Multispecific antibodies

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. Further, 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 a nucleic acid encoding said multispecific antibody or a specific domain thereof, a vector comprising said nucleic acid, a host cell comprising said nucleic acid or said vector, and a method of producing said multispecific antibody or a specific domain thereof.

Background Art

[0002] Cancer still has significant unmet medical needs despite substantial progress in its treatment. Some of the most substantial recent advances in cancer treatment have come with the emergence of various molecular classes of immunotherapy, such as, 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 anti-tumor immunity by a) actively inducing immune effector cells to tumor resident cells, and / or b) stimulating immune effector cells, and / or c) reducing tumor-mediated immunosuppression. Generally, these immunotherapies direct their pharmacological activity to tumors by taking advantage of the fact that specific antigens by tumor resident cells (e.g., malignant cells, cells of the tumor vasculature, stromal cells, immune cells, etc.) are overexpressed relative to extra-tumor sites. Among these antigens, tumor-associated antigens (TAAs) include cell surface proteins that are selectively overexpressed in malignant cells. By binding to TAAs with high affinity, immunotherapy can to some extent confine immunomodulatory activity to the immune synapse between tumor cells and immune effector cells.

[0003] A common class of TAA-conjugated immunotherapies are mAbs that opsonize tumor cells and induce anti-tumor immunity by causing antibody-dependent cell-mediated cytotoxicity (ADCC) by Fcγ receptor (FcγR) expressing cells, primarily natural killer (NK) cells. Similar to bsAbs (TAA / CD3 bsAbs) that conjugate simultaneously with the T cell antigen CD3, other TAA-conjugated immunotherapies also utilize cytotoxic T lymphocytes (CTLs) to induce targeted depletion of malignant cells such as CAR-T cells. While the therapeutic utility of TAA-redirecting CTLs has been clinically validated, such utility may be limited to cases where tumor-mediated immunosuppression impairs CTL activation / stimulation. Even in tumors with a very high number of tumor-infiltrating lymphocytes (TILs) (i.e., "inflammatory" or "hot" tumors), tumor immune evasion can be induced by various means, including through the expression of immune checkpoint ligands / receptors (e.g., PD-1, PD-L1, CTLA-4) and the 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. PD-L1 is expressed by activated T and B cells and is a surface glycoprotein ligand for PD-1, a key immune checkpoint receptor that mediates immunosuppression. PD-L1 is involved in suppressing immune responses in chronic infections, pregnancy, tissue transplantation, autoimmune diseases, and cancer. PD-L1 is found in both antigen-presenting cells and human cancer cells such as head and neck squamous cell carcinoma, melanoma, brain tumors, thyroid, thymus, esophagus, lung, breast, gastrointestinal tract, colon, liver, pancreas, kidney, adrenal cortex, bladder, urothelium, ovaries, 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 in normal tissues but is inductively expressed in 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). When PD-1 is activated by PD-L1, it may provide an immune-tolerant environment for tumor development and growth.PD-L1 also negatively modulates T cell function through its interaction with another receptor, B7.1 (B7-1, CD80).

[0006] Inhibition of PD-L1 / PD-1 interaction enables 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, New Jersey; pizilizumab, CureTech, Yavne, Israel) and those targeting PD-L1 (MPDL3280A, Genentech, South San Francisco, California; MEDI4736, MedImmune / AstraZeneca; BMS-936559, Bristol-Myers Squibb; MSB0010718C, EMD Serono, Rockland, Massachusetts) (see Postow MA et al., J Clin Oncol. Jun 10;33(17):1974-82 (2015) for a review). Targeting PD-L1 can produce different biological effects compared to targeting PD-1. PD-1 antibodies inhibit the interaction between PD-1 and both its ligands, PD-L1 and PD-L2. While the effects of the interaction between PD-1 and PD-L2 are not yet understood, PD-L1 antibodies do not inhibit this interaction. However, PD-L1 antibodies are thought to inhibit not only the interaction with PD-1 but also the interaction between PD-L1 and B7-1 (Butte MJ, et al., Immunity 27:111-122, (2007)), thus exerting a negative signal on T cells. Blocking PD-L1 has shown promising initial data, and currently, the following four clinical anti-PD-L1 mAbs are under trial. 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] Novel and emerging therapies often combine TAA-targeted immunotherapy with one or more additional immunotherapies targeting immune checkpoint pathways to further mitigate or overcome tumor-mediated immunosuppression. Monoclonal antibodies that block immunosuppressive antigens such as CTLA-4 (e.g., ipiriumumab), PD-1 (e.g., nivolumab, pembrolizumab), and PD-L1 (e.g., avelumab, atezolizumab) have shown excellent response rates in patients with diverse tumor histological features. Early outcomes of treatments combining immune checkpoint modulators with TAA-conjugated immunotherapy have been promising. For example, trastuzumab (Herceptin®, Genentech), a HER2-targeting mAb, 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) (National Clinical Trial (NCT) 03409848). Similarly, blinatumomab (Blincyto®, Amgen), a CD19 / CD3 bsAb, 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 (NCT 03512405 and NCT 02879695, respectively).

[0008] Furthermore, combination therapies targeting immune checkpoints and T cell costimulatory receptors are being evaluated. The combination of anti-PD-L1 antibody and anti-CD137 antibody extended overall survival and enhanced T cell effector function in an 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 currently being investigated in a Phase I / II trial in both solid tumors and B-cell non-Hodgkin lymphoma (NCT02253992), while PF-05082566 (anti-CD137) is being investigated in a Phase Ib trial with pembrolizumab (anti-PD-1) in patients with solid tumors (NCT02179918) (Chester C., et al., Cancer Immunol Immunother Oct;65(10):1243-8 (2016)).

[0009] In recent years, the effects of polyvalent and multispecific fusion polypeptides that bind to PD-L1 and CD137 on T cell activation and proliferation have been evaluated in vitro. In an autologous in vitro co-culture system introducing immature DCs and donor-matched T cells, the multispecific, polyvalent polypeptide INBRX-105, possessing two PD-L1 binding domains, two CD137 binding domains, and an Fc region, has been shown to be superior in stimulating interferon-gamma production in the induction of IFNγ or mediating the proliferation and activation of CD8+ T cells compared to monospecific PD-L1 sd-Ab-Fc fusion proteins, CD137 sdAb-Fc fusion proteins, combinations thereof, the anti-PD-L1 antibody atezolizumab, the anti-CD137 antibody utomirumab (PF-05082566), or the anti-PD-L1 antibody pembrolizumab (prembrolizumab), and combinations thereof (International Publication No. 2017 / 123650). Furthermore, in International Publication No. 2016 / 149201, a specific antibody against PD-L1 is disclosed, suggesting the creation of a bispecific antibody construct that includes further T cell-inducing antibodies, with CD137 included in a non-exclusive list of more than 20 potential T cell targets.

[0010] Combination immunotherapies have shown potential to enhance antitumor responses through additive or synergistic activity, but are always accompanied by two consistent limitations: 1) the challenges of clinical development due to the complexity of adjusting the dosages of multiple constituent therapies across different patient cohorts, and 2) the impact of reliance on two or more separate manufacturing processes for the constituent therapies and the associated high cost of goods sold (COGS) and pricing. These limitations become more severe as the number of immunotherapies included in a combination regimen increases. Furthermore, even in treatment regimens containing a single immunotherapy, dose-limiting toxicity (DLT) often prevents administration at the maximum effective dose (MED) or leads to treatment discontinuation, resulting in limited efficacy. Unfortunately, drug-related toxicities induced by each constituent immunotherapy in a combination regimen, as well as their antitumor activity, also tend to be additive or synergistic.

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

[0012] While the exact pathways by which such DLTs arise may vary, the risks of immunotherapy-related toxicity can typically be minimized or eliminated by enhancing the tumor localization of pharmacological activity. Extratumor activity of immunotherapy can lead to the secretion of pro-inflammatory cytokines in healthy tissue, resulting in an undesirable safety profile. Leveraging T-cell-induced bsAbs that require binding to TAAs to elicit immunomodulatory activity is a promising strategy to limit such cytokine release to cytolytic / immunological synapses between tumor-resident cells and T cells. However, conventional TAA / CD3 bsAbs also carry toxicity, such as cytokine release syndrome (CRS), which is usually presumed to be due to excessive activity of the anti-CD3 domain. Furthermore, while TAA / CD3 bsAbs potently reduce TAA-overexpressing cells, they 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-responsive T cells). Therefore, TAA / CD3 bsAbs may stimulate CTLs somewhat indiscriminately rather than stimulate or reactivate the host's intrinsic anti-tumor immunity, potentially posing a safety risk and leading to inadequate anti-cancer 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 of the advantages of antitumor T cell stimulation via such targets is presumed to be their transient expression during TCR signaling. Therefore, their expression tends to be selectively upregulated in inflammatory TMEs, particularly tumor-responsive T cells, whose TCRs are constantly stimulated by extensive interactions with major histocompatibility complexes (MHC) expressed by malignant cells and antigen-presenting cells (APCs). Thus, targeting costimulatory receptors using mAbs and bsAbs, for example, should stimulate and expand existing antitumor T cells more selectively than approaches targeting CD3, potentially making such biologics safer and more effective.

[0014] Among the co-stimulatory receptors, 4-1BB (CD137, TNF receptor superfamily 9, TNFRSF9) has emerged as particularly promising due to its expression profile and its role as a pluripotent mediator of antitumor immunity (Bartkowiak and Curran 2015, Yonezawa et al. 2015). 4-1BB is an inducible T cell co-stimulatory 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 shown in tumor vascular systems (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 co-stimulates T cells to perform effector functions such as eradication of established tumors, expansion of the primary CD8+ T cell response, and enhancement of the memory pool of antigen-specific CD8+ T cells. In vivo efficacy studies in mice have shown that 4-1BB agonist mAbs, whether administered as monotherapy or as part of a combination regimen, induce antitumor-protective T cell memory responses and tumor regression in multiple tumor models. Furthermore, two 4-1BB agonist mAbs are currently in clinical use: ulurumab (PF-05082566, Pfizer), a fully humanized IgG4 mAb, and utomirumab (Chester C., et al., Cancer Immunol Immunother Oct;65 (10):1243-8 (2016)), a fully humanized IgG2 mAb. While the use of 4-1BB agonist mAbs is a very promising therapeutic strategy, the clinical data collected to date suggests that approaches based on mAbs that stimulate 4-1BBs involve a trade-off between efficacy and safety. Specifically, highly active 4-1BB agonist mAbs induce DLTs that reduce therapeutic efficacy, while less active 4-1BB agonist mAbs, although well-tolerated, appear to be less effective, including in their predicted MEDs.

[0016] Highly active 4-1BB agonist mAbs induce alterations in 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 elevated aspartate aminotransferase and alanine aminotransferase levels consistent with hepatic 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 on the therapeutic use of 4-1BB agonist mAbs in humans have shown elevated liver enzymes and an increased incidence of 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 the Phase II CD137 trial of Bristol-Myers Squibb (BMS) for previously treated stage III / IV melanoma (NCT 00612664). That trial and several others (NCT00803374, NCT00309023, NCT00461110, NCT00351325) were discontinued due to adverse events (Chester C., et al., Cancer Immunol Immunother Oct;65(10):1243-8 (2016)). These adverse events were almost certainly due to systemic hyperstimulation of T cells.

[0017] Similar to TAA / CD3 bsAbs, TAA / 4-1BB bsAbs are designed to selectively activate 4-1BB in the immunological synaptic context between tumor-resident cells and immune effector cells, thereby preventing toxicity associated with extratumor T cell stimulation. As an example, the 5T4 / 4-1BB bsAb (APV-527) (International Publication No. 2017182672(A1) brochure), jointly developed by Aptevo Therapeutics and Alligator Biosciences, is designed to induce targeted T cell costimulation by immobilizing 5T4, a TAA expressed by various solid tumors. Preclinical data of 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 tumor mesenter and inhibiting tumor growth in 5T4+ tumor models. This same tumor localization strategy could potentially leverage a variety of clinically validated TAAs whose therapeutic targets have been proven effective and safe.

[0018] HER2 is established as a target-mediated antimicrobial agent (TAA) that can effectively and safely target 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). While trastuzumab and pertuzumab are similar in that they act partially by opsonizing HER2+ cells and inducing ADCC, the two antibodies differ in their mechanisms of inhibiting proliferative HER2 signaling. In the case of trastuzumab, its binding to the epitope prevents HER2 homodimerization, thereby inhibiting HER2 signaling. However, in some patients, compensatory HER3 overexpression and HER2 / HER3 heterodimer formation lead to enhanced signaling, and such patients become refractory to trastuzumab treatment. On the other hand, pertuzumab binds to an epitope that inhibits HER2 / HER3 heterodimerization and similarly inhibits proliferative signaling. Due to the complementarity of this 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 exhibit a HER2+ tumor phenotype that is very weak in response to conventional antibody treatments. This is because, in some cases, certain HER2+ tumors are independent of HER2 signaling in terms of proliferation and negate the primary mechanism of action of trastuzumab / pertuzumab. This has led to the hypothesis that a more potent targeted cytotoxic approach than ADCC may be highly beneficial. The validity of this concept has been partially demonstrated by the commercial approval of the ADC trastuzumab-emtansine (Kadcyla®, Genentech). Similarly, several companies are currently developing HER2 / CD3 bsAbs that stimulate redirected T cells to induce potent targeted cytotoxicity. Furthermore, in some patients who are primary or secondary unresponsive to HER2-targeted mAbs, the HER2+ tumor phenotype includes increased expression of ligands / receptors (e.g., PD-L1) that actively suppress the antitumor immune response. As expected, this has led to the combination of HER2-targeted mAbs and immune checkpoint-modulating mAbs, which have shown some clinical success. TAAs that are expressed almost exclusively 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 levels than on cancer cells) are called non-clean TAAs. Non-clean TAAs are problematic because the TAA / CD3 bsAb approach, due to its very high potency, can lead 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, at low levels, on various other tissues. Therefore, there is a need for novel therapies that improve the selectivity of the TAA / CD3 bsAb approach to tumor tissue.

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

[0021] As mentioned earlier, an unavoidable drawback of combination therapies is that their clinical development can become burdensomely complex and therefore expensive, especially as the number of constituent therapies increases. Development costs further increase due to the need to develop multiple manufacturing processes, and COGS increases. Incorporating more than two specificities into a single molecule (e.g., triplicate or quadruplicate antibodies) could theoretically address many of the aforementioned limitations regarding safety, efficacy, and cost. Triplicate / quadrispecificate molecules targeting TAAs could theoretically localize highly to tumors and induce synergistic antitumor modulation of multiple immune checkpoint pathways, potentially providing safer and more effective treatments for various cancers. Furthermore, such molecules would further limit the need to administer additional immunotherapies in conjunction to enhance patient response, boosting ease of development and minimizing treatment costs. However, the practical use of triplicate / quadrispecificate antibodies for therapeutic purposes is complicated by issues of their molecular structure, the characteristics of the antigen-binding domains they comprise, and / or their poor biophysical properties. Therefore, 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 to tumor-associated immune checkpoint antigens, TAAs, and / or T-cell costimulatory receptors already exist, the complex and specific requirements of such triple or quadruple-specific antibodies necessitate the development of novel antibody domains with tailor-made properties.

[0023] Thus, while numerous treatment options exist for cancer patients, there remains a need for effective and safe therapeutic agents and for their preferential use in a more targeted manner. Immunomodulatory biological agents offer a promising approach to cancer treatment due to their mechanism of action, but the lack of global immune stimulation and limitation of this immunomodulation to pathologically relevant cells and sites can lead to numerous side effects and significant toxicity, potentially resulting in increased morbidity and mortality in patients. Therefore, the object of this invention is to provide pharmaceuticals for improving the treatment of proliferative disorders, particularly cancer. [Overview of the project]

[0024] The object of the present invention is to provide pharmaceuticals for improving the treatment of proliferative disorders, particularly cancer. The present invention addresses the need for highly precise therapeutic agents for immuno-oncology that target only disease-associated cells.

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

[0026] The present invention further relates to a multispecific antibody comprising at least a first domain that specifically binds to tumor-associated immune checkpoint antigens with low affinity, at least a second domain that specifically binds to tumor-associated antigens (TAAs), and at least a third domain that specifically binds to immune cell antigens, wherein the immune cell antigen is present on T cells or NK cells.

[0027] More specifically, the present invention relates to a multispecific antibody in which 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 of (i) the multispecific antibody of the present invention and (ii) a second compound selected from (iia) an antibody against TAA, particularly an antibody against HER2, particularly trastuzumab, (iib) a regulator of immune checkpoint antigens that are not particularly tumor-associated immune checkpoint antigens and / or particularly present on T cells or NK cells, and (iic) a regulator of angiogenesis.

[0029] In another embodiment, the present invention relates to a pharmaceutical composition comprising the multispecific antibody of the present invention and a pharmaceutically acceptable carrier.

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

[0031] In a further embodiment, the present invention provides multispecific antibodies or pharmaceutical compositions of the present invention for use in the treatment of cancer in subjects requiring such treatment.

[0032] In one embodiment, the present invention provides the use of the multispecific antibodies or pharmaceutical compositions of the present invention for treating cancer in a subject requiring such treatment.

[0033] In one embodiment, the present invention provides the use of the multispecific antibody or pharmaceutical composition of the present invention in the manufacture of a pharmaceutical for the treatment of cancer in a subject that requires it.

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

[0035] In a further embodiment, the present invention provides a nucleic acid comprising a nucleotide sequence encoding the multispecific antibody of the present invention; in a further embodiment, the present invention provides a vector comprising the nucleic acid; and in a further embodiment, the present invention provides a host cell comprising the nucleic acid or the vector.

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

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

[0038] 1. A multispecific antibody, (a) A first domain that specifically binds to tumor-associated immune checkpoint antigens, in all cases measured by SPR in scFv format (monovalent affinity), and which binds to the tumor-associated immune checkpoint antigen with a dissociation constant (KD) greater than 50 nM, particularly with a dissociation constant of 50 nM to 1 μM, particularly greater than 100 nM, particularly with a dissociation constant of 100 nM to 900 nM, particularly greater than 200 nM, particularly A first domain that binds with a dissociation constant of 200 nM to 800 nM, especially greater than 300 nM, especially greater than 400 nM, especially greater than 400 nM, especially greater than 600 nM, especially greater than 450 nM, especially greater than 450 nM, especially greater than 550 nM, especially greater than 475 nM, especially greater than 525 nM, especially with a dissociation constant of approximately 500 nM (KD), and (b) A second domain that specifically binds to tumor-associated antigens (TAAs). A multispecific antibody containing [specific antibody].

[0039] 2. A multispecific antibody of item 1, measured by SPR in scFv format (monovalent affinity), wherein the second domain binds to the TAA 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, and particularly less than 0.5 nM.

[0040] 3. A multispecific antibody of item 1 or item 2, in which both the tumor-associated immune checkpoint antigen and the TAA are present on the same tumor cells.

[0041] 4. A multispecific antibody selected from item 1 to 3, which is one of the tumor-associated immune checkpoint antigens selected from the group consisting of PD-L1, PD-L2, CD80, CD86, CD276 (B7-H3), and VTCN1 (B7-H4).

[0042] 5. A multispecific antibody of item 4 in which the tumor-associated immune checkpoint antigen is PD-L1.

[0043] 6. A multispecific antibody from any one of items 1 to 5, wherein the first domain is a tumor-associated immune checkpoint antigen inhibitor.

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

[0045] 8. The TAAs mentioned above are EGFRvIII, 5T4, CD19, CD20, CD22, CD38, BCMA, IL4RA, mesoserine, GD2, Tn antigen, sTn antigen, Tn-O-glycopeptide, sTn-O-glycopeptide, PSMA, CD97, TAG72, CD44v6, CEA, EPCAM, KIT, IL-13Ra2, Legman, 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, Regmaine, HPV A multispecific antibody selected from any one of items 1-7 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, β-human chorionic gonadotropin, AFP, thyroglobulin, PLAC1, globoH, 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 TAA is a multispecific antibody of item 8, selected from HER2 and mesoserine, and particularly HER2.

[0047] 10. A multispecific antibody comprising a third domain that specifically binds to an immune cell antigen, particularly one of items 1 to 8 present on T cells or NK cells.

[0048] 11. A multispecific antibody of item 10, wherein the third domain specifically binds to an immune cell antigen that is a stimulating or co-stimulating molecule of the immune cell.

[0049] 12. A multispecific antibody according to item 11, wherein the third domain is an agonist and the immune cell antigen is a stimulating immune cell antigen.

[0050] 13. A multispecific antibody of item 12, selected from the group consisting of CD3 and CD16 as the stimulating immune cell antigen.

[0051] 14. A multispecific antibody of item 13 in which the stimulating immune cell antigen is CD3, particularly CD3ε.

[0052] 15. A multispecific antibody according to item 11, wherein the third domain is an agonist and the immune cell antigen is a co-stimulated immune cell antigen.

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

[0054] 17. A multispecific antibody of item 16 in which the stimulating immune cell antigen is CD137.

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

[0056] 19. A 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 suppressive immune cell antigen is a multispecific antibody of item 19 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. A multispecific antibody from any one of items 1-20, further containing a domain that specifically binds to human serum albumin (HSA).

[0059] 22. A multispecific antibody from any one of items 1 to 21, in which the domain can simultaneously bind to each of those antigens.

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

[0061] 24. The multispecific antibody may be a single-chain diabody (scDb), tandem scDb (Tandab), linear dimer scDb (LD-scDb), cyclic dimer scDb (CD-scDb), bispecific T cell-inducing antibody (BiTE; tandem di-scFv), tandem tri-scFv, tribody (Fab-(scFv)2) or vibody (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, 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) One of the multispecific antibodies from item 1 to 23, in a format selected from the group consisting of a linked scFv, Ts2Ab (dsscFv linked to the C-terminus of the heavy chain), bispecific antibodies based on a heterodimeric Fc domain, such as Knob-into-Hole antibodies (KiHs); scDb, tandem-di-scFv, tandem-tri-scFv, Fab-(scFv)2, Fab-(scFv)1, Fab, Fab-Fv2, COVD, MATCH, and DuoBodies fused to the N- and / or C-terminus of any of the chains of a heterodimeric Fc domain or any other heterodimerized domain.

[0062] 25. The antibody described above does not contain an immunoglobulin Fc region polypeptide and, optionally, does not contain the CH1 and / or CL region, is a multispecific antibody from any one of items 1 to 24.

[0063] 26. A multispecific antibody from any one of items 1 to 25, comprising the CH1 and / or CL region and optionally comprising an immunoglobulin Fc region polypeptide.

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

[0065] 28. A multispecific antibody from any one of items 1 to 27, wherein the antibody is scDb-scFv, tribody, DVD-tribody, or MATCH, and is in MATCH or tribody format, more specifically in MATCH format, and more specifically in MATCH3 or MATCH4.

[0066] 29. One of the multispecific antibodies listed in items 5-28, a. The first domain, as measured particularly by SPR, 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, and more preferably 400 nM to 600 nM. b. When the first domain is in scFv format, it does not bind to cells expressing PD-L1, particularly when measured by flow cytometry, and especially when the scFv concentration is less than 100 μg / ml. c. If the first domain described above is in scFv format, (i) In particular, as measured by an NFAT reporter gene assay, it does not neutralize PD-L1 binding to PD-1, or (ii) The relative potency of avelumab, measured by an NFAT reporter gene assay, is less than 0.001, preferably less than 0.0005, and neutralizes PD-L1 binding to PD-1, and the relative potency is measured by the IC of avelumab using an NFAT reporter gene assay. 50 The measured IC value (in units of ng / ml) of the scFv by the NFAT reporter gene assay. 50 This is a ratio to the value (in units of ng / ml). Multispecific antibodies.

[0067] 30. One of the multispecific antibodies from items 5 to 29, (i) In the presence of TAA- / PD-L1+ cells, the relative potency relative to the potency of avelumab, as measured by flow cytometry assay, is less than 0.001, preferably less than 0.0005, and the drug has the ability to block the interaction between PD-L1 and PD-1, wherein the relative potency is measured by the IC of avelumab by flow cytometry assay. 50 The IC value (in units of ng / ml) of the multispecific antibody measured by flow cytometry assay. 50 It is a ratio to the value (in ng / ml), and (ii) In the presence of TAA+ / PD-L1+ cells, the relative potency measured by flow cytometry assay is greater than 0.01, preferably greater than 0.05, more preferably 0.1, and the drug has the ability to block the interaction between PD-L1 and PD-1, wherein the relative potency is measured by flow cytometry assay of avelumab. 50 The measured IC value (in units of ng / ml) of the multispecific antibody by flow cytometry assay. 50 This is a ratio to the value (in units of ng / ml). Multispecific antibodies.

[0068] 31. One of the multispecific antibodies from items 5 to 30, a. When the first domain is in scFv format, the melting temperature (Tm) is measured by differential scanning fluorescence quantification to be at least 65°C, preferably at least 70°C, and in particular when the scFv is formulated in pH 6.4, 50 mM citrate phosphate buffer, and 150 mM NaCl, b. If the first domain is in scFv format, the monomer content decreases by less than 3%, preferably less than 1%, after four consecutive freeze-thaw cycles, the scFv has an initial concentration of 10 mg / ml, and in particular the scFv is formulated in 50 mM citrate phosphate buffer containing 150 mM NaCl at pH 6.4, and c. If the first domain is in scFv format, after storage at 4°C for at least two weeks, particularly at least four weeks, the monomer content decreases by less than 10%, for example, less than 9%, less than 8%, less than 7%, less than 6%, preferably less than 5%, and the scFv is at an initial concentration of 10 mg / ml, and in particular the scFv is formulated in 50 mM citrate phosphate buffer containing 150 mM NaCl at pH 6.4. multispecific antibodies

[0069] 32. A multispecific antibody that is one of items 1 to 31 in which each domain contains a heavy chain variable region (VH) and a light chain variable region (VL), a. The VH includes, in order, three complementarity determination regions: HCDR1, HCDR2, and HCDR3. b. The VL includes, in order, three complementarity determination regions: LCDR1, LCDR2, and LCDR3. Multispecific antibodies.

[0070] 33. A multispecific antibody according to any one of items 1 to 32, wherein the light chain variable region (VL) comprises a second domain that specifically binds to the tumor-associated immune checkpoint antigen and / or the TAA to the first domain, and optionally a third domain that specifically binds to the immune cell antigen, and optionally a further domain that specifically binds to human serum albumin (HSA), The VL includes a framework R4 selected from Vλ FR4, which contains an amino acid sequence having at least 80 percent, particularly at least 90 percent, identity with an amino acid sequence 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 described in any of SEQ ID NOs: 145-152, preferably Vλ FR4 described in any of SEQ ID NOs: 145-152, preferably Vλ FR4 described in SEQ ID NOs: 145, 146, or 152, more preferably Vλ FR4 described in SEQ ID NOs: 146 or 152. Multispecific antibodies.

[0071] 34. A multispecific antibody according to any one of items 1 to 33, 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 7, and one or more of the CDR sequences optionally contain one or two mutations, particularly mutations to an alanine residue, more specifically, (i) the LCDR3 contains Q108A (according to AHo numbering), (ii) the LCDR3 contains G109A (according to AHo numbering), or (iii) the LCDR3 contains Q108A and G109A (according to AHo numbering) and / or (iv) the HCDR3 contains Y112A (according to AHo numbering).

[0072] 35. A 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. A 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. A multispecific antibody of item 35 or item 36, wherein the first domain includes a heavy chain variable region (VH), and the VH is VH3 or VH4, preferably VH4.

[0075] 38. A multispecific antibody of item 35, comprising VH containing an amino acid sequence that is at least 90 percent, particularly at least 95 percent, identical to amino acid sequence sequence number 11, and VL containing an amino acid sequence that is at least 90 percent, particularly at least 95 percent, identical to amino acid sequence sequence number 15.

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

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

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

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

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

[0081] 44. A 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, particularly the VH sequence of SEQ ID NO: 34 and the VL sequence of SEQ ID NO: 36.

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

[0083] 46. ​​A multispecific antibody of item 45, wherein the third domain comprises (i) a VH directed to CD3 and containing an amino acid sequence that is at least 90 percent, particularly at least 95 percent, identical to amino acid sequence sequence number 43, and a VL containing an amino acid sequence that is at least 90 percent, particularly at least 95 percent, identical to amino acid sequence sequence number 44, or (ii) a VH directed to CD137 and containing an amino acid sequence that is at least 90 percent, particularly at least 95 percent, identical to amino acid sequence sequence number 77, and a VL containing an amino acid sequence that is at least 90 percent, particularly at least 95 percent, identical to amino acid sequence sequence number 78.

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

[0085] 48. A multispecific antibody according to any one of items 21-28, wherein the domain that specifically binds to the HSA comprises (i) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs. 45, 46, and 47, 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, 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, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs. 64, 65, and 66, respectively, and comprises one or more optionally selected CDR sequences containing one or two mutations, particularly mutations to alanine residues.

[0086] 49. A multispecific antibody of item 48, wherein the domain that specifically binds to the HSA comprises (i) a VH containing an amino acid sequence that is at least 90 percent, particularly at least 95 percent, identical to amino acid sequence sequence number 51, and a VL containing an amino acid sequence that is at least 90 percent, particularly at least 95 percent, identical to amino acid sequence sequence number 52; (i) a VH containing an amino acid sequence that is at least 90 percent, particularly at least 95 percent, identical to amino acid sequence number 59, and a VL containing an amino acid sequence that is at least 90 percent, particularly at least 95 percent, identical to amino acid sequence number 60; (i) a VH containing an amino acid sequence that is at least 90 percent, particularly at least 95 percent, identical to amino acid sequence number 67, and a VL containing an amino acid sequence that is at least 90 percent, particularly at least 95 percent, identical to amino acid sequence number 69; or (i) a VH containing an amino acid sequence that is at least 90 percent, particularly at least 95 percent, identical to amino acid sequence number 68, and a VL containing an amino acid sequence that is at least 90 percent, particularly at least 95 percent, identical to amino acid sequence number 70.

[0087] 50. A 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. A multispecific antibody of any of items 1 to 50, wherein each chain is (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 sequence numbers 125 and 126, 127 and 128, 129 and 130, 131 and 132, 133 and 134, and 135 and 136, or a combination of sequences contained in one of the sequences selected from sequence numbers 115 to 136, particularly the combination of chains of sequence numbers 123 and 124 or 127 and 128, comprising a combination of two chains having an amino acid sequence with at least 80% identity, particularly at least 90% identity, more specifically at least 95% identity, for example 100% identity, (i) a first domain that specifically binds to PD-L1, comprising the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs. 1, 2, and 3, or SEQ ID NOs. 1, 2, and 4, respectively, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs. 5, 6, and 7, SEQ ID NOs. 5, 6, and 8, SEQ ID NOs. 5, 6, and 9, respectively, or in particular SEQ ID NOs. 5, 6, and 10, respectively, and (ii)(i) a second domain that specifically binds to HER2, comprising the HCDR1, HCDR2, and HCDR3 sequences of sequence numbers 17, 18, and 19, respectively, and the LCDR1, LCDR2, and LCDR3 sequences of sequence numbers 20, 21, and 22, respectively, or in particular, (ii) the HCDR1, HCDR2, and HCDR3 sequences of sequence numbers 27, 28, and 29, respectively, and the LCDR1, LCDR2, and LCDR3 sequences of sequence numbers 30, 31, and 32, respectively. (iii) optionally, a third domain that specifically binds to (i) CD3 containing the HCDR1, HCDR2, and HCDR3 sequences of sequence numbers 37, 38, and 39, and the LCDR1, LCDR2, and LCDR3 sequences of sequence numbers 40, 41, and 42, respectively, or (ii) CD137 containing the HCDR1, HCDR2, and HCDR3 sequences of sequence numbers 71, 72, and 73, respectively, and the LCDR1, LCDR2, and LCDR3 sequences of sequence numbers 74, 75, and 76, respectively. (iv) Optionally, further domains that specifically bind to an HSA containing (i) the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs. 45, 46, and 47, 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. A multispecific antibody containing [specific antibody].

[0089] 52.(i) a multispecific antibody from any one of items 1 to 51, and (ii) a combination of a second compound selected from (iia) an antibody against TAA, particularly an antibody against HER2, particularly trastuzumab, (iib) an immune checkpoint antigen, particularly the immune checkpoint antigen that is not a tumor-associated immune checkpoint antigen, and / or a regulator of the immune checkpoint antigen that is particularly present on T cells or NK cells, and (iic) a regulator of angiogenesis.

[0090] 53. A combination of item 52 in which the regulatory factor is an antibody.

[0091] 54. A combination of item 52 or 53 in which the regulatory factor is an agonist and the immune checkpoint antigen is an immune cell antigen.

[0092] 55. A combination of 54 items in which 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 activating gene 3, and T cell immunoglobulin mucin-3, BTLA, TIM3, TIGIT, CD160, LAG3, LAIR1, B7-1, and B7-H1.

[0093] 56. A combination of item 55 in which the stimulating immune cell antigen is CD3 or CD137, particularly CD3.

[0094] 57. A combination of item 52 or 53 in which the regulatory factor is an inhibitor and the immune cell antigen is an inhibitory immune checkpoint antigen.

[0095] 58. The combination of item 57 wherein the suppressive 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 the suppressive immune checkpoint antigen is cytotoxic T lymphocyte-associated protein 4 (CTLA4), and more preferably the regulatory factor is ipilimumab.

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

[0097] 60. The combination of item 59 in which the TAA is selected from HER2 and mesoserine, and in particular HER2, and in particular the antibody is trastuzumab.

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

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

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

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

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

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

[0104] 67. One multispecific antibody from items 1-51, one combination from items 52-60, or one binding domain from items 62-66, for use as a pharmaceutical product.

[0105] For use in the treatment of cancer in a subject that requires it, any one of the multispecific antibodies of items 1 to 51, any one of the combinations of items 52 to 60, the pharmaceutical composition of item 61, or any one of the binding domains of items 62 to 66.

[0106] Use of any one of the multispecific antibodies of items 1 to 51, any one of the combinations of items 52 to 60, the pharmaceutical composition of item 61, or any one of the binding domains of items 62 to 66 for treating cancer in a subject that requires it.

[0107] Use of any one of the multispecific antibodies of items 1 to 51, any one of the combinations of items 52 to 60, the pharmaceutical composition of item 61, or any one of the binding domains of items 62 to 66 in the manufacture of a medicament for the treatment of cancer in a subject that requires it.

[0108] 71. A method for treating cancer in a subject that requires it, comprising administering to the subject a therapeutically effective amount of any one of the multispecific antibodies of items 1 to 51, 67 and 68, any one of the combinations of items 52 to 60, the pharmaceutical composition of item 61, or any one of the binding domains of items 62 to 66, or the use of item 69 or 70, wherein the cancer is a cancer that is positive for the TAA and the tumor-associated immune checkpoint antigen, particularly wherein the cancer is TAA + / PDL + and more particularly wherein the cancer is HER2 + / PD-L1 + and is a method as such.

[0109] 72. The cancer is a cancer that is positive for HER2 and PD-L1, and the use of the multispecific antibody of item 六十八, the use of item 六十九 or 七十, or the method of item 七十一, wherein the cancer is refractory to standard of care therapy, particularly trastuzumab.

[0110] [[ID=二十九]] 73. A multispecific antibody from any of items 1-51 or a nucleic acid encoding a binding domain from any of items 62-66.

[0111] 74. A vector containing nucleic acids from item 73.

[0112] 75. Host cells containing the nucleic acid of item 73 or the vector of item 74.

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

[0114] 77. A kit comprising one multispecific antibody from items 1-51, one combination from items 52-60, a pharmaceutical composition from item 61, or one binding domain from items 62-66.

[0115] 78. A multispecific antibody of any of items 1 to 50, wherein the sequence numbers are: 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, sequence numbers An antibody comprising a combination of two chains selected from sequence numbers 111 and 112, sequence numbers 113 and 114, sequence numbers 123 and 124, sequence numbers 125 and 126, sequence numbers 127 and 128, sequence numbers 129 and 130, sequence numbers 131 and 132, sequence numbers 133 and 134, and sequence numbers 135 and 136, or a combination of sequences contained in one of the sequences selected from sequence numbers 115 to 136, particularly the combination of chains of sequence numbers 123 and 124 or sequence numbers 127 and 128. [Brief explanation of the drawing]

[0116] [Figure 1] Figure 1 shows the binding of (A) PRO1434 and (B) PRO1494 to PD-L1-expressing cells as evaluated by flow cytometry. PRO830 was used as a reference. PRO1434 showed a signal only at 100 μg / ml, while binding was observed for PRO1494 at 3.5 μg / ml. [Figure 2] Figure 2 shows the blockade of PD-1 / PD-L1 interaction in an NFAT reporter gene assay. (A) PD-L1 neutralization by PRO1434 and (B) PRO1494. Avelumab was used as a reference. Both molecules tested showed only partial neutralization of the PD-1-PD-L1 interaction at the highest concentration tested (162 μg / ml). [Figure 3] Figure 3 shows the structure of the multispecific molecule of the present invention. Schematic diagrams and descriptions of the three different multispecific formats: Tribody, DVD-Tribody, and MATCH-4. Table 14 lists the domains contained in each constructed molecule and their arrangement within that 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 shows the blockade of PD-1 / PD-L1 interaction in Her2-expressing cells. Inhibition of PD-1 binding to (A) PD-L1 and Her2-expressing cells (HCC1954) or (B) PD-L1-expressing cells without significant Her2 expression (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 205 ng / ml in PD-L1 / high-Her2-expressing cells (HCC1954) and with an IC50 of 1204 ng / ml in PD-L1-expressing cells (HCC827). PRO1497, containing an anti-PD-L1 domain at 50-fold lower affinity, inhibited PD-1 binding to Her2 / PD-L1-expressing cells with comparable efficacy to PRO1454. However, in cells expressing only PD-L1, it showed only very weak inhibition of PD-1 binding at the tested concentrations. PRO1456, which does not contain an anti-PD-L1 domain, did not affect PD-1 binding in either cell line. Data were fitted using sigmoidal 4PL fit (graph pad prism). [Figure 5] Figure 5 shows the blockade of PD-1 / PD-L1 interactions in Her2-expressing cells in the presence of human serum albumin. A) Inhibition of PD-1 binding to PD-L1 and Her2-expressing cells (HCC1954) in the presence of increasing concentrations of avelumab, nivolumab, PRO1543 (Her2×CD3×HSA×PD-L1 low affinity), and PRO1546 (HER2×CD3×HSA×IL23R). B) Inhibition of PD-1 binding to PD-L1-expressing cells (HCC827) without significant Her2 expression. PRO1543 inhibited PD-1 binding to PD-L1 only in PD-L1 / high-Her2-expressing cells, with an IC50 value of 600 ng / ml. PRO1546, which does not contain an anti-PD-L1 domain, did not affect PD-1 binding in either cell line. Data were fitted using sigmoid 4PL fit (graph pad prism). [Figure 6]Figure 6 shows CD3 activation and simultaneous PD-L1 blockade by PRO1454 or PRO1497, as evaluated by an 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 at 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 an anti-PD-L1 domain. This suggests that PRO1454 blocks PD-L1 in the presence of cells co-expressing Her2 and PD-L1, while simultaneously activating CD3 within the immune synapse. Weaker activation was observed with PRO1455, a molecule lacking an anti-Her2 domain and possessing a low-affinity anti-PD-L1 domain (33-03-G02 G109A). B) Tribody molecule PRO1497, containing anti-PD-L1 domains with two alanine mutations (Q108A and G109A) that have at least 50-fold lower affinity than the molecules tested in A and the domains incorporated in the corresponding reference, was tested. In the case of these molecules, PRO1497, containing a low-affinity anti-PD-L1 domain, induced higher maximal activation than PRO1456, which instead contains an anti-IL23R domain, thus simultaneously observing PD-L1 blockade and CD3 activation. Compared to PRO1455, PRO1498 induced very weak activation due to the very low affinity of its incorporated anti-PD-L1 domain. Luminescence was read 5 hours after the addition of jarcut reporter cells, and the data were fitted using a sigmoid 4PL fit (graph pad prism). [Figure 7]Figure 7 shows the CD3 activation and simultaneous PD-L1 blockade by PRO1543, as evaluated by an 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 Jurcat cells at 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 an anti-PD-L1 domain. This suggests that PRO1543 blocks PD-L1 in the presence of cells co-expressing Her2 and PD-L1, while simultaneously activating CD3 within the immune synapse. These observations were further supported by the fact that the addition of 1 μg / ml nivolumab to all molecules resulted in maximal activation similar to that of PRO1543 alone in the presence of PRO1557 and PRO1543, demonstrating complete PD-L1 / PD-1 blockade by PRO1543. No activation was observed with PRO1546, a molecule that lacks an anti-Her2 domain but possesses a low-affinity anti-PD-L1 domain. B) In the presence of PD-L1-expressing CHO cells, no activation was observed with molecules possessing an anti-PD-L1 domain (PRO1543 and PRO1546), regardless of the presence or absence of an anti-Her2 domain. No activation was observed with PRO1557, as this molecule does not contain an anti-PD-L1 domain. Luminescence was read 5 hours after the addition of Jarcut reporter cells, and the data were fitted using a sigmoid 4PL fit (graph pad prism). [Figure 8] Figure 8 shows the activation of CD8+ T cells, measured by CD69 upregulation, in the presence of PRO1543, PRO1895, and the control molecule PRO2290, after co-incubation with HCC827 tumor cells (low HER2, PD-L1+). [Figure 9] Figure 9 shows the activation of CD8+ T cells, measured by CD69 upregulation, 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 shows the viability of CD4+ and CD8+ T cells. CD4+ T cell (A) and CD8+ T cell (B) viability decreased by only 5-10% at the highest concentrations of the tested molecule. Cytotoxicity of PBMCs in the presence of PD-L1 / high-Her2 expressing cancer cells (HCC1954) was assessed by staining CD4+ and CD8+ T cells with a fluorescently labeled antibody 40 hours after the start of the disease cycle and analyzing the results by flow cytometry. Similar data were obtained at 16 hours. [Figure 11] Figure 11 shows T cell-mediated cytotoxicity of target cells and activation of CD8+ cells in the presence of A) Her2+++ / PD-L1+ HCC1954 and B) Her2+ / PD-L1-MCF-7. In this assay, fresh isolated human PBMCs were co-cultured for 16 hours with the target cells shown in the figure in the presence of various test molecules. Compared to Her2 / CD3 scDb(PRO957), PRO1543 showed only slightly different efficacy in PD-L1-negative cells, while it showed 50-100 times better efficacy on Her2+++ / PD-L1+ cells. Since the EC50 for PD-L1 blockade in these cells was considerably higher than the EC50 for target cell lysis, it is highly likely that the improved activity of molecules containing the anti-PD-L1 domain arose from increased avidity. Consequently, binding to Her2 and PD-L1 bipositive cells was stronger than binding to Her2-expressing PD-L1-negative cells. This avidity binding selectively enhances efficacy against tumor cells (Her2 / PD-L1 double-positive) but does not enhance efficacy against PD-L1-negative Her2-expressing normal cells, thus broadening the therapeutic window. [Figure 12]Figure 12 shows T cell-mediated cytotoxicity of target cells and CD8+ cell activation in the presence of A) Her2+ / - / PD-L1+ HCC827 and B) Her2- / PD-L1+ CHO PD-L1 cells. Fresh isolated human PBMCs were co-cultured for 16 hours with the target cells shown in the figure in the presence of various test molecules. In Her2+ / - / PD-L1+ cells, PRO1543 showed 20 times better efficacy than Her2 / CD3 scDb. In the presence of Her2-negative PD-L1-positive cells, only slight cytotoxicity of target cells and CD8+ cell activation were observed at high concentrations of MATCH4, which has a low-affinity PD-L1 domain, allowing for a very broad therapeutic range. [Figure 13] Figure 13 shows human PBMC-substituted NOG mice transplanted with HCC1954 ductal carcinoma cells (n=8 each). Mice were administered the treatment on days 0, 5, 10, 15, 20, 25, and 30 (dotted vertical lines). Tumor growth and body weight were recorded twice weekly. PRO1678 scMATCH3 exhibited antitumor effects similar to nivolumab and showed efficient tumor-targeted PD-L1 blockade. PRO1543 MATCH4 therapy resulted in a higher antitumor effect than the nivolumab / trastuzumab combination. [Figure 14] Figure 14 shows the design of a multispecific molecule. Schematic diagrams and descriptions of three different multispecificity formats: tribody, DVD-tribody, and MATCH-4. Table 25 lists the domain composition and intramolecular arrangement of each constructed 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 shows the effect of PRO1993 on CD137 signaling activity in NF-κB Jurkat reporter cells. PRO1993 and Jurkat cells were incubated 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), and CD137 signaling activity was evaluated by detecting luminescence. PRO1993 activated CD137 signaling in the presence of HCC1954 cells with high Her2 expression, while only slight activation of CD137 signaling was observed in the presence of HCC827 cells with low Her2 expression. Data were fitted using a sigmoid 4PL fit (graph pad prism). [Figure 16] Figure 16 shows the blockade of the PD-1 / PD-L1 interaction in Her2-expressing cells in the presence of human serum albumin. The 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 × CD137 × HSA × low affinity for PD-L1). PRO1993 inhibited PD-1 binding to PD-L1 in PD-L1 / high-Her2 expressing cells, with an IC50 of 166.7 ng / ml, while no inhibition of PD-1 binding was observed in HCC827 cells. In comparison, avelumab inhibited this interaction, with IC50 values ​​of 127.7 ng / ml (HCC1954) and 46.03 ng / ml (HCC827). The data was fitted using a sigmoid 4PL fit (graph pad prism). [Figure 17] Figure 17 shows the design of the scDb-scFv molecule. Schematic diagram and description of the scDb-scFv molecule. [Figure 18]Figure 18 shows the blockade of the 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 Her2 levels (HCC1954) or (B) cells expressing PD-L1 and lower levels of Her2 (HCC827) in the presence of increasing concentrations of avelumab and PRO1678 (Her2 × HSA × low affinity for PD-L1). PRO1678 inhibited PD-1 binding to PD-L1 with 100 times better efficacy in PD-L1 / high-Her2-expressing cells than in PD-L1-expressing cells (HCC827), with an IC50 of 428.2 ng / ml. In comparison, avelumab inhibited this interaction, with IC50 values ​​of 127.7 ng / ml (HCC1954) and 46.03 ng / ml (HCC827). The data was fitted using a sigmoid 4PL fit (graph pad prism). [Figure 19] Figure 19 shows the cell membrane binding of MATCH4 molecules PRO1543 and PRO1895 to SK-OV3, MCF-7, and CHO PD-L1 cells. The concentration-response curves for MATCH4 molecules PRO1543 and PRO1895, as well as clinical-stage anti-HER2 antibodies trastuzumab and pertuzumab, are shown for SK-OV3 (top left), MCF-7 (top right), and CHO PD-L1 (bottom right). The 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. However, when evaluating binding to MCF-7 cells, the apparent binding affinity of the MATCH4 molecule was inferior to that of trastuzumab and pertuzumab. [Figure 20]Figure 20 shows the 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, as well as 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 flow cytometry testing. When evaluating binding to HCC1954 cells expressing high levels of HER2 and PD-L1, the apparent binding affinity of MATCH4 molecules was inferior to that of 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 an apparent binding affinity similar to that of the clinical-stage antibodies trastuzumab and pertuzumab (right) (right) (right). Note that in the graph shown on the right, the highest concentrations of PRO1543 and PRO1895 were not used for curve fitting. [Figure 21] Figure 21 shows the cell membrane binding of MATCH4 molecules PRO1543 and PRO1895 to SK-OV3 cells in the presence of trastuzumab and pertuzumab. The concentration-response curves of MATCH4 molecules PRO1543 (left) and PRO1895 (right) are shown with and without the addition of trastuzumab or pertuzumab. Cells were incubated with 50 nM trastuzumab or pertuzumab for 1 hour before the addition of the 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 in the presence of pertuzumab. On the other hand, PRO1895 showed binding to cells in the presence of trastuzumab and when tested alone. When PRO1543 and PRO1895 were tested in the presence of trastuzumab or pertuzumab, respectively, no binding was observed. [Modes for carrying out the invention]

[0117] The use of therapeutic antibodies that inhibit the interaction between tumor-associated immune checkpoint antigens such as PD-L1 and their homologous ligands such as PD-1 is a very promising therapeutic strategy, but it is accompanied by difficulties such as high toxicity and adverse events. Therefore, there is a need in the medical field for novel approaches that inhibit the interaction between tumor-associated immune checkpoint antigens and their homologous ligands with lower dose-limiting toxicity and adverse event rates than currently available approaches.

[0118] The present invention provides a multispecific antibody comprising at least a first domain that specifically binds to tumor-associated immune checkpoint antigens with low affinity, and at least a second domain that specifically binds to tumor-associated antigens (TAAs). The multispecific antibody of this disclosure can bind to target cells that present TAAs due to the first domain that specifically binds to TAAs, and the low-affinity binding domain can simultaneously bind to the tumor-associated immune checkpoint antigens present on the same target cells by avidity, thereby inhibiting the interaction of the tumor-associated immune checkpoint antigens. Due to the low affinity of the first domain, specific binding to non-target cells that present only the tumor-associated immune checkpoint antigens and do not present TAAs does not occur to a significant degree. Therefore, the multispecific antibody of the present invention does not interact with non-target cells, and by its ability to mediate, for example stimulate, potent signaling of the tumor-associated immune checkpoint antigens on target cells, treatment with the multispecific antibody of the present invention does not lead to the depletion of cells that do not express TAAs.

[0119] Furthermore, surprisingly, the multispecific antibodies of this disclosure, comprising (a) at least the first domain, (b) at least the second domain, and (c) at least a third domain that specifically binds to an immune cell antigen, have been found to exhibit further beneficial properties, as shown in the examples and accompanying figures. Moreover, the optional addition of an anti-HSA domain that extends the half-life should not only enable convenient administration but also facilitate the delivery of molecules to the tumor microenvironment.

[0120] Therefore, the multispecific antibodies of the present invention offer therapeutic advantages that differ from conventional compositions and treatments.

[0121] All technical and scientific terms used herein have the same meanings as those generally understood by those skilled in the art to which the present invention pertains, unless otherwise defined.

[0122] In this specification, unless otherwise specified, the terms “comprising” and “including” are used in their open-ended and non-restrictive sense. Therefore, with respect to such latter embodiments, the term “comprising” includes the narrower term “consisting of.”

[0123] In the context describing the present invention (particularly in the context of the following claims), the terms "a," "an," and "the," and similar references, should be interpreted as encompassing both singular and plural, unless otherwise indicated herein or the context clearly indicates otherwise. For example, the term "a cell" includes multiple cells and mixtures thereof. Where the plural form is used for compounds, salts, and similar items, this also means a single compound, salt, or similar item.

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

[0125] As used herein, the term “antibody” and similar entities include whole antibodies or their single chains, any antigen-binding fragment (i.e., “antigen-binding portion”) or its single chain, and molecules containing antibody CDR, VH region, or VL region (including, but not limited to, multispecific antibodies). Naturally occurring “whole antibodies” are glycoproteins in which at least two heavy chains (H) and two light chains (L) are interconnected by disulfide bonds. Each heavy chain consists of a heavy chain variable region (abbreviated herein as VH) and a heavy chain constant region. The heavy chain constant region consists of three domains: CH1, CH2, and CH3. Each light chain consists of a light chain variable region (abbreviated herein as VL) and a light chain constant region. The light chain constant region consists of one domain: CL. The VH and VL regions are further subdivided into a highly variable region called the complementarity-determining region (CDR) and a more conserved region called the framework region (FR). Each VH and VL consists of three CDRs and four FRs arranged from the amino terminus to the carboxyl terminus in the following order: FR1, CDR1, FR2, CDR2, FR3, CDR3, FR4. The variable regions of the heavy and light chains contain binding domains that interact with the antigen. The constant region of the antibody can mediate the binding of immunoglobulins to various cells of the immune system (e.g., effector cells) and to host tissues or factors including the first component (C1q) of the classical complement system.

[0126] As used herein, the terms “binding domain,” “antigen-binding fragment,” “antigen-binding portion,” and similar terms refer to one or more fragments of an intact antibody that possess 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 the multispecific antibody of the present invention is a Fab fragment, which is a monovalent fragment consisting of VL, VH, CL, and CH1 domains; an F(ab)2 fragment, which is a bivalent fragment containing two Fab fragments linked by disulfide crosslinking at a hinge region; an Fd fragment consisting of VH and CH1 domains; an Fv fragment consisting of the single-arm VL and VH domains of the antibody; and a single-domain antibody (dAb) fragment consisting of the VH domain (Ward et al., 1989 Nature). 341:544-546); Selected from the group consisting of isolated complementarity-determining regions (CDRs), dsFv, scAb, STAB, single-domain antibodies (sdAb or dAb), single-domain heavy-chain antibodies, and single-domain light-chain antibodies, VHH, VNAR, single-domain antibodies based on shark-derived VNAR structures, and binding domains based on alternative scaffolds, for example, but not limited to, domains based on ankyrin, fynomers, avimers, antikalin, fibronectin, and binding sites incorporated into the constant region of an antibody (e.g., f-star technology (F-star's Modular Antibody Technology®))). Preferably, the binding domain of the present invention is a single-stranded Fv fragment (scFv) or a single-antibody variable domain. In preferred embodiments, the binding domain of the present invention is a single-stranded Fv fragment (scFv).

[0127] The term "complementarity-determining region" ("CDR") refers to an amino acid sequence with a boundary determined using one of several well-known schemes, including those described below. As 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). The numbering scheme described in 657-670 ("AHo" numbering). 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), 95-102 (HCDR3), and the CDR amino acid residues of the light chain variable domain (VL) are numbered 24-34 (LCDR1), 50-56 (LCDR2), 89-97 (LCDR3). Under Chothia, the CDR amino acids of VH are numbered 26-32 (HCDR1), 52- The amino acid residues in the VL are numbered 56 (HCDR2) and 95-102 (HCDR3), respectively, while those in the VL are numbered 24-34 (LCDR1), 50-56 (LCDR2), and 89-97 (LCDR3). Combining the CDR definitions of both Kabat and Chothia, the CDR consists of amino acid residues 26-35 (HCDR1), 50-65 (HCDR2), and 95-102 (HCDR3) in human VH, and amino acid residues 24-34 (LCDR1), 50-56 (LCDR2), and 89-97 (LCDR3) in human VL.Under IMGT, the CDR amino acid residues of VH are numbered approximately 26-35 (HCDR1), 51-57 (HCDR2), and 93-102 (HCDR3), while the CDR amino acid residues of VL are numbered approximately 27-32 (LCDR1), 50-52 (LCDR2), and 89-97 (LCDR3) (numbering by "Kabat"). Under IMGT, the CDR of an antibody can be determined using the IMGT / DomainGap Align program.

[0128] In the context of this invention, unless otherwise specifically mentioned, 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 CDRs 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. To clarify, the numbering system according to Honegger & Plueckthun takes into account the sequence diversity found in various VH and VL subfamilies, particularly in naturally occurring antibodies present in CDRs, and introduces gaps within the sequence. Therefore, a given antibody variable domain is not typically occupied entirely by amino acid residues from position 1 to 149.

[0129] As used herein, the term “binding specificity” refers to the ability of an individual antibody to react with a certain antigenic determinant and not with a different antigenic determinant. As used herein, the terms “specifically binding” or “specific” refer to a measurable and reproducible interaction, such as binding between a target and an antibody, which 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 for a longer duration than an antibody that binds to other targets. In its most common form (and unless a defined reference is mentioned), “specific binding” refers to the ability of an antibody to distinguish a target of interest from an unrelated molecule, as determined, for example, by specificity assays known in the art. Such methods include, but are not limited to, Western blotting, ELISA, RIA, ECL, IRMA, SPR (surface plasmon resonance) assays, and peptide scans. For example, a standard ELISA assay can be performed. Scoring can be performed by standard color development (e.g., using a secondary antibody with horseradish-derived peroxidase and tetramethylbenzidine, hydrogen peroxide). The reaction in a well is scored by optical density at, for example, 450 nm. A typical background (= negative reaction) can be about 0.1 OD, and 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 of at least 10-fold, preferably at least 100-fold, between the background and signal indicates specific binding. Typically, the determination of binding specificity is performed by using a set of about 3-5 unrelated molecules, such as powdered milk, transferrin, or similar molecules, rather than a single reference molecule.

[0130] Preferably, the antibody of the present invention is an isolated antibody. As used herein, the term “isolated antibody” means an antibody that substantially does not contain other antibodies having different antigen specificities (for example, an isolated antibody that specifically binds to PD-L1 and HER2 substantially does not contain antibodies that specifically bind to antigens other than PD-L1 and HER2; for example, an isolated antibody that specifically binds to PD-L1, HER2 and human serum albumin substantially does not contain antibodies that specifically bind to antigens other than PD-L1, HER2 and human serum albumin). Furthermore, an isolated antibody may substantially not contain other cellular substances and / or chemical substances.

[0131] Preferably, the antibody of the present invention is a monoclonal antibody. As used herein, the terms “monoclonal antibody” or “monoclonal antibody composition” refer to antibodies having substantially identical amino acid sequences or derived from the same gene source. Monoclonal antibody compositions exhibit binding specificity and affinity to a specific epitope, or binding specificity and affinity to a specific 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 its antigen-binding fragment) refers to an antibody molecule in which (a) the constant region, or a portion thereof, is modified, substituted, or exchanged so that the antigen-binding site (variable region) binds to a different or modified class, effector function, and / or species, or to a completely different molecule that confers new characteristics to the chimeric antibody, such as an enzyme, toxin, hormone, growth factor, drug, etc., or (b) the variable region, or a portion thereof, is modified, substituted, or exchanged with a variable region having a different or modified antigen specificity. For example, a mouse antibody can be modified by replacing its constant region with a constant region from human immunoglobulin. By substituting with a human constant region, the chimeric antibody can reduce its antigenicity in humans compared to the original mouse antibody while retaining its specificity for recognizing the antigen.

[0134] As used herein, the term “human antibody” (or its antigen-binding fragment) is intended to include antibodies (and their antigen-binding fragments) having a variable region in which both the framework and the CDR region are derived from human sequences. Furthermore, if a constant region is included, the constant region is also derived from such a human sequence, e.g., a human germline sequence, or a variant version of a human germline sequence. The human antibodies and their antigen-binding fragments of the present invention may include amino acid residues not encoded by human sequences (e.g., mutations introduced by random or site-directed mutagenesis in vitro or by somatic mutation in vivo). Humanized antibodies containing non-human antigen-binding residues are particularly excluded in this definition of human antibodies. Human antibodies can be prepared 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)). Furthermore, 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) and Boemer et al, J. Immunol, 147(l):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 the antigen to transgenic animals in which the endogenous locus has been deactivated, for example, by immunized xenomies (xenomices), which have been modified to produce such antibodies in response to antigen challenge (see, for example, U.S. Patents 6,075,181 and 6,150,584, with respect to XENOMOUSE® technology). For example, regarding human antibodies produced by human B-cell hybridoma technology, see also Li et al, Proc. Natl. Acad. Sci. USA, 103:3557-3562 (2006).

[0135] As used herein, a “humanized” antibody (or its antigen-binding fragment) is an antibody (or its antigen-binding fragment) that retains the reactivity of a non-human antibody while exhibiting low immunogenicity in humans. This can be achieved, for example, by retaining the non-human CDR region and replacing the rest of the antibody with the corresponding human portion (i.e., the framework portion of the constant and variable regions). Additional framework region modifications can be made within the human framework sequence and within the CDR sequence derived from the germline of other mammalian species. The humanized antibodies of the present invention may contain amino acid residues not encoded by human sequences (e.g., mutations introduced by random or site-directed mutagenesis in vitro or somatic mutation in vivo, or conservative substitutions to enhance stability or production). See, for example, 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 technology include, but are not limited to, the Xoma technology disclosed in U.S. Patent No. 5,766,886.

[0136] As used herein, the term “recombinant humanized antibody” encompasses all human antibodies prepared, expressed, produced, or isolated by recombinant means, including antibodies isolated from host cells transformed to express humanized antibodies, such as transfectomas, and antibodies prepared, expressed, produced, or isolated by any other means, including splicing all or part of a human immunoglobulin gene or sequence with another DNA sequence.

[0137] Preferably, the antibody or antigen-binding fragment of the present invention is humanized. Preferably, the antibody or antigen-binding fragment of the present invention is humanized and contains rabbit-derived CDR.

[0138] As used herein, the term “multispecific antibody” refers to an antibody that binds to two or more different epitopes (e.g., PD-L1 and HER2) on at least two or more different targets. The term “multispecific antibody” includes bispecificity, triplicity, quadruplicity, quinticity, and sextaspecificity. As used herein, the term “bispecific antibody” refers to an antibody that binds to two different epitopes on at least two different targets (e.g., PD-L1 and HER2). As used herein, the term “trispecific antibody” 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 determinant of a protein that can specifically bind to an antibody. Epitopes typically consist of groups of surfaces on chemically active molecules, such as amino acid or sugar side chains, and usually possess distinctive three-dimensional structural features and specific charge properties. "Stereostructural" epitopes and "linear" epitopes are distinguished in that binding to the former is lost in the presence of a denaturing solvent, while binding to the latter is not.

[0140] As used herein, the term “structural epitope” refers to the amino acid residues of an antigen that come together on the surface when a polypeptide chain folds to form a native protein.

[0141] The term "linear epitope" refers to an epitope where all interaction points of molecules (such as antibodies) that interact with a protein occur linearly (continuously) along the primary amino acid sequence of the protein.

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

[0143] As used herein, the term "recognize" refers to antibodies and antigen-binding fragments that locate their three-dimensional epitopes and interact with (e.g., bind to) them.

[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's "arms" interact with the antigen at numerous sites through weak non-covalent bonds, and the more interactions there are, the stronger the affinity.

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

[0146] As used herein, the terms “Kassoc,” “Ka,” or “Kon” are intended to refer to the association rate of a particular antibody-antigen interaction, while 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). In one embodiment, “KD” or “KD value” or “KD” or “KD value” according to the present invention is measured by surface plasmon resonance assay. 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 molecular affinity for human serum albumin (HSA) and mouse serum albumin (MSA) was determined by SPR measurement as described in Section

[0179] .

[0147] Preferably, the multispecific antibody of the present invention is monovalent, divalent, or polyvalent with respect to PD-L1 specificity. In one embodiment, the multispecific antibody of the present invention is divalent with respect to PD-L1 specificity. In a preferred embodiment, the multispecific antibody of the present invention is monovalent with respect to PD-L1 specificity.

[0148] A preferred PD-L1-BD for use in the multispecific antibodies of the present invention is the binding domain provided in this disclosure. Examples of PD-L1-BDs of the present invention include, but are not limited to, the humanized monoclonal antibodies whose sequences are listed in Table 1.

[0149] Preferably, the multispecific antibody of the present invention is monovalent, divalent, or polyvalent with respect to HER2 specificity. In one embodiment, the multispecific antibody of the present invention is divalent with respect to HER2 specificity. In a preferred embodiment, the multispecific antibody of the present invention is monovalent with respect to HER2 specificity.

[0150] A suitable HER2-BD for use in the multispecific antibodies of the present invention is the binding domain provided herein. Examples of HER2-BDs of the present invention include, but are not limited to, the humanized monoclonal antibodies whose sequences are listed in Table 2.

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

[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. Furthermore, as used herein, the terms "binding domain" or "monovalent binding domain" refer 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 this invention have now, surprisingly, found that the addition of the triplicate molecule PRO1678 (anti-HSA × PDL1 × HER2) resulted in a significant reduction in tumor growth in the HCC1954 xenograft NOG mouse model compared to an equivalent dose of nivolumab (the same activity determined in vitro). A five-fold lower dose of PRO1678 produced the same reduction in tumor growth in this model (see Figure 13). Even more surprisingly, the inventors found that a fourth CD3-BD-containing quadruspecific molecule, such as PRO1543 (anti-CD3 × HSA × PDL1 × HER2), resulted in complete tumor regression in the HCC1954 xenograft NOG mouse model. This finding is surprising because, in the complex, multi-target, multicellular in vivo environment, it is not expected from previous experience that all four binding domains would maintain their function without sterically or otherwise inhibiting each other. EC of PD-L1 blockade in these cells 50 This is the EC of target cell lysis. 50 Since this is considerably higher, the increased activity of molecules containing the anti-PD-L1 domain is very likely to result from increased avidity. As a result, the binding of Her2 and PD-L1 to bipositive cells is stronger than the binding to Her2-expressing PD-L1-negative cells. This avidity binding selectively enhances efficacy against tumor cells (Her2 / PD-L1 bipositive) but not against PD-L1-negative normal Her2-expressing cells, thus broadening the therapeutic range.

[0155] The term "tumor-associated immune checkpoint antigen" refers to transmembrane proteins expressed by tumors that suppress the activity of immune cells, and is specifically selected from the group of PD-L1, PD-L2, CD80, CD86, CD276 (B7-H3), and VTCN1 (B7-H4), and more specifically to antibodies that are PD-L1.

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

[0157] The term "tumor-associated antigen (TAA)" refers to an antigen expressed on the surface of tumor cells. In certain embodiments, TAAs are antigens that are preferentially expressed on tumor cells compared to non-tumor cells, and in particular, the expression of TAAs on tumor cells is at least 5 times, at least 10 times, at least 20 times, at least 50 times, or at least 100 times higher than on non-tumor cells from the same organism or patient. Specifically, TAAs are selected from the following groups: EGFRvIII, Mesoserine, GD2, Tn antigen, sTn antigen, Tn-O-glycopeptide, sTn-O-glycopeptide, PSMA, CD97, TAG72, CD44v6, CEA, EPCAM, KIT, IL-13Ra2, Regman, 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, Regmaine, HPV E6 or E7, ML-IAP, CLDN6, TSHR, GPRC5D, ALK, polysialic acid, Fos-related antigen, neutrophil elastase, TRP-2, CYP1B1, sperm protein 17, β-human chorionic gonadotropin, AFP, thyroglobulin, PLAC1, GloboH, 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 antigens present on immune cells, particularly selected immune cells from T cells, NK cells, and myeloid cells. In particular, this term relates to proteins that are stimulating or co-stimulating molecules of the aforementioned immune cells.

[0159] In the context of this invention, the term "stimulating molecule of the immune cells" refers to molecules such as CD3 and CD16.

[0160] In the context of this invention, the term "co-stimulatory molecule" refers to molecules such as those included 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 present invention further comprises (i) a binding domain to CD3, or (ii) a binding domain to CD137.

[0162] A preferred CD3-BD for use in the multispecific antibodies of the present invention is the CD3-BD binding domain provided herein. Examples of CD3-BDs of the present invention include, but are not limited to, the humanized monoclonal antibodies whose sequences are listed in Table 3.

[0163] A preferred CD137-BD for use in the multispecific antibodies of the present invention is the CD137-BD binding domain provided herein. Examples of CD137-BDs of the present invention include, but are not limited to, the humanized monoclonal antibodies whose sequences are listed in Table 5.

[0164] Preferably, the multispecific antibody of the present invention has two different specificities (PD-L1 and HER2). Preferably, the multispecific antibody of the present invention is a bispecific antibody. The multispecific antibody of the present invention may include further specificities (triple specificity) or multiple specificities (quadrispecific, quintuple specificity, or hexaspecificity antibody). In one embodiment, the multispecific antibody is triplicate. In another embodiment, the multispecific antibody is quadruplespecific.

[0165] In one embodiment, the multispecific antibody of the present invention comprises an immunoglobulin Fc region polypeptide. Hereinafter, 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. Preferred native sequence Fc regions include human IgG1, IgG2 (IgG2A, IgG2B), IgG3, and IgG4. “Fc receptor” or “FcR” describes a receptor that binds to the Fc region of an antibody. A preferred FcR is the native sequence of a human FcR. Furthermore, preferred FcRs bind to IgG antibodies (gamma receptors) and include receptors of the FcγRI, FcγRII, and FcγRIII subclasses, including allelic variants and alternatively spliced ​​forms of these receptors. The FcγRII receptor includes FcγRIIA ("activating receptor") and FcγRIIB ("inhibiting receptor"), which have different but similar amino acid sequences, primarily in their cytoplasmic domains. The activating receptor FcγRIIA contains an immunoreceptor tyrosine-based activating motif (ITAM) in its cytoplasmic domain. The 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 outlined 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 to be identified in the future, are included in the term “FcR” herein. The term “Fc receptor” or “FcR” also includes FcRn, the neonatal receptor 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, for example, 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), and International Publication No. 2004 / 92219 (Hinton et al).). The in vivo binding and serum half-life of human FcRn high-affinity binding polypeptides to FcRn can be assayed, for example, in transgenic mice expressing human FcRn, transfected human cell lines, or primates administered with polypeptides containing a variant Fc region. International Publication No. 2004 / 42072 (Presta) describes antibody variants that enhance or reduce binding to FcR. See also, for example, Shields et al., J. Biol. Chem. 9(2): 6591-6604 (2001).

[0166] In another embodiment, the antibody of the present invention does not contain an immunoglobulin Fc region polypeptide.

[0167] To increase the number of specificities / functionalities at the same or low molecular weight, it is advantageous to use antibodies containing 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. While such fragments appear to offer many advantages over whole immunoglobulin, they also have the problem of increased clearance from serum due to the absence of the Fc domain that confers a long half-life in vivo (Medasan et al., 1997, J. Immunol. 158:2211-2217). Molecules with smaller molecular weights penetrate target tissues more efficiently, and therefore, improved efficacy can be expected at the same or lower doses.

[0168] The inventors of this invention have surprisingly found that adding a human serum albumin-binding domain (HSA-BD) to the multispecific antibody of this invention does not impair the ability of the other binding domains to their respective targets. This finding is surprising because, in a complex, multi-target, multicellular in vitro environment, it is not predictable from previous experience that all four binding domains would maintain their function without sterically or otherwise inhibiting each other.

[0169] Preferably, the multispecific antibody of the present invention may include further binding domains 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. Preferably, 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, abundant in human serum (50% of total protein) (Sugio, Protein Eng, Vol. 12, 1999, 439-446). The multifunctional nature of the HSA protein is due to its structure, which allows it to bind and transport many metabolites, including fatty acids, metal ions, bilirubin, and certain drugs (Fanali, Molecular Aspects of Medicine, Vol. 33, 2012, 209-290). Serum HSA concentrations are approximately 3.5–5 g / dL. Albumin-binding antibodies and their fragments can be used, for example, to extend the in vivo serum half-life of drugs or proteins bound to them.

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

[0172] A suitable HSA-BD for use in the multispecific antibodies of the present invention is the binding domain provided herein. Examples of HSA-BDs of the present 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 contains a VH CDR having one of the amino acid sequences of the VH CDRs listed in Table 4. In particular, the present invention provides an HSA-BD containing one, two, three or more VH CDRs having one of the amino acid sequences of the VH CDRs listed in Table 4.

[0174] The present invention also provides HSA-BDs comprising VL CDRs having one of the amino acid sequences 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 one of the amino acid sequences of the VL CDRs listed in Table 4.

[0175] In further embodiments, the present invention provides HSA-BD that specifically binds to human serum albumin, wherein the binding domain comprises a VH domain and a VL main.

[0176] Other suitable HSA-BDs for use in the multispecific antibodies of the present invention are (i) polypeptides that bind to serum albumin (see, for example, Smith et al., 2001, Bioconjugate Chem. 12:750-756, European Patent No. 0486525, U.S. Patent No. 6267964, International Publication No. 2004 / 001064, International Publication No. 2002 / 076489, and International Publication No. 2001 / 45746), (ii) Holt et al., Protein Engineering, Design & Selection, vol 21, 5, (iii) an antibody selected from the group consisting of an antiserum albumin-binding monovariable domain described in pp. 283-288, International Publication No. 2004 / 003019, International Publication No. 2008 / 096158, International Publication No. 2005 / 118642, International Publication No. 2006 / 0591056 and International Publication No. 2011 / 006915, and (iii) an antiserum albumin antibody described in International Publication No. 2009 / 040562, International Publication No. 2010 / 035012 and International Publication No. 2011 / 086091.

[0177] In certain embodiments, the multispecific antibody of the present invention comprises an HSA-binding domain having the CDR sequence described in SEQ ID NOs. 61-66 and the VH / VL sequence described in SEQ ID NOs. 67-70.

[0178] These HSA-BDs exhibit particularly advantageous properties, such as high stability and cross-reactivity with cynomolgus monkey serum albumin (CSA) and mouse serum albumin (MSA), further enhancing the already advantageous properties of the multispecific antibodies of the present invention. More specifically, the HSA-BDs are characterized by one or more of the following parameters. a. By measuring with surface plasmon resonance (SPR), the dissociation constant (K) of monovalent plasmons less than 20 nM was found at a pH of approximately 5.5. D ), especially K at 0.01~20nM, especially 0.05~10nM, and especially 0.1~5nM DIt binds to human serum albumin (HSA), and in particular, the hSA-BD is in scFv format; b. By measuring with surface plasmon resonance (SPR), the monovalent dissociation constant (K) of less than 20 nM was found at both pH values ​​of approximately 5.5 and approximately 7.4. D ), especially K at 0.01~20nM, especially 0.05~10nM, and especially 0.1~5nM D It binds to human serum albumin (HSA), and in particular, the HSA-BD is in scFv format; c. Macaca fascicularis (cynomolgus monkey) serum albumin (CSA) is cross-reactive, particularly when measured by SPR, at pH values ​​of approximately 5.5 or 7.4, especially monovalent K15 nM. D , especially K 0.01~15nM, especially 0.05~7nM, especially 0.1~4nM D It is then combined with the CSA, and in particular the HSA-BD is in scFv format; d. Mus musculus (mouse) cross-reactive with serum albumin (MSA), particularly when measured by SPR at a pH of approximately 5.5, especially monovalent K20 nM. D In particular, K at 0.01-20 nM, 0.05-10 nM, and 0.1-5 nM D It is then combined with the MSA, and in particular the HSA-BD is in scFv format; e. HSA bound to the antibody retains its ability to bind to FcRn; If in f.scFv format, the melting temperature (Tm) is measured by differential scanning fluorescence (DSF) and is at least 72°C, preferably at least 75°C, and more preferably at least 78°C, and in particular the HSA-BD is formulated in 50 mM citrate phosphate buffer containing 150 mM NaCl at pH 6.4; If in g.scFv format, after storage at 4°C for at least two weeks, the monomer content decreases by less than 5%, for example, less than 4%, less than 3%, less than 2%, preferably less than 1%, and the antigen-binding fragment is at an initial concentration of 50 mg / ml, and in particular, the HSA-BD is formulated in 50 mM citrate phosphate buffer containing 150 mM NaCl at pH 6.4; If in h.scFv format, after storage at 40°C for at least two weeks, the monomer content decreases by less than 11%, for example, less than 8%, less than 5%, less than 2%, preferably less than 1%, and the antigen-binding fragment is at an initial concentration of 10 mg / ml, and in particular, the HSA-BD is formulated in 50 mM citrate phosphate buffer containing 150 mM NaCl at pH 6.4; and / or If in i.scFv format, after storage at 40°C for at least 4 weeks, the decrease in protein content is less than 5%, for example, less than 4%, less than 3%, less than 2%, preferably less than 1%, and the antigen-binding fragment is at an initial concentration of 10 mg / ml, and in particular, the HSA-BD is formulated in 50 mM citrate phosphate buffer containing 150 mM NaCl at pH 6.4.

[0179] Other variable domains of the present invention include amino acids that have been mutated but have at least 60, 70, 80, 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identity in the CDR region with the CDR region shown in the sequences listed in Tables 1-5. Other variable domains of the present invention include mutant amino acid sequences in which 1, 2, 3, 4, or 5 or fewer amino acids are mutated in the CDR region compared to the CDR region shown in the sequences listed in Tables 1-5.

[0180] Preferably, the VH domain of the binding domain of the present invention belongs to the VH3 or VH4 family. In one embodiment, the binding domain of the present invention includes 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-FR2 exhibit the highest degree of homology to the VHx family (or VLx, respectively). Examples of the VH and VL families are described in Knappik et al., J. Mol. Biol. 296 (2000) 57-86, or in brochure International Publication No. 2019 / 057787. A specific example of a VH domain belonging to the VH3 family is shown in Sequence ID No. 142, and a specific example of a VH domain belonging to the VH4 family is shown in Sequence ID No. 143. In particular, the framework regions FR1-FR4 taken from Sequence ID No. 142 belong to the VH3 family (Table 7, regions marked in non-bold). Preferably, as used herein, a VH belonging to the VH3 family is a VH containing FR1-FR4 having at least 85%, preferably at least 90%, more preferably at least 95% sequence identity with FR1-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 in International Publication No. 2019 / 057787. Preferably, the HSA-BD of the present invention includes: Vκ framework FR1, FR2 and FR3, particularly Vκ1 or Vκ3 framework, preferably Vκ1 framework FR1-3, and Vκ FR4, particularly framework FR4 selected from Vκ1 FR4, Vκ3 FR4, and Vλ FR4. Preferred Vκ1 framework FR1-3 are described in SEQ ID NO: 144 (Table 7, FR region shown in non-bold). Examples of alternative Vκ1 sequences and examples of Vκ2, Vκ3, or Vκ4 sequences can be found in Knappik et al., J. Mol. Biol. 296 (2000) 57-86.Preferred Vκ1 frameworks FR1-FR3 correspond to FR1-FR3 and include amino acid sequences having at least 60, 70, 80, and 90 percent identity with the amino acid sequence taken from SEQ ID NO: 144 (Table 7, FR regions are shown in non-bold). Preferred Vλ FR4 is as described in SEQ ID NOs: 145-152. In one embodiment, the VL domain of the present invention includes a Vλ FR4 that includes an amino acid sequence having at least 60, 70, 80, and 90 percent identity with an amino acid sequence selected from any of SEQ ID NOs: 145-152, preferably SEQ ID NOs: 146 or 152.

[0181] The binding domains of the present invention include VH domains listed in Tables 1 to 5. Preferably, the binding domains of the present invention include a VH amino acid sequence listed in one of Tables 1 to 5, in which approximately 10 or fewer amino acids of the framework sequence (e.g., a sequence other than the CDR) are mutated (mutations are, in various non-limiting examples, additions, substitutions, or deletions). Preferably, the binding domains of the present invention include a VH amino acid sequence listed in one of Tables 1 to 5, in which approximately 20 or fewer amino acids of the framework sequence (e.g., a sequence other than the CDR) are mutated (mutations are, in various non-limiting examples, additions, substitutions, or deletions). Other binding domains of the present invention include an amino acid sequence in which mutations are introduced, but which has at least 60, 70, 80, 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identity in the VH region with the VH region shown in the corresponding sequence listed in one of Tables 1 to 5.

[0182] In particular, the binding domain of the present invention includes a VL domain listed in one of Tables 1 to 5. Preferably, the binding domain of the present invention includes a VL amino acid sequence listed in one of Tables 1 to 5, in which approximately 10 or fewer amino acids of the framework sequence (e.g., a sequence that is not a CDR) are mutated (mutations are, in various non-limiting examples, additions, substitutions, or deletions). Preferably, the binding domain of the present invention includes a VL amino acid sequence listed in one of Tables 1 to 5, in which approximately 20 or fewer amino acids of the framework sequence (e.g., a sequence that is not a CDR) are mutated (mutations are, in various non-limiting examples, additions, substitutions, or deletions). Other binding domains of the present invention include an amino acid sequence in which mutations are introduced, but which has at least 60, 70, 80, 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99 percent identity in the VL region with the VL region shown in the sequences listed in Tables 1 to 5.

[0183] In the context of the present invention, the term “binding domain of the present invention” refers to both such binding domains, i.e., those unrelated to the context of multispecificity, and in particular to one of the binding domains contained in a multispecificity construct, e.g., one of the binding domains contained in a bispecificity, triplicity, or quadruplicity construct.

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

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

[0186] The multispecific antibody of the present invention can be in any preferred format.

[0187] Preferably, the binding domains of the multispecific antibody are operably linked. The binding domains of the multispecific antibody of the present invention can simultaneously bind to each antigen or receptor.

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

[0189] As used herein, the term “operably linked” indicates that two molecules (e.g., polypeptides, domains, binding domains) are linked in such a way that they each retain functional activity. Two molecules can be “operably linked” whether they are linked directly or indirectly (e.g., via a linker, via a molar, or via a linker to a molar). The term “linker” refers to a peptide or other molar optionally placed between the binding domain or antibody fragment of the present invention. Many methods can be used to covalently link molecules together. These include, but are not limited to, polypeptide linkages between the N-terminus and C-terminus of a protein or protein domain, linkages via disulfide bonds, and linkages via chemical crosslinking reagents. In one aspect of this embodiment, the linker is a peptide bond and is produced by recombinant technology or peptide synthesis. The selection of a linker suitable for a particular case in which two polypeptide chains are linked depends on a variety of parameters, such as, but are not limited to, the properties of the two polypeptide chains (e.g., whether they are naturally oligomerized), the distance between the N-terminus and C-terminus to be linked, if known, and the stability of the linker against proteolysis and oxidation. Furthermore, the linker may contain amino acid residues that provide flexibility.

[0190] In the context of this 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 attached to one end of the linker. The polypeptide linker should be long enough to link two molecules so that they assume the correct three-dimensional structure relative to each other and consequently maintain 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 to be included in the polypeptide linker should exhibit properties that do not significantly interfere with the activity of the polypeptide. Therefore, the linker peptide as a whole should not exhibit a charge that is unlikely to be compatible with the polypeptide activity, nor should it interfere with internal folding, nor should it form bindings or other interactions with one or more amino acid residues in the monomer that are likely to significantly inhibit the binding of the receptor monomer domain. In certain embodiments, the polypeptide linker is a non-structured polypeptide. Useful linkers include glycine-serine, or GS linkers. "Gly-Ser" or "GS" linkers mean polymers of glycine and serine in series (e.g., (Gly-Ser)n, (GSGGS)n, (GGGGS)n, and (GGGS)n, where n is at least an integer of 1), glycine-alanine polymers, alanine-serine polymers, and flexible linkers such as tethers for shaker potassium channels, as well as many other flexible linkers, as will be understood by those skilled in the art. Glycine-serine polymers are preferred because both of these amino acids are relatively unstructured and therefore can potentially function as a neutral tether between the components. Secondly, serine is hydrophilic and therefore can solubilize what may be spherical glycine chains. Thirdly, similar chains have been shown to be effective in binding subunits of recombinant proteins, such as single-chain antibodies.

[0191] Preferably, the multispecific antibody is in a format selected from any suitable multispecificity known in the art, for example, a bispecificity format, for example, a non-limiting example, a format based on a single-chain diabody (scDb), tandem scDb (Tandab), linear dimer scDb (LD-scDb), cyclic dimer scDb (CD-scDb), bispecificity T cell inducing antibody (BiTE; tandem di-scFv), tandem tri-scFv, tribody (Fab-(scFv)2) or vibody (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 chain) IgG-scFv fusions such as dsscFv) linked to the C-terminus of the chain; bispecific antibodies based on heterodimer 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 (International Publication No. 2016 / 0202457, Egan) fused to the N- and / or C-terminus of any heterodimer Fc domain or any other heterodimerizing domain. Examples include T., et al., mAbs 9 (2017) 68-84), and duobody (bispecific IgG prepared by duobody technology) (MAbs. 2017 Feb / Mar;9(2):182-212. doi: 10.1080 / 19420862.2016.1268307).The single-stranded diabody (scDb) or scDb-scFv is particularly suitable for use in this specification.

[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 is the single-stranded diabody (scDb), in particular the bispecific monomer scDb. Also particularly suitable for use herein is scDb-scFv, in particular the form of scDb, where the CD137-BD and the PD-L1-BD are in the form of scDb, and the HSA-BD is operably linked to the scDb.

[0193] The term "diabody" refers to an antibody fragment having two antigen-binding sites, specifically a fragment containing VH linked to VL within the same polypeptide chain (VH-VL). By using a linker that is too short for two domains on the same chain to pair, the domains are forced to pair with complementary domains on another chain, thus creating two antigen-binding sites. Diabodies can be bivalent or bispecific. Diabodies are described more extensively, for example, in European Patent No. 404097, International Publication No. 93 / 01161, Hudson et al., Nat. Med. 9:129-134 (2003), and Hollinger et al., Proc. Natl. Acad. Sci. USA 90: 6444-6448 (1993). Furthermore, triabodies and tetrabodies are described in Hudson et al., Nat. Med. 9:129-134 (2003).

[0194] Bispecific scDb, in particular bispecific monomers scDb, are formed by linkers L1, L2, and L3, resulting in 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 chain domains (VH) or fragments thereof linked in the order of 2-VLA-L3-VLB, VLB-L1-VLA-L2-VHA-L3-VHB, or VLB-L1-VHA-L2-VLA-L3-VHB, and two variable light chain domains (VL) or fragments thereof, 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 a peptide of 2 to 10 amino acids, more specifically 3 to 7 amino acids, most specifically 5 amino acids, and linker L3 is a peptide of 1 to 10 amino acids, more specifically 2 to 7 amino acids, most specifically 5 amino acids. In certain embodiments, linker L1 and / or L3 comprises one or two (GGGGS)n units of four (4) glycine amino acid residues and one (1) serine amino acid residue, where n=1 or 2, preferably n=1.

[0196] The intermediate linker L2 is a peptide of 10 to 40 amino acids, more specifically 15 to 30 amino acids, and most specifically 20 to 25 amino acids. In certain embodiments, the linker L2 comprises one or more (GGGGS)n units 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 present invention is scDb-scFv. The term "scDb-scFv" refers to an antibody format in which a single-stranded Fv (scFv) fragment is fused to a single-stranded diabody (scDb) by a flexible Gly-Ser linker. In one embodiment, the flexible Gly-Ser linker is a peptide of 2 to 40 amino acids, e.g., 2 to 35, 2 to 30, 2 to 25, 2 to 20, 2 to 15, 2 to 10 amino acids, particularly 10 amino acids. In a particular embodiment, the 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 present invention, the multispecific antibody of the present invention is in the MATCH format described in International Publication No. 2016 / 0202457, Egan T., et al., mAbs 9 (2017) 68-84.

[0199] The multispecific antibodies of the present invention can be prepared using any simple antibody production method known in the art (for example, for the preparation of bispecific constructs, see Fischer, N. & Leger, O., Pathobiology 74 (2007) 3-14; for bispecific diabodies and tandem scFv, see Hornig, N. & Faerber-Schwarz, A., Methods Mol. Biol. 907 (2012) 713-727, and International Publication No. 99 / 57150). Specific examples of methods suitable for the preparation of the bispecific constructs of the present invention include, in particular, the 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) techniques. Methods for producing bispecific antibodies containing a functional antibody Fc moiety are also known in the art (see, for example, 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 by, for example, using hybridoma technology to fuse mouse-derived myeloma cells immunized with a 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 (repertory cloning or phage display / yeast display) (see, e.g., Chames & Baty, FEMS Microbiol. Letters 189 (2000) 1-8), and by combining the antigen-binding domains or fragments or parts thereof of two or more different monoclonal antibodies using known molecular cloning techniques to obtain bispecific or multispecific constructs.

[0201] The multispecific molecules of the present invention can be prepared by binding 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 bound together. When the binding specificity is a protein or peptide, various coupling agents or crosslinking agents can be used for covalent bonding. Examples of crosslinking 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)cyclohaxane-l-carboxylate (sulfo-SMCC) (see, for example, Karpovsky et al., 1984 J. Exp. Med. 160: 1686 and Liu, MA et al., 1985 Proc. Natl. Acad. Sci. USA 82:8648). Other methods include those described in 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). The conjugating agents are SATA and sulfo-SMCC, both of which are available from Pierce Chemical Co. (Rockford, Illinois).

[0202] When the binding specificity substance is an antibody, it can be bound by sulfhydryl bonds in the C-terminal hinge regions of the two heavy chains. In certain embodiments, the hinge region is modified to include an odd number, for example, one sulfhydryl residue, before binding.

[0203] Alternatively, two or more binding specificities can be encoded in the same vector, expressed in the same host cell, and assembled. This method is particularly useful when the bispecific molecule is an mAb×mAb, mAb×Fab, Fab×F(ab')2, or ligand×Fab fusion protein. The multispecific antibody of the present invention may be a single-chain molecule containing one single-chain antibody and a binding determinant, or a single-chain multispecific antibody containing two binding determinants. The multispecific antibody may contain at least two single-chain molecules. Methods for preparing multispecific antibodies and molecules are described, for example, in U.S. Patent 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] The binding of multispecific antibodies to their specific targets can be confirmed, for example, by enzyme-linked immunosorbent assay (ELISA), radioimmunoassay (REA), FACS analysis, bioassay (e.g., growth inhibition), or Western blotting assay. Each of these assays generally detects the presence of a protein-antibody complex of particular interest by using a reagent (e.g., antibody) specifically labeled for that complex.

[0205] In a further embodiment, the present invention provides nucleic acids encoding the multispecific antibodies or fragments thereof or their binding domains. Such sequence nucleic acids can be optimized for expression in mammalian cells.

[0206] In this specification, the term “nucleic acid” is used interchangeably with the term “polynucleotide,” and refers to one or more deoxyribonucleotides or ribonucleotides and their polymers in either single-stranded or double-stranded forms. The term encompasses nucleic acids containing known nucleotide analogs or modified skeletal residues or linkages, which are synthetic, naturally occurring, and unnatural, possess similar binding properties to the reference nucleic acid, and are metabolized in a similar manner to the reference nucleic acid. Examples of such analogs, but not limited to, include phosphorothioates, phosphoramides, methylphosphonates, chiral methylphosphorates, 2-O-methylribonucleotides, and peptide nucleic acids (PNAs). Unless otherwise indicated, a particular sequence nucleic acid also implicitly includes its conservatively modified variants (e.g., degenerate codon substitutions) and complementary sequences, as well as the explicitly indicated sequence. Specifically, as will be described in detail later, degenerate codon substitution can be achieved by generating sequences in which the third position of one or more selected (or all) codons is replaced with a mixed base and / or a deoxyinosine residue (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 containing fragments or domains of the multispecific antibodies described above. When expressed from a suitable expression vector, the polypeptides encoded by these nucleic acid molecules can exhibit the antigen-binding ability or capacity(s) of the multispecific antibodies of the present invention.

[0208] Furthermore, the present invention also provides polynucleotides that encode at least one CDR region, and usually all three CDR regions, of the binding domains of the multispecific antibodies of the present invention described in Tables 1 to 4. Due to coding degeneracy, various nucleic acid sequences encode their respective immunoglobulin amino acid sequences.

[0209] Polynucleotide sequences can be prepared by de novo solid-phase DNA synthesis or by PCR mutagenesis of existing sequences encoding the multispecific antibodies of the present invention, or fragments thereof, or their binding domains (for example, sequences described in the examples below). 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 diethylphosphoramidite method of Beaucage et al., Tetra. Lett., 22:1859, 1981, and the solid support method of U.S. Patent No. 4,458,066. Introducing mutations into polynucleotide sequences by PCR is described in, for example, *DNA Amplification*, HA 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, fragments thereof, or binding domains thereof.

[0211] The term "vector" is intended to refer to a polynucleotide molecule that can transport another polynucleotide to which it is linked. One type of vector is a "plasmid," which refers to a circular double-stranded DNA loop to which further DNA fragments can be ligated. Another type of vector is a viral vector, to which further DNA fragments can be ligated into a viral genome. Certain types of vectors can autonomously replicate within the host cell into which they are introduced (e.g., bacterial vectors with bacterial origins of replication and episomal mammalian vectors). Other vectors (e.g., non-episomal mammalian vectors) can be incorporated into the host cell's genome upon introduction into the host cell, thereby replicating along with the host genome.

[0212] Furthermore, certain types of vectors can induce the expression of genes to which they are operably ligated. In this specification, such vectors are referred to as “recombinant expression vectors” (or simply “expression vectors”). Generally, expression vectors useful in recombinant DNA technology are often in the form of plasmids. In this specification, “plasmid” and “vector” can be used interchangeably, as plasmids are the most commonly used form of vectors. However, the present invention intends to include other forms of expression vectors, such as viral vectors (e.g., replication-deficient retroviruses, adenoviruses, adeno-associated viruses), that perform equivalent functions. In this particular context, the term “operably ligated” means a functional relationship between two or more polynucleotide (e.g., DNA) fragments. Typically, this refers to a functional relationship between a transcriptional regulatory sequence and a sequence being transcribed. For example, a promoter or enhancer sequence is operably ligated to a coding sequence if it stimulates or modulates the transcription of the coding sequence in a suitable host cell or other expression system. Generally, a promoter transcriptional regulatory sequence operably ligated to a sequence being transcribed is physically continuous with the sequence being transcribed, i.e., cis-acting. However, some transcriptional regulatory sequences, such as enhancers, do not need to be physically adjacent to or in close proximity to the coding sequence that promotes their transcription.

[0213] Various expression vectors can be employed to express polynucleotides encoding multispecific antibody chains or binding fragments. Both virus-based and non-viral expression vectors can be used to produce antibodies in mammalian host cells. Non-viral vectors and systems typically include episomal vectors with expression cassettes for plasmids, 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, California), MPS V vector, and numerous other vectors known in the art for expressing other proteins. Useful viral vectors include retrovirus, adenovirus, adeno-associated virus, herpesvirus-based vectors, SV40, papillomavirus, HBP Epstein-Barr virus-based vectors, vaccinia virus vectors, and Semryki Forest Virus (SFV). See Brent et al., ibid., 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 cells in which the vector is intended to be expressed. Typically, an expression vector contains a promoter and other regulatory sequences (e.g., enhancers) operably linked to a polynucleotide encoding a multispecific antibody chain or fragment. In one embodiment, an inducible promoter is employed to ensure that the insert sequence is not expressed except under inducible conditions. Examples of inducible promoters include arabinose, lacZ, metallothionein promoters, or heat shock promoters. Cultures of transformed organisms can be grown under non-inducible conditions without biasing the population towards sequences that the host cells are better able to tolerate in their expression products. In addition to promoters, other regulatory elements may also be required or desired for the efficient expression of multispecific antibody chains or fragments. Typically, these elements include the ATG start codon and adjacent ribosome-binding sites or other sequences. Furthermore, the efficiency of expression can be enhanced by including an enhancer suitable for the cell system being used (see, for example, Scharf et al., Results Probl. Cell Differ. 20: 125, 1994, and Bittner et al., Meth. Enzymol., 153:516, 1987). For example, SV40 enhancers or CMV enhancers can be used to increase expression in mammalian host cells.

[0215] The expression vector can also provide a secretion signal sequence position for forming a fusion protein having a polypeptide encoded by the inserted multispecific antibody or fragment thereof or its binding domain sequence of the present invention. More often, the inserted multispecific antibody or fragment thereof or its binding domain sequence is ligated to a signal sequence before being included in the vector. The vector used to accept the sequence encoding the binding domain of the multispecific antibody light chain and heavy chain variable domain sometimes also encodes a constant region or a portion thereof. Such a vector allows the variable region to be expressed as a fusion protein with the constant region, thereby enabling the production of an intact antibody and its antigen-binding fragment. Typically, such a constant region is human.

[0216] The term “recombinant host cell” (or simply “host cell”) refers to a cell into which a recombinant expression vector has been introduced. Such terminology should be understood as intended to refer not only to a specific target cell but also to the offspring of such cells. Because certain changes can occur in the progeny due to mutation or environmental influences, such offspring may not actually be identical to the parent cells, but they still fall within the scope of the term “host cell” as used herein.

[0217] The host cells for possessing and expressing the multispecific antibodies, fragments thereof, or binding domains thereof of the present invention may be prokaryotes or eukaryotes. Escherichia coli is one prokaryotic host useful for the cloning and expression of the polynucleotides of the present invention. Other suitable microbial hosts include rod-shaped bacteria such as Bacillus subtilis, and other Enterobacteriaceae such as Salmonella, Serratia, and various Pseudomonas species. Expression vectors can also be made in these prokaryotic hosts, typically containing expression regulatory sequences (e.g., origins of replication) adapted to the host cell. Furthermore, there are many well-known promoters, such as lactose promoters, tryptophan (trp) promoters, β-lactamase promoters, and phage-lambda-derived promoters. Typically, promoters control expression and optionally have operator sequences, as well as ribosome-binding site sequences and similar sequences for initiating and completing transcription and translation. Other microorganisms, such as yeast, can also be used for the expression of the CD137-binding polypeptide of the present invention. Furthermore, 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, fragments thereof, or their binding domains of the present invention. For example, these cells may be hybridoma cell lines expressing endogenous immunoglobulin genes, or mammalian cell lines possessing extracellular expression vectors. These include cells from normally lethal, normal, or abnormally immortal animals or humans. 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 cultures for polypeptide expression is generally discussed, for example, in Winnacker, FROM GENES TO CLONES, VCH Publishers, NY, NY, 1987. Expression vectors for mammalian host cells may include origins of replication, regulatory expression sequences such as promoters and 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 typically contain promoters derived from mammalian genes or mammalian viruses. Preferred promoters may be constitutive, cell type-specific, stage-specific, and / or regulatory or controllable. Useful promoters include, but are not limited to, the metallothionein promoter, the constitutive adenovirus major late promoter, the dexamethasone-inducible MMTV promoter, the SV40 promoter, the MRP polIII promoter, the constitutive MPS V promoter, the tetracycline-inducible CMV promoter (such as the human early CMV promoter), the constitutive CMV promoter, and promoter-enhancer combinations known in the art.

[0219] The method for introducing an expression vector containing the target polynucleotide sequence varies depending on the type of cell host. For example, calcium chloride transfection is widely used in prokaryotic cells, while calcium phosphate treatment or electroporation can be used in other cell hosts (see Sambrook, et al., cited above for the whole picture). 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 the herpesvirus structural protein VP22 (Elliot and O'Hare, Cell 88:223, 1997), drug-enhanced uptake of DNA, and extracellular transduction. To produce recombinant proteins in high yield over long periods, stable expression is often desired. For example, cell lines that stably express the multispecific antibody or its fragment or its binding domain of the present invention can be prepared using the expression vector of the present invention containing a viral replication origin or endogenous expression element and a select marker gene. Following the introduction of the vector, the cells may be cultured in concentrated medium for 1-2 days before being switched to selective medium. The purpose of the selective marker is to confer resistance to selection, and its presence allows cells that successfully express the introduced sequence to be cultured in selective medium. Resistant, stably transfected cells can be grown using tissue culture techniques appropriate to the cell type. Accordingly, the present invention provides a method for producing the antibody or antigen-binding fragment thereof, the method comprising the step of culturing host cells containing nucleic acids or vectors encoding the antibody or antigen-binding fragment thereof, thereby expressing the antibody or fragment thereof of the present disclosure.

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

[0221] In a further embodiment, the present invention relates to a pharmaceutical composition comprising the multispecific antibody of the present invention and a pharmaceutically acceptable carrier. The pharmaceutically acceptable carrier enhances or stabilizes the composition or facilitates the preparation of the composition. Examples of pharmaceutically acceptable carriers include physiologically compatible solvents, dispersions, coatings, antimicrobial and antifungal agents, isotonic and absorption retardants, and the like.

[0222] The pharmaceutical compositions of the present invention can be administered by various methods known in the art. The route and / or mode of administration will vary depending on the desired effect. Administration may be intravenous, intramuscular, intraperitoneal, or subcutaneous, or may be performed in close proximity to the target site. Pharmaceutically acceptable carriers should be suitable for intravenous, intramuscular, subcutaneous, parenteral, spinal, or cutaneous 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, may 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 customarily 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, JR Robinson, ed., Marcel Dekker, Inc., New York, 1978. The pharmaceutical compositions are preferably manufactured under GMP conditions. Typically, a therapeutically effective dose or effective dose of the multispecific antibody of the present invention is used in the pharmaceutical composition of the present invention. The multispecific antibody of the present invention is formulated into a pharmaceutically acceptable dosage form by conventional methods known to those skilled in the art. The administration regimen is adjusted to produce the optimal desired response (e.g., therapeutic response). For example, a single bolus dose may be administered, or it may be administered in several divided doses over time, or the dose may be proportionally increased or decreased as appropriate in the emergency situation of the treatment. For ease of administration and uniformity of dosage, it is particularly advantageous to formulate parenteral compositions in the form of dosage units. As used herein, the form of a dosage unit refers to a physically individual unit suitable for unit administration to the target of treatment, each unit containing a predetermined amount of the active compound calculated to exert the desired therapeutic effect with the necessary pharmaceutical carrier.

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

[0225] The multispecific antibodies of the present invention are typically administered in multiple doses. Single dose intervals can be weekly, monthly, or yearly. Irregular intervals are also possible, as can be adjusted by measuring the blood concentration of the multispecific antibodies in the patient. Alternatively, the multispecific antibodies can be administered as a sustained-release formulation, in which case a lower frequency of administration is required. Dosage and frequency of administration vary depending on the half-life of the antibody in the patient. Generally, humanized antibodies exhibit longer half-lives than chimeric and unhumanized antibodies. Dosage and frequency of administration vary depending on whether the administration is prophylactic or therapeutic. For prophylactic use, relatively low doses are administered over a long period at relatively infrequent intervals. Some patients continue treatment for life. For therapeutic use, relatively high doses at relatively short intervals are sometimes required until disease progression is reduced or halted, preferably until the patient shows partial or complete improvement in the symptoms of the disease. Afterward, the patient can be administered in a prophylactic regime.

[0226] In one embodiment, the present invention relates to a multispecific antibody or pharmaceutical composition of the present invention for use as a pharmaceutical. In a preferred embodiment, the present invention provides a multispecific antibody or pharmaceutical composition for use in the treatment of proliferative disorders, particularly cancer, in subjects requiring such treatment.

[0227] In another aspect, the present invention provides multispecific antibodies or pharmaceutical compositions for use in the manufacture of pharmaceuticals for the treatment of proliferative disorders, particularly cancer.

[0228] In another aspect, the present invention relates to the use of multispecific antibodies or pharmaceutical compositions for treating proliferative disorders, particularly cancer, in subjects requiring such treatment.

[0229] In a further embodiment, the present invention relates to the use of multispecific antibodies or pharmaceutical compositions in the manufacture of pharmaceuticals for the treatment of proliferative disorders, particularly cancer, in subjects where such use is required.

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

[0231] The term "subject" encompasses both humans and non-human animals. Non-human animals include all vertebrates, such as mammals and non-mammals, such as non-human primates, sheep, dogs, cattle, chickens, amphibians, and reptiles. Unless otherwise specified, the terms "patient" and "subject" are interchangeable herein.

[0232] In this specification, the terms “treatment,” “to treat,” “to treat,” “to be treated,” and similar terms 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 the disease, or delays the progression of the disease. As used herein, “treatment” encompasses the treatment of any disease in mammals, e.g., humans, and includes (a) suppressing the disease, i.e., preventing its onset, and (b) alleviating the disease, i.e., reducing the disease.

[0233] The term "therapeutic dose" or "effective dose" refers to the amount of a drug administered to a mammal or other subject to treat a disease that is sufficient to provide such treatment. The therapeutic dose will vary depending on the drug, the disease and its severity, and the age and weight of the subject being treated.

[0234] In one embodiment, the proliferative disorder is cancer. The term “cancer” refers to a disease characterized by the rapid and uncontrolled proliferation of abnormal cells. Cancer cells can metastasize locally or to other parts of the body through the bloodstream and lymphatic system. The terms “tumor” and “cancer” are used interchangeably herein, and for example, both terms encompass solid and liquid tumors, e.g., diffuse or circulating tumors. Where used herein, the terms “cancer” or “tumor” include precancerous and 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. 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, especially colon cancer or colorectal adenoma, tumors of the neck and head, endometrial cancer, Cowden syndrome, Lhermitt-Dukuro disease, Banayan-Zonana syndrome, benign prostatic hyperplasia, neoplasms, especially epithelial ones, 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 lymphoma, plasma cell myeloma, Hodgkin 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 or compositions of the present invention inhibit the growth of solid tumors, but also inhibit the growth of liquid tumors. In further embodiments, the proliferative disorder is a solid tumor. The term "solid tumor" means, in particular, breast cancer, uterine cancer, colon cancer, rectal cancer, prostate cancer, stomach cancer (especially gastric cancer), cervical cancer, lung cancer (e.g., non-small cell lung cancer and small cell lung cancer), and tumors of the head and neck. Furthermore, depending on the type of tumor and the specific combination used, a reduction in tumor volume may be achieved. The multispecific antibodies or compositions of the present invention are also suitable for preventing metastatic spread and the growth or development of micrometastases in subjects with cancer.

[0236] In one embodiment, the cancer is PD-L1-positive, and preferably, the cancer expresses a higher level of PD-L1 compared to healthy tissue, and in particular, the cancer expresses a level of PD-L1 (mRNA or protein, respectively) that is at least 2 times, at least 3 times, at least 4 times, at least 5 times, at least 6 times, at least 7 times, at least 8 times, at least 9 times, at least 10 times, at least 15 times, at least 20 times, at least 30 times, at least 40 times, at least 50 times, at least 60 times, at least 70 times, at least 80 times, at least 90 times, or at least 100 times higher than the PD-L1 expression (mRNA or protein, respectively) 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 present invention or a pharmaceutical composition of the present invention. The kit can include one or more other components including the following: instructions for use; other reagents, such as agents useful for labeling, therapeutic agents, or antibodies to chelate or otherwise bind to the labeling or therapeutic agent, or radiation protection compositions; devices or other materials for preparing antibody molecules for administration; pharmaceutically acceptable carriers; and devices or other materials for administration to a subject. In a specific embodiment, the kit comprises a multispecific antibody of the present invention in a pharmaceutically effective amount. In a further embodiment, the kit comprises a multispecific antibody of the present invention in a pharmaceutically effective amount in a lyophilized form, a diluent, and optionally instructions for use. The kit may further include a filter needle for reconstitution and an injection needle.

[0238]

Table 1-1

Table 1-2

[0239]

Table 2-1

Table 2-2

[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] In the event of any inconsistency between the text of the specification (e.g., Tables 1-6) and the sequence listings throughout the entire text of this application, the text of the specification shall prevail.

[0246] For clarity, it is understood that certain features of the Invention described in the context of separate embodiments can also be provided in combination in a single embodiment. Conversely, for brevity, various features of the Invention described in the context of a single embodiment can also be provided separately or in any preferred subcombination. All combinations of embodiments relating to the Invention are specifically encompassed by the Invention and are disclosed herein as if each combination were individually and expressly disclosed. Furthermore, all subcombinations of various embodiments and their elements are also specifically encompassed by the Invention and are disclosed herein as if each such subcombination were individually and expressly disclosed.

[0247] The present invention is not limited in scope by the specific embodiments described herein. In fact, various modifications of the present invention, in addition to those described herein, will be 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 incorporated herein by reference.

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

[0250] Example 1: Preparation and testing of low-affinity anti-PD-L1 molecules: Project objectives The objective of this project is to identify low-affinity anti-PD-L1 antibody fragments that neutralize the interaction between PD-L1 and PD-1. Ultimately, this domain will be conjugated in a multispecific molecule with a high-affinity domain to a selected tumor-associated antigen (TAA) co-expressed with PD-L1 on tumor cells, enabling the targeted and neutralization of PD-L1 specifically in those cancer cells. Two groups of long-acting molecules corresponding to two projects were designed. The molecules in each group differ primarily 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, although one group additionally contains an anti-CD3ε-binding domain to induce T cell activation. This example describes the construction and characterization of the low-affinity domain and the multispecific molecules designed in these projects.

[0251] Design and fabrication of scFv To generate a low-affinity PD-L1 antibody that neutralizes the interaction between PD-L1 and PD-1, we introduced a single alanine substitution into the CDR region of the previously identified clone 33-03-G02 high-affinity neutralizing anti-PD-L1 domain. As a first step, we mutated each amino acid in the CDR3 region of the high-affinity domain (which has the greatest amino acid diversity) to alanine, yielding 21 mutants. Three single mutants showed a more than 100-fold decrease in affinity to PD-L1 compared to the original domain. Therefore, we combined the single mutants to generate two double mutants. In parallel, we designed nine more 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, presumably exposed residues, that only slightly reduced the affinity of the parent domain. Furthermore, three additional mutants containing up to three alanine substitutions of expected exposed residues were expressed. The data obtained for the five most interesting molecules with affinity reductions of 100-6,500-fold are shown below.

[0252] method: scFv creation Heterologous expression of the protein was induced in small quantities overnight in E. coli to produce insoluble inclusion bodies (as shown in Table 8 below). Inclusion complexes were isolated from the homogenized cell pellet by a centrifugation protocol that included several washing steps to remove cell residue and other host cell impurities. The purified inclusion bodies were dissolved in denaturation buffer, and the scFv were refolded using a scalable refolding protocol that produced milligrams of natively folded monomer scFv. At that point, the scFv was purified using a standardized protocol. The refolded product was captured by affinity chromatography to obtain purified scFv. Table 8 summarizes the preparation of the scFv molecule. Mammalian constructs were expressed in CHO-S cells using the CHOgro transient transfection kit (Mirus) (see Table 8). Five to seven days after expression (cell viability <70%), the culture was collected by centrifugation at 37°C. Proteins were purified from the clarified culture supernatant by protein L affinity chromatography, followed by a final purification step (polishing) by size exclusion chromatography as needed. Standard analytical methods such as SE-HPLC, UV280, and SDS-PAGE were used for quality control of the prepared materials.

[0253] Freeze-thaw stability test The suitability of the best-performing scFv molecules was evaluated in terms of freeze-thaw (F / T) cycles (colloidal stability). For F / T stability evaluation, the same analytical methods (SE-HPLC, UV absorbance at 280 nm) and parameters as those used in the storage stability test were applied to monitor molecular quality over multiple F / T cycles. Since a dedicated freeze-thaw test was not performed, freeze-thaw data was extracted from -80°C samples obtained from the storage stability test over 28 days (with a maximum of 7 days of storage between F / T cycles).

[0254] Differential scanning fluorescence (DSF) The midpoint of the thermal unfolding transition of the scFv construct was determined by differential scanning fluorescence (SMRI) quantification using the fluorescent dye SYPRO® Orange. The sample was prepared in 100 μl total volume in a 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 the prepared sample were added in triple strips to AB gene PCR plates with white walls. The assay was performed using a qPCR instrument as a thermal cycler, and fluorescence emission was detected using a custom dye calibration routine in the software. The PCR plates containing the test samples were subjected to heating from 25°C to 96°C in 1°C increments, with a 30-second pause after each heating step. The total assay time was approximately 2 hours. The second-order derivative method was used to determine the inflection point of the curve, and Tm was calculated using a software graphpad prism. The reported Tm is the average of three measurements.

[0255] [Table 8]

[0256] Affinity for PD-L1 by SPR method: Affinity for PD-L1 was measured by surface plasmon resonance (SPR) assay using a ViaCore T200 instrument (GE Healthcare). All measurements were performed at 25°C. In this experiment, Fc-tagged human PD-L1 extracellular domain (ECD, Sinobiological, catalog 10084-H02H) was captured using a GE Healthcare human antibody capture kit (catalog BR-1008-39). After each analyte injection cycle, anti-human Fc-specific IgG was regenerated to capture new antigen. For affinity measurement, a dose-response multicycle kinetic assay was used, injecting scFv as the analyte at analyte concentrations ranging from 6.86 to 5000 nM (3-fold serial dilution) diluted in running buffer (10 mM HEPES, 150 mM NaCl, and 0.05% Tween 20, pH 7.4). Association time and dissociation time were set to 300 seconds and 720 seconds, respectively. Apparent dissociation (k d ) and meetings (k a ) Rate constant and apparent dissociation equilibrium constant (K D The curve fitting quality was monitored using Chi2 and U values, which are measures of curve fitting quality. In addition to kinetic measurements, the curve fitting was performed by fitting a plot of the equilibrium response to the concentration. D The value obtained was (steady-state affinity measurement).

[0257] result: As shown in Table 9, binding to human PD-L1 was confirmed for the tested humanized scFv. The affinity for PD-L1 was 6,500 times lower for PRO1434 compared to the parent scFv PRO830.

[0258] [Table 9]

[0259] Binding of PD-L1 expressing cells by FC method: CHO-K1 (control cells that do not express PD-L1) and CHO-PD-L1 (Amsbio) cells with high PD-L1 expression levels were collected, and the cell count was measured. The cell suspension was centrifuged at 400×g for 5 minutes, and 100 μl (10,000 cells) of the cell suspension diluted with PBS-EB (1×DPBS, 2% BCS HI, 2 mM EDTA) was added to the designated wells of an unbound plate. After three washing steps with PBS-EB, the cells were centrifuged, and the washing buffer was aspirated. Subsequently, 100 μl of serial dilutions of the test sample and the positive control were added directly to the plate. The positive control sample (PRO830, 33-03-G02) was diluted in PBS-EB to concentrations ranging from 3500 to 0.22 ng / ml, and the sample dilutions ranged from 1000 to 0.064 μg / ml. After incubation at 4°C for 1 hour, the plate was washed three times with 100 μl of PBS-EB. The cell pellet was resuspended in 100 μl of Protein L-APC at a concentration of 2 μg / ml and incubated at 4°C for 1 hour. Next, the cells were washed three times again with 100 μl of PBS-EB. The cell pellet was resuspended in 50 μl of PBS-EB and analyzed using a NovoCyte 2060 flow cytometer. For each sample, the fluorescence intensity of APC was recorded for 5,000 events, and the geometric mean of the fluorescence intensity MFI was calculated. Data were corrected for nonspecific antibody binding (blank and CHO-K1 cell binding).

[0260] result: The efficacy of binding to PD-L1-expressing cells was evaluated using flow cytometry as described above. Serial dilutions of each molecule under test and reference PRO830 were added to the plate. Individual ICs from each plate were evaluated. 50 The IC values ​​were obtained from the IC of the reference molecule PRO830 (high affinity PD-L1 domain) that was incorporated together on each plate. 50 Calibrated against (relative EC) 50 :EC 50 PRO830 / IC 50(Test scFv). The efficacy is summarized in Table 10. PRO1434 and PRO1494 were shown to have the weakest binding. The dose-response curves obtained for PRO1434 and PRO1494 are shown in Figure 1.

[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 interactions when both interacting molecules are expressed on the cell surface was tested using CHO / PD-L1 / TCR (T cell receptor) activator cells and Jurcutt / PD-1 cells. In this assay, CHO cells stably expressing PD-L1 and TCR activators were incubated with Jurcutt T cells stably expressing firefly luciferase under the control of an NFAT response element as a reporter gene to monitor PD-1 and T cell activation. When TCR activators on CHO cells bind to Jurcutt T cells, TCR signaling induces NFAT-induced firefly luciferase expression. However, the interaction between PD-L1 and PD-1 negatively regulates such TCR signaling and therefore reduces firefly luciferase expression. 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) was added to the inner wells of a white cell culture plate and incubated at 37°C and 5% CO2 for 16–20 hours. The following day, the cell culture medium was removed from each well, and 50 μl of 2-fold concentrated serial dilutions of each molecule of interest (final concentration 162–0.025 μg / ml) and the reference molecule avelumab (final concentration 3,000–0.46 ng / ml) were added. Then, 50 μl of effector jarcut cells diluted to 400,000 cells / ml in assay buffer (RPMI1640 containing 10% FCS) was added to each well, and the plate was incubated at 37°C and 5% CO2 for 6 hours. Then, 50 μL of luciferase substrate (BPS Biosciences), prepared according to the manufacturer's protocol, was added to each well, the plate was incubated in the dark for 30 minutes, and luminescence was measured using a Flexstation III multimode microplate reader.

[0264] result: The efficacy of neutralizing the binding of PD-L1 to PD-1 was evaluated using a cell-based reporter gene assay as described above. Serial dilutions of each molecule under test and a reference avelumab were added to the plate. Individual ICs were obtained from each plate. 50 The values ​​were taken together on each plate by the reference molecule avelumab (high affinity PD-L1 domain IC). 50 Calibrated against (relative IC) 50 :I C 50 avelumab / IC 50 (Test scFv). The efficacy is summarized in Table 11. PRO1434 and PRO1494 were shown to have the lowest efficacy. The titration curves obtained for PRO1434 and PRO1494 are shown in Figure 2.

[0265] [Table 11]

[0266] Storage stability test method: Humanized scFv was formulated at a concentration of 10 mg / ml in aqueous buffer (final buffer, 50 mM NaCiP, 150 mM NaCl, pH 6.4) and subjected to stability tests, including a 4-week stability test where the product was stored for 4 weeks at -80°C, 4°C, and 40°C. At a minimum, the protein concentration was measured by UV absorbance at 280 nm, and the monomer and oligomer fractions in the formulation were evaluated by integrating the SE-HPLC peak area at 1 week, 2 weeks, and the end of each test. Parameters such as % monomer content, % monomer decrease, content, and % content decrease were recorded over time.

[0267] result: Tables 12 and 13 compare the endpoint measurements obtained at d0, d7, d14, and d28 of the tests at 4°C and 40°C. At 4°C, all molecules except PRO1075 and PRO1076 showed a decrease in monomer content of less than 10% over 28 days. At 40°C, only two molecules showed 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 a next-generation multispecific immuno-oncology program targeting HER2-expressing malignancies. HER2 is a clinically effective target in several cancer types with unmet medical needs, most notably breast and gastric cancer. However, while the spectrum of tumors (over)expressing HER2 is very broad, conventional antibodies such as trastuzumab are inaccessible for mechanistic reasons. This approach, however, is acceptable because it is designed to not only effectively mediate T-cell-induced lysis in HER2-expressing tumors but also to evade tumor immune evasion by simultaneously blocking immunosuppressive PD-L1 signaling. Locally restricting two additive, and possibly concurrent, mechanisms of action to tumor tissue is expected to give compounds using this approach a substantially extended efficacy profile, potentially enabling clinical efficacy even in HER2-expressing tumors that are primary or secondary refractory to standard anti-HER2 therapies. This approach should result in i) a more selective blockade of PD-1 / PD-L1 interactions that is at least as potent as avelumab / bavencio®. Specifically, the compounds developed using this approach should efficiently block the binding of PD-1 to PD-L1 in HER2 / PD-L1 co-expressing target cells, while their ability to block the PD-1 / PD-L1 interaction in cells that do not express Her2 should be far weaker than that of avelumab / bavencio®. ii) Compared to trastuzumab / herceptin®, avelumab / bavencio®, and combinations thereof, these should result in more selective lysis of HER2 / PD-L1 co-expressing cells by peripheral blood mononuclear cells with at least equivalent efficacy. More specifically, the compounds developed using this approach should potently lyse cells co-expressing both HER2 and PD-L1, while not lysing cells expressing only PD-L1.

[0269] Design and fabrication of tribodies, DVD-tribodies, and MATCH-4 molecules Figure 3 shows descriptions of the different formats (tribody, DVD-tribody, and MATCH-4) designed within this approach. Table 14 details the data for the construction of all molecules, and describes the domain composition of the different molecules constructed 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: Tribodies, DVD-tribodies, and the MATCH-4 construct were expressed in CHO-S cells using the CHOgro transient transfection kit (Mirus). Five to seven days after expression (cell viability <70%), the cultures were collected by centrifugation at 37°C. Proteins were purified from the clarified culture supernatant by protein L-affinity chromatography, followed by final purification by size exclusion chromatography as needed. Standard analytical methods such as SE-HPLC, UV280, and SDS-PAGE were used for quality control of the prepared materials.

[0274] Affinity of human PD-L1, IL-23R, Her2, CD3ε, and human and mouse serum albumin (SA) based on SPR method: The affinity for PD-L1 was measured using surface plasmon resonance (SPR) measurements with a via-core T200 instrument (GE Healthcare) as described above. Apparent dissociation (k d ) and meetings (k a ) Rate constant and apparent dissociation equilibrium constant (K D The K2 value was calculated using a 1:1 Langmuir-coupled model with viacore analysis software (BIAevaluation, GE Healthcare). The quality of the fitting was monitored based on Chi2 and U values. In addition to kinetic measurements, the K2 value was calculated by fitting a plot of the equilibrium response against concentration. D The value obtained was (steady-state affinity measurement).

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

[0276] The affinity of the molecule to human serum albumin (HSA) and mouse serum albumin (MSA) was measured by SPR measurement using a BiaCore T200 instrument (GE Healthcare). SA was directly coupled to a CM5 sensor chip (GE Healthcare) using amine coupling chemistry. After finding the optimal assay conditions by regeneration scouting and surface performance testing, the dose-response of the target molecule was investigated at concentrations ranging from 0.7 to 180 nM. The assay was performed in PBS-Tween buffer at pH 5.5. The association time and dissociation time were set to 180 seconds and 720 seconds, respectively. The obtained coupling curves were double-referenced (empty reference channel and zero analyte injection) and fitted using BiaEvaluation software (GE Healthcare) and a 1:1 Langmuir coupling model. The apparent dissociation equilibrium constant (K) was determined using the obtained kinetic parameters. 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 measurement of low-affinity domains by SPR is extremely difficult due to the large amount of protein that needs to be injected to cover the concentration range corresponding to the molecular affinity. 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 is observed, introducing artifacts into kinetic analysis, suggesting that steady-state analysis of SPR measurements may be more appropriate. Affinity measurement for human PD-L1 was valid for only one tribody, PRO1498. For the MATCH-4 molecule, valid measurements were obtained using steady-state analysis. The lowest affinity for PD-L1 was around 900 nM for PRO1544, PRO1545, and PRO1547. PRO1543 and PRO1546 showed similar affinity around 300 nM. MATCH-4 proteins (PRO1543, PRO1544, PRO1557, and PRO1558) containing lambda-capped trastuzumab G100C / G172C mutant anti-Her2 domains exhibited similar affinity (300-400 pM) to Her2, which is 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] Neutralization of PD-L1 / PD-1 by FC method: These assays were performed to evaluate the ability of a multispecific low-affinity PD-L1× / Her2 construct to neutralize the interaction between PD-1 and PD-L1 expressed on HCC1954 cells that 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 its ability to block the PD-1 / PD-L1 interaction should be far weaker than that of 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 avelumab, the reference IgG. Furthermore, HCC827 cells, which express PD-L1 at levels comparable to HCC1954 cells but lack significant Her2 expression, were used as an additional control.

[0281] HCC1954 and HCC827 cells were stimulated with 10 ng / ml human IFNγ for 24 hours to further induce PD-L1 expression. The following day, HCC827 and HCC1954 cells were detached, centrifuged at 200 g for 4 minutes, resuspended in PBS / 2% FCS / 2 mM EDTA (staining buffer), and seeded in 96-well PP microplates (50 μl / well). Dilution series of multispecific molecules and avelumab were prepared in 3-fold steps, starting at 20 μg / ml and 5 μg / ml, respectively, using staining buffer containing 500 ng / ml biotinylated PD1. Plates of HCC827 and HCC1954 were centrifuged at 200 g for 4 minutes, the dilution series were added to the cells (100 μl / well), and incubated at room temperature (RT) for 30 minutes. Next, the cells were washed 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 minutes. As the next step, the cells were washed again as described 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 were fitted using a sigmoid 4PL fit (graph pad prism). Individual ICs of each plate were measured. 50 The IC values ​​were captured together on each plate for the reference molecule avelumab. 50 Calibrated against (relative IC) 50 :I C 50 avelumab / IC 50 (Test molecule). Furthermore, relative IC was observed in high-Her2-expressing HCC1954 cells and low-Her2-expressing HCC827 cells. 50 The ratio of the values ​​was calculated.

[0282] result: Table 17 summarizes the efficacy obtained from constructs using this approach to neutralize PD-1 / PD-L1 interactions. In HCC1954 cells expressing high levels of Her2 and PD-L1, PRO1454, a tribody with a single alanine mutant PD-L1 domain (33-03-G02 G109A), showed similar efficacy to avelumab. However, in cells expressing only PD-L1 (HCC827), the efficacy of PRO1454 was 14 times lower. PRO1497, a tribody with a double alanine mutant 33-03-G02 Q108A / G109A, showed similar relative IC in Her2 / PD-L1 positive cells (HCC1954). 50 As indicated by the values, PRO1497 neutralized the PD-1 / PD-L1 interaction with almost the same efficiency as PRO1454 (PRO1454: 0.66, PRO1497: 0.38). However, in contrast to PRO1454, PRO1497 showed only very weak neutralizing efficacy against PD-L1 expressing cells (HCC827) (Figure 4). This result suggests that molecules such as PRO1497, which contain a PD-L1 domain with comparable affinity to the domain with both mutations (Q108A / G109A) compared to a domain with similar affinity to a single alanine mutation (G109A), have a potentially broader therapeutic range.

[0283] All MATCH-4 molecules containing the trastuzumab anti-Her2 domain and the anti-PD-L1 domain with both mutations (Q108A / G109A) exhibited similar efficacy 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 the line, DVD-tribody PRO1547, containing a trastuzumab anti-Her2 domain and an anti-PD-L1 domain with both mutations (Q108A / G109A), neutralized the PD-1 / PD-L1 interaction with similar potency to PRO1497 (relative IC50 value, PRO1547: 0.25).

[0284] [Table 17]

[0285] CD3 activation and PD-L1 / PD-1 interaction blockade by NFAT reporter gene assay. method: To evaluate the simultaneous effects of the molecule on CD3 crosslinking and PD-1 / PD-L1 blockade, T cell activation was tested using the NFAT (activated T cell nuclear factor) assay. Specifically, the molecule should efficiently induce CD3 activation and block PD-1 binding to PD-L1 in Her2 / PD-L1 co-expressing target cells (HCC1954), but its efficacy in CD3 activation and PD-1 / PD-L1 interaction blockade should be much weaker in cells that do not express Her2 (CHO-PD-L1). The Jarcut PD-1 NFAT reporter T cell line expresses the luciferase reporter gene under the control of an NFAT response element derived from the IL-2 promoter. The transcription factor NFAT is activated by crosslinking CD3ε and induces many genes involved in T cell activation. In this system, CD3ε crosslinking induces the expression of the luciferase reporter gene. Furthermore, the interaction between PD-L1 and PD-1 negatively regulates such CD3ε signaling and therefore reduces firefly luciferase expression. Thus, blocking the PD-L1 / PD-1 interaction in this system leads to increased 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 in 96-well culture plates at a density of 25,000 cells per well (50 μl per well). Serial dilutions of the target molecule were prepared in assay medium containing 50 mg / ml HSA, and 25 μl were added to the cells (final concentration in the range of 250 nM to 0.026 pM). PD-1-expressing Jarcut 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 adding a luciferase reagent and read using a luminescence reader 5 or 22 hours after the addition of Jarcut PD-1 NFAT reporter cells. Relative luminescence units (RLU) are shown. HER2×PD-L1 高KD The potency of PRO1497, which is ×CD3 scDb-scFv, was used as a reference for calculating the relative potency of half-life extension molecules.

[0286] result: The efficacy of simultaneously inducing CD3 activation and PD-L1 / PD-1 interaction blockade was evaluated using an NFAT reporter gene assay. The results are shown in Table 18. The assay was performed for each of the target molecules and for the reference scDb-scFv Her2×CD3×PD-L1. 低親和性 A serial dilution of (PRO1497) was added to the plate. 50 The IC50 value and maximum activation of the reference molecule PRO1497 in the presence of Her2 and PD-L1 expressing cells HCC1954. 50 and normalized for maximum activation (relative IC) 50 Or maximum activation: IC 50 Alternatively, maximum activation, PRO1497 / IC with HCC1954 50 (or maximum activation, test molecule). In the presence of PD-L1 / Her2-expressing cells (HCC1954), the tribodies PRO1454, PRO1497, and PRO1456 similarly mediated CD3 signaling in Jarcut cells. 50While activation occurred, the maximum activation was higher for PRO1454 and PRO1497, i.e., molecules with a low-affinity anti-PD-L1 domain, 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 within the immune synapse and simultaneously activate CD3 in the presence of cells co-expressing Her2 and PD-L1. PRO1455 showed 27-fold weaker activation compared to PRO1454, and PRO1498 showed an even greater 475-fold weaker activation compared to PRO1497, i.e., molecules that lacked an anti-Her2 domain but possessed a low-affinity anti-PD-L1 domain, 33-03-G02 G109A or 33-03-G02 Q108A / G109A, respectively. These results indicate that molecules containing a PD-L1 domain with comparable affinity to a domain with both mutations (Q108A / G109A) have a potentially broader therapeutic range compared to a domain with similar affinity to a single alanine mutation (G109A), and that both molecules exhibit similar efficacy to a molecule without a PD-L1 domain (PRO1456). Efficacy is summarized in Table 18. PRO1543 showed the best efficacy in Her2 / PD-L1 expressing cells (HCC1954) and was the lowest in the presence of PD-L1 / very low Her2 expressing cells (HCC827). Titration curves obtained for PRO1543 and control molecules PRO1546 and PRO1557 in the presence or absence of 1 μg / ml nivolumab and PRO1557 combined with the low-affinity PD-L1 domain PRO1434 are shown in Figure 7.

[0287] [Table 18]

[0288] Cytotoxic assay (T cell-driven target cell depletion) To evaluate the ability of the compounds in this approach to selectively induce T cells to co-express Her2 and PD-L1, a cytotoxicity assay was performed using a Her2 and PD-L1-positive cell line (HCC1954) in the presence of human PBMCs. When the compounds in this approach bind simultaneously to both targets, CD3ε on T cells is cross-linked, activating a signaling cascade that leads to T cell activation (CD69 upregulation, cytokine secretion) and release of cytotoxic granules, ultimately resulting in target cell toxicity. In contrast to the avelumab / bavencio® and trastuzumab / herceptin® combinations, the compounds in this approach should selectively lyse cells co-expressing PD-L1 and HER2, while lysing cells expressing only PD-L1 (PD-L1 transfectant CHO cells) should not be observed.

[0289] method: Blood cell fractionation: Peripheral blood mononuclear cells (PBMCs) were isolated from fresh blood of healthy volunteers using Lymphoprep (Stemcell Technologies) lymphocyte isolation medium 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 monkey PBMC isolation buffer (PBS, 5% FCS, 2 mM EDTA) and added to Leucosep tubes containing the recommended volume of Lymphoprep medium. The Leucosep tubes were centrifuged at room temperature for 30 minutes without braking at 800 g (human blood) or 2000 g (cynomolgus monkey blood). The cell layer containing the PBMCs was then collected and washed twice with human PBMC isolation buffer, and the erythrocytes were lysed at room temperature for 5 minutes using erythrocyte lysis buffer. Subsequently, the isolated human cells were washed once with their respective isolation buffers and once with assay medium (RPMI-1640, 10% FCS). After removing platelets, the isolated PBMCs were added to the assay medium at a rate of 3 × 10⁶ per ml. 6 The cells were resuspended at a density of individual living cells.

[0290] Flow cytometry-based in vitro cytotoxicity assay (FC assay), CD8+ T cell activation, and CD11c+, CD4+ T cell, and CD8+ T cell survival rates. Two cell lines were used as target cells: HCC1954 cells (high co-expression of HER2 and PD-L1) and HCC827 cells (low co-expression of HER2 and PD-L1) stimulated with 10 ng / ml IFNy for 24 hours to increase PD-L1 expression. CHO-PD-L1 (Her2-free, high PD-L1) and CHO-K1 cell lines were used as negative control cell lines. 5,000 viable target cells, pre-labeled with PKH67 and diluted in 75 μl of assay medium (RPMI-1640, 10% FCS), were added to a 96-well plate. The test protein, concentrated 6-fold, was diluted in assay medium, and 25 μl 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), the plate was mixed at room temperature on an orbital mixer (nutating mixer), and then incubated at 37°C with 5% CO2. After 16 or 40 hours, the cells were trypsinized, resuspended in staining buffer (PBS, 2% BCS, 2 mM EDTA), and transferred to an unbound plate.

[0291] Cells were stained with different markers, including CD69, CD8, CD4, CD11c, and Annexin-V. Analysis focused on apoptotic and dead target cells, as well as activated CD8+ T cells. Target cells were identified using 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 antibody was incubated at room temperature for 30 minutes with gentle agitation. Cells were washed once with staining buffer and once with Annexin-binding buffer, and Annexin-V staining was performed at room temperature for 30 minutes with agitation. The cells were washed once with annexin-V binding buffer, and flow cytometry analysis was performed using 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 surviving CD4+, CD8+ T cells, and CD11c+ cells correspond to the percentages of annexin-V-negative cells within different cell populations.

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

[0297] [Table 19]

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

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

[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 level of HER2 expression and high level of PD-L1 expression), MCF-7 (low level of HER2 and very low level of PD-L1 expression) stimulated with 10 ng / ml IFNy for 24 hours to increase PD-L1 expression, and CHO-PD-L1 (no HER2, high PD-L1, purchased from BPS Bioscience). Note that this cell line expresses PD-L1 at approximately 9 times lower levels compared to 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 indicate that PRO1543, a molecule containing a low-affinity PD-L1 domain, is more potent than PRO1557 and PRO957 in killing cells expressing both HER2 and PD-L1, while it does not kill cells that do not express HER2. Co-expression of HER2 and PD-L1 occurs only in cancer cells. Furthermore, PRO1543 more potently mediates the lysis of HER2 / PD-L1 double-positive cancer cells than normal cells expressing only HER2. Therefore, in contrast to molecules lacking a PD-L1 binding domain, PRO1543 clearly exhibits selectivity for double-positive cells, thus allowing normal cells expressing HER2 to escape. Lacking a PD-L1 binding domain, PRO1557 and PRO957 do not possess such selectivity, and their potency is determined solely 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 CD3 local stimulation in the human cell line-derived ductal carcinoma xenograft model HCC1954. In human HCC1954 tubular breast cancer xenografts using Taconic's immunodeficient NOG mouse strain and allogeneic human peripheral blood mononuclear cells, the antitumor activity of the compounds in this approach was compared with anti-PD-1 therapy and anti-PD-1 / anti-Her2 combination therapy. The effects of PRO1678 (scMATCH3) and PRO1543 (MATCH4) on tumor volume were compared with treatment with an anti-PD-1 antibody (nivolumab) and a nivolumab / anti-Her2 antibody (trastuzumab) combo. An unrelated IgG, palivizumab, was used as a control IgG. Animal body weight was also tracked.

[0305] Study setup and dosing schedule Female NOG mice were injected unilaterally with 5 × 10⁶ HCC1954 cells. The cells were injected in a mixture of 50% cell suspension and 50% Matrigel in PBS, with a total injection volume of 100 μl. After successful tumor engraftment in NOG mice (median tumor volume in the group was 80–100 mm²), the tumor cells were injected into the mice. 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. Administration was started 1-2 hours after PBMC infusion and administered as follows.

[0306] [Table 22]

[0307] Body weight (Table 24) and tumor volume measurement using a caliber (Table 23 and Figure 13) were performed twice a week. Animal rearing was discontinued on day 33. Some animals in the group 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 examples of co-stimulating immune cell antigens). Design and fabrication of the MATCH4 molecule Figure 14 shows a description of the MATCH-4 molecule designed within this part of the approach. Its molecular composition is shown in Table 25.

[0311] [Table 25]

[0312] method: The MATCH4 construct was expressed in CHO-S cells using the CHOgro transient transfection kit (Mirus). Five to seven days after expression (cell viability <70%), the culture was collected by centrifugation at 37°C. The protein was purified from the clarified culture supernatant by protein L affinity chromatography, followed by final purification by size exclusion chromatography as needed. Standard analytical methods such as SE-HPLC, UV280, and SDS-PAGE were used for quality control of the prepared materials.

[0313] Anti-Her2×CD137×HSA×PD-L1 (低親和性) Evaluation of CD137 agonist effects using a cell-based assay with a transgenic NF-κB jarcut reporter cell line expressing the CD137 molecule of MATCH4. This assay evaluated the activation of CD137 signaling in Jurcut cells. CD137 signaling activity is reported by measuring luciferase expression driven by CD137-induced NF-κB activation in Jurcut reporter cell lines. Luciferase expression directly correlates with CD137 activity. Furthermore, CD137 clustering is required for activation of the signaling pathway and is facilitated through the formation of immunological synapses between Jurcut cells and Her2-expressing cell lines. Therefore, 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 level of Her2 but high expression level of PD-L1) were seeded in 96-well culture plates at a rate of 25,000 cells per well. The seeded cells were then stimulated with 10 ng / ml IFNy for 24 hours, or left unstimulated. Next, serial dilutions of the target MATCH4 molecule and the internal reference molecule PRO1186 or PRO1430 (both anti-CD137 × HSA × PD-L1 (high affinity) scMATCH3) were prepared and added to the cells. After the addition of the target molecule, Jarcut reporter cells were prepared in assay medium containing 25 mg / ml of HSA and added at a cell density of 40,000 cells per well. Luciferase expression was detected by adding a luciferase reagent and read with a luminescence reader 24 hours after the addition of the Jarcut cells. The data is presented by plotting the relative luminescence units (RLU) of the test samples as a function of the test sample concentration, and then fitting the data using a sigmoid 4PL fit (graph pad prism).

[0315] result: As shown in Figure 15, PRO1993 was found to induce CD137 signaling in the Jarcut 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. Data from the NF-κB reporter gene assay of the ND029 molecule are shown in Table 26.

[0316] [Table 26]

[0317] Neutralization of PD-L1 / PD-1 by FC method: These assays were performed to evaluate the ability of a multispecific low-affinity PD-L1× / Her2 construct to neutralize the interaction between PD-1 and PD-L1 expressed on HCC1954 cells that also express Her2. Specifically, the molecule should efficiently block the binding of PD-1 to PD-L1 in Her2 / PD-L1 co-expressing target cells (HCC1954), but its ability to block the PD-1 / PD-L1 interaction should be far weaker than that of avelumab / bavencio® or nivolumab / Opdivo® in cells that do not express Her2 (HCC827). Blocking 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 under test and avelumab as a reference were added to plates. Individual ICs of each plate were analyzed. 50 The IC values ​​were captured together on each plate for the reference molecule avelumab. 50 Calibrated against (relative IC) 50 :I C 50 avelumab / IC 50 (Test molecule). Furthermore, relative IC50 was observed in high-Her2-expressing HCC1954 cells and low-Her2-expressing HCC827 cells. 50 The ratio of the values ​​was calculated.

[0318] result: The efficacy of the MATCH4 construct in this approach is summarized in Table 27. The PD-L1 inhibition curve obtained for PRO1993 is shown in Figure 16.

[0319] [Table 27]

[0320] Example 4: Generation and testing of a multispecific construct targeting HER2 (as an example of TAA) and PD-L1. Design and fabrication of scDb-scFv Figure 17 shows descriptions of various scDb-scFv(scMATCH3) molecules designed within this aspect of the approach. Table 28 details the data for the fabrication of all molecules, showing the domains constituting the various molecules and their arrangement within those molecules.

[0321] method: The scDb-scFv construct was expressed in CHO-S cells using the CHOgro transient transfection kit (Mirus). Five to seven days after expression (cell viability <70%), the culture was collected by centrifugation at 37°C. The protein was purified from the clarified culture supernatant by protein L affinity chromatography, followed by final purification by size exclusion chromatography as needed. Standard analytical methods such as SE-HPLC, UV280, and SDS-PAGE were used for quality control of the prepared materials.

[0322] [Table 28]

[0323] Affinity of human Her2, PD-L1, and human and mouse serum albumin (SA) based on SPR method: The affinity for human PD-L1, Her2, and human and mouse serum albumin (SA) was measured by SPR using a ViaCore T200 instrument (GE Healthcare) as described above. The obtained SPR sensorgrams were double-referenced (empty reference channel and zero analyte injection) and fitted using BiaEvaluation software (GE Healthcare) and a 1:1 Langmuir coupled model. The quality of the fitting was monitored based on Chi2 and U values ​​(ViaCore). The apparent dissociation equilibrium constant (K) was determined using the acquired kinetic parameters. D The K of low affinity PD-L1 was calculated by fitting a plot of the equilibrium response against the concentration in addition to kinetic measurements. D The value obtained was (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 equivalent (K D Values: PRO1678: 156 nM, PRO1679: 122 nM).

[0325] [Table 29]

[0326] Neutralization of PD-L1 / PD-1 by FC method: The assay was performed to evaluate the ability of a multispecific low-affinity PD-L1× / Her2 construct to neutralize the interaction between PD-1 and PD-L1 expressed on HCC1954 cells that also express Her2. Specifically, the molecule should efficiently block the binding of PD-1 to PD-L1 in Her2 / PD-L1 co-expressing target cells (HCC1954), but its ability to block the PD-1 / PD-L1 interaction should be far weaker than that of avelumab / bavencio® or nivolumab / Opdivo® in cells that do not express Her2 (HCC827). The 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 50 The IC values ​​were captured together on each plate for the reference molecule avelumab. 50 Calibrated against (relative IC) 50 :I C 50 avelumab / IC 50 (Test molecule). Furthermore, relative IC50 was observed in high-Her2-expressing HCC1954 cells and low-Her2-expressing HCC827 cells. 50 The ratio of the values ​​was calculated.

[0327] result: The effectiveness of the scMATCH3 construct in 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 (derived 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 (derived from Amsbio, control cells expressing high levels of PD-L1 and not expressing human Her2) were collected and their cell counts were measured. The cell suspensions were centrifuged at 400×g for 5 minutes, and 100 μl (50,000 cells) of the cell suspension diluted with PBS-EB (1×DPBS, 2% BCS HI, 2 mM EDTA) was added to the designated wells of an unbound plate. After three washing steps with PBS-EB, the cells were centrifuged and the washing buffer was aspirated. The test samples (MATCH4 molecules: PRO1543 based on anti-HER2 trastuzumab and PRO1895 based on anti-HER2 pertuzumab) and 100 μl of serial dilutions of the reference antibodies trastuzumab and pertuzumab, starting at a concentration of 50 nM, were added directly to the plate. Dilution series of the multispecific molecules and avelumab were prepared in 3-fold steps, starting at 20 μg / ml and 5 μg / ml, respectively, with a staining buffer containing 500 ng / ml biotinylated PD1. After incubation at 4°C for 1 hour, the plate was washed three times with 100 μl of PBS-EB. The cell pellet incubated with the MATCH4 molecules was resuspended with 100 μl of Numab framework-specific detection antibody, followed by detection with APC-labeled anti-rabbit IgG antibody at a concentration of 2 μg / ml, and incubated at 4°C for 1 hour. Cell pellets incubated with reference antibodies trastuzumab and pertuzumab molecules were resuspended in 100 μl of RPE-labeled anti-human Fc antibody at a concentration of 5 μg / ml and incubated at 4°C for 1 hour. Next, the cells were washed three times with 100 μl of PBS-EB. The cell pellet was resuspended in 50 μl of PBS-EB and analyzed using a NovoCyte 2060 flow cytometer. For each sample, the fluorescence intensity of the APC and RPE channels was recorded for 5,000 events, and the geometric mean of the fluorescence intensity MFI was calculated. The data were single-referenced (subtracting the fluorescence intensity observed in cells incubated with buffer and detection antibody only), and the resulting concentration-response curves were fitted using 4-PL fitting (GraphPad Prism software). result: The apparent binding affinity to SK-OV3, MCF-7, and CHO PD-L1 was evaluated using flow cytometry. Individual EC of each plate 50 The values ​​are the EC values ​​for each reference molecule (i.e., trastuzumab for PRO1543 and pertuzumab for PRO1895). 50 It was calibrated against. The obtained EC 50 and relative EC 50 The maximum binding in flow cytometry is shown in Table 31. The concentration-response curve is shown in Figure 19. As shown in Table 31, the MATCH4 molecule binds to SK-OV3 cells expressing high levels of HER2 with an apparent binding affinity comparable to that of the clinical-stage antibodies trastuzumab and pertuzumab. This result indicates that, when tested for binding to cells expressing high levels of HER2 and PD-L1 (even at very low levels), the MATCH4 molecule can match the apparent binding affinity of the bivalent anti-HER2 antibodies trastuzumab and pertuzumab through avidity effects (binding to HER2 and PD-L1). On the other hand, when evaluating binding to cells expressing both antigens at very low levels (e.g., MCF-7 cells), 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 for trastuzumab and the trastuzumab-based MATCH4 molecule PRO1543. It can be inferred that pertuzumab can bind to hamster HER2, while 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 only PD-L1. Binding to Her2-expressing HCC1954 and HCC827 cells by flow cytometry. method: The apparent affinity of PRO1543 and PRO1895, as well as the clinical-stage anti-HER2 antibodies trastuzumab and pertuzumab, to HCC1954 (high levels of Her2 and PD-L1 expression) and HCC827 (low levels of Her2 expression but high levels of PD-L1 expression) was evaluated by flow cytometry. After further increasing PD-L1 expression by stimulating HCC827 and HCC1954 cells with IFNy for 24 hours, the cells were tested by flow cytometry as described above, except that the protein dilution series was started at 150 nM.

[0329] [Table 31]

[0330] result: Individual EC of each plate 50 The values ​​are the EC values ​​for each reference molecule (i.e., trastuzumab for PRO1543 and pertuzumab for PRO1895). 50 The value was calibrated. The obtained EC 50 and relative EC 50The maximum binding in flow cytometry is shown in Table 32. The concentration-response curve is shown in Figure 20. As shown in Table 32, the MATCH4 molecule bound to IFNy-stimulated HCC827 cells expressing high levels of PD-L1 and low levels of HER2 with apparent binding affinity equivalent to that of the clinical-stage antibodies trastuzumab and pertuzumab. On the other hand, when the binding of MATCH4 to cells expressing both antigens at high levels (IFNy-stimulated HCC1954 cells) was evaluated, the apparent binding affinity of MATCH4 was inferior to that of the clinical-stage antibodies. Simultaneous binding of PRO1543 and pertuzumab, and PRO1895 and trastuzumab to SK-OV3 cells in flow cytometry. method: The ability of the MATCH4 molecule to bind to HER2-expressing SK-OV3 cells in the presence of saturated concentrations of trastuzumab and pertuzumab was evaluated 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 the MATCH4 molecule. After incubating the cells with trastuzumab or pertuzumab at 4°C for 1 hour, serial dilutions of the MATCH4 molecule, starting from 50 nM, were added to the cells. Subsequently, the cell membrane binding of the MATCH4 molecules PRO1543 and PRO1895 was evaluated by flow cytometry as described above. result: Individual EC of each plate 50 The values ​​are the EC values ​​for each reference (i.e., the binding of the MATCH4 molecule without the anti-HER2 antibodies trastuzumab and pertuzumab). 50 The value was calibrated. The obtained EC 50 and relative EC 50 The maximum binding in flow cytometry is shown in Table 33. The concentration-response curve is shown in Figure 21. As shown in Table 33 and Figure 21, the MATCH4 molecule bound to SK-OV3 cells with apparent binding affinity comparable to that found in the aforementioned experiment (see Table 31), even in the absence of the antibodies trastuzumab and pertuzumab. Notably, PRO1543 bound to SK-OV3 cells in the presence of pertuzumab, and PRO1895 bound to SK-OV3 cells in the presence of trastuzumab, exhibiting similar binding affinity (relative EC2) to that obtained in the absence of each antibody. 50 (The values ​​were compared) and it was also possible to bind to cells. On the other hand, when SK-OV3 cells were incubated with trastuzumab and pertuzumab, respectively, before the addition of the MATCH4 molecule, no binding of PRO1543 and PRO1895 was observed. Since PRO1543 contains the anti-HER2 binding moiety of trastuzumab and PRO1895 contains the anti-HER2 binding moiety of pertuzumab, binding of PRO1543 in the presence of pertuzumab and PRO1895 in the presence of trastuzumab was expected and demonstrated 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 human serum albumin, wherein the antibody or antigen-binding fragment comprises the HCDR1, HCDR2, and HCDR3 sequences of SEQ ID NOs. 61, 62, and 63, and the LCDR1, LCDR2, and LCDR3 sequences of SEQ ID NOs. 64, 65, and 66, respectively. An antibody or its antigen-binding fragment, including the above.

2. The antibody or its antigen-binding fragment is as follows: (i) A VH domain having an amino acid sequence that is at least 95 percent identical to the amino acid sequence of SEQ ID NO: 67 within the framework sequence; and a VL domain having an amino acid sequence that is at least 95 percent identical to the amino acid sequence of SEQ ID NO: 69 within the framework sequence; or (ii) A VH domain having an amino acid sequence that is at least 95 percent identical to the amino acid sequence of SEQ ID NO: 68 within the framework sequence; and a VL domain having an amino acid sequence that is at least 95 percent identical to the amino acid sequence of SEQ ID NO: 70 within the framework sequence. The antibody or antigen-binding fragment thereof according to claim 1, comprising:

3. below: (i) the VH domain of SEQ ID NO: 67 and the VL domain of SEQ ID NO: 69; or (ii) The VH domain of sequence number 68 and the VL domain of sequence number 70, The antibody or antigen-binding fragment thereof according to claim 2, comprising:

4. A multispecific antibody comprising (i) at least one antibody or antigen-binding fragment thereof as described in any one of claims 1 to 3, and (ii) at least one antibody or antigen-binding fragment thereof having specificity to a target different from human serum albumin.

5. The multispecific antibody according to claim 4, wherein the target, which is different from human serum albumin, is a tumor-associated immune checkpoint antigen or tumor-associated antigen (TAA).

6. The multispecific antibody according to claim 5, 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).

7. The multispecific antibody according to claim 5, 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.

8. The aforementioned TAAs include EGFRvIII, 5T4, CD19, CD20, CD22, CD38, BCMA, IL4RA, mesoserine, GD2, Tn antigen, sTn antigen, Tn-O-glycopeptide, sTn-O-glycopeptide, PSMA, CD97, TAG72, CD44v6, CEA, EPCAM, KIT, IL-13Ra2, Legman, GD3, CD171, IL-11Ra, IL-13RA2, ROR1, PSCA, MAD-CT-1, MAD-CT-2, VE 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, Regmaine, HPV E6 or E7, ML-IAP, CLDN6, TSHR, GPRC5D, ALK, polysialic acid, Fos-related antigen, neutrophil elastase, TRP-2, CYP1B1, sperm protein 17, β-human chorionic gonadotropin, AFP, thyroglobulin, PLAC1, globoH, RAGE1, MN-CA IX, human telomerase reverse transcriptase, intestinal carboxylesterase, mut A multispecific antibody according to claim 5, 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.

9. (iii) A multispecific antibody according to any one of claims 4 to 8, further comprising at least one antibody or antigen-binding fragment thereof that is specific to an immune cell antigen.

10. (i) The immune cell antigen is a stimulating immune cell antigen selected from the group consisting of CD3 and CD16, or (ii) The multispecific antibody according to claim 9, wherein the immune cell antigen is a co-stimulating 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.

11. A pharmaceutical composition comprising a multispecific antibody according to any one of claims 4 to 10, and a pharmaceutically acceptable carrier.

12. A nucleic acid encoding an antibody or antigen-binding fragment thereof according to any one of claims 1 to 3, or a multispecific antibody according to any one of claims 4 to 10.

13. A vector comprising the nucleic acid described in claim 12.

14. A host cell comprising the nucleic acid described in claim 12 or the vector described in claim 13.

15. A method for producing an antibody or antigen-binding fragment thereof according to any one of claims 1 to 3, or a multispecific antibody according to any one of claims 4 to 10, the following: (i) the step of culturing the host cells according to claim 14, and (ii) A step of expressing an antibody or antigen-binding fragment thereof according to any one of claims 1 to 3 or a multispecific antibody according to any one of claims 4 to 10. Methods that include...

Images