Multispecific antibody

Abstract

To provide a medicament to improve treatment of a proliferative disease, particularly a cancer, and also to address the need for high precision therapeutics for immuno-oncology that target only the disease-related cells.SOLUTION: The present invention relates to a multispecific antibody comprising at least one domain specifically binding to a tumor-associated immune checkpoint antigen with low affinity, and at least one domain specifically binding to a tumor-associated antigen (TAA), and pharmaceutical compositions and methods of use thereof. The present invention further relates to a nucleic acid encoding the multispecific antibody, a vector comprising the nucleic acid, a host cell comprising the nucleic acid or the vector, and a method of producing the multispecific antibody.SELECTED DRAWING: Figure 13

Description

[Technical field]

[0001] The present invention relates to a multispecific antibody comprising at least one domain that specifically binds with low affinity to a tumor-associated immune checkpoint antigen, at least one domain that specifically binds to a tumor-associated antigen (TAA) and optionally at least one domain that specifically binds to an immune cell antigen. Furthermore, the present invention relates to specific domains for use in said multispecific antibody, as well as pharmaceutical compositions and methods of use thereof. The present invention further relates to nucleic acids encoding said multispecific antibody or its specific domains, vectors comprising said nucleic acids, host cells comprising said nucleic acids or said vectors, and methods of making said multispecific antibody or its specific domains. [Background technology]

[0002] Despite considerable progress in its treatment, cancer remains a major unmet medical need. Some of the most substantial advances in cancer treatment in recent years have come with the advent of various molecular classes of immunotherapies, including but not limited to monoclonal antibodies (mAbs), bispecific antibodies (bsAbs), recombinant proteins, and chimeric antigen receptor-T cell (CAR-T cell) therapies. Such therapies induce antitumor immunity by a) actively attracting immune effector cells to tumor-resident cells, and / or b) stimulating immune effector cells, and / or c) alleviating tumor-mediated immune suppression. In general, these immunotherapies exploit the overexpression of specific antigens by tumor-resident cells (e.g., malignant cells, cells of the tumor vasculature, stromal cells, immune cells, etc.) compared to extratumoral sites to direct their pharmacological activity to the tumor. Among these antigens, tumor-associated antigens (TAAs) include cell surface proteins that are selectively overexpressed on malignant cells. By binding to TAAs with high affinity, immunotherapies can limit their immunomodulatory activity to some extent at the immune synapse between tumor cells and immune effector cells.

[0003] A common class of TAA-conjugated immunotherapies are mAbs that induce antitumor immunity by opsonizing tumor cells and triggering antibody-dependent cell-mediated cytotoxicity (ADCC) by Fcγ receptor (FcγR)-expressing cells, primarily natural killer (NK) cells. Similar to bsAbs that simultaneously bind the T cell antigen CD3 (TAA / CD3 bsAb), other TAA-conjugated immunotherapies also utilize cytotoxic T lymphocytes (CTLs) to induce targeted depletion of malignant cells, such as CAR-T cells. Although the therapeutic utility of TAA (re)directed CTLs has been clinically validated, such utility may be limited when tumor-mediated immune suppression impairs CTL activation / stimulation. Even in tumors with high abundance of tumor-infiltrating lymphocytes (TILs) (i.e., "inflammatory" or "hot" tumors), tumor immune evasion can be induced by various means, such as through expression of immune checkpoint ligands / receptors (e.g., PD-1, PD-L1, CTLA-4) and recruitment of regulatory T cells (Tregs) and myeloid-derived suppressor cells (MDSCs).

[0004] Immune checkpoints are regulators of the immune system and are involved in processes such as self-tolerance and immunosuppression in cancer.

[0005] PD-L1 (CD274, B7-H1) is a 40 kDa type I transmembrane protein. It is expressed by activated T cells and B cells and is the surface glycoprotein ligand for PD-1, a key immune checkpoint receptor that mediates immunosuppression. PD-L1 is involved in suppressing immune system responses during chronic infections, pregnancy, tissue transplantation, autoimmune diseases, and cancer. PD-L1 is found on both antigen-presenting cells and human cancer cells, including head and neck squamous cell carcinoma, melanoma, brain tumors, thyroid, thymus, esophagus, lung, breast, gastrointestinal tract, colon, liver, pancreas, kidney, adrenal cortex, bladder, urothelium, ovary, and skin (Katsuya Y, et al., Lung Cancer.88(2):154-159 (2015); Nakanishi J, et al., Cancer Immunol Immunother. 56(8):1173-1182 (2007); Nomi T, et al., Clin Cancer Res. 13(7):2151-2157 (2007); Fay AP, et al., J Immunother Cancer. 3:3 (2015); Strome SE, et al., Cancer Res. 63(19):6501-6505 (2003); Jacobs JF, et al. Neuro Oncol.11(4):394-402 (2009), Wilmotte R, et al. Neuroreport. 16(10):1081-1085 (2005)). PD-L1 is rarely expressed on normal tissues, but is inducibly expressed on tumor sites (Dong H, et al., Nat Med. 8(8):793-800 (2002), Wang et al., Onco Targets Ther. 9: 5023-5039 (2016)). PD-L1 downregulates T cell activation and cytokine secretion by binding to PD-1 (Freeman et al., 2000, Latchman et al, 2001). PD-1, when activated by PD-L1, may provide an immune-permissive environment for tumor development and growth.PD-L1 also negatively regulates T cell function through interaction with another receptor, B7.1 (B7-1, CD80).

[0006] Inhibition of PD-L1 / PD-1 interaction allows potent antitumor activity. Numerous antibodies that disrupt PD-1 signaling are in clinical development. These antibodies belong to two main categories: those targeting PD-1 (nivolumab, Bristol-Myers Squibb; pembrolizumab, Merck, Whitehorse Station, NJ; pidilizumab, CureTech, Yavne, Israel) and those targeting PD-L1 (MPDL3280A, Genentech, South San Francisco, CA; MEDI4736, MedImmune / AstraZeneca; BMS-936559, Bristol-Myers Squibb; MSB0010718C, EMD Serono, Rockland, MA) (for review, see Postow MA et al., J Clin Oncol. Jun 10;33(17):1974-82 (2015)). Targeting PD-L1 may have different biological effects as opposed to targeting PD-1. PD-1 antibodies prevent the interaction of PD-1 with both its ligands, PD-L1 and PD-L2. The effect of PD-1 interaction with PDL2 is not yet known, but PD-L1 antibodies do not prevent that interaction. However, PD-L1 antibodies inhibit not only the interaction with PD-1, but also the interaction of PD-L1 with B7-1 (Butte MJ, et al., Immunity 27:111-122, (2007)), which is thought to exert a negative signal on T cells. Blockade of PD-L1 has shown promising early data, and currently four clinical anti-PD-L1 mAbs are in testing: Atezolizumab and MEDI4736 (both Fc null variants of human IgG1), MSB001078C (IgG1), and BMS-936559 (IgG4) (Chester C., et al., Cancer Immunol Immunother Oct;65(10):1243-8 (2016)).

[0007] New and emerging treatments often combine TAA-targeted immunotherapy with one or more additional immunotherapies targeting immune checkpoint pathways to further alleviate or overcome tumor-mediated immune suppression. Monoclonal antibodies that block immune suppressive antigens such as CTLA-4 (e.g., ipiriumab), PD-1 (e.g., nivolumab, pembrolizumab), and PD-L1 (e.g., avelumab, atezolizumab) have elicited impressive response rates in patients with various tumor histologies. Early results of combined treatment with immune checkpoint regulators and TAA-conjugated immunotherapies have been promising. For example, the HER2-targeted mAb trastuzumab (Herceptin®, Genentech Genentech) is currently undergoing clinical evaluation (Phase II) in combination with nivolumab (Opdivo®, Bristol-Myers Squibb) and in combination with nivolumab and ipilimumab (Yervoy®, Bristol-Myers Squibb) (US National Clinical Trial (NCT) 03409848). Similarly, the CD19 / CD3 bsAb blinatumomab (Blincyto®, Amgen) is currently undergoing clinical evaluation (Phase I / II) in combination with pembrolizumab (Keytruda®, Merck) and is in Phase I trials as part of a triple immunotherapy combination with both nivolumab and ipilimumab (NCT03512405 and NCT02879695, respectively).

[0008] In addition, combination therapies targeting immune checkpoints and T cell costimulatory receptors are being evaluated. The combination of anti-PD-L1 and anti-CD137 antibodies improved overall survival and enhanced T cell effector function in the ID-8 ovarian adenocarcinoma model (Duraiswamy J, et al., Cancer Res 73:6900-6912 (2013)). The combination of urelumab (anti-CD137) and nivolumab (anti-PD-1) is being tested in a Phase I / II study in both solid tumors and B-cell non-Hodgkin lymphoma (NCT02253992), while PF-05082566 (anti-CD137) is being tested in patients with solid tumors in a Phase Ib trial with pembrolizumab (anti-PD-1) (NCT02179918) (Chester C., et al., Cancer Immunol Immunother Oct;65(10):1243-8 (2016)).

[0009] Recently, multivalent and multispecific fusion polypeptides that bind PD-L1 and CD137 have been evaluated for their effects on T cell activation and proliferation in vitro. Using an autologous in vitro co-culture system transduced with immature DCs and donor-matched T cells, it has been shown that the multispecific, multivalent polypeptide INBRX-105, which has two PD-L1-binding domains, two CD137-binding domains and an Fc region, is superior in stimulating interferon-gamma production compared to the monospecific PD-L1 sd-Ab-Fc fusion protein, the CD137 sdAb-Fc fusion protein, a combination of the two, the anti-PD-L1 antibody atezolizumab, the anti-CD137 antibody utomirumab (PF-05082566), or the anti-PD-L1 antibody pembrolizumab, and combinations thereof, in inducing IFNγ or mediating the proliferation and activation of CD8+ T cells (WO 2017 / 123650). Furthermore, WO 2016 / 149201 discloses certain antibodies against PD-L1 and suggests creating bispecific antibody constructs further comprising a T cell engaging antibody, with CD137 in a non-exclusive list of over 20 potential T cell targets.

[0010] Immunotherapy combinations have shown the potential to enhance antitumor responses through additive or synergistic activity, but have always been subject to two consistent limitations: 1) clinical development challenges due to the complexity of adjusting the dosing of multiple constituent therapies across different patient cohorts, and 2) reliance on two or more separate manufacturing processes for the constituent therapies with attendant high cost of goods (COGS) and pricing implications. These limitations become more severe as the number of immunotherapies included in the combination regimen increases. Moreover, even in the case of treatment regimens that include a single immunotherapy, dose-limiting toxicities (DLTs) often preclude administration at the maximally effective dose (MED) or lead to treatment interruption, resulting in limited efficacy. Unfortunately, like their antitumor activity, the drug-related toxicities induced by each constituent immunotherapy in a combination regimen also tend to be additive or synergistic.

[0011] Thus, despite the promising opportunities offered by inhibiting the interaction between PD-1 and PD-L1, the above applications have often resulted in toxicity caused by binding of anti-PD-L1 antibodies to PD-L1 expressed on non-target cells (for review, see Wang et al., Cancer J. 24 (2018) 36-40).

[0012] Although the exact pathways by which such DLTs arise may vary, the risk of immunotherapy-related toxicity can typically be minimized or eliminated by enhancing tumor localization of pharmacological activity. Extratumoral activity of immunotherapy can result in the secretion of proinflammatory cytokines in healthy tissues, resulting in an undesirable safety profile. Utilizing T cell-guiding bsAbs that require binding to TAAs to elicit immunomodulatory activity is a promising strategy to restrict such cytokine release to the cytolytic / immune synapse between tumor-resident cells and T cells. However, conventional TAA / CD3 bsAbs are also associated with toxicities such as cytokine release syndrome (CRS), usually presumed to result from excessive activity of the anti-CD3 domain. Moreover, TAA / CD3 bsAbs potently deplete TAA-overexpressing cells, but do so by recruiting and stimulating CTLs, regardless of whether such cells express T cell receptors (TCRs) that recognize tumor antigens (i.e., are tumor-reactive T cells). Thus, rather than stimulating or reactivating the host's innate antitumor immunity, TAA / CD3 bsAbs may stimulate CTLs in a somewhat indiscriminate manner, potentially posing safety risks and leading to inadequate anticancer immune memory formation.

[0013] In addition to CD3, T cell costimulatory receptors (e.g., 4-1BB, OX40, ICOS, GITR) are currently being clinically evaluated as targets for therapeutic stimulation of T cells in cancer. One advantage of antitumor T cell stimulation via such targets is presumed to be their transient expression upon TCR signaling. Thus, their expression tends to be selectively upregulated in the inflammatory TME, especially in tumor-reactive T cells, whose TCRs are constantly stimulated by the abundant interactions with major histocompatibility complexes (MHC) expressed by malignant cells and antigen-presenting cells (APCs). Thus, targeting costimulatory receptors, for example with mAbs and bsAbs, should more selectively stimulate and expand existing antitumor T cells than approaches targeting CD3, potentially making such biologics safer and more durable in their effects.

[0014] Among the costimulatory receptors, 4-1BB (CD137, TNF-receptor superfamily 9, TNFRSF9) has emerged as particularly promising due to its expression profile and role as a pluripotent mediator of antitumor immunity (Bartkowiak and Curran 2015; Yonezawa et al. 2015). 4-1BB is an inducible T cell costimulatory receptor. Its expression is activation-dependent and includes a broad subset of immune cells, including activated CD8+ T cells, CD4+ T cells, NK and NKT cells, Tregs, dendritic cells (DCs), including follicular DCs, stimulated mast cells, differentiated myeloid cells, monocytes, neutrophils, eosinophils (Wang et al, Immunol Rev. 229(1): 192-215 (2009)), and activated B cells (Zhang et al, J Immunol. 184(2):787-795 (2010)). Furthermore, 4-1BB expression has also been demonstrated in tumor vasculature (Broil K et al., Am J Clin Pathol. 115(4):543-549 (2001); Seaman et al, Cancer Cell 11(6):539-554 (2007)) and atherosclerotic endothelium (Olofsson et al, Circulation 117(10): 1292 1301 (2008)).

[0015] 4-1BB costimulates T cells to perform effector functions such as eradication of established tumors, expansion of primary CD8+ T cell responses, and enhancement of the memory pool of antigen-specific CD8+ T cells. In vivo efficacy studies in mice have demonstrated that 4-1BB agonist mAbs, administered either as monotherapy or as a component of combination regimens, induce antitumor-protective T cell memory responses and tumor regression in multiple tumor models. In addition, two 4-1BB agonist mAbs are currently in clinical use: urulumab (PF-05082566, Pfizer), a fully humanized IgG4 mAb, and utomilumab, a fully human IgG2 mAb (Chester C., et al., Cancer Immunol Immunother Oct;65 (10) :1243-8 (2016)). Although the use of 4-1BB agonist mAbs is a very promising therapeutic strategy, the clinical data gathered so far suggest that approaches based on 4-1BB-stimulating mAbs are subject to a trade-off between efficacy and safety: highly active 4-1BB agonist mAbs induce DLTs that attenuate therapeutic efficacy, whereas less active 4-1BB agonist mAbs, although well tolerated, appear to be less effective, including at their predicted MED.

[0016] Highly active 4-1BB agonist mAbs result in alterations of the immune system and organ function, increasing the risk of toxicity. High doses of such mAbs in naive and tumor-bearing mice have been reported to induce T cell infiltration into the liver and elevations in aspartate aminotransferase and alanine aminotransferase, consistent with liver inflammation (Niu L, et al. J Immunol 178 (7) :4194-4213 (2007); Dubrot J, et al., Int J Cancer 128 (1) :105-118 (2011)). Early clinical studies of human therapeutic use of 4-1BB agonist mAbs also showed increased incidence of liver enzyme elevations and hepatitis (Sznol M., et al., J Clin Oncol 26(115S):3007 (2008); Ascierto PA, et al., Semin Oncol 37(5):508-516 (2010); Chester C., et al, Cancer Immunol Immunother Oct;65(10):1243-8(2016)). Potentially fatal hepatitis was observed in a Bristol-Myers Squibb (BMS) phase II anti-CD137 trial (US Clinical Trial (NCT) 00612664) for pre-treated stage III / IV melanoma. That trial and several others (NCT00803374, NCT00309023, NCT00461110, NCT00351325) were stopped due to adverse events (Chester C., et al., Cancer Immunol Immunother Oct;65(10):1243-8 (2016)), most probably due to systemic overstimulation of T cells.

[0017] Similar to TAA / CD3 bsAbs, TAA / 4-1BB bsAbs are designed to selectively agonize 4-1BB in the context of the immune synapse between tumor-resident cells and immune effector cells, thereby preventing toxicity associated with extratumoral T cell stimulation. As an example, 5T4 / 4-1BB bsAb (APV-527) (WO 2017182672(A1)), which is being co-developed by Aptevo Therapeutics and Alligator Biosciences, is designed to induce targeted costimulation of T cells by locking onto 5T4, a TAA expressed by a variety of solid tumors. Preclinical data for APV-527 suggest that conditional stimulation of 4-1BB in the presence of 5T4 effectively localizes T cell costimulation to the tumor, leading to a substantial enhancement of T cell activation in the TME and inhibiting tumor growth in 5T4+ tumor models. This same tumor localization strategy could potentially leverage a variety of clinically validated TAAs that have demonstrated therapeutic targeting to be effective and safe.

[0018] HER2 has been established as a TAA that can be effectively and safely targeted to address HER2+ cancers. The most notable HER2-targeted therapies approved for use in patients with HER2+ tumors are the mAbs trastuzumab (Herceptin®, Genentech) and pertuzumab (Perjeta®, Genentech). Trastuzumab and pertuzumab are similar in that they act in part by opsonizing HER2+ cells and inducing ADCC, but the two antibodies differ in the means by which they inhibit growth-promoting HER2 signaling. In the case of trastuzumab, binding to its epitope prevents HER2 homodimerization, thereby inhibiting HER2 signaling. However, in some patients, compensatory HER3 overexpression and formation of HER2 / HER3 heterodimers lead to enhanced signaling, rendering such patients refractory to treatment with trastuzumab. On the other hand, pertuzumab binds to an epitope that inhibits HER2 / HER3 heterodimerization, similarly inhibiting growth-promoting signaling. Due to this complementary mechanism of action (MoA), pertuzumab and trastuzumab are synergistic, and their combination is approved for the treatment of HER2+ breast cancer.

[0019] While the combination of HER2 signaling inhibition and ADCC-mediated HER2+ cell depletion is effective in many patients, many others display a HER2+ tumor phenotype that responds very poorly to treatment with conventional antibodies. This is because, in some cases, certain HER2+ tumors do not depend on HER2 signaling for growth, rendering the primary mechanism of action of trastuzumab / pertuzumab ineffective. This has given rise to the hypothesis that a more potent targeted cytotoxic approach than ADCC could be highly beneficial. The validity of this concept has been somewhat clarified by the marketing approval of the ADC trastuzumab-emtansine (Kadcyla®, Genentech). In line with this, several companies are currently developing HER2 / CD3 bsAbs that stimulate redirected T cells to induce potent targeted cytotoxicity. Moreover, in some patients who are primarily or secondarily non-responsive to HER2-targeted mAbs, the HER2+ tumor phenotype includes increased expression of ligands / receptors (e.g., PD-L1) that actively suppress antitumor immune responses. As expected, this has led to the combination of HER2-targeting mAbs with mAbs modulating immune checkpoints, which have had some success in the clinic. TAAs that are almost exclusively expressed on cancer cells, such as oncofetal tumor antigens, are called clean TAAs. TAAs that are also expressed on normal non-cancer cells (typically at lower expression levels than on cancer cells) are non-clean TAAs. Non-clean TAAs are problematic because the TAA / CD3 bsAb approach, due to its very high potency, also leads to the depletion of non-tumor cells expressing the TAA. A well-known example of a non-clean TAA is HER2, which is expressed not only on tumor cells but also, albeit at low levels, in a variety of other tissues. Thus, novel therapeutic approaches that improve the selectivity of the TAA / CD3 bsAb approach for tumor tissues are needed.

[0020] There is also precedent for using HER2 as a target to localize 4-1BB stimulation by bispecific molecules to the tumor. Pieris Pharmaceuticals has initiated a clinical trial to evaluate a HER2 / 4-1BB bispecific fusion protein (PRS-343) (NCT03330561). PRS-343 contains an IgG4 variant of trastuzumab fused to a bivalent 4-1BB-binding anticalin. Preclinical and clinical evidence supporting 1) the potential benefit of PD-(L)1 blockade and 4-1BB stimulation, 2) the benefit of combining HER2-targeted immunotherapy with PD-(L)1 blocking immunotherapy, and 3) the synergistic effects of trastuzumab and pertuzumab suggests that it may be beneficial to combine such a HER2 / 4-1BB bispecific molecule with up to two booster immunotherapies in a single treatment. Indeed, PRS-343 is also currently being clinically evaluated in combination with the PD-L1 blocking mAb atezolizumab (Tecentriq®, Genentech) (NCT03650348).

[0021] As mentioned above, an unavoidable drawback of combination therapies is that their clinical development can be burdensomely complex and therefore expensive, especially as the number of constituent therapies increases. The need to develop multiple manufacturing processes further increases development costs and increases COGS. The inclusion of more than two specificities in one molecule (e.g., tri- or tetra-specific antibodies) could theoretically address many of the aforementioned limitations regarding safety, efficacy, and cost. Tri- / tetra-specific molecules targeting TAAs could theoretically be highly localized to tumors and induce synergistic anti-tumor modulation of multiple immune checkpoint pathways, providing safer and more effective treatments for various cancers. Moreover, such molecules would further limit the need for co-administration of additional immunotherapies to enhance patient responses, supporting ease of development and minimization of treatment costs. However, the practical use of tri- / tetra-specific antibodies for therapeutic applications has been complicated due to issues with their molecular structure, the properties of the constituent antigen-binding domains, and / or poor biophysical properties. Thus, there remains a clear need for novel tri- / tetra-specific antibodies that localize to tumors, exert synergistic immunomodulatory effects, and possess biophysical properties suitable for drug development.

[0022] Furthermore, despite the fact that numerous antibodies specific for tumor-associated immune checkpoint antigens, TAAs and / or T-cell costimulatory receptors already exist, the complex and specific requirements of such tri- or tetraspecific antibodies necessitate the development of novel antibody domains with tailor-made properties.

[0023] Thus, although there are many treatment options for cancer patients, there is still a need for effective and safe therapeutic agents and for their preferential use in a more targeted manner.Immunomodulatory biologics provide a promising approach in the treatment of cancer due to their mechanism of action, but the lack of global immune stimulation and the restriction of this immune modulation to pathologically relevant cells and sites may cause many side effects and significant toxicity, leading to increased morbidity and mortality of patients.Therefore, the object of the present invention is to provide a pharmaceutical agent for improving the treatment of proliferative diseases, particularly cancer. Summary of the Invention

[0024] The object of the present invention is to provide a medicament for improving the treatment of proliferative diseases, in particular cancer. The present invention addresses the need for precision therapeutics for immuno-oncology that target only disease-related cells.

[0025] In one aspect, the present invention relates to a multispecific antibody comprising at least a first domain that specifically binds with low affinity to a tumor-associated immune checkpoint antigen and at least a second domain that specifically binds to a tumor-associated antigen (TAA).

[0026] The present invention further relates to multispecific antibodies comprising at least a first domain that specifically binds with low affinity to a tumor-associated immune checkpoint antigen, at least a second domain that specifically binds to a tumor-associated antigen (TAA), and at least a third domain that specifically binds to an immune cell antigen, in particular where said immune cell antigen is present on T cells or NK cells.

[0027] More specifically, the present invention relates to a multispecific antibody, wherein a first domain that specifically binds to PD-L1 comprises the VH sequence of SEQ ID NO:11 and the VL sequence of SEQ ID NO:16.

[0028] The present invention further relates to a combination comprising (i) a multispecific antibody of the invention and (ii) a second compound selected from (iia) an antibody against a TAA, in particular an antibody against HER2, in particular an antibody which is trastuzumab, (iib) a modulator of an immune checkpoint antigen which is in particular not a tumor-associated immune checkpoint antigen and / or which is in particular present on T cells or NK cells, and (iic) a modulator of angiogenesis.

[0029] In another aspect, the invention relates to a pharmaceutical composition comprising the multispecific antibody of the invention and a pharma- ceutically acceptable carrier.

[0030] In a further aspect, the present invention provides a multispecific antibody of the invention or a pharmaceutical composition of the invention for use as a medicament.

[0031] In a further aspect, the present invention provides a multispecific antibody of the invention or a pharmaceutical composition of the invention for use in the treatment of cancer in a subject in need thereof.

[0032] In one aspect, the present invention provides the use of a multispecific antibody of the invention or a pharmaceutical composition of the invention for treating cancer in a subject in need thereof.

[0033] In one aspect, the present invention provides the use of a multispecific antibody of the invention or a pharmaceutical composition of the invention in the manufacture of a medicament for the treatment of cancer in a subject in need thereof.

[0034] In yet another aspect, the present invention provides a method for treating cancer in a subject in need thereof, comprising administering to the subject a therapeutically effective amount of a multispecific antibody of the invention or a pharmaceutical composition of the invention.

[0035] In a further aspect, the invention provides a nucleic acid comprising a nucleotide sequence encoding a multispecific antibody of the invention. In another aspect, the invention provides a vector comprising said nucleic acid. In another aspect, the invention provides a host cell comprising said nucleic acid or said vector.

[0036] In yet another aspect, the invention provides a method of making a multispecific antibody of the invention or a binding domain or fragment thereof, the method comprising culturing a host cell comprising a nucleic acid or vector encoding a multispecific antibody of the invention or a binding domain or fragment thereof.

[0037] The aspects, advantageous features and preferred embodiments of the present invention summarized in the following items each alone or in combination further contribute to solving the object of the present invention.

[0038] 1. A multispecific antibody comprising: (a) a first domain which specifically binds to a tumor-associated immune checkpoint antigen, in each case in scFv format (monovalent affinity), and which binds to said tumor-associated immune checkpoint antigen with a dissociation constant (KD) of more than 50 nM, in particular with a dissociation constant between 50 nM and 1 μM, in particular with a dissociation constant of more than 100 nM, in particular with a dissociation constant between 100 nM and 900 nM, in particular with a dissociation constant of more than 200 nM, in particular with a dissociation constant of more than 200 nM, in particular with a dissociation constant between 100 nM and 9 ... a first domain which binds to the ribosome with a dissociation constant of between 200 nM and 800 nM, particularly with a dissociation constant of greater than 300 nM, particularly with a dissociation constant of between 300 nM and 700 nM, particularly with a dissociation constant of greater than 400 nM, particularly with a dissociation constant of between 400 nM and 600 nM, particularly with a dissociation constant of greater than 450 nM, particularly with a dissociation constant of between 450 nM and 550 nM, particularly with a dissociation constant of greater than 475 nM, particularly with a dissociation constant of between 475 nM and 525 nM, particularly with a dissociation constant of about 500 nM (KD); (b) a second domain that specifically binds to a tumor-associated antigen (TAA); 2. A multispecific antibody comprising:

[0039] 2. The multispecific antibody of item 1, wherein said second domain binds to said TAA in scFv format (monovalent affinity) with a dissociation constant (KD) of less than 50 nM, particularly less than 20 nM, particularly less than 10 nM, particularly less than 5 nM, particularly less than 2 nM, particularly less than 1 nM, particularly less than 0.5 nM, measured by SPR.

[0040] 3. The multispecific antibody of item 1 or item 2, wherein said tumor-associated immune checkpoint antigen and said TAA are both present on the same tumor cell.

[0041] 4. The multispecific antibody of any one of items 1 to 3, wherein said tumor-associated immune checkpoint antigen is selected from the group consisting of PD-L1, PD-L2, CD80, CD86, CD276 (B7-H3), and VTCN1 (B7-H4).

[0042] 5. The multispecific antibody of item 4, wherein the tumor-associated immune checkpoint antigen is PD-L1.

[0043] 6. The multispecific antibody of any one of items 1 to 5, wherein said first domain is an inhibitor of said tumor-associated immune checkpoint antigen.

[0044] 7. The multispecific antibody of any one of items 1 to 6, wherein said TAA is not PD-L1.

[0045] 8. The TAA is selected from the group consisting of EGFRvIII, 5T4, CD19, CD20, CD22, CD38, BCMA, IL4RA, mesothelin, GD2, Tn antigen, sTn antigen, Tn-O-glycopeptide, sTn-O-glycopeptide, PSMA, CD97, TAG72, CD44v6, CEA, EPCAM, KIT, IL-13Ra2, leguman, GD3, CD171, IL-11Ra, IL-13RA2, ROR1, PSCA, MAD-CT-1, MAD-CT-2, V EGFR2, CLEC12A, Lewis Y, CD24, PDGFR-β, SSEA-4, folate receptor α, ERBB (e.g., ERBB2), Her2 / neu (HER2), MUC1, MUC16, EGFR, NCAM, ephrin B2, CAIX, LMP2, sLe, HMWMAA, o-acetyl-GD2, folate receptor β, TEM1 / CD248, CD33, CD123, CD133, CD135, TEM7R, FAP, legumain, HPV 8. The polyspecific antibody of any one of items 1 to 7 selected from the group consisting of E6 or E7, ML-IAP, CLDN6, TSHR, GPRC5D, ALK, polysialic acid, Fos-related antigen, neutrophil elastase, TRP-2, CYP1B1, sperm protein 17, beta human chorionic gonadotropin, AFP, thyroglobulin, PLAC1, globo H, RAGE1, MN-CA IX, human telomerase reverse transcriptase, intestinal carboxylesterase, mut hsp70-2, NA-17, NY-BR-1, UPK2, HAVCR1, ADRB3, PANX3, NY-ESO-1, GPR20, Ly6k, OR51E2, TARP, GFRα4, GPC3, CDH3, B7H3, FGFR1, SSTR2, CECAM6, GA733, and gp120.

[0046] 9. The multispecific antibody of item 8, wherein said TAA is selected from HER2 and mesothelin, in particular HER2.

[0047] 10. The multispecific antibody of any one of items 1 to 8, further comprising a third domain that specifically binds to an immune cell antigen, in particular wherein said immune cell antigen is present on a T cell or a NK cell.

[0048] 11. The multispecific antibody of item 10, wherein said third domain specifically binds to an immune cell antigen that is a stimulatory molecule or a costimulatory molecule of said immune cell.

[0049] 12. The multispecific antibody of item 11, wherein the third domain is an agonist and the immune cell antigen is a stimulatory immune cell antigen.

[0050] 13. The multispecific antibody of item 12, wherein said stimulatory immune cell antigen is selected from the group consisting of CD3 and CD16.

[0051] 14. The multispecific antibody of item 13, wherein the stimulatory immune cell antigen is CD3, in particular CD3ε.

[0052] 15. The multispecific antibody of item 11, wherein the third domain is an agonist and the immune cell antigen is a costimulatory immune cell antigen.

[0053] 16. The multispecific antibody of item 15, wherein said costimulatory immune cell antigen is selected from the group consisting of CD137, CD28, ICOS, HVEM, CD27, OX40, DR3, GITR, CD30, SLAM, CD2, 2B4, TIM1, TIM2, and CD226.

[0054] 17. The multispecific antibody of item 16, wherein the stimulatory immune cell antigen is CD137.

[0055] 18. The multispecific antibody of item 17, wherein the third domain specifically binds to CD137 at an epitope comprised in the distal part of the extracellular domain of CD137, in particular within the cysteine-rich domains CRD1 and / or CRD2, more particularly within amino acid residues 24-86 of SEQ ID NO: 153, with the proviso that amino acid residue Asn42 of CD137 is not a critical residue for binding.

[0056] 19. The multispecific antibody of item 11, wherein the third domain is an inhibitor and the immune checkpoint antigen is an inhibitory immune cell antigen.

[0057] 20. The multispecific antibody of item 19, wherein the inhibitory immune cell antigen is selected from the group consisting of cytotoxic T-lymphocyte-associated protein 4 (CTLA4), PD-1, lymphocyte-activation gene 3, and T-cell immunoglobulin mucin-3, BTLA, TIM3, TIGIT, CD160, LAG3, LAIR1, B7-1, and B7-H1.

[0058] 21. The multispecific antibody of any one of items 1 to 20, further comprising a domain that specifically binds to human serum albumin (HSA).

[0059] 22. The multispecific antibody of any one of items 1 to 21, wherein the domains are capable of binding simultaneously to their respective antigens.

[0060] 23. The multispecific antibody of any one of items 1 to 22, wherein the domains are independently selected from the group consisting of Fab, Fv, scFv, dsFv, scAb, STAB, single domain antibodies (sdAb or dAb), single domain heavy chain antibodies, single domain light chain antibodies, VHH, and single domain antibodies based on VNAR structures derived from sharks.

[0061] 24. The multispecific antibody is a single chain diabody (scDb), a tandem scDb (Tandab), a linear dimeric scDb (LD-scDb), a cyclic dimeric scDb (CD-scDb), a bispecific T cell-inducing antibody (BiTE; tandem di-scFv), a tandem tri-scFv, a tribody (Fab-(scFv)2) or a bibody (Fab-(scFv)1), a Fab, a Fab-Fv2, a Morrison (IgG CH3-scFv fusion (Morrison L) or an IgG CL-scFv fusion (Morrison H)), triabody, scDb-scFv, bispecific Fab2, di-miniantibody, tetrabody, scFv-Fc-scFv fusion, scFv-HSA-scFv fusion, di-diabody, DVD-Ig, COVD, IgG-scFab, scFab-dsscFv, Fv2-Fc, IgG-scFv fusion, e.g., bsAb (scFv linked to the C-terminus of the light chain), Bs1Ab (scFv linked to the N-terminus of the light chain), Bs2Ab (scFv linked to the N-terminus of the heavy chain), Bs3Ab (scFv linked to the C-terminus of the heavy chain), Ts1Ab (scFv linked to the N-terminus of both the heavy and light chains), 24. The multispecific antibody according to any one of items 1 to 23, wherein the format is selected from the group consisting of: a fused scFv to the N- and / or C-terminus of either chain of a heavy chain, a Ts2Ab (dsscFv linked to the C-terminus of the heavy chain), a bispecific antibody based on a heterodimeric Fc domain, such as a Knob-into-Hole antibody (KiHs); a scDb, a tandem-di-scFv, a tandem tri-scFv, Fab-(scFv)2, Fab-(scFv)1, Fab, Fab-Fv2, a heterodimeric Fc domain or any other heterodimerization domain fused to the N- and / or C-terminus of either chain of COVD, MATCH and DuoBodies.

[0062] 25. The multispecific antibody of any one of items 1 to 24, wherein said antibody does not comprise an immunoglobulin Fc region polypeptide, and optionally does not comprise a CH1 and / or CL region.

[0063] 26. The multispecific antibody of any one of items 1 to 25, wherein said antibody comprises a CH1 and / or CL region and optionally comprises an immunoglobulin Fc region polypeptide.

[0064] 27. The multispecific antibody of any one of items 1 to 26, wherein said antibody is monovalent for each specificity.

[0065] 28. The multispecific antibody of any one of items 1 to 27, wherein the antibody is a scDb-scFv, a tribody, a DVD-tribody, MATCH, in particular in MATCH or tribody format, more particularly in MATCH format, more particularly MATCH3 or MATCH4.

[0066] 29. The multispecific antibody according to any one of items 5 to 28, the first domain binds to human PD-L1 with a dissociation constant (KD) of 100 nM to 1000 nM, for example, 100 nM to 900 nM, 150 nM to 850 nM, 200 nM to 800 nM, 250 nM to 750 nM, 300 nM to 700 nM, preferably 350 nM to 650 nM, more preferably 400 nM to 600 nM, particularly as measured by SPR; b. the first domain, when in scFv format, does not bind to cells expressing PD-L1, particularly as measured by flow cytometry, particularly at a concentration of the scFv less than 100 μg / ml; c. When the first domain is in scFv format, (i) does not neutralize PD-L1 binding to PD-1, particularly as measured by an NFAT reporter gene assay; or (ii) neutralizes PD-L1 binding to PD-1 with a potency relative to the potency of avelumab (relative potency) of less than 0.001, preferably less than 0.0005, as measured by an NFAT reporter gene assay, wherein the relative potency is determined by the IC50 of avelumab measured by the NFAT reporter gene assay. 50 IC values ​​(in ng / ml) of the scFvs measured by NFAT reporter gene assay 50 The ratio is based on the value (unit: ng / ml). Multispecific antibodies.

[0067] 30. The multispecific antibody according to any one of items 5 to 29, (i) has the ability to block the interaction between PD-L1 and PD-1 in the presence of TAA- / PD-L1+ cells with a potency relative to the potency of avelumab (relative potency) of less than 0.001, preferably less than 0.0005, as measured by a flow cytometry assay, said relative potency being determined by the IC50 of avelumab measured by a flow cytometry assay; 50 IC values ​​(in ng / ml) of the polyspecific antibodies measured by flow cytometry assay 50 value (in ng / ml), and (ii) has the ability to block the interaction between PD-L1 and PD-1 in the presence of TAA+ / PD-L1+ cells with a potency relative to the potency of avelumab (relative potency) of greater than 0.01, preferably greater than 0.05, more preferably greater than 0.1, as measured by a flow cytometry assay, wherein the relative potency is greater than the IC50 of avelumab measured by a flow cytometry assay. 50 IC values ​​(in ng / ml) of the polyspecific antibodies measured by flow cytometry assay 50 The ratio is based on the value (unit: ng / ml). Multispecific antibodies.

[0068] 31. The multispecific antibody according to any one of items 5 to 30, a. the first domain, when in scFv format, has a melting temperature (Tm) of at least 65° C., preferably at least 70° C., as measured by differential scanning fluorimetry, particularly when the scFv is formulated in 50 mM citrate phosphate buffer, 150 mM NaCl, pH 6.4; b. the first domain, when in scFv format, has less than 3% loss in monomer content, preferably less than 1%, after four successive freeze-thaw cycles, the scFv having a starting concentration of 10 mg / ml, in particular the scFv being formulated in 50 mM citrate phosphate buffer, pH 6.4, containing 150 mM NaCl, and c. the first domain, when in scFv format, exhibits less than 10% loss in monomer content, e.g., less than 9%, less than 8%, less than 7%, less than 6%, preferably less than 5%, after storage at 4° C. for at least 2 weeks, particularly at least 4 weeks, and the scFv has a starting concentration of 10 mg / ml, particularly the scFv is formulated in 50 mM citrate phosphate buffer, pH 6.4, containing 150 mM NaCl; multispecific antibodies

[0069] 32. The multispecific antibody according to any one of items 1 to 31, wherein each domain comprises a heavy chain variable region (VH) and a light chain variable region (VL), a. the VH comprises, in order, three complementarity determining regions, HCDR1, HCDR2 and HCDR3; b. said VL comprises, in order, three complementarity determining regions, LCDR1, LCDR2 and LCDR3 Multispecific antibodies.

[0070] 33. The multispecific antibody of any one of items 1 to 32, wherein the second domain specifically binds to the tumor-associated immune checkpoint antigen and / or the TAA that specifically binds to the first domain, and optionally a third domain specifically binds to the immune cell antigen, and optionally a further domain specifically binds to human serum albumin (HSA) comprise a light chain variable region (VL), The VL comprises a framework R4 selected from Vκ frameworks FR1, FR2 and FR3, particularly Vκ1 or Vκ3 FR1-FR3, preferably Vκ1 FR1-FR3 and Vκ FR4, particularly Vκ1 FR4, Vκ3 FR4 and Vλ FR4, particularly Vλ FR4 comprising an amino acid sequence having at least 80, particularly at least 90 percent identity to an amino acid sequence selected from any of SEQ ID NOs: 145 to 152, preferably any of SEQ ID NOs: 145 to 152, preferably Vλ FR4 set forth in SEQ ID NOs: 145, 146 or 152, more preferably Vλ FR4 set forth in SEQ ID NO: 146 or 152. Multispecific antibodies.

[0071] 34. The multispecific antibody of any one of items 1 to 33, wherein said first domain comprises HCDR1, HCDR2 and HCDR3 sequences of SEQ ID NOs: 1, 2 and 3, and LCDR1, LCDR2 and LCDR3 sequences of SEQ ID NOs: 5, 6 and 7, respectively, and one or more of said CDR sequences optionally comprise one or two mutations, in particular mutations to alanine residues, more particularly (i) said LCDR3 comprises Q108A (according to AHo numbering), (ii) said LCDR3 comprises G109A (according to AHo numbering), or (iii) said LCDR3 comprises Q108A and G109A (according to AHo numbering), and / or (iv) said HCDR3 comprises Y112A (according to AHo numbering).

[0072] 35. The multispecific antibody of item 34, wherein the first domain comprises the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 1, 2, and 3, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 5, 6, and 9, respectively.

[0073] 36. The multispecific antibody of item 34, wherein the first domain comprises the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 1, 2, and 3, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 5, 6, and 10, respectively.

[0074] 37. The multispecific antibody of item 35 or item 36, wherein the first domain comprises a heavy chain variable region (VH), wherein the VH is VH3 or VH4, preferably VH4.

[0075] 38. The multispecific antibody of item 35, comprising a VH comprising an amino acid sequence that is at least 90 percent, in particular at least 95 percent, identical to the amino acid sequence SEQ ID NO: 11, and a VL comprising an amino acid sequence that is at least 90 percent, in particular at least 95 percent, identical to the amino acid sequence SEQ ID NO: 15.

[0076] 39. The multispecific antibody of item 38, comprising the VH sequence of SEQ ID NO: 11 and the VL sequence of SEQ ID NO: 15.

[0077] 40. The multispecific antibody of item 36, comprising a VH comprising an amino acid sequence that is at least 90 percent, in particular at least 95 percent, identical to the amino acid sequence SEQ ID NO: 11, and a VL comprising an amino acid sequence that is at least 90 percent, in particular at least 95 percent, identical to the amino acid sequence SEQ ID NO: 16.

[0078] 41. The multispecific antibody of item 40, comprising the VH sequence of SEQ ID NO: 11 and the VL sequence of SEQ ID NO: 16.

[0079] 42. The multispecific antibody according to any one of items 1 to 41, wherein said second domain comprises (i) HCDR1, HCDR2 and HCDR3 sequences of SEQ ID NOs: 17, 18 and 19, respectively, and LCDR1, LCDR2 and LCDR3 sequences of SEQ ID NOs: 20, 21 and 22, or, in particular, (ii) HCDR1, HCDR2 and HCDR3 sequences of SEQ ID NOs: 27, 28 and 29, respectively, and LCDR1, LCDR2 and LCDR3 sequences of SEQ ID NOs: 30, 31 and 32, wherein one or more of said CDR sequences optionally comprises one or two mutations, in particular a mutation to an alanine residue.

[0080] 43. The multispecific antibody according to item 42, wherein the second domain comprises (i) a VH comprising an amino acid sequence that is at least 90 percent, particularly at least 95 percent, identical to the amino acid sequence SEQ ID NO: 23, and a VL comprising an amino acid sequence that is at least 90 percent, particularly at least 95 percent, identical to the amino acid sequence SEQ ID NO: 25, in particular wherein the VH comprises a Cys at position 51 and the VL comprises a Cys at position 141 (AHo numbering), or in particular (ii) a VL comprising an amino acid sequence that is at least 90 percent, particularly at least 95 percent, identical to the amino acid sequence SEQ ID NO: 33 and at least 90 percent, particularly at least 95 percent, identical to the amino acid sequence SEQ ID NO: 35, in particular wherein the VH comprises a Cys at position 51 and the VL comprises a Cys at position 141 (AHo numbering).

[0081] 44. The multispecific antibody of item 43, comprising (i) the VH sequence of SEQ ID NO: 24 and the VL sequence of SEQ ID NO: 26, or (ii) the VH sequence of SEQ ID NO: 34 and the VL sequence of SEQ ID NO: 36, in particular the VH sequence of SEQ ID NO: 34 and the VL sequence of SEQ ID NO: 36.

[0082] 45. The multispecific antibody according to any one of items 10 to 44, wherein said third domain (i) is directed against CD3, in particular wherein said third domain comprises the HCDR1, HCDR2 and HCDR3 sequences of SEQ ID NOs: 37, 38 and 39, respectively, and the LCDR1, LCDR2 and LCDR3 sequences of SEQ ID NOs: 40, 41 and 42, or (ii) is directed against CD137, in particular wherein said third domain comprises the HCDR1, HCDR2 and HCDR3 sequences of SEQ ID NOs: 71, 72 and 73, respectively, and the LCDR1, LCDR2 and LCDR3 sequences of SEQ ID NOs: 74, 75 and 76, and wherein one or more of said CDR sequences optionally comprises one or two mutations, in particular a mutation to an alanine residue.

[0083] 46. ​​The multispecific antibody of item 45, wherein the third domain (i) is directed against CD3 and comprises a VH comprising an amino acid sequence that is at least 90 percent, in particular at least 95 percent identical to the amino acid sequence SEQ ID NO: 43, and a VL comprising an amino acid sequence that is at least 90 percent, in particular at least 95 percent identical to the amino acid sequence SEQ ID NO: 44, or (ii) is directed against CD137 and comprises a VH comprising an amino acid sequence that is at least 90 percent, in particular at least 95 percent identical to the amino acid sequence SEQ ID NO: 77, and a VL comprising an amino acid sequence that is at least 90 percent, in particular at least 95 percent identical to the amino acid sequence SEQ ID NO: 78.

[0084] 47. The multispecific antibody of item 46 is (i) directed against CD3 and comprises the VH sequence of SEQ ID NO: 43 and the VL sequence of SEQ ID NO: 44, or (ii) directed against CD137 and comprises the VH sequence of SEQ ID NO: 77 and the VL sequence of SEQ ID NO: 78.

[0085] 48. The multispecific antibody of any one of items 21 to 28, wherein the domain specifically binding to HSA comprises (i) HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 45, 46 and 47, respectively, and LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 48, 49 and 50, respectively, (ii) HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 53, 54 and 55, respectively, and LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 56, 57 and 58, respectively, or (iii) HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 61, 62 and 63, and LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 64, 65 and 66, respectively, wherein one or more of said CDR sequences optionally comprises one or two mutations, in particular a mutation to an alanine residue.

[0086] 49. The multispecific antibody of item 48, wherein the domain that specifically binds to HSA comprises (i) a VH comprising an amino acid sequence that is at least 90 percent, particularly at least 95 percent identical to the amino acid sequence SEQ ID NO: 51, and a VL comprising an amino acid sequence that is at least 90 percent, particularly at least 95 percent identical to the amino acid sequence SEQ ID NO: 52, (i) a VH comprising an amino acid sequence that is at least 90 percent, particularly at least 95 percent identical to the amino acid sequence SEQ ID NO: 59, and a VL comprising an amino acid sequence that is at least 90 percent, particularly at least 95 percent identical to the amino acid sequence SEQ ID NO: 60, (i) a VH comprising an amino acid sequence that is at least 90 percent, particularly at least 95 percent identical to the amino acid sequence SEQ ID NO: 67, and a VL comprising an amino acid sequence that is at least 90 percent, particularly at least 95 percent identical to the amino acid sequence SEQ ID NO: 69, or (i) a VH comprising an amino acid sequence that is at least 90 percent, particularly at least 95 percent identical to the amino acid sequence SEQ ID NO: 68, and a VL comprising an amino acid sequence that is at least 90 percent, particularly at least 95 percent identical to the amino acid sequence SEQ ID NO: 70.

[0087] 50. The multispecific antibody of item 49, comprising: (i) the VH sequence of SEQ ID NO: 51 and the VL sequence of SEQ ID NO: 52; (ii) the VH sequence of SEQ ID NO: 59 and the VL sequence of SEQ ID NO: 60, (iii) the VH sequence of SEQ ID NO: 67 and the VL sequence of SEQ ID NO: 69, or (iv) the VH sequence of SEQ ID NO: 68 and the VL sequence of SEQ ID NO: 70.

[0088] 51. The multispecific antibody according to any one of items 1 to 50, wherein each chain is selected from the group consisting of: (i) SEQ ID NOs: 79 and 80, SEQ ID NOs: 81 and 82, SEQ ID NOs: 83 and 84, SEQ ID NOs: 85 and 86, SEQ ID NOs: 87 and 88, SEQ ID NOs: 89 and 90, SEQ ID NOs: 91 and 92, SEQ ID NOs: 93 and 94, SEQ ID NOs: 95 and 96, SEQ ID NOs: 97 and 98, SEQ ID NOs: 99 and 100, SEQ ID NOs: 101 and 102, SEQ ID NOs: 103 and 104, SEQ ID NOs: 105 and 106, SEQ ID NOs: 107 and 108, SEQ ID NOs: 109 and 110, SEQ ID NOs: 111 and 112, SEQ ID NOs: 113 and 114, SEQ ID NOs: 123 and 124, A combination of chains selected from SEQ ID NOs: 125 and 126, SEQ ID NOs: 127 and 128, SEQ ID NOs: 129 and 130, SEQ ID NOs: 131 and 132, SEQ ID NOs: 133 and 134, and SEQ ID NOs: 135 and 136, or a combination of sequences included in one of the sequences selected from SEQ ID NOs: 115 to 136, in particular a combination of two chains having an amino acid sequence with at least 80% identity, in particular at least 90% identity, more particularly at least 95% identity, for example 100% identity, to the sequences of the combination of chains of SEQ ID NOs: 123 and 124 or SEQ ID NOs: 127 and 128, (i) a first domain that specifically binds PD-L1 comprising the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 1, 2, and 3, respectively, or SEQ ID NOs: 1, 2, and 4, respectively, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 5, 6, and 7, respectively, SEQ ID NOs: 5, 6, and 8, respectively, SEQ ID NOs: 5, 6, and 9, respectively, or particularly SEQ ID NOs: 5, 6, and 10, respectively; and (ii) a second domain that specifically binds to HER2 comprising (i) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 17, 18, and 19, respectively, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 20, 21, and 22, respectively, or in particular (ii) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 27, 28, and 29, respectively, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 30, 31, and 32, respectively; (iii) optionally, a third domain that specifically binds to (i) CD3 comprising the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 37, 38, and 39, respectively, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 40, 41, and 42, respectively, or (ii) CD137 comprising the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 71, 72, and 73, respectively, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 74, 75, and 76, respectively; (iv) optionally, a further domain that specifically binds to HSA comprising: (i) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 45, 46, and 47, respectively, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 48, 49, and 50, respectively; (ii) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 53, 54, and 55, respectively, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 56, 57, and 58, respectively; or (iii) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 61, 62, and 63, respectively, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 64, 65, and 66, respectively. 2. A multispecific antibody comprising:

[0089] 52. A combination comprising (i) a multispecific antibody according to any one of items 1 to 51 and (ii) a second compound selected from (iia) an antibody against a TAA, in particular an antibody against HER2, in particular an antibody which is trastuzumab, (iib) an immune checkpoint antigen, in particular said immune checkpoint antigen which is not a tumour associated immune checkpoint antigen, and / or a modulator of said immune checkpoint antigen, in particular present on T cells or NK cells, and (iic) a modulator of angiogenesis.

[0090] 53. The combination of item 52, wherein the modulator is an antibody.

[0091] 54. The combination of items 52 or 53, wherein the modulator is an agonist and the immune checkpoint antigen is an immune cell antigen.

[0092] 55. The combination of item 54, wherein the immune cell antigen is selected from the group consisting of CD28, ICOS, HVEM, CD27, OX40, DR3, GITR, CD30, SLAM, CD2, 2B4, TIM1, TIM2, CD226, CTLA4, PD-1, lymphocyte activation gene 3, and T cell immunoglobulin mucin-3, BTLA, TIM3, TIGIT, CD160, LAG3, LAIR1, B7-1, and B7-H1.

[0093] 56. The combination of item 55, wherein the stimulatory immune cell antigen is CD3 or CD137, in particular CD3.

[0094] 57. The combination of items 52 or 53, wherein the modulator is an inhibitor and the immune cell antigen is an inhibitory immune checkpoint antigen.

[0095] 58. The combination of item 57, wherein said inhibitory immune cell antigen is selected from the group consisting of cytotoxic T-lymphocyte-associated protein 4 (CTLA4), PD-1, lymphocyte-activation gene 3, and T-cell immunoglobulin mucin-3, preferably said inhibitory immune checkpoint antigen is cytotoxic T-lymphocyte-associated protein 4 (CTLA4), more preferably said modulator is ipilimumab.

[0096] 59. A combination comprising (i) a multispecific antibody according to any one of items 1 to 51 and (ii) an antibody against a TAA.

[0097] 60. The combination of item 59, wherein said TAA is selected from HER2 and mesothelin, in particular HER2, and in particular said antibody is trastuzumab.

[0098] 61. A pharmaceutical composition comprising the multispecific antibody of any one of items 1 to 51, or the combination of any one of items 52 to 60, and a pharma- ceutically acceptable carrier.

[0099] 62. A PD-L1 binding domain defined in any one of items 29, 30, and 34 to 41.

[0100] 63. A HER2-binding domain defined in any one of items 42 to 44.

[0101] 64. A CD3-binding domain defined in any one of items 45(i) to 47(i).

[0102] 65. A CD137-binding domain defined in any one of items 45(ii) to 47(ii).

[0103] 66. An HSA-binding domain defined in any one of items 48 to 50.

[0104] 67. The multispecific antibody of any one of items 1 to 51, the combination of any one of items 52 to 60, or the binding domain of any one of items 62 to 66 for use as a medicament.

[0105] 68. The multispecific antibody of any one of items 1 to 51, the combination of any one of items 52 to 60, the pharmaceutical composition of item 61, or the binding domain of any one of items 62 to 66 for use in the treatment of cancer in a subject in need thereof.

[0106] 69. Use of the multispecific antibody of any one of items 1 to 51, the combination of any one of items 52 to 60, the pharmaceutical composition of item 61, or the binding domain of any one of items 62 to 66 for treating cancer in a subject in need thereof.

[0107] 70. Use of a multispecific antibody according to any one of items 1 to 51, a combination according to any one of items 52 to 60, a pharmaceutical composition according to item 61, or a binding domain according to any one of items 62 to 66 in the manufacture of a medicament for the treatment of cancer in a subject in need thereof.

[0108] 71. A method for treating cancer in a subject in need thereof, comprising administering to the subject a therapeutically effective amount of a multispecific antibody according to any one of items 1 to 51, 67 and 68, a combination according to any one of items 52 to 60, a pharmaceutical composition according to item 61, or a binding domain according to any one of items 62 to 66, or the use of item 69 or 70, wherein the cancer is a cancer positive for the TAA and the tumor-associated immune checkpoint antigen, and in particular the cancer is a cancer positive for the TAA. + / PDL + and more specifically, the cancer is HER2 + / PD-L1 + How to be.

[0109] 72. The use of the multispecific antibody of item 68, of item 69 or of item 70, or the method of item 71, wherein the cancer is a HER2- and PD-L1-positive cancer and the cancer is refractory to standard of care therapy, in particular trastuzumab.

[0110] 73. A nucleic acid encoding the multispecific antibody of any one of items 1 to 51 or the binding domain of any one of items 62 to 66.

[0111] 74. A vector comprising the nucleic acid of item 73.

[0112] 75. A host cell comprising the nucleic acid of item 73 or the vector of item 74.

[0113] 76. A method for producing a multispecific antibody according to any one of items 1 to 51 or a binding domain according to any one of items 62 to 66, the method comprising a step of culturing a host cell containing a nucleic acid or a vector encoding the multispecific antibody according to any one of items 1 to 51 or the binding domain according to any one of items 62 to 66.

[0114] 77. A kit comprising the multispecific antibody of any one of items 1 to 51, the combination of any one of items 52 to 60, the pharmaceutical composition of item 61, or the binding domain of any one of items 62 to 66.

[0115] 78. The multispecific antibody according to any one of items 1 to 50, comprising SEQ ID NOs: 79 and 80, SEQ ID NOs: 81 and 82, SEQ ID NOs: 83 and 84, SEQ ID NOs: 85 and 86, SEQ ID NOs: 87 and 88, SEQ ID NOs: 89 and 90, SEQ ID NOs: 91 and 92, SEQ ID NOs: 93 and 94, SEQ ID NOs: 95 and 96, SEQ ID NOs: 97 and 98, SEQ ID NOs: 99 and 100, SEQ ID NOs: 101 and 102, SEQ ID NOs: 103 and 104, SEQ ID NOs: 105 and 106, SEQ ID NOs: 107 and 108, SEQ ID NOs: 109 and 110, or the sequences An antibody comprising a combination of two chains selected from SEQ ID NOs: 111 and 112, SEQ ID NOs: 113 and 114, SEQ ID NOs: 123 and 124, SEQ ID NOs: 125 and 126, SEQ ID NOs: 127 and 128, SEQ ID NOs: 129 and 130, SEQ ID NOs: 131 and 132, SEQ ID NOs: 133 and 134, and SEQ ID NOs: 135 and 136, or a combination of sequences contained in one of the sequences selected from SEQ ID NOs: 115 to 136, in particular a combination of chains of SEQ ID NOs: 123 and 124 or SEQ ID NOs: 127 and 128. [Brief description of the drawings]

[0116] [Figure 1] Figure 1. Binding to PD-L1 expressing cells assessed by flow cytometry. Binding of (A) PRO1434 and (B) PRO1494 to PD-L1 expressing cells. PRO830 was used as a reference. PRO1434 only showed a signal at 100 μg / ml, while binding was observed at 3.5 μg / ml for PRO1494. [Diagram 2] Figure 2. Blockade of PD-1 / PD-L1 interaction in an NFAT reporter gene assay. PD-L1 neutralization by (A) PRO1434 and (B) PRO1494. Avelumab was used as a reference. With both molecules tested, only partial neutralization of the PD-1 and PD-L1 interaction was observed at the highest concentration tested (162 μg / ml). [Diagram 3] Figure 3 shows the structure of the multispecific molecules of the invention. Schematic diagram and description of the three different multispecific formats, tribody, DVD-tribody and MATCH-4. Table 14 lists the domains contained in each molecule that was created and their location within the molecule. The targets of each domain are as follows: Trastuzumab: Her2; Clone 14-11-D07: IL23R; Clone 23-13-A01: Human / Mouse Serum Albumin; Clone 28-21-D09: CD3e; Clone 33-02-G02 and its variants: PD-L1. The Gly-Ser linker sequences connecting the individual domains are shown in the figure. [Figure 4]Figure 4. Blockade of PD-1 / PD-L1 interaction in Her2 expressing cells. Inhibition of PD-1 binding to (A) cells expressing PD-L1 and Her2 (HCC1954) or (B) cells expressing PD-L1 without significant expression of Her2 (HCC827) in the presence of increasing concentrations of avelumab, PRO1454, PRO1456 and PRO1497. PRO1454 inhibits PD-1 binding to PD-L1 with an IC50 of 205ng / ml in PD-L1 / high Her2 expressing cells (HCC1954) and an IC50 of 1204ng / ml in PD-L1 expressing cells (HCC827). PRO1497, which contains an anti-PD-L1 domain with 50-fold lower affinity, inhibited PD-1 binding to Her2 / PD-L1 expressing cells with comparable potency as PRO1454, but showed very weak inhibition of PD-1 binding to cells expressing PD-L1 alone at the concentrations tested. PRO1456, which does not contain the anti-PD-L1 domain, had no effect on PD-1 binding in both cell lines. Data were fitted using a sigmoidal 4PL fit (GraphPad Prism). [Diagram 5] Figure 5. Blockade of PD-1 / PD-L1 interaction in Her2 expressing cells in the presence of human serum albumin. A) Inhibition of PD-1 binding to cells expressing PD-L1 and Her2 (HCC1954) in the presence of increasing concentrations of avelumab, nivolumab, PRO1543 (Her2 x CD3 x HSA x PD-L1 low affinity) and PRO1546 (HER2 x CD3 x HSA x IL23R). B) Inhibition of PD-1 binding to cells expressing PD-L1 without significant expression of Her2 (HCC827). PRO1543 inhibits PD-1 binding to PD-L1 with an IC50 value of 600ng / ml only in PD-L1 / high Her2 expressing cells. PRO1546, which does not contain the anti-PD-L1 domain, had no effect on PD-1 binding in both cell lines. Data were fitted using sigmoidal 4PL fit (GraphPad Prism). [Figure 6]Figure 6. CD3 activation and concomitant PD-L1 blockade by PRO1454 or PRO1497 assessed by NFAT-luciferase reporter gene assay in the presence of human serum albumin. A) In the presence of PD-L1 / Her2 expressing cells (HCC1954), PRO1454 and PRO1456 activated CD3 signaling in Jurkat cells with similar EC50, but maximal activation was higher with PRO1454, a molecule with a low affinity anti-PD-L1 domain, compared to PRO1456, which contains an anti-IL23R dummy domain instead of the anti-PD-L1 domain. This suggests that PRO1454 blocks PD-L1 and concomitantly activates CD3 within the immune synapse in the presence of cells co-expressing Her2 and PD-L1. Weaker activation was observed with PRO1455, a molecule without an anti-Her2 domain and with a low affinity anti-PD-L1 domain (33-03-G02 G109A). B) Tribody molecule PRO1497 was tested, which contains an anti-PD-L1 domain with two alanine mutations (Q108A and G109A) with at least 50-fold lower affinity than the molecules tested in A and the corresponding reference incorporated domain. For these molecules, PD-L1 blockade and CD3 activation were observed simultaneously, as PRO1497 contained a low affinity anti-PD-L1 domain and induced higher maximum activation than PRO1456, which contained an anti-IL23R domain instead. Compared to PRO1455, PRO1498 caused very weak activation due to the very low affinity of the incorporated anti-PD-L1 domain. Luminescence was read 5 hours after addition of Jurkat reporter cells and data were fitted using sigmoidal 4PL fit (GraphPad Prism). [Figure 7]Figure 7. CD3 activation and concomitant PD-L1 blockade by PRO1543 assessed by NFAT-luciferase reporter gene assay in the presence of human serum albumin. A) In the presence of PD-L1 / Her2 expressing cells (HCC1954), PRO1543 and PRO1557 activated CD3 signaling in Jurkat cells with similar EC50, but maximal activation was higher with PRO1543, a molecule with a low affinity anti-PD-L1 domain, compared to PRO1557, which contains an anti-IL23R dummy domain instead of the anti-PD-L1 domain. This suggests that PRO1543 blocks PD-L1 and concomitantly activates CD3 within the immune synapse in the presence of cells co-expressing Her2 and PD-L1. This observation was further supported by the addition of 1 μg / ml nivolumab to all molecules, which resulted in similar maximal activation in the presence of PRO1557 and PRO1543 as PRO1543 alone, demonstrating complete PD-L1 / PD-1 blockade by PRO1543. No activation was observed with PRO1546, a molecule lacking an anti-Her2 domain but with a low affinity anti-PD-L1 domain. B) In the presence of PD-L1 expressing CHO cells, no activation was observed with molecules with an anti-PD-L1 domain (PRO1543 and PRO1546), with or without an anti-Her2 domain. No activation was seen with PRO1557, as this molecule does not contain an anti-PD-L1 domain. Luminescence was read 5 hours after addition of Jurkat reporter cells and data was fitted using a sigmoidal 4PL fit (GraphPad Prism). [Figure 8] FIG. 8. Activation of CD8+ T cells as measured by upregulation of CD69 in the presence of PRO1543, PRO1895 and the control molecule PRO2290 after co-incubation with HCC827 tumor cells (low HER2, PD-L1+). [Figure 9] FIG. 9. Activation of CD8+ T cells measured by upregulation of CD69 in the presence of PRO1543, PRO1895 and the control molecule PRO2290 after co-incubation with HCC1954 tumor cells (high HER2, PD-L1+). [Figure 10] Figure 10. CD4+ and CD8+ T cell viability. CD4+ T cell (A) and CD8+ T cell (B) viability was only reduced by 5-10% at the highest concentration of molecules tested. Cytotoxicity assessment of PBMCs in the presence of PD-L1 / high Her2 expressing cancer cells (HCC1954) was performed 40 hours after initiation by staining CD4+ and CD8+ T cells with fluorescently labeled antibodies and analyzing by flow cytometry. Similar data were obtained after 16 hours. [Figure 11] Figure 11. T cell-mediated target cell killing and CD8+ cell activation in the presence of A) Her2+++ / PD-L1+HCC1954 and B) Her2+ / PD-L1-MCF-7. In this assay, freshly isolated human PBMCs were co-cultured with the indicated target cells for 16 hours in the presence of various test molecules. Compared to Her2 / CD3 scDb (PRO957), PRO1543 showed 50-100-fold better potency on Her2+++ / PD-L1+, whereas only slightly different potency was observed on PD-L1-negative cells. The improved activity of molecules containing anti-PD-L1 domains most likely resulted from increased avidity, since the EC50 of PD-L1 blockade on these cells was significantly higher than the EC50 of target cell lysis. As a result, binding to Her2 and PD-L1 double positive cells was stronger than binding to PD-L1-negative cells expressing Her2. This avidity binding selectively enhances efficacy against tumor cells (Her2 / PD-L1 double positive) but not against PD-L1 negative, Her2-expressing normal cells, thereby broadening the therapeutic window. [Figure 12]Figure 12. T cell-mediated target cell killing and CD8+ cell activation in the presence of A) Her2+ / - / PD-L1+ HCC827 and B) Her2- / PD-L1+ CHO PD-L1 cells. Freshly isolated human PBMCs were co-cultured with the indicated target cells for 16 hours in the presence of various test molecules. In Her2+ / - / PD-L1+, PRO1543 showed 20-fold better efficacy than Her2 / CD3 scDb. In the presence of Her2-negative PD-L1-positive cells, only little target cell killing and CD8+ cell activation was observed at high concentrations of MATCH4 with a low affinity PD-L1 domain, allowing for a very wide therapeutic window. [Figure 13] Figure 13: Human PBMC-replaced NOG mice were implanted with HCC1954 ductal carcinoma cells (n=8 each). Mice were dosed on days 0, 5, 10, 15, 20, 25, and 30 (dotted vertical lines). Tumor growth and body weight were recorded twice weekly. PRO1678 scMATCH3 had antitumor effects similar to nivolumab, demonstrating efficient tumor-targeted PD-L1 blockade. PRO1543 MATCH4 therapy produced a higher antitumor effect than the nivolumab / trastuzumab combination. [Figure 14] Figure 14: Design of multispecific molecules. Schematic diagram and description of three different multispecific formats, tribody, DVD-tribody and MATCH-4. Table 25 describes the composition of the domains of each molecule created and their arrangement within the molecule. The targets of each domain are as follows: Trastuzumab: Her2; Clone 14-11-D07: IL23R; Clone 23-13-A01: Human / Mouse SA; Clone 28-21-D09: CD3e; Clone 33-02-G02 and its variants: PD-L1. The Gly-Ser linker sequences connecting the individual domains are shown in the figure. [Figure 15]Figure 15: Effect of PRO1993 on CD137 signaling activity in NF-kB Jurkat reporter cells. After incubating Jurkat cells with PRO1993 for 24 hours in the presence of HCC1954 (high expression of Her2 and PD-L1) and HCC827 (low expression of Her2, high expression of PD-L1), the activity of CD137 signaling was evaluated by detecting luminescence. PRO1993 activated CD137 signaling in the presence of HCC1954 cells, which highly express Her2, whereas only slight activation of CD137 signaling was observed in the presence of HCC827, which expresses low levels of Her2. Data was fitted using sigmoidal 4PL fit (GraphPad Prism). [Figure 16] Figure 16. Blockade of PD-1 / PD-L1 interaction in Her2 expressing cells in the presence of human serum albumin. PD-1 binding levels to A) cells expressing PD-L1 and high levels of Her2 (HCC1954) or B) cells expressing PD-L1 and low levels of Her2 (HCC827) in the presence of increasing concentrations of avelumab and PRO1993 (Her2 x CD137 x HSA x PD-L1 low affinity). PRO1993 inhibited PD-1 binding to PD-L1 with an IC50 value of 166.7ng / ml in PD-L1 / high Her2 expressing cells, whereas no inhibition of PD-1 binding was observed in HCC827 cells. In comparison, avelumab inhibited this interaction with IC50s of 127.7ng / ml (HCC1954) and 46.03ng / ml (HCC827). Data were fitted using sigmoidal 4PL fit (GraphPad Prism). [Figure 17] Figure 17. Design of scDb-scFv molecules. Schematic and illustration of scDb-scFv molecules. [Figure 18]Figure 18. Blockade of PD-1 / PD-L1 interaction in Her2-expressing cells in the presence of human serum albumin. A) PD-1 binding levels to cells expressing PD-L1 and high levels of Her2 (HCC1954) or (B) cells expressing PD-L1 and lower levels of Her2 (HCC827) in the presence of increasing concentrations of avelumab and PRO1678 (Her2 x HSA x PD-L1 low affinity). PRO1678 inhibits PD-1 binding to PD-L1 in PD-L1 / high Her2 expressing cells with 100-fold better potency than in PD-L1 expressing cells (HCC827), with an IC50 value of 428.2ng / ml. In comparison, avelumab inhibits this interaction with IC50s of 127.7ng / ml (HCC1954) and 46.03ng / ml (HCC827). Data were fitted using sigmoidal 4PL fit (GraphPad Prism). [Figure 19] Figure 19 shows cell membrane binding of MATCH4 molecules PRO1543 and PRO1895 to SK-OV3, MCF-7 and CHO PD-L1 cells. Concentration response curves of MATCH4 molecules PRO1543 and PRO1895 and clinical-stage anti-HER2 antibodies trastuzumab and pertuzumab to SK-OV3 (top left), MCF-7 (top right) and CHO PD-L1 (bottom right). MATCH4 molecules bound to SK-OV3 cells expressing high levels of HER2 with apparent binding affinity comparable to that of the clinical-stage antibodies trastuzumab and pertuzumab. On the other hand, when binding to MCF-7 cells was evaluated, the apparent binding affinity of MATCH4 molecules was inferior to that of trastuzumab and pertuzumab. [Figure 20]Figure 20: Cell membrane binding of MATCH4 molecules PRO1543 and PRO1895 to IFNy-stimulated HCC1954 and HCC827 cells. Concentration-response curves of MATCH4 molecules PRO1543 and PRO1895 and clinical-stage anti-HER2 antibodies trastuzumab and pertuzumab to HCC1954 (left) and HCC827 (right). Cells were stimulated with 10 ng / ml IFNy for 24 hours before testing by flow cytometry. When evaluating binding to HCC1954 cells expressing high levels of HER2 and PD-L1, the apparent binding affinity of MATCH4 molecules is inferior to trastuzumab and pertuzumab. On the other hand, MATCH4 molecules bound to HCC827 cells expressing low levels of HER2 and high levels of PD-L1 with apparent binding affinity similar to that of clinical-stage antibodies trastuzumab and pertuzumab (right). Please note that in the graph shown to the right, the highest concentrations of PRO1543 and PRO1895 were not used in fitting the curves. [Figure 21] Figure 21: Cell membrane binding of MATCH4 molecules PRO1543 and PRO1895 to SK-OV3 cells in the presence of trastuzumab and pertuzumab. Concentration response curves of MATCH4 molecules PRO1543 (left) and PRO1895 (right) with or without the addition of trastuzumab or pertuzumab are shown. Cells were incubated with 50 nM trastuzumab or pertuzumab for 1 hour before the addition of MATCH4 molecules. Cell membrane binding of PRO1543 and PRO1895 was then evaluated by flow cytometry. PRO1543 showed binding to HER2-expressing SK-OV3 cells when applied alone and when tested in the presence of pertuzumab. On the other hand, PRO1895 showed binding to cells in the presence of trastuzumab and when tested alone. No binding was observed when PRO1543 and PRO1895 were tested in the presence of trastuzumab or pertuzumab, respectively. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0117] The use of therapeutic antibodies that block the interaction of tumor-associated immune checkpoint antigens, such as PD-L1, with their cognate ligands, such as PD-1, is a very promising treatment strategy, but is associated with challenges, including high toxicity and adverse events. Thus, there is a medical need for novel approaches to block the interaction of tumor-associated immune checkpoint antigens with their cognate ligands that have lower rates of dose-limiting toxicity and adverse events than currently available approaches.

[0118] The present invention provides a multispecific antibody comprising at least a first domain that specifically binds to a tumor-associated immune checkpoint antigen with low affinity and at least a second domain that specifically binds to a tumor-associated antigen (TAA). The multispecific antibody of the present disclosure can bind to a target cell presenting the TAA by the first domain that specifically binds to the TAA, and the low affinity binding domain can simultaneously bind to the tumor-associated immune checkpoint antigen present on the same target cell by the avidity effect, thereby inhibiting the interaction of the tumor-associated immune checkpoint antigen. Due to the low affinity of the first domain, specific binding to non-target cells that present only the tumor-associated immune checkpoint antigen and not the TAA does not occur to a relevant extent. Thus, due to its ability to mediate, e.g. stimulate, strong signaling of the tumor-associated immune checkpoint antigen on the target cell without interacting with non-target cells, treatment with the multispecific antibody of the present invention does not lead to depletion of cells that do not express the TAA.

[0119] Moreover, it has surprisingly been found that the multispecific antibodies of the present disclosure, comprising (a) at least said first domain, (b) at least said second domain, and (c) at least a third domain that specifically binds to an immune cell antigen, exhibit further beneficial properties, as shown in the examples and the accompanying figures. Furthermore, the optional addition of a half-life extending anti-HSA domain should facilitate delivery of the molecule to the tumor microenvironment as well as allow convenient administration.

[0120] Thus, the multispecific antibodies of the present invention offer distinct therapeutic advantages over conventional compositions and treatments.

[0121] All technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs unless otherwise defined.

[0122] As used herein, unless otherwise indicated, the terms "comprising" and "including" are used in their open-ended and non-limiting sense. Thus, with respect to such latter embodiments, the term "comprising" includes the narrower term "consisting of."

[0123] The terms "a" and "an" and "the" and similar references in the context of describing the present invention (particularly in the context of the claims below) should be construed to encompass both the singular and the plural, unless otherwise indicated herein or clearly contradicted by context. For example, the term "a cell" includes a plurality of cells, including mixtures thereof. It should be noted that when the plural forms are used for compounds, salts, and the like, this is intended to refer to a single compound, salt, or the like.

[0124] In one aspect, the present invention relates to a multispecific antibody comprising at least a first domain that specifically binds with low affinity to a tumor-associated immune checkpoint antigen and at least a second domain that specifically binds to a tumor-associated antigen (TAA).

[0125] As used herein, the term "antibody" and cognates include whole antibodies or single chains thereof, and any antigen-binding fragment (i.e., "antigen-binding portion") or single chains thereof, and molecules comprising antibody CDRs, VH regions or VL regions, including, but not limited to, multispecific antibodies. A naturally occurring "whole antibody" is a glycoprotein with at least two heavy chains (H) and two light chains (L) interconnected by disulfide bonds. Each heavy chain is composed of a heavy chain variable region (abbreviated herein as VH) and a heavy chain constant region. The heavy chain constant region is composed of three domains, CH1, CH2 and CH3. Each light chain is composed of a light chain variable region (abbreviated herein as VL) and a light chain constant region. The light chain constant region is composed of one domain, CL. The VH and VL regions are further subdivided into regions of hypervariability called complementarity determining regions (CDRs) and more conserved regions called framework regions (FRs). Each VH and VL is composed of three CDRs and four FRs arranged from amino-terminus to carboxy-terminus in the following order: FR1, CDR1, FR2, CDR2, FR3, CDR3, FR4. The variable regions of the heavy and light chains contain a binding domain that interacts with an antigen. The constant region of the antibody can mediate the binding of the immunoglobulin to host tissues or factors, including various cells of the immune system (e.g., effector cells) and the first component (C1q) of the classical complement system.

[0126] As used herein, the terms "binding domain", "antigen-binding fragment thereof", "antigen-binding portion" of an antibody, and the like, refer to one or more fragments of an intact antibody that retain the ability to specifically bind to a given antigen (e.g., CD137, PD-L1, HSA). The antigen-binding function of an antibody can be performed by fragments of an intact antibody. In some embodiments, the binding domain of a multispecific antibody of the invention is a Fab fragment, which is a monovalent fragment consisting of the VL, VH, CL and CH1 domains; a F(ab)2 fragment, which is a bivalent fragment comprising two Fab fragments linked by a disulfide bridge at the hinge region; an Fd fragment consisting of the VH and CH1 domains; an Fv fragment consisting of a single arm VL and VH domains of an antibody; a single domain antibody (dAb) fragment consisting of the VH domain (Ward et al., 1989 Nature 103:131-135). 341:544-546); isolated complementarity determining regions (CDRs), dsFvs, scAbs, STABs, single domain antibodies (sdAbs or dAbs), single domain heavy chain antibodies, and single domain light chain antibodies, VHHs, VNARs, single domain antibodies based on shark-derived VNAR structures, and binding domains based on alternative scaffolds, including but not limited to ankyrin-based domains, fynomers, avimers, anticalins, fibronectins, binding sites integrated into the constant region of an antibody (e.g. f-star technology (F-star's Modular Antibody Technology™)). Suitably, the binding domain of the invention is a single chain Fv fragment (scFv) or a single antibody variable domain. In a preferred embodiment, the binding domain of the invention is a single chain Fv fragment (scFv).

[0127] The term "complementarity determining region" ("CDR") is an amino acid sequence having boundaries determined using any of a number of well-known schemes, including those described below. Numbering schemes include those described in Kabat et al. (1991), "Sequences of Proteins of Immunological Interest," 5th Ed. Public Health Service, National Institutes of Health, Bethesda, MD ("Kabat" numbering scheme), Al-Lazikani et al., (1997) JMB 273, 927-948 ("Chothia" numbering scheme), ImMunoGenTics (IMGT) numbering (Lefranc, M.-P., The Immunologist, 7, 132-136 (1999); Lefranc, M.-P. et al., Dev. Comp. Immunol., 27, 55-77 (2003) ("IMGT" numbering scheme), and Honegger & Plueckthun, J. Mol. Biol. 309 (2001) For example, in the classical format, under Kabat, the CDR amino acid residues of the heavy chain variable domain (VH) are numbered 31-35 (HCDR1), 50-65 (HCDR2), and 95-102 (HCDR3), and the CDR amino acid residues of the light chain variable domain (VL) are numbered 24-34 (LCDR1), 50-56 (LCDR2), and 89-97 (LCDR3). Under Chothia, the CDR amino acids of the VH are numbered 26-32 (HCDR1), 52- The CDRs are numbered 56 (HCDR2), 95-102 (HCDR3) in the human VH and 24-34 (LCDR1), 50-56 (LCDR2), 89-97 (LCDR3) in the VL. Combining both Kabat and Chothia CDR definitions, the CDRs consist of amino acid residues 26-35 (HCDR1), 50-65 (HCDR2), and 95-102 (HCDR3) in the human VH and amino acid residues 24-34 (LCDR1), 50-56 (LCDR2), and 89-97 (LCDR3) in the human VL.Under IMGT, the CDR amino acid residues of the VH are numbered approximately 26-35 (HCDR1), 51-57 (HCDR2) and 93-102 (HCDR3), and the CDR amino acid residues of the VL are numbered approximately 27-32 (LCDR1), 50-52 (LCDR2) and 89-97 (LCDR3) ("Kabat" numbering). Under IMGT, the CDRs of an antibody can be determined using the program IMGT / DomainGap Align.

[0128] In the context of the present invention, unless otherwise specifically stated, the numbering system proposed by Honegger & Plueckthun ("AHo") is used (Honegger & Plueckthun, J. Mol. Biol. 309 (2001) 657-670).Furthermore, the following residues are defined as CDR according to the AHo numbering scheme: LCDR1 (also called CDR-L1): L24-L42; LCDR2 (also called CDR-L2): L58-L72; LCDR3 (also called CDR-L3): L107-L138; HCDR1 (also called CDR-H1): H27-H42; HCDR2 (also called CDR-H2): H57-H76; HCDR3 (also called CDR-H3): H108-H138. For clarity, the numbering system according to Honegger & Plueckthun takes into account the sequence diversity found in naturally occurring antibodies in the various VH and VL subfamilies, especially in the CDRs, and provides gaps in the sequence, such that a given antibody variable domain will not usually have all amino acid residues from positions 1 to 149 occupied.

[0129] The term "binding specificity" as used herein refers to the ability of an individual antibody to react with one antigenic determinant and not with a different antigenic determinant. As used herein, the term "specifically binds" or "specific" refers to a measurable and reproducible interaction, such as binding between a target and an antibody, that determines the presence of the target in the presence of a heterogeneous population of molecules, including biomolecules. For example, an antibody that specifically binds to a target (which may be an epitope) is an antibody that binds to this target with higher affinity, avidity, more readily, and / or with longer duration than it binds to other targets. In its most general form (and unless a defined reference is made), "specific binding" refers to the ability of an antibody to discriminate the target of interest from unrelated molecules, for example as determined according to a specificity assay method known in the art. Such methods include, but are not limited to, Western blot, ELISA, RIA, ECL, IRMA, SPR (surface plasmon resonance) test, and peptide scan. For example, a standard ELISA assay can be performed. Scoring can be done by standard color development (e.g. using a secondary antibody with horseradish peroxidase and tetramethylbenzidine, hydrogen peroxide). Reactions in a well are scored by optical density at, for example, 450 nm. A typical background (=negative reaction) can be about 0.1 OD, a typical positive reaction can be about 1 OD. This means that the ratio between positive and negative scores can be 10-fold or more. In a further example, an SPR assay can be performed, where a difference between background and signal of at least 10-fold, preferably at least 100-fold, indicates specific binding. Typically, the determination of binding specificity is done by using not a single reference molecule, but a set of about 3-5 unrelated molecules, such as milk powder, transferrin or the like.

[0130] Suitably, the antibodies of the present invention are isolated antibodies. As used herein, the term "isolated antibody" refers to an antibody that is substantially free of other antibodies having different antigen specificity (e.g., an isolated antibody that specifically binds PD-L1 and HER2 is substantially free of antibodies that specifically bind to antigens other than PD-L1 and HER2, e.g., an isolated antibody that specifically binds PD-L1, HER2 and human serum albumin is substantially free of antibodies that specifically bind to antigens other than PD-L1, HER2 and human serum albumin). Furthermore, an isolated antibody can be substantially free of other cellular material and / or chemicals.

[0131] Preferably, the antibody of the present invention is a monoclonal antibody. As used herein, the term "monoclonal antibody" or "monoclonal antibody composition" refers to antibodies whose amino acid sequences are substantially identical or derived from the same genetic source. A monoclonal antibody composition exhibits binding specificity and affinity for a particular epitope or binding specificity and affinity for a particular epitope.

[0132] The antibodies of the present invention include, but are not limited to, chimeric, human and humanized antibodies.

[0133] The term "chimeric antibody" (or antigen-binding fragment thereof) refers to an antibody molecule in which (a) the constant region, or a portion thereof, has been altered, replaced or exchanged so that the antigen-binding site (variable region) is linked to a different or altered class, effector function and / or species, or an entirely different molecule that confers new properties to the chimeric antibody, such as an enzyme, toxin, hormone, growth factor, drug, etc., or (b) the variable region, or a portion thereof, has been altered, replaced or exchanged with a variable region having a different or altered antigen specificity. For example, a murine antibody can be modified by replacing its constant region with a constant region from a human immunoglobulin. The replacement with a human constant region allows the chimeric antibody to retain the specificity of recognizing the antigen, while reducing its antigenicity in humans compared to the original murine antibody.

[0134] As used herein, the term "human antibody" (or antigen-binding fragment thereof) is intended to include antibodies (and antigen-binding fragments thereof) having variable regions in which both framework and CDR regions are derived from sequences of human origin. Furthermore, if they contain a constant region, the constant region is also derived from such a human sequence, e.g., a human germline sequence, or a mutated version of a human germline sequence. The human antibodies and antigen-binding fragments thereof of the present invention can include amino acid residues not encoded by human sequences (e.g., mutations introduced by random or site-specific mutagenesis in vitro or by somatic mutation in vivo). This definition of a human antibody specifically excludes humanized antibodies that include non-human antigen-binding residues. Human antibodies can be generated using a variety of techniques known in the art, including phage display libraries (Hoogenboom and Winter, J. Mol. Biol, 227:381 (1991); Marks et al, J. Mol. Biol, 222:581 (1991)). Also, methods for preparing human monoclonal antibodies are available as described in Cole et al, Monoclonal Antibodies and Cancer Therapy, Alan R. Liss, p. 77 (1985); Boemer et al, J. Immunol, 147(1):86-95 (1991). See also van Dijk and van de Winkel, Curr. Opin. Pharmacol, 5: 368-74 (2001). Human antibodies can be prepared by administering antigen to transgenic animals that have been engineered to produce such antibodies in response to antigen challenge, but have been disabled at their endogenous locus, e.g., immunizing xenomice (see, e.g., U.S. Pat. Nos. 6,075,181 and 6,150,584 for XENOMOUSE™ technology). See also, e.g., Li et al., Proc. Natl. Acad. Sci. USA, 103:3557- 3562 (2006) regarding human antibodies produced by human B cell hybridoma technology.

[0135] As used herein, a "humanized" antibody (or antigen-binding fragment thereof) is an antibody (or antigen-binding fragment thereof) that retains the reactivity of a non-human antibody while being less immunogenic in humans. This can be achieved, for example, by retaining the non-human CDR regions and replacing the remainder of the antibody with their human counterparts (i.e., the constant and variable region framework portions). Additional framework region modifications can be made within the human framework sequences and within the CDR sequences derived from the germline of other mammalian species. The humanized antibodies of the invention can include amino acid residues not encoded by human sequences (e.g., mutations introduced by random or site-specific mutagenesis in vitro or somatic mutation in vivo, or conservative substitutions to facilitate stability or manufacturing). See, e.g., Morrison et al., Proc. Natl. Acad. Sci. USA, 81:6851-6855, 1984; Morrison and Oi, Adv. Immunol., 44:65-92, 1988; Verhoeyen et al., Science, 239: 1534-1536, 1988; Padlan, Molec. Immun., 28:489-498, 1991; and Padlan, Molec. Immun., 31: 169-217, 1994. Other examples of human engineering techniques include, but are not limited to, the Xoma technology disclosed in U.S. Patent No. 5,766,886.

[0136] The term "recombinant humanized antibody" as used herein encompasses all human antibodies that have been prepared, expressed, produced or isolated by recombinant means, such as antibodies isolated from a host cell, e.g., a transfectoma, that has been transformed to express the humanized antibody, and antibodies prepared, expressed, produced or isolated by any other means, including splicing all or part of a human immunoglobulin gene, sequence, into other DNA sequences.

[0137] Preferably, the antibody or antigen-binding fragment thereof of the present invention is humanized. Preferably, the antibody or antigen-binding fragment thereof of the present invention is humanized and comprises CDRs from rabbit.

[0138] The term "multispecific antibody" as used herein refers to an antibody that binds to two or more different epitopes on at least two or more different targets (e.g., PD-L1 and HER2). The term "multispecific antibody" includes bispecific, trispecific, tetraspecific, pentaspecific and hexaspecific. The term "bispecific antibody" as used herein refers to an antibody that binds to two different epitopes on at least two different targets (e.g., PD-L1 and HER2). The term "trispecific antibody" as used herein refers to an antibody that binds to three different epitopes on at least three different targets (e.g., PD-L1, HER2 and HSA).

[0139] The term "epitope" refers to a protein determinant capable of specific binding to an antibody. Epitopes usually consist of chemically active surface groupings of molecules such as amino acids or sugar side chains and usually have specific three-dimensional structural features as well as specific charge characteristics. "Conformational" and "linear" epitopes are distinguished in that the binding to the former but not the latter is lost in the presence of denaturing solvents.

[0140] The term "conformational epitope" as used herein refers to amino acid residues of an antigen that come together on the surface when the polypeptide chain folds to form the native protein.

[0141] The term "linear epitope" refers to an epitope in which all of the interaction points of a molecule (such as an antibody) that interacts with a protein occur linearly (continuous) along the primary amino acid sequence of the protein.

[0142] The term "distal epitope" refers to an epitope contained in a region of the extracellular portion of a cell-associated antigen, away from the cell surface.

[0143] The term "recognize" as used herein refers to antibodies and antigen-binding fragments that find and interact with (eg, bind to) that conformational epitope.

[0144] As used herein, the term "affinity" refers to the strength of the interaction between an antibody and an antigen at a single antigenic site. Within each antigenic site, the variable regions of the antibody "arms" interact with the antigen at multiple sites through weak non-covalent forces; the more interactions, the stronger the affinity.

[0145] "Binding affinity" generally refers to the strength of the sum of non-covalent interactions between a single binding site of a molecule (e.g., an antibody) and its binding partner (e.g., an antigen). Unless otherwise indicated, as used herein, "binding affinity", "binds to", "binds to" or "binding to" refers to the intrinsic binding affinity that reflects a 1:1 interaction between members of a binding pair (e.g., an antibody fragment and an antigen). The affinity of a molecule X for its partner Y can generally be expressed as a dissociation constant (KD). Affinity can be measured by common methods known in the art, including those described herein. Low affinity antibodies generally bind antigens slowly and tend to dissociate easily, whereas high affinity antibodies generally bind antigens faster and tend to remain bound longer. Various methods for measuring binding affinity are known in the art, any of which can be used for purposes of the present invention. Specific exemplary and preferred embodiments for measuring binding affinity, i.e., binding strength, are described below.

[0146] As used herein, the terms "Kassoc", "Ka" or "Kon" are intended to refer to the association rate rate of a particular antibody-antigen interaction, whereas the terms "Kdis", "Kd" or "Koff" are intended to refer to the dissociation rate of a particular antibody-antigen interaction. In one embodiment, as used herein, the term "KD" is intended to refer to the dissociation constant obtained from the ratio of Kd to Ka (i.e., Kd / Ka) and expressed as a molar concentration (M). The "KD" or "KD value" or "KD" or "KD value" according to the present invention is, in one embodiment, measured by using a surface plasmon resonance assay. The affinity to PD-L1 was measured by surface plasmon resonance (SPR) measurement as described in section

[0165] . The binding affinity of the multispecific constructs to recombinant human CD3ε ECD, recombinant human IL-23R and recombinant human Her2 ECD was measured by SPR as described in section

[0178] . The affinity of the molecules for human serum albumin (HSA) and mouse serum albumin (MSA) was determined by SPR measurements as described in section

[0179] .

[0147] Suitably, multispecific antibodies of the invention are monovalent, bivalent or multivalent with respect to PD-L1 specificity. In one embodiment, multispecific antibodies of the invention are bivalent with respect to PD-L1 specificity. In a preferred embodiment, multispecific antibodies of the invention are monovalent with respect to PD-L1 specificity.

[0148] Suitable PD-L1-BDs for use in the multispecific antibodies of the invention are the binding domains provided in this disclosure. The PD-L1-BDs of the invention include, but are not limited to, the humanized monoclonal antibodies whose sequences are listed in Table 1.

[0149] Suitably, the multispecific antibodies of the invention are monovalent, bivalent or multivalent for HER2 specificity. In one embodiment, the multispecific antibodies of the invention are bivalent for HER2 specificity. In a preferred embodiment, the multispecific antibodies of the invention are monovalent for HER2 specificity.

[0150] Suitable HER2-BDs for use in the multispecific antibodies of the invention are the binding domains provided in this disclosure. HER2-BDs of the invention include, but are not limited to, the humanized monoclonal antibodies whose sequences are listed in Table 2.

[0151] The term "multivalent antibody" refers to a single binding molecule with more than one valency, where "valency" is described as the number of antigen-binding moieties that bind to an epitope on the same target molecule. Thus, one binding molecule can bind to multiple binding sites on a target molecule. Examples of multivalent antibodies include, but are not limited to, bivalent antibodies, trivalent antibodies, tetravalent antibodies, pentavalent antibodies, and the like.

[0152] As used herein, the term "monovalent antibody" refers to an antibody that binds to a single epitope on a target molecule, such as PD-L1. Also, as used herein, the term "binding domain" or "monovalent binding domain" refers to a binding domain that binds to a single epitope on a target molecule, such as PD-L1.

[0153] As used herein, the term "bivalent antibody" refers to an antibody that binds to two epitopes on at least two identical target molecules, such as a PD-L1 target molecule.

[0154] The inventors of the present invention have now surprisingly found that the addition of the trispecific molecule PRO1678 (anti-HSA x PDL1 x HER2) results in a significant reduction in tumor growth in the HCC1954 xenograft NOG mouse model compared to an equipotent dose of Nivolumab (same activity as determined in vitro). A five-fold lower dose of PRO1678 resulted in the same tumor growth reduction in this model (see FIG. 13). The inventors have further surprisingly found that a tetraspecific molecule comprising a fourth CD3-BD, such as PRO1543 (anti-CD3 x HSA x PDL1 x HER2), results in complete tumor regression in the HCC1954 xenograft NOG mouse model. This finding is surprising because it is not expected from previous experience that in a complex, multi-target, multi-cellular in vivo context, all four binding domains would remain functional without sterically or otherwise inhibiting each other. The EC of PD-L1 blockade in these cells 50 EC of target cell lysis 50 The improved activity of molecules containing anti-PD-L1 domains most likely results from increased avidity, since the binding of Her2 and PD-L1 double positive cells is therefore stronger than that of PD-L1 negative cells expressing Her2. This avidity binding selectively improves efficacy against tumor cells (Her2 / PD-L1 double positive) but not against PD-L1 negative Her2 expressing normal cells, thereby widening the therapeutic window.

[0155] The term "tumor-associated immune checkpoint antigen" refers to a transmembrane protein expressed by tumors that suppresses the activity of immune cells, in particular an antibody selected from the group of PD-L1, PD-L2, CD80, CD86, CD276 (B7-H3), and VTCN1 (B7-H4), more specifically PD-L1.

[0156] The term "low affinity" refers to a binding domain that binds to its cognate target with a dissociation constant of between 50 nM and 2000 nM, preferably between 100 nM and 1000 nM.

[0157] The term "tumor associated antigen (TAA)" refers to an antigen expressed on the surface of a tumor cell. In certain embodiments, a TAA is an antigen that is preferentially expressed on tumor cells compared to non-tumor cells, in particular the expression of the TAA on tumor cells is at least 5-fold higher, at least 10-fold higher, at least 20-fold higher, at least 50-fold higher, or at least 100-fold higher than on non-tumor cells from the same organism or patient. In particular, the TAA is selected from the group consisting of: EGFRvIII, mesothelin, GD2, Tn antigen, sTn antigen, Tn-O-glycopeptide, sTn-O-glycopeptide, PSMA, CD97, TAG72, CD44v6, CEA, EPCAM, KIT, IL-13Ra2, legumain, GD3, CD171, IL-11Ra, PSCA, MAD-CT-1, MAD-CT-2, VEGFR2, Lewis Y, CD24, PDGFR-β, SSEA-4, folate receptor α, ERBB (e.g., ERBB2), Her2 / neu (HER2), MUC1, EGFR, NCAM, ephrin B2, CAIX, LMP2, sLe, HMWMAA, o-acetyl-GD2, folate receptor β, TEM1 / CD248, TEM7R, FAP, legumain, HPV E6 or E7, ML-IAP, CLDN6, TSHR, GPRC5D, ALK, polysialic acid, Fos-related antigen, neutrophil elastase, TRP-2, CYP1B1, sperm protein 17, beta human chorionic gonadotropin, AFP, thyroglobulin, PLAC1, globo H, RAGE1, MN-CA IX, human telomerase reverse transcriptase, intestinal carboxylesterase, mut hsp70-2, NA-17, NY-BR-1, UPK2, HAVCR1, ADRB3, PANX3, NY-ESO-1, GPR20, Ly6k, OR51E2, TARP, and GFRα4.

[0158] The term "immune cell antigen" refers to an antigen present on an immune cell, in particular an immune cell selected from T cells, NK cells and myeloid cells. In particular, this term relates to a protein that is a stimulatory or co-stimulatory molecule of said immune cell.

[0159] In the context of the present invention, the term "stimulatory molecules of said immune cells" relates to molecules such as CD3 and CD16.

[0160] In the context of the present invention, the term "costimulatory molecule" relates to molecules such as those comprised in the group of molecules consisting of CD137, CD28, ICOS, HVEM, CD27, OX40, DR3, GITR, CD30, SLAM, CD2, 2B4, TIM1, TIM2 and CD226.

[0161] In certain embodiments, the multispecific antibody of the invention further comprises (i) a binding domain for CD3, or (ii) a binding domain for CD137.

[0162] Suitable CD3-BDs for use in the multispecific antibodies of the invention are the disclosed binding domains provided in this disclosure. CD3-BDs of the invention include, but are not limited to, the humanized monoclonal antibodies whose sequences are listed in Table 3.

[0163] Suitable CD137-BDs for use in the multispecific antibodies of the invention are the disclosed binding domains provided in this disclosure. CD137-BDs of the invention include, but are not limited to, the humanized monoclonal antibodies whose sequences are listed in Table 5.

[0164] Suitably, the multispecific antibody of the invention has two different specificities (PD-L1 and HER2). Suitably, the multispecific antibody of the invention is a bispecific antibody. The multispecific antibody of the invention may comprise further specificities (trispecific) or multiple specificities (tetraspecific, pentaspecific or hexaspecific antibodies). In one embodiment, the multispecific antibody is trispecific. In another embodiment, the multispecific antibody is tetraspecific.

[0165] In one embodiment, the multispecific antibodies of the invention comprise an immunoglobulin Fc region polypeptide. As used herein, the term "Fc region" is used to define the C-terminal region of an immunoglobulin heavy chain, including native sequence Fc regions and variant Fc regions. Suitable native sequence Fc regions include human lgG1, lgG2 (lgG2A, lgG2B), lgG3 and lgG4. "Fc receptor" or "FcR" describes a receptor that binds to the Fc region of an antibody. A preferred FcR is the native sequence human FcR. Moreover, a preferred FcR binds an IgG antibody (a gamma receptor) and includes receptors of the FcγRI, FcγRII, and FcγRIII subclasses, including allelic variants and alternatively spliced ​​forms of these receptors; FcγRII receptors include FcγRIIA (an "activating receptor") and FcγRIIB (an "inhibiting receptor"), which have similar amino acid sequences that differ primarily in their cytoplasmic domains. Activating receptor FcγRIIA contains an immunoreceptor tyrosine-based activation motif (ITAM) in its cytoplasmic domain. Inhibiting receptor FcγRIIB contains an immunoreceptor tyrosine-based inhibitory motif (ITIM) in its cytoplasmic domain (see M. Daeron, Annu. Rev. Immunol. 5:203-234 (1997)). FcRs are reviewed in Ravetch and Kinet, Annu. Rev. Immunol. 9: 457-92 (1991), Capet et al, Immunomethods 4: 25-34 (1994), and de Haas et al, J. Lab. Clin. Med. 126: 330-41 (1995). Other FcRs, including those identified in the future, are encompassed by the term "FcR" herein. The term "Fc receptor" or "FcR" also encompasses the neonatal receptor, FcRn, responsible for the transfer of maternal IgG to the fetus. Guyer et al., J. Immunol. 117: 587 (1976) and Kim et al., J. Immunol. 24: 249 (1994).Methods for measuring binding to FcRn are known (see, e.g., Ghetie and Ward, Immunol. Today 18: (12): 592-8 (1997); Ghetie et al., Nature Biotechnology 15 (7): 637-40 (1997); Hinton et al., J. Biol. Chem. TJI (8): 6213-6 (2004); WO 2004 / 92219 (Hinton et al)). The in vivo binding to FcRn and serum half-life of human FcRn high affinity binding polypeptides can be assayed, for example, in transgenic mice or transfected human cell lines expressing human FcRn, or in primates to which polypeptides having variant Fc regions are administered. WO 2004 / 42072 (Presta) describes antibody variants with improved or diminished binding to FcR. See, e.g., Shields et al., J. Biol. Chem. 9(2): 6591-6604 (2001).

[0166] In another embodiment, the antibodies of the invention do not comprise an immunoglobulin Fc region polypeptide.

[0167] To increase the number of specificities / functionalities at the same or lower molecular weight, it is advantageous to use antibodies, including antibody fragments such as Fv, Fab, Fab', F(ab')2 fragments, and other antibody fragments. These small molecules retain the antigen-binding activity of the whole antibody and can exhibit improved tissue penetration and pharmacokinetic properties compared to whole immunoglobulin molecules. Although such fragments appear to offer many advantages over whole immunoglobulins, they suffer from an increased clearance rate from serum due to the lack of an Fc domain that confers a long half-life in vivo (Medasan et al., 1997, J. Immunol. 158:2211-2217). Molecules with smaller molecular weights can penetrate target tissues more efficiently and therefore be expected to be more effective at the same or lower doses.

[0168] The inventors of the present invention have surprisingly found that the addition of a human serum albumin binding domain (HSA-BD) to a multispecific antibody of the present invention does not interfere with the ability of the other binding domains to bind to their respective targets. This finding is surprising because previous experience would not predict that all four binding domains would remain functional without sterically or otherwise interfering with each other in a complex, multi-target, multi-cellular in vitro context.

[0169] Suitably, the multispecific antibody of the present invention may comprise a further binding domain having specificity for human serum albumin. In one embodiment, the multispecific antibody comprises (i) at least one PD-L1-BD, (ii) at least one HER2-BD, and (iii) at least one HSA-BD. Suitably, the multispecific antibody of the present invention comprises (i) one PD-L1-BD, (ii) at least one HER2-BD, preferably one PD-L1-BD or two PD-L1-BDs, more preferably one PD-L1-BD, and (iii) at least one HSA-BD, preferably one HSA-BD.

[0170] The term "HSA" specifically refers to human serum albumin with UniProt ID number P02768. Human serum albumin (HSA) is a 66.4 kDa protein composed of 585 amino acids and abundant in human serum (50% of total protein) (Sugio, Protein Eng, Vol. 12, 1999, 439-446). The multifunctional HSA protein is related to its structure that allows it to bind and transport many metabolites, such as fatty acids, metal ions, bilirubin, and some drugs (Fanali, Molecular Aspects of Medicine, Vol. 33, 2012, 209-290). The concentration of HSA in serum is around 3.5-5 g / dL. Albumin-binding antibodies and fragments thereof can be used, for example, to extend the in vivo serum half-life of drugs or proteins bound thereto.

[0171] In some embodiments, the HSA-BD is derived from a monoclonal antibody or antibody fragment.

[0172] Suitable HSA-BDs for use in the multispecific antibodies of the invention are the binding domains provided in this disclosure. HSA-BDs of the invention include, but are not limited to, the humanized monoclonal antibodies whose sequences are listed in Table 4.

[0173] In particular, the HSA-BD of the present invention specifically binds to human serum albumin. The HSA-BD of the present invention comprises a VH CDR having the amino acid sequence of any one of the VH CDRs listed in Table 4. In particular, the present invention provides an HSA-BD comprising one, two, three or more VH CDRs having the amino acid sequence of any of the VH CDRs listed in Table 4.

[0174] The present invention also provides HSA-BDs comprising a VL CDR having the amino acid sequence of any one of the VL CDRs listed in Table 4. In particular, the present invention provides HSA-BDs comprising one, two, three or more VL CDRs having the amino acid sequence of any of the VL CDRs listed in Table 4.

[0175] In a further embodiment, the present invention provides an HSA-BD that specifically binds to human serum albumin, said binding domain comprising a VH domain and a VL domain.

[0176] Another suitable HSA-BD for use in the multispecific antibodies of the invention is (i) a polypeptide that binds serum albumin (see, for example, Smith et al., 2001, Bioconjugate Chem. 12:750-756, EP 0486525, U.S. Pat. No. 6,267,964, WO 2004 / 001064, WO 2002 / 076489, and WO 2001 / 45746); (ii) a polypeptide that binds serum albumin (see, for example, Smith et al., Protein Engineering, Design & Selection, vol 21, 5, pp 283-288, WO 2004 / 003019, WO 2008 / 096158, WO 2005 / 118642, WO 2006 / 0591056 and WO 2011 / 006915, (iii) an antibody selected from the group consisting of the anti-serum albumin binding single variable domains described in WO 2009 / 040562, WO 2010 / 035012 and WO 2011 / 086091.

[0177] In a particular embodiment, the multispecific antibody of the invention comprises an HSA-binding domain having the CDR sequences set forth in SEQ ID NOs: 61-66 and the VH / VL sequences set forth in SEQ ID NOs: 67-70.

[0178] These HSA-BDs exhibit particularly advantageous properties, such as high stability and cross-reactivity, especially towards Cynomolgus Serum Albumin (CSA) and Mouse Serum Albumin (MSA), further improving the already advantageous properties of the multispecific antibodies of the invention. More specifically, said HSA-BDs are characterized by one or more of the following parameters: a. A monovalent dissociation constant (K) of less than 20 nM at a pH value of about 5.5, as measured by surface plasmon resonance (SPR) D ), especially 0.01 to 20 nM, especially 0.05 to 10 nM, especially 0.1 to 5 nM K Dwhich binds to human serum albumin (HSA), in particular said hSA-BD is in scFv format; b. A monovalent dissociation constant (K) of less than 20 nM at pH values ​​of about 5.5 and about 7.4, as measured by surface plasmon resonance (SPR). D ), especially 0.01 to 20 nM, especially 0.05 to 10 nM, especially 0.1 to 5 nM K D which binds to human serum albumin (HSA), in particular said HSA-BD is in scFv format; c. Cross-reactive with Macaca fascicularis (cynomolgus monkey) serum albumin (CSA), particularly at pH values ​​of about 5.5 and about 7.4, and particularly at a monovalent K of less than 15 nM, as measured by SPR. D , especially 0.01 to 15 nM, especially 0.05 to 7 nM, especially 0.1 to 4 nM D binds to CSA in a scFv format, in particular said HSA-BD; d. Cross-reactive with Mus musculus (mouse) serum albumin (MSA), particularly at pH values ​​of about 5.5, and with a monovalent K of less than 20 nM, as measured by SPR. D , especially 0.01 to 20 nM, especially 0.05 to 10 nM, especially 0.1 to 5 nM D binds to MSA with HSA-BD, particularly said HSA-BD is in scFv format; e. HSA bound to antibody retains the ability to bind to FcRn; if in f.scFv format, it has a melting temperature (Tm) of at least 72°C, preferably at least 75°C, more preferably at least 78°C, as measured by differential scanning fluorimetry (DSF), particularly where said HSA-BD is formulated in 50 mM citrate phosphate buffer, pH 6.4, containing 150 mM NaCl; In the case of the g.scFv format, after storage at 4°C for at least two weeks, the decrease in monomer content is less than 5%, such as less than 4%, less than 3%, less than 2%, preferably less than 1%, the antigen-binding fragment has an initial concentration of 50 mg / ml, and in particular the HSA-BD is formulated in a 50 mM citrate phosphate buffer containing pH 6.4 and 150 mM NaCl; In the case of the h.scFv format, after storage at 40°C for at least two weeks, the decrease in monomer content is less than 11%, such as less than 8%, less than 5%, less than 2%, preferably less than 1%, the antigen-binding fragment has an initial concentration of 10 mg / ml, and in particular the HSA-BD is formulated in a 50 mM citrate phosphate buffer containing pH 6.4 and 150 mM NaCl; and / or In the case of the i.scFv format, after storage at 40°C for at least four weeks, the decrease in protein content is less than 5%, such as less than 4%, less than 3%, less than 2%, preferably less than 1%, the antigen-binding fragment has an initial concentration of 10 mg / ml, and in particular the HSA-BD is formulated in a 50 mM citrate phosphate buffer containing pH 6.4 and 150 mM NaCl.

[0179] Other variable domains of the present invention contain amino acids having mutations but having at least 60, 70, 80, 90, 91, 92, 93, 94, 95, 96, 97, 98 or 99 percent identity to the CDR regions shown in the sequences described in Tables 1 to 5 in the CDR regions. Other variable domains of the present invention contain mutant amino acid sequences in which 1, 2, 3, 4 or 5 or fewer amino acids are mutated in the CDR regions as compared to the CDR regions shown in the sequences described in Tables 1 to 5.

[0180] Suitably, the VH domain of the binding domain of the invention belongs to the VH3 or VH4 family. In one embodiment, the binding domain of the invention comprises a VH domain belonging to the VH3 family. In the context of the present invention, the term "belonging to the VHx family (or VLx family)" means that the framework sequences FR1 to FR2 show the highest degree of homology to said VHx family (or VLx, respectively). Examples of VH and VL families are described in Knappik et al., J. Mol. Biol. 296 (2000) 57-86, or in WO 2019 / 057787. A specific example of a VH domain belonging to the VH3 family is shown in SEQ ID NO: 142, and a specific example of a VH domain belonging to the VH4 family is shown in SEQ ID NO: 143. In particular, the framework regions FR1 to FR4 taken from SEQ ID NO: 142 belong to the VH3 family (Table 7, regions marked in non-bold). Suitably, as used herein, a VH belonging to the VH3 family is a VH comprising FR1 to FR4 having at least 85%, preferably at least 90%, more preferably at least 95% sequence identity to FR1 to FR4 of SEQ ID NO: 142. Examples of alternative VH3 sequences and examples of VH4 sequences can be found in Knappik et al., J. Mol. Biol. 296 (2000) 57-86 or WO 2019 / 057787. Suitably, the HSA-BD of the invention comprises: Vκ framework FR1, FR2 and FR3, in particular Vκ1 or Vκ3 frameworks, preferably Vκ1 framework FR1 to 3, and Vκ FR4, in particular a framework FR4 selected from Vκ1 FR4, Vκ3 FR4, and Vλ FR4. A suitable Vκ1 framework FR1 to 3 is set out in SEQ ID NO: 144 (Table 7, FR regions shown in non-bold). Alternative examples of Vκ1 sequences and examples of Vκ2, Vκ3 or Vκ4 sequences can be found in Knappik et al., J. Mol. Biol. 296 (2000) 57-86.Suitable Vκ1 frameworks FR1-3 correspond to FR1-3 and comprise an amino acid sequence having at least 60, 70, 80, 90 percent identity to an amino acid sequence taken from SEQ ID NO: 144 (Table 7, FR regions shown in non-bold). Suitable Vλ FR4s are as set forth in SEQ ID NO: 145-152. In one embodiment, the VL domain of the present invention comprises a Vλ FR4 comprising an amino acid sequence having at least 60, 70, 80, 90 percent identity to an amino acid sequence selected from any of SEQ ID NO: 145-152, preferably SEQ ID NO: 146 or 152.

[0181] Binding domains of the invention include VH domains listed in Tables 1-5. Suitably, binding domains of the invention include a VH amino acid sequence listed in one of Tables 1-5, in which about 10 or less amino acids in the framework sequences (e.g., sequences that are not CDRs) have been mutated (mutations can be, for various non-limiting examples, additions, substitutions, or deletions). Suitably, binding domains of the invention include a VH amino acid sequence listed in one of Tables 1-5, in which about 20 or less amino acids in the framework sequences (e.g., sequences that are not CDRs) have been mutated (mutations can be, for various non-limiting examples, additions, substitutions, or deletions). Other binding domains of the invention include amino acid sequences in which mutations have been introduced, but which have at least 60, 70, 80, 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identity in the VH region to the VH region shown in the corresponding sequence in one of Tables 1-5.

[0182] In particular, the binding domain of the invention comprises a VL domain listed in one of Tables 1-5. Suitably, the binding domain of the invention comprises a VL amino acid sequence listed in one of Tables 1-5, in which about 10 or less amino acids in the framework sequences (e.g., sequences that are not CDRs) have been mutated (mutations can be, for various non-limiting examples, additions, substitutions, or deletions). Suitably, the binding domain of the invention comprises a VL amino acid sequence listed in one of Tables 1-5, in which about 20 or less amino acids in the framework sequences (e.g., sequences that are not CDRs) have been mutated (mutations can be, for various non-limiting examples, additions, substitutions, or deletions). Other binding domains of the invention comprise amino acid sequences in which mutations have been introduced but which have at least 60, 70, 80, 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identity in the VL region to the VL region shown in the sequence in Tables 1-5.

[0183] In the context of the present invention, the term "binding domain of the invention" relates both to such a binding domain, i.e. outside the context of multispecificity, and in particular to a binding domain comprised in a multispecific construct, e.g. one of the binding domains comprised in a bispecific, trispecific or tetraspecific construct.

[0184] Suitably, the binding domain of the present invention is selected from the group consisting of a Fab, Fv, scFv, dsFv, scAb, and STAB.

[0185] Suitably, the binding domain of the present invention is an scFv antibody fragment.

[0186] The multispecific antibodies of the invention can be in any suitable format.

[0187] Suitably, the binding domains of a multispecific antibody are operatively linked. The binding domains of a multispecific antibody of the invention are capable of binding simultaneously to their respective antigens or receptors.

[0188] In one embodiment, a multispecific antibody of the invention comprises at least one PD-L1-BD, at least one HER2-BD, and (i) the PD-L1-BD and the HER2-BD are both operably linked to each other. In one embodiment, a multispecific antibody of the invention comprises at least one PD-L1-BD, at least one HER2-BD, and at least one HSA-BD, and (i) the PD-L1-BD and the HER2-BD are both operably linked to the HSA-BD, or (ii) the PD-L1-BD and the HSA-BD are both operably linked to the HER2-BD, or (iii) the HER2-BD and the HSA-BD are both operably linked to the PD-L1-BD. In a preferred embodiment, the multispecific antibody of the invention comprises at least one PD-L1-BD, at least one HER2-BD and at least one HSA-BD, wherein said PD-L1-BD and said HSA-BD are both operably linked to said HER2-BD.

[0189] As used herein, the term "operably linked" refers to two molecules (e.g., polypeptides, domains, binding domains) linked together such that each retains functional activity. Two molecules can be "operably linked" whether they are directly or indirectly linked (e.g., via a linker, via a moiety, via a linker to a moiety). The term "linker" refers to a peptide or other moiety that is optionally placed between the binding domains or antibody fragments of the invention. Many methods can be used to covalently link molecules together. These include, but are not limited to, a polypeptide bond between the N-terminus and C-terminus of a protein or protein domain, linkage via a disulfide bond, and linkage by a chemical cross-linking reagent. In one aspect of this embodiment, the linker is a peptide bond and is generated by recombinant techniques or peptide synthesis. The selection of a linker appropriate for a particular case in which two polypeptide chains are to be joined depends on various parameters, including, but not limited to, the nature of the two polypeptide chains (e.g., whether they naturally oligomerize), the distance between the N-terminus and C-terminus to be connected, if known, and the stability of the linker against proteolysis and oxidation. Additionally, the linker can contain amino acid residues that provide flexibility.

[0190] In the context of the present invention, the term "polypeptide linker" refers to a linker consisting of a chain of amino acid residues linked by peptide bonds connecting two domains, each of which is attached to one end of the linker. The polypeptide linker should be of sufficient length to link two molecules such that they adopt the correct conformation relative to each other and thus retain the desired activity. In certain embodiments, the polypeptide linker has a continuous chain of 2 to 30 amino acid residues (e.g., 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30 amino acid residues). Furthermore, the amino acid residues selected for inclusion in the polypeptide linker should exhibit properties that do not significantly interfere with the activity of the polypeptide. Thus, the linker peptide as a whole should not exhibit charges that would be incompatible with the activity of the polypeptide, or should not interfere with internal folding, or should not form bonds or other interactions with amino acid residues in one or more monomers that would significantly impede the binding of the receptor monomer domains. In certain embodiments, the polypeptide linker is a non-structured polypeptide. Useful linkers include glycine-serine, or GS linkers. By "Gly-Ser" or "GS" linker is meant flexible linkers such as tandem glycine and serine polymers (e.g., (Gly-Ser)n, (GSGGS)n, (GGGGS)n, and (GGGS)n, where n is an integer of at least 1), glycine-alanine polymers, alanine-serine polymers, and tethers for Shaker potassium channels, as well as many other flexible linkers, as would be understood by one of skill in the art. Glycine-serine polymers are preferred because both of these amino acids are relatively unstructured and therefore may function as neutral tethers between components. Second, serine is hydrophilic and therefore can solubilize what might be globular glycine chains. Third, similar chains have been shown to be effective in joining subunits of recombinant proteins such as single chain antibodies.

[0191] Suitably, the multispecific antibody is in a format selected from any suitable multispecificity known in the art, such as a bispecific format, such as, by way of non-limiting example, a format based on single chain diabody (scDb), tandem scDb (Tandab), linear dimeric scDb (LD-scDb), cyclic dimeric scDb (CD-scDb), bispecific T cell engaging antibody (BiTE; tandem di-scFv), tandem tri-scFv, tribody (Fab-(scFv)2) or bibody (Fab-(scFv)1), Fab, Fab-Fv2, Morrison (IgG CH3-scFv fusion (Morrison L) or IgG CL-scFv fusion (Morrison H)), triabody, scDb-scFv, bispecific Fab2, di-miniantibody, tetrabody, scFv-Fc-scFv fusion, scFv-HSA-scFv fusion, di-diabody, DVD-Ig, COVD, IgG-scFab, scFab-dsscFv, Fv2-Fc, bsAb (scFv linked to the C-terminus of the light chain), Bs1Ab (scFv linked to the N-terminus of the light chain), Bs2Ab (scFv linked to the N-terminus of the heavy chain), Bs3Ab (scFv linked to the C-terminus of the heavy chain), Ts1Ab (scFv linked to the N-terminus of both the heavy and light chains), Ts2Ab (heavy bispecific antibodies based on heterodimeric Fc domains such as Knob-into-Hole antibodies (KiHs) (bispecific IgG prepared by KiH technology); Fv, scFv, scDb, tandem-di-scFv, tandem tri-scFv, Fab-(scFv)2, Fab-(scFv)1, Fab, Fab-Fv2, COVD, MATCH (WO 2016 / 0202457, Egan et al.) fused to the N- and / or C-terminus of either chain of a heterodimeric Fc domain or any other heterodimerization domain. T., et al., mAbs 9 (2017) 68-84) as well as duobodies (bispecific IgG prepared by duobody technology) (MAbs. 2017 Feb / Mar;9(2):182-212. doi: 10.1080 / 19420862.2016.1268307).Particularly suitable for use herein are single chain diabodies (scDb) or scDb-scFv.

[0192] In one embodiment, the multispecific antibody of the present invention is in a format selected from the list consisting of scDb (diabody), scDb-scFv, triabody, and tribody. Particularly suitable for use herein are single chain diabodies (scDb), especially bispecific monomeric scDb. Also particularly suitable for use herein are scDb-scFv, especially where the CD137-BD and the PD-L1-BD are in the form of a scDb, and the HSA-BD is operably linked to the scDb.

[0193] The term "diabody" refers to an antibody fragment with two antigen-binding sites, a fragment comprising a VH linked to a VL in the same polypeptide chain (VH-VL). By using a linker that is too short to allow pairing between the two domains on the same chain, the domains are forced to pair with complementary domains on another chain, creating two antigen-binding sites. Diabodies can be bivalent or bispecific. Diabodies are described more fully in, for example, EP 404097, WO 93 / 01161, Hudson et al., Nat. Med. 9:129-134 (2003), and Hollinger et al., Proc. Natl. Acad. Sci. USA 90: 6444-6448 (1993). Triabodies and tetrabodies are also described in Hudson et al., Nat. Med. 9:129-134 (2003).

[0194] Bispecific scDbs, in particular bispecific monomeric scDbs, can be represented in particular by the linkers L1, L2 and L3 as VHA-L1-VLB-L2-VHB-L3-VLA, VHA-L1-VHB-L2-VLB-L3-VLA, VLA-L1-VLB-L2-VHB-L3-VHA, VLA-L1-VHB-L2-VLB-L3-VHA, VHB-L1-VLA-L2-VHA-L3-VLB, VHB-L1-VHA-L It comprises two variable heavy domains (VH) or fragments thereof and two variable light domains (VL) or fragments thereof linked in the order 2-VLA-L3-VLB, VLB-L1-VLA-L2-VHA-L3-VHB or VLB-L1-VHA-L2-VLA-L3-VHB, wherein the VLA and VHA domains together form an antigen-binding site for a first antigen and the VLB and VHB together form an antigen-binding site for a second antigen.

[0195] Linker L1 is in particular a peptide of 2 to 10 amino acids, more in particular 3 to 7 amino acids, most in particular 5 amino acids, and linker L3 is in particular a peptide of 1 to 10 amino acids, more in particular 2 to 7 amino acids, most in particular 5 amino acids. In a particular embodiment, linker L1 and / or L3 comprises one or two units of four (4) glycine amino acid residues and one (1) serine amino acid residue (GGGGS)n, where n=1 or 2, preferably n=1.

[0196] The intermediate linker L2 is in particular a peptide of 10 to 40 amino acids, more particularly 15 to 30 amino acids, most particularly 20 to 25 amino acids. In a particular embodiment, said linker L2 comprises one or more units (GGGGS)n of four (4) glycine amino acid residues and one (1) serine amino acid residue, where n=1, 2, 3, 4, 5, 6, 7 or 8, preferably n=4.

[0197] In one embodiment, the multispecific antibody of the invention is a scDb-scFv. The term "scDb-scFv" refers to an antibody format in which a single chain Fv (scFv) fragment is fused to a single chain diabody (scDb) by a flexible Gly-Ser linker. In one embodiment, said flexible Gly-Ser linker is a peptide of 2-40 amino acids, such as 2-35, 2-30, 2-25, 2-20, 2-15, 2-10 amino acids, in particular 10 amino acids. In a particular embodiment, said linker comprises one or more units (GGGGS)n of four (4) glycine amino acid residues and one (1) serine amino acid residue, where n=1, 2, 3, 4, 5, 6, 7 or 8, preferably n=2.

[0198] In one embodiment of the invention, the multispecific antibody of the invention is in the MATCH format as described in WO 2016 / 0202457, Egan T., et al., mAbs 9 (2017) 68-84.

[0199] The multispecific antibodies of the invention can be produced using any convenient antibody production method known in the art (see, for example, Fischer, N. & Leger, O., Pathobiology 74 (2007) 3-14 for bispecific construct production; Hornig, N. & Faerber-Schwarz, A., Methods Mol. Biol. 907 (2012) 713-727 for bispecific diabodies and tandem scFvs, and WO 99 / 57150). Specific examples of methods suitable for the preparation of bispecific constructs of the invention further include, inter alia, Genmab (see Labrijn et al., Proc. Natl. Acad. Sci. USA 110 (2013) 5145-5150) and Merus (see de Kruif et al., Biotechnol. Bioeng. 106 (2010) 741-750) technologies. Methods for making bispecific antibodies containing functional antibody Fc portions are also known in the art (see, e.g., Zhu et al., Cancer Lett. 86 (1994) 127-134, and Suresh et al., Methods Enzymol. 121 (1986) 210-228).

[0200] Typically, these methods involve generating monoclonal antibodies, for example by using hybridoma technology to fuse myeloma cells from mice immunized with the desired antigen with spleen cells (see, e.g., Yokoyama et al., Curr. Protoc. Immunol. Chapter 2, Unit 2.5, 2006) or by recombinant antibody engineering (repertoire cloning or phage display / yeast display) (see, e.g., Chames & Baty, FEMS Microbiol. Letters 189 (2000) 1-8), and combining the antigen-binding domains, or fragments or portions thereof, of two or more different monoclonal antibodies using known molecular cloning techniques to obtain bispecific or multispecific constructs.

[0201] Multispecific molecules of the invention can be prepared by conjugating the constituent binding specificities using methods known in the art. For example, each binding specificity of a bispecific molecule can be generated separately and then conjugated to one another. When the binding specificities are proteins or peptides, a variety of coupling or cross-linking agents can be used for covalent conjugation. Examples of cross-linking agents include protein A, carbodiimide, N-succinimidyl-5-acetyl-thioacetate (SATA), 5,5'-dithiobis(2-nitrobenzoic acid) (DTNB), o-phenylenedimaleimide (oPDM), N-succinimidyl-3-(2-pyridyldithio)propionate (SPDP), and sulfosuccinimidyl 4-(N-maleimidomethyl)cyclohexane-1-carboxylate (sulfo-SMCC) (see, e.g., Karpovsky et al., 1984 J. Exp. Med. 160: 1686; Liu, MA et al., 1985 Proc. Natl. Acad. Sci. USA 82:8648). Other methods include those described by Paulus, 1985 Behring Ins. Mitt. No. 78, 118-132; Brennan et al., 1985 Science 229:81-83; and Glennie et al., 1987 J. Immunol. 139: 2367-2375. Conjugating agents are SATA and sulfo-SMCC, both available from Pierce Chemical Co. (Rockford, Ill.).

[0202] When the binding specificity is an antibody, it can be attached by sulfhydryl bonding at the C-terminal hinge regions of the two heavy chains, hi certain embodiments, the hinge region is modified to contain an odd number of sulfhydryl residues, e.g., one, prior to attachment.

[0203] Alternatively, two or more binding specificities can be encoded in the same vector and expressed and assembled in the same host cell. This method is particularly useful when the bispecific molecule is a mAb x mAb, mAb x Fab, Fab x F(ab')2 or ligand x Fab fusion protein. The multispecific antibody of the invention can be a single chain molecule comprising one single chain antibody and a binding determinant, or a single chain multispecific antibody comprising two binding determinants. The multispecific antibody can comprise at least two single chain molecules. Methods for preparing multispecific antibodies and molecules are described, for example, in U.S. Pat. Nos. 5,260,203, 5,455,030, 4,881,175, 5,132,405, 5,091,513, 5,476,786, 5,013,653, 5,258,498, and 5,482,858.

[0204] Binding of a multispecific antibody to its specific targets can be confirmed, for example, by enzyme-linked immunosorbent assay (ELISA), radioimmunoassay (REA), FACS analysis, bioassay (e.g., growth inhibition), or Western blot assay. Each of these assays generally detects the presence of a protein-antibody complex of particular interest by using a labeled reagent (e.g., an antibody) specific for the complex of interest.

[0205] In a further aspect, the present invention provides a nucleic acid encoding a multispecific antibody of the invention or a fragment thereof or a binding domain thereof. Such a sequence nucleic acid can be optimized for expression in a mammalian cell.

[0206] As used herein, the term "nucleic acid" is used interchangeably with the term "polynucleotide(s)" and refers to one or more deoxyribonucleotides or ribonucleotides and polymers thereof in either single-stranded or double-stranded form. The term encompasses nucleic acids containing known nucleotide analogs or modified backbone residues or linkages, which are synthetic, naturally occurring, and non-naturally occurring, have similar binding properties as the reference nucleic acid, and are metabolized in a similar manner as the reference nucleotide. Examples of such analogs include, but are not limited to, phosphorothioates, phosphoramidates, methyl phosphonates, chiral-methyl phosphorates, 2-O-methyl ribonucleotides, peptide-nucleic acids (PNAs), and the like. Unless otherwise indicated, a particular sequence nucleic acid also implicitly encompasses its conservatively modified variants (e.g., degenerate codon substitutions) and complementary sequences, as well as the sequence explicitly indicated. Specifically, as described in more detail below, degenerate codon substitutions can be achieved by generating sequences in which the third position of one or more selected (or all) codons is substituted with mixed-base and / or deoxyinosine residues (Batzer et al., Nucleic Acid Res. 19:5081, 1991; Ohtsuka et al., J. Biol. Chem. 260:2605-2608, 1985; and Rossolini et al., Mol. Cell. Probes 8:91-98, 1994).

[0207] The present invention provides substantially purified nucleic acid molecules encoding polypeptides comprising fragments or domains of the multispecific antibodies described above. When expressed from a suitable expression vector, the polypeptides encoded by these nucleic acid molecules are capable of exhibiting the antigen-binding capacity or capacities of the multispecific antibodies of the invention.

[0208] The present invention also provides polynucleotides encoding at least one CDR region, and typically all three CDR regions, of the binding domains of the multispecific antibodies of the present invention listed in Tables 1 to 4. Due to the degeneracy of the code, different nucleic acid sequences will encode each immunoglobulin amino acid sequence.

[0209] Polynucleotide sequences can be generated by de novo solid-phase DNA synthesis or by PCR mutagenesis of existing sequences (e.g., sequences described in the Examples below) encoding the multispecific antibodies of the invention or fragments thereof or binding domains thereof. Direct chemical synthesis of nucleic acids can be achieved by methods known in the art, such as the phosphotriester method of Narang et al., 1979, Meth. Enzymol. 68:90, the phosphodiester method of Brown et al., Meth. Enzymol. 68: 109, 1979, the diethyl phosphoramidite method of Beaucage et al., Tetra. Lett., 22: 1859, 1981, and the solid support method of U.S. Pat. No. 4,458,066. Introducing mutations into polynucleotide sequences by PCR is described, for example, in DNA Amplification, H. A. Erlich (Ed.), Freeman Press, NY, NY, 1992, PCR Protocols: A Guide to Methods and Applications, Innis et al. (Ed.), Academic Press, San Diego, Calif, 1990, Mattila et al., Nucleic Acids Res. 19:967, 1991, and Eckert et al., PCR Methods and Applications 1:17, 1991.

[0210] The present invention also provides expression vectors and host cells for producing the multispecific antibodies of the present invention, or fragments thereof, or binding domains thereof.

[0211] The term "vector" is intended to refer to a polynucleotide molecule capable of transporting another polynucleotide to which it is linked. One type of vector is a "plasmid", which refers to a circular double-stranded DNA loop into which additional DNA segments can be ligated. Another type of vector is a viral vector, into which additional DNA segments can be ligated into the viral genome. Certain vectors can replicate autonomously in a host cell into which they are introduced (e.g., bacterial vectors having a bacterial origin of replication and episomal mammalian vectors). Other vectors (e.g., non-episomal mammalian vectors) can integrate into the genome of the host cell upon introduction into the host cell, and thereby are replicated along with the host genome.

[0212] Moreover, certain vectors are capable of inducing the expression of genes to which they are operatively linked. Such vectors are referred to herein as "recombinant expression vectors" (or simply "expression vectors"). In general, expression vectors useful in recombinant DNA techniques are often in the form of plasmids. As used herein, "plasmid" and "vector" may be used interchangeably, as the plasmid is the most commonly used form of vector. However, the invention is intended to include such other forms of expression vectors, such as viral vectors (e.g., replication defective retroviruses, adenoviruses, adeno-associated viruses), which perform equivalent functions. In this particular context, the term "operably linked" refers to a functional relationship between two or more polynucleotide (e.g., DNA) fragments. Typically, it refers to the functional relationship between a transcriptional regulatory sequence and a transcribed sequence. For example, a promoter or enhancer sequence is operably linked to a coding sequence if it stimulates or regulates the transcription of the coding sequence in an appropriate host cell or other expression system. Generally, a promoter transcriptional regulatory sequence operably linked to a transcribed sequence is physically contiguous to the transcribed sequence, i.e., cis-acting. However, some transcriptional regulatory sequences, such as enhancers, need not be physically contiguous or located in close proximity to the coding sequences whose transcription they promote.

[0213] A variety of expression vectors can be employed to express polynucleotides encoding multispecific antibody chains or binding fragments. Both viral-based and non-viral expression vectors can be used to produce antibodies in mammalian host cells. Non-viral vectors and systems include plasmids, episomal vectors that typically carry expression cassettes for expressing proteins or RNA, and human artificial chromosomes (see, e.g., Harrington et al., Nat Genet. 15:345, 1997). For example, non-viral vectors useful for expressing CD137-binding polynucleotides and polypeptides in mammalian (e.g., human) cells include pThioHis A, B and C, pcDNA3.1 / His, pEBVHis A, B and C, (Invitrogen, San Diego, Calif.), MPS V vectors, and many other vectors known in the art for expressing other proteins. Useful viral vectors include vectors based on retroviruses, adenoviruses, adeno-associated viruses, herpes viruses, SV40, papilloma viruses, HBP Epstein-Barr virus-based vectors, vaccinia virus vectors and Semliki Forest virus (SFV). See Brent et al., supra; Smith, Annu. Rev. Microbiol. 49:807, 1995; and Rosenfeld et al., Cell 68: 143, 1992.

[0214] The choice of expression vector depends on the host cell in which the vector is to be expressed. Typically, expression vectors contain a promoter and other regulatory sequences (e.g., enhancers) operably linked to the polynucleotide encoding the multispecific antibody chain or fragment. In one embodiment, an inducible promoter is employed to ensure that the inserted sequence is not expressed except under inducing conditions. Inducible promoters include, for example, arabinose, lacZ, metallothionein promoters or heat shock promoters. Cultures of transformed organisms can be grown under non-inducing conditions without biasing the population towards coding sequences whose expression products are better tolerated by the host cell. In addition to promoters, other regulatory elements may also be required or desired for efficient expression of the multispecific antibody chain or fragment. Typically, these elements include an ATG initiation codon and adjacent ribosome binding sites or other sequences. Furthermore, the efficiency of expression can be increased by including enhancers appropriate for the cell system used (see, e.g., Scharf et al., Results Probl. Cell Differ. 20: 125, 1994, and Bittner et al., Meth. Enzymol., 153:516, 1987). For example, the SV40 enhancer or CMV enhancer can be used to increase expression in mammalian host cells.

[0215] The expression vector may also provide a secretion signal sequence position to form a fusion protein with the polypeptide encoded by the inserted multispecific antibody of the invention or a fragment thereof or a binding domain sequence thereof. More often, the inserted multispecific antibody of the invention or a fragment thereof or a binding domain sequence thereof is linked to a signal sequence before inclusion in the vector. The vector used to receive the sequences encoding the binding domains of the multispecific antibody light and heavy chain variable domains sometimes also encodes the constant regions or parts thereof. Such vectors allow the variable regions to be expressed as fusion proteins with the constant regions, thereby producing intact antibodies and antigen-binding fragments thereof. Typically, such constant regions are human.

[0216] The term "recombinant host cell" (or simply "host cell") refers to a cell into which a recombinant expression vector has been introduced. It should be understood that such terms are intended to refer not only to the particular subject cell but to the progeny of such a cell. Because certain changes may occur in the progeny, either due to mutation or environmental influences, such progeny may not in fact be identical to the parent cell, but are still included within the scope of the term "host cell" as used herein.

[0217] Host cells for harboring and expressing the multispecific antibodies or fragments thereof or binding domains thereof of the invention can be prokaryotic or eukaryotic. Escherichia coli is one prokaryotic host useful for cloning and expressing the polynucleotides of the invention. Other microbial hosts suitable for use include bacilli, such as Bacillus subtilis, and other Enterobacteriaceae, such as Salmonella, Serratia, and various Pseudomonas species. In these prokaryotic hosts, expression vectors can also be made, which typically contain expression control sequences (e.g., an origin of replication) compatible with the host cell. In addition, there will be a large number of different well-known promoters, such as the lactose promoter system, the tryptophan (trp) promoter system, the β-lactamase promoter system, promoter systems from phage lambda, and the like. Typically, the promoter controls expression, optionally has an operator sequence, and has ribosome binding site sequences and the like for initiating and completing transcription and translation. Other microorganisms, such as yeast, can also be employed for the expression of the CD137-binding polypeptides of the invention. Additionally, insect cells can also be used in combination with baculovirus vectors.

[0218] In one embodiment, mammalian host cells are used to express and produce the multispecific antibodies or fragments thereof or binding domains thereof of the invention. For example, the cells may be either hybridoma cell lines expressing endogenous immunoglobulin genes or mammalian cell lines carrying exogenous expression vectors. These include normal mortal or normal or abnormal immortal animal or human cells. Numerous suitable host cell lines capable of secreting intact immunoglobulins have been developed, including, for example, CHO cell lines, various Cos cell lines, HeLa cells, myeloma cell lines, transformed B cells, and hybridomas. The use of mammalian tissue cell culture to express polypeptides is generally discussed in, for example, Winnacker, FROM GENES TO CLONES, VCH Publishers, NY, NY, 1987. Expression vectors for mammalian host cells can include an origin of replication, expression control sequences such as promoters, enhancers (see, e.g., Queen, et al., Immunol. Rev. 89:49-68, 1986), ribosome binding sites, and necessary processing information sites such as RNA splice sites, polyadenylation sites, and transcription termination sequences. These expression vectors usually contain a promoter derived from a mammalian gene or a mammalian virus. Suitable promoters can be constitutive, cell type-specific, stage-specific, and / or regulatable or controllable. Useful promoters include, but are not limited to, metallothionein promoter, constitutive adenovirus major late promoter, dexamethasone-inducible MMTV promoter, SV40 promoter, MRP polIII promoter, constitutive MPS V promoter, tetracycline-inducible CMV promoter (such as human immediate early CMV promoter), constitutive CMV promoter, and promoter-enhancer combinations known in the art.

[0219] Methods for introducing expression vectors containing polynucleotide sequences of interest vary depending on the type of cellular host. For example, calcium chloride transfection is widely used for prokaryotic cells, while calcium phosphate treatment or electroporation may be used for other cellular hosts (see generally Sambrook, et al., supra). Other methods include, for example, electroporation, calcium phosphate treatment, liposome-mediated transformation, injection and microinjection, ballistic methods, virosomes, immunoliposomes, polycation / nucleic acid conjugates, naked DNA, artificial virions, fusion to herpes virus structural protein VP22 (Elliot and O'Hare, Cell 88:223, 1997), drug-enhanced uptake of DNA, and extracellular transduction. For long-term, high-yield production of recombinant proteins, stable expression will often be desired. For example, cell lines stably expressing the multispecific antibodies of the invention or fragments thereof or binding domains thereof can be prepared using the expression vectors of the invention containing viral origins of replication or endogenous expression elements and a selectable marker gene. Following introduction of the vector, the cells may be cultured in enriched medium for 1-2 days and then switched to selective medium. The purpose of the selection marker is to confer resistance to selection, such that its presence allows cells that successfully express the introduced sequences to be cultured in selective medium. Resistant, stably transfected cells can be propagated using tissue culture techniques appropriate to the cell type. Thus, the invention provides a method of making an antibody or antigen-binding fragment thereof of the invention, the method comprising the step of culturing a host cell comprising a nucleic acid or vector encoding an antibody or antigen-binding fragment thereof of the invention, whereby the antibody or fragment thereof of the present disclosure is expressed.

[0220] In one aspect, the invention relates to a method for making a multispecific antibody of the invention or a binding domain or fragment thereof, the method comprising culturing a host cell expressing a nucleic acid encoding a multispecific antibody of the invention or a binding domain or fragment thereof.

[0221] In a further aspect, the invention relates to a pharmaceutical composition comprising a multispecific antibody of the invention and a pharma- ceutically acceptable carrier, which enhances or stabilizes the composition or facilitates preparation of the composition. Pharmaceutically acceptable carriers include physiologically compatible solvents, dispersion media, coatings, antibacterial and antifungal agents, isotonic and absorption delaying agents, and the like.

[0222] The pharmaceutical composition of the present invention can be administered by various methods known in the art. The route and / or mode of administration varies depending on the desired effect. Administration can be intravenous, intramuscular, intraperitoneal, or subcutaneous, or can be performed in close proximity to the target site. The pharmaceutically acceptable carrier should be suitable for intravenous, intramuscular, subcutaneous, parenteral, spinal or epidermal administration (e.g., by injection or infusion). Depending on the route of administration, the active compound, i.e., the multispecific antibody of the present invention, can be coated with a material to protect it from the action of acids and other natural conditions that may inactivate the compound.

[0223] The pharmaceutical compositions of the present invention can be prepared according to methods well known and routinely practiced in the art. See, for example, Remington: The Science and Practice of Pharmacy, Mack Publishing Co., 20th ed., 2000, and Sustained and Controlled Release Drug Delivery Systems, J.R. Robinson, ed., Marcel Dekker, Inc., New York, 1978. The pharmaceutical compositions are preferably manufactured under GMP conditions. Typically, a therapeutically effective dose or an efficacious dose of the multispecific antibodies of the present invention is employed in the pharmaceutical compositions of the present invention. The multispecific antibodies of the present invention are formulated into pharma- ceutically acceptable dosage forms by conventional methods known to those skilled in the art. The dosage regimen is adjusted to provide the optimum desired response (e.g., therapeutic response). For example, a single bolus may be administered or several divided doses may be administered over time, and the dosage may be proportionally increased or decreased as indicated by the exigencies of the therapeutic situation. For ease of administration and uniformity of dosage, it is particularly advantageous to formulate parenteral compositions in dosage unit form.Dosage unit form as used herein refers to physically discrete units suitable for unitary administration to treated subjects, each unit containing a predetermined amount of active compound calculated to produce the desired therapeutic effect, together with the necessary pharmaceutical carrier.

[0224] The actual dosage of the active ingredient of the pharmaceutical composition of the present invention can be varied to obtain an amount of the active ingredient effective to achieve the desired therapeutic response for a particular patient, composition, and mode of administration, without causing toxicity to the patient. The selected dosage level will depend on a variety of pharmacokinetic factors, including the activity of the particular composition of the present invention, or its esters, salts, or amides, employed, the route of administration, the duration of administration, the excretion rate of the particular compound employed, the duration of treatment, other drugs, compounds and / or substances used in combination with the particular composition employed, the age, sex, weight, condition, general health and past medical history of the patient being treated, and the like.

[0225] The multispecific antibodies of the invention are usually administered in multiple doses. The intervals between single doses can be weekly, monthly, or yearly. Irregular intervals are also possible, as indicated by measuring the blood concentration of the multispecific antibodies of the invention in the patient. Alternatively, the multispecific antibodies of the invention can be administered as sustained release formulations, in which case less frequent administration is required. The dosage and frequency of administration depend on the half-life of the antibody in the patient. In general, humanized antibodies exhibit a longer half-life than chimeric and non-humanized antibodies. The dosage and frequency of administration depend on whether the administration is prophylactic or therapeutic. In prophylactic applications, relatively low dosages are administered relatively infrequently over a long period of time. Some patients continue to receive treatment for the rest of their lives. In therapeutic applications, relatively high dosages are sometimes required at relatively short intervals until the progression of the disease is reduced or stopped, preferably until the patient shows partial or complete improvement of the symptoms of the disease. The patient can then be administered in a prophylactic regime.

[0226] In one aspect, the invention relates to a multispecific antibody of the invention or a pharmaceutical composition of the invention for use as a medicament. In a preferred embodiment, the invention provides a multispecific antibody or a pharmaceutical composition for use in the treatment of a proliferative disease, in particular cancer, in a subject in need thereof.

[0227] In another aspect, the present invention provides a multispecific antibody or a pharmaceutical composition for use in the manufacture of a medicament for the treatment of a proliferative disease, in particular cancer.

[0228] In another aspect, the present invention relates to the use of the multispecific antibody or the pharmaceutical composition for treating a proliferative disease, in particular cancer, in a subject in need thereof.

[0229] In a further aspect, the present invention relates to the use of the multispecific antibody or the pharmaceutical composition in the manufacture of a medicament for the treatment of a proliferative disease, in particular cancer, in a subject in need thereof.

[0230] In another aspect, the present invention relates to a method of treating a subject, comprising administering to the subject a therapeutically effective amount of a multispecific antibody of the invention. In a preferred embodiment, the present invention relates to a method of treating a proliferative disease, in particular cancer, in a subject, comprising administering to the subject a therapeutically effective amount of a multispecific antibody of the invention.

[0231] The term "subject" includes humans and non-human animals. Non-human animals include all vertebrates, e.g., mammals and non-mammals, such as non-human primates, sheep, dogs, cows, chickens, amphibians, and reptiles. Unless otherwise specified, the terms "patient" and "subject" are used interchangeably herein.

[0232] As used herein, the terms "treatment," "treating," "treat," "treated," and the like refer to obtaining a desired pharmacological and / or physiological effect. The effect may be therapeutic in that it partially or completely cures a disease and / or adverse effects caused by a disease, or slows the progression of a disease. As used herein, "treatment" encompasses any treatment of a disease in a mammal, e.g., a human, and includes (a) inhibiting the disease, i.e., preventing its occurrence, and (b) palliating the disease, i.e., relieving the disease.

[0233] The term "therapeutically effective amount" or "effective amount" refers to the amount of a drug that, when administered to a mammal or other subject for treating a disease, is sufficient to effect such treatment for the disease. A "therapeutically effective amount" will vary depending on the drug, the disease and its severity, and the age, weight, etc., of the subject being treated.

[0234] In one embodiment, the proliferative disease is cancer. The term "cancer" refers to a disease characterized by rapid and uncontrolled proliferation of abnormal cells. Cancer cells can metastasize locally or through the bloodstream and lymphatic system to other parts of the body. The terms "tumor" and "cancer" are used interchangeably herein, e.g., both terms encompass solid and liquid, e.g., diffuse or circulating tumors. As used herein, the term "cancer" or "tumor" includes precancerous, as well as malignant cancers and tumors. The term "cancer" is used herein to mean a broad range of tumors, including all solid and hematological malignancies. Examples of such tumors include, but are not limited to: benign or particularly malignant tumors, solid tumors, brain tumors, kidney cancer, liver cancer, adrenal cancer, bladder cancer, breast cancer, stomach cancer, and the like. cancer) (e.g. gastric tumors), esophageal cancer, uterine cancer, cervical cancer, colon cancer, rectal cancer, prostate cancer, pancreatic cancer, lung cancer (e.g. non-small cell lung cancer and small cell lung cancer), vaginal cancer, thyroid cancer, melanoma (e.g. unresectable or metastatic melanoma), renal cell carcinoma, sarcoma, glioblastoma, multiple myeloma or gastrointestinal cancer, in particular colon cancer or colorectal adenoma, neck and head tumors, endometrial cancer, Cowden syndrome, Lhermitte-Duclos disease, Banayan-Zonana syndrome, benign prostatic hyperplasia, neoplasms, in particular those of epithelial nature, preferably breast cancer or squamous cell carcinoma, chronic lymphocytic leukemia, chronic myeloid leukemia (e.g. Philadelphia chromosome positive chronic myeloid leukemia), acute lymphoblastic leukemia (e.g. Philadelphia chromosome positive acute lymphoblastic leukemia), non-Hodgkin's lymphoma, plasma cell myeloma, Hodgkin's lymphoma, leukemia, and any combination thereof. In a preferred embodiment, the cancer is lung cancer, preferably non-small cell lung cancer (NSCLC). In another embodiment, the cancer is colorectal cancer.

[0235] The multispecific antibodies of the invention or the compositions of the invention inhibit the growth of solid tumors, but also the growth of liquid tumors. In a further embodiment, the proliferative disease is a solid tumor. The term "solid tumor" refers in particular to breast, uterine, colon, rectal, prostate, stomach cancer (especially gastric cancer), cervical, lung cancer (e.g. non-small cell lung cancer and small cell lung cancer), and head and neck tumors. Furthermore, depending on the type of tumor and the particular combination used, a reduction in tumor volume may be obtained. The multispecific antibodies of the invention or the compositions of the invention are also suitable for preventing the metastatic spread of tumors and the growth or occurrence of micrometastases in subjects with cancer.

[0236] In one embodiment, the cancer is PD-L1-positive, preferably the cancer expresses elevated levels of PD-L1 compared to healthy tissue, in particular the cancer expresses at least 2-fold, at least 3-fold, at least 4-fold, at least 5-fold, at least 6-fold, at least 7-fold, at least 8-fold, at least 9-fold, at least 10-fold, at least 15-fold, at least 20-fold, at least 30-fold, at least 40-fold, at least 50-fold, at least 60-fold, at least 70-fold, at least 80-fold, at least 90-fold, at least 100-fold elevated levels of PD-L1 (mRNA or protein, respectively) compared to PD-L1 expression (mRNA or protein) in healthy tissue. In some embodiments, the cancer is malignant. In some embodiments, the cancer is benign. In some embodiments, the cancer is primary. In some embodiments, the cancer is secondary. In one embodiment, the cancer is lung cancer, preferably non-small cell lung cancer (NSCLC). In another embodiment, the cancer is colorectal cancer.

[0237] In one aspect, the present invention relates to a kit comprising a multispecific antibody of the invention or a pharmaceutical composition of the invention. The kit may comprise one or more other components, including: instructions for use; other reagents, such as a label, a therapeutic agent, or an agent useful for chelating or otherwise binding the antibody to a label or a therapeutic agent, or a radioprotective composition; equipment or other materials for preparing the antibody molecule for administration; a pharma- ceutically acceptable carrier; and equipment or other materials for administration to a subject. In a specific embodiment, the kit comprises a pharma- ceutically effective amount of the multispecific antibody of the invention. In a further embodiment, the kit comprises a pharma- ceutical effective amount of the multispecific antibody of the invention in lyophilized form and a diluent and, optionally, instructions for use. The kit may further comprise a filter needle for reconstitution and a needle for injection.

[0238] [Table 1-1] [Table 1-2]

[0239] [Table 2-1] [Table 2-2]

[0240] [Table 3]

[0241] [Table 4-1] [Table 4-2]

[0242] [Table 5]

[0243]

Table 6-1

Table 6-2

Table 6-3

Table 6-4

Table 6-5

Table 6-6

Table 6-7

Table 6-8

Table 6-9

Table 6-10

Table 6-11

Table 6-12

Table 6-13

Table 6-14

Table 6-15

Table 6-16

Table 6-17

Table 6-18

Table 6-19

Table 6-20

Table 6-21

Table 6-22

Table 6-23

Table 6-24

Table 6-25

Table 6-26

Table 6-27

Table 6-28

Table 6-29

[0244]

Table 7-1

Table 7-2

[0245] Throughout the text of this application, in the event of a conflict between the text of the specification (e.g., Tables 1-6) and the Sequence Listing, the text of the specification shall control.

[0246] It is understood that certain features of the invention that are described for clarity in the context of separate embodiments can also be provided in combination in a single embodiment. Conversely, various features of the invention that are described for brevity in the context of a single embodiment can also be provided separately or in any suitable subcombination. All combinations of the embodiments of the invention are specifically embraced by the invention and are disclosed herein as if each combination were individually and explicitly disclosed. Moreover, all subcombinations of the various embodiments and elements thereof are also specifically embraced by the invention and are disclosed herein as if each such subcombination were individually and explicitly disclosed.

[0247] The present invention is not limited in scope by the specific embodiments described herein. Indeed, various modifications of the invention in addition to those described herein will become apparent to those skilled in the art from the foregoing description. Such modifications are intended to be included within the scope of the appended claims.

[0248] To the extent possible under the respective patent laws, all patents, applications, publications, test methods, literature, and other materials cited herein are hereby incorporated by reference.

[0249] The following examples illustrate the invention described above, but are not intended to limit the scope of the invention in any way. Also, other test models known to those skilled in the art can determine the beneficial effects of the claimed invention. EXAMPLES

[0250] Example 1: Generation and testing of low affinity anti-PD-L1 molecules: Project Objective The goal of this project is to identify low affinity anti-PD-L1 antibody fragments that neutralize the interaction of PD-L1 with PD-1. Ultimately, this domain will be combined in a multispecific molecule with a high affinity domain against selected tumor-associated antigens (TAA) that are co-expressed with PD-L1 on tumor cells, allowing for targeting and neutralization of PD-L1 specifically on those cancer cells. Two groups of long-acting molecules were designed corresponding to the two projects. The molecules in each group differ mainly in the number of specificities (3 or 4) and their format. Both groups of molecules contain a Her2 domain as a TAA, a low affinity PD-L1 domain, and a human serum albumin binding domain for half-life extension, while one group additionally contains an anti-CD3ε binding domain to trigger T cell activation. This example describes the generation and characterization of the low affinity domains and the multispecific molecules designed in these projects.

[0251] Design and construction of scFv To generate a low affinity PD-L1 antibody that neutralizes the interaction between PD-L1 and PD-1, we introduced only one alanine substitution in the CDR region of the previously identified clone 33-03-G02 high affinity neutralizing anti-PD-L1 domain. As a first step, each amino acid in the CDR3 region of the high affinity domain (where the greatest amino acid diversity exists) was mutated to alanine, resulting in 21 mutants. Three single mutants showed a greater than 100-fold reduction in affinity to PD-L1 compared to the original domain. Therefore, the single mutants were combined to generate two double mutants. In parallel, we designed a further nine single mutants of the most diverse residues in the CDR1 and CDR2 regions, as well as two mutants containing combinations of other single mutations in the CD3 region that only slightly reduced the affinity of the parent domain, presumably exposed residues. Furthermore, three mutants were additionally expressed that contained up to three alanine substitutions of predicted exposed residues. The data obtained for the five most interesting molecules, with affinities reduced by 100-fold to 6'500-fold, are presented below.

[0252] method: scFv production Heterologous expression of proteins was performed in E. coli at small scale with overnight induction of expression as insoluble inclusion bodies (as shown in Table 8 below). Inclusion bodies were isolated from homogenized cell pellets by a centrifugation protocol that included several washing steps to remove cell debris and other host cell impurities. The purified inclusion bodies were dissolved in denaturing buffer and the scFvs were refolded by a scalable refolding protocol that generated milligram quantities of natively folded monomeric scFvs. At that point, a standardized protocol was employed to purify the scFvs. The refolded products were captured by affinity chromatography to obtain purified scFvs. Table 8 summarizes the production of scFv molecules. Expression of mammalian constructs was performed in CHO-S cells using the CHOgro transient transfection kit (Mirus) (see Table 8). After 5-7 days of expression (cell viability <70%), cultures were harvested by centrifugation at 37°C and proteins were purified from clarified culture supernatants by Protein L affinity chromatography followed, if necessary, by a polishing step by size-exclusion chromatography. Standard analytical methods such as SE-HPLC, UV280 and SDS-PAGE were used for quality control of the produced material.

[0253] Freeze-thaw stability test The suitability of the best performing scFv molecules was evaluated for freeze-thaw (F / T) cycling (colloidal stability). For F / T stability evaluation, the same analytical methods (SE-HPLC, UV absorbance at 280 nm) and parameters as for storage stability testing were applied to monitor the quality of the molecules over multiple F / T cycles. As no dedicated freeze-thaw study was performed, freeze-thaw data were extracted from -80°C samples from the storage stability study obtained over a period of 28 days (maximum of 7 days storage between F / T cycles).

[0254] Differential Scanning Fluorometry (DSF) The midpoint of the thermal unfolding transition of scFv constructs was determined by differential scanning fluorimetry using the fluorescent dye SYPRO® Orange. Samples were prepared in a total volume of 100 μl in final buffer (50 mM NaCiP, 150 mM NaCl, pH 6.4) containing a final concentration of 5× SYPRO® Orange at a final protein concentration of 50 μg / mL. 25 microliters of prepared samples were added in triplicate to a white-walled AB gene PCR plate. The assay was performed on a qPCR machine used as a thermal cycler, and fluorescence emission was detected using a custom dye calibration routine in the software. The PCR plate containing the test samples was subjected to a 1°C temperature ramp from 25°C to 96°C with a 30 second pause after each ramp. The total assay time was approximately 2 hours. The Tm was calculated with the software GraphPad Prism using the mathematical second derivative method to find the inflection point of the curve. The reported Tm is the average of triplicate measurements.

[0255] [Table 8]

[0256] Affinity for PD-L1 by SPR method: Affinity for PD-L1 was measured by surface plasmon resonance (SPR) measurements using a Biacore T200 instrument (GE Healthcare). All measurements were performed at 25 °C. In this experiment, Fc-tagged human PD-L1 extracellular domain (ECD, Sino Biological, Cat. 10084-H02H) was captured using a human antibody capture kit (Cat. BR-1008-39) from GE Healthcare. After each analyte injection cycle, anti-human Fc-specific IgG was regenerated and new antigen was captured. For affinity measurements, scFv was injected as analyte at analyte concentrations ranging from 6.86 to 5000 nM (3-fold serial dilutions) diluted in running buffer (10 mM HEPES, 150 mM NaCl, and 0.05% Tween 20, pH 7.4) using a dose-response multi-cycle kinetic assay. Association and dissociation times were 300 and 720 s, respectively. Apparent dissociation (k d ) and meeting (k a ) rate constant and apparent dissociation equilibrium constant (K D ) was calculated using a one-to-one Langmuir binding model with Biacore analysis software (BIAevaluation, GE Healthcare) and monitored based on Chi2 and U values, which are measures of the quality of the curve fitting. In addition to kinetic measurements, K was calculated by fitting a plot of the equilibrium response versus concentration. D was obtained (steady-state affinity measurement).

[0257] result: Binding to human PD-L1 was confirmed for the humanized scFvs tested, as shown in Table 9. The affinity for PD-L1 was reduced 6,500-fold for PRO1434 compared to the parent scFv, PRO830.

[0258] [Table 9]

[0259] Binding of FCs to PD-L1-expressing cells method: CHO-K1 (control cells not expressing PD-L1) and CHO-PD-L1 (Amsbio), which express high levels of PD-L1, were harvested and cell counts were determined. The cell suspension was centrifuged at 400×g for 5 min, and 100 μl (10'000 cells) of the cell suspension diluted in PBS-EB (1×DPBS, 2% BCS HI, 2 mM EDTA) was added to the designated wells of the non-binding plate. After three washing steps with PBS-EB, the cells were centrifuged and the wash buffer was aspirated. Then, 100 μl of serial dilutions of the samples to be tested and the positive control were added directly to the plate. The positive control sample (PRO830, 33-03-G02) was diluted in PBS-EB at concentrations ranging from 3500 to 0.22 ng / ml, and the dilutions of the samples ranged from 1000 to 0.064 μg / ml. After 1 h incubation at 4°C, the plates were washed three times using 100 μl PBS-EB. The cell pellet was resuspended in 100 μl Protein L-APC at a concentration of 2 μg / ml and incubated for 1 h at 4°C. The cells were then washed again three times using 100 μl PBS-EB. The cell pellet was resuspended in 50 μl PBS-EB and analyzed on a NovoCyte2060 flow cytometer instrument. For each sample, the fluorescence intensity of APC for 5'000 events was recorded and the geometric mean of the fluorescence intensity MFI was calculated. Data were corrected for non-specific antibody binding (blank and CHO-K1 cell binding).

[0260] result: The potency of binding to cells expressing PD-L1 was assessed using flow cytometry as described above. Serial dilutions of each molecule being tested and the reference PRO830 were added to the plate. The individual IC 50 Values ​​are expressed as the IC of the reference molecule PRO830 (high affinity PD-L1 domain) co-incorporated on each plate. 50 Calibrated against (relative EC 50 :EC 50 , PRO830 / IC 50, tested scFv). The potencies are summarized in Table 10. PRO1434 and PRO1494 are shown to have the weakest binding. The dose response curves obtained for PRO1434 and PRO1494 are shown in FIG.

[0261] [Table 10]

[0262] PD-L1 / PD-1 neutralization by NFAT reporter gene assay method: The ability of scFv to neutralize PD-L1 / PD-1 interaction when both interacting molecules are expressed on the cell surface was tested using CHO / PD-L1 / TCR (T cell receptor) activator cells and Jurkat / PD-1 cells. In this assay, CHO cells stably expressing PD-L1 and TCR activators are incubated with Jurkat T cells stably expressing PD-1 and firefly luciferase under the control of an NFAT response element as a reporter gene to monitor T cell activation. When TCR activators on CHO cells bind to Jurkat T cells, TCR signaling induces NFAT-induced firefly luciferase expression. However, the interaction between PD-L1 and PD-1 negatively regulates such TCR signaling, thus reducing the expression of firefly luciferase. Thus, blocking the PD-L1 / PD-1 interaction in this system restores luciferase activity.

[0263] 100 μl of 35'000 CHO / PDL1 / TCR activator cells in cell culture medium (DMEM / F12, 10% FCS) were added to the inner wells of a white cell culture plate and incubated at 37°C and 5% CO 2The plates were incubated at 37 °C and 5% CO for 16–20 h. The next day, cell culture medium was removed from each well and 50 μl of two-fold concentrated serial dilutions of each molecule to be tested (final concentrations 162–0.025 μg / ml) and the reference molecule avelumab (final concentrations 3 000–0.46 ng / ml) were added. Then, 50 μl of effector Jurkat cells diluted at 400 000 cells / ml in assay buffer (RPMI 1640 with 10% FCS) were added to each well and the plates were incubated at 37 °C and 5% CO for 16–20 h. 2 Plates were incubated at 4°C for 6 h at RT. Then, 50 μL of luciferase substrate (BPS Biosciences) prepared according to the manufacturer's protocol was added per well, plates were incubated in the dark for 30 min, and luminescence was measured using a Flexstation III multimode microplate reader.

[0264] result: The potency of neutralizing PD-L1 binding to PD-1 was assessed using a cell-based reporter gene assay as described above. Serial dilutions of each molecule to be tested and the reference avelumab were added to the plate. The individual IC 50 Values ​​were calculated using the IC of the reference molecule avelumab (high affinity PD-L1 domain) co-incorporated on each plate. 50 Calibrated against (relative IC 50 :I C 50 , avelumab / IC 50 , tested scFv). The potencies are summarized in Table 11. PRO1434 and PRO1494 are shown to have the lowest potencies. The titration curves obtained for PRO1434 and PRO1494 are shown in FIG. 2.

[0265] [Table 11]

[0266] Storage stability test method: The humanized scFv was formulated at 10 mg / ml in aqueous buffer (final buffer, 50 mM NaCiP, 150 mM NaCl, pH 6.4) and subjected to stability studies, including a 4 week stability study where the samples were stored for 4 weeks at -80°C, 4°C and 40°C. At a minimum, protein concentration was determined by measuring UV absorbance at 280 nm and monomer and oligomer fractions in the formulations were assessed by integration of SE-HPLC peak areas after 1 week, 2 weeks and at the end of each study. Parameters such as % monomer content, % monomer loss, content and % content loss were recorded over time.

[0267] result: Tables 12 and 13 compare the endpoint measurements obtained on d0, d7, d14 and d28 of the study at 4° C. and 40° C. At 4° C., all molecules show a decrease in monomer content of less than 10% over 28 days, except for PRO1075 and PRO1076. At 40° C., only two molecules show a monomer content of more than 85% after 28 days: PRO1434 (86%) and PRO1494 (90%).

[0268] Example 2: Generation and testing of multispecific constructs targeting HER2 (as an example of a TAA), PD-L1 and CD3 (as examples of immune cell antigens): This approach is directed towards next generation polyspecific immuno-oncology programs targeting HER2 expressing malignancies. HER2 is a clinically validated target in several cancer types with unmet medical need, most notably breast and gastric cancer. However, the spectrum of tumors (over)expressing HER2 is very broad, but for mechanistic reasons inaccessible to conventional antibodies such as trastuzumab, and is amenable to the present approach, designed not only to effectively mediate T cell-induced lysis of HER2 expressing tumors, but also to avoid tumor immune escape by simultaneously blocking immunosuppressive PD-L1 signaling. The local restriction of two additive and possibly simultaneous mechanisms of action to tumor tissue is expected to give compounds according to this approach a substantially expanded efficacy profile that could be clinically effective even in HER2 expressing tumors that are primarily or secondarily refractory to standard anti-HER2 therapy. This approach should i) result in a more selective blockade of PD-1 / PD-L1 interaction that is at least as potent as avelumab / BAVENCIO®; Specifically, compounds according to this approach should efficiently block the binding of PD-1 to PD-L1 in HER2 / PD-L1 co-expressing target cells, while blocking PD-1 / PD-L1 interaction in cells not expressing Her2 much less potently than Avelumab / BAVENCIO®, and ii) result in at least equally potent and more selective lysis of HER2 / PD-L1 co-expressing cells by peripheral blood mononuclear cells compared to Trastuzumab / Herceptin®, Avelumab / BAVENCIO® and the combination of the two. More specifically, compounds according to this approach should potently lyse cells co-expressing HER2 and PD-L1, while not resulting in lysis of cells expressing only PD-L1.

[0269] Design and construction of tribodies, DVD-tribodies and MATCH-4 molecules A description of the different formats designed within this approach (Tribody, DVD-Tribody and MATCH-4) is shown in Figure 3. Data regarding the generation of all molecules is detailed in Table 14, which gives a description of the domain composition of the different molecules generated and their arrangement within the molecule. The targets of each domain are as follows: Trastuzumab: Her2; Clone 14-11-D07: IL23R; Clone 23-13-A01: Human / Mouse SA; Clone 28-21-D09: CD3e; Clone 33-02-G02 and its variants: PD-L1.

[0270] [Table 12]

[0271] [Table 13]

[0272] [Table 14-1] [Table 14-2]

[0273] method: Expression of tribodies, DVD-tribodies and MATCH-4 constructs was performed in CHO-S cells using the CHOgro transient transfection kit (Mirus). After 5-7 days of expression (cell viability <70%), cultures were harvested by centrifugation at 37°C and proteins were purified from clarified culture supernatants by Protein L affinity chromatography followed, if required, by a polishing purification step by size exclusion chromatography. Standard analytical methods such as SE-HPLC, UV280 and SDS-PAGE were used for quality control of the produced material.

[0274] Affinity to human PD-L1, IL-23R, Her2, CD3ε, and human and mouse serum albumin (SA) by SPR method: The affinity for PD-L1 was measured by surface plasmon resonance (SPR) measurements using a Biacore T200 instrument (GE Healthcare) as described above. d ) and meeting (k a ) rate constant and apparent dissociation equilibrium constant (K D ) was calculated using a one-to-one Langmuir binding model with Biacore analysis software (BIAevaluation, GE Healthcare). The quality of the fit was monitored based on Chi2 and U values. In addition to kinetic measurements, K was calculated by fitting a plot of equilibrium response versus concentration. D was obtained (steady-state affinity measurement).

[0275] The binding affinity of the multispecific constructs to recombinant human CD3ε ECD (Sino Biological, Catalog 10977-H08H), recombinant human IL-23R (custom made by Trenzyme) and recombinant human Her2 ECD (Sino Biological, Catalog 10004-HCCH) was measured by SPR using a Biacore T200 instrument. All measurements were performed at 25 °C. The different proteins were immobilized on a sensor chip (CM5 Sensor Chip, GE Healthcare) by amine coupling such that the immobilization level reached approximately 100 response units (RU). Serial dilutions of the multispecific molecules ranging from 0.35 to 90 nM (2-fold serial dilutions) in running buffer were injected into the flow cell at a flow rate of 30 to 50 μl / min for 5 to 7 min. Dissociation of the multispecific constructs from CD3ε, IL-23R and Her2 on the CM5 chip was allowed to proceed for 12 min. After each injection cycle, the surface was regenerated with one injection of 10 mM glycine HCl, pH 2. The resulting binding curves were double-referenced (empty reference channel and zero analyte injection) and the kd, ka, and K were calculated using a one-to-one Langmuir binding model in the Biacore analysis software. DThe value of K was calculated and the quality of the fit was monitored based on Chi2 and U values. In the case of CD3ε, fitting using a one-to-one Langmuir binding model showed that the quality of the curve fit was not optimal, so a two-state reaction model was further used to calculate K D was calculated. This model describes the conformational changes that occur following 1:1 binding of the analyte to the immobilized ligand, stabilizing the complex.

[0276] The affinity of the molecules for human serum albumin (HSA) and mouse serum albumin (MSA) was measured by SPR measurements using a Biacore T200 instrument (GE Healthcare). SA was directly coupled to a CM5 sensor chip (GE Healthcare) using amine coupling chemistry. After regeneration scouting and surface performance tests to find the optimal assay conditions, dose-response of the molecules of interest was examined at concentrations ranging from 0.7 to 180 nM. The assay was performed in PBS-Tween buffer at pH 5.5. The association and dissociation times were set to 180 and 720 s, respectively. The resulting binding curves were double-referenced (empty reference channel and zero analyte injection) and fitted using BiaEvaluation software (GE Healthcare) and a one-to-one Langmuir binding model. The obtained kinetic parameters were used to calculate the apparent dissociation equilibrium constant (K D ) was calculated.

[0277] result: As shown in Tables 15 and 16, binding to CD3ε, human and mouse serum albumin was similar for all molecules tested, and molecules containing anti-IL23R domains showed comparable affinity for IL23R. Affinity measurements by SPR for low affinity domains are very challenging due to the large amount of protein that needs to be injected to cover the concentration range corresponding to the affinity of the molecule. Therefore, reliable measurements for PD-L1 could not be obtained for some molecules. Furthermore, in the case of low affinity domains, a large bulk shift was observed, introducing artifacts into the kinetic analysis, so steady-state analysis of SPR measurements may be more appropriate. Affinity measurements for human PD-L1 were valid for only one tribody, PRO1498. For MATCH-4 molecules, valid measurements were obtained using steady-state analysis. The lowest affinities for PD-L1 were around 900 nM for PRO1544, PRO1545 and PRO1547. PRO1543 and PRO1546 showed similar affinities around 300 nM. MATCH-4 proteins bearing the lambda capped trastuzumab G100C / G172C mutant anti-Her2 domain (PRO1543, PRO1544, PRO1557 and PRO1558) showed similar affinity for Her2 (300-400 pM), which is also comparable to the affinity of the trastuzumab anti-Her2 domain incorporated into the tribody (PRO1497) and DVD-tribody (PRO1547 and PRO1548).

[0278] [Table 15]

[0279] [Table 16]

[0280] PD-L1 / PD-1 neutralization by FC method: These assays were performed to evaluate the ability of the multispecific low affinity PD-L1x / Her2 construct to neutralize the interaction between PD-1 and PD-L1 expressed on HCC1954 cells, which also express Her2. Specifically, the molecule should efficiently block PD-1 binding to PD-L1 in Her2 / PD-L1 co-expressing target cells (HCC1954), but should block PD-1 / PD-L1 interaction much less potently than avelumab / BAVENCIO® or nivolumab / OPDIVO® in cells that do not express Her2 (HCC827). Blockade of PD-1 binding to cells was analyzed by FC and compared to the reference IgG, avelumab. Additionally, HCC827 cells were used as an additional control, as these cells express PD-L1 at levels comparable to HCC1954 cells, but lack significant expression of Her2.

[0281] HCC1954 and HCC827 cells were stimulated with 10ng / ml human IFNγ for 24 hours to further induce PD-L1 expression. The next day, HCC827 and HCC1954 cells were detached, centrifuged at 200g for 4 minutes, resuspended in PBS / 2%FCS / 2mM EDTA (staining buffer), and plated (50μl / well) in 96-well PP microplates. Three-fold step dilution series of multispecific molecules and avelumab were prepared in staining buffer containing 500ng / ml biotinylated PD1, starting at 20μg / ml and 5μg / ml, respectively. HCC827 and HCC1954 plates were centrifuged at 200g for 4 minutes, and the dilution series was added to the cells (100μl / well) and incubated at room temperature (RT) for 30 minutes. Then, after washing once with 150 μl of staining buffer, streptavidin-PE solution was added to the cells (100 μl / well) and incubated at 4°C for 30 min. As a next step, the cells were washed again as above, and then resuspended in 100 μl of staining buffer. The resuspended cells were then processed for fluorescence measurement using a NovoCyte flow cytometer (ACEA Bioscience Inc.). The mean fluorescence intensity of PE-labeled PD-1 was recorded and the data was fitted using sigmoidal 4PL fit (GraphPad Prism). Individual ICs of each plate were analyzed using a 30-μl flow cytometer (GraphPad Prism). 50 Values ​​were calculated based on the IC of the reference molecule avelumab co-incorporated on each plate. 50 Calibrated against (relative IC 50 :I C 50 , avelumab / IC 50 , test molecule). Furthermore, the relative IC observed in high Her2-expressing HCC1954 cells and low Her2-expressing HCC827 cells 50 The ratio of values ​​was calculated.

[0282] result: The resulting potency of constructs from this approach to neutralize PD-1 / PD-L1 interaction is summarized in Table 17. In HCC1954 cells, which express high levels of Her2 and PD-L1, the tribody PRO1454 with a single alanine mutation PD-L1 domain (33-03-G02 G109A) showed similar potency as avelumab, whereas in cells expressing only PD-L1 (HCC827), PRO1454 was 14-fold less potent. PRO1497, a tribody with the double alanine mutation 33-03-G02 Q108A / G109A, showed similar relative IC 50 As shown by the values, PRO1497 neutralized the PD-1 / PD-L1 interaction almost as efficiently as PRO1454 (PRO1454: 0.66, PRO1497: 0.38). However, in contrast to PRO1454, PRO1497 showed a very weak neutralization potency against PD-L1 expressing cells (HCC827) (Figure 4). This result indicates the potentially broader therapeutic window of molecules such as PRO1497 that contain a PD-L1 domain with similar affinity to a domain with both mutations (Q108A / G109A) compared to a domain with similar affinity to a single alanine mutation (G109A).

[0283] All MATCH-4 molecules containing the trastuzumab anti-Her2 domain and the anti-PD-L1 domain with both mutations (Q108A / G109A) had similar potency compared to the tribody PRO1497 (relative IC 50 values, PRO1543: 0.28, PRO1544: 0.24). The titration curves obtained for PRO1543 and PRO1546 are shown in Figure 5. Along those lines, the DVD-tribody PRO1547, which contains the trastuzumab anti-Her2 domain and the anti-PD-L1 domain with both mutations (Q108A / G109A), neutralized the PD-1 / PD-L1 interaction with a similar potency as PRO1497 (relative IC50 value, PRO1547: 0.25).

[0284] [Table 17]

[0285] Blockade of CD3 activation and PD-L1 / PD-1 interaction by NFAT reporter gene assay method: To evaluate the simultaneous effect of the molecules on CD3 cross-linking and PD-1 / PD-L1 blockade, T cell activation was tested in an NFAT (nuclear factor of activated T cells) assay. Specifically, the molecules should efficiently induce CD3 activation and block PD-1 binding to PD-L1 in Her2 / PD-L1 co-expressing target cells (HCC1954), but should be much less potent in blocking CD3 activation and PD-1 / PD-L1 interaction in cells not expressing Her2 (CHO-PD-L1). The Jurkat PD-1 NFAT reporter T cell line expresses a luciferase reporter gene under the control of an NFAT response element derived from the IL-2 promoter. The transcription factor NFAT is activated upon cross-linking of CD3ε and induces many genes involved in T cell activation. In this system, cross-linking of CD3ε induces expression of a luciferase reporter gene. Moreover, the interaction between PD-L1 and PD-1 negatively regulates such CD3ε signaling, thus decreasing the expression of firefly luciferase. Thus, blocking the PD-L1 / PD-1 interaction in this system results in an increase in luciferase activity. HCC1954 cells and PD-L1-expressing CHO-K1 cells (clone A2), stimulated with 10 ng / ml IFNy for 24 hours to increase PD-L1 expression, were used as target cells and seeded at 50 μl and 25'000 cells per well in a 96-well culture plate. Serial dilutions of the molecules to be tested were prepared in assay medium containing 50 mg / ml HSA, and 25 μl was added to the cells (final concentrations ranging from 250 nM to 0.026 pM). PD-1-expressing Jurkat NFAT reporter cells were prepared in assay medium containing 50 mg / ml HSA, and added at a cell density of 50'000 cells per well. Luciferase expression was detected by addition of luciferase reagent and read in a luminescence reader 5 or 22 hours after addition of Jurkat PD-1 NFAT reporter cells. Relative light units (RLU) are shown. HER2×PD-L1 高KD The potency of the xCD3 scDb-scFv, PRO1497, was used as a reference for calculation of the relative potency of the half-life extenders.

[0286] result: The efficacy of simultaneously inducing CD3 activation and blocking PD-L1 / PD-1 interaction was evaluated by NFAT reporter gene assay, the results of which are shown in Table 18. Each molecule tested and the reference scDb-scFv Her2xCD3xPD-L1 低親和性 Serial dilutions of (PRO1497) were added to the plate. Individual IC 50 The IC value and maximum activation were calculated using the IC of the reference molecule PRO1497 in the presence of Her2 and PD-L1 expressing cells HCC1954. 50 and normalized to maximum activation (relative IC 50 or maximum activation: IC 50 or maximum activation, PRO1497 / IC in HCC1954 50 or maximal activation, test molecule). In the presence of PD-L1 / Her2 expressing cells (HCC1954), tribodies PRO1454, PRO1497 and PRO1456 inhibited CD3 signaling in Jurkat cells at similar EC 50Although the maximum activation was higher for PRO1454 and PRO1497, i.e. molecules with low affinity anti-PD-L1 domains, compared to PRO1456, which contains an anti-IL-23R dummy domain instead of an anti-PD-L1 domain (Figure 6). This suggests that PRO1454 and PRO1497 block PD-L1 in the immune synapse and simultaneously activate CD3 in the presence of cells co-expressing Her2 and PD-L1. A 27-fold weaker activation was observed for PRO1455 compared to PRO1454, and an even 475-fold weaker for PRO1498 compared to PRO1497, i.e. molecules without an anti-Her2 domain but with the low affinity anti-PD-L1 domains 33-03-G02 G109A or 33-03-G02 Q108A / G109A, respectively. This result indicates that the therapeutic window of a molecule containing a PD-L1 domain with comparable affinity to a domain with both mutations (Q108A / G109A) is potentially broader than that of a domain with similar affinity to a single alanine mutation (G109A), and both molecules are as potent as a molecule without a PD-L1 domain (PRO1456). The potencies are summarized in Table 18. PRO1543 has been shown to have the best potency in Her2 / PD-L1 expressing cells (HCC1954) and one of the lowest in the presence of PD-L1 / very low Her2 expressing cells (HCC827). The titration curves obtained for PRO1543 and the control molecules PRO1546 and PRO1557 in the presence or absence of 1 μg / ml nivolumab and PRO1557 in combination with the low affinity PD-L1 domain PRO1434 are shown in Figure 7.

[0287] [Table 18]

[0288] Cytotoxicity assay (T cell-driven target cell depletion) To evaluate the ability of the compounds of this approach to selectively direct T cells to Her2 and PD-L1 co-expressing cells, a cytotoxicity assay using a Her2 and PD-L1 positive cell line (HCC1954) was performed in the presence of human PBMC. When the compounds of this approach simultaneously bind to both targets, CD3ε on T cells is cross-linked, activating a signal cascade that leads to T cell activation (CD69 upregulation, cytokine secretion) and release of cytotoxic granules, ultimately resulting in target cell killing. In contrast to the combination of avelumab / BAVENCIO® and trastuzumab / HERCEPTIN®, the compounds of this approach should selectively lyse cells co-expressing PD-L1 and HER2, while no lysis of cells expressing only PD-L1 (PD-L1 transfectants of CHO cells) should be observed.

[0289] method: Blood Cell Fractions: Peripheral blood mononuclear cells (PBMCs) were isolated from fresh blood of healthy volunteers using lymphocyte separation medium Lymphoprep (Stemcell technologies) according to the manufacturer's instructions. Briefly, blood was diluted 1:2 with human PBMC isolation buffer (PBS, 2% FCS, 2 mM EDTA) or cynomolgus PBMC isolation buffer (PBS, 5% FCS, 2 mM EDTA) and added to Leucosep tubes containing the recommended volume of Lymphoprep medium. Leucosep tubes were centrifuged at 800g (human blood) or 2000g (cynomolgus blood) for 30 min without brake at room temperature. The cell layer containing PBMCs was then collected and washed twice with human PBMC isolation buffer, and red blood cells were lysed for 5 min at room temperature using red blood cell lysis buffer. The isolated human cells were then washed once with the respective isolation buffer and once with assay medium (RPMI-1640, 10% FCS). After platelet removal, isolated PBMCs were cultured at 3 × 10 per ml in assay medium. 6 The cells were resuspended at a density of 10 viable cells.

[0290] Flow cytometry-based in vitro cytotoxicity assay (FC assay), CD8+ T cell activation, and CD11c+, CD4+, and CD8+ T cell viability Two cell lines were used as target cells: HCC1954 cells (highly co-expressing HER2 and PD-L1) and HCC827 cells (lowly co-expressing HER2 and PD-L1) stimulated with 10ng / ml IFNy for 24 hours to increase PD-L1 expression. CHO-PD-L1 (no Her2, high PD-L1) and CHO-K1 cell lines were used as negative control cell lines. 5'000 viable target cells previously labeled with PKH67 and diluted in 75μl assay medium (RPMI-1640, 10% FCS) were added to a 96-well plate. 6-fold concentrated test proteins were diluted in assay medium and 25μl was added to the appropriate wells. 150,000 viable effector cells (PBMCs) diluted in 50 μl of assay medium were added to each well (E:T ratio of 30:1), and the plate was mixed on an orbital mixer at room temperature and then incubated at 37°C and 5% CO 2 After 16 or 40 hours, cells were trypsinized, resuspended in staining buffer (PBS, 2% BCS, 2 mM EDTA) and transferred to non-binding plates.

[0291] Cells were stained for different markers, including CD69, CD8, CD4, CD11c and Annexin-V. The analysis focused on apoptotic and dead target cells as well as activated CD8+ T cells. Thereby, target cells were identified by green fluorescence (PKH67) and their viability was analyzed with Annexin-V APC. Effector cells (CD8+ cells) were identified by detecting CD8 on their surface (anti-CD8 PerCP-Cy5.5). Finally, activation of CD8+ T cells was detected by quantifying CD69 expression (anti-CD69 PE). CD4 was used to better distinguish between CD8+ and CD4+ T cells. CD11c was used to stain monocytes and dendritic cells. For each marker except Annexin-V antibody, the antibodies were incubated for 30 min at room temperature under gentle agitation. Cells were washed once with staining buffer and once with Annexin binding buffer and Annexin-V staining was performed for 30 min at room temperature under agitation. Cells were washed once with Annexin-V binding buffer and flow cytometric analysis was performed on a Novocyte flow cytometer.

[0292] The percentage of specific target cell lysis was calculated according to the following formula:

[0293]

number

[0294] The percentage of activated CD8+ T cells corresponds to the percentage of CD69+CD8+ T cells.

[0295] The percentages of viable CD4+, CD8+ T cells and CD11c+ cells correspond to the proportion of Annexin-V negative cells within the different cell populations.

[0296] result: The cytotoxic potential of selected MATCH-4 molecules PRO1543 and PRO1895 was evaluated using a flow cytometry-based cytotoxicity assay. The resulting data are shown in Table 19, and the titration curves of PRO1543 and PRO1895 and the control molecule PRO2290 are shown in Figures 8 and 9. PRO1543 and PRO1895, which have anti-Her2 and low affinity anti-PD-L1 domains, are similarly potent to each other and more potent than PRO2290, which has an anti-Her2 domain of the pertuzumab epitope but does not contain the low affinity anti-PD-L1 domain. This data indicates an additive effect of targeting PD-L1 within the immune synapse, resulting in improved target cell lysis.

[0297] [Table 19]

[0298] CD4+ and CD8+ T cell viability was analyzed after 16 and 40 hours. This provides a safety read-out since activated CD4+ and CD8+ T cells express PD-L1 but not Her2 and can be targeted by PRO1543 as well. As summarized in Table 21 and shown in Figure 10, PRO1543 reduced CD4+ and CD8+ T cell viability by 5-10% only at the highest concentration tested. This can be considered a small effect on cell viability.

[0299] Cytotoxicity assays (T cell-driven target cell depletion) against cells expressing different Her2 and PD-L1 levels

[0300] method: The same method as in the above examples was used. Four cell lines were used as target cells: HCC1954 (high co-expression of HER2 and PD-L1), HCC1827 (very low expression of HER2 and high expression of PD-L1), MCF-7 (low expression of HER2 and very low expression of PD-L1) stimulated with 10ng / ml IFNy for 24 hours to increase PD-L1 expression, and CHO-PD-L1 (no HER2, high PD-L1, purchased from BPS Biosciences). It should be noted that this cell line expresses PD-L1 at approximately 9-fold lower levels than the CHO-PD-L1 clone A2 cell line used in the experiments shown in Tables 18 and 19 and Figures 8 and 9.

[0301] result: The results of the cytotoxicity assay are shown in Table 22 and Figures 11 and 12. The results show that PRO1543, a molecule containing a low affinity PD-L1 domain, is more potent than PRO1557 and PRO957 in killing cells expressing HER2 and PD-L1, while not killing cells that do not express HER2. Co-expression of HER2 and PD-L1 occurs only in cancer cells. Furthermore, PRO1543 mediates the lysis of HER2 / PD-L1 double positive cancer cells more potently than normal cells expressing HER2 alone. Thus, in contrast to molecules that do not contain a PD-L1 binding domain, PRO1543 clearly exhibits selectivity for double positive cells, thus sparing normal cells expressing HER2. Lacking a PD-L1 binding domain, PR01557 and PRO957 do not have such selectivity, and their potency is determined only by the HER2 expression level of the target cells.

[0302] [Table 20]

[0303] [Table 21]

[0304] Evaluation of the antitumor efficacy of PD-L1 blockade and simultaneous local CD3 stimulation in the human cell line-derived breast ductal carcinoma xenograft model HCC1954 The antitumor activity of the compounds of this approach was compared to anti-PD-1 therapy and anti-PD-1 anti-Her2 combination therapy in human HCC1954 ductal breast cancer xenografts using Taconic's immunodeficient NOG mouse line and allogeneic human peripheral blood mononuclear cells. The effect of PRO1678 (scMATCH3) and PRO1543 (MATCH4) on tumor volume was compared to treatment with an anti-PD-1 antibody (nivolumab) and a nivolumab / anti-Her2 antibody (trastuzumab) combo. The irrelevant IgG palivizumab was used as a control IgG. Animal weights were also followed.

[0305] Study Setup and Dosing Schedule Female NOG mice were subjected to unilateral injection of 5 × 106 HCC1954 cells. Cells were injected in a mixture of 50% cell suspension in PBS and 50% Matrigel with a total injection volume of 100 μl. After tumor cell injection into NOG mice and successful tumor engraftment (median tumor volume for groups was 80–100 mm), 10% HCC1954 cells were injected. 3 ), mice were replaced with 5 × 106 human PBMCs by intravenous injection. On the day of randomization, four mice in each group were reconstituted with PBMCs from donor A and another four mice were reconstituted with PBMCs from donor B. Dosing began 1–2 hours after injection of PBMCs and was applied as follows:

[0306] [Table 22]

[0307] Body weight (Table 24) and tumor volume measurements by caliber (Table 23 and Figure 13) were performed twice weekly. Animals were terminated on day 33. Some animals in some groups had already died after day 27. No weight loss was observed.

[0308] [Table 23]

[0309] [Table 24]

[0310] Example 3: Generation and testing of multispecific constructs targeting HER2 (as an example of a TAA), PD-L1 and CD173 (as an example of a costimulatory immune cell antigen) Design and construction of MATCH4 molecules A representation of the MATCH-4 molecule designed within this part of the approach is shown in Figure 14. The molecular composition is shown in Table 25.

[0311] [Table 25]

[0312] method: Expression of MATCH4 constructs was performed in CHO-S cells using the CHOgro transient transfection kit (Mirus). After 5-7 days of expression (cell viability <70%), cultures were harvested by centrifugation at 37°C and proteins were purified from clarified culture supernatants by protein L affinity chromatography followed, if required, by a polishing purification step by size exclusion chromatography. Standard analytical methods such as SE-HPLC, UV280 and SDS-PAGE were used for quality control of the produced material.

[0313] Anti-Her2×CD137×HSA×PD-L1 (低親和性) Evaluation of CD137 agonist effects by using a cell-based assay of a transgenic NF-kB Jurkat reporter cell line expressing the MATCH4 molecule CD137 In this assay, activation of CD137 signaling in Jurkat cells was evaluated. The activity of CD137 signaling is reported by measuring luciferase expression, which is driven by CD137-induced NF-kB activation in Jurkat reporter cell lines. Luciferase expression directly correlates with CD137 activity. Furthermore, CD137 clustering is required for signal pathway activation and is facilitated through the formation of an immune synapse between Jurkat cells and Her2-expressing cell lines. Thus, Her2 expression is required for CD137 clustering and activation in reporter cell lines.

[0314] method: Cancer cell lines HCC1954 (high expression levels of Her2 and PD-L1) and HCC827 (low expression levels of Her2 but high expression levels of PD-L1) were seeded at 25'000 cells per well in 96-well culture plates. The seeded cells were then stimulated with 10ng / ml IFNy for 24 hours or left unstimulated. Serial dilutions of the MATCH4 molecule of interest and the internal reference molecule PRO1186 or PRO1430 (both anti-CD137xHSAxPD-L1 (high affinity) scMATCH3) were then prepared and added to the cells. After addition of the molecule of interest, Jurkat reporter cells were prepared in assay medium containing HSA at 25mg / ml and added at a cell density of 40'000 cells per well. Luciferase expression was detected by addition of luciferase reagent and read in a luminescence reader 24 hours after addition of Jurkat cells. Data was presented by plotting the relative luminescence units (RLU) of the test samples as a function of test sample concentration and fitted using a sigmoidal 4PL fit (GraphPad Prism).

[0315] result: As shown in Figure 15, PRO1993 was found to induce CD137 signaling in Jurkat reporter cell line when cultured in the presence of HCC1954 cancer cells, while only slight activation of CD137 was observed in the presence of HCC827 cancer cells. NF-kB reporter gene assay data for ND029 molecule are shown in Table 26.

[0316] [Table 26]

[0317] PD-L1 / PD-1 neutralization by FC method: These assays were performed to evaluate the ability of the multispecific low affinity PD-L1x / Her2 construct to neutralize the interaction between PD-1 and PD-L1 expressed on HCC1954 cells, which also express Her2. Specifically, the molecule should efficiently block PD-1 binding to PD-L1 on Her2 / PD-L1 co-expressing target cells (HCC1954), but should block PD-1 / PD-L1 interaction much less potently than avelumab / BAVENCIO® or nivolumab / OPDIVO® on cells that do not express Her2. Blockade of PD-1 binding to cells was analyzed by flow cytometry in the presence of human SA as described above. Serial dilutions of each molecule to be tested and the reference avelumab were added to the plate. The individual ICs of each plate were analyzed by ELISA using the ELISA kit. 50 Values ​​were calculated based on the IC of the reference molecule avelumab co-incorporated on each plate. 50 Calibrated against (relative IC 50 :I C 50 , avelumab / IC 50 , test molecule). Furthermore, the relative IC observed in high Her2-expressing HCC1954 cells and low Her2-expressing HCC827 cells 50 The ratio of values ​​was calculated.

[0318] result: The potency of the MATCH4 constructs of this approach is summarized in Table 27. The PD-L1 inhibition curves obtained for PRO1993 are shown in Figure 16.

[0319] [Table 27]

[0320] Example 4: Generation and testing of multispecific constructs targeting HER2 (as an example of a TAA) and PD-L1 Design and construction of scDb-scFv An illustration of the various scDb-scFv (scMATCH3) molecules designed within this aspect of the approach is shown in Figure 17. Data regarding the generation of all molecules is detailed in Table 28, showing the domains that make up the various molecules generated and their arrangement within the molecule.

[0321] method: Expression of scDb-scFv constructs was performed in CHO-S cells using the CHOgro transient transfection kit (Mirus). After 5-7 days of expression (cell viability <70%), cultures were harvested by centrifugation at 37°C and proteins were purified from clarified culture supernatants by protein L affinity chromatography followed, if required, by a polishing purification step by size exclusion chromatography. Standard analytical methods such as SE-HPLC, UV280 and SDS-PAGE were used for quality control of the produced material.

[0322] [Table 28]

[0323] Affinity to human Her2, PD-L1, and human and mouse serum albumin (SA) by SPR method: Affinities for human PD-L1, Her2, and human and mouse serum albumin (SA) were measured by SPR using a Biacore T200 instrument (GE Healthcare) as described above. The resulting SPR sensorgrams were double-referenced (empty reference channel and zero analyte injection) and fitted using BiaEvaluation software (GE Healthcare) and a one-to-one Langmuir binding model. The quality of the fit was monitored based on Chi2 and U values ​​(Biacore). The obtained kinetic parameters were used to calculate the apparent dissociation equilibrium constant (K D In addition to the kinetic measurements, the K for low affinity PD-L1 assays was calculated by fitting a plot of equilibrium response versus concentration. D was obtained (steady-state affinity measurement).

[0324] result: As shown in Table 29, binding to Her2 and human SA was similar for all molecules tested. Affinity for PD-L1 could only be determined for molecules PRO1678 and PRO1679 using steady-state analysis, and the affinities found for both proteins were comparable (K D values, PRO1678: 156 nM, PRO1679: 122 nM).

[0325] [Table 29]

[0326] PD-L1 / PD-1 neutralization by FC method: The assay was performed to evaluate the ability of the multispecific low affinity PD-L1x / Her2 construct to neutralize the interaction between PD-1 and PD-L1 expressed on HCC1954 cells, which also express Her2. Specifically, the molecule should efficiently block PD-1 binding to PD-L1 on Her2 / PD-L1 co-expressing target cells (HCC1954), but should block PD-1 / PD-L1 interaction much less potently than avelumab / BAVENCIO® or nivolumab / OPDIVO® on cells that do not express Her2. Blockade of PD-1 binding to cells was analyzed by flow cytometry in the presence of human SA as described above. Individual ICs from each plate were analyzed by ELISA using the ELISA kit. 50 Values ​​were calculated based on the IC of the reference molecule avelumab, which was co-incorporated on each plate. 50 Calibrated against (relative IC 50 :I C 50 , avelumab / IC 50 , test molecule). Furthermore, the relative IC observed in high Her2-expressing HCC1954 cells and low Her2-expressing HCC827 cells 50 The ratio of values ​​was calculated.

[0327] result: The efficacy of the scMATCH3 constructs of this approach is summarized in Table 30.

[0328] [Table 30] Binding to Her2-expressing SK-OV3 and MCF-7 cells by flow cytometry method: SK-OV3 cells (from ATCC, catalog HTB-77, human ovarian adenocarcinoma cells expressing very high levels of HER2 and very low levels of PD-L1), MCF-7 (human breast cancer cells expressing low levels of HER2 and very low levels of PD-L1) and CHO PD-L1 (from Amsbio, control cells expressing high levels of PD-L1 and no human Her2) were collected and cell counts were performed. The cell suspension was centrifuged at 400×g for 5 min and 100 μl (50'000 cells) of the cell suspension diluted in PBS-EB (1×DPBS, 2% BCS HI, 2 mM EDTA) was added to the designated wells of the non-binding plate. After three washing steps with PBS-EB, the cells were centrifuged and the washing buffer was aspirated. 100 μl of the samples to be tested (MATCH4 molecules: PRO1543 based on anti-HER2 trastuzumab and PRO1895 based on anti-HER2 pertuzumab) and serial dilutions of the reference antibodies trastuzumab and pertuzumab starting at a concentration of 50 nM were added directly to the plate. A 3-fold step dilution series of the multispecific molecules and avelumab starting at 20 μg / ml and 5 μg / ml, respectively, was prepared in staining buffer containing 500 ng / ml biotinylated PD1. After 1 h of incubation at 4° C., the plate was washed 3 times with 100 μl PBS-EB. The cell pellets incubated with MATCH4 molecules were resuspended in 100 μl of framework-specific detection antibody from Numab, followed by detection by adding an APC-labeled anti-rabbit IgG antibody at a concentration of 2 μg / ml and incubated at 4° C. for 1 h. Cell pellets incubated with the reference antibodies Trastuzumab and Pertuzumab molecules were resuspended with 100 μl of RPE-labeled anti-human Fc antibody at a concentration of 5 μg / ml and incubated for 1 h at 4°C. Cells were then washed three times using 100 μl of PBS-EB. Cell pellets were resuspended in 50 μl of PBS-EB and analyzed on a NovoCyte 2060 flow cytometer instrument. For each sample, the fluorescence intensity of the APC and RPE channels of 5,000 events was recorded and the geometric mean of the fluorescence intensity MFI was calculated. Data were single referenced (subtracted for the fluorescence intensity seen in cells incubated with buffer and detection antibody alone) and the resulting concentration-response curves were fitted using 4-PL fit (GraphPad Prism software). result: The apparent binding affinity of SK-OV3, MCF-7 and CHO PD-L1 was assessed using flow cytometry. 50 Values ​​are expressed as the EC 50 The obtained EC 50 and relative EC 50 , as well as maximum binding in flow cytometry are shown in Table 31. Concentration response curves are shown in FIG. As shown in Table 31, MATCH4 molecules bind to SK-OV3 cells expressing high levels of HER2 with apparent binding affinity comparable to that of the clinical stage antibodies Trastuzumab and Pertuzumab. This result indicates that MATCH4 molecules are able to catch up with the apparent binding affinity of the bivalent anti-HER2 antibodies Trastuzumab and Pertuzumab due to the avidity effect (binding to HER2 and PD-L1) when tested for binding to cells expressing high levels of HER2 and PD-L1 (even at very low expression). On the other hand, when the binding to cells expressing both antigens at very low levels (e.g., MCF-7 cells) was evaluated, the apparent binding affinity of MATCH4 was inferior to that of clinical-stage antibodies due to the lack of avidity of the MATCH4 molecule. Residual binding to CHO PD-L1 cells was observed for pertuzumab and the pertuzumab-based MATCH4 molecule PRO1895, in contrast to the lack of binding of trastuzumab and the trastuzumab-based MATCH4 molecule PRO1543. It may be speculated that pertuzumab can bind to hamster HER2, whereas trastuzumab, which binds to a different epitope, cannot. In any case, these data indicate that the low-affinity anti-PD-L1 moiety incorporated into the MATCH4 molecule cannot bind to cells expressing PD-L1 alone. Binding to Her2-expressing HCC1954 and HCC827 cells by flow cytometry method: Apparent affinity assessment of PRO1543 and PRO1895 and the clinical stage anti-HER2 antibodies trastuzumab and pertuzumab binding to HCC1954 (which expresses high levels of Her2 and PD-L1) and HCC827 (which expresses low levels of Her2 but high levels of PD-L1) was performed by flow cytometry. HCC827 and HCC1954 cells were stimulated with IFNy for 24 hours to further increase PD-L1 expression and then tested in flow cytometry experiments as described above, except that a dilution series of proteins was started at 150 nM.

[0329] [Table 31]

[0330] result: Individual ECs from each plate 50 Values ​​are expressed as the EC 50 The obtained EC 50 and relative EC 50, as well as maximum binding in flow cytometry are shown in Table 32. Concentration response curves are shown in FIG. As shown in Table 32, MATCH4 molecule bound to IFNy-stimulated HCC827 cells expressing high levels of PD-L1 and low levels of HER2 with apparent binding affinity comparable to that of clinical-stage antibodies Trastuzumab and Pertuzumab, whereas the apparent binding affinity of MATCH4 was inferior to that of the clinical-stage antibodies when binding to cells expressing high levels of both antigens (IFNy-stimulated HCC1954 cells) was assessed. Simultaneous binding of PRO1543 and pertuzumab and PRO1895 and trastuzumab to SK-OV3 cells by flow cytometry method: The ability of MATCH4 molecules to bind to HER2-expressing SK-OV3 cells in the presence of saturating concentrations of trastuzumab and pertuzumab was assessed by flow cytometry. Instead of testing serial dilutions of trastuzumab and pertuzumab, a high concentration of antibody (50 nM) that resulted in saturation of binding to SK-OV3 cells was added to the cells before the addition of MATCH4 molecules. After incubating cells with trastuzumab or pertuzumab for 1 h at 4°C, serial dilutions of MATCH4 molecules starting from 50 nM were added to the cells. Cell membrane binding of the MATCH4 molecules PRO1543 and PRO1895 was then assessed by flow cytometry as described above. result: Individual ECs from each plate 50 Values ​​are expressed as the EC50 of each reference (i.e., binding of MATCH4 molecules in the absence of the anti-HER2 antibodies trastuzumab and pertuzumab). 50 The obtained EC 50 and relative EC 50 , as well as maximum binding in flow cytometry are shown in Table 33. Concentration response curves are shown in FIG. As shown in Table 33 and Figure 21, the MATCH4 molecules bound to SK-OV3 cells in the absence of the antibodies Trastuzumab and Pertuzumab with apparent binding affinities comparable to those found in the previous experiments (see Table 31). Of note, PRO1543 bound to SK-OV3 cells in the presence of Pertuzumab and PRO1895 bound to SK-OV3 cells in the presence of Trastuzumab with binding affinities (relative EC 50 The results showed that PRO1543 and PRO1895 could also bind to cells at 100% HER2 / 200 / 200 / 100 / 200 / 100 / 200 (compare values). In contrast, no binding of PRO1543 and PRO1895 was observed when SK-OV3 cells were incubated with trastuzumab and pertuzumab, respectively, prior to the addition of MATCH4 molecules. Since PRO1543 contains the anti-HER2 binding portion of trastuzumab and PRO1895 contains the anti-HER2 binding portion of pertuzumab, the binding of PRO1543 in the presence of pertuzumab and PRO1895 in the presence of trastuzumab was expected and shown in this experiment. Trastuzumab and pertuzumab bind to different, non-overlapping epitopes on human HER2.

[0331] [Table 32]

[0332] [Table 33]

Claims

1. An antibody or antigen-binding fragment thereof that specifically binds to HAS, the antibody or antigen-binding fragment thereof comprising: (i) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 45, 46, and 47, respectively, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 48, 49, and 50, respectively; (ii) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 53, 54, and 55, respectively, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 56, 57, and 58, respectively; or (iii) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs: 61, 62, and 63, respectively, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs: 64, 65, and 66, respectively; Including; wherein one or more of said CDR sequences optionally contain one or two mutations, an antibody or antigen-binding fragment thereof.

2. The antibody or antigen-binding fragment thereof described in claim 1, wherein the one or two mutations are mutations in alanine residues.

3. The antibody or antigen-binding fragment thereof, (i) a VH domain having an amino acid sequence within its framework sequences that is at least 95 percent identical to the amino acid sequence of SEQ ID NO: 51; and a VL domain having an amino acid sequence within its framework sequences that is at least 95 percent identical to the amino acid sequence of SEQ ID NO: 52; (ii) a VH domain having an amino acid sequence within its framework sequences that is at least 95 percent identical to the amino acid sequence of SEQ ID NO: 59; and a VL domain having an amino acid sequence within its framework sequences that is at least 95 percent identical to the amino acid sequence of SEQ ID NO: 60; (iii) a VH domain having an amino acid sequence within its framework sequences that is at least 95 percent identical to the amino acid sequence of SEQ ID NO: 67; and a VL domain having an amino acid sequence within its framework sequences that is at least 95 percent identical to the amino acid sequence of SEQ ID NO: 69; or (iv) a VH domain having an amino acid sequence within its framework sequences that is at least 95 percent identical to the amino acid sequence of SEQ ID NO: 68; and a VL domain having an amino acid sequence within its framework sequences that is at least 95 percent identical to the amino acid sequence of SEQ ID NO:

70. The antibody or antigen-binding fragment thereof of claim 1 .

4. below: (i) a VH domain of SEQ ID NO: 51 and a VL domain of SEQ ID NO: 52; (ii) a VH domain of SEQ ID NO:59 and a VL domain of SEQ ID NO:60; (iii) a VH domain of SEQ ID NO: 67 and a VL domain of SEQ ID NO: 69; or (iv) a VH domain of SEQ ID NO: 68 and a VL domain of SEQ ID NO: 70; The antibody or antigen-binding fragment thereof of claim 3 .

5. A multispecific antibody comprising: (i) at least one antibody or antigen-binding fragment thereof according to any one of claims 1 to 4; and (ii) at least one antibody or antigen-binding fragment thereof having specificity for a target different from HAS.

6. The multispecific antibody of claim 5, wherein the target distinct from HAS is a tumor-associated immune checkpoint antigen or a tumor-associated antigen (TAA).

7. The multispecific antibody of claim 6, wherein the tumor-associated immune checkpoint antigen is selected from the group consisting of PD-L1, PD-L2, CD80, CD86, CD276 (B7-H3), and VTCN1 (B7-H4).

8. The multispecific antibody of claim 6, wherein the tumor-associated immune checkpoint antigen is an inhibitory immune cell antigen selected from the group consisting of CTLA4, PD-1, lymphocyte activation gene 3, T cell immunoglobulin mucin-3, BTLA, TIM3, TIGIT, CD160, LAG3, LAIR1, B7-1, and B7-H1.

9. The TAA is selected from the group consisting of EGFRvIII, 5T4, CD19, CD20, CD22, CD38, BCMA, IL4RA, mesothelin, GD2, Tn antigen, sTn antigen, Tn-O-glycopeptide, sTn-O-glycopeptide, PSMA, CD97, TAG72, CD44v6, CEA, EPCAM, KIT, IL-13Ra2, leguman, GD3, CD171, IL-11Ra, IL-13RA2, ROR1, PSCA, MAD-CT-1, MAD-CT-2, VEGF, VEGF-β ... GFR2, CLEC12A, Lewis Y, CD24, PDGFR-β, SSEA-4, folate receptor α, ERBB (e.g., ERBB2), Her2 / neu (HER2), MUC1, MUC16, EGFR, NCAM, ephrin B2, CAIX, LMP2, sLe, HMWMAA, o-acetyl-GD2, folate receptor β, TEM1 / CD248, CD33, CD123, CD133, CD135, TEM7R, FAP, legumain, HPV E6 or E7, ML-IAP, CLDN6, TSHR, GPRC5D, ALK, polysialic acid, Fos-related antigen, neutrophil elastase, TRP-2, CYP1B1, sperm protein 17, beta human chorionic gonadotropin, AFP, thyroglobulin, PLAC1, globo H, RAGE1, MN-CA IX, human telomerase reverse transcriptase, intestinal carboxylesterase, mut 7. The multispecific antibody of claim 6, selected from the group consisting of hsp70-2, NA-17, NY-BR-1, UPK2, HAVCR1, ADRB3, PANX3, NY-ESO-1, GPR20, Ly6k, OR51E2, TARP, GFRα4, GPC3, CDH3, B7H3, FGFR1, SSTR2, CECAM6, GA733, and gp120.

10. The multispecific antibody according to any one of claims 5 to 9, further comprising (iii) at least one antibody or antigen-binding fragment thereof having specificity for an immune cell antigen.

11. (i) the immune cell antigen is a stimulatory immune cell antigen selected from the group consisting of CD3 and CD16; or 11. The multispecific antibody of claim 10, wherein the immune cell antigen is a costimulatory immune cell antigen selected from the group consisting of CD137, CD28, ICOS, HVEM, CD27, OX40, DR3, GITR, CD30, SLAM, CD2, 2B4, TIM1, TIM2, and CD226.

12. A pharmaceutical composition comprising the multispecific antibody of any one of claims 5 to 11 and a pharma- ceutically acceptable carrier.

13. A nucleic acid encoding an antibody or antigen-binding fragment thereof according to any one of claims 1 to 4 or a multispecific antibody according to any one of claims 5 to 11.

14. A vector comprising the nucleic acid of claim 13.

15. A host cell comprising a nucleic acid according to claim 13 or a vector according to claim 14.

16. A method for producing an antibody or antigen-binding fragment thereof according to any one of claims 1 to 4 or a multispecific antibody according to any one of claims 5 to 11, comprising: (i) culturing the host cell of claim 15; and (ii) expressing the antibody or antigen-binding fragment thereof according to any one of claims 1 to 4 or the multispecific antibody according to any one of claims 5 to 11, A method comprising:

Images