0554] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQQKPGQAPRLLIYGASSRAT

[0555] GIPDRFSGSGSGTDFTLTISRLEPEDFAVYYCQQSHYWNPFTFGQGTKLEIK

[0556] PD-L1 10 variable heavy chain SEQ ID NO: 150

[0557] EVQLLESGGGLVQPGGSLRLSCAASGFTFSHYDMGWVRQAPGKGLEWVSGISGY

[0558] GGSTYYADSVKGRFTISRDNSKNTLYLQMNSLRAEDTAVYYCARGSYLDYWGQG TLVTVSS

[0559] PD-L1 10 variable light chain SEQ ID NO: 160

[0560] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQQKPGQAPRLLIYGASSRAT

[0561] GIPDRFSGSGSGTDFTLTISRLEPEDFAVYYCQQDSYFYPYTFGQGTKLEIK

[0562] Polypeptide chains of selected bispecific anti-PVRIGxPD-Ll RUBY antibodies Hl = VH (mab) -CH 1-CH2-Ch3 -linker- VL (Fab) -Ckappa (= ’fusion chain’)

[0563] LI = ’mAb position’ VL-Ckappa

[0564] H2 = ’Fab position’ VH-CH1

[0565] Multi-216

[0566] Hl (SEQ ID NO: 161)

[0567] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSHGAIHWVRRAPGQGLEWMGGIIHIYGTAHYAQKFQ GRVTITADESTSTAYMELSSLRSEDTAVYYCARGSAWFTDLDYWGQGTLVTVSSASTKGPSVFPLAPS SKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVATGPAVLQSSGLYSLSSVVTVPSSSLGTQ.TYIC NVNHKPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHE

[0568] DPEVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTI SKAKGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSF FLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGKGGGGSGGGGSGGGGSEIVLTQSP GTLSLSPGERATLSCRASQSVSSSYLAWYQRKPGQAPRLLIYGASSRATGIPDRFSGSGSGTDFTLTISRL

[0569] EPEDFAVYYCQQNTWTLLTFGQGTKLEIKRPVAAPAVFIFPPSDEQLKSGTASVVCLLKNFYPREAKVQ WKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFN RGEC

[0570] LI (SEQ ID NO: 162)

[0571] DIQMTQSPSSLSASVGDRVTITCRASQSISSYLNWYQEKPGKAPKLLIYAASSLQSGVPSRFSGSGSGTD FTLTISSLQPEDFATYYCQQGWYLHTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSGTASVVCYLNNFYP REAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLWSTLTLSKADYEKHKVYACEVTHQGLSSPVTKS FNRGEC

[0572] H2 (SEQ ID NO: 163)

[0573] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSSDSIHWVREAPGQGLEWMGGIIPYFDTADYAQKFQG RVTITADESTSTAYMELSSLRSEDTAVYYCARYPSYGYIDYWGQGTLVTVSSASTKGPSVFPLAPSSKST SGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLTSVVEVPSSSLGTQ.TYICNVNH KPSNTKVDKKVEPKSC

[0574] Multi-246

[0575] Hl (SEQ ID NO: 164)

[0576] EVQLLESGGGLVQPGGSLRLSCAASGFTFDDYAMGWVRRAPGKGLEWVSAISDYGGGTYYADSVKG RFTISRDNSKNTLYLQMNSLRAEDTAVYYCARRFWLDYWGQGTLVTVSSASTKGPSVFPLAPSSKSTS GGTAALGCLVKDYFPEPVTVSWNSGALTSGVATGPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVNH KPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPEV

[0577] KFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISKAK GQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSK LTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGKGGGGSGGGGSGGGGSEIVLTQSPGTLSL SPGERATLSCRASQSVSSSYLAWYQRKPGQAPRLLIYGASSRATGIPDRFSGSGSGTDFTLTISRLEPED

[0578] FAVYYCQQNTWTLLTFGQGTKLEIKRPVAAPAVFIFPPSDEQLKSGTASVVCLLKNFYPREAKVQWKV DNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNRGEC

[0579] LI (SEQ ID NO: 165)

[0580] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQEKPGQAPRLLIYGASSRATGIPDRFSGSGSGT DFTLTISRLEPEDFAVYYCQQDWHHPLTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSGTASVVCYLNNF YPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLWSTLTLSKADYEKHKVYACEVTHQGLSSPVT KSFNRGEC

[0581] H2 (SEQ ID NO: 166)

[0582] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSSDSIHWVREAPGQGLEWMGGIIPYFDTADYAQKFQG

[0583] RVTITADESTSTAYMELSSLRSEDTAVYYCARYPSYGYIDYWGQGTLVTVSSASTKGPSVFPLAPSSKST

[0584] SGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLTSVVEVPSSSLGTQ.TYICNVNH KPSNTKVDKKVEPKSC

[0585] Multi-256

[0586] Hl (SEQ ID NO: 167)

[0587] EVQLLESGGGLVQPGGSLRLSCAASGFTFSSHDMGWVRRAPGKGLEWVSHISSGGGSTYYADSVKG

[0588] RFTISRDNSKNTLYLQMNSLRAEDTAVYYCARPFWFFDAHFDYWGQGTLVTVSSASTKGPSVFPLAPS

[0589] SKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVATGPAVLQSSGLYSLSSVVTVPSSSLGTQ.TYIC

[0590] NVNHKPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHE

[0591] DPEVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTI

[0592] SKAKGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSF

[0593] FLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGKGGGGSGGGGSGGGGSEIVLTQSP

[0594] GTLSLSPGERATLSCRASQSVSSSYLAWYQRKPGQAPRLLIYGASSRATGIPDRFSGSGSGTDFTLTISRL

[0595] EPEDFAVYYCQQNTWTLLTFGQGTKLEIKRPVAAPAVFIFPPSDEQLKSGTASVVCLLKNFYPREAKVQ WKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFN RGEC

[0596] LI (SEQ ID NO: 168)

[0597] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQEKPGQAPRLLIYGASSRATGIPDRFSGSGSGT

[0598] DFTLTISRLEPEDFAVYYCQQTWSPFTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSGTASVVCYLNNFY

[0599] PREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLWSTLTLSKADYEKHKVYACEVTHQGLSSPVTK SFNRGEC

[0600] H2 (SEQ ID NO: 169)

[0601] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSSDSIHWVREAPGQGLEWMGGIIPYFDTADYAQKFQG

[0602] RVTITADESTSTAYMELSSLRSEDTAVYYCARYPSYGYIDYWGQGTLVTVSSASTKGPSVFPLAPSSKST

[0603] SGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLTSVVEVPSSSLGTQ.TYICNVNH KPSNTKVDKKVEPKSC

[0604] Amino acid sequences of reference anti-PVRIG and anti-PDL-1 antibodies

[0605] Anti-human PVRIG antibody (with human IgGl constant domains)

[0606] Light chain (SEQ ID NO: 177):

[0607] DIQMTOSPSSLSASVGDRVTITCRASOSISSYLNWYOQKPGKAPKLLIYAASSLOSGVPS

[0608] RFSGSGSGTDFTLTISSLQPEDFATYYCOOSDILYTFGGGTKFEIKRTVAAPSVF FPPS

[0609] DEQLKSGTASVVCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSL

[0610] SSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNRGEC

[0611] Italics: VL sequence (SEQ ID NO: 170). Underlining = CDR sequences.

[0612] Heavy chain (SEQ ID NO: 178): OVOLVOSGAEVKKPGSSVKVSCKASGGTFSSAAISWVROAPGOGLEWMGNIIPIVGIA NYAQKFOGRVTITADESTSTAYMELSSLRSEDTAVYYCARDTGRGYTRHFWFDPWGQG 7 / .F7INS'ASTI<GPSVFPLAPSSI<STSGGTAALGCLVI<DYFPEPVTVSWNSGALTSGV HTFPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVNHKPSNTKVDKKVEPKSCDKT HTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPEVKFNWYV DGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIE KTISKAKGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPEN NYKTTPPVLDSDGSFFLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLS PGK

[0613] Italics: VH sequence (SEQ ID NO: 171). Underlining = CDR sequences.

[0614] Anti-PD-Ll antibody with human kappa light chain & hlgGl constant domains

[0615] Light chain (SEQ ID NO: 179):

[0616] DIQMTQSPSSLSASVGDRVTITCRASODVSTA VA WYOQKPGKAPKLLIYSASFLYSGVP SRFSGSGSGTDFTLTISSLQPEDFATYYCOOYLYHPATFGQGTKVEIKRTVAAPSVFIFP PSDEQLKSGTASWCLLNNFYPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLSST LTLSKADYEKHKVYACEVTHQGLSSPVTKSFNRGEC

[0617] Italics: VL sequence (SEQ ID NO: 172). Underlining = CDR sequences.

[0618] Heavy chain (SEQ ID NO: 180):

[0619] EVOLVESGGGLVQPGGSLRLSCAASGFTFSDSWIHWVROAPGKGLEWVAWISPYGGS TYYADSVKGRFTISADTSKNTAYLQMNSLRAEDTAVYYCARRHWPGGFDYWGQGTLV 7FN4ASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHT FP AVLQS SGLYSLS S VVTVPS S SLGTQT YICNVNHKPSNTKVDKKVEPKSCDKTHT CPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPEVKFNWYVDG VEVHNAI<TI<PREEQYNSTYRVVSVLTVLHQDWLNGI<EYI<CI<VSNI<ALPAPIEI<TI SKAKGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNY KTTPPVLDSDGSFFLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPG K

[0620] Italics: VH sequence (SEQ ID NO: 173). Underlining = CDR sequences.

[0621] Multispecific antigen-binding protein for PVRIG x PD-L1 based on reference antibody sequences

[0622] Hl = VH (mab) -CH 1-CH2-Ch3 -linker- VL (Fab) -Ckappa (= ’fusion chain’)

[0623] LI = ’mAb position’ VL-Ckappa

[0624] H2 = ’Fab position’ VH-CH1

[0625] Amino acid sequences of further bispecific anti-PVRIGxPD-Ll RUBY antibodies

[0626] Multi-61

[0627] Hl (SEQ ID NO: 174)

[0628] EVQLVESGGGLVQPGGSLRLSCAASGFTFSDSWIHWVRRAPGKGLEWVAWISPY GGSTYYADSVKGRFTISADTSKNTAYLQMNSLRAEDTAVYYCARRHWPGGFDY WGQGTLVTVSSASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWNSG ALTSGVATGPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVNHKPSNTKVDKKVEP KSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPEVK FNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKA LPAPIEI<TISI<AI<GQPREPQVYTLPPSRDELTI<NQVSLTCLVI<GFYPSDIAVEWESN

[0629] GQPENNYKTTPPVLDSDGSFFLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQ KSLSLSPGKGGGGSGGGGSGGGGSDIQMTQSPSSLSASVGDRVTITCRASQSISSYL NWYQRKPGKAPKLLIYAASSLQSGVPSRFSGSGSGTDFTLTISSLQPEDFATYYCQ QSDILYTFGGGTKVEIKRPVAAPAVFIFPPSDEQLKSGTASVVCLLKNFYPREAKVQ WKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGL

[0630] SSPVTKSFNRGEC

[0631] LI (SEQ ID NO: 175)

[0632] DIQMTQSPSSLSASVGDRVTITCRASQDVSTAVAWYQEKPGKAPKLLIYSASFLYS GVPSRFSGSGSGTDFTLTISSLQPEDFATYYCQQYLYHPATFGQGTKVEIKRTVAAP SVFIFPPSDEQLKSGTASVVCYLNNFYPREAKVQWKVDNALQSGNSQESVTEQDS KDSTYSLWSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNRGEC

[0633] H2 (SEQ ID NO: 176)

[0634] Q VQLVQSGAEVKKPGS S VKVSCKASGGTF S S AAISWVREAPGQGLEWMGNIIPIV GIANYAQKFQGRVTITADESTSTAYMELSSLRSEDTAVYYCARDTGRGYTRHFWF DPWGQGTLVTVSSASTKGPSVFPLAPSSKSTSGGTAALGCLVKDYFPEPVTVSWN SGALTSGVHTFPAVLQSSGLYSLTSVVEVPSSSLGTQTYICNVNHKPSNTKVDKKV EPKSC

[0635] Multi-108

[0636] Hl (SEQ ID NO: 181)

[0637] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSSDSIHWVRRAPGQGLEWMGGIIPYFDTADYAQKFQG RVTITADESTSTAYMELSSLRSEDTAVYYCARYPSYGYIDYWGQGTLVTVSSASTKGPSVFPLAPSSKST SGGTAALGCLVKDYFPEPVTVSWNSGALTSGVATGPAVLQSSGLYSLSSVVTVPSSSLGTQ.TYICNVN HKPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPE VKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISKA

[0638] KGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYS KLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGKGGGGSGGGGSGGGGSEIVLTQSPGTLS LSPGERATLSCRASQSVSSSYLAWYQRKPGQAPRLLIYGASSRATGIPDRFSGSGSGTDFTLTISRLEPED FAVYYCQQDWAGPPTFGQGTKLEIKRPVAAPAVFIFPPSDEQLKSGTASVVCLLKNFYPREAKVQWK VDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNRGEC

[0639] LI (SEQ ID NO: 182)

[0640] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQEKPGQAPRLLIYGASSRATGIPDRFSGSGSGT DFTLTISRLEPEDFAVYYCQQNTWTLLTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSGTASVVCYLNNF YPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLWSTLTLSKADYEKHKVYACEVTHQGLSSPVT KSFNRGEC H2 (SEQ ID NO: 183)

[0641] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSGSAIHWVREAPGQGLEWMGGIIHYFGTAYYAQKFQG

[0642] RVTITADESTSTAYMELSSLRSEDTAVYYCARSGAWAAWIDYWGQGTLVTVSSASTKGPSVFPLAPSS

[0643] KSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLTSVVEVPSSSLGTQ.TYICN

[0644] VNHKPSNTKVDKKVEPKSC

[0645] Multi-121

[0646] Hl (SEQ ID NO: 184)

[0647] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSDYDIHWVRRAPGQGLEWMGGIDPYFGGANYAQKFQ

[0648] GRVTITADESTSTAYMELSSLRSEDTAVYYCARSVWWASSIDYWGQGTLVTVSSASTKGPSVFPLAPS

[0649] SKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVATGPAVLQSSGLYSLSSVVTVPSSSLGTQ.TYIC

[0650] NVNHKPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHE

[0651] DPEVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTI

[0652] SKAKGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSF

[0653] FLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGKGGGGSGGGGSGGGGSEIVLTQSP

[0654] GTLSLSPGERATLSCRASQSVSSSYLAWYQRKPGQAPRLLIYGASSRATGIPDRFSGSGSGTDFTLTISRL

[0655] EPEDFAVYYCQQTWSPFTFGQGTKLEIKRPVAAPAVFIFPPSDEQLKSGTASVVCLLKNFYPREAKVQ

[0656] WKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFN RGEC

[0657] LI (SEQ ID NO: 185)

[0658] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQEKPGQAPRLLIYGASSRATGIPDRFSGSGSGT

[0659] DFTLTISRLEPEDFAVYYCQQGGYSLFTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSGTASVVCYLNNFY

[0660] PREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLWSTLTLSKADYEKHKVYACEVTHQGLSSPVTK SFNRGEC

[0661] H2 (SEQ ID NO: 186)

[0662] EVQLLESGGGLVQPGGSLRLSCAASGFTFSSHDMGWVREAPGKGLEWVSHISSGGGSTYYADSVKG

[0663] RFTISRDNSKNTLYLQMNSLRAEDTAVYYCARPFWFFDAHFDYWGQGTLVTVSSASTKGPSVFPLAPS

[0664] SKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLTSVVEVPSSSLGTQTYIC

[0665] NVNHKPSNTKVDKKVEPKSC

[0666] Multi-215

[0667] Hl (SEQ ID NO: 187)

[0668] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSHGAIHWVRRAPGQGLEWMGGIIHIYGTAHYAQKFQ

[0669] GRVTITADESTSTAYMELSSLRSEDTAVYYCARGSAWFTDLDYWGQGTLVTVSSASTKGPSVFPLAPS

[0670] SKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVATGPAVLQSSGLYSLSSVVTVPSSSLGTQ.TYIC

[0671] NVNHKPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHE

[0672] DPEVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTI

[0673] SKAKGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSF

[0674] FLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGKGGGGSGGGGSGGGGSDIQMTQS

[0675] PSSLSASVGDRVTITCRASQSISSYLNWYQRKPGKAPKLLIYAASSLQSGVPSRFSGSGSGTDFTLTISSL

[0676] QPEDFATYYCQQGYYFYPYTFGQGTKLEIKRPVAAPAVFIFPPSDEQLKSGTASVVCLLKNFYPREAKV

[0677] QWKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNRGEC LI (SEQ ID NO: 188)

[0678] DIQMTQSPSSLSASVGDRVTITCRASQSISSYLNWYQEKPGKAPKLLIYAASSLQSGVPSRFSGSGSGTD FTLTISSLQPEDFATYYCQQGWYLHTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSGTASVVCYLNNFYP REAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLWSTLTLSKADYEKHKVYACEVTHQGLSSPVTKS FNRGEC

[0679] H2 (SEQ ID NO: 189)

[0680] EVQLLESGGGLVQPGGSLRLSCAASGFTFSSYAMSWVREAPGKGLEWVSGISYGYSSTYYADSVKGRF TISRDNSKNTLYLQMNSLRAEDTAVYYCARSANSGPGYSSYIDYWGQGTLVTVSSASTKGPSVFPLAPS SKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLTSVVEVPSSSLGTQ.TYIC NVNHKPSNTKVDKKVEPKSC

[0681] Multi-217

[0682] Hl (SEQ ID NO: 190)

[0683] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSHGAIHWVRRAPGQGLEWMGGIIHIYGTAHYAQKFQ GRVTITADESTSTAYMELSSLRSEDTAVYYCARGSAWFTDLDYWGQGTLVTVSSASTKGPSVFPLAPS SKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVATGPAVLQSSGLYSLSSVVTVPSSSLGTQ.TYIC NVNHKPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHE

[0684] DPEVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTI SKAKGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSF FLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGKGGGGSGGGGSGGGGSEIVLTQSP GTLSLSPGERATLSCRASQSVSSSYLAWYQRKPGQAPRLLIYGASSRATGIPDRFSGSGSGTDFTLTISRL

[0685] EPEDFAVYYCQQGGYSLFTFGQGTKLEIKRPVAAPAVFIFPPSDEQLKSGTASVVCLLKNFYPREAKVQ WKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFN RGEC

[0686] LI (SEQ ID NO: 191)

[0687] DIQMTQSPSSLSASVGDRVTITCRASQSISSYLNWYQEKPGKAPKLLIYAASSLQSGVPSRFSGSGSGTD FTLTISSLQPEDFATYYCQQGWYLHTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSGTASVVCYLNNFYP REAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLWSTLTLSKADYEKHKVYACEVTHQGLSSPVTKS FNRGEC

[0688] H2 (SEQ ID NO: 192)

[0689] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSDYDIHWVREAPGQGLEWMGGIDPYFGGANYAQKFQ GRVTITADESTSTAYMELSSLRSEDTAVYYCARSVWWASSIDYWGQGTLVTVSSASTKGPSVFPLAPS SKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLTSVVEVPSSSLGTQTYIC NVNHKPSNTKVDKKVEPKSC

[0690] Multi-224

[0691] Hl (SEQ ID NO: 193)

[0692] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSGSAIHWVRRAPGQGLEWMGGIIHYFGTAYYAQKFQG

[0693] RVTITADESTSTAYMELSSLRSEDTAVYYCARSGAWAAWIDYWGQGTLVTVSSASTKGPSVFPLAPSS KSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVATGPAVLQSSGLYSLSSVVTVPSSSLGTQTYICN

[0694] VNHKPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHED

[0695] PEVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISK

[0696] AKGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFL

[0697] YSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGKGGGGSGGGGSGGGGSDIQMTQSPS

[0698] SLSASVGDRVTITCRASQGIRDYLNWYQRKPGKAPKLLIYDASSLQSGVPSRFSGSGSGTDFTLTISSLQ

[0699] PEDFATYYCQQHHYSSPYTFGQGTKLEIKRPVAAPAVFIFPPSDEQLKSGTASVVCLLKNFYPREAKVQ

[0700] WKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFN RGEC

[0701] LI (SEQ ID NO: 194)

[0702] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQEKPGQAPRLLIYGASSRATGIPDRFSGSGSGT

[0703] DFTLTISRLEPEDFAVYYCQQDWAGPPTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSGTASVVCYLNN

[0704] FYPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLWSTLTLSKADYEKHKVYACEVTHQGLSSPV

[0705] TKSFNRGEC

[0706] H2 (SEQ ID NO: 195)

[0707] EVQLLESGGGLVQPGGSLRLSCAASGFTFSSYAMSWVREAPGKGLEWVSYISYSYGGTYYADSVKGRF

[0708] TISRDNSKNTLYLQMNSLRAEDTAVYYCARHVYGPGDGALDYWGQGTLVTVSSASTKGPSVFPLAPS

[0709] SKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLTSVVEVPSSSLGTQTYIC

[0710] NVNHKPSNTKVDKKVEPKSC

[0711] Multi-226

[0712] Hl (SEQ ID NO: 196)

[0713] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSGSAIHWVRRAPGQGLEWMGGIIHYFGTAYYAQKFQG

[0714] RVTITADESTSTAYMELSSLRSEDTAVYYCARSGAWAAWIDYWGQGTLVTVSSASTKGPSVFPLAPSS

[0715] KSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVATGPAVLQSSGLYSLSSVVTVPSSSLGTQTYICN

[0716] VNHKPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHED

[0717] PEVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISK

[0718] AKGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFL

[0719] YSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGKGGGGSGGGGSGGGGSEIVLTQSPGT

[0720] LSLSPGERATLSCRASQSVSSSYLAWYQRKPGQAPRLLIYGASSRATGIPDRFSGSGSGTDFTLTISRLEP

[0721] EDFAVYYCQQNTWTLLTFGQGTKLEIKRPVAAPAVFIFPPSDEQLKSGTASVVCLLKNFYPREAKVQW

[0722] KVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNRGEC

[0723] LI (SEQ ID NO: 197)

[0724] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQEKPGQAPRLLIYGASSRATGIPDRFSGSGSGT

[0725] DFTLTISRLEPEDFAVYYCQQDWAGPPTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSGTASVVCYLNN

[0726] FYPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLWSTLTLSKADYEKHKVYACEVTHQGLSSPV

[0727] TKSFNRGEC

[0728] H2 (SEQ ID NO: 198)

[0729] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSSDSIHWVREAPGQGLEWMGGIIPYFDTADYAQKFQG

[0730] RVTITADESTSTAYMELSSLRSEDTAVYYCARYPSYGYIDYWGQGTLVTVSSASTKGPSVFPLAPSSKST SGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLTSVVEVPSSSLGTQTYICNVNH KPSNTKVDKKVEPKSC

[0731] Multi-229

[0732] Hl (SEQ ID NO: 199)

[0733] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSGSAIHWVRRAPGQGLEWMGGIIHYFGTAYYAQKFQG

[0734] RVTITADESTSTAYMELSSLRSEDTAVYYCARSGAWAAWIDYWGQGTLVTVSSASTKGPSVFPLAPSS

[0735] KSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVATGPAVLQSSGLYSLSSVVTVPSSSLGTQTYICN

[0736] VNHKPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHED

[0737] PEVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISK

[0738] AKGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFL

[0739] YSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGKGGGGSGGGGSGGGGSEIVLTQSPGT

[0740] LSLSPGERATLSCRASQSVSSSYLAWYQRKPGQAPRLLIYGASSRATGIPDRFSGSGSGTDFTLTISRLEP EDFAVYYCQQDGYWNPYTFGQGTKLEIKRPVAAPAVFIFPPSDEQLKSGTASVVCLLKNFYPREAKVQ WKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFN RGEC

[0741] LI (SEQ ID NO: 200)

[0742] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQEKPGQAPRLLIYGASSRATGIPDRFSGSGSGT

[0743] DFTLTISRLEPEDFAVYYCQQDWAGPPTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSGTASVVCYLNN FYPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLWSTLTLSKADYEKHKVYACEVTHQGLSSPV TKSFNRGEC

[0744] H2 (SEQ ID NO: 201)

[0745] QVQLVQSGAEVKKPGSSVKVSCKASGGTFYHSAISWVREAPGQGLEWMGGIIPIFDDAHYAQKFQG

[0746] RVTITADESTSTAYMELSSLRSEDTAVYYCARNDVSGSVFDYWGQGTLVTVSSASTKGPSVFPLAPSSK

[0747] STSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLTSVVEVPSSSLGTQ.TYICNV NHKPSNTKVDKKVEPKSC

[0748] Multi-236

[0749] Hl (SEQ ID NO: 202)

[0750] EVQLLESGGGLVQPGGSLRLSCAASGFTFSYDSMSWVRRAPGKGLEWVSAISGDGGSTSYADSVKGR

[0751] FTISRDNSKNTLYLQMNSLRAEDTAVYYCARSYRFYMDYWGQGTLVTVSSASTKGPSVFPLAPSSKST

[0752] SGGTAALGCLVKDYFPEPVTVSWNSGALTSGVATGPAVLQSSGLYSLSSVVTVPSSSLGTQ.TYICNVN

[0753] HKPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPE

[0754] VKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISKA KGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYS KLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGKGGGGSGGGGSGGGGSEIVLTQSPGTLS

[0755] LSPGERATLSCRASQSVSSSYLAWYQRKPGQAPRLLIYGASSRATGIPDRFSGSGSGTDFTLTISRLEPED FAVYYCQQNTWTLLTFGQGTKLEIKRPVAAPAVFIFPPSDEQLKSGTASVVCLLKNFYPREAKVQWKV DNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNRGEC LI (SEQ ID NO: 203)

[0756] DIQMTQSPSSLSASVGDRVTITCRASQSISSYLNWYQEKPGKAPKLLIYAASSLQSGVPSRFSGSGSGTD FTLTISSLQPEDFATYYCQQPFHPLTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSGTASVVCYLNNFYPR EAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLWSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSF NRGEC

[0757] H2 (SEQ ID NO: 204)

[0758] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSSDSIHWVREAPGQGLEWMGGIIPYFDTADYAQKFQG RVTITADESTSTAYMELSSLRSEDTAVYYCARYPSYGYIDYWGQGTLVTVSSASTKGPSVFPLAPSSKST SGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLTSVVEVPSSSLGTQ.TYICNVNH KPSNTKVDKKVEPKSC

[0759] Multi-239

[0760] Hl (SEQ ID NO: 205)

[0761] EVQLLESGGGLVQPGGSLRLSCAASGFTFSYDSMSWVRRAPGKGLEWVSAISGDGGSTSYADSVKGR FTISRDNSKNTLYLQMNSLRAEDTAVYYCARSYRFYMDYWGQGTLVTVSSASTKGPSVFPLAPSSKST SGGTAALGCLVKDYFPEPVTVSWNSGALTSGVATGPAVLQSSGLYSLSSVVTVPSSSLGTQ.TYICNVN HKPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPE

[0762] VKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISKA KGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYS KLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGKGGGGSGGGGSGGGGSEIVLTQSPGTLS LSPGERATLSCRASQSVSSSYLAWYQRKPGQAPRLLIYGASSRATGIPDRFSGSGSGTDFTLTISRLEPED

[0763] FAVYYCQQDGYWNPYTFGQGTKLEIKRPVAAPAVFIFPPSDEQLKSGTASVVCLLKNFYPREAKVQW KVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNRGEC

[0764] LI (SEQ ID NO: 206)

[0765] DIQMTQSPSSLSASVGDRVTITCRASQSISSYLNWYQEKPGKAPKLLIYAASSLQSGVPSRFSGSGSGTD FTLTISSLQPEDFATYYCQQPFHPLTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSGTASVVCYLNNFYPR EAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLWSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSF NRGEC

[0766] H2 (SEQ ID NO: 207)

[0767] QVQLVQSGAEVKKPGSSVKVSCKASGGTFYHSAISWVREAPGQGLEWMGGIIPIFDDAHYAQKFQG RVTITADESTSTAYMELSSLRSEDTAVYYCARNDVSGSVFDYWGQGTLVTVSSASTKGPSVFPLAPSSK STSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLTSVVEVPSSSLGTQ.TYICNV NHKPSNTKVDKKVEPKSC

[0768] Multi-244

[0769] Hl (SEQ ID NO: 208)

[0770] EVQLLESGGGLVQPGGSLRLSCAASGFTFDDYAMGWVRRAPGKGLEWVSAISDYGGGTYYADSVKG RFTISRDNSKNTLYLQMNSLRAEDTAVYYCARRFWLDYWGQGTLVTVSSASTKGPSVFPLAPSSKSTS GGTAALGCLVKDYFPEPVTVSWNSGALTSGVATGPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVNH KPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPEV KFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISKAK

[0771] GQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSK

[0772] LTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGKGGGGSGGGGSGGGGSDIQMTQSPSSLS

[0773] ASVGDRVTITCRASQGIRDYLNWYQRKPGKAPKLLIYDASSLQSGVPSRFSGSGSGTDFTLTISSLQPED

[0774] FATYYCQQHHYSSPYTFGQGTKLEIKRPVAAPAVFIFPPSDEQLKSGTASVVCLLKNFYPREAKVQWKV

[0775] DNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNRGEC

[0776] LI (SEQ ID NO: 209)

[0777] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQEKPGQAPRLLIYGASSRATGIPDRFSGSGSGT

[0778] DFTLTISRLEPEDFAVYYCQQDWHHPLTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSGTASVVCYLNNF

[0779] YPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLWSTLTLSKADYEKHKVYACEVTHQGLSSPVT

[0780] KSFNRGEC

[0781] H2 (SEQ ID NO: 210)

[0782] EVQLLESGGGLVQPGGSLRLSCAASGFTFSSYAMSWVREAPGKGLEWVSYISYSYGGTYYADSVKGRF

[0783] TISRDNSKNTLYLQMNSLRAEDTAVYYCARHVYGPGDGALDYWGQGTLVTVSSASTKGPSVFPLAPS

[0784] SKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLTSVVEVPSSSLGTQ.TYIC

[0785] NVNHKPSNTKVDKKVEPKSC

[0786] Multi-247

[0787] Hl (SEQ ID NO: 211)

[0788] EVQLLESGGGLVQPGGSLRLSCAASGFTFDDYAMGWVRRAPGKGLEWVSAISDYGGGTYYADSVKG

[0789] RFTISRDNSKNTLYLQMNSLRAEDTAVYYCARRFWLDYWGQGTLVTVSSASTKGPSVFPLAPSSKSTS

[0790] GGTAALGCLVKDYFPEPVTVSWNSGALTSGVATGPAVLQSSGLYSLSSVVTVPSSSLGTQTYICNVNH

[0791] KPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHEDPEV

[0792] KFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISKAK

[0793] GQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSK

[0794] LTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGKGGGGSGGGGSGGGGSEIVLTQSPGTLSL

[0795] SPGERATLSCRASQSVSSSYLAWYQRKPGQAPRLLIYGASSRATGIPDRFSGSGSGTDFTLTISRLEPED

[0796] FAVYYCQQGGYSLFTFGQGTKLEIKRPVAAPAVFIFPPSDEQLKSGTASVVCLLKNFYPREAKVQWKV

[0797] DNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNRGEC

[0798] LI (SEQ ID NO: 212)

[0799] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQEKPGQAPRLLIYGASSRATGIPDRFSGSGSGT

[0800] DFTLTISRLEPEDFAVYYCQQDWHHPLTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSGTASVVCYLNNF

[0801] YPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLWSTLTLSKADYEKHKVYACEVTHQGLSSPVT

[0802] KSFNRGEC

[0803] H2 (SEQ ID NO: 213)

[0804] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSDYDIHWVREAPGQGLEWMGGIDPYFGGANYAQKFQ

[0805] GRVTITADESTSTAYMELSSLRSEDTAVYYCARSVWWASSIDYWGQGTLVTVSSASTKGPSVFPLAPS

[0806] SKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLTSVVEVPSSSLGTQTYIC

[0807] NVNHKPSNTKVDKKVEPKSC Multi-257

[0808] Hl (SEQ ID NO: 214)

[0809] EVQLLESGGGLVQPGGSLRLSCAASGFTFSSHDMGWVRRAPGKGLEWVSHISSGGGSTYYADSVKG

[0810] RFTISRDNSKNTLYLQMNSLRAEDTAVYYCARPFWFFDAHFDYWGQGTLVTVSSASTKGPSVFPLAPS

[0811] SKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVATGPAVLQSSGLYSLSSVVTVPSSSLGTQ.TYIC

[0812] NVNHKPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHE

[0813] DPEVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTI

[0814] SKAKGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSF

[0815] FLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGKGGGGSGGGGSGGGGSEIVLTQSP

[0816] GTLSLSPGERATLSCRASQSVSSSYLAWYQRKPGQAPRLLIYGASSRATGIPDRFSGSGSGTDFTLTISRL

[0817] EPEDFAVYYCQQGGYSLFTFGQGTKLEIKRPVAAPAVFIFPPSDEQLKSGTASVVCLLKNFYPREAKVQ

[0818] WKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFN RGEC

[0819] LI (SEQ ID NO: 215)

[0820] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQEKPGQAPRLLIYGASSRATGIPDRFSGSGSGT

[0821] DFTLTISRLEPEDFAVYYCQQTWSPFTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSGTASVVCYLNNFY

[0822] PREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLWSTLTLSKADYEKHKVYACEVTHQGLSSPVTK

[0823] SFNRGEC

[0824] H2 (SEQ ID NO: 216)

[0825] QVQLVQSGAEVKKPGSSVKVSCKASGGTFSDYDIHWVREAPGQGLEWMGGIDPYFGGANYAQKFQ

[0826] GRVTITADESTSTAYMELSSLRSEDTAVYYCARSVWWASSIDYWGQGTLVTVSSASTKGPSVFPLAPS

[0827] SKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLTSVVEVPSSSLGTQTYIC

[0828] NVNHKPSNTKVDKKVEPKSC

[0829] Multi-259

[0830] Hl (SEQ ID NO: 217)

[0831] EVQLLESGGGLVQPGGSLRLSCAASGFTFSSHDMGWVRRAPGKGLEWVSHISSGGGSTYYADSVKG

[0832] RFTISRDNSKNTLYLQMNSLRAEDTAVYYCARPFWFFDAHFDYWGQGTLVTVSSASTKGPSVFPLAPS

[0833] SKSTSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVATGPAVLQSSGLYSLSSVVTVPSSSLGTQ.TYIC

[0834] NVNHKPSNTKVDKKVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTLMISRTPEVTCVVVDVSHE

[0835] DPEVKFNWYVDGVEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTI

[0836] SKAKGQPREPQVYTLPPSRDELTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSF

[0837] FLYSKLTVDKSRWQQGNVFSCSVMHEALHNHYTQKSLSLSPGKGGGGSGGGGSGGGGSEIVLTQSP

[0838] GTLSLSPGERATLSCRASQSVSSSYLAWYQRKPGQAPRLLIYGASSRATGIPDRFSGSGSGTDFTLTISRL

[0839] EPEDFAVYYCQQDGYWNPYTFGQGTKLEIKRPVAAPAVFIFPPSDEQLKSGTASVVCLLKNFYPREAK

[0840] VQWKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQGLSSPVTKSFNRGE

[0841] C

[0842] LI (SEQ ID NO: 218)

[0843] EIVLTQSPGTLSLSPGERATLSCRASQSVSSSYLAWYQEKPGQAPRLLIYGASSRATGIPDRFSGSGSGT

[0844] DFTLTISRLEPEDFAVYYCQQ.TWSPFTFGQGTKLEIKRTVAAPSVFIFPPSDEQLKSGTASVVCYLNNFY PREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLWSTLTLSKADYEKHKVYACEVTHQGLSSPVTK SFN RGEC

[0845] H2 (SEQ ID N0: 219)

[0846] QVQLVQSGAEVKKPGSSVKVSCKASGGTFYHSAISWVREAPGQGLEWMGGII PIFDDAHYAQKFQG RVTITADESTSTAYMELSSLRSEDTAVYYCARNDVSGSVFDYWGQGTLVTVSSASTKGPSVFPLAPSSK STSGGTAALGCLVKDYFPEPVTVSWNSGALTSGVHTFPAVLQSSGLYSLTSVVEVPSSSLGTQ.TYICNV NHKPSNTKVDKKVEPKSC

[0847] Nucleic acid sequences for selected bispecific anti-PVRIGxPD-Ll RUBY antibodies

[0848] Multi-216

[0849] Hl (SEQ ID NO: 220)

[0850] CAGGTTCAGCTGGTTCAGAGCGGTGCAGAAGTTAAAAAACCGGGTAGCAGCGTTAAAGTTAGCT GTAAAGCAAGCGGTGGCACCTTTAGCCACGGAGCAATTCACTGGGTTCGTAGAGCACCTGGTCA AGGTCTGGAATGGATGGGTGGTATTATTCATATTTATGGCACCGCACACTATGCCCAGAAATTTC AGGGTCGTGTTACCATTACCGCAGATGAAAGCACCAGCACCGCATATATGGAACTGAGCAGCCT GCGTAGCGAAGATACCGCAGTGTATTATTGTGCACGTGGTAGTGCTTGGTTTACGGATCTGGACT ACTGGGGCCAGGGAACCCTGGTCACCGTCTCCTCAGCCTCTACAAAGGGCCCCTCCGTTTTTCCAC TGG CTCCTAG CAG CAAG AG CACAAG CG G AG G AAC AG CCGCTCTG GG CTGTCTG GTCAAG G ATTA CTTTCCCGAGCCTGTGACCGTGTCCTGGAATTCTGGTGCTCTGACCAGCGGAGTGGCTACAGGAC CTG CTGTGCTG CAGTCTAG CG GCCTGTATTCTCTG AG CAG CGTG GTCAC AGTG CCTAG CTCTAG C CTGGGCACCCAGACCTACATCTGCAACGTGAACCACAAGCCTAGCAACACCAAGGTGGACAAGA AGGTGGAACCCAAGAGCTGCGACAAGACCCACACCTGTCCTCCATGTCCTGCTCCAGAACTGCTC GGCGGACCCTCCGTGTTTCTGTTCCCTCCAAAGCCTAAGGACACCCTGATGATCAGCAGAACCCCT GAAGTGACCTGCGTGGTGGTGGATGTGTCTCACGAGGACCCCGAAGTGAAGTTCAATTGGTACG TGGACGGCGTGGAAGTGCACAACGCCAAGACCAAGCCTAGAGAGGAACAGTACAACAGCACCT ACAGAGTGGTGTCCGTGCTGACCGTGCTGCACCAGGATTGGCTGAACGGCAAAGAGTACAAGTG CAAGGTGTCCAACAAGGCCCTGCCTGCTCCTATCGAGAAAACCATCAGCAAGGCCAAGGGCCAG CCAAGAGAACCCCAGGTTTACACACTGCCTCCAAGCAGGGACGAGCTGACCAAGAATCAGGTGT CCCTGACCTGCCTGGTTAAGGGCTTCTACCCTTCCGATATCGCCGTGGAATGGGAGAGCAATGGA CAGCCCGAGAACAACTACAAGACAACCCCTCCTGTGCTGGACAGCGACGGCTCATTCTTCCTGTA CAGCAAGCTGACAGTGGACAAGTCCAGATGGCAGCAGGGCAACGTGTTCAGCTGTTCTGTGATG CACGAGGCCCTGCACAACCACTACACCCAGAAGTCTCTGTCTCTGAGCCCTGGAAAAGGCGGAG GTGGAAGCGGAGGCGGAGGAAGTGGTGGCGGCGGATCTGAAATTGTTCTGACCCAGAGTCCGG GTACACTGAGCCTGTCACCGGGTGAACGTGCAACCCTGAGCTGTCGTGCAAGCCAGAGCGTTAG CAGCAGCTATCTGGCATGGTATCAGAGAAAACCTGGTCAGGCACCGCGTCTGCTGATTTATGGTG CAAGCAGCCGTGCAACCGGTATTCCGGATCGTTTTAGCGGTAGCGGTAGTGGCACCGATTTTACC CTGACCATTAGCCGTCTGGAACCGGAAGATTTTGCAGTGTATTATTGTCAGCAGAATACTTGGAC

[0851] CTTGCTGACTTTTGGCCAGGGGACCAAGCTGGAGATCAAACGACCCGTGGCTGCCCCTGCCGTGT TCATCTTCCCACCTTCCGACGAGCAGCTGAAGTCTGGCACAGCCTCTGTCGTGTGCCTGCTGAAGA ACTTCTACCCCAGAGAAGCCAAGGTGCAGTGGAAGGTGGACAACGCCCTGCAGAGCGGCAATAG CCAAGAGAGCGTGACCGAGCAGGACAGCAAGGACTCTACCTACAGCCTGAGCAGCACCCTGACA CTGAGCAAGGCCGACTACGAGAAGCACAAAGTGTACGCCTGCGAAGTGACCCACCAGGGCCTTT CTAG CCCTGTG ACCAAG AG CTTC AACCG GG GCG AATGT

[0852] LI (SEQ ID NO: 221)

[0853] GACATCCAGATGACCCAGTCTCCATCCTCCCTGTCTGCATCTGTAGGAGACAGAGTCACCATCACT TGCCGGGCAAGTCAGAGCATTAGCAGCTATTTAAATTGGTATCAGGAGAAACCAGGGAAAGCCC CTAAGCTCCTGATCTATGCTGCATCCAGTTTGCAAAGTGGGGTCCCATCAAGGTTCAGTGGCAGT GGATCTGGGACAGATTTCACTCTCACCATCAGCAGTCTGCAACCTGAAGATTTTGCAACTTACTAC TGTCAACAGGGCTGGTACCTGCATACTTTTGGCCAGGGGACCAAGCTGGAGATCAAAAGGACAG TGGCCGCTCCTAGCGTGTTCATCTTTCCACCTAGCGACGAGCAGCTGAAAAGCGGCACAGCCAGC GTCGTGTGTTACCTGAACAACTTCTACCCCAGAGAAGCCAAGGTGCAGTGGAAGGTGGACAATG CCCTGCAGAGCGGCAATAGCCAAGAGAGCGTGACCGAGCAGGACAGCAAGGACTCTACCTACA GCCTGTGGTCCACACTGACCCTGAGCAAGGCCGACTACGAGAAGCACAAAGTGTACGCCTGCGA AGTGACACACCAGGGACTGAGCAGCCCTGTGACCAAGAGCTTCAACAGAGGCGAGTGC

[0854] H2 (SEQ ID NO: 222)

[0855] CAGGTTCAGCTGGTTCAGAGCGGTGCAGAAGTTAAAAAACCGGGTAGCAGCGTTAAAGTTAGCT GTAAAGCAAGCGGTGGCACCTTTAGCAGCGATAGCATTCACTGGGTTCGTGAGGCACCTGGTCA AGGTCTGGAATGGATGGGTGGTATTATTCCGTATTTTGACACCGCAGACTATGCCCAGAAATTTC AGGGTCGTGTTACCATTACCGCAGATGAAAGCACCAGCACCGCATATATGGAACTGAGCAGCCT GCGTAGCGAAGATACCGCAGTGTATTATTGTGCACGTTATCCGAGTTATGGCTATATCGACTACT GGGGCCAGGGAACCCTGGTCACCGTCTCCTCAGCCTCTACAAAGGGCCCTAGTGTGTTCCCTCTG GCTCCCAGCAGCAAGTCTACATCTGGCGGAACAGCCGCTCTGGGCTGCCTGGTCAAGGATTACTT TCCCGAGCCTGTGACCGTGTCCTGGAATTCTGGCGCTCTGACAAGCGGCGTGCACACCTTTCCAG CTGTGCTGCAAAGCAGCGGCCTGTACTCTCTGACCAGCGTGGTCGAGGTGCCTAGCTCTAGCCTG GGCACCCAGACCTACATCTGCAATGTGAACCACAAGCCTAGCAACACCAAGGTGGACAAGAAGG

[0856] TG G AACCC AAG AG CTGC

[0857] Multi-246

[0858] Hl (SEQ ID NO: 223)

[0859] GAGGTGCAGCTGTTGGAGTCTGGGGGAGGCTTGGTACAGCCTGGGGGGTCCCTGAGACTCTCCT GTGCAGCCTCTGGATTCACCTTTGATGACTATGCCATGGGATGGGTCCGCAGAGCTCCAGGGAAG GGGCTGGAGTGGGTCTCAGCTATTAGTGACTACGGTGGTGGTACATACTACGCAGACTCCGTGA AGGGCCGGTTCACCATCTCCAGAGACAATTCCAAGAACACGCTGTATCTGCAAATGAACAGCCTG AGAGCCGAGGACACGGCCGTATATTACTGTGCGCGCCGTTTTTGGCTGGACTACTGGGGCCAGG G AACCCTGGTC ACCGTCTCCTCAG CCTCTAC AAAG G GCCCCTCCGTTTTTCC ACTGG CTCCTAG CA GCAAGAGCACAAGCGGAGGAACAGCCGCTCTGGGCTGTCTGGTCAAGGATTACTTTCCCGAGCC TGTGACCGTGTCCTGGAATTCTGGTGCTCTGACCAGCGGAGTGGCTACAGGACCTGCTGTGCTGC AGTCTAG CGG CCTGTATTCTCTG AG CAG CGTG GTCACAGTG CCTAG CTCTAGCCTG GG CACCC AG ACCTACATCTGCAACGTGAACCACAAGCCTAGCAACACCAAGGTGGACAAGAAGGTGGAACCCA

[0860] AGAGCTGCGACAAGACCCACACCTGTCCTCCATGTCCTGCTCCAGAACTGCTCGGCGGACCCTCC GTGTTTCTGTTCCCTCCAAAGCCTAAGGACACCCTGATGATCAGCAGAACCCCTGAAGTGACCTGC GTGGTGGTGGATGTGTCTCACGAGGACCCCGAAGTGAAGTTCAATTGGTACGTGGACGGCGTGG AAGTGCACAACGCCAAGACCAAGCCTAGAGAGGAACAGTACAACAGCACCTACAGAGTGGTGTC CGTGCTGACCGTGCTGCACCAGGATTGGCTGAACGGCAAAGAGTACAAGTGCAAGGTGTCCAAC AAGGCCCTGCCTGCTCCTATCGAGAAAACCATCAGCAAGGCCAAGGGCCAGCCAAGAGAACCCC AGGTTTACACACTGCCTCCAAGCAGGGACGAGCTGACCAAGAATCAGGTGTCCCTGACCTGCCTG GTTAAGGGCTTCTACCCTTCCGATATCGCCGTGGAATGGGAGAGCAATGGACAGCCCGAGAACA ACTACAAGACAACCCCTCCTGTGCTGGACAGCGACGGCTCATTCTTCCTGTACAGCAAGCTGACA GTGGACAAGTCCAGATGGCAGCAGGGCAACGTGTTCAGCTGTTCTGTGATGCACGAGGCCCTGC ACAACCACTACACCCAGAAGTCTCTGTCTCTGAGCCCTGGAAAAGGCGGAGGTGGAAGCGGAGG CG G AG G AAGTG GTGG CG G CG G ATCTG AAATTGTTCTG ACCCAG AGTCCGG GTAC ACTG AG CCTG TCACCGGGTGAACGTGCAACCCTGAGCTGTCGTGCAAGCCAGAGCGTTAGCAGCAGCTATCTGG CATGGTATCAGAGAAAACCTGGTCAGGCACCGCGTCTGCTGATTTATGGTGCAAGCAGCCGTGCA ACCGGTATTCCGGATCGTTTTAGCGGTAGCGGTAGTGGCACCGATTTTACCCTGACCATTAGCCG TCTGGAACCGGAAGATTTTGCAGTGTATTATTGTCAGCAGAATACTTGGACCTTGCTGACTTTTGG CCAGGGGACCAAGCTGGAGATCAAACGACCCGTGGCTGCCCCTGCCGTGTTCATCTTCCCACCTT CCGACGAGCAGCTGAAGTCTGGCACAGCCTCTGTCGTGTGCCTGCTGAAGAACTTCTACCCCAGA GAAGCCAAGGTGCAGTGGAAGGTGGACAACGCCCTGCAGAGCGGCAATAGCCAAGAGAGCGTG ACCG AG CAG G AC AGC AAG G ACTCTACCTACAG CCTG AG CAG CACCCTG AC ACTG AG CAAG GCCG ACTACGAGAAGCACAAAGTGTACGCCTGCGAAGTGACCCACCAGGGCCTTTCTAGCCCTGTGACC AAG AG CTTC AACCGG GG CG AATGT

[0861] LI (SEQ ID NO: 224) GAAATTGTTCTGACCCAGAGTCCGGGTACACTGAGCCTGTCACCGGGTGAACGTGCAACCCTGAG CTGTCGTGCAAGCCAGAGCGTTAGCAGCAGCTATCTGGCATGGTATCAGGAGAAACCTGGTCAG GCACCGCGTCTGCTGATTTATGGTGCAAGCAGCCGTGCAACCGGTATTCCGGATCGTTTTAGCGG TAG CG GTAGTGG CACCG ATTTTACCCTG ACCATTAGCCGTCTG G AACCG G AAG ATTTTG CAGTGT ATTATTGTCAGCAGGATTGGCATCACCCGCTTACTTTTGGCCAGGGGACCAAGCTGGAGATCAAA AGGACAGTGGCCGCTCCTAGCGTGTTCATCTTTCCACCTAGCGACGAGCAGCTGAAAAGCGGCAC AGCCAGCGTCGTGTGTTACCTGAACAACTTCTACCCCAGAGAAGCCAAGGTGCAGTGGAAGGTG GACAATGCCCTGCAGAGCGGCAATAGCCAAGAGAGCGTGACCGAGCAGGACAGCAAGGACTCT ACCTACAGCCTGTGGTCCACACTGACCCTGAGCAAGGCCGACTACGAGAAGCACAAAGTGTACG CCTGCGAAGTGACACACCAGGGACTGAGCAGCCCTGTGACCAAGAGCTTCAACAGAGGCGAGTG C

[0862] H2 (SEQ ID NO: 225) CAGGTTCAGCTGGTTCAGAGCGGTGCAGAAGTTAAAAAACCGGGTAGCAGCGTTAAAGTTAGCT GTAAAGCAAGCGGTGGCACCTTTAGCAGCGATAGCATTCACTGGGTTCGTGAGGCACCTGGTCA AGGTCTGGAATGGATGGGTGGTATTATTCCGTATTTTGACACCGCAGACTATGCCCAGAAATTTC AGGGTCGTGTTACCATTACCGCAGATGAAAGCACCAGCACCGCATATATGGAACTGAGCAGCCT GCGTAGCGAAGATACCGCAGTGTATTATTGTGCACGTTATCCGAGTTATGGCTATATCGACTACT GGGGCCAGGGAACCCTGGTCACCGTCTCCTCAGCCTCTACAAAGGGCCCTAGTGTGTTCCCTCTG GCTCCCAGCAGCAAGTCTACATCTGGCGGAACAGCCGCTCTGGGCTGCCTGGTCAAGGATTACTT TCCCGAGCCTGTGACCGTGTCCTGGAATTCTGGCGCTCTGACAAGCGGCGTGCACACCTTTCCAG CTGTGCTGCAAAGCAGCGGCCTGTACTCTCTGACCAGCGTGGTCGAGGTGCCTAGCTCTAGCCTG GGCACCCAGACCTACATCTGCAATGTGAACCACAAGCCTAGCAACACCAAGGTGGACAAGAAGG TG G AACCC AAG AG CTGC Multi-256

[0863] Hl (SEQ ID NO: 226)

[0864] GAGGTGCAGCTGTTGGAGTCTGGGGGAGGCTTGGTACAGCCTGGGGGGTCCCTGAGACTCTCCT GTGCAGCCTCTGGATTCACCTTTAGCAGCCATGATATGGGGTGGGTCCGCAGAGCTCCAGGGAA GGGGCTGGAGTGGGTCTCACATATTAGTTCGGGCGGTGGTAGCACATACTACGCAGACTCCGTG AAGGGCCGGTTCACCATCTCCAGAGACAATTCCAAGAACACGCTGTATCTGCAAATGAACAGCCT GAGAGCCGAGGACACGGCCGTATATTACTGTGCGCGCCCGTTTTGGTTTTTCGACGCACATTTTG ACTACTG GG G CCAG G G A ACCCTG GTCACCGTCTCCTCAG CCTCTAC AAAGG GCCCCTCCGTTTTTC CACTGGCTCCTAGCAGCAAGAGCACAAGCGGAGGAACAGCCGCTCTGGGCTGTCTGGTCAAGGA TTACTTTCCCGAGCCTGTGACCGTGTCCTGGAATTCTGGTGCTCTGACCAGCGGAGTGGCTACAG GACCTGCTGTGCTGCAGTCTAGCGGCCTGTATTCTCTGAGCAGCGTGGTCACAGTGCCTAGCTCT AGCCTGGGCACCCAGACCTACATCTGCAACGTGAACCACAAGCCTAGCAACACCAAGGTGGACA AGAAGGTGGAACCCAAGAGCTGCGACAAGACCCACACCTGTCCTCCATGTCCTGCTCCAGAACTG CTCGGCGGACCCTCCGTGTTTCTGTTCCCTCCAAAGCCTAAGGACACCCTGATGATCAGCAGAACC CCTGAAGTGACCTGCGTGGTGGTGGATGTGTCTCACGAGGACCCCGAAGTGAAGTTCAATTGGT ACGTGGACGGCGTGGAAGTGCACAACGCCAAGACCAAGCCTAGAGAGGAACAGTACAACAGCA CCTACAGAGTGGTGTCCGTGCTGACCGTGCTGCACCAGGATTGGCTGAACGGCAAAGAGTACAA GTGCAAGGTGTCCAACAAGGCCCTGCCTGCTCCTATCGAGAAAACCATCAGCAAGGCCAAGGGC CAG CCA AG AG AACCCCAGGTTTAC ACACTG CCTCC AAG CAG GG ACG AG CTG ACC AAG AATCAG G TGTCCCTGACCTGCCTGGTTAAGGGCTTCTACCCTTCCGATATCGCCGTGGAATGGGAGAGCAAT GGACAGCCCGAGAACAACTACAAGACAACCCCTCCTGTGCTGGACAGCGACGGCTCATTCTTCCT GTACAGCAAGCTGACAGTGGACAAGTCCAGATGGCAGCAGGGCAACGTGTTCAGCTGTTCTGTG ATGCACGAGGCCCTGCACAACCACTACACCCAGAAGTCTCTGTCTCTGAGCCCTGGAAAAGGCGG AGGTGGAAGCGGAGGCGGAGGAAGTGGTGGCGGCGGATCTGAAATTGTTCTGACCCAGAGTCC GGGTACACTGAGCCTGTCACCGGGTGAACGTGCAACCCTGAGCTGTCGTGCAAGCCAGAGCGTT AGCAGCAGCTATCTGGCATGGTATCAGAGAAAACCTGGTCAGGCACCGCGTCTGCTGATTTATGG TGCAAGCAGCCGTGCAACCGGTATTCCGGATCGTTTTAGCGGTAGCGGTAGTGGCACCGATTTTA CCCTGACCATTAGCCGTCTGGAACCGGAAGATTTTGCAGTGTATTATTGTCAGCAGAATACTTGG ACCTTGCTGACTTTTGGCCAGGGGACCAAGCTGGAGATCAAACGACCCGTGGCTGCCCCTGCCGT GTTCATCTTCCCACCTTCCGACGAGCAGCTGAAGTCTGGCACAGCCTCTGTCGTGTGCCTGCTGAA GAACTTCTACCCCAGAGAAGCCAAGGTGCAGTGGAAGGTGGACAACGCCCTGCAGAGCGGCAAT AGCCAAGAGAGCGTGACCGAGCAGGACAGCAAGGACTCTACCTACAGCCTGAGCAGCACCCTGA CACTGAGCAAGGCCGACTACGAGAAGCACAAAGTGTACGCCTGCGAAGTGACCCACCAGGGCCT TTCTAGCCCTGTGACCAAGAGCTTCAACCGGGGCGAATGT

[0865] LI (SEQ ID NO: 227)

[0866] GAAATTGTTCTGACCCAGAGTCCGGGTACACTGAGCCTGTCACCGGGTGAACGTGCAACCCTGAG CTGTCGTGCAAGCCAGAGCGTTAGCAGCAGCTATCTGGCATGGTATCAGGAGAAACCTGGTCAG GCACCGCGTCTGCTGATTTATGGTGCAAGCAGCCGTGCAACCGGTATTCCGGATCGTTTTAGCGG TAG CG GTAGTGG CACCG ATTTTACCCTG ACCATTAGCCGTCTG G AACCG G AAG ATTTTG CAGTGT ATTATTGTCAGCAGACTTGGAGTCCTTTCACTTTTGGCCAGGGGACCAAGCTGGAGATCAAAAGG ACAGTG GCCG CTCCTAG CGTGTTC ATCTTTCCACCTAGCG ACG AG CAG CTG AAA AGCGG CAC AG C CAGCGTCGTGTGTTACCTGAACAACTTCTACCCCAGAGAAGCCAAGGTGCAGTGGAAGGTGGAC AATGCCCTGCAGAGCGGCAATAGCCAAGAGAGCGTGACCGAGCAGGACAGCAAGGACTCTACCT ACAGCCTGTGGTCCACACTGACCCTGAGCAAGGCCGACTACGAGAAGCACAAAGTGTACGCCTG CGAAGTGACACACCAGGGACTGAGCAGCCCTGTGACCAAGAGCTTCAACAGAGGCGAGTGC

[0867] H2 (SEQ ID NO: 228)

[0868] CAGGTTCAGCTGGTTCAGAGCGGTGCAGAAGTTAAAAAACCGGGTAGCAGCGTTAAAGTTAGCT GTAAAGCAAGCGGTGGCACCTTTAGCAGCGATAGCATTCACTGGGTTCGTGAGGCACCTGGTCA AGGTCTGGAATGGATGGGTGGTATTATTCCGTATTTTGACACCGCAGACTATGCCCAGAAATTTC AGGGTCGTGTTACCATTACCGCAGATGAAAGCACCAGCACCGCATATATGGAACTGAGCAGCCT

[0869] GCGTAGCGAAGATACCGCAGTGTATTATTGTGCACGTTATCCGAGTTATGGCTATATCGACTACT GGGGCCAGGGAACCCTGGTCACCGTCTCCTCAGCCTCTACAAAGGGCCCTAGTGTGTTCCCTCTG GCTCCCAGCAGCAAGTCTACATCTGGCGGAACAGCCGCTCTGGGCTGCCTGGTCAAGGATTACTT TCCCGAGCCTGTGACCGTGTCCTGGAATTCTGGCGCTCTGACAAGCGGCGTGCACACCTTTCCAG

[0870] CTGTGCTGCAAAGCAGCGGCCTGTACTCTCTGACCAGCGTGGTCGAGGTGCCTAGCTCTAGCCTG GGCACCCAGACCTACATCTGCAATGTGAACCACAAGCCTAGCAACACCAAGGTGGACAAGAAGG TG G AACCC AAG AG CTGC

[0871] Amino acid sequence for human PD-L1 (SEQ ID NO: 230), amino acids 19-238 of UniProt accession number Q9NZQ7:

[0872] FTVTVPKDLYVVEYGSNMTIECKFPVEKQLDLAALIVYWEMEDKNIIQFVHGEEDLKVQHSSYRQRAR LLKDQLSLGNAALQITDVKLQDAGVYRCMISYGGADYKRITVKVNAPYNKINQRILVVDPVTSEHELTC QAEGYPKAEVIWTSSDHQVLSGKTTTTNSKREEKLFNVTSTLRINTTTNEIFYCTFRRLDPEENHTAELVI PELPLAHPPNER

[0873] Amino acid sequence for human PVRIG (SEQ ID NO: 231), amino acids 41-172 of UniProt accession number Q6DKI7

[0874] TPEVWVQVRMEATELSSFTIRCGFLGSGSISLVTVSWGGPNGAGGTTLAVLHPERGIRQWAPARQA RWETQSSISLILEGSGASSPCANTTFCCKFASFPEGSWEACGSLPPSSDPGLSAPPTPAPILRADL

[0875] References

[0876] Hu, S. et al. (2024) Structural basis for the immune recognition and selectivity of the immune receptor PVRIG for ligand Nectin-2. Structure, 32, 1-12.

[0877] Kabat, E.(1991) Sequences of proteins of immunological interest. No. 91. US Department of Health and Human Services, Public Health Service, National Institutes of Health.

[0878] Kazuki et al. (2011) Mol. Ther. 19(9): 1591-1601.

[0879] Kouprina et al. (2014) Expert Opinion on Drug Delivery 11(4): 517-535.

[0880] Lefranc, Marie-Paule et al. (2003) IMGT unique numbering for immunoglobulin and T cell receptor variable domains and Ig superfamily V-like domains. Developmental & Comparative Immunology 27.1, 55-77

[0881] Li, J. et al. (2021) PVRIG is a novel natural killer cell immune checkpoint receptor in acute myeloid leukemia. Haematologica. 106(12): 3115-3124.

[0882] Medina-Cucurella, A. V. and T. A. Whitehead (2018). "Characterizing Protein-Protein Interactions Using Deep Sequencing Coupled to Yeast Surface Display." Methods Mol Biol 1764: 101-121.

[0883] Nyesiga, B. et al. (2024) RUBY® - a tetravalent (2+2) bispecific antibody format with excellent functionality and IgG-like stability, pharmacology and developability properties. mAbs, 16: 1, 2330113.

[0884] Plaisance, K. et al. (2000) Effect of antipyretic therapy on the duration of illness in experimental influenza A, Shigella sonnei, and Rickettsia rickettsii infections. Pharmacotherapy 20(12): 1417-1422.

[0885] Pruvost, T. et al. (2023) Deciphering cross-species reactivity of LAMP-1 antibodies using deep mutational epitope mapping and AlphaFold. MAbs 15(1): 2175311.

[0886] Sugiura et al., 2019. Restriction of PD-1 function by cis-PD-Ll / CD80 interactions is required for optimal T cell responses. Science 364 (6440): 558-566.

[0887] Van Blarcom, T. et al. (2015) Precise and efficient antibody epitope determination through library design, yeast display and next-generation sequencing. J Mol Biol 427(6 Pt B): 1513-1534.

[0888] Whelan, S. et al (2019) PVRIG and PVRL2 Are Induced in Cancer and Inhibit CD8+ T- cell Function. Cancer Immunol Res. 7(2):257-268.

[0889] Yang, C. et al (2023) Nonredundant Upregulation of CD112R (PVRIG) and PD-1 on Cytotoxic T Lymphocytes Located in T Cell Nests of Colorectal Cancer. Mod Pathol. 36(4): 100089. Zhu, Y. et al. (2016) Identification of CD112R as a novel checkpoint for human T cells. J Exp Med. 213(2): 167-76.

Claims

CLAIMS1. A multispecific antigen-binding protein comprising a first antigen-binding domain specific for PD-L1 and a second antigen-binding domain specific for PVRIG.

2. The multispecific antigen-binding protein according to claim 1, which is a bispecific antibody_comprising a first antigen-binding domain specific for PD-L1 and a second antigen-binding domain specific for PVRIG.

3. The multispecific antigen-binding protein according to claim 1 or 2, which binds to PD-L1 on a cancer cell or an antigen presenting cell.

4. The multispecific antigen-binding protein according to any one of claims 1 to 3, which binds to PVRIG on an immune cell.

5. The multispecific antigen-binding protein according to claim 4, wherein the immune cell is a T cell or an NK cell.

6. The multispecific antigen-binding protein according to any one of the preceding claims, which activates T cells, optionally wherein T cell activation is PD-L1 dependent.

7. The multispecific antigen-binding protein according to any one of the preceding claims, which activates NK cells, optionally wherein NK cell activation is PVRG- dependent.

8. The multispecific antigen-binding protein according to any one of the preceding claims, which induces ADCC against cancer cells.

9. The multispecific antigen-binding protein according to any one of the preceding claims, which binds to PD-L1 on a cancer cell and PVRIG on an immune cell, such as a T cell or an NK cell, and which induces clustering of said cancer cells and immune cells.

10. The multispecific antigen binding protein according to any one of the preceding claims, which blocks binding of CD80 and / or PD-1 to PD-L1.

11. The multispecific antigen binding protein according to any one of the preceding claims, which blocks binding of PVRL2 to PVRIG.

12. The multispecific antigen binding protein according to any one of the preceding claims, which induces an anti-tumour response without inducing T cell exhaustion,optionally which does not increase expression of one or more genes associated with T cell exhaustion, such as TIGIT, LAG3 and / or CTLA4.

13. The multispecific antigen-binding protein according to any one of the preceding claims, which binds to an epitope of human PD-L1 comprising at least one amino acid residue selected from V23, D26, E58, Ml 15, D122, Y123, K124, and R125 or to an epitope of human PD-L1 comprising at least one amino acid residue selected from Y56, E58, R113, Ml 15, SI 17, A121, and D122.

14. The multispecific antigen-binding protein according to claim 13, which binds to an epitope of human PD-L1 comprising the amino acid residues D26, D122, Y123, K124, and R125 or to an epitope of human PD-L1 comprising the amino acid residues E58, M115, and A121.

15. The multispecific antigen-binding protein according to any one of the preceding claims, which binds to an epitope of human PVRIG comprising at least one amino acid residue selected from S71, L72, V90, H92, P93, R95, G96, R98, W100, A137, F139, P140, G142, and S143.

16. The multispecific antigen-binding protein according to claim 15, which binds to an epitope of human PVRIG comprising the amino acid residues L72, V90, R95, G96, R98, W100, A137, F139, P140, and G142.

17. The multispecific antigen-binding protein according to any one of the preceding claims, wherein the first and second antigen-binding domains are selected from antibodies or antigen-binding fragments thereof.

18. The multispecific antigen-binding protein according to any one of the preceding claims, wherein the first and second antigen-binding domains are fused, optionally wherein a heavy chain polypeptide of the first antigen-binding domain is fused to a light chain polypeptide of the second antigen-binding domain, optionally wherein the fusion is via a linker.

19. The multispecific antigen-binding protein according to any one of the preceding claims, wherein the first antigen-binding domain is an antibody comprising an Fc region.

20. The multispecific antigen-binding protein according to any one of claims 17-19, wherein the second antigen-binding domain is a Fab fragment or an scFv fragment.

21. The multispecific antigen-binding protein according to any one of claims 18-20, comprising a first polypeptide chain comprising a heavy chain polypeptide of the first antigen-binding domain fused to a light chain polypeptide of the second antigenbinding domain, optionally via linker, a second polypeptide chain comprising a light chain that pairs with the heavy chain polypeptide of the first antigen-binding domain, and a third polypeptide chain that comprises a heavy chain polypeptide that pairs with the light chain polypeptide of the second antigen-binding domain, optionally wherein the multispecific antigen-binding protein comprises two copies of each of the three different polypeptide chains.

22. The multispecific antigen-binding protein according to any one of the preceding claims, comprising an IgGl Fc region.

23. The multispecific antigen-binding protein according to any one of the preceding claims, wherein the first antigen-binding domain comprises the three heavy chain CDRs and three light chain CDRs contained within a HCVR / LCVR pair selected from SEQ ID NOs 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156, 147 / 157, 148 / 158, 149 / 159, and 150 / 160.

24. The multispecific antigen-binding protein according to claim 23, wherein the first antigen-binding domain comprises an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from SEQ ID NOs 61-66, SEQ ID NOs 67-72, SEQ ID NOs 73-78, SEQ ID NOs 79-84, SEQ ID NOs 85-90, SEQ ID NOs 91-96, SEQ ID NOs 97-102, SEQ ID NOs 103-108, SEQ ID NOs 109-114 and SEQ ID NOs 115-120, optionally wherein the first antigen-binding domain comprises a HCVR / LCVR pair having at least 95% identity to a HCVR / LCVR pair selected from SEQ ID NOs 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156, 147 / 157, 148 / 158, 149 / 159, and 150 / 160.

25. The multispecific antigen-binding protein according to claim 24, wherein the first antigen-binding domain comprises an HCVR / LCVR pair selected from SEQ ID NOs 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156, 147 / 157, 148 / 158, 149 / 159, and 150 / 160.

26. The multispecific antigen-binding protein according to claim 24, wherein the first antigen-binding domain comprises anHCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from SEQID NOs 79-84, SEQ ID NOs 97-102 and SEQ ID NOs 103-108, optionally wherein the first antigen-binding domain comprises a HCVR / LCVR pair having at least 95% identity to a HCVR / LCVR pair selected from SEQ ID NOs 144 / 154, 147 / 157 and 148 / 158.

27. The multispecific antigen-binding protein according to claim 26, wherein the first antigen-binding domain comprises an HCVR / LCVR pair selected from SEQ ID NOs 144 / 154, 147 / 157 and 148 / 158.

28. The multispecific antigen-binding protein according to any one of the preceding claims, wherein the second antigen-binding domain comprises the three heavy chain CDRs and three light chain CDRs contained within a HCVR / LCVR pair selected from SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140.

29. The multispecific antigen-binding protein according to claim 28, wherein the second antigen-binding domain comprises an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from SEQ ID NOs 1-6, SEQ ID NOs 7-12, SEQ ID NOs 13-18, SEQ ID NOs 19-24, SEQ ID NOs 25-30, SEQ ID NOs 31-36, SEQ ID NOs 37-42, SEQ ID NOs 43-48, SEQ ID NOs 49-54 and SEQ ID NOs 55-60, optionally wherein the second antigen-binding domain comprises a HCVR / LCVR pair having at least 95% identity to a HCVR / LCVR pair selected from SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140.

30. The multispecific antigen-binding protein according to claim 29, wherein the second antigen-binding domain comprises an HCVR / LCVR pair selected from SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140.

31. The multispecific antigen-binding protein according to claim 29 wherein the second antigen-binding domain comprises the HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination of SEQ ID NOs 13- 18, optionally wherein the second antigen-binding domain comprises a HCVR / LCVR pair having at least 95% identity to SEQ ID NOs 123 / 133.

32. The multispecific antigen-binding protein according to claim 31, wherein the second antigen-binding domain comprises the HCVR / LCVR pair of SEQ ID NOs 123 / 133.

33. An antigen-binding protein specific for PD-L1, comprising the three heavy chain CDRs and three light chain CDRs contained within a HCVR / LCVR pair selected from SEQ ID NOs 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156, 147 / 157, 148 / 158, 149 / 159, and 150 / 160, optionally wherein:(a) the antigen-binding protein comprises an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from the following combinations of CDR sequences: SEQ ID NOs 61-66, SEQ ID NOs 67-72, SEQ ID NOs 73-78, SEQ ID NOs 79-84, SEQ ID NOs 85-90, SEQ ID NOs 91-96, SEQ ID NOs 97-102, SEQ ID NOs 103-108, SEQ ID NOs 109-114 and SEQ ID NOs 115-120, optionally wherein the antigen-binding protein comprises a HCVR / LCVR pair having at least 95% identity to a HCVR / LCVR pair selected from SEQ ID NOs 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156, 147 / 157, 148 / 158, 149 / 159, and 150 / 160; and / or(b) the antigen-binding protein comprises an HCVR / LCVR pair selected from SEQ ID NOs 141 / 151, 142 / 152, 143 / 153, 144 / 154, 145 / 155, 146 / 156, 147 / 157, 148 / 158, 149 / 159, and 150 / 160.

34. An antigen-binding protein specific for PVRIG, comprising the three heavy chain CDRs and three light chain CDRs contained within a HCVR / LCVR pair selected from SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140, optionally wherein: the antigen-binding protein s comprises an HCDR1 / HCDR2 / HCDR3 / LCDR1 / LCDR2 / LCDR3 combination selected from the following combinations of CDR sequences: SEQ ID NOs 1-6, SEQ ID NOs 7-12, SEQ ID NOs 13-18, SEQ ID NOs 19-24, SEQ ID NOs 25-30, SEQ ID NOs 31-36, SEQ ID NOs 37-42, SEQ ID NOs 43-48, SEQ ID NOs 49-54 and SEQ ID NOs 55-60, optionally wherein the antigen-binding protein comprises a HCVR / LCVR pair having at least 95% identity to a HCVR / LCVR pair selected from SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140; and / or(b) the antigen-binding protein comprises an HCVR / LCVR pair selected from SEQ ID NOs 121 / 131, 122 / 132, 123 / 133, 124 / 134, 125 / 135, 126 / 136, 127 / 137, 128 / 138, 129 / 139, and 130 / 140.

35. A combination of an antigen-binding protein specific for PD-L1 as defined in claim 33 and an antigen-binding protein specific for PVRIG as defined in claim 34.

36. One or more nucleic acids encoding a multispecific antigen-binding protein as defined in any one of claims 1-32, an antigen-binding protein specific for PD-L1 as defined in claim 33, or an antigen-binding protein specific for PVRIG as defined in claim 34.

37. A pharmaceutical composition comprising a multispecific antigen-binding protein as defined in any one of claims 1-32, an antigen-binding protein specific for PD-L1 as defined in claim 33, an antigen-binding protein specific for PVRIG as defined in claim 34, a combination of antigen-binding proteins as defined in claim 35, or one or more encoding nucleic acids as defined in claim 36, and a pharmaceutically acceptable excipient.

38. A multispecific antigen-binding protein as defined in any one of claims 1-32, an antigen-binding protein specific for PD-L1 as defined in claim 33, an antigen-binding protein specific for PVRIG as defined in claim 34, a combination of antigen-binding proteins as defined in claim 35, one or more encoding nucleic acids as defined in claim 36, or a pharmaceutical composition according to claim 37, for use in a method of treatment of the human or animal body by therapy.

39. A multispecific antigen-binding protein as defined in any one of claims 1-32, an antigen-binding protein specific for PD-L1 as defined in claim 33, an antigen-binding protein specific for PVRIG as defined in claim 34, a combination of antigen-binding proteins as defined in claim 35, one or more encoding nucleic acids as defined in claim 36, or a pharmaceutical composition according to claim 37, for use in a method of treatment of a cancer or tumour, optionally wherein:(a) the cancer or tumour is a solid cancer or tumour or a haematological cancer; and / or(b)the cancer or tumour is refractory to treatment with an antagonist of PD-L1 or an antagonist of PVRIG.

Images