ANTI CD25 FC GAMMA RECEPTOR BISPECIFIC ANTIBODIES FOR TUMOR SPECIFIC CELL DEPLETION

Abstract

The present disclosure relates to a method of treating a solid tumour, wherein said method involves the use of an antibody to CD25. In particular, the antibody to CD25 is optimized for depletion of regulatory T cells (Treg) within tumours. The present invention also provides novel anti-CD25 antibodies and their combination with other anti-cancer drugs, such as immune checkpoint inhibitors, compounds that target cancer antigens or the inhibitory Fc receptor FcyRllb (CD32b).

Claims

1. A method of treating a solid tumor in a human subject comprising of administering an anti-CD25 antibody and administering an immune checkpoint inhibitor to the subject, wherein the anti-CD25 antibody is a human IgG1 antibody that binds to at least one activating Fcγ receptor selected from FcγRI, FcγRIIc, and FCγRIIIa with high affinity, and depletes tumour-infiltrating regulatory T cells.

2. The method according to claim 1, wherein the anti-CD25 antibody has a dissociation constant (K.sub.d) for CD25 of less than 10.sup.−8 M, and/or a dissociation constant for at least one activating Fcγ receptor of less than about 10.sup.−6 M.

3. The method according to claim 1, wherein the anti-CD25 antibody: (a) binds to Fcγ receptors with an activatory to inhibitory ratio (A/I) superior to 1; and/or (b) binds to at least one of FcγRI, FcγRIIc, and FcγRIIIa with higher affinity than it binds to FcγRIIb.

4. The method according to claim 1, wherein the anti-CD25 antibody is a monoclonal antibody.

5. The method according to claim 1, wherein the anti-CD25 antibody elicits an enhanced CDC, ADCC and/or ADCP response.

6. The method according to claim 5, wherein the anti-CD25 antibody elicits an enhanced ADCC response.

7. The method according to claim 1, wherein said anti-CD25 antibody is administered to a subject who has an established tumour.

8. The method according to claim 1, wherein said method further comprises the step of identifying a subject who has a solid tumour.

9-17. (canceled)

18. A kit for use in the treatment of cancer comprising an anti-CD25 antibody, as defined in claim 1 and an immune checkpoint inhibitor.

19. A pharmaceutical composition comprising an anti-CD25 antibody and an immune checkpoint inhibitor in a pharmaceutically acceptable medium.

21-29. (canceled)

30. The method according to claim 1 wherein said immune checkpoint inhibitor is a PD-1 antagonist.

31. The method according to claim 30, wherein said PD-1 antagonist is an anti-PD-1 antibody.

32. The method according to claim 31, wherein said anti-PD-1 antibody is Nivolumab or Pembrolizumab

33. The method according to claim 30, wherein said PD-1 antagonist is an anti-PD-L1 antibody.

34. The method according to claim 33, wherein said anti-PD-L1 antibody is Atezolizumab.

35. The method according to claim 1, wherein the anti-CD25 antibody and the immune checkpoint inhibitor are administered simultaneously, separately or sequentially.

Description

[0129] The invention will now be further described by way of the following Examples, which are meant to serve to assist one of ordinary skill in the art in carrying out the invention and are not intended in any way to limit the scope of the invention, with reference to the drawings in which:

[0130] FIG. 1—shows the expression of CD25pos regulatory T cells to blood and lymph node (A) Expression of CD25 (detection antibody clone 7D4; anti-mouse CD25, IgM isotype) on the surface of T cell subsets that are present in lymph nodes (LN) and tumour infiltrating lymphocytes (TIL) of different tumour models. Histograms are representative of one mouse for each tumour model. (B) Percentage of CD25 positive cells and MFI of CD25 in PBMC and T cell subsets from pooled data (n=10) of individual experiments using the MCA205 tumour model. The same evaluation (restricted to T cell subsets) have been performed in MC38, B16, and CT26 tumour models in (C) and (D). Error bars represent standard error (SE) of the mean. Statistical relevance between CD4-positive, Foxp3-positive cells and CD8-positive or CD4-positive/Foxp3-negative cells is indicated.

[0131] FIG. 2—shows the restriction of anti-CD25 (αCD25)-mediated depletion of CD25-positive, regulatory T cells to blood and lymph node in MCA205 tumour model. (A) Expression of CD25 (detection antibody clone 7D4; anti-mouse CD25, IgM isotype) and FoxP3 in CD4-positive T cells. (B) Mean fluorescence intensity of CD25 on Treg (gated on CD4-positive, FoxP3-positive T cells). Tumor-bearing mice were injected with 200 μg of anti-CD25-r1 (αCD25-r1; anti-CD25 rat IgG1), anti-CD25-m2a (αCD25-m2a; anti-CD25 murine IgG2a), anti-CTLA-4 (αCTLA-4; anti-CTLA4 clone B56), or not treated (no tx) on days 5 and 7 after s.c. inoculation with 5×10.sup.5 MCA205 cells. Peripheral blood mononuclear cells (PBMC), lymph nodes (LN) and tumor infiltrating lymhocytes (TIL) were harvested on day 9, processed and stained for flow cytometry analysis.

[0132] FIG. 3—shows the anti-CD25 (αCD25)-mediated effects on T cells sub-populations in the MCA205 tumour model of FIG. 2. (A) Percentage of FoxP3-positive cells from total CD4-positive T cells and (B) absolute number of CD4-positive, FoxP3-positive T cells in PBMC (number of cells/mL), LN (total number of cells in three draining lymph nodes) and TIL (number of cells/g of tumour) are shown in parallel to CD4-positive, FoxP3-negative T cells (C) the ratio of effector CD4-positive, FoxP3-positive T cells (Treg cells) and (D) the ratio of effector CD8-positive T cells)/Treg cells, gated on CD4-positive FoxP3-negative and CD8-positive T cells.

[0133] FIG. 4—shows representative histograms for the expression of FcγRs on B cells (CD19-positive), T cells (CD3-positive, CD5-positive), NK cells (NK1.1-positive), granulocytes (CD11b+Ly6G+), conventional dendritic cells (cDC; CD11c-high MHCII-positive) and monocyte/macrophages (Mono/Mϕ; CD11b-positive, Ly6G-negative, NK1.1-negative, CD11c-low/negative). as assessed by flow cytometry in untreated MCA205 tumour model (see FIG. 2) 10 days after tumour challenge. Error bars represent SEM (n=3); data corresponds to one of three separate experiments across which findings were consistent.

[0134] FIG. 5—shows how Treg depletion depends on expression of activatory Fc-gamma receptors. C57BL/6 wild type mice (wt) and Fcer1g−/− mice were injected subcutaneously with 5×10.sup.5 MCA205 cells on day 0 and then injected with 200 μg of anti-CD25 on days 5 and 7. Tumours, draining lymph nodes and blood were harvested on day 9, processed and stained for flow cytometry analysis. Regulatory T cells were identified by CD4 and FoxP3 expression in PBMC, LN, and TIL. Percentage of Foxp3+ from total CD4+ cells are shown (A). The same approach was applied in wild-type (wt), Fcgr3−/−, Fcgr4−/− or Fcgr2b−/−, demonstrating the inhibition of αCD25-r1-mediated Treg depletion in tumors by FcγRIIb. The plots show quantification of the percentage of Treg (CD4+Foxp3+) from total CD4+ T cells in TIL only (B).

[0135] FIG. 6—shows the synergistic effect of anti-CD25-m2a and anti-PD-1 combination results in eradication of established tumours. Growth curves of individual mice (A) and mean of MCA205 tumour volume for each treatment group over time (B) are shown. The number of tumor-free survivors after 100 days or the statistical significance is indicated in each graph. Error bars represent SE of the mean. Kaplan-Meier survival curves with cumulative data of two separate experiments are also shown. Survival curves of mice injected with MC38 (C) or CT26 (D) tumour cells and treated as described in the MCA205 model (n=10 per condition) are also shown. Mice were injected subcutaneously with 5×10.sup.5 MCA205, MC38, or CT26 cells and then treated with the indicated anti-CD25 (200 μg i.p.) on day 5, followed (or not) by the administration of anti-PD-1 (αPD-1, anti-PD1, clone RMP1-14; 100 μg i.p.) on days 6, 9 and 12. Tumour size was measured twice a week and mice were euthanized when any orthogonal diameter reached 150 mm.

[0136] FIG. 7—shows functional analyses of MCA205 tumour model that was established as described in FIG. 6, using immune cells that were harvested on day 14. Proportion (%) of Ki67+ cells in tumour-infiltrating CD4+Foxp3− and in CD8+ T cells in MCA205 tumours. (A) and the CD4-positive, FoxP3-negative Teff/Treg ratio and CD8-positive/Treg ratio (B) in the tumour is shown for each treatment group. The intracellular staining of tumour-infiltrating lymphocytes for IFNg expression following ex vivo re-stimulation with PMA and ionomycin (C) and frequency of interferon gamma (IFNγ)-producing effector T cells (D) is also shown for the same treatment groups in CD4-positive and CD8-positive cells. Histograms in (B) correspond to a representative mouse per treatment group. Representative plots from two separate experiments (n=10) and statistical significance are provided in (A), (B), and (D).

[0137] FIG. 8—shows that tumour elimination by anti-CD25-m2a/anti-PD1 is CD8+ T cell dependent. MCA205 tumour growth curves of individual mice not treated (no tx, A), treated with a combination of anti-CD25-m2a with anti-PD-1 (αPD-1+αCD25-m2a; B), or the same combination further including anti-CD8 (αPD-1+αCD25-m2a+αCD8; C). The number of survivors after 40 days for each treatment group (n=7) is indicated in each graph. The corresponding Kaplan-Meier survival curves were also generated (D). Mice were injected s.c. with 5×10.sup.5 MCA205 cells and treated with 200 μg of anti-CD25-m2a (αCD25-m2a, clone PC61, mouse IgG2a isotype) on day 5 followed by 100 μg of anti-PD-1 (αPD-1, clone RMP1-14) i.p. on days 6, 9 and 12. In the indicated group of mice, CD8-positive cells were depleted by injecting 200 μg of anti-CD8 (αCD8, clone 2.43) i.p. on days 4, 9, 12 and 17. Tumour sized was measured twice a week and mice were euthanized when any orthogonal diameter reached 150 mm.

[0138] FIG. 9—shows that anti-CD25-m2a/anti-PD-1 therapy induces at least partial tumour control against B16 melanoma tumours. B16 tumour growth curves of individual mice treated with Gvax alone or in combination with the indicated antibodies, as defined in the FIG. 6(A). The corresponding Kaplan-Meier survival curves were also generated (B). Mice were injected with 5×10.sup.4 B16 melanoma cells intra-dermally (i.d.) and then treated with 200 μg of anti-CD25 (αCD25-r1, clone PC61 rat IgG1 isotype or αCD25-m2a, clone PC61 mouse IgG2a isotype) on day 5 followed by 200 μg of anti-PD-1 (αPD-1, clone RMP1-14) i.p. and 1×10.sup.6 irradiated (150 Gy) B16-Gvax i.d. on days 6, 9 and 12. Tumour growth was followed up and the mice euthanized when any orthogonal diameter reached 150 mm or on day 80 of the study, whichever came first. The median survival in days for the different groups (n, number of mice) was: 21d for Gvax only (n=14), 27d for Gvax+αPD-1 (n=15), 21d for Gvax+αCD25-r1 (n=7), 33d for Gvax+αCD25-m2a (n=8), 29d for Gvax+αPD-1+αCD25-r1 (n=13), and 39d for Gvax+αPD-1+αCD25-m2a (n=12).

[0139] FIG. 10—shows CT26 tumour growth curves of individual mice not treated (PBS, vehicle only), treated with anti-mouse CD25 having either IgG1 (PC61m1; mouse IgG1 isotype, thus with low FcReceptor-mediated killing activity, low ADCC and CDC activity) or IgG2a (PC61m2; mouse IgG2a, thus with high Fc Receptor mediated activity, high ADCC, and CDC activity), and further combined or not with anti-mouse PD1 (αPD1 RMP1-14). CT26 cells used for implantation were harvested during log phase growth and re-suspended in cold PBS. On Day 1 of the study, each mouse was injected subcutaneously in the right flank with 3×10.sup.5 cells in 0.1 mL cell suspension. The anti-mouse CD25 was injected i.p. (10 mg/kg) at Day 6 (when palpable tumours were detected). The anti-mouse PD1 was injected i.p. (100 μg/injection) at Day 7, Day 10, Day 14, and Day 17. Tumours were calipered in two dimensions twice weekly to monitor growth. Tumour size, in mm.sup.3, was calculated as follows: Tumour Volume=(w2×l)/2 where w=width and l=length, in mm, of the tumour. The study endpoint was a tumour volume of 2000 mm.sup.3 or 60 days, whichever came first.

[0140] FIG. 11—shows CT26 tumour growth curves of individual mice not treated (PBS, vehicle only), treated with anti-mouse CD25 having either IgG1 or IgG2a (PC61m1, and PC61m2 with, respectively), and further combined or not with anti-mouse PD-L1 clone 10F.9G2 (aPDL1 10F.9G2). Model, regimen, and analysis was performed as for the αPD1-based combination experiment of FIG. 10.

[0141] FIG. 12—shows MC38 tumour growth curves of individual mice not treated (PBS, vehicle only), treated with anti-mouse CD25 having either IgG1 or IgG2a (PC61m1, and PC61m2, respectively), and further combined or not with anti-mouse PD1 clone RMP1-14 (αPD1 RMP1-14), as described for CT26 tumour model in FIG. 10. The MC38 colon carcinoma cells used for implantation were harvested during log phase growth and re-suspended in cold PBS. Each mouse was injected subcutaneously in the right flank with 5×10.sup.5 tumour cells in a 0.1 mL cell suspension. Tumours were monitored as their volumes approached the target range of 100 to 150 mm.sup.3. Twenty-two days after tumour implantation, on Day 1 of the study, animals with individual tumour volumes ranging from 63-196 mm.sup.3 were sorted into nine groups (n=10) with group mean tumour volumes ranging from 104-108 mm.sup.3. Treatments began on D1 in mice bearing established MC38 tumours. The effects of each treatment were compared to a vehicle-treated control group that received PBS intraperitoneally (i.p.) on Day 1, Day 2, Day 5, Day 9, and Day 12. Anti-PD1 was administered i.p. at 100 μg/animal, twice weekly for two weeks, beginning on Day 2. PC61-ml and PC61-m2a were administered i.p. once on Day 1 at 200 μg/animal. Tumour measurements were taken twice weekly until Day 45 with individual animals exiting the study upon reaching the tumour volume endpoint of 1000 mm.sup.3.

[0142] FIG. 13—shows MC38 tumour growth curves of individual mice not treated (PBS, vehicle only), treated with anti-mouse CD25 having either IgG1 or IgG2a (PC61m1, and PC61m2 with, respectively), and further combined or not with anti-mouse PD-L1 clone 10F.9G2 (aPDL1 10F.9G2). Model, regimen, and analysis was performed as for the αPD1-based combination experiment of FIG. 12.

[0143] FIG. 14—shows CD25 expression in peripheral of tumour-localized immune cells in samples from distinct types of human cancers. Representative histograms demonstrate CD25 expression in TIL subsets from a stage IV human ovarian carcinoma (peritoneal metastasis; A) and in a human bladder cancer (B). Representative histograms were also obtained for individual CD8-positive, CD4-positive, FoxP3-negative and CD4-positive, FoxP3-positive T cell subsets within PBMC and TIL that are isolated from other types of cancer (C). Quantification of CD25 expression as percentage (%) and mean fluorescence intensity (MFI) on individual T cell subsets within each studied patient cohort for melanoma (upper panel), NSCLC (middle panel) and RCC (lower panel) is also shown (D).

[0144] FIG. 15—shows data on CD25 expression in patients treated with anti-PD-1 Multiplex immunohistochemical (IHC) analysis of a subcutaneous melanoma metastasis prior to anti-PD-1 therapy (‘Baseline’) and following two infusions (‘Week 6’) is shown in parallel to the quantification of CD8 and FoxP3 IHC staining at baseline and week 6 in two patients, one responding and one non-responding to therapy at week 6 (B; mean count per ×40 high power field is displayed). Percentage (%) of CD8-positive, CD25-positive, and FoxP3-positive, CD25-positive, double-stained cells at baseline and on therapy (at week 6) is shown for Melanoma and RCC patients treated with anti-PD1 (C).

[0145] FIG. 16—shows structure and binding activity of bispecfiic anti-IgG1-, anti-PD-L1-based Duobody (bs CD25/PD-L1) that have been generated by using the antigen binding region of anti-mouse CD25 (PC61) and anti-mouse/human PD-L1 (clone S70), both having a human IgG1 isotype and mutated in a specific amino acid (K409R for PC61-IgG1 and F405L for S70-IgG1; A). The specificity of bs CD25/PD-L1 for CD25 have been tested using a cell line (SUP-T1 cells, human T lymphoblasts; SUP-T1 [VB] ATCC® CRL-1942™) that has been transfected with a vector expressing either mouse CD25 (CD25+ cell line) or mouse PD-L1 (PD-L1+ cell line). The original cell line and the other two resulting cell lines have been used to compare the binding ability of bs CD25/PD-L1 to binding of the related monospecific antibody (aCD25, clone PC61; aPD-L1, clone S70). The CD25+ cell line and PD-L1+ cell lines are mixed at 1:1 ratio with each other (or each separately with untransfected, control cells), then incubated with bs CD25/PD-L1, aCD25, aPD-L1, or without any antibody (NO antibody) for 30 minutes. After the incubation, the three groups of cell samples are analysed in flow cytometry to calculate the percentage of double positive cells in the different cell samples (B). The specificity of bs CD25/PD-L1 (BsAb) has been confirmed using the CD25+ cell line and PD-L1+ cell lines separately. Each cell line have been labelled with either the bs CD25/PDL1, or the respective monospecific Ab (MsAb, anti-mouse CD25 IgG1 for CD25+ cells and anti-mouse PD-L1 for PD-L1+ cells) as primary antibody, or with buffer only. Cells were then incubated with aHuman AF647 (aHuman) as secondary antibody in FACS buffer for 30 mins as well as fixable viability dye. Cells incubated with the secondary antibody only (aHuman AF647) or cells incubated with neither primary nor secondary antibody (unstained) are used as negative controls. Cells are then analysed by flow cytometry to calculate the percentage of positive cells obtained with BsAb compared to MsAb (indicated in the right of each panel; C).

[0146] FIG. 17—shows the impact of bispecific IgG1-based Duobody (Bs CD25 PD-L1), the anti-mouse CD25 (aCD25) IgG1 and the anti-mouse PD-L1 (aPD-L1) IgG1 monospecific antibodies, separately or mixed together (aCD25&aPD-L1), or isotype IgG1 control, as described in FIG. 16, on effector and regulatory T cells in LN and tumour in the MCA205 tumour mouse model (established as described in FIG. 3; with four or five mice for each group). The samples were used for isolating Tumor Infiltrating Lymphocytes (TIL) or in lymph node cells (LN) were isolated and analysed for the presence of effector and regulatory T cells th in each treatment group on the basis of FoxP3, CD3, and CD4 positivity (A) or of CD8 positivity/Treg (FoxP3-positive, CD4-positive) ratio (B). The in vivo effect of each treatment on the ability of tumor-infiltrating CD4-positive T cells to respond to stimulation was also evaluated. TIL were re-stimulated in vitro using PMA and ionomycin, in the presence of a Golgi plug protein inhibitor and then stained extracellularly for CD5 and CD4 and intracellularly after fixation for Interferon-γ (IFNg). The percentage of CD5-positive and CD4-positive T cells also positive for IFNg was analysed by flow cytometry (C).

EXAMPLES

Materials & Methods

Mice

[0147] C57BL/6 and BALB/c mice were obtained from Charles River Laboratories. Fcer1g.sup.−/− and Fcgr3.sup.−/− mice were kindly provided by S. Beers (University of Southampton, UK). Fcgr4.sup.−/− and Fcgr2b.sup.−/− mice were a kind gift from J. V. Ravetch (The Rockefeller University, New York, USA). All animal studies were performed under University College of London and UK Home Office ethical approval and regulations.

Cell Lines and Tissue Culture

[0148] MC38, B16, CT26, and MCA205 tumour cells (3-methylcholanthrene-induced weakly immunogenic fibrosarcoma cells; from G. Kroemer, Gustave Roussy Cancer Institute) and 293T cells used for retrovirus production were cultured in Dulbecco's modified Eagle medium (DMEM, Sigma) supplemented with 10% fetal calf serum (FCS, Sigma), 100 U/mL penicillin, 100 μg/mL streptomycin and 2 mM L-glutamine (all from Gibco). K562 cells used for antibody production were cultured in phenol red-free Iscove modified Dulbecco medium (IMDM) supplemented with 10% IgG-depleted FCS (Life Technologies). B16 (mouse skin melanoma cells) and CT26 (N-nitroso-N-methylurethane-induced, undifferentiated colon carcinoma cell line) cells and related culture conditions are available through ATCC.

Antibody Production

[0149] The sequence of the variable regions of the heavy and light chains of anti-CD25 were resolved from the PC-61.5.3 hybridoma by rapid amplification of cDNA ends (RACE) and then cloned into the constant regions of murine IgG2a and κ chains sourced from the pFUSEss-CHIg-mG2A and pFUSE2ss-CLIg-mk plasmids (Invivogen). Each antibody chain was then sub-cloned into a murine leukemia virus (MLV)-derived retroviral vector. For preliminary experiments, antibody was produced using K562 cells transduced with vectors encoding both the heavy and the light chains. The re-cloned, anti-CD25 heavy variable DNA sequence from PC-61.5.3 antibody encodes for the following protein sequence:

TABLE-US-00002 METDTLLLWVLLLWVPGSTGEVQLQQSGAELVRPGTSVKLSCKVSGDTIT AYYIHFVKQRPGQGLEWIGRIDPEDDSTEYAEKFKNKATITANTSSNTAH LKYSRLTSEDTATYFCTTDNMGATEFVYWGQGTLVTVSS

[0150] The re-cloned, anti-CD25 light variable DNA sequence from PC-61.5.3 antibody encodes for the following protein sequence:

TABLE-US-00003 METDTLLLWVLLLWVPGSTGQVVLTQPKSVSASLESTVKLSCKLNSGNIG SYYMHWYQQREGRSPTNLIYRDDKRPDGAPDRFSGSIDISSNSAFLTINN VQTEDEAMYFCHSYDGRMYIFGGGTKLTVL

[0151] The antibody was purified from supernatants using a protein G HiTrap MabSelect column (GE Healthcare), dialyzed in phosphate-buffered saline (PBS), concentrated and filter-sterilized. For further experiments, antibody production was outsourced to Evitria AG. Commercial anti-CD25 clone PC-61 was purchased from BioXcell. The published anti-PDL1 (MPDL3280A/RG7446) variable heavy and light DNA sequences have been recloned and expressed as recombinant antibodies.

In Vivo Tumour Experiments

[0152] Cultured tumour cells were trypsinized, washed and resuspended in PBS and injected subcutaneously (s.c.) in the flank (5×10.sup.5 cells for MCA205 and MC38 models in C57BL/6 mice; 2.5×10.sup.5 cells for B16 model in C57BL/6 mice, 5×10.sup.5 cells for CT26 models in BALB/c mice) cells). Antibodies were injected intraperitoneally (i.p.) at the time points described in the figure legends. For functional experiments, 10 days later the tumors, draining lymph nodes, and tissues were harvested and processed for analysis by flow cytometry as described in Simpson et al. (2013) J Exp Med 210, 1695-710. For therapeutic experiments, tumours were measured twice weekly and volumes calculated as the product of three orthogonal diameters. Mice were humanely euthanized when any diameter reached 150 mm. Tumor-bearing mice were treated with 200 μg of anti-CD25-r1 (αCD25-r1), anti-CD25-m2a (αCD25-m2a) or anti-CTLA-4 (αCTLA-4) on days 5 and 7 and 100 μg of anti-PD-1 on days 6, 9 and 12. For the therapeutics experiments, the mice were only treated on day 5, for the phenotyping and depletion on day 5 and 7. Tumour size was measured twice a week and mice were euthanized when any tumour dimension reached 150 mm. Peripheral blood mononuclear cells (PBMC), lymph nodes (LN) and tumors (TIL) were harvested on day 9, processed and stained for flow cytometry analysis.

Flow Cytometry

[0153] Acquisition was performed with a BD LSR II Fortessa (BD Biosciences). The following directly conjugated antibodies were used: anti-CD25 (7D4)-FITC, CD4 (RM4-5)-v500 (BD Biosciences); anti-IFNγ (XMG1.2)-AlexaFluor488, anti-PD-1 (J43)-PerCP-Cy5.5, anti-Foxp3 (FJK-16s)-PE, anti-CD3 (145-2C11)-PE-Cy7, anti-Ki67 (SolA15)-eFluor450, anti-CD5 (53-7.3)-eFluor450, fixable viability dye-eFluor780 (eBioscience); anti-CD8 (53-6.7)-BrilliantViolet650 (BioLegend); and anti-granzyme B (GB11)-APC (Invitrogen). The following antibodies were used to stain human cells: anti-CD25 (BC96)-BrilliantViolet650 (Biolegend), anti-CD4 (OKT4)-AlexaFluor700 (eBioscience), anti-CD8 (SK1)-V500, anti-Ki67 (B56)-FITC (BD Biosciences); anti-FoxP3 (PCH101)-PerCP-Cy5.5 (eBioscience); anti-CD3 (OKT3)-BrilliantViolet785 (Biolegend). Intranuclear staining of Foxp3 was done using the Foxp3 Transcription Factor Staining Buffer Set (eBioscience). For intracellular staining of cytokines, cells were re-stimulated with phorbol 12-myristate 13-acetate (PMA, 20 ng/mL) and ionomycin (500 ng/mL) (Sigma Aldrich) for 4 hours at 37 C in the presence of GolgiPlug (BD Biosciences) and then stained using Cytofix/Cytoperm buffer set (BD Biosciences). For quantification of absolute number of cells, a defined number of fluorescent beads (Cell Sorting Set-up Beads for UV Lasers, ThermoFisher) was added to each sample before acquisition and used as counting reference.

Human Tissues

[0154] Peripheral blood (PBMCs) and tumor-infiltrating lymphocytes (TIL) were studied in three separate cohorts of patients with advanced melanoma (n=10, 12 lesions), early-stage non-small cell lung cancer (NSCLC) (n=8) and renal cell carcinoma (RCC) (n=5). Presented human data derives from three separate, ethically approved, translational studies (melanoma REC no. 11/LO/0003, NSCLC—REC no. 13/LO/1546, RCC—REC no. 11/LO/1996). Written, informed consent was obtained in all cases.

Isolation of Tumour-Infiltrating Lymphocytes (TILs)

[0155] Tumours were taken directly from the operating theatre to the department of pathology, where tumour representative areas were isolated. Samples were subsequently minced under sterile conditions followed by enzymatic digestion (RPMI-1640 (Sigma) with Liberase TL research grade (Roche) and DNAse I (Roche)) at 37° C. for 30 minutes before mechanical dissociation using gentleMACS (Miltenyi Biotech). Resulting single cell suspensions were filtered and enriched for leukocytes by passage through a Ficoll-paque (GE Healthcare) gradient. Live cells were counted and frozen in human AB serum (Sigma) with 10% dimethyl sulfoxide at −80° C. before transfer to liquid nitrogen.

Phenotypic Analysis of TILs and PBMCs by Multi-Parametric Flow Cytometry

[0156] Tumour samples and PBMCs were thawed, washed in complete RPMI, re-suspended in FACS buffer (500 mL PBS, 2% FCS, 2 nM EDTA) and placed in round-bottomed 96 well plates. A mastermix of surface antibodies was prepared at the manufacturer's recommended dilution: CD8-V500, SK1 clone (BD Biosciences), PD-1-BV605, EH12.2H7 clone (Biolegend), CD3-BV785. A fixable viability dye (eFlour780, eBioscience) was also included the surface mastermix. Following permeablisation for 20 minutes with use of an intracellular fixation and permeabilization buffer set (eBioscience), an intracellular staining panel was applied consisting of the following antibodies used at the manufacturers recommended dilution: granzyme B-V450, GB11 clone (BD Biosciences), FoxP3-PerCP-Cy5.5, PCH101 clone (eBioscience), Ki67-FITC, clone B56 (BD Biosciences) and CTLA-4-APC, L3D10 clone (Biolegend).

Multiplex Immunohistochemistry

[0157] Tumour samples were fixed in buffered formalin and embedded in paraffin. 2-5 μm tissue sections were cut and stained with the following antibodies for immunohistochemistry: anti-CD8 (SP239), anti-CD4 (SP35) (Spring Biosciences Inc.), anti-FoxP3 (236A/E7) (a gift from Dr. G. Roncador CNIO, Madrid, Spain) and anti-CD25 (4C9) (Leica Biosystems). For multiple staining, paraffin-embedded tissue sections were incubated with the primary antibodies for 30 min after antigen retrieval by using cell conditioning 1 reagent (Ventana Medical Systems, Inc.) and hydrogen peroxide for inactivation of endogenous peroxidase. Detection was performed using a peroxidase-based detection reagent (OptiView DAB IHC Detection Kit Ventana Medical Systems, Inc.) and an alkaline phosphatase detection reagent (UltraView Universal Alkaline Phosphatase Red Detection Kit, Ventana Medical Systems, Inc.). A further cycle of immuno-alkaline phosphatase was performed by using an alternative substrate (Fast Blue if Fast Red had been used previously, or vice versa). Immunohistochemistry and protein reactivity patterns were assessed. Scoring of multiple immuno-staining was also performed. Approval for this study was obtained from the National Research Ethics Service, Research Ethics Committee 4 (REC Reference number 09/H0715/64).

Construction and Validation of Anti-CD25- and Anti-PD-L1-Based Bispecific Duobody

[0158] The Bs CD25 PD-L1 Duobody has been generated and produced in accordance to technology described in the literature starting from two parental IgG1s containing single matching point mutations in the CH3 domain that allow Fab exchange (Labrijn A F et al., Nat Protoc. 2014, 9:2450-63). Briefly, each of the anti-mouse CD25 (PC61; mouse IgG1 isotype, as described above) and the anti-mouse/human PD-L1 (clone S70, also known as Atezolizumab, MPDL3280A, RG7446, or clone YW243.55.570; see WO2010077634 and Herbst R et al., 2014, Nature 515:563-7) is cloned in mammalian expression vectors (504865|UCOE® Expression Vector—Mouse 3.2 kb Puro Set—Novagen) with K409R mutation (for PC61-IgG1) and F405L mutation (for S70-IgG1) in CH3 domain, while light chains are maintained identical, and produced as separate recombinant proteins in mammalian cells. These parental IgG1s are mixed in vitro in equimolar amounts, under permissive redox conditions (e.g. 75 mM 2-MEA; 5h incubation) in order to enable recombination of half-molecules. Following the removal of the reductant to allow reoxidation of interchain disulfide bonds, resulting heteromeric proteins are analysed for exchange efficiency using SDS-PAGE chromatography-based or mass spectrometry-based methods. In the case of Bs CD25 PD-L1, the mass spectrometry has confirmed that the molecular weight of the heterodimeric proteins was 151 Kd, corresponding to the addition of the molecular weight of Clone S70 single Heavy Chain and Light Chain (74 Kd) and PD61-IgG1 single Heavy Chain and Light Chain (77 Kd) and showing that half of each parental IgG1 have been combined in a single molecule.

[0159] The specificity of Bs CD25 PD-L1 has been further confirmed by flow cytometry as described in Example 5, using the parental antibodies as control, and IgG1-recognizing detections antibodies (aHuman, Alexa Fluor® 647, AffiniPure Goat Anti-Human IgG, Fcγ Fragment Specific; Jackson Labs 109-605-098), that are used according to literature and manufacturer's instructions diluted in FACS buffer (PBS+2% FCS+2 mM EDTA). Additional flow cytometry and cell biology materials are fixable viability dye eFluor780 (Ebioscience 65086514), PMA (50 ng/ml; Santa Cruz Biotechnology, sc-3576) and ionomycin (400 ng/ml; Sigma 10634), and Golgi plug protein inhibitor (BD Bioscience, 512301KZ).

[0160] The validation of Bs CD25 PD-L1 in MCA205 models has been performed by using the same approach shown in previous Examples, with isotype control, monospecific antibodies (100 μg each) or bispecific Duobody (200 μg each) administered at day 7 after MCA205 injection and mouse tissue obtained and prepared at day 10.

Example 1—High Expression of CD25 in Treg Makes it a Suitable Target for their Depletion

[0161] The interleukin-2 high affinity receptor alpha (IL2Rα), CD25, has historically been used as a bona fide surface marker of Treg and therefore a target for antibody-mediated Treg depletion. Because there has been controversy as to whether anti-CD25 (αCD25) can also result in elimination of activated effector T cells, the expression of CD25 was analysed in lymphocyte subpopulations in tumours and peripheral lymphoid organs.

[0162] Mice were injected subcutaneously (s.c.) in the flank with MCA205 (5×10.sup.5 cells, C57BL/6 mice), B16 (2.5×10.sup.5 cells, C7BL/6 mice) or CT26 (5×10.sup.5 cells, BALB/c mice) cells and 10 days later the tumours (TIL) and draining lymph nodes were harvested and processed for analysis by flow cytometry.

[0163] We sought to evaluate the relative expression of CD25 by individual T lymphocyte subpopulations within tumours, draining lymph nodes and the blood of tumour-bearing mice 10 days after tumour challenge. The results are shown in FIG. 1. Across different models of transplantable tumour cell lines (including MCA205 sarcoma, MC38 colon adenocarcinoma, B16 melanoma and CT26 colorectal carcinoma), CD25 expression was consistently high in CD4-positive, Foxp3-positive T cells (Treg) and minimal in CD4+Foxp3− and CD8+ T cells (FIG. 1(A)), as has been previously described (Sakaguchi et al. 1995. J Immunol; Shimizu et al. 1999. J Immunol). Because of its immunogenicity and higher T cell infiltration, the effects on Treg depletion in the MCA205 tumour model were studied in more detail (FIG. 1(B-C)). Contrary to in vitro studies, minimal expression of CD25 on the effector compartment (CD4.sup.+FoxP3.sup.− and CD8.sup.+ T cells) was observed in vivo. Although CD25 was slightly upregulated on tumor-infiltrating CD8.sup.+ and CD4.sup.+FoxP3.sup.− T effector cells (Teff). The percentage of CD25-positive cells (3.08%-8.35% CD8+, 14.11-26.87% CD4-positive, Foxp3-negative cells) and the expression levels on a per cell basis (mean fluorescence intensity (MFI) 166.6 in CD8-positive and 134 in CD4-positive, Foxp3-negative cells) were considerably lower than in Treg (83.66-90.23%, MFI 1051.9; p<0.001). Finally, CD25 was also expressed on the Treg present in draining lymph nodes and blood, although the level of expression based on mean fluorescence intensity (MFI) was higher on tumor-infiltrating Treg. The considerably lower expression of CD25 on Teff cells compared to Treg cells indicate that CD25 is a suitable and attractive target for Treg depletion in the tumour where expression levels on Treg are significantly higher.

Example 2—Isotype Swapping is Necessary for the Effective and Safe Intratumoural Treg Depletion with Anti-CD25

[0164] Traditionally, the anti-CD25 antibody (αCD25) clone PC-61 (rat IgG1,κ) (αCD25-r1) has been used for Treg depletion in mouse models, in which it has been repeatedly shown to result in elimination of Treg in peripheral lymphoid organs. To avoid the inter-species differences in FcγR engagement, the constant regions of PC-61 were swapped with the murine IgG2a, κ (αCD25-m2a)—the classical mouse depleting isotype—and the number of Treg both in the periphery and in the tumour were quantified and compared to the effect of anti-CTLA4 (αCTLA4, clone 9H10), which is known to result in depletion of tumour-infiltrating Treg.

[0165] Based on previous evidence demonstrating the importance of intra-tumoral Treg depletion in co-defining the activity of immune modulatory antibodies, we sought to compare the effect of αCD25-r1 on the frequency of Teff and Treg in the blood, draining lymph nodes (LN) and tumour-infiltrating lymphocytes (TILs) in the MCA205 mouse model, because of its higher immunogenicity and for evaluating any potential negative impact of anti-CD25 on activated Teff within tumours.

[0166] Tumour-bearing mice were injected with 200 μg of anti-CD25-r1 (αCD25-r1), anti-CD25-m2a (αCD25-m2a) or anti-CTLA-4 (αCTLA-4) on days 5 and 7 after s.c. inoculation with 5×10.sup.5 MCA205 cells. Peripheral blood mononuclear cells (PBMC), lymph nodes (LN) and tumours (TIL) were harvested on day 9, processed and stained for flow cytometry analysis. The results are shown in FIGS. 2 and 3.

[0167] In vivo administration of αCD25 decreased the number of CD25+ cells in lymph nodes and particularly in blood, independently of the antibody isotype. When quantifying the number of Treg by expression of their signature transcription factor, Foxp3, both isotypes were again equally effective in the periphery but, surprisingly, only the mouse IgG2a isotype resulted in a significant reduction in the frequency and absolute number of tumour-infiltrating Treg to levels comparable to those observed with αCTLA4. Although CD25 expression is upregulated in a small proportion of tumour-infiltrating effector T cells (see Example 1), we observed no significant reduction in the number of CD8+ and CD4+Foxp3− in the periphery or in the tumour. As a consequence, both αCD25 isotypes resulted in an increased Teff/Treg ratio in the periphery. However, only αCD25-m2a increased this ratio in a similar way to anti-CTLA4, which is known to preferentially deplete Treg in the tumour site but not the periphery. This potentially explains the lack of efficacy observed against established tumors in previous studies. Thus, only anti-CD25 (mouse IgG2a) reduces the number of Treg in lymph node and blood and depletes tumour-infiltrating Treg. Importantly, despite a reduction in the number of circulating and LN-resident Treg, no macroscopic, evidence of toxicity was observed in the skin, lungs and liver following multiple doses of αCD25-m2a. This type of anti-CD25 therapy was not associated other major problems due to its toxicity in mice during such experiments, since no statistically relevant differences in the general health status and total body weight, as well in serum levels of lactate dehydrogenase (LDH) and liver enzymes (AST, aspartate aminotransferase; ALT, alanine amino-transferase) were measured among the different treatment groups.

[0168] The expression levels of both activatory and inhibitory FcγRs on different leukocyte subpopulations in the blood, spleen, LN and tumor of mice bearing subcutaneous MCA205 tumors was also determined (FIG. 4). FcγRs appeared more expressed on tumor-infiltrating myeloid cells (granulocytic cells, conventional dendritic cells and monocyte/macrophages), relative to all other studied organs. The binding affinity of the two Fc variants of anti-CD25 to FcγRs was also determined by surface plasmon resonance (Table 1).

TABLE-US-00004 TABLE 1 rIgG1 mIgG2a FcγRI n.b. 1.1 × 10.sup.−8 FcγRIIb 2.6 × 10.sup.−6 4.2 × 10.sup.−6 FcγRIII 2.5 × 10.sup.−6 4.5 × 10.sup.−6 FcγRIV n.b. 2.2 × 10.sup.−7

[0169] These data demonstrate that mIgG2a isotype binds to all FcγR subtypes with a high activatory to inhibitory ratio (A/I). In contrast, the rIgG1 isotype binds with a similar affinity to a single activatory FcγR, FcγRIII, as well as the inhibitory FcγRIIb, resulting in a low A/I ratio (<1).

[0170] The number of tumor-infiltrating Treg in mice lacking expression of different FcγRs was established in different mouse models to distinguish which specific FcγRs were involved in anti-CD25-mediated Treg depletion (FIG. 5). C57BL/6 control mice and Fcer1g−/− mice were injected subcutaneously MCA205 cells and tumours, draining lymph nodes and blood were harvested, processed and stained for flow cytometry analysis. Regulatory T cells were identified by CD4 and FoxP3 expression. Percentage of Foxp3-positive from total CD4-positive cells shows how the anti-CD25 effect is due to the expression of Fcer1g gene. Analysis of Fcer1g.sup.−/− mice, which do not express any of the activating FcγRs (I, Ill and IV), demonstrated a complete absence of Treg depletion. Treg elimination by αCD25-r1 in the periphery and αCD25-m2a in the periphery and tumor therefore results from FcγR-mediated ADCC and not blocking of IL-2 binding to CD25. Depletion by αCD25-m2a was not dependent on any individual activatory FcγR, with Treg elimination maintained in both Fcgr3.sup.−/− and Fcgr4.sup.−/− mice. Thus, depletion of peripheral Treg by αCD25-r1 fails to deplete in the tumor despite high intra-tumoral expression of this receptor. Intra-tumoral Treg depletion is however effectively restored in mice lacking expression of the inhibitory receptor FcγRIIb. In this setting, intra-tumoral Treg depletion is comparable between αCD25-r1 and αCD25-m2a. Therefore, the lack of Treg depletion by αCD25-r1 in the tumor can be explained by its low A/I binding ratio and high intra-tumoral expression of FcγRIIb, which inhibits ADCC mediated by the single activatory receptor engaged by this isotype.

Example 3—Anti-CD25 Therapy Synergizes with Anti-PD-1, Eradicates Established Tumours and Increases Survival of Tumour-Bearing Mice

[0171] Because of its better efficiency in intra-tumoural Treg depletion, it was hypothesized that αCD25-m2a could have a better therapeutic outcome in the treatment of established tumours. The anti-tumor activity of αCD25-m2a and -r1 against established tumours was evaluated by administering a single dose of αCD25 five days after subcutaneous implantation of MCA205 cells, when tumours were established. The results are provided in FIG. 6.

[0172] Consistent with the observed lack of capacity to deplete intra-tumoral Treg, a single dose of αCD25 given to mice with established tumours (day 5) resulted in no protection with αCD25-r1. On the other hand, growth delay and long term survival of mice given αCD25-m2a was observed (15.4%). Because of the clinical relevance of agents targeting the co-inhibitory receptor PD-1 as immunotherapeutic target and PD-1 key role in controlling T cell regulation within the tumor microenvironment, we hypothesized that depletion of CD25.sup.+ Treg cells and PD-1 blockade might be synergistic in combination. In the same model, the combination of αCD25 with PD-1 blockade using anti-PD-1 (αPD-1, clone RMP1-14; at a dose of 100 μg every three days) was tested. αPD-1 as a monotherapy is not effective in the treatment of established MCA205 tumour model and combination with αCD25-r1 did not improve its effect. However, a single dose of αCD25-m2a followed by αPD-1 therapy eradicated established tumours in 78.5% of the mice resulting in long-term survival of more than 100 days. A similar result was observed in MC38 and CT26 tumour models, where αCD25-m2a had a partial therapeutic effect that synergized with αPD-1 therapy, in contrast to combination with αCD25-r1 which failed to deplete tumour-infiltrating Treg in these tumors. Thus, this combined administration allowed an efficient tumour elimination dramatically improved long-term survival of different tumour mice models.

[0173] To understand the mechanism of action underlying the synergism with the αCD25-m2a and αPD-1 combination, we evaluated the phenotype and function of tumor-infiltrating lymphocytes (TILs) present in the MCA205 tumour microenvironment at the end of the treatment protocol, 24 hours after the third dose of αPD-1 (FIG. 7). Monotherapy with αPD-1 did not impact on Teff proliferation nor on the magnitude of Teff infiltration in the tumor, where we also observed a persisting high frequency of Treg (data not shown), and low ratio of Teff/Treg in keeping with the lack of therapeutic activity. Conversely, intra-tumoral Treg depletion with αCD25-m2a resulted in a higher proportion of proliferating and interferon-γ (IFN-γ)-producing CD4.sup.+FoxP3.sup.− T cells in the tumor, corresponding to a high Teff/Treg ratio and anti-tumor response. This effect was further enhanced in combination with anti-PD-1, which yielded even higher proliferation and a 1.6-fold increase in the number of IFN-γ-producing CD4-positive, FoxP3-negative T cells compared to monotherapy with anti-CD25-m2a. In contrast, the observed lack of Treg depletion with anti-CD25-r1 resulted in no change in Teff proliferation or IFN-γ production, when used as monotherapy or in combination with anti-PD-1.

[0174] The data that have been generated with PC61 having either the original mouse IgG1 isotype or the mouse IgG2a isotype that allow efficient Treg depletion suggest the such anti-CD25, alone or in combination of anti-cancer antibodies may be effective at rejecting established tumours, particularly for those tumours requiring efficient intra-tumoural Treg depletion.

[0175] As shown above, the administration of a single dose of αCD25-m2a, followed by αPD-1 therapy had a positive effect on both tumour size and mice survival in the MCA205 murine model. This therapeutic effect due to anti-CD25-m2a/anti-PD1 is dependent from the activity of CD8-positive T cell since the further administration of an anti-CD8 antibody brought tumour size and mice survival to the levels observed in untreated animals (FIG. 8). Thus, the MCA205 tumour elimination depends on the impact of αPD-1/αCD25 synergism on both CD8-positive and Treg cell populations, and that overall effector T cell responses are not negatively impacted by a depleting anti-CD25 antibody.

[0176] Such a synergy was also observed against the poorly immunogenic B16 melanoma tumour model when αCD25-m2a and αPD-1 were combined with Gvax, a GM-CSF-expressing whole tumour cell vaccine (FIG. 9). In this system, neither Gvax therapy alone nor the combination of Gvax with αPD-1 or αCD25-r1 are able to block tumour growth or to extend survival of tumour-bearing mice. In this setting, only the combination of Gvax with αCD25-m2a (alone or together with αPD-1). Such improved was not observed in any treatment group where αCD25-r1 was administered.

[0177] A similar result about the synergism of an immune checkpoint inhibitor with αCD25-m2a was observed in MC38 tumour model both when an αPD-1 (FIG. 10) or an αPD-L1 (FIG. 11) is administered. Also the CT26 tumour models confirmed the therapeutic effect of these combinations (FIGS. 12 and 13). Thus, where αCD25-m2a had already a partial therapeutic effect due to the Treg depletion, this advantageous property give the possibility to surprisingly improve the response to therapies based on immune checkpoint inhibitors.

Example 4—CD25 is Highly Expressed on Treg Infiltrating Human Tumours and Anti-PD-1 Therapy Induces Infiltration of CD25-Expressing Treg in Human Tumours

[0178] CD25 appears an attractive target for Treg depletion and combination immunotherapeutic approaches based on mouse models. In order to validate CD25 as a possible target for Treg depletion in humans, its expression levels in peripheral blood and tumour-infiltrating lymphocytes were compared using biological samples obtained from ovarian cancer, bladder cancer, melanoma, non-small cell carcinoma (NSCS) and renal cell carcinoma (RCC) patients by flow cytometry and immunohistochemistry (IHC). The number of Treg and CD25 expression in tumour samples from patients with RCC before and after receiving αPD-1 therapy with Pembrolizumab were also quantified. Results are shown in FIGS. 14 and 15.

[0179] Independently of the anatomical location, tumour type or stage, it was observed that CD25 expression in Treg is significantly higher (50-85%) than in CD4+Foxp3− (10-15%) and CD8+ (<5%) T cells. Similar to murine models, the level of CD25 expression, as assessed by MFI, was significantly higher on CD4+FoxP3+ Treg relative to CD4+FoxP3− and CD8+ Teff within all studied tumour subtypes.

[0180] These observations were further supported by multiplex immunohistochemistry (IHC). Analysis of melanoma, NSCLC and RCC tumours from the same patient cohorts demonstrated that even in areas of dense CD8-positive, T cell infiltrate, CD25 expression remained restricted to FoxP3-positive cells. Strikingly, this expression profile remained consistent, regardless of tumour subtype, stage, resection site, current or prior therapy and they are in keeping with the data obtained in mouse models.

[0181] In addition, in contrast with the high proportion of Treg observed in subcutaneous murine tumours, RCC samples showed a scarce number of Treg in untreated tumours. However, anti-PD-1 therapy resulted in a significant infiltration both by CD8+ T cells and Treg (Foxp3-positive cells). Moreover, data generated from melanoma and RCC samples confirm that CD25 was highly expressed by Foxp3-positive cells, while its expression was minimal in Foxp3-negative, CD8-positive cells.

[0182] When CD25 expression is assessed in the context of therapeutic immune modulation. Core biopsies were performed on the same lesion at baseline and following 4 cycles of either Nivolumab or 2 cycles of Pembrolizumab) in patients with advanced kidney cancer and melanoma respectively. Despite systemic immune modulation, CD25 expression remained restricted to FoxP3-positive Treg, even in areas of dense CD8-positive T cell infiltrate evaluated by multiplex IHC.

[0183] These findings confirm the translational value of the described pre-clinical data and lend further support to the concept of selective therapeutic targeting of Treg via CD25 in human cancers. Moreover, CD25 expression profiles in human solid cancers in connection with anti-PD1 treatment provides a rationale for the therapeutic combination of anti-human CD25 antibodies having CD25 binding and Fcgamma receptor specificity comparably to those shown for anti-mouse CD25 PC61(IgG2a) with immune checkpoint inhibitors such as a PD-1 antagonist.

Example 5—Anti-CD25- and Anti-PD-L1-Based Bispecific Antibodies and Combination of Antibodies Present Efficient Treg Depletion and Cytokine Inducing Properties

[0184] The previous examples have shown that the Treg depleting, CD25-binding properties of antibodies based on PC61 and with appropriate isotype can be exploited in combination with other anti-cancer compounds such as antibodies targeting immune checkpoint proteins such as a PD-1 antagonist (being an anti-PD-1 or an anti-PD-L1 antibody). These findings suggest the construction of bispecific antibodies combining the two antigen-binding properties and the relevant isotype (e.g. IgG1).

[0185] This approach has been validated by using Duobody technology that allows the efficient association of single Heavy and light chain from two distinct monospecific antibodies that are produced separately, within a single heteromeric protein that is named Bs CD25 PD-L1 (FIG. 16A). The binding specificity of this antibody has been validated using two genetically modified human cell lines, each expressing either mouse CD25 or mouse PD-L1, and compared with those of the initial monospecific antibodies (FIGS. 16B and C). These cell lines have been tested by flow cytometry, separately or mixed in equivalent amounts, showing that Bs CD25 PD-L1 retains its dual CD25, PD-L1 specificity, even allowing detecting complex of double positive cell complexes that formed by binding of Bs CD25 PD-L1 to both CD25-positive and PD-L1-positive cells at the same time.

[0186] The functional properties of BsAb CD25 PD-L1 have been evaluated in vivo by using models of cell interaction and depletion that were used for validating PC61 in previous Examples. The MCA205 model was used for evaluating the impact of the BsAb on effector and regulatory T cells in tumour and LN. In this model, BsAb CD25 PD-L1 can recognize and deplete CD4positive, Foxp3positive regulatory T cells and increase the CD8-positive, Foxp3-positive regulatory T cells ratio in tumours and LN with equivalent efficacy to anti-CD25 (PC61-m2a) or combination of monospecific anti-CD25 and anti-PD-L1 antibodies (FIG. 17AB). Moreover BsAb CD25 PD-L1 increase the number of Interferon gamma expressing, CD4-positive, CD5-positive cells at a level that is at least similar to the combination of monospecific anti-CD25 and anti-PD-L1, and possibly superior to the one of anti-CD25 m2a antibody alone (FIG. 17C).

[0187] The data shows how the use of a PC61-based, Treg depleting, anti-human CD25 antibody for treating cancer can be not only improved by selecting the appropriate isotype but also efficiently combined with other anti-cancer drug, in particular with anti-cancer antibodies that bind to a different cell surface antigen. This approach can be pursued by producing and administering the two products as a novel mixture of monospecific antibodies or as novel bispecific antibodies that are associated and produced in order to maintain the Treg depleting, CD25 binding and other binding properties of the parent monoclonal antibodies.

[0188] All documents referred to herein are hereby incorporated by reference in their entirety, with special attention to the subject matter for which they are referred Various modifications and variations of the described methods and system of the invention will be apparent to those skilled in the art without departing from the scope and spirit of the invention. Although the invention has been described in connection with specific preferred embodiments, it should be understood that the invention as claimed should not be unduly limited to such specific embodiments. Indeed, various modifications of the described modes for carrying out the invention which are obvious to those skilled in molecular biology, cellular immunology or related fields are intended to be within the scope of the following claims.