Non-activating antibody variants

Abstract

Provided herein are proteins including Fc regions, such as monoclonal antibodies, bispecific antibodies, and multispecific antibodies, where the Fc region has been modified to eliminate or strongly reduce Fc-mediated effector function for therapeutic purposes and where effector function is undesirable, while at the same time allowing for good developability.

Description

[Technical field]

[0001] The present invention relates to polypeptides comprising Fc regions, such as monoclonal antibodies, bispecific antibodies and multispecific antibodies, where the Fc region has been modified to eliminate or strongly reduce Fc-mediated effector function for therapeutic purposes and in cases where effector function is undesirable, while at the same time allowing for good developability. [Background technology]

[0002] Antibodies have proven successful as therapeutic molecules, especially for the treatment of cancer and immune regulation. Tumor target-specific antibodies can achieve tumor cell cytotoxicity typically through Fc-mediated effector functions such as complement-dependent cytotoxicity (CDC), antibody-dependent cell-mediated cytotoxicity (ADCC) or antibody-dependent cell-mediated phagocytosis (ADCP). Immune cell targeting antibodies can enhance cells of the immune system, such as T cells and macrophages, and promote tumor cell cytotoxicity. Antibodies that target components of the immune system can be used to regulate immune system functions.

[0003] In certain scenarios, activation of the immune system or its components by antibody therapy may be undesirable, such as when applied to (i) neutralize cytokines systemically, (ii) block specific immune cell receptors, or (iii) use bispecific antibodies to redirect effector cells to target diseased tissues (Kang et al. Exp. Mol. Med. 2019;51(11):1-9). For example, when an immune cell-targeting antibody engages complement component C1q via its Fc region, activation of the classical complement system may be initiated, resulting, for example, in CDC of undesirable target immune cells. In addition, antibody Fc regions may also bind to Fc receptors expressed on a range of immune cells, resulting in undesirable depletion of effector cells or inducing immune-related toxicity via high levels of cytokine secretion.

[0004] To avoid unwanted Fc-mediated effector functions, antibodies have been engineered to carry mutations in the Fc portion (also called deactivating mutations) that reduce or eliminate some or all of the Fc-mediated effector mechanisms.

[0005] A wide range of different non-activated antibody formats have been developed that introduce amino acid substitutions and combinations thereof into the constant heavy chain region of IgG1 isotype antibodies to eliminate Fc-mediated effector functions (see, e.g., Chiu et al., Antibodies 2019 Dec;8(4):55; Liu et al., Antibodies, 2020 Nov 17;9(4):64;29(10):457-66; Shields et al., J Biol Chem., 2001 Mar 2;276(9):6591-604). Examples of such substitutions include the introduction of an N297G inactivating mutation (Tao and Morrison, J Immunol 1989; 143(8): 2595-601), an E233P-L234V-L235A-delG236-S267K inactivating mutation (Moore at al., Methods 2019; 154: 38-50), an L234A-L235A-P329G inactivating mutation (Schlothauer et al., Protein Eng. Design and Selection 2016; 29(10): 457-66), or an L234F-L235E-D265A inactivating mutation (also referred to herein as FEA or FEA format, Engelberts et al., EBioMedicine 2020; 52: 102625; U.S. Pat. No. 10,590,206). Other non-activated formats have been developed using IgG4, an IgG subclass with reduced effector function, in combination with amino acid substitutions in the constant heavy chain region of the antibody to further eliminate Fc-mediated effector functions (e.g. introduction of E233P-F234V-L235A-G236del non-activating mutations as described in WO 2015 / 143079, or introduction of F234A-L235A non-activating mutations as described by Vafa et al. Methods 2014;65:114-126).

[0006] However, such antibody engineering can result in developability and manufacturability issues that can limit the potential development of antibodies as therapeutics. For example, viral inactivation constitutes a critical step in the development of potential antibody therapeutics. Standard viral inactivation protocols require low pH maintenance. However, engineered proteins such as antibodies may respond differently to low pH conditions, potentially resulting in protein instability. For example, other conditions that need to be evaluated in the development and manufacturability of antibodies are repeated freeze-thaw cycles and fluctuations in storage temperature. It is therefore desirable to provide engineering of antibody therapeutics that allows them to retain protein stability when exposed to such conditions and the like, as this allows for ease of manufacture and distribution of such products as well as clinical use.

[0007] Thus, there is a need to provide additional non-activated formats suitable for antibodies that lack Fc-mediated effector functions and can be easily developed and produced for therapeutic use. [Prior art documents] [Patent documents]

[0008] [Patent Document 1] U.S. Pat. No. 1,059,0206 [Patent Document 2] International Publication No. 2015 / 143079 [Non-patent literature]

[0009] [Non-Patent Document 1] Kang et al.Exp.Mol.Med.2019;51(11):1-9 [Non-Patent Document 2] Chiu et al., Antibodies 2019 Dec;8(4):55 [Non-Patent Document 3] Liu et al.,Antibodies,2020 Nov 17;9(4):64;29(10):457-66 [Non-Patent Document 4] Shields et al., J Biol Chem.,2001 Mar 2;276(9):6591-604 [Non-Patent Document 5] Tao and Morrison, J Immunol 1989;143(8):2595-601 [Non-Patent Document 6] Moore at al.,Methods 2019;154:38-50 [Non-Patent Document 7] Schlothauer et al., Protein Eng. Design and Selection 2016;29(10):457-66 [Non-Patent Document 8] Engelberts et al.,EBioMedicine 2020;52:102625 [Non-Patent Document 9] Vafa et al.Methods 2014;65:114-126 Summary of the Invention [Problem to be solved by the invention]

[0010] The L234F-L235E-D265A inactivating mutation (also referred to herein as FEA or FEA format) has been shown to have an excellent safety profile and the ability to strongly suppress Fc-mediated effector functions. Nevertheless, it was observed that IgG1 antibodies that are potent inducers of complement-dependent cytotoxicity (CDC) with the FEA mutation may exhibit some residual CDC (see, inter alia, Examples 3 and 5). Furthermore, it was observed that recombinantly expressed antibodies with the FEA format may exhibit increased glycosylation heterogeneity compared to wild-type IgG1 Fc regions as a result of further processing of their N-glycans (data not shown, see also Example 14), and were also shown to be more susceptible to aggregation induced by low pH conditions (see, e.g., Example 20).

[0011] Therefore, the inventors have attempted to provide an improved inactivated format that can avoid residual CDC activity, provide a wild-type-like glycosylation profile, and / or may be more resistant to low pH conditions. Surprisingly, when the inventors combined the mutations L234F, L235E and G236R (herein also referred to as FER or FER format) in an IgG1 antibody, this resulted in an improved inactivated format that can avoid potential residual CDC activity, provide wild-type-like glycosylation and have improved resistance to low pH conditions. Thereby, the inventors now provide a highly advantageous further inactivated antibody format that is well suited for clinical development and use. As shown in the Examples section, this further inactivated format can be considered as a best-in-class inactivated format, although it may also be useful in other contexts than antibodies, such as fusion proteins. [Means for solving the problem]

[0012] Thus, the present invention provides a new inactive format for polypeptides having a human IgG1 Fc region, said region having substitutions at positions 234, 235 and 236 according to the Eu numbering, preferably with the substitutions F, E and R, respectively. This inactive format is particularly useful for monospecific and bispecific antibodies. Such polypeptides, e.g. monospecific or bispecific antibodies, having such substitutions according to the invention may be referred to as having a non-activated IgG1 Fc region.

[0013] In the case of bispecific antibodies, this inactive format may also be combined in a heterodimeric format with respect to the inactive format substitution, e.g., a bispecific antibody may be composed of one chain with an inactive format substitution according to the invention, and the other chain may contain a different inactive format substitution, e.g., FEA. Thus, the inactive format according to the invention is well suited to be combined with existing candidate antibodies, e.g., already undergoing development for clinical use, without the need to redesign and rerun all the required assays, thereby allowing the rapid generation of bispecific antibodies therewith, utilizing techniques such as controlled Fab arm exchange.

[0014] Thus, in one embodiment there is provided a protein comprising a first polypeptide and a second polypeptide, each of which comprises at least the hinge, CH2 and CH3 regions, respectively, of a human IgG1 immunoglobulin heavy chain, and wherein at least one of the first and second polypeptides is modified to include substitutions of amino acids corresponding to amino acids at positions L234, L235 and G236, the amino acid positions being as defined by Eu numbering. As such, preferably the amino acids at positions L234, L235 and G236 of at least one of the first and second polypeptides are substituted with F, E and R, respectively.

[0015] In another embodiment there is provided a protein according to the invention, wherein one of the first and second polypeptides comprises amino acid substitutions corresponding to the amino acids at positions L234, L235 and G236, and the other is modified and comprises amino acid substitutions corresponding to the amino acids at positions L234, L235 and D265, preferably the substitutions are F, E and A, respectively.

[0016] In another further embodiment there is provided the protein according to the invention, wherein both the first and second polypeptides comprise such substitutions of amino acids corresponding to amino acids L234, L235 and G236.

[0017] It will be appreciated that a protein according to the invention can be any protein that can benefit from having an Fc region that is inactive. This can include, for example, fusion proteins in which a functional domain is fused to an Fc region, thereby providing the functional domain with, for example, an improved plasma half-life.

[0018] As such, a protein according to the invention preferably comprises a first and a second binding domain. An exemplary and preferred protein according to the invention having a first and a second binding domain is an antibody.

[0019] Thus, in another further embodiment, the protein according to the invention is an antibody. In another further embodiment, the first and second polypeptides of the protein are identical in the protein or antibody according to the invention. In another embodiment, such antibodies are monospecific antibodies, since they will typically have the same binding domain. Such monospecific antibodies can then be used to generate bispecific antibodies according to the invention.

[0020] Thus, in yet another further embodiment, the protein according to the invention is a bispecific antibody. Preferably, a bispecific antibody according to the invention is provided, wherein said first and second polypeptides comprise further substitutions in their respective CH2 and CH3 regions such that the sequences of the respective CH2 and CH3 regions from said first and second polypeptides are different, said substitutions making it possible to obtain said polypeptide comprising said first and second polypeptides. A preferred example of such a substitution is having one of the first and second polypeptides comprising a substitution of the amino acid at position F405 by L and a substitution of the amino acid at position K409 by R. Alternatively, other substitutions or methods making it possible to provide heterodimers (notably combining different first and second polypeptides to provide bispecific antibodies) can be envisaged according to the invention.

[0021] Thus, in one embodiment there is provided a method for preparing a bispecific antibody according to the invention, the method comprising the steps of: a) providing a first antibody comprising: a. an immunoglobulin heavy chain comprising at least the hinge region, the CH2 region and the CH3 region, respectively, of a human IgG1 immunoglobulin heavy chain, the immunoglobulin heavy chain comprising substitutions of amino acids at positions L234, L235 and G236 with F, E and R, respectively; and b. an immunoglobulin light chain; b) providing a second antibody comprising: a. an immunoglobulin heavy chain comprising at least the hinge, CH2 and CH3 regions, respectively, of a human IgG1 immunoglobulin heavy chain, wherein the immunoglobulin heavy chain comprises: - at positions L234, L235 and D265, preferably said substitutions being F, E and A, respectively, or - substitutions of amino acids at positions L234, L235 and G236 by F, E and R, respectively; and b. an immunoglobulin light chain; c) wherein the sequences of the first and second CH3 regions of each of the first and second antibodies are different such that a heterodimeric interaction between the first and second CH3 regions is stronger than a homodimeric interaction of the first and second CH3 regions, respectively; d) incubating the first antibody with the second antibody under reducing conditions sufficient to allow cysteines in the hinge region to undergo disulfide bond isomerization; e) obtaining the bispecific antibody comprising the first immunoglobulin heavy chain and the first immunoglobulin light chain of the first antibody and the second immunoglobulin heavy chain and the second immunoglobulin light chain of the second antibody; Includes. [Brief description of the drawings]

[0022] [Figure 1]Figure 1. Target binding by anti-human CD20 antibody variants. Binding to CD20 present on Raji cells by IgG1 and IgG4 antibody variants with non-activating mutations in the heavy chain is shown. Binding is shown as antibody concentration vs. median MFI-PE (dataset split into A and B). Data are means ± SEM from four independent replicates.

[0023] The variants tested were IgG1-FEA-K409R, IgG1-FER-K409R, IgG1-AAG-K409R, IgG1-RR-K409R, IgG1lh2-S267K-K409R, IgG1-N297G-K409R, IgG1-AEASS-K409R, IgG1, IgG1-K409R, IgG1-b12, IgG4lh2-S228P, IgG4-PAA, IgG4, IgG4-S228P, where FEA:L234 F-L235E-D265A, FER:L234F-L235E-G236R, AAG:L234A-L235A-P329G, RR:G236R-L328R, IgG1lh2:E233P-L234V-L235A-G23 6del, AEASS:L234A-L235E-G237A-A330S-P331S, IgG4lh2:E233P-F234V-L235A-G236del, and PAA:S228P-F234A-L235A. [Diagram 2]Figure 1: CDC of Raji cells by anti-human CD20 antibody variants with non-activating mutations in the constant heavy chain region. CDC of CD20 positive Raji cells induced by IgG1 and IgG4 antibody variants with non-activating mutations in the constant heavy chain region was assessed using NHS as a source of complement. Cytolysis is determined by analysis of the percentage of PI positive cells by flow cytometry. CDC is shown as area under the curve (AUC) normalized against the non-binding control antibodies IgG1-b12 (0%) and wild type IgG1 (100%). Data are mean ± SEM from three independent repeats. The variants tested were IgG1-FEA-K409R, IgG1-FER-K409R, IgG1-AAG-K409R, IgG1-RR-K409R, IgG1lh2-S267K-K409R, IgG1-N297G-K409R, IgG1-AEASS-K409R, IgG1, IgG1-K409R, IgG1-b12, IgG4lh2-S228P, IgG4-PAA, IgG4, IgG4-S228P, where FEA:L234 F-L235E-D265A, FER:L234F-L235E-G236R, AAG:L234A-L235A-P329G, RR:G236R-L328R, IgG1lh2:E233P-L234V-L235A-G23 6del, AEASS:L234A-L235E-G237A-A330S-P331S, IgG4lh2:E233P-F234V-L235A-G236del, and PAA:S228P-F234A-L235A. [Diagram 3] Figure 1 shows C1q binding by anti-human CD20 IgG1 antibody variants with non-activating mutations in the constant heavy chain region. Binding is shown as antibody concentration vs. median MFI-FITC. Data are means (±SD) from triplicate determinations in one experiment. Tested variants were: IgG1-FEA-K409R, IgG1-FER-K409R, IgG1, IgG1-b12, where FEA: L234F-L235E-D265A and FER: L234F-L235E-G236R. [Figure 4]Figure 1 shows CDC of Raji cells induced by anti-human HLA-DR antibody variants with inactivating mutations in the constant heavy chain region. CDC of Raji cells induced by IgG1-HLA-DR-4 (A) and IgG1-HLA-DR-1D09C3 (B) antibody variants with inactivating mutations in the constant heavy chain region and HLA-DR targeting F(ab')2 fragments was evaluated in an in vitro CDC assay using NHS as a source of complement. Cytolysis is determined by analysis of the percentage of PI positive cells by flow cytometry. CDC is shown as the area under the curve (AUC) normalized to an IgG1 antibody with a K409R mutation (IgG1-K409R, 100%). Data are means ± SEM from five (wild type and L234F-L235E-D265A-K409R variants) or two (L234F-L235E-G236R-K409R variants, or F(ab')2 fragments) independent replicates. The variants tested were IgG1-FEA-K409R, IgG1-FER-K409R, IgG1-K409R, F(ab')2, where FEA: L234F-L235E-D265A and FER: L234F-L235E-G236R. [Diagram 5]Figure 1. Capture of anti-human CD20 IgG1 and IgG4 antibody variants to ELISA plates. Immobilization of IgG1 and IgG4 variants with non-activating mutations in the heavy chain region by anti-human IgG F(ab')2 fragments to ELISA plates. Binding is shown as area under the curve (AUC) normalized to wild type IgG1 (100%). Data are means (± SEM) from three independent replicates. Detection was performed with anti-human IgG-Fcγ-HRP and ABTS. The variants tested were IgG1-FEA-K409R, IgG1-FER-K409R, IgG1-AAG-K409R, IgG1-RR-K409R, IgG1lh2-S267K-K409R, IgG1-N297G-K409R, IgG1-AEASS-K409R, IgG1, IgG1-K409R, IgG4lh2-S228P, IgG4-PAA, IgG4, IgG4-S228P, where FEA:L234F-L23 5E-D265A, FER:L234F-L235E-G236R, AAG:L234A-L235A-P329G, RR:G236R-L328R, IgG1lh2:E233P-L234V-L235A-G236d el, AEASS:L234A-L235E-G237A-A330S-P331S, IgG4lh2:E233P-F234V-L235A-G236del, and PAA:S228P-F234A-L235A. [Figure 6]Figure 1 shows binding of anti-human CD20 IgG1 and IgG4 antibody variants with non-activating mutations in the constant heavy chain region to FcγRIa, FcγRIIa (allotypes 131H and 131R), FcγRIIb and FcγRIIIa (allotypes 158F and 158V). Binding of immobilized IgG1 and IgG4 variants with non-activating mutations to His-tagged biotinylated extracellular domains (ECDs) of FcγRIa (A), FcγRIIa allotype 131H (B), FcγRIIa allotype 131R (C), FcγRIIb (D), FcγRIIIa allotype 158F (E) and FcγRIIIa allotype 158V (F) monomers and dimers as tested in an ELISA assay. Binding is shown as area under the curve (AUC) normalized to wild type IgG1 (100%). Data are means (±SEM) from three independent replicates. Detection was performed with streptavidin-polyHRP and ABTS. The variants tested were IgG1-FEA-K409R, IgG1-FER-K409R, IgG1-AAG-K409R, IgG1-RR-K409R, IgG1lh2-S267K-K409R, IgG1-N297G-K409R, IgG1-AEASS-K409R, IgG1, IgG1-K409R, IgG4lh2-S228P, IgG4-PAA, IgG4, IgG4-S228P, where FEA:L234F-L23 5E-D265A, FER:L234F-L235E-G236R, AAG:L234A-L235A-P329G, RR:G236R-L328R, IgG1lh2:E233P-L234V-L235A-G236d el, AEASS:L234A-L235E-G237A-A330S-P331S, IgG4lh2:E233P-F234V-L235A-G236del, and PAA:S228P-F234A-L235A. [Figure 7]Figure 1 shows FcγR activation by anti-human CD20 IgG1 and IgG4 antibody variants with non-activating mutations in the heavy chain region as measured using target-expressing Raji cells and FcyR-expressing reporter cells. (A-F) Activation of Jurkat reporter cell lines stably expressing either (A) FcγRIa, (B) FcγRIIa allotype 131H, (C) FcγRIIa allotype 131R, (D) FcγRIIb, (E) FcγRIIIa allotype 158F, or (F) FcγRIIIa allotype 158V when co-cultured with Raji cells expressing CD20 and different concentrations of IgG1 and IgG4 antibody variants as measured by luminescence levels (RLU). Activation is shown per experimental replicate as the area under the dose-response curve (AUC) normalized to the non-binding control IgG1-b12 (0%) and wild-type IgG1 (100%). Data are means (±SEM) from three independent experimental replicates. The variants tested were IgG1-FEA-K409R, IgG1-FER-K409R, IgG1-AAG-K409R, IgG1-RR-K409R, IgG1lh2-S267K-K409R, IgG1-N297G-K409R, IgG1-AEASS-K409R, IgG1, IgG1-K409R, IgG1-b12, IgG4lh2-S228P, IgG4-PAA, IgG4, IgG4-S228P, where FEA:L234 F-L235E-D265A, FER:L234F-L235E-G236R, AAG:L234A-L235A-P329G, RR:G236R-L328R, IgG1lh2:E233P-L234V-L235A-G23 6del, AEASS:L234A-L235E-G237A-A330S-P331S, IgG4lh2:E233P-F234V-L235A-G236del, and PAA:S228P-F234A-L235A. [Figure 8]Figure 1 shows ADCC induced by anti-human CD20 IgG1 and IgG4 antibody variants with non-activating mutations in the heavy chain region as measured using the DELFIA® EuTDA TRF cytotoxicity assay. (A-B) NK cell-mediated ADCC was measured by the release level of EuTDA reagent from peripheral blood mononuclear cells (PBMCs) and BATDA-labeled CD20-expressing Raji cells upon co-incubation with anti-human CD20 IgG1 and IgG4 antibody variants. For some variants, ADCC is shown as the area under the curve (AUC) normalized to the non-binding control IgG1-b12 (0%) and wild-type IgG1 (100%) for each experimental replicate (A). Data are means (± SEM) obtained from four (wild type and K409R variants) or two (L234F-L235E-D265A-K409R and L234F-L235E-G236R-K409R variants) independent replicates. For some variants, ADCC is shown as (B) percentage lysis at an antibody concentration of 10 μg / ml compared to non-binding control IgG1-b12 (0%) and wild type IgG1 (100%) per experimental replicate. Data are means (± SEM) obtained from six donors from two independent experiments. The variants tested were IgG1-FEA-K409R, IgG1-FER-K409R, IgG1-AAG-K409R, IgG1-RR-K409R, IgG1lh2-S267K-K409R, IgG1-N297G-K409R, IgG1-AEASS-K409R, IgG1, IgG1-K409R, IgG1-b12, IgG4lh2-S228P, IgG4-PAA, IgG4, IgG4-S228P, where FEA:L234 F-L235E-D265A, FER:L234F-L235E-G236R, AAG:L234A-L235A-P329G, RR:G236R-L328R, IgG1lh2:E233P-L234V-L235A-G23 6del, AEASS:L234A-L235E-G237A-A330S-P331S, IgG4lh2:E233P-F234V-L235A-G236del, and PAA:S228P-F234A-L235A. [Figure 9]Figure 1: T cell activation by variants of anti-human CD3 antibodies IgG1 or IgG4-huCLB-T3 / 4 with non-activating mutations in the constant heavy chain region. (A-B) Upregulation of CD69 expression as a measure of early T cell activation (measured by flow cytometric analysis) on T cells in PBMC co-cultures induced by anti-human CD3 IgG1 and IgG4 antibody variants with the indicated mutations. CD69 upregulation is shown (A) as a dose response to the percentage of CD69+ T cells (CD28+) or (B) as the area under the dose-response curve (AUC) normalized to the wild-type antibody variant IgG1-F405L (100%) per donor and experimental replicate. Data are mean values ​​(± SEM) from three independent experimental replicates (two independent donors per experimental replicate). The variants tested were IgG1-FEA-F405L, IgG1-FER-F405L, IgG1-AAG-F405L, IgG1-RR-F405L, IgG1lh2-S267K-F405L, IgG1-N297G-F405L, IgG1-AEASS-F405L, IgG1-F405L, IgG4lh2-S228P-F405L-R409K, IgG4-PAA-F405L-R409K, IgG4-S228P, where FEA:L2 34F-L235E-D265A, FER:L234F-L235E-G236R, AAG:L234A-L235A-P329G, RR:G236R-L328R, IgG1lh2:E233P-L234V-L235A-G2 36del, AEASS:L234A-L235E-G237A-A330S-P331S, IgG4lh2:E233P-F234V-L235A-G236del, and PAA:S228P-F234A-L235A. [Figure 10]Figure 1. In vitro T cell-mediated cytotoxicity by non-activating bispecific antibody variants. (A-C) T cell-mediated cytotoxicity of HER2-positive SK-OV-3 cells in PBMC co-cultures was assessed using bispecific antibody variants, CD3xHER2 (A), CD3xb12 (B; does not bind target cells) or b12xHER2 (C; no binding to T cells), carrying non-activating mutations in the constant heavy chain region using Alamar blue. Absorbance at 590 nm was measured using an Envision plate reader and the percentage of viable cells was calculated for each donor and experimental replicate, with staurosporine-treated cells representing 100% cytotoxicity and media control (no antibody, no PBMC) representing 0% cytotoxicity. Data are presented as dose-response curves vs. % viable SK-OV-3 cells. Data are mean values ​​(± SEM) from three independent experimental replicates (two independent donors per experimental replicate). The bispecific antibody variants tested were BisG1F405LxK409R, BisG1FEA-F405LxFEA-K409R, BisG1FER-F405LxFER-K409R, BisG1AAG-F405LxAAG-K409R, BisG1RR-F405LxRR-K409R, BisG1lh2S267K-F405LxS267K-K409R, BisG4lh2S228P-F405L-R409KxS228P, BisG1N297G-F405LxN297G-K409R, BisG1AEASS-F405LxAEAS S-K409R, BisG4PAA-F405L-R409KxPAA, where FEA: L234F-L235E-D265A, FER: L234F-L235E-G236R, AAG: L234A-L235A-P329G, RR: G236R-L328R, G1lh2: E233P-L234V-L235A-G236del, AEASS: L234A-L235E-G237A-A330S-P331S, G4lh2: E233P-F234V-L235A-G236del, and PAA: S228P-F234A-L235A. [Figure 11]Figure 1 shows total human IgG (hIgG) concentrations measured in blood samples taken from mice injected with anti-human CD20 IgG1 or anti-human CD3 IgG1 (huCLB-T3 / 4) antibody variants. (A) Total hIgG concentrations in blood samples taken from mice injected with wild-type anti-human CD20 IgG1, IgG1-CD20-K409R, IgG1-CD20-L234F-L235E-D265A-K409R and IgG1-CD20-L234F-L235E-G236R-K409R at different time points after injection. Data are means (±SEM) from 3 mice per group, except for IgG1-FER-K409R (2 mice). (B) Total hIgG concentrations in blood samples taken from mice injected with wild-type anti-human CD3 IgG1, IgG1-CD3-F405L, IgG1-CD3-L234F-L235E-D265A-F405L, and IgG1-CD3-L234F-L235E-G236R-F405L at different times after injection. Data are mean values ​​(±SEM) obtained from 3 mice per group. (C) Clearance up to 21 days after antibody administration was determined according to the formula D*1000 / AUC, with D, injected dose, and AUC, area under the curve of the concentration-time curve. In all figures, the dotted line represents the predicted IgG1 concentration over time for wild-type IgG1 antibodies in SCID mice. Data shown are from 3 mice per group, except for IgG1-FER-K409R (2 mice). The variants tested were: IgG1-FEA-K409R, IgG1-FER-K409R, IgG1-K409R, IgG1-FEA-F405L, IgG1-FER-F405L, IgG1-F405L, IgG1, where FEA: L234F-L235E-D265A and FER: L234F-L235E-G236R. [Figure 12] FIG. 1 shows a schematic representation of the different glycan species detected in the IgG1 antibody variants tested in Example 14. [Figure 13]Figure 1 shows the efficiency of controlled Fab arm exchange (cFAE) for the generation of bsAb variants. A bispecific antibody (designated as BisG1) is generated by cFAE where one monospecific antibody (designated as IgG1-A) has a F405L mutation and another monospecific antibody (designated as IgG1-B) has a K409R mutation. The efficiency of cFAE for the generation of bsAb variants is shown in (A; 20 data points) where both monospecific antibodies carry the L234F-L235E-G236R (FER) inactivating mutations in addition to the F405L and K409R mutations, (B; 16 data points) where the first monospecific antibody carries the L234F-L235E-G236R (FER) mutation in addition to the F405L mutation and the second monospecific Variants were evaluated in which the antibody harbors the L234F-L235E-D265A (FEA) mutation in addition to the K409R mutation, (C; 12 data points) the first monospecific antibody harbors the L234F-L235E-D265A (FEA) mutation in addition to the F405L mutation, and the second monospecific antibody harbors the L234F-L235E-G236R (FER) mutation in addition to the K409R mutation. The percentage of bsAb or residual monospecific antibody variants (IgG1-A or IgG1-B) is shown and was determined using an Orbitrap Q-Exactive Plus mass spectrometer. FEA: L234F-L235E-D265A and FER: L234F-L235E-G236R. [Figure 14]Figure 1 shows the production levels of antibody variants with either L234F-L235E-G236R (FER) or L234F-L235E-D265A (FEA) deactivating mutations in the constant heavy chain region in addition to either F405L or K409R. Antibody variants were produced in Expi293F cells. Production titers are expressed as mg / L in a scatter dot plot and mean values ​​(±SEM) are shown. Each dot represents production yield data for a particular antibody clone with the indicated mutation (mean value when more production data was available for that particular clone). To allow comparison, production titers of matched clones for the L234F-L235E-D265A-F405L antibody variant (FEA-F405L; open circles) and the L234F-L235E-G236R-F405L antibody variant (FER-F405L; closed circles) are shown. Similarly, production titers of matched clones for the L235E-D265A-K409R antibody variant (FEA-K409R, open boxes) and the L234F-L235E-G236R-K409R antibody variant (FER-K409R, closed boxes) are shown. [Figure 15] 1 shows exemplary schematic diagrams of monospecific antibodies (A) and bispecific antibodies (B). (A) Heavy chain shown in black; light chain shown in white. Heavy and light chain domains of individual antibodies are shown as CH1, CH2, CH3 and VH (constant heavy chain domains (H1, H2, H3) and variable heavy chain domain (VH)), CL and VL (CL, VL, constant and variable light chain domains). (B) Bispecific antibody consisting of two half molecules with two different specificities of Fab arms (one half molecule shown as heavy and light chain in black and white, respectively; one half molecule shown as heavy and light chain stripes), e.g., a bispecific antibody generated by controlled Fab arm exchange. Hinge region, Fab arms and Fc domains are as shown. [Figure 16]Figure 1 shows C1q binding by anti-human CD20 IgG1 antibody variants with non-activating mutations in the heavy chain constant region. Binding is shown as area under the curve (AUC) normalized to non-binding control antibody IgG1-b12 (0%) and wild type IgG1-CD20 (100%). Data are means ± SEM from three independent experiments. The antibody variants tested were IgG1-CD20 wild type (IgG1) and its variants, IgG1-FEA-F405L, IgG1-FEA-K409R, IgG1-FER-F405L, IgG1-FER-K409R, BisG1FEA-F405LxFEA-K409R, BisG1FER-F405LxFER-K409R, BisG1FER-F405LxFEA-K409R, BisG1FEA-F405LxFER-K409R, where FER: L234F-L235E-G236R and FEA: L234F-L235E-D265A. [Figure 17] Figure 1: CDC of Raji cells by anti-human CD20 IgG1 antibody variants with non-activating mutations in the heavy chain constant region. CDC of CD20 positive Raji cells induced by IgG1-CD20 antibody variants with non-activating mutations in the heavy chain constant region was assessed using NHS as a source of complement. Cytolysis is determined by analysis of the percentage of PI positive cells by flow cytometry. CDC is shown as area under the curve (AUC) normalized against the non-binding control antibody IgG1-b12 (0%) and wild type IgG1-CD20 (100%). Data are mean ± SEM from three independent repeats. The antibody variants tested were IgG1-CD20 wild type (IgG1) and its variants, IgG1-FEA-F405L, IgG1-FEA-K409R, IgG1-FER-F405L, IgG1-FER-K409R, BisG1FEA-F405LxFEA-K409R, BisG1FER-F405LxFER-K409R, BisG1FER-F405LxFEA-K409R, BisG1FEA-F405LxFER-K409R, where FER: L234F-L235E-G236R and FEA: L234F-L235E-D265A. [Figure 18] Figure 1. In vitro T cell-mediated cytotoxicity by non-activating bispecific antibody variants. (A-B) Alamar Blue was used to assess T cell-mediated cytotoxicity of HER2-positive SK-OV-3 cells in PBMC co-cultures using bispecific antibody variants CD3xHER2 (A) or CD3xb12 (B; does not bind target cells) with asymmetric non-activating mutations in the Fc region. The absorbance at 590 nm was measured using an Envision plate reader and the percentage of viable cells was calculated for each donor and experimental replicate, with staurosporine-treated SK-OV-3 cells representing 100% cytotoxicity and medium control (SK-OV-3 cells, no antibody, no PBMC) representing 0% cytotoxicity. Data are presented as dose-response curves for the percentage of viable SK-OV-3 cells. Data are mean ± SEM obtained from four donors from two independent experiments (two donors per independent experiment). The CD3xHER2 and CD3xb12 bispecific antibody variants tested were BisG1F405LxK409R, BisG1FER-F405LxK409R, BisG1FER-F405LxFEA-K409R, BisG1FER-F405LxAAG-K409R, and BisG1FER-F405LxN297G-K409R, where FER: L234F-L235E-G236R, FEA: L234F-L235E-D265A, and AAG: L234A-L235A-P329G. [Figure 19]Figure 1 shows T cell activation by non-activating bispecific CD3xHER2 antibody variants. Upregulation of CD69 expression (measured by flow cytometric analysis) as a measure of early T cell activation on T cells in human PBMC co-cultures was assessed using wild-type-like CD3xHer2 bispecific antibody variants and their variants with the indicated symmetric or asymmetric non-activating mutations in the Fc region. CD69 upregulation is shown as area under the dose-response curve (AUC) normalized to the AUC values ​​measured for the non-binding negative control IgG1-b12 (0%) and the wild-type-like IgG1 bispecific antibody variant (BisG1F405LxK409R, 100%) per donor and experimental replicate. Data are mean values ​​(±SEM) obtained from four donors in two independent experiments (two donors per independent experiment). The CD3xHER2 bispecific antibody variants tested were BisG1F405LxK409R, BisG1FER-F405LxK409R, BisG1FER-F405LxFEA-K409R, BisG1FER-F405LxAAG-K409R, and BisG1FER-F405LxN297G-K409R, where FER: L234F-L235E-G236R, FEA: L234F-L235E-D265A, and AAG: L234A-L235A-P329G. [Figure 20]Figure 1 shows CDC of Raji cells induced by anti-human CD20 antibody variants with non-activating mutations in the heavy chain constant region, assessed in an in vitro CDC assay using NHS as a source of complement. The ability to induce CDC was compared between variants produced to contain or lack a C-terminal lysine in the heavy chain constant region. Cytolysis is determined by analysis of the percentage of PI positive cells by flow cytometry. CDC is shown as the area under the curve (AUC) normalized against wild type IgG1-CD20 antibody (IgG1; 100%) and no antibody control sample (0%). Data are mean ± SEM from three independent experiments. The variants tested were: IgG1, IgG1-delK, IgG1-FEA, IgG1-FEA-delK, IgG1-FER, IgG1-FER-delK, where FEA: L234F-L235E-D265A, FER: L234F-L235E-G236R, delK: recombinant deletion of the HC C-terminal lysine. [Figure 21]Figure 1 shows human FcγR activation by anti-human CD20 antibody variants with non-activating mutations in the heavy chain constant region as measured using target-expressing Raji cells and FcγR-expressing reporter cells. The ability to induce FcγR activation was compared between variants produced to contain or lack a C-terminal lysine in the heavy chain constant region. (A-D) Activation of Jurkat reporter cell lines stably expressing either (A) FcγRIa, (B) FcγRIIa allotype 131H, (C) FcγRIIb, or (D) FcγRIIIa allotype 158V as measured by luminescence levels upon co-culture with CD20-expressing Raji cells and various concentrations of IgG1-CD20 or IgG1-CD20-delK antibody variants. Activation is shown per experimental replicate as the area under the dose-response curve (AUC) normalized to non-binding control IgG1-b12 (0%) and wild-type IgG1 (100%). Data shown are the mean ± SEM of two independent replicates. The variants tested were IgG1, IgG1-delK, IgG1-FEA, IgG1-FEA-delK, IgG1-FER, IgG1-FER-delK, where FEA: L234F-L235E-D265A, FER: L234F-L235E-G236R, delK: recombinant deletion of the HC C-terminal lysine. [Figure 22] Figure 1 shows CDC of Raji cells induced by allotypic variants of wild-type anti-human CD20 IgG1 antibody and its variants with non-activating mutations in the heavy chain constant region, assessed in an in vitro CDC assay using NHS as a complement source. Cell lysis is determined by analysis of the percentage of PI-positive cells by flow cytometry. CDC is shown as the area under the curve (AUC) normalized against wild-type anti-human CD20 antibody (allotype IgG1(f); 100%) and against no antibody control sample (0%). Data are mean ± SEM from three independent experiments. The variants tested were IgG1(fa), IgG1(zax), IgG1(zav), IgG1(za) and IgG1(f), where FEA: L234F-L235E-D265A, FER: L234F-L235E-G236R. [Diagram 23] Figure 1 shows human FcγR activation by anti-human CD20 IgG1 antibody variants with non-activating mutations in the heavy chain constant region of different IgG1 allotype variants as measured using target-expressing Raji cells and FcyR-expressing reporter cells. (A-D) Activation of Jurkat reporter cell lines stably expressing either (A) FcγRIa, (B) FcγRIIa allotype 131H, (C) FcγRIIb, or (D) FcγRIIIa allotype 158V as measured by luminescence levels when co-cultured with CD20-expressing Raji cells and various concentrations of IgG1-CD20 antibody variants. Activation is shown per experimental replicate as area under the dose-response curve (AUC) normalized to non-binding control IgG1-b12 (0%) and wild-type IgG1 (f) (100%). Data shown are the mean ± SEM of two independent repeats. The variants tested were IgG1(f), IgG1(za), IgG1(zav), IgG1(zax), IgG1(fa) and variants with FER or FEA mutations, where FER: L234F-L235E-G236R and FEA: L234F-L235E-D265A. [Figure 24]Figure 1 shows CDC of Raji cells induced by subclass variants of wild-type anti-human CD20 antibodies and their variants with non-activating mutations in the heavy chain constant region, assessed in an in vitro CDC assay using NHS as a complement source. (A) CDC induced by wild-type anti-human CD20 IgG1 and IgG3 antibodies (allotypes IGHG3*01 [IgG3] and IGHG3*04 [IgG3rch2]) and their non-activating variants. (B) CDC induced by wild-type anti-human CD20 IgG1 and IgG4 antibodies and their non-activating variants. Cytolysis is determined by analysis of the percentage of PI-positive cells by flow cytometry. CDC is shown as the area under the curve (AUC) normalized against wild-type anti-human CD20 IgG1 antibody (IgG1; 100%) and against no antibody control sample (0%). Data are mean ± SEM from three independent experiments. FEA:L234F-L235E-D265A, FER:L234F-L235E-G236R, EA:L235E-D265A, and ER:L235E-G236R. [Diagram 25]Figure 1 shows human FcγR activation by anti-human CD20 IgG1, IgG3 and IgG4 antibody variants with non-activating mutations in the heavy chain constant region as measured using target-expressing Raji cells and FcyR-expressing reporter cells. (A-D) Activation of Jurkat reporter cell lines stably expressing either (A) FcγRIa, (B) FcγRIIa allotype 131H, (C) FcγRIIb or (D) FcγRIIIa allotype 158V as measured by luminescence levels when co-cultured with CD20-expressing Raji cells and various concentrations of IgG-CD20 antibody variants. Activation is shown per experimental replicate as area under the dose-response curve (AUC) normalized to non-binding control IgG1-b12 (0%) and wild-type IgG1 (100%). Data shown are the mean ± SEM of two independent repeats. The variants tested were IgG1, IgG3 (IGHG3*01), IgG3rch2 (IGHG3*04), IgG4 and variants thereof harboring ER, EA, FER, or FEA mutations, where ER:L235E-G236R, EA:L235E-D265A, FER:L234F-L235E-G236R, and FEA:L234F-L235E-D265A. [Figure 26]Figure 1 shows human FcγR activation by anti-human CD20 mouse IgG2a antibody variants with non-activating mutations in the heavy chain constant region as measured using target-expressing Raji cells and FcyR-expressing reporter cells. (A-D) Activation of Jurkat reporter cell lines stably expressing either (A) FcγRIa, (B) FcγRIIa allotype 131H, (C) FcγRIIb, or (D) FcγRIIIa allotype 158V as measured by luminescence levels when co-cultured with CD20-expressing Raji cells and various concentrations of mouse IgG2a-CD20 antibody variants. Activation is presented as area under the dose-response curve (AUC) normalized to non-binding control IgG2a-b12 (0%) and wild-type IgG2a-CD20 (100%) per experimental replicate. Data shown are the mean ± SEM of two independent replicates. The variants tested were IgG2a, IgG2a-FER, IgG2a-LALA, and IgG2a-LALAPG, where FER: L234F-L235E-G236R, LALA: L234A-L235A, and LALAPG: L234A-L235A-P329G. [Figure 27] Figure 1 shows C1q binding by anti-human CD20 mouse IgG2a antibody variants with non-activating mutations in the heavy chain constant region when CD20-positive Raji cells were opsonized with normal human serum (NHS) as a source of C1q. Binding is shown as area under the curve (AUC) normalized to the non-binding control antibody IgG2a-b12 (0%) and wild-type mouse IgG2a-CD20 (100%). Data are means ± SEM from three independent experiments. Antibody variants tested were wild-type IgG2a, IgG2a-FER, IgG2a-LALA, and IgG2a-LALAPG, where FER: L234F-L235E-G236R, LALA: L234A-L235A, and LALAPG: L234A-L235A-P329G. [Figure 28]Figure 1. CDC of Raji cells by anti-human CD20 mouse IgG2a antibody variants with non-activating mutations in the heavy chain constant region. CDC of CD20 positive Raji cells induced by IgG2a-CD20 antibody variants with non-activating mutations in the heavy chain constant region was assessed using normal human serum (NHS) as a source of complement. Cytolysis is determined by analysis of the percentage of PI positive cells by flow cytometry. CDC is shown as area under the curve (AUC) normalized against the non-binding control antibody IgG2a-b12 (0%) and wild type mouse IgG2a-CD20 (100%). Data are mean ± SEM from three independent repeats. The antibody variants tested were wild type IgG2a, IgG2a-FER, IgG2a-LALA, and IgG2a-LALAPG, where FER: L234F-L235E-G236R, LALA: L234A-L235A, and LALAPG: L234A-L235A-P329G. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0024] As described herein, certain modifications at amino acid positions in the Fc region of an antibody have been shown to be non-activating modifications that render the protein essentially inactive with respect to Fc function, while at the same time possessing advantageous properties from a manufacturing standpoint.

[0025] The term "non-activating" as used herein is intended to refer to the inhibition or abolishment of the interaction of the protein according to the present invention with Fc receptors (FcR) or C1q present on a wide range of effector cells, such as monocytes, to activate the complement pathway. As used herein, "non-activating" includes reduced CDC activity, reduced C1q binding, reduced ADCC, reduced or absent binding to human FcγRIa, FcγRIIa(H), FcγRIIa(R), FcγRIIb, FcγRIIIa(F) and FcγRIIIa(V), reduced or absent activation and signaling via human FcγRIa, FcγRIIa(H), FcγRIIa(R), FcγRIIb, FcγRIIIa(F) and FcγRIIIa(V). "Non-activating", when used in the context of targeting CD3 (e.g., in monospecific antibodies targeting CD3 or in the context of bispecific or multispecific antibodies in a tumor-associated antigen-independent manner), also includes not inducing T-cell activation. It is understood that such "non-activating" characteristics should preferably be assessed in comparison to a protein that is not "non-activating", e.g., comparing an antibody having an unmodified Fc region with wild-type-like function as described herein to a modified Fc region according to the invention. The term "Fc region" as used herein is intended to refer to a region comprising, in the N-terminal to C-terminal direction, at least the hinge region, the CH2 region and the CH3 region.

[0026] The term "protein" as used herein is intended to refer to a large biological molecule comprising one or more chains of amino acids covalently linked to each other. Such bonds may be via peptide bonds and / or disulfide bridges. A single chain of amino acids may also be called a "polypeptide". Thus, a protein in the context of the present invention may consist of one or more polypeptides. A protein according to the present invention may be any kind of protein, such as an antibody or a variant of a parent antibody, or a fusion protein.

[0027] The term "antibody" as used herein is intended to refer to an immunoglobulin molecule, a fragment of an immunoglobulin molecule, or any derivative thereof, that has the ability to specifically bind to an antigen under typical physiological conditions with a half-life of a significant period, such as at least about 30 minutes, at least about 45 minutes, at least about 1 hour, at least about 2 hours, at least about 4 hours, at least about 8 hours, at least about 12 hours, about 24 hours or more, about 48 hours or more, about 3, 4, 5, 6, 7 days or more, etc., or any other relevant functionally defined period (e.g., a time sufficient to induce, promote, enhance and / or modulate a physiological response associated with antibody binding to the antigen and / or a time sufficient for the antibody to recruit effector activity). The binding region (or binding domain, which may be used herein with the same meaning) that interacts with the antigen includes the variable regions of both the heavy and light chains of the immunoglobulin molecule. The constant region of an antibody (Ab) may mediate the binding of the immunoglobulin to host tissues or factors, including various cells of the immune system (such as effector cells) and components of the complement system, such as C1q, the first component in the classical pathway of complement activation.

[0028] As stated above, the term antibody herein includes fragments of antibodies that retain the ability to specifically interact with, e.g., bind to, an antigen, unless otherwise specified or clearly contradicted by context. It has been shown that the antigen-binding function of an antibody can be performed by fragments of a full-length antibody. Examples of binding fragments encompassed by the term "antibody" include (i) Fab' or Fab fragments, V L , V H , C L and C H (ii) a monovalent fragment consisting of one domain, or a monovalent antibody as described in WO2007059782 (Genmab A / S); 2 (iii) a V fragment, a bivalent fragment comprising two Fab fragments linked by a disulfide bridge at the hinge region; H Domain and C H (iv) a Fd fragment consisting essentially of one domain, V of a single arm of an antibody; L Domain and V H(v) an Fv fragment consisting essentially of a V H These include dAb fragments, which consist essentially of domains and are also called domain antibodies (Ward et al., Nature 341, 544-546 (1989)) (Holt et al; Trends Biotechnol. 2003 Nov; 21(11):484-90); (vi) camelids or nanobodies (Revets et al; Expert Opin Biol Ther. 2005 Jan; 5(1):111-24); and (vii) isolated complementarity determining regions (CDRs). In addition, the two domains of the Fv fragment, V L and V H are encoded by separate genes, they can be synthesized using recombinant methods to L and V H The regions may be linked by synthetic linkers that allow them to be produced as a single protein chain that pairs to form a monovalent molecule (known as single chain antibodies or single chain Fvs (scFvs), see, e.g., Bird et al., Science 242, 423-426 (1988) and Huston et al., PNAS USA 85, 5879-5883 (1988)). Such single chain antibodies are encompassed within the term antibody unless otherwise stated or clearly indicated by the context. Although such fragments are generally included within the meaning of antibody, they collectively and each independently are unique features of the present invention and exhibit different biological properties and utilities. These and other useful antibody fragments in the context of the present invention are discussed further herein. The term antibody, unless otherwise specified, should also be understood to include polyclonal antibodies, monoclonal antibodies (mAbs), antibody-like polypeptides such as chimeric and humanized antibodies, as well as antibody fragments (antigen-binding fragments) that retain the ability to specifically bind to the antigen provided by any known technique, such as enzymatic cleavage, peptide synthesis, and recombinant techniques. The antibodies generated can have any isotype.

[0029] It should be understood within the context of the present invention that when the antibody is a fragment, such as a binding fragment, the fragment is fused to an Fc region as described herein. The antibody may thereby be a fusion protein falling within the scope of the present invention. Thus, in one embodiment, the protein is a fusion protein.

[0030] The term "humanized" as used herein in relation to antibodies refers to a genetically engineered non-human antibody comprising a human antibody constant domain and a non-human variable domain modified to contain a high level of sequence homology to the human variable domain. This can be achieved by grafting the non-human antibody complementarity determining regions (CDRs), which together form the antigen binding site, onto a homologous human acceptor framework region (FR) (see, inter alia, WO 92 / 22653 and EP 0629240). Substitution (backmutation) of framework residues from the parent antibody (i.e., non-human antibody) into the human framework region may be required to fully reconstitute the binding affinity and binding specificity of the parent antibody binding region. Structural homology modeling can help to identify amino acid residues within the framework region that are important for the binding properties of the antibody binding region. Thus, a humanized variable region or antibody may comprise non-human CDR sequences, primarily human framework regions that may contain one or more amino acid backmutations to the non-human amino acid sequences. Further amino acid modifications, not necessarily back mutations, may be applied to obtain humanized antibodies or humanized variable regions with favorable properties, such as particular useful affinity and biochemical properties, for example to include modifications that avoid deamidation, provide an "inactive Fc region", enhance heterodimerization, and / or improve manufacturing.

[0031] The term "human" as used herein in reference to antibody variable regions and antibodies is intended to include antibodies that may be engineered with variable and framework regions derived from human germline immunoglobulin sequences and constant domains derived from human immunoglobulin constant domains. The human variable regions or human antibodies of the invention may include amino acid residues not encoded by human germline immunoglobulin sequences (e.g., mutations, insertions or deletions introduced by random or site-specific mutagenesis in vitro or somatic mutation in vivo). A "human antibody" may incorporate VH and VL sequences generated from human germline immunoglobulin sequences in transgenic animals, such as those described in Lee et al. Nature Biotech, 32(4) 2014 pp 355-63 and Macdonald et al., PNAS April 8, 2014 111(14) 5147-52. Such VH and VL sequences are considered human VH and VL sequences that may be fused to constant domains derived from human immunoglobulin constant domains. Further amino acid modifications, not necessarily back mutations, may be applied to obtain human antibodies or human variable regions with favorable properties, such as particular useful affinities and biochemical properties, for example to avoid deamidation, provide an "inactive Fc region", enhance heterodimerization, and / or include modifications that improve manufacturing. Thus, a "human antibody" can include engineered antibodies.

[0032] As used herein, the term "complement dependent cytotoxicity" ("CDC") is intended to refer to the process of antibody-mediated complement activation that leads to lysis of a cell or virion as a result of membrane holes created by MAC assembly when an antibody binds to its target on a cell or virion.

[0033] As used herein, the term "antibody-dependent cell-mediated cytotoxicity" ("ADCC") is intended to refer to the mechanism of killing of antibody-coated target cells or virions by cells expressing Fc receptors that recognize the constant region of a bound antibody.

[0034] The term "immunoglobulin heavy chain" or "heavy chain of an immunoglobulin" as used herein is intended to refer to one of the heavy chains of an immunoglobulin. A heavy chain is typically composed of a heavy chain variable region (abbreviated herein as VH) that defines the immunoglobulin isotype, and a heavy chain constant region (abbreviated herein as CH). The heavy chain constant region is typically composed of three domains, CH1, CH2, and CH3. CH1 and CH2 are typically linked via a hinge region.

[0035] The term "immunoglobulin" as used herein is intended to refer to a class of structurally related glycoproteins consisting of two pairs of polypeptide chains, one pair of light (L) low molecular weight chains and one pair of heavy (H) chains, all four of which may be interconnected by disulfide bonds. The structure of immunoglobulins has been well characterized (see, e.g., Fundamental Immunology Ch. 7 (Paul, W., ed., 2nd ed. Raven Press, NY (1989)). Within the structure of an immunoglobulin, the two heavy chains are linked to each other via disulfide bonds at the so-called "hinge region". Like the heavy chains, each light chain is typically composed of several regions: a light chain variable region (abbreviated herein as VL) and a light chain constant region. The light chain constant region is typically composed of one domain, CL. Furthermore, the VH and VL regions can be further subdivided into regions of hypervariable regions (or hypervariable regions that may be hypervariable in sequence and / or in the form of structurally defined loops), also called complementarity determining regions (CDRs), interspersed with more conserved regions called framework regions (FRs). Each VH and VL is typically composed of three CDRs and four FRs, arranged from the amino terminus to the carboxy terminus in the following order: FR1, CDR1, FR2, CDR2, FR3, CDR3, FR4 (Lefranc MP et al, Dev Comp Immunol (See also Jan:27(1):55-77 (2003)).

[0036] As used herein, the terms "first polypeptide" and "second polypeptide" refer to a pair of polypeptides whose amino acid sequences may be identical or different. Thus, the first and second polypeptides may form homodimers or heterodimers. The first and second polypeptides may associate with additional polypeptides.

[0037] The term "isotype" as used herein refers to the immunoglobulin isotype (e.g., IgG, IgD, IgA, IgE or IgM) or its subclass (IgG1, IgG2, IgG3, IgG4) or any allotype thereof encoded by the heavy chain constant region genes. Allotypes of IgG1 include IgG1m(za) and IgG1m(f). Thus, in one embodiment, the protein comprises a heavy chain of an IgG1 class of immunoglobulin or any allotype thereof. Furthermore, each heavy chain isotype can be combined with a kappa (κ) and / or lambda (λ) light chain, or any allotype thereof.

[0038] The term "hinge region" as used herein refers to the hinge region of an immunoglobulin heavy chain. Thus, for example, the hinge region of a human IgG1 antibody is the region of the hinge region of an immunoglobulin heavy chain as described by Kabat (Kabat, EA et al., Sequences of proteins of immunological interest. 5 th This corresponds to amino acids 216 to 230 according to the Eu numbering as set forth in the US Pat. No. 6,313,662 (Eu)-US Department of Health and Human Services, NIH publication No. 91-3242, pp 662, 680, 689 (1991)).

[0039] The VH and VL regions may be "human VH and VL regions" or "humanized VH and VL regions". With respect to human VH and / or human VL regions, it is understood that such regions are typically composed of three CDRs and four FRs arranged from amino-terminus to carboxy-terminus in the following order: FR1, CDR1, FR2, CDR2, FR3, CDR3, FR4, as derived from or found in a human germline sequence. Such VH and VL regions may be derived from a humanized animal model or from a human. For example, human monoclonal antibodies may be produced by a hybridoma comprising B cells obtained from a transgenic or transchromosomal non-human animal, such as a transgenic mouse having a genome comprising human heavy and light chain transgenes, fused to an immortalized cell. Human monoclonal antibodies may be derived from human B cells or plasma cells. The term "humanized VH and VL regions" as used herein refers to engineered VH and VL regions derived from a non-human antibody that have been modified to contain a high level of sequence homology to the human variable domain. This can be achieved by grafting the non-human antibody complementarity determining regions (CDRs), which together form the antigen binding site, onto homologous human acceptor framework regions (FRs) (see, inter alia, WO 92 / 22653 and EP 0629240). Substitution (backmutation) of framework residues from the parent antibody (i.e., non-human antibody) into the human framework regions may be required to fully reconstitute the binding affinity and binding specificity of the parent antibody. Structural homology modeling can help identify amino acid residues within the framework regions that are important for the binding properties of the antibody. Thus, humanized VH and VL regions may include non-human CDR sequences, primarily human framework regions that may include one or more amino acid backmutations to the non-human amino acid sequences. Further amino acid modifications, not necessarily back mutations, may be applied, e.g. to include modifications to avoid deamidation and / or to improve manufacturing, to obtain humanized or human VH and VL regions with favourable characteristics, such as particular useful affinities and biochemical properties.

[0040] The term "CH2 region" or "CH2 domain" as used herein refers to the CH2 region of an immunoglobulin heavy chain. Thus, for example, the CH2 region of a human IgG1 antibody corresponds to amino acids 231-340 according to the Eu numbering system. However, the CH2 region may also be of any of the other subtypes as described herein.

[0041] The term "CH3 region" or "CH3 domain" as used herein refers to the CH3 region of an immunoglobulin heavy chain. Thus, for example, the CH3 region of a human IgG1 antibody corresponds to amino acids 341-447 according to the Eu numbering system. However, the CH3 region may also be of any of the other subtypes as described herein.

[0042] The term "full-length antibody" as used herein refers to an antibody (e.g., parent antibody or variant antibody) that includes all heavy and light chain constant and variable domains corresponding to those normally found in a wild-type antibody, i.e., has a VH, CH1, linker, CH2, CH3 region, respectively, in the heavy chain, and a VL and CL region, respectively, in the light chain, such as a human (or humanized) IgG1 heavy chain, or a human (or humanized) kappa or lambda light chain. It is understood that a bispecific antibody may also be a full-length antibody, i.e., one that includes different heavy and / or light chains, such as those normally found in a wild-type antibody. A full-length antibody may, for example, be engineered to include substitutions or modifications as defined herein in accordance with the present invention.

[0043] As used herein, the term "amino acid corresponding to position" refers to the amino acid position number of the human IgG1 heavy chain. Unless otherwise indicated or contradicted by context, amino acids in constant region sequences are numbered as follows in the Eu index (Kabat, EA et al., Sequences of proteins of immunological interest.5 thEdition-US Department of Health and Human Services, NIH publication No. 91-3242, pp 662, 680, 689 (1991). Thus, an amino acid or segment in one sequence that "corresponds to" an amino acid or segment in another sequence, typically aligned with the other amino acid or segment using ALIGN, ClustalW or similar standard sequence alignment programs with default settings, has at least 50%, at least 80%, at least 90%, or at least 95% identity to the human IgG1 heavy chain. Methods for aligning sequences or segments in sequences and thereby determining corresponding positions in sequences for amino acid positions according to the invention are believed to be well known in the art.

[0044] In the context of the present invention, amino acids may be defined by conservative or non-conservative classes. The classes of amino acids may therefore be reflected in one or more of the following tables: [Table 1] [Table 2] In the context of the present invention, substitutions in proteins are indicated as follows:

[0045] Original amino acid-position-substitute amino acid; Referring to the well-recognized nomenclature for amino acids, three-letter or one-letter codes are used, including the codes Xaa and X to denote any amino acid residue. Thus, the designation "L234F" or "Leu234Phe" means that the protein contains a substitution of leucine for phenylalanine at the protein amino acid position corresponding to amino acid position 234 of the wild-type protein.

[0046] Substitution of an amino acid at a given position with any other amino acid is referred to as follows:

[0047] The original amino acid-position; or for example "L234".

[0048] In cases where the original amino acid(s) and / or the replacement amino acid(s) may include multiple amino acids but not all amino acids(s), the multiple amino acids may be separated by "," or " / ". For example, a substitution of phenylalanine, arginine, lysine or tryptophan at position 234 with leucine is "Leu234Phe,Arg,Lys,Trp" or "L234F,R,K,W" or "L234F / R / K / W" or "L234 to F,R,K or W".

[0049] Such designations may be used interchangeably in the context of the present invention and have the same meaning and purpose.

[0050] Furthermore, the terms "substitution" or "mutation", which can be used interchangeably, encompass substitutions with any one of the other 19 natural amino acids or with other amino acids, such as unnatural amino acids. For example, substitutions of the amino acid L at position 234 include each of the following substitutions: 234A, 234C, 234D, 234E, 234F, 234G, 234H, 234I, 234K, 234M, 234N, 234Q, 234R, 234S, 234T, 234V, 234W, 234P, and 234Y. This corresponds, by way of example, to the designation 234X, where X denotes any amino acid other than the original amino acid. These substitutions may also be referred to as L234A, L234C, etc., or L234A, C, etc., or L234A / C / , etc. The same applies by analogy to each and every position mentioned herein, and specifically includes herein any one of such substitutions. It is well known in the art that when an amino acid sequence includes "X" or "Xaa", the X or Xaa represents any amino acid. Thus, X or Xaa can typically represent any of the 20 naturally occurring amino acids. As used herein, the term "naturally occurring" refers to any one of the following amino acid residues: glycine, alanine, valine, leucine, isoleucine, serine, threonine, lysine, arginine, histidine, aspartic acid, asparagine, glutamic acid, glutamine, proline, tryptophan, phenylalanine, tyrosine, methionine, and cysteine. The terms "amino acid" and "amino acid residue" may be used interchangeably. For the purposes of the present invention, sequence identity between two amino acid sequences can be determined using the Needleman-Wunsch algorithm (Needleman and Wunsch, 1970, J. Mol. Biol. 48:443-453) as implemented in the Needle program in the EMBOSS package (EMBOSS: The European Molecular Biology Open Software Suite, Rice et al., 2000, Trends Genet. 16:276-277), preferably version 5.0.0 or later.The parameters used are a gap open penalty of 10, a gap extension penalty of 0.5, and the EBLOSUM62 (the EMBOSS version of BLOSUM62) substitution matrix. The output of the needle labeled "longest identity" (obtained using the -nobrief option) is used as the percent identity, calculated as follows:

[0051] (identical residues × 100) / (length of alignment − total number of gaps in the alignment) Retention of similar residues may also, or alternatively, be measured by a similarity score determined by use of a BLAST program (e.g., BLAST 2.2.8 available from NCBI using standard settings BLOSUM62, Open Gap=11, and Extended Gap=1).

[0052] In one embodiment, in at least one of said first and second polypeptides, the amino acids at positions corresponding to positions L234, L235 and G236 of a human IgG1 heavy chain are not L, L and G, respectively. A protein is provided comprising a first polypeptide and a second polypeptide, each of said first and second polypeptides comprising at least the hinge region, CH2 region and CH3 region, respectively, of a human IgG1 immunoglobulin heavy chain, and at least one of said first and second polypeptides is modified to comprise a substitution of an amino acid corresponding to the amino acid at positions L234, L235 and G236, as the amino acid positions are defined by EU numbering. Preferably, the amino acids at positions L234, L235 and G236 of at least one of said first and second polypeptides are substituted with F, E and R, respectively.

[0053] With respect to amino acid positions used herein, they are numbered according to the Eu numbering system, which follows the Eu index of numbering as set forth in Kabat, EA et al., Sequences of proteins of immunological interest. 5th Edition-US Department of Health and Human Services, NIH publication No. 91-3242, pp 662, 680, 689 (1991). Sequences of amino acid sequences useful according to the invention are also provided herein with the indicated amino acid modifications in bold (see Table 1).

[0054] For example, as shown in the Examples section, an example of a protein according to the invention can be an antibody consisting of two identical heavy chains (corresponding to a first and a second polypeptide) and two identical light chains. However, it is not a requirement that both the first and the second polypeptide have the same substitutions, for example one of the first and the second polypeptide can have said substitutions at positions L234, L235 and G236, while the other can have, for example, a different substitution. Thus, the other chain can have, for example, a substitution of another inactive format, for example the FEA format. In the context of the present invention, a further advantage of the non-activating substitutions at positions L234, L235 and G236 is considered to be that they effectively suppress Fc-mediated effector functions even when other polypeptides are present in an "asymmetric" manner with a different inactive format, for example the FEA format. This allows the production of multispecific antibodies by combining the newly developed antibodies and the FER inactive format with previously produced antibodies with other non-activating substitutions, such as the FEA format.

[0055] Thus, in a further embodiment there is provided a protein according to the invention, wherein one of the first and second polypeptides comprises amino acid substitutions corresponding to the amino acids at positions L234, L235 and G236, and the other is modified and comprises amino acid substitutions corresponding to the amino acids at positions L234, L235 and D265, preferably said substitutions being F, E and A, respectively.

[0056] In another embodiment, in the protein according to the invention, both said first and said second polypeptides comprise said substitutions of amino acids corresponding to amino acids L234, L235 and G236, which are preferably substitutions by F, E and R, respectively.

[0057] In a further embodiment, each of the first and second polypeptides comprises an immunoglobulin CH1 region. The CH1 region is preferably linked to a hinge region, i.e. providing the polypeptide with the CH1 region, hinge region, CH2 region and CH3 region, respectively, of a human IgG1 immunoglobulin heavy chain. Preferably, the CH1 region is of a human IgG1 immunoglobulin heavy chain. For example, the CH1 region may be a sequence having the sequence listed in SEQ ID NO: 4. The CH1 region, hinge region, CH2 region and CH3 region defined herein may be the sequences listed in SEQ ID NO: 5. Such sequences may have substitutions as described herein, for example provided with FER substitutions and / or further substitutions as defined herein.

[0058] In another embodiment, the protein according to the invention comprises a first and a second binding region. Although any binding region may suffice, it may be preferred that the binding region is derived from an immunoglobulin binding region, such as a human or humanized antibody.

[0059] The term "binding region" or "binding domain" as used herein refers to a region of a protein capable of binding to an antigen, such as a polypeptide present on a cell, e.g. a cancer cell, a bacterium or a virion. The binding region can be a polypeptide sequence, e.g. a protein, a protein ligand, a receptor, an antigen-binding region, or a ligand-binding region capable of binding to a cell, a bacterium or a virion. In particular, the binding region is an antigen-binding region. When the binding region is, for example, a receptor, the protein may be prepared as a fusion protein of the receptor with the Fc domain of an immunoglobulin. When the binding region is an antigen-binding region, the protein according to the invention may be an antibody, a chimeric antibody, or an antibody with a humanized or human binding region antibody or a heavy chain only antibody or an ScFv-Fc fusion.

[0060] As used herein, the term "binding" refers to the binding of an antibody to an antigen as the analyte, as determined by biolayer interferometry. D 1E -6 Below M, for example 5E -7 M or less, 1E -7 Below M, for example 5E -8 Below M, for example 1E -8 Below M, for example 5E -9 M or less, or for example 1E -9 The antibody binds to a given antigen or target with a binding affinity corresponding to M or less, and K D "antigen" refers to a antibody that binds to a given antigen with an affinity that is at least 10 times lower, such as at least 100 times lower, such as at least 1,000 times lower, such as at least 10,000 times lower, such as at least 100,000 times lower, than its affinity for binding to a non-specific antigen other than the given antigen or a closely related antigen (e.g., BSA, casein).

[0061] Thus, in a further embodiment, the protein according to the invention comprises a first and a second binding region comprising a first immunoglobulin heavy chain variable region (VH) and a first immunoglobulin light chain variable region (VL), respectively, and the second binding region comprises a second immunoglobulin heavy chain variable region and a second immunoglobulin light chain variable region. It is understood that these regions are preferably human or humanized VH and VL regions. Thus, the VH and VL regions have framework regions of human or human origin and may have CDR1-3 sequences of human or human origin or of other species origin, such as mouse or rat. Thus, in another further embodiment, the immunoglobulin heavy and light chain variable regions are human or humanized immunoglobulin heavy and light chain variable regions.

[0062] As mentioned above, the protein according to the invention comprises an antibody. Thus, in another embodiment, the first and second polypeptides are immunoglobulin heavy chains, and the first and second polypeptides comprise the first and second immunoglobulin heavy chain variable regions, respectively. It is understood that such heavy chains may be preferably human or humanized heavy chains when they comprise a human variable region, which is understood to comprise a human or humanized variable region and a human constant region. Furthermore, the protein may comprise a first immunoglobulin light chain constant region and a second immunoglobulin light chain constant region, more preferably, the protein comprises a first and a second immunoglobulin light chain, and the immunoglobulin light chain comprises the respective first and second immunoglobulin light chain variable regions and the respective first and second immunoglobulin constant light chain regions. It is understood that such light chains are highly preferably human or humanized light chains when they comprise a non-human derived CDR region. The light chain may have a light chain variable region and a human kappa light chain constant region or a human lambda light chain constant region. In a further embodiment, the human kappa light chain constant region is as listed in SEQ ID NO: 6. In another further embodiment, the human lambda light chain constant region is as listed in SEQ ID NO: 7. Such light chains may comprise kappa or lambda light chains, or both, e.g., the protein according to the invention may comprise two different light chains, such that the protein comprises one kappa light chain and one lambda light chain. Thus, such light chains may be either human or humanized kappa or lambda light chains, or both, e.g., the protein according to the invention may comprise two different light chains, such that the protein comprises one human kappa light chain and one human lambda light chain.

[0063] It is understood that the human or humanized heavy and light chains, in addition to the mutations described herein, may contain further modifications to provide preferred characteristics, such as modifications to avoid deamidation and / or enhance heterodimerization, modifications to improve production and separation, and the like, to provide certain useful affinity and biochemical properties.

[0064] Thus, in a preferred embodiment, a protein according to the invention comprising said first and second polypeptides consists of or comprises a first and a second immunoglobulin light chain and a first and a second heavy chain, the latter corresponding to the first and the second polypeptide. Such a protein according to the invention may have a first immunoglobulin light chain linked to said first immunoglobulin heavy chain via a disulfide bridge and said second immunoglobulin light chain linked to said second immunoglobulin heavy chain via a disulfide bridge, thereby forming said first binding region and said second binding region, respectively, said first and second immunoglobulin heavy chains being linked likewise via a disulfide bridge. As used herein, the term "disulfide bridge" refers to a covalent bond between two cysteine ​​residues, i.e. said interaction may also be referred to as a Cys-Cys interaction.

[0065] In a further embodiment, which may be a preferred embodiment, the protein according to the invention is an antibody, preferably a full length antibody. In yet another further embodiment, the full length antibody is of or is derived from the human IgG1 isotype.

[0066] In accordance with the present invention, it is understood that substitutions at such positions, when included in, for example, an antibody, result in reduced Fc-mediated effector functions. Such reduced Fc-mediated effector functions include reduced CDC activity (see, inter alia, Examples 3 and 5), reduced C1q binding (see, inter alia, Example 4), reduced ADCC (see, inter alia, Example 9), the absence of detectable binding to human FcγRIa, FcγRIIa(H), FcγRIIa(R), FcγRIIb, FcγRIIIa(F) and FcγRIIIa(V) (see, inter alia, Example 6), and reduced binding to human FcγRIa, FcγRIIa(H), FcγRIIa(R), FcγRIIb, FcγRIIIa(F) and and FcγRIIIa(V) (Example 7), do not induce tumor-associated antigen-independent T cell activation in terms of targeting CD3 (especially Example 10), and have pharmacokinetics, particularly due to human FcRn binding characteristics similar to wild-type human IgG1, glycosylation very similar to wild-type human IgG1 antibodies, etc., when such substations are included in the context of antibodies (see, e.g., Examples 12 and 14). Furthermore, these substitutions can provide improved pH stability at low pH (especially Examples 20-23). The proteins according to the invention are not only useful in the context of antibodies, which may be a highly preferred embodiment, but other formats of proteins are also contemplated, such as fusion proteins, having said first and second polypeptides according to the invention comprising at least one hinge region, CH2 region and CH3 region, wherein at least one of said first and second polypeptides is modified to include substitutions of amino acids corresponding to amino acids at positions L234, L235 and G236, wherein the amino acid positions are as defined by EU numbering.

[0067] The term "reduced" or "undetectable" as used herein when referring to Fc-mediated effector functions relates to the ability of a protein according to the invention, such as an antibody, to induce CDC activity, ADCC, C1q binding, FcγRIa, FcγRIIa(H), FcγRIIa(R), FcγRIIb, FcγRIIIa(F), and FcγRIIIa(V) binding, activate and signal via FcγRIa, FcγRIIa(H), FcγRIIa(R), FcγRIIb, FcγRIIIa(F), and FcγRIIIa(V), and activate T cells in terms of targeting CD3, compared to the same protein having a wild-type IgG1 region, etc., which has full ability to induce said effector functions, e.g., in the context of an antibody, etc. as described in the Examples section. Exemplary assays for determining "reduced" or "undetectable binding" for these features are known in the art and are described in detail, for example, in the Examples section throughout. The Examples section describes the properties of these features in the context of antibodies, such as full-length antibodies. It is understood that the advantageous properties of each of the modified hinge, CH2 and CH3 regions of the human IgG1 immunoglobulin heavy chain according to the invention may also be useful in other formats, such as fusion proteins, so that the highly preferred properties are determined in the context as described in the Examples section, and this is not meant to mean that the proteins according to the invention are to be understood as being limited to antibodies. Nevertheless, the invention finds use, particularly when applied in the context of antibodies, and is presented herein.

[0068] The ability of a protein to induce complement-dependent cytotoxicity (CDC) can be determined using methods known in the art. Briefly, a reference protein with an unmodified fully functional Fc region, for example an IgG1 antibody, which is a strong CDC inducer, is incubated in an in vitro assay with cells presenting a target antigen on the cell surface in the presence of human serum, and the percentage of cell lysis is subsequently determined and set to 100%. The percentage of cell lysis of a protein according to the invention with a modified Fc region, for example an antibody with a FER Fc region, is compared to the cell lysis that occurs under the same conditions with a control protein that does not target cells or does not have an Fc region (such as F(ab')2). The percentage of lysis is determined relative to the unmodified reference, which is set to 100%. Examples of suitable methods that can be used are as described in Example 3 or Example 5. The protein according to the invention has reduced CDC activity when compared to the same protein with an Fc region instead having the FEA format and / or has similar CDC activity when compared to, for example, F(ab')2.

[0069] The ability of a protein to reduce C1q binding can be determined using methods known in the art. Briefly, a protein, e.g. an antibody, with a fully functional (unmodified) Fc region is incubated in an in vitro assay with cells presenting the target antigen on the cell surface in the presence of human serum, after which the percentage of C1q binding is determined, e.g. by binding with a polyclonal rabbit anti-human C1q complement FITC antibody (Dako, Cat. No. F0254, Agilent Technologies) and FACS analysis according to the manufacturer's instructions. The detected signal of the protein with a fully functional (unmodified) Fc region is compared to a protein with a modified Fc region, e.g. an IgG1 antibody, which is a strong CDC inducer. A detailed example of a suitable method according to the invention that can be used is described in Example 4. The protein according to the invention has reduced CDC activity when compared to the same protein with an Fc region instead having the FEA format, and / or has similar CDC activity when compared to, e.g., a non-binding control antibody. The protein according to the invention preferably has 15% or less C1q binding activity in the context of a full-length IgG1 antibody and FER, as described in Example 4, when compared to an antibody having the same sequence but without FER.

[0070] The ability to reduce antigen-dependent cytotoxicity can be determined using methods known in the art. For example, to determine the ADCC ability of wild-type antibodies and their non-activated variants, the DELFIA® EuTDA TRF (Time-Resolved Fluorescence) Cytotoxicity Kit (Cat. No. AD0116, Perkin Elmer) can be used according to the manufacturer's instructions. Briefly, cells presenting the target antigen are intracellularly labeled using, for example, Bis(acetoxymethyl) 2,2':6',2"-terpyridine-6,6"-dicarboxylate reagent solution (DELFIA BATDA Reagent, Cat. No. C136-100, Perkin Elmer) according to the manufacturer's instructions. These cells are then incubated in the presence of PBMC with PBMC cells presenting the target antigen on their cell surface, with a protein having a (unmodified) fully functional Fc region, for example, a wild-type antibody against the target antigen and its non-activated variant. The assessment of NK-mediated ADCC is determined with reference to a fully functional control IgG1 antibody (set to 100%) and a non-binding negative IgG1 control antibody (set to 0%). A detailed example of a suitable method that may be used is described in Example 9. The protein according to the invention preferably has 35% or less residual ADCC activity when compared to an antibody having the same sequence but without FER, for example when comparing a full-length IgG1 antibody with FER, as described in Example 9. The protein according to the invention has a similarly reduced ADCC activity when compared to the same protein with an Fc region having the FEA format instead.

[0071] Binding to human FcγRIa, FcγRIIa(H), FcγRIIa(R), FcγRIIb, FcγRIIIa(F) and FcγRIIIa(V) can be measured using ELISA. Binding of the proteins according to the invention can be assessed in an ELISA assay by determining binding to His-tagged C-terminally biotinylated FcγR, monomeric ECD of FcγRIa (SEQ ID NO: 15) (monomer), or dimeric ECD of FcγRIIa allotype 131H (SEQ ID NO: 16), FcγRIIa allotype 131R (SEQ ID NO: 17), FcγRIIb (SEQ ID NO: 18), FcγRIIIa allotype 158F (SEQ ID NO: 19), and FcγRIIIa allotype 158V (SEQ ID NO: 20). Briefly, for example, an IgG1 antibody with an Fc region is bound to a plate coated with an anti-human F(ab')2 antibody, followed by incubation with each of the respective extracellular domains, followed by quantification of the binding using streptavidin-polyHRP (CLB, Cat. No. M2032, 1:10.000). A detailed example of a suitable method according to the invention that can be used is described in Example 6. The protein, e.g., antibody, according to the invention does not detectably bind to the Fcγ receptor in an ELISA assay utilizing the His-tagged C-terminally biotinylated FcγR monomer ECD. The protein according to the invention has a similar undetectable binding to the Fcγ receptor as observed when comparing an IgG1 antibody with an Fc region containing FER, e.g., FEA, as shown in Example 6.

[0072] Concerning activation and signaling via FcγRIa, FcγRIIa(H), FcγRIIa(R), FcγRIIb, FcγRIIIa(F) and FcγRIIIa(V), this can be determined by commercially available reporter assays. For example, reporter assays can be used to determine activation and binding of the proteins according to the invention using target expressing cells and Jurkat reporter cell lines expressing the indicated FcγRs (Promega, FcγRIa: Catalog No. CS1781C08; FcγRIIa allotype 131H: Catalog No. G9991; FcγRIIa allotype 131R: Catalog No. CS1781B08; FcγRIIb: Catalog No. CS1781E04; FcγRIIIa allotype 158F: Catalog No. G9790; FcγRIIIa allotype 158V: Catalog No. G7010). For example, in the case of CD20 targeting antibodies, CD20 expressing Raji cells can be used as target cells. Proteins according to the invention, such as antibodies, have similar irremovable activation and signaling through the Fcγ receptor, e.g., as observed when comparing antibodies with Fc regions according to the invention, such as FER, and FEA, e.g., as shown in Example 7.

[0073] With regard to the reduction of T cell activation in the context of targeting CD3, it is understood that such applies to the proteins according to the invention, which comprise a binding region that binds to human CD3 on human T cells, for example an exemplary bivalent monospecific antibody that binds to human CD3 as described in the Examples herein. The reduction of T cell activation can be determined, for example, by incubating a dose response series of anti-CD3 antibodies, which comprise amino acid substitutions corresponding to amino acids at positions L234, L235 and G236 according to EU numbering, respectively, in the hinge, CH2 and CH3 regions of the human IgG1 immunoglobulin heavy chain, with freshly isolated PBMCs, in both of the two polypeptides according to the invention, followed by staining the cells with mouse anti-human CD28-PE (cat. no. 130-092-921; Miltenyi Biotec; T cell marker) and mouse anti-human CD69-APC antibodies (cat. no. 340560; BD Biosciences). This allows the determination of CD69 upregulation on T cells, which is a measure of T cell activation. Details of such an assay are described in Example 10. Proteins according to the invention, such as antibodies targeting human CD3, can prevent or highly reduce CD69 upregulation compared to an IgG1 antibody targeting human CD3 with a wild-type-like Fc region.

[0074] Concerning the proteins according to the invention, i.e. proteins comprising the hinge, CH2 and CH3 regions of the human IgG1 immunoglobulin heavy chain at least substituted with amino acids corresponding to the amino acids at positions L234, L235 and G236, preferably F, E and R, respectively, they preferably have glycosylation very similar to the wild-type IgG1 sequence. More specifically, the galactosylation and / or the presence of charged glycans of the proteins according to the invention is preferably in the same range as found for the wild-type IgG1 amino acid sequence produced using the same cell line and the same production conditions. Mammalian host cells suitable for production, including CHO cell lines, are well known in the art (see, inter alia, Butler and Spearman, 2014, Curr Opin Biotechno, Dec;30:107-12). The total galactosylation rate is preferably in the range of plus or minus 20% compared to the sum of the galactosylation rates observed in the same protein comprising the wild-type IgG1 sequence, such as SEQ ID NO:1. For example, if the galactosylation rate of a wild-type IgG1 sequence, etc. is 25%, then the total galactosylation rate can be in the range of 5%-45%. The total galactosylation rate is preferably in the range of plus or minus 20% compared to the total galactosylation rate observed in a protein according to the invention that does not contain the FER format. The total percentage of charged glycans is preferably in the range of plus or minus 3% compared to the total percentage of charged glycans observed in a protein comprising a wild-type IgG1 sequence, such as SEQ ID NO: 1. For example, if a wild-type IgG1 sequence, etc., has a percentage of charged glycans of 1%, then the total percentage of charged glycans can be in the range of 0%-4%. The total percentage of charged glycans is preferably in the range of plus or minus 3% compared to the total percentage of charged glycans observed in a protein according to the invention that does not contain the FER format. The total percentage of charged glycans and / or the percentage of galactosylation of a protein according to the invention, such as an antibody, is preferably within the range of ±3% and ±20% of the total percentage of charged glycans compared to the total percentage of galactosylation observed for the same protein comprising a wild-type IgG1 sequence, such as SEQ ID NO:1.The total percentage of charged glycans and / or galactosylation of a protein according to the invention, such as an antibody, is preferably within the range of ±3% and ±20% of the total percentage of charged glycans compared to the total percentage of galactosylation observed in a protein according to the invention that does not contain the FER format.

[0075] The percentage of galactosylation and / or charged glycans of a protein according to the invention, such as an antibody, can be determined using methods known in the art. Such methods are described, for example, in Example 14. Suitable methods include 2-aminobenzamide labeling followed by HPLC analysis as described in Example 14, or LC-MC using an Orbitrap Q-Extractive Pluss mass spectrometer. Here, the percentage of galactosylation and charged glycans is calculated as the occupancy of galactose or charged glycans in the oligosaccharide relative to all glycans having the A2F glycan structure, respectively. The amounts referred to as percentages herein refer to molar amounts, i.e., molecular numbers rather than masses. The percentage of charged glycans and / or galactosylation can be determined for a protein according to the invention, such as an antibody, when produced in Expi293F cells. For example, as shown in the Examples section, the percentage of charged glycans and the percentage of galactosylation of a protein produced in Expi293F cells with a wild-type IgG1 sequence are about 0.5% and about 15%-25%, respectively. Thus, the protein according to the invention, when produced in Expi293F cells, preferably has a percentage of charged glycans and a percentage of galactosylation, which is preferably 0-4% to 5-45%, respectively.

[0076] With regard to the proteins according to the invention comprising the hinge, CH2 and CH3 regions of a human IgG1 immunoglobulin heavy chain with substitutions of amino acids corresponding to at least the amino acids at positions L234, L235 and G236, preferably with F, E and R, respectively, these preferably have human FcRn binding similar to that of the wild-type human IgG1 Fc region. It is understood that the substitutions selected herein may be substitutions that do not affect the FcRn binding function. Thus, the pharmacokinetics of such proteins are similar to those of the corresponding proteins with the wild-type IgG1 Fc region, for example as described, inter alia, in the Examples section. However, it is understood that in scenarios where it is useful to modify the FcRn function, for example to extend or shorten the half-life, it may be contemplated to include further substitutions for that purpose. For antibodies, for example full-length monospecific and bispecific antibodies, in one embodiment the human FcRn binding properties are not different from the same antibodies with the wild-type human IgG1 Fc region. Such binding properties are known in the art and can be determined as described in the Examples section herein. Thus, FcRn binding at pH 6.0 occurs similarly to that observed for the corresponding wild-type human IgG1 Fc region, with no detectable binding at pH 7.4.

[0077] In a further embodiment, the protein according to the invention comprising said first and second polypeptides has the same amino acid sequence. This includes proteins produced from a single expression cassette encoding one polypeptide in a host cell, i.e. it is understood that the first and second polypeptides may be homodimers of that one polypeptide. Examples of such proteins include antibodies, such as antibodies having two heavy chains and two light chains (see FIG. 15A), both of which are identical, as are both of the light chains. Such antibodies are bivalent, having two binding regions, each capable of binding to the same target antigen, i.e. the same epitope. Thus, in a further embodiment, the protein of the invention comprises a first and a second polypeptide, which are immunoglobulin heavy chains with identical amino acid sequences, and further comprise first and second immunoglobulin light chains with identical amino acid sequences.

[0078] In a further embodiment, the protein according to the invention comprises further substitutions. Preferred further substitutions according to the invention include modifications that allow the formation of heterodimers, i.e. to provide a protein comprising a first and a second polypeptide, the first and second polypeptides being different. Examples of such proteins include bispecific antibodies, e.g. antibodies with two heavy chains and two light chains (see FIG. 15B), where at least two heavy chains are not identical and each pair of heavy and light chains in the antibody can target a different antigen. It may not be necessary to introduce differentiation modifications in the hinge, CH2 and CH3 regions of the IgG1 sequences of the first and second polypeptides, where these regions of the first and second polypeptide sequences can be identical. In that way, a mixture of homodimers and heterodimers can be formed, the use of which may be advantageous in itself, or the heterodimers and homodimers can be easily separated from the mixture (e.g. via differences in size and / or charge, affinity, etc.). However, it is preferred to introduce further modifications that allow or drive the generation of bispecific antibodies and the like. Thus, the protein according to the invention may comprise further substitutions in the hinge, CH2 and CH3 regions of the IgG1 sequence of the first and second polypeptide. In this way, first and second polypeptides which comprise different sequences with respect to the hinge, CH2 and CH3 regions of the IgG1 sequence can be advantageously combined in a protein according to the invention.Examples of such proteins, i.e. immunoglobulins or immunoglobulin-like proteins with such substitutions, include Triomab / Quadroma (Trion Pharma / Fresenius Biotech; Roche, WO 2011069104), Knobs-into-Holes (Genentech, WO 9850431), Cross-MAb (Roche, WO 2011117329) and electrostatically matched (Amgen, EP 1870459 and WO 2009089004; Chugai, US 201000155133; Oncomed, WO 2010129304), LUZ-Y (Genentech), DIG bodies and PIG bodies (Pharmabcine), strand-exchange engineered domain bodies (SEED bodies) (EMD Examples of such antibodies include, but are not limited to, proteins with complementary CH3 domains such as Serono, WO 2007110205), Biclonics (Merus), FcΔAdp (Regeneron, WO 201015792), bispecific IgG1 and IgG2 (Pfizer / Rinat, WO 11143545), azimetric scaffolds (Zymeworks / Merck, WO 2012058768), mAb-Fv (Xencor, WO 2011028952), bivalent bispecific antibodies (Roche) and DuoBody (Genmab A / S, WO 2011131746).

[0079] In a particular embodiment, the protein according to the invention, said first and second polypeptides, comprises a further amino acid substitution, preferably an amino acid substitution selected from the group consisting of T366, L368, K370, D399, F405, Y407 and K409, e.g. F405L or K409R. It is understood that if the protein is a homodimer, e.g. having identical first and second polypeptides, both may have the same substitution. Such proteins according to the invention, e.g. monospecific antibodies, can be very advantageously used for the preparation of bispecific antibodies as described in the Examples section and e.g. in WO2011131746.

[0080] Thus, in a further embodiment, the protein according to the invention is a monospecific antibody. In the context of the present invention, "monospecific" refers to a protein that binds, i.e. is capable of binding, to the same epitope with its binding region. In particular, such a monospecific protein or antibody preferably binds to an antigen selected from target molecules, cellular targets and pathogens.

[0081] Contemplated target molecules include molecules such as cytokines, growth factors, ligands, etc. Cellular targets may be considered to include cell surface molecules such as receptors or adhesion molecules present on cancer cells, tumor cells, effector cells (e.g., macrophages, monocytes, NK cells, and T cells). Pathogens that may be targeted include viruses, bacteria, protozoa, parasites, etc. Thus, contemplated proteins according to the present invention include proteins, such as monospecific antibodies, having a binding region for an antigen or target selected from the group of cytokines, growth factors, ligands, cancer cells, tumor cells, effector cells, viruses, and bacteria.

[0082] However, the present invention is not limited to monospecific proteins such as monospecific antibodies, but also relates to multispecific proteins such as bispecific antibodies. In a further embodiment, the protein according to the invention is a bispecific or multispecific antibody. This means that the binding regions of the protein according to the invention do not bind to the same epitope, but to different epitopes. Thus, the first and second binding regions of the protein according to the invention are different, for example with respect to the amino acid sequence. The different epitopes may originate from different epitopes presented by the same target entity, for example presented by the same target molecule and / or presented by the same target cell, but may originate from different target entities.

[0083] Highly advantageous bispecific proteins, such as bispecific antibodies, according to the invention comprise proteins in which one of the first and second binding regions targets an effector cell and the other of the first and second binding regions targets a cancer antigen. Such multispecific antibodies can be used to engage a specific class of effector cells with cancer cells, thereby inducing, for example, the killing of the cancer cells by the effector cells. In one embodiment, a bispecific antibody according to the invention is provided, in which one of the binding regions binds to a cancer antigen. In another embodiment, one of the binding regions binds to an effector cell, such as a T cell, a NK cell, a macrophage, a dendritic cell-calling, a monocyte or a neutrophil. In yet another embodiment, one of the binding regions binds to an effector cell, such as a T cell or a NK cell, and the other binding region binds to a cancer antigen.

[0084] There are applications where Fc binding of the Fc region of a therapeutic antibody to effector cells or complement is unnecessary or undesirable as it may contribute to undesirable cytotoxicity, such as receptor inhibition or T cell recruitment by bispecific antibodies. Thus, bispecific antibodies that bind with one binding region to the human T cell receptor can advantageously recruit human cytotoxic T cells. Thus, bispecific antibodies are provided herein that bind with one binding region to human CD3 and can recruit cytotoxic T cells. CD3 antibodies, including bispecific antibodies with an activating IgG Fc region, can induce undesirable agonism in the absence of tumor cells through crosslinking by FcγR expressing cells, inappropriate activation of FcγR expressing cells and subsequent cytokine storm and associated toxic effects, or platelet aggregation. Thus, CD3 bispecific antibodies with a non-activating Fc region are advantageous in preventing potential undesirable cell activation.

[0085] Thus, in one embodiment, at least one of the first and second binding regions binds to CD3, i.e. is capable of binding to CD3. In a particular embodiment, the first binding region binds to CD3 and the second binding region binds to, i.e. is capable of binding to, any other target of interest. Such other target may be a cancer antigen. Such other target may be a tumor-specific target or a cancer-specific target. Preferably, the protein according to the invention is a bispecific antibody. The protein bispecific antibody preferably has a first binding region capable of binding to CD3 and a second binding region capable of binding to a cancer-specific target.

[0086] It is understood that antibodies, including monospecific, bispecific or multispecific antibodies, may comprise, for example, an Fc region comprising the hinge region, CH2 and CH3 regions of a human IgG1 antibody according to the invention. Although antibody formats, such as, inter alia, hereinafter will not comprise such an IgG1 Fc region, such antibodies may be provided therewith, for example, by replacing the Fc region of such antibodies with a hinge region, CH2 and CH3 region comprising FER according to the invention, or, if such antibodies do not comprise an Fc region, by providing such antibodies therewith, for example, by fusion and / or conjugation. Thus, any antibody format may be contemplated according to the invention, as long as the antibody comprises the hinge region, CH2 and CH3 region of a human IgG1 antibody according to the invention.

[0087] Examples of bispecific antibody molecules that can be used in the present invention include (i) single antibodies with two arms containing different antigen-binding regions, (ii) single chain antibodies with specificity for two different epitopes, for example, via two scFvs linked in tandem by an extra peptide linker; (iii) dual variable domain antibodies (DVD-Ig), in which each light and heavy chain contains two variable domains in tandem via a short peptide linkage (Wu et al., Generation and Characterization of a Dual Variable Domain Immunoglobulin (DVD-Ig™) Molecule, In: Antibody Engineering, Springer Berlin Heidelberg (2010)); (iv) chemically linked bispecifics (Fab'). 2 (v) Tandabs, which are the fusion of two single chain diabodies resulting in a tetravalent bispecific antibody with two binding sites for each of the target antigens; (vi) Flexa, which is a combination of an scFv and a diabody resulting in a multivalent molecule; (vii) the so-called "dock and lock" molecules, based on the "dimerization and docking domain" of protein kinase A, which when applied to Fabs can result in a trivalent bispecific binding protein consisting of two identical Fab fragments linked to different Fab fragments; (viii) the so-called Scorpion molecules, which contain, for example, two scFvs fused to both ends of a human Fab arm; and (ix) diabodies.

[0088] In one embodiment, a bispecific antibody of the invention is a diabody, crossbody or bispecific antibody obtained via controlled Fab arm exchange as described in the present invention (e.g. as described in WO 11 / 131746).

[0089] Examples of different classes of bispecific antibodies include, but are not limited to, (i) IgG-like molecules with complementary CH3 domains that force heterodimerization; (ii) recombinant IgG-like dual targeting molecules, where each side of the molecule contains a Fab fragment or a portion of a Fab fragment of at least two different antibodies; (iii) IgG fusion molecules, where a full-length IgG antibody is fused to an extra Fab fragment or a portion of a Fab fragment; (iv) Fc fusion molecules, where a single chain Fv molecule or a stabilized diabody is fused to a heavy chain constant domain, an Fc region or a portion thereof; (v) Fab fusion molecules, where different Fab fragments are fused together and fused to a heavy chain constant domain, an Fc region or a portion thereof; (vi) ScFv and diabody systems and heavy chain antibodies (e.g., domain antibodies, nanobodies) where different single chain Fv molecules or different diabodies or different heavy chain antibodies (e.g., domain antibodies, nanobodies) are fused to each other or to another protein or carrier molecule fused to a heavy chain constant domain, an Fc region or a portion thereof.

[0090] Examples of bispecific IgG-like molecules with complementary CH3 domain molecules include triomabs / quadromas (Trion Pharma / Fresenius Biotech; Roche, WO 2011069104), knobs-into-holes (Genentech, WO 9850431), cross-MAbs (Roche, WO 2011117329) and electrostatically matched (Amgen, EP 1870459 and WO 2009089004; Chugai, US 201000155133; Oncomed, WO 2010129304), LUZ-Y (Genentech), DIG bodies and PIG bodies (Pharmabcine), strand-exchange engineered domain bodies (SEED bodies) (EMD Serono, WO 2007110205), Biclonics (Merus), FcΔAdp (Regeneron, WO 201015792), bispecific IgG1 and IgG2 (Pfizer / Rinat, WO 11143545), azimetric scaffolds (Zymeworks / Merck, WO 2012058768), mAb-Fv (Xencor, WO 2011028952), bivalent bispecific antibodies (Roche) and DuoBody (Genmab A / S, WO 2011131746).

[0091] Examples of recombinant IgG-like dual targeting molecules include, but are not limited to, Dual Targeting (DT)-Ig (GSK / Domantis), Two-in-one Antibody (Genentech), Cross-linked Mab (Karmanos Cancer Center), mAb 2 (F-Star, WO2008003116), Zybodies (Zyngenia), the common light chain approach (Crucell / Merus, U.S. Pat. No. 7,262,028), κλBodies (NovImmune) and CovX-body (CovX / Pfizer).

[0092] Examples of IgG fusion molecules include, but are not limited to, dual variable domain (DVD)-Ig (Abbott, US Patent No. 7,612,181), dual domain double-headed antibodies (Unilever; SanofiAventis, WO20100226923), IgG-like bispecifics (ImClone / Eli Lilly), Ts2Ab (MedImmune / AZ) and BsAb (Zymogenetics), HERCURES (BiogenIdec, US Patent No. 007951918), scFv fusions (Novartis), scFv fusions (Changzhou Adam Biotech Inc, China Patent No. 102250246) and TvAb (Roche, WO2012025525, WO2012025530).

[0093] Examples of Fc fusion molecules include ScFv / Fc Fusions (Academic Institution), SCORPION (Emergent BioSolutions / Trubion, Zymogenetics / BMS), Dual Affinity Retargeting Technology (Fc-DART) (MacroGenics, WO 2008157379, WO 2010 / 080538) and Dual(ScFv) 2 -Fab (National Research Center for Antibody Medicine-China).

[0094] An example of a Fab-fused bispecific antibody is F(ab) 2 (Medarex / AMGEN), Dual-Action or Bis-Fab (Genentech), Dock-and-Lock (DNL) (ImmunoMedics), Bivalent Bispecific (Biotecnol) and Fab-Fv (UCB-Celltech).

[0095] Examples of ScFv systems, diabody systems and domain antibodies include, but are not limited to, bispecific T cell engagers (BiTEs) (Micromet, tandem diabodies (Tandab) (Affimed), dual affinity retargeting technology (DART) (MacroGenics), single chain diabodies (Academic), TCR-like antibodies (AIT, ReceptorLogics), human serum albumin ScFv fusions (Merrimack) and COMBODY (Epigen Biotech), dual targeting nanobodies (Ablynx), dual targeting heavy chain only domain antibodies.

[0096] In a further embodiment, a bispecific antibody according to the invention is provided, wherein said first and second polypeptides comprise further substitutions in their respective CH2 and CH3 regions such that the sequences of the respective CH2 and CH3 regions from said first and second polypeptides are different, said substitutions making it possible to obtain said polypeptide comprising said first and second polypeptides. In a particular embodiment, in said first polypeptide at least one amino acid at a position corresponding to a position selected from the group consisting of T366, L368, K370, D399, F405, Y407 and K409 of a human IgG1 heavy chain is substituted and in said second polypeptide at least one amino acid at a position corresponding to a position selected from the group consisting of T366, L368, K370, D399, F405, Y407 and K409 of a human IgG1 heavy chain is substituted, said substitutions in said first and said second polypeptide not being at the same position. In this context, the term "substituted" refers to an amino acid at a particular amino acid position which has been replaced with an amino acid of another type. Thus, a "substituted" amino acid at a position that corresponds to a position in a human IgG1 heavy chain means that the amino acid at the particular position is different from the naturally occurring amino acid at that position in an IgG1 heavy chain.

[0097] Further provided is a bispecific antibody according to the present invention, wherein in said first polypeptide at least one amino acid at a position corresponding to a position selected from the group consisting of T366, L368, K370, D399, F405, Y407 and K409 in a human IgG1 heavy chain is substituted, and in said second polypeptide at least one amino acid at a position corresponding to a position selected from the group consisting of T366, L368, K370, D399, F405, Y407 and K409 in a human IgG1 heavy chain is substituted, and said substitutions in said first and said second polypeptide are not at the same position.

[0098] In a further embodiment, the amino acid at the position corresponding to F405 of a human IgG1 heavy chain is L in said first polypeptide and the amino acid at the position corresponding to K409 of a human IgG1 heavy chain is R in said second polypeptide, or vice versa.

[0099] In another embodiment a bispecific antibody according to the invention is provided, wherein the amino acid at the position corresponding to F405 is L in said first polypeptide and the amino acid at the position corresponding to K409 is R in said second polypeptide, or vice versa. In another embodiment a bispecific antibody according to the invention is provided, wherein in said first polypeptide the amino acids at the positions corresponding to F405 and K409 are L and K, respectively, and in said second polypeptide the amino acids at the positions corresponding to F405 and K409 are F and R, respectively, or vice versa.

[0100] In yet another embodiment a bispecific antibody according to the invention is provided, wherein said bispecific antibody has in both said first and second polypeptide a modification comprising substitution of amino acids at positions L234, L235 and G236 by F, E and R and substitution of amino acid at positions F405 by L in said first polypeptide and K409 by R in said second polypeptide, or vice versa.

[0101] In one embodiment a bispecific antibody according to the invention is provided, wherein said bispecific antibody has in one of said first and second polypeptides a modification comprising substitution of amino acids at positions L234, L235 and G236 by F, E and R and substitution of amino acid at positions F405 by L in said first polypeptide and K409 by R in said second polypeptide, or vice versa.

[0102] In yet another embodiment a bispecific antibody according to the invention is provided, wherein said bispecific antibody has modifications in said first and second polypeptide comprising substitutions of amino acids at positions L234, L235 and G236 in the first second polypeptide by F, E and R, and substitutions of amino acids at positions L234, L235 and D265 in said second polypeptide by F, E and A, and substitution of the amino acid at position F405 in said first polypeptide by L, and substitution of K409 in said second polypeptide by R.

[0103] In yet another embodiment a bispecific antibody according to the invention is provided, wherein said bispecific antibody has modifications in said first and second polypeptides comprising substitutions of amino acids at positions L234, L235 and G236 in the first second polypeptide by F, E and R, and substitutions of amino acids at positions L234, L235 and D265 in said second polypeptide by F, E and A, and substitution of the amino acid at position F405 in said second polypeptide by L, and substitution of K409 in said first polypeptide by R.

[0104] In yet another embodiment a bispecific antibody according to the invention is provided, wherein said bispecific antibody has in both said first and second polypeptide a modification comprising substitution of amino acids at positions L234, L235 and G236 by F, E and R and substitution of amino acid at positions F405 by L in said first polypeptide and K409 by R in said second polypeptide, or vice versa.

[0105] In one embodiment a bispecific antibody according to the invention is provided, wherein said bispecific antibody has in one of said first and second polypeptides a modification comprising substitution of amino acids at positions L234, L235 and G236 by F, E and R and substitution of amino acid at positions F405 by L in said first polypeptide and K409 by R in said second polypeptide, or vice versa.

[0106] In yet another embodiment a bispecific antibody according to the invention is provided, wherein said bispecific antibody has modifications in said first and second polypeptide comprising substitutions of amino acids at positions L234, L235 and G236 in the first second polypeptide by F, E and R, and substitutions of amino acids at positions L234, L235 and D265 in said second polypeptide by F, E and A, and substitution of the amino acid at position F405 in said first polypeptide by L, and substitution of K409 in said second polypeptide by R.

[0107] In yet another embodiment a bispecific antibody according to the invention is provided, wherein said bispecific antibody has modifications in said first and second polypeptides comprising substitutions of amino acids at positions L234, L235 and G236 in the first second polypeptide by F, E and R, and substitutions of amino acids at positions L234, L235 and D265 in said second polypeptide by F, E and A, and substitution of the amino acid at position F405 in said second polypeptide by L, and substitution of K409 in said first polypeptide by R.

[0108] The protein according to the invention can be based on the hinge, CH2 and CH3 regions of human IgG1 as defined in SEQ ID NO: 1. The hinge, CH2 and CH3 regions of human IgG1 correspond to the allotype IgG1m(f). Of course, any other human IgG1 allotype within the IgG1 immunoglobulin class (e.g. IgG1m(za), IgG1m(zax), IgG1m(zav) or IgG1m(fa); see inter alia Vidarsson et al., 2014, Front. Immunol., 20 Oct. and as provided in the IMGT database (www.imgt.org)) can be envisaged and is equally suitable.

[0109] The Fc region may have a lysine at its C-terminus. The source of this lysine is the naturally occurring sequence found in the human from which these Fc regions are derived. During cell culture production of recombinant antibodies, this terminal lysine may be proteolytically cleaved by endogenous carboxypeptidase(s), resulting in a constant region with the same sequence but lacking the C-terminal lysine. For antibody manufacturing purposes, the DNA encoding this terminal lysine can be omitted from the sequence such that the antibody is produced without the lysine. For example, when using antibodies produced in a CHO-based production system, the extent of processing of the terminal lysine is typically high, so that antibodies produced from nucleic acid sequences that do or do not encode the terminal lysine are substantially identical in sequence and function (Dick, L Wet al. Biotechnol. Bioeng. 2008;100:1132-1143). The constant region sequences listed herein list the terminal lysine (K) (see especially SEQ ID NO: 1), and sequences encoding the terminal lysine (K) were used in the Examples section herein. It is therefore understood that proteins according to the present invention, such as antibodies, can be produced without encoding or having terminal lysines as recited herein in SEQ ID NOs: 1-3, 5 and 9-14.

[0110] Thus, a protein according to the invention, or a bispecific antibody according to the invention, may comprise a first and a second polypeptide comprising an amino acid sequence as defined herein according to SEQ ID NO: 1, said first and second proteins having an amino acid substitution as defined herein. A protein according to the invention, or a bispecific antibody according to the invention, wherein said first and second polypeptides preferably comprise an amino acid sequence according to SEQ ID NO: 1, said amino acid sequences comprised in said first and second polypeptides having an amino acid substitution as defined herein. Furthermore, said amino acid sequence defined by SEQ ID NO: 1 having a substitution as defined herein may not comprise a terminal lysine.

[0111] Thus, a protein, which may be a full-length antibody or a monospecific or bispecific antibody according to the invention, may comprise the amino acid sequence defined in SEQ ID NO: 2. The protein according to the invention, in addition to comprising said substitutions of L234, L235 and G236 by F, E and R in the hinge, CH2 and CH3 region sequence, may comprise further substitutions therein. Preferably, the number of further substitutions is in the range of up to 5 further substitutions in the hinge, CH2 and CH3 region of a human IgG1 immunoglobulin heavy chain. Thus, in one embodiment, a protein comprising a sequence defined in SEQ ID NO: 2 comprises further substitutions in said sequence defined in SEQ ID NO: 2, the number of further substitutions consisting of up to 5 substitutions. In another embodiment, a protein according to the invention is provided which comprises a sequence defined in SEQ ID NO: 2 and comprises further substitutions in said sequence defined in SEQ ID NO: 2, the number of further substitutions consisting of up to 10 substitutions. In another embodiment, the protein according to the invention may comprise the sequence defined in SEQ ID NO: 1 or the corresponding sequence of a human IgG1 of another allotype, in which within the hinge, CH2 and CH3 region sequences, the substitutions of L234, L235 and G236 by F, E and R may be provided and may also comprise further substitutions, for example up to 5 further substitutions. Examples of such proteins comprising the polypeptide defined in SEQ ID NO: 2 with further substitutions, highly suitable for having further substitutions, such as with an R / L substitution as defined herein, are as defined by the amino acid sequences defined in SEQ ID NO: 11 and 12. Furthermore, the amino acid sequences defined by SEQ ID NO: 2 or 11 and 12 with any further substitutions as defined herein may be missing the terminal lysine.

[0112] Further provided is a protein according to the invention which may be an antibody or a full length antibody as defined herein, wherein both the first and second polypeptides comprise the amino acid sequence as defined in SEQ ID NO: 2, 11 or 12.

[0113] In another embodiment there is provided a protein which may be a bispecific antibody as defined herein, wherein the first and second polypeptides comprise the amino acid sequences defined in SEQ ID NOs: 2 and 3, respectively. In another embodiment there is provided a protein which may be a bispecific antibody as defined herein, wherein the first and second polypeptides comprise the amino acid sequences defined in SEQ ID NOs: 11 and 12, or 11 and 14, or 12 and 13, respectively. Such polypeptides may comprise a CH1 region adjacent to the hinge region, for example a human CH1 region defined by SEQ ID NO: 4.

[0114] In a further embodiment, the protein or monospecific or bispecific antibody provided according to the invention comprises an amino acid sequence having at least 85% sequence identity, such as at least 90% sequence identity, at least 95% sequence identity, at least 97% sequence identity, at least 98% sequence identity, at least 99% sequence identity, or having 100% sequence identity to the sequence defined in SEQ ID NO: 2, wherein the amino acid residues at positions corresponding to 234, 235 and 236 as defined by Eu numbering are F, E and R, respectively.

[0115] In a further embodiment, a protein or a monospecific or bispecific antibody provided according to the invention comprises an amino acid sequence having at least 85% sequence identity, such as at least 90% sequence identity, at least 95% sequence identity, at least 97% sequence identity, at least 98% sequence identity, at least 99% sequence identity, or having 100% sequence identity to the sequence defined in SEQ ID NO: 11, wherein the amino acid residues at positions corresponding to 234, 235 and 236 are F, E and R, respectively, and the amino acid residue at position corresponding to 405 is L, and wherein the amino acid numbers are as defined by Eu numbering.

[0116] In a further embodiment, a protein or a monospecific or bispecific antibody provided according to the invention comprises an amino acid sequence having at least 85% sequence identity, such as at least 90% sequence identity, at least 95% sequence identity, at least 97% sequence identity, at least 98% sequence identity, at least 99% sequence identity, or having 100% sequence identity to the sequence defined in SEQ ID NO: 12, wherein the amino acid residues at positions corresponding to 234, 235 and 236 are F, E and R, respectively, and the amino acid residue at position corresponding to 409 is R, and wherein the amino acid numbers are as defined by Eu numbering.

[0117] In some embodiments, a protein or bispecific antibody provided according to the invention comprises a first and a second polypeptide, wherein: the first polypeptide comprises an amino acid sequence having at least 85% sequence identity, e.g., at least 90% sequence identity, at least 95% sequence identity, at least 97% sequence identity, at least 98% sequence identity, at least 99% sequence identity, or 100% sequence identity, to the sequence set forth in SEQ ID NO:2, wherein the amino acid residues at positions corresponding to 234, 235, and 236 are F, E, and R, respectively; and the second polypeptide comprises an amino acid sequence having at least 85% sequence identity, e.g., at least 90% sequence identity, at least 95% sequence identity, at least 97% sequence identity, at least 98% sequence identity, at least 99% sequence identity, or 100% sequence identity, to the sequence set forth in SEQ ID NO:3, wherein the amino acid residues at positions corresponding to 234, 235, and 265 are F, E, and A, respectively; Amino acid numbers are as defined by Eu numbering.

[0118] In yet a further embodiment, the protein or bispecific antibody provided according to the invention comprises a first and a second polypeptide, wherein the first polypeptide comprises an amino acid sequence having at least 85% sequence identity, e.g., at least 90% sequence identity, at least 95% sequence identity, at least 97% sequence identity, at least 98% sequence identity, at least 99% sequence identity, or 100% sequence identity, to the sequence set forth in SEQ ID NO:11, wherein the amino acid residues at positions corresponding to 234, 235, and 236 are F, E, and R, respectively, and the amino acid residue at position corresponding to 405 is L; and the second polypeptide comprises an amino acid sequence having at least 85% sequence identity, e.g. at least 90% sequence identity, at least 95% sequence identity, at least 97% sequence identity, at least 98% sequence identity, at least 99% sequence identity, or 100% sequence identity, to the sequence set forth in SEQ ID NO: 12, wherein the amino acid residues at positions corresponding to 234, 235 and 236 are F, E and R, respectively, and the amino acid residue at position corresponding to 409 is R; Amino acid numbers are as defined by Eu numbering.

[0119] In yet a further embodiment, the protein or bispecific antibody provided according to the invention comprises a first and a second polypeptide, wherein the first polypeptide comprises an amino acid sequence having at least 85% sequence identity, e.g., at least 90% sequence identity, at least 95% sequence identity, at least 97% sequence identity, at least 98% sequence identity, at least 99% sequence identity, or 100% sequence identity, to the sequence set forth in SEQ ID NO:12, wherein the amino acid residues at positions corresponding to 234, 235, and 236 are F, E, and R, respectively, and the amino acid residue at position corresponding to 409 is R; and the second polypeptide comprises an amino acid sequence having at least 85% sequence identity, e.g., at least 90% sequence identity, at least 95% sequence identity, at least 97% sequence identity, at least 98% sequence identity, at least 99% sequence identity, or 100% sequence identity, to the sequence set forth in SEQ ID NO: 13, wherein the amino acid residues at positions corresponding to 234, 235 and 236 are F, E and A, respectively, and the amino acid residue at position corresponding to 405 is L; Amino acid numbers are as defined by Eu numbering.

[0120] In yet a further embodiment, the protein or bispecific antibody provided according to the invention comprises a first and a second polypeptide, wherein the first polypeptide comprises an amino acid sequence having at least 85% sequence identity, e.g., at least 90% sequence identity, at least 95% sequence identity, at least 97% sequence identity, at least 98% sequence identity, at least 99% sequence identity, or 100% sequence identity, to the sequence set forth in SEQ ID NO:11, wherein the amino acid residues at positions corresponding to 234, 235, and 236 are F, E, and R, respectively, and the amino acid residue at position corresponding to 405 is L; and the second polypeptide comprises an amino acid sequence having at least 85% sequence identity, e.g. at least 90% sequence identity, at least 95% sequence identity, at least 97% sequence identity, at least 98% sequence identity, at least 99% sequence identity, or 100% sequence identity, to the sequence set forth in SEQ ID NO: 14, wherein the amino acid residues at positions corresponding to 234, 235 and 265 are F, E and A, respectively, and the amino acid residue at position corresponding to 409 is R; Amino acid numbers are as defined by Eu numbering.

[0121] The sequence identified herein as SEQ ID NO:2 comprises the hinge, CH2 and CH3 regions of human IgG1 allotype G1m(f), in which L234, L235 and G236 have been replaced with F, E and R. The corresponding sequences of the constant regions (each including the CH1 region) of other allotypes of human IgG1 with said substitutions by F, E and R at L234, L235 and G236 in the hinge, their CH2 and CH3 region sequences are provided as SEQ ID NO: 27 (CH1, hinge, CH2 and CH3 regions of human IgG1 allotype G1m(fa) with a FER substitution), SEQ ID NO: 29 (CH1, hinge, CH2 and CH3 regions of human IgG1 allotype G1m(za) with a FER substitution), SEQ ID NO: 31 (CH1, hinge, CH2 and CH3 regions of human IgG1 allotype G1m(zav) with a FER substitution), SEQ ID NO: 33 (CH1, hinge, CH2 and CH3 regions of human IgG1 allotype G1m(zax) with a FER substitution). In another embodiment, there is provided a nucleic acid encoding said first or second polypeptide as defined herein, wherein both said first and second polypeptides comprise said substitutions of amino acids corresponding to amino acids L234, L235 and G236, most preferably said substitutions at positions L234, L235 and G236 are F, E and R, respectively. In another further embodiment, there is provided a nucleic acid encoding said first or second polypeptide as defined herein, wherein both said first or second polypeptides comprise said substitutions of amino acids corresponding to amino acids L234, L235 and G236, preferably said substitutions at positions L234, L235 and G236 are F, E and R, respectively.

[0122] In one embodiment, a nucleic acid is provided encoding a first or second polypeptide, the first or second polypeptide comprising an amino acid sequence defined by SEQ ID NO: 2, 11, or 12. Such a nucleic acid may further encode a polypeptide comprising a CH1 region adjacent to the hinge region, e.g., a CH1 region defined by SEQ ID NO: 4. Preferably, the nucleic acid encodes an immunoglobulin heavy chain. Most preferably, such an immunoglobulin heavy chain comprises a human immunoglobulin variable region or a humanized immunoglobulin variable region. Such nucleic acids encoding the first or second polypeptide may have a terminal lysine deleted from the coding sequence. These nucleic acids may be combined with nucleic acids encoding immunoglobulin light chains, etc.

[0123] In one embodiment, there is provided a method of producing a construct for producing a protein according to the invention having a non-activated Fc region, the protein comprising first and second polypeptides having an Fc region including a hinge region, CH2 and CH3 regions as defined in SEQ ID NO:1, the method comprising: a) providing one or more constructs for expression of a first and / or a second polypeptide comprising a nucleic acid sequence encoding a first and / or a second polypeptide sequence comprising a hinge region, a CH2 region and a CH3 region; b) modifying the nucleic acid sequence encoding amino acids corresponding to amino acids L234, L235, and G236 in the hinge, CH2, and CH3 regions such that they encode F, E, and R, respectively; c) thereby providing one or more constructs for producing a protein with an inactivated Fc region; Includes.

[0124] In another embodiment, a method is provided for producing a construct for producing a protein having an inactivated Fc region, the protein comprising first and second polypeptides having an Fc region comprising a hinge region, a CH2 and a CH3 region, such as that defined in SEQ ID NO:1, the method comprising: a) providing a nucleic acid sequence encoding a hinge region, a CH2 region and a CH3 region, such as that defined in SEQ ID NO:1; b) modifying the nucleic acid sequence encoding amino acids corresponding to amino acids L234, L235, and G236 in the hinge, CH2, and CH3 regions such that they encode F, E, and R, respectively; c) generating a construct using the modified sequence obtained in step b) for the expression of a first and / or second polypeptide of the protein according to the invention, comprising a nucleic acid sequence encoding said first and / or second polypeptide comprising a hinge region, a CH2 region and a CH3 region, wherein the amino acids L234, L235 and G236 in the hinge region, the CH2 region and the CH3 region have been modified to code for F, E and R, respectively, thereby providing a construct for producing a protein according to the invention with an inactivated Fc region; Includes.

[0125] In a further embodiment, a method is provided for providing a construct for producing an antibody having a non-activated Fc region, the antibody comprising a heavy chain having an Fc region including a hinge region, CH2 and CH3 regions as defined in SEQ ID NO:1, the method comprising: a) providing a construct for expression of an antibody heavy chain comprising a nucleic acid sequence encoding an antibody heavy chain sequence comprising a hinge region, a CH2 region and a CH3 region; b) modifying the nucleic acid sequences encoding amino acids corresponding to amino acids L234, L235, and G236 in the hinge, CH2, and CH3 regions such that they encode F, E, and R, respectively; c) thereby providing a construct for producing an antibody with a non-activated Fc region; Includes.

[0126] In yet another embodiment, a method is provided for providing a construct for producing an antibody having a non-activated Fc region, the antibody comprising a heavy chain having an Fc region including a hinge region, CH2 and CH3 regions as defined in SEQ ID NO:1, the method comprising: a) providing a nucleic acid sequence encoding a hinge region, a CH2 region and a CH3 region as defined in SEQ ID NO:1; b) modifying the nucleic acid sequences encoding amino acids corresponding to amino acids L234, L235, and G236 in the hinge, CH2, and CH3 regions such that they encode F, E, and R, respectively; c) producing a construct using the modified sequence obtained in step b) for the expression of an antibody heavy chain comprising a nucleic acid sequence encoding an antibody heavy chain sequence comprising a hinge region, a CH2 region and a CH3 region, said hinge region, the CH2 region and the CH3 region having the amino acids L234, L235 and G236 modified to encode F, E and R, respectively, thereby providing a construct for producing an antibody with a non-activated Fc region; Includes.

[0127] In the method of providing the construct as above, it is understood that the polypeptide or heavy chain having an Fc region comprising a hinge region, a CH2 and a CH3 region can be any such Fc region comprising a hinge region, a CH2 and a CH3 region as defined herein in accordance with the present invention. Such a method is particularly useful for improving the safety profile of an antibody by introducing a FER substitution or for suppressing Fc-mediated effector function. Such a method is particularly useful for improving the safety profile of a protein, antibody, etc. by introducing said FER substitution. Such a method is also useful for suppressing Fc-mediated effector function of a protein or antibody, etc. by introducing a FER substitution. Such a method is particularly useful for improving the safety profile and suppressing Fc-mediated effector function of a protein, antibody, etc. by introducing a FER substitution. By the method, a nucleic acid for obtaining a protein having an Fc region comprising a hinge region, a CH2 and a CH3 region can be provided in accordance with the present invention.

[0128] Such nucleic acids are particularly useful for producing proteins according to the invention, for example antibodies comprising heavy and light chains comprising the first and second polypeptides according to the invention as defined herein. Such antibodies may be monospecific or bispecific. Thus, in another aspect, there is provided a nucleic acid encoding said first or second polypeptide according to the invention, for use in, for example, an expression vector encoding an antibody sequence. Thus, there is provided a host cell comprising such an expression vector, including a hybridoma capable of producing the antibody, and a method for producing such an antibody by culturing such a host cell or hybridoma under suitable conditions, whereby a protein according to the invention, such as an antibody, is produced and optionally recovered.

[0129] Thus, a host cell may be provided with a nucleic acid according to the invention, which is incorporated into an expression vector, such as those described in the Examples section. An expression vector in the context of the present invention may be any suitable vector, including chromosomal, non-chromosomal and synthetic nucleic acid vectors (nucleic acid sequences comprising an appropriate set of expression control elements). Examples of such vectors include derivatives of SV40, bacterial plasmids, phage DNA, baculovirus, yeast plasmids, vectors derived from combinations of plasmids and phage DNA, and viral or non-viral nucleic acid (RNA or DNA) vectors. In one embodiment, the antibody-encoding nucleic acid is expressed in a naked DNA or RNA vector, including, for example, a linear expression element (e.g., as described in Sykes and Johnston, Nat Biotech 17, 355-59 (1997)), a compacted nucleic acid vector (e.g., as described in U.S. Pat. No. 6,077,835 and / or WO 00 / 70087), a plasmid vector, such as pcDNA3.3 (described herein), pBR322, pUC19 / 18, or pUC118 / 119, a "middle" minimal size nucleic acid vector (e.g., as described in Schakowski et al., Mol Ther 3, 793-800 (2001)), or a CaPO 4 -These include precipitated nucleic acid vector constructs, such as precipitated constructs (e.g., as described in WO00 / 46147, Benvenisty and Reshef, PNAS USA 83, 9551-55 (1986), Wigler et al., Cell 14, 725 (1978), and Coraro and Pearson, Somatic Cell Genetics 7, 603 (1981)). Such nucleic acid vectors and methods for their use are well known in the art (see, e.g., U.S. Pat. Nos. 5,589,466 and 5,973,972).

[0130] In one embodiment, the vector is suitable for expression in bacterial cells. Examples of such vectors include expression vectors such as BlueScript (Stratagene), pIN vectors (Van Heeke & Schuster, J Biol Chem 264, 5503-5509 (1989)), pET vectors (Novagen, Madison WI), etc. The expression vector may also, or alternatively, be a vector suitable for expression in a yeast system. Any vector suitable for expression in a yeast system may be used. Suitable vectors include, for example, vectors containing constitutive or inducible promoters such as alpha factor, alcohol oxidase, and PGH (reviewed in F. Ausubel et al., ed. Current Protocols in Molecular Biology, Greene Publishing and Wiley InterScience New York (1987), and Grant et al., Methods in Enzymol 153, 516-544 (1987)).

[0131] The nucleic acid and / or vector may also include a nucleic acid sequence encoding a secretion / localization sequence capable of targeting a polypeptide, such as a nascent polypeptide chain, to the periplasmic space or cell culture medium. Such sequences are known in the art and include secretion leader or signal peptides, organelle targeting sequences (e.g., nuclear localization sequences, ER retention signals, mitochondrial import sequences, chloroplast import sequences), membrane localization / anchor sequences (e.g., stop transfer sequences, GPI anchor sequences), and the like.

[0132] In the expression vector of the present invention, the nucleic acid according to the present invention may comprise or be associated with any suitable promoter, enhancer and other expression promoting elements. Examples of such elements include a strong expression promoter (e.g., human CMV IE promoter / enhancer and RSV, SV40, SL3-3, MMTV and HIV LTR promoters), an effective poly(A) termination sequence, an origin of replication of the plasmid product in E. coli, an antibiotic resistance gene as a selection marker, and / or a convenient cloning site (e.g., a polylinker). The nucleic acid may also comprise an inducible promoter as opposed to a constitutive promoter such as CMV IE (the skilled artisan will recognize that such a term is an actual descriptor of the degree of gene expression under certain conditions).

[0133] Thus, a host cell is provided comprising a nucleic acid sequence encoding an antibody according to the invention, said antibody comprising an immunoglobulin heavy chain comprising at least the hinge, CH2 and CH3 regions of a human IgG1 immunoglobulin heavy chain, respectively, and comprising substitutions of amino acids at positions L234, L235 and G236 with F, E and R, respectively, and an immunoglobulin light chain. Such a host cell may be particularly useful for the production of monospecific antibodies having two identical heavy chains and two identical light chains comprising said first and second polypeptides as described herein. In particular, when two of such antibodies are provided in which the sequences of the heavy chains differ, allowing the exchange of arms, such two antibodies may be very advantageously used for the preparation of bispecific antibodies, as described, for example, in the Examples section.

[0134] As mentioned above, it may not be required that both the first and second polypeptides of the bispecific antibody have the same substitutions at F, E and R at positions L234, L235 and G236, but may have different substitutions, for example. This allows for convenient production of e.g. bispecific antibodies since it is not necessarily necessary to generate new constructs and / or cell lines each time in order to screen different combinations of antibodies that can be generated and / or developed as new products, thereby saving considerable time and effort.

[0135] Thus, in one embodiment there is provided a method for preparing a bispecific antibody according to the invention, the method comprising: a) a. an immunoglobulin heavy chain comprising at least the hinge region, the CH2 region, and the CH3 region of a human IgG1 immunoglobulin heavy chain comprising substitutions of amino acids at positions L234, L235, and G236 with F, E, and R, respectively; B an immunoglobulin light chain; providing a first antibody comprising: b) a. an immunoglobulin heavy chain comprising at least the hinge region, CH2 region, and CH3 region, respectively, of a human IgG1 immunoglobulin heavy chain; - Amino acid substitutions at positions L234, L235 and D265, preferably said substitutions being F, E and A, respectively. Or, - replacement of the amino acids at positions L234, L235 and G236 by F, E and R, respectively an immunoglobulin heavy chain comprising B an immunoglobulin light chain; providing a second antibody comprising: c) wherein the sequences of the first and second CH3 regions of each of the first and second antibodies are distinct and a heterodimeric interaction between the first and second CH3 regions is stronger than a homodimeric interaction of the first and second CH3 regions, respectively; d) incubating said first antibody with said second antibody under reducing conditions sufficient to allow cysteines in the hinge region to undergo disulfide bond isomerization; e) obtaining the bispecific antibody comprising the first immunoglobulin heavy chain and the first immunoglobulin light chain of the first antibody and the second immunoglobulin heavy chain and the second immunoglobulin light chain of the second antibody; Includes.

[0136] In a further embodiment, instead of bispecific antibodies, multispecific antibodies can be produced.

[0137] It is understood that the first and second antibodies provided in steps a) and b) are preferably monospecific antibodies. More preferably, said monospecific antibodies are full-length antibodies. Most preferably, said antibodies comprise human or humanized variable regions and have human constant regions comprising substitutions as defined herein. Exemplary first and second antibodies highly suitable for the above method are first and second antibodies comprising the amino acid sequences defined in SEQ ID NOs: 11 and 12, 11 and 14, or 12 and 13, respectively. Such methods for preparing bispecific antibodies are described inter alia in the examples herein and are also fully described in (Labrijn et al., 2014, Nature Protocols Oct;9(10):2450-63). Of course, other suitable differences as described herein may be envisaged in step c).

[0138] Pharmaceutical Compositions In one aspect, the invention provides a pharmaceutical composition comprising a protein, such as an antibody, as defined in any of the aspects and embodiments described herein, and a pharma- ceutically acceptable carrier.

[0139] Pharmaceutical compositions can be formulated with pharma- ceutically acceptable carriers or diluents and any other known adjuvants and excipients according to conventional techniques, such as those disclosed in Remington: The Science and Practice of Pharmacy, 19th Edition, Gennaro, Ed., Mack Publishing Co., Easton, PA, 1995.

[0140] Pharmaceutically acceptable carriers or diluents, as well as any other known adjuvants and excipients, should be suitable for the protein, variant or antibody of the invention and the selected mode of administration. Suitability of carriers and other components of pharmaceutical compositions is determined based on the lack of significant adverse effects on the desired biological properties of the selected compound or pharmaceutical composition of the invention (e.g., less than a substantial impact (10% or less relative inhibition, 5% or less relative inhibition, etc.)) on antigen binding.

[0141] Pharmaceutical compositions of the present invention may also include diluents, fillers, salts, buffers, detergents (e.g., non-ionic detergents such as Tween-20 or Tween-80), stabilizers (e.g., sugars or protein-free amino acids), preservatives, tissue fixatives, solubilizing agents, and / or other materials suitable for inclusion in a pharmaceutical composition.

[0142] The actual dosage level of the active ingredient in the pharmaceutical composition of the present invention can be varied to obtain an amount of the active ingredient effective to achieve the desired therapeutic response for a particular patient, composition, and mode of administration without being toxic to the patient. The dosage level selected will depend on various pharmacokinetic factors, including the activity of the particular composition of the present invention or its amide used, the route of administration, the time of administration, the rate of excretion of the particular compound used, the duration of treatment, other drugs, compounds and / or materials used in combination with the particular composition used, the age, sex, weight, condition, general health and previous medical history of the patient being treated, and similar factors well known in the medical arts.

[0143] Pharmaceutical compositions can be administered by any suitable route and mode. Suitable routes of administration of the proteins, variants or antibodies of the invention in vivo and in vitro are well known in the art and can be selected by the skilled artisan.

[0144] In one embodiment, the pharmaceutical compositions of the invention are administered parenterally.

[0145] The phrases "parenteral administration" and "administered parenterally" as used herein refer to modes of administration other than enteral and topical administration, usually by injection, and include epidermal, intravenous, intramuscular, intraarterial, intrathecal, intracapsular, intraorbital, intracardiac, intradermal, intraperitoneal, intratendinous, transtracheal, subcutaneous, subcuticular, intraarticular, subcapsular, subarachnoid, intraspinal, intracranial, intrathoracic, epidural and intrasternal injection and infusion.

[0146] In one embodiment, the pharmaceutical composition is administered by intravenous or subcutaneous injection or infusion.

[0147] Injectable pharmaceutical compositions must typically be sterile and stable under the conditions of manufacture and storage. The compositions may be formulated as solutions, microemulsions, liposomes, or other ordered structures suitable for high drug concentration. The carrier may be an aqueous or non-aqueous solvent or dispersion medium containing, for example, water, ethanol, polyol (e.g., glycerol, propylene glycol, polyethylene glycol, etc.), and suitable mixtures thereof, vegetable oils such as olive oil, and injectable organic esters such as ethyl oleate. Proper fluidity may be maintained, for example, by the use of a coating such as lecithin, by the maintenance of the required particle size in the case of dispersions, and by the use of surfactants. In many cases, it is preferable to include an isotonic agent, for example, a polyalcohol such as sugar, glycerol, mannitol, sorbitol, or sodium chloride, in the composition. Prolonged absorption of an injectable composition may be brought about by including an agent that delays absorption, for example, monostearate salts and gelatin, in the composition. Sterile injectable solutions may be prepared by incorporating the required amount of active compound in an appropriate solvent, with, for example, one or a combination of the ingredients listed above, as needed, followed by sterile microfiltration. Generally, dispersions are prepared by incorporating the active compound into a sterile vehicle containing a basic dispersion medium and the required other ingredients, e.g., from those enumerated above. In the case of sterile powders for the preparation of sterile injectable solutions, examples of preparation methods are vacuum drying and freeze-drying (lyophilization), which yields a powder of the active ingredient plus any additional desired ingredients from a previously sterile-filtered solution thereof.

[0148] Sterile injection can be prepared by incorporating the required amount of active compound into a suitable solvent with one or combination of the above-listed components as required, followed by sterilization precision filtration.Generally, dispersion is prepared by incorporating active compound into a sterile vehicle that contains a basic dispersion medium and other components required from the above-listed components.For the preparation of sterile powder for sterile injection, the example of preparation method is vacuum drying and freeze-drying (lyophilization), which obtains the powder of active ingredient + any additional desired components from its solution that has been previously sterilized and filtered.

[0149] therapeutic use In another aspect, the invention relates to a protein, such as an antibody, as defined in any aspect or embodiment described herein, or a pharmaceutical composition of the invention, for use as a medicament.

[0150] In another aspect, the invention relates to a protein, such as an antibody, as defined in any aspect or embodiment described herein, or a pharmaceutical composition of the invention, for use in the treatment of a disease.

[0151] In a further aspect, said treatment of a disease comprises the treatment of cancer, an infectious disease, an inflammatory disease, or an autoimmune disease. In another aspect, the present invention relates to a use, wherein the disease is cancer. It is understood that such a use highly preferably includes use in humans.

[0152] In another aspect, the invention relates to a method of treatment comprising administering to a human subject a protein, such as an antibody or a pharmaceutical composition of the invention as defined in any aspect or embodiment described herein.

[0153] In another aspect, the invention relates to a method of treatment comprising administering a protein, such as an antibody or a pharmaceutical composition of the invention as defined in any aspect or embodiment described herein to a human subject suffering from a disease. In a further aspect, the disease comprises cancer, an inflammatory, infectious disease or an autoimmune disease. In another aspect, the disease is cancer.

[0154] The proteins, variants, antibodies or pharmaceutical compositions of the invention can be used therapeutically as well when the immune effector functions of the antibody IgG1 Fc region as found in wild-type antibodies are undesirable. For example, the proteins, variants or antibodies can be administered to cells in culture, e.g., in vitro or ex vivo, or to a human subject, e.g., in vivo, to treat or prevent a disorder such as cancer, an infectious disease, an inflammatory or autoimmune disorder. As used herein, the term "subject" is typically a human who responds to the protein, variant, antibody, or pharmaceutical composition. A subject can include, for example, a human patient having a disorder that can be corrected or ameliorated by directly or indirectly modulating a target function or by causing cell death.

[0155] In another aspect, the present invention provides a method for treating or preventing a disorder, such as cancer, in which the recruitment of T cells contributes to the treatment or prevention, comprising administering a therapeutically effective amount of a protein, variant, antibody or pharmaceutical composition of the present invention to a subject in need thereof. For example, such a protein according to the present invention can engage cytotoxic T cells, such as a bispecific antibody targeting CD3 and a cancer antigen. Cells that overexpress tumor-specific targets are particularly good targets for such proteins, variants or antibodies of the present invention, since recruitment of T cells by one of the two binding regions of the protein, variant or antibody can trigger the cytotoxic activity of the T cells. This mechanism is usually difficult to obtain, since the induction of cytotoxic activity may not function properly in the elimination of cancer cells.

[0156] In another embodiment, the proteins, including the antibodies and bispecific antibodies according to the present invention as described herein, are conjugated to another molecule. Such proteins can be produced by chemically linking other molecules to the N- or C-terminal side of the protein or antibody or fragment thereof (see, for example, Antibody Engineering Handbook, edited by Osamu Kanemitsu, published by Chijin Shokan (1994)). Such conjugated antibody derivatives can also be produced by conjugation at internal residues or sugars, as appropriate. In a preferred embodiment, the proteins, including the antibodies and bispecific antibodies according to the present invention, are conjugated to a therapeutic molecule. Suitable therapeutic molecules can include, for example, nucleic acids such as aptamers, ribozymes, antisense molecules, or RNAi-inducing agents. Other therapeutic molecules that may be contemplated include cytotoxins, chemotherapeutic drugs, immunosuppressants, or radioisotopes. Such conjugates can be referred to as immunoconjugates. Immunoconjugates that include one or more cytotoxins can be referred to as immunotoxins.

[0157] Suitable therapeutic molecules for forming an immunoconjugate of the invention include taxol, cytochalasin B, gramicidin D, ethidium bromide, emetine, mitomycin, etoposide, tenoposide, vincristine, vinblastine, colchicine, doxorubicin, daunorubicin, dihydroxyanthracin dione, mitoxantrone, mithramycin, actinomycin D, glucocorticoids, procaine, tetracaine, lidocaine, propranolol and puromycin, antimetabolites (e.g., methotrexate, 6-mercaptopurine, 6-thioguanine, cytarabine, fludarabine, 5-fluorouracil, decarbazine, hydroxyurea, asparaginase, gemcitabine, cladribine), alkylating agents (e.g., mechlorethamine, thioepa, chloramphenicol, tetracaine, lidocaine, propranolol, and puromycin), Bucil, melphalan, carmustine (BSNU), lomustine (CCNU), cyclophosphamide, busulfan, dibromomannitol, streptozotocin, dacarbazine (DTIC), procarbazine, mitomycin C, cisplatin and other platinum derivatives, e.g., carboplatin), antibiotics (e.g., dactinomycin (formerly actinomycin), bleomycin, daunorubicin (formerly daunomycin), doxorubicin, idarubicin, mithramycin, mitomycin, mitoxantrone, plicamycin, anthramycin (AMC)), diphtheria toxin and related molecules (e.g., diphtheria A chain and its active fragments and hybrid molecules), ricin toxin (e.g., ricin A or deglycosylated ricin A chain toxin), cholera toxin, Shiga-like toxin (SLT I, SLT II, ​​SLT IIV), LT toxin, C3 toxin, Shiga toxin, pertussis toxin, tetanus toxin, soybean Bowman-Birk protease inhibitor, Pseudomonas exotoxin, allorin, saporin, modeccin, gelatin, abrin A chain, modeccin A chain, alpha-sarcin, Aleurites fordii proteins, dianthin proteins, Phytolacca americana proteins (PAPI, PAPII, and PAP S), momordica charantia inhibitor, curcin, crotin, sapaonariaofficinalis inhibitors, gelonin, mitogenin, restrictocin, phenomycin, or enomycin toxin.

[0158] The methods typically involve administering to the subject an amount of the protein, variant or antibody effective to treat or prevent the disorder.

[0159] The effective dosage and administration regimen of the protein, variant or antibody depends on the disease or condition being treated and can be determined by one skilled in the art.

[0160] For example, an "effective amount" for therapeutic use may be measured by its ability to stabilize disease progression. The ability of a compound to inhibit cancer may be evaluated, for example, in an animal model system predictive of efficacy in human tumors. Alternatively, this property of a composition may be evaluated by examining the ability of the protein, variant or antibody to inhibit cell proliferation or induce cytotoxicity by in vitro assays known to those skilled in the art. A therapeutically effective amount of a therapeutic compound, i.e., a therapeutic protein, variant, antibody or pharmaceutical composition according to the invention, may reduce tumor size or otherwise ameliorate symptoms in a subject.

[0161] In one embodiment, the protein, antibody or variant may be administered by maintenance therapy, for example at regular intervals over a defined period of time or until disease progression. The protein, antibody or variant may also be administered prophylactically to reduce the risk of developing cancer, delay the occurrence of events in the progression of cancer, and / or reduce the risk of recurrence when the cancer is in remission.

[0162] The proteins, variants, antibodies, or antibodies may also be administered prophylactically to reduce the risk of developing cancer, delay the occurrence of events in the progression of cancer, and / or reduce the risk of recurrence when the cancer is in remission.

[0163] Diagnostic Applications The non-activated proteins of the present invention may also be useful for diagnostic purposes, using compositions comprising the proteins described herein. Thus, the present invention provides diagnostic methods and compositions using the proteins described herein. Such methods and compositions can be used for purely diagnostic purposes, for example, for detecting or identifying disease, as well as for monitoring the progress of therapeutic treatment, monitoring disease progression, assessing the condition after treatment, monitoring disease recurrence, assessing the risk of developing disease, etc. By using such proteins according to the present invention, this makes it possible to avoid any undesirable effects exerted by the Fc region, which may, for example, interfere with diagnostic applications.

[0164] In one aspect, the protein of the present invention is used ex vivo, such as to diagnose a disease, where cells expressing a particular target of interest and to which the protein binds are indicative of the disease or involved in pathogenesis, by detecting the level of the target of interest or cells expressing the target of interest on the cell surface in a sample taken from a patient. This can be achieved, for example, by contacting the sample to be tested, optionally together with a control sample, with the protein according to the present invention under conditions that allow binding of the protein to the target. Complex formation can then be detected (e.g., using ELISA). When a control sample is used along with the test sample, the levels of the protein or protein-target complex are analyzed in both samples, and a statistically significant higher level of the protein or protein-target complex in the test sample indicates a higher level of the target in the test sample compared to the control sample.

[0165] Examples of conventional immunoassays in which the proteins of the present invention can be used include, but are not limited to, ELISA, RIA, FACS assays, plasmon resonance assays, chromatographic assays, tissue immunohistochemistry, Western blots and / or immunoprecipitations.

[0166] In one embodiment, the invention provides a method for detecting the presence of a target or a cell expressing a target in a sample, comprising the steps of: - contacting the sample with a protein of the invention under conditions that allow binding of the protein to a target in the sample; - analyzing whether a complex is formed; and Typically the sample is a biological sample.

[0167] In one embodiment, the sample is a tissue sample known or suspected to contain a particular target and / or cells expressing the target. For example, in situ detection of target expression can be achieved by removing a histological specimen from a patient and providing such specimen with a protein of the invention. The protein can be provided by applying the protein to the specimen or by overlaying the protein on the specimen, and then detected using appropriate means. It is then possible to determine not only the presence of the target or target-expressing cells, but also the distribution of the target or target-expressing cells in the examined tissue (e.g., in the context of assessing the spread of cancer cells). Using the present invention, one skilled in the art will readily recognize that any of a wide variety of histological methods (such as staining procedures) can be modified to achieve such in situ detection.

[0168] In the above assays, the protein can be labeled with a detectable substance to detect the bound protein, or the bound (primary) specific protein can be detected by an antibody that is labeled with a detectable substance and binds to the primary specific protein.

[0169] The level of target in a sample can also be estimated by a competitive immunoassay that utilizes a target standard labeled with a detectable substance and an unlabeled target-specific protein. In this type of assay, the biological sample, the labeled target standard(s) and the target-specific protein are combined, and the amount of labeled target standard bound to the unlabeled target-specific protein is determined. The amount of target in the biological sample is inversely proportional to the amount of labeled target standard bound to the target-specific protein.

[0170] Suitable labels for target-specific proteins, secondary antibodies and / or target standards used in in vitro diagnostic techniques include, but are not limited to, various enzymes, prosthetic groups, fluorescent materials, luminescent materials and radioactive materials. Examples of suitable enzymes include horseradish peroxidase, alkaline phosphatase, β-galactosidase and acetylcholinesterase; examples of suitable prosthetic group complexes include streptavidin / biotin and avidin / biotin; examples of suitable fluorescent materials include umbelliferone, fluorescein, fluorescein isothiocyanate, rhodamine, dichlorotriazinylamine fluorescein, dansyl chloride and phycoerythrin; luminescent materials include, for example, luminol; examples of suitable radioactive materials include, for example, 125 I, 131 I, 35 S, and 3 Examples include H.

[0171] In one embodiment, the invention provides an in vivo imaging method in which a target-specific protein of the invention is conjugated to a detection-enhancing radio-opacifying agent, the conjugated protein is administered to a host, for example by injection into the bloodstream, and the presence and location of the labeled protein in the host is assayed. Through this technique and any other diagnostic methods provided herein, the invention provides methods for screening for the presence of disease-associated cells in a human patient or a biological sample taken from a human patient, and / or for assessing the distribution of target-specific proteins prior to target-specific ADC therapy.

[0172] For diagnostic imaging, radioisotopes can be bound directly or indirectly to target-specific proteins by using intermediate functional groups. Useful intermediate functional groups include chelating agents such as ethylenediaminetetraacetic acid and diethylenetriaminepentaacetic acid (see, e.g., U.S. Pat. No. 5,057,313).

[0173] In addition to radioisotopes and radio-opaque agents, diagnostic methods can be performed using target-specific proteins conjugated to dyes (such as biotin-streptavidin complexes), contrast agents, fluorescent compounds or molecules and enhancing agents (e.g., paramagnetic ions) for magnetic resonance imaging (MRI) (see, e.g., U.S. Pat. No. 6,331,175, which describes MRI techniques and the preparation of proteins conjugated to MRI enhancing agents). Such diagnostic / detection agents can be selected from agents and fluorescent compounds for use in MRI. Thus, the present invention provides diagnostic target-specific proteins, which are conjugated to contrast agents (e.g., for magnetic resonance imaging, computed tomography, or ultrasound contrast agents) or radionuclides, which can be, for example, gamma, beta, alpha, Auger electron or positron emitting isotopes.

[0174] In a further aspect, the present invention provides a kit for detecting the presence of a target antigen or target-expressing cells in a sample, comprising: a target-specific antibody of the invention, -Instructions for use of the kit, The present invention relates to a kit comprising:

[0175] In one embodiment, the present invention provides a kit for diagnosing cancer, comprising a container containing a target-specific protein and one or more reagents for detecting the binding of the target-specific protein to a target. The reagents may include, for example, fluorescent tags, enzymatic tags, or other detectable tags. The reagents may also include secondary or tertiary antibodies or reagents for enzymatic reactions that generate products that can be visualized. In one embodiment, the present invention provides a diagnostic kit comprising one or more target-specific proteins of the present invention in labeled or unlabeled form in appropriate container(s), depending on the nature of the label, reagents for incubation for indirect assays, and substrates or derivatizing agents for detection in such assays. Control reagent(s) and instructions for use may also be included.

[0176] Diagnostic kits may also be provided for use with target-specific proteins, such as labeled target-specific proteins, to detect the presence of a target in a tissue sample or a host. In such diagnostic kits, as well as in kits for therapeutic use described elsewhere herein, the target-specific protein may typically be provided in lyophilized form in a container, either alone or in conjunction with additional antibodies specific for the target cell or peptide. Typically, a pharma- ceutical acceptable carrier (e.g., an inert diluent) and / or its components, such as Tris, phosphate or carbonate buffer, stabilizers, preservatives, biocides, inert proteins, such as serum albumin, etc. (typically contained in a separate container for mixing), as well as additional reagents (also typically in separate container(s)). Certain kits also include a secondary antibody capable of binding to the target-specific protein, typically present in a separate container. The second antibody is typically conjugated to a label and formulated in a similar manner to the target-specific protein of the invention. Using the methods described above and elsewhere herein, target-specific proteins can be used to define subsets of cancer / tumor cells and to characterize such cells and associated tumor tissue. [Table 3] TIFF2024512386000004.tif235159TIFF2024512386000005.tif238159TIFF2024512386000006.tif237158TIFF2024512386000007.tif237159TIFF2024512386000008.tif238158TIFF2024512386000009.tif233159TIFF2024512386000010.tif236157TIFF2024512386000011.tif106160[Example] Introduction Utilizing the FEA-inactivated format (L234F-L235E-D265A) in antibodies of the IgG1 isotype, which are highly efficient in inducing complement-mediated cytotoxicity (CDC) as wild-type IgG1, strongly suppresses CDC activity. However, investigative experiments have occasionally shown that IgG1 antibodies with FEA substitutions still retain low residual CDC activity, which may not be desirable when the adverse effects or dose-limiting toxicity of a putative therapeutic antibody may be highly dependent on Fc-mediated effector functions such as CDC. Furthermore, the inventors have also observed that glycosylation heterogeneity is increased in antibodies with the FEA-inactivated format, with increased presence of galactosylation and charged glycans compared to the wild-type-like format. Changes in the glycosylation profile may affect efficacy and pharmacokinetic properties, which need to be monitored and controlled in manufacturing. When evaluating the glycan profile of antibodies with the FEA-inactivated format, it was observed that increased glycosylation heterogeneity, increased galactosylation levels, and increased charged glycan levels can be assigned to the D265A substitution. Considering the possible residual CDC activity of FEA, and considering the glycosylation profile of FEA, the inventors sought to provide an improved format.

[0177] In search of improvement, it was attempted to combine L234F, L235E and G236R substitutions. However, combining any mutation, let alone any selection of a specific mutation, is inherently unpredictable. Moreover, combining L234F, L235E and G236R mutations, i.e., multiple modifications packed together, is not easy, since the structural and / or functional impact of such multiple changes cannot be predicted. Therefore, in pilot experiments, the inventors tested the ability of the combination L234F, L235E and G236R (FER) to silence CDC and its effect on the glycosylation profile of the antibody. Surprisingly, these initial experiments showed that the ability to silence CDC by introducing the FER mutation was greatly improved compared to the individual substitutions and compared to the FEA substitution. Furthermore, evaluation of the glycosylation profile revealed that the antibody now with the FER inactive format has a glycan profile similar to the corresponding wild-type-like antibody. These initial experiments therefore prompted us to fully analyze the clinical suitability of the FER inactivated format, which is described in the following examples. Surprisingly, in all characteristics tested, highly favorable properties were observed for the FER inactivated format when compared to the FEA inactivated format, as well as when compared to other inactivated formats. This indicates that the new FER format is well suited for clinical development and use. This inactivated format, which may also be useful in situations other than antibodies, such as fusion proteins, is highly advantageous and can be considered a best-in-class inactivated format.

[0178] [Example 1] Production and purification of antibody variants and production of bispecific antibody variants Antibody expression constructs For expression of antibodies with human or humanized binding domains, antibody constructs comprising sequences of variable heavy (VH) and variable light (VL) chain domains were prepared by de novo gene synthesis (GeneArt Gene Synthesis; ThermoFisher Scientific, Germany) and expressed as appropriate for the selected binding domain: human IgG1m(f) allotype (also referred to as IgG1 throughout the Examples section of this application; SEQ ID NO: 5), or human IgG1m(fa) allotype (SEQ ID NO: 23), or human IgG1m(za) allotype (SEQ ID NO: 24), or human IgG1m(zax) allotype (SEQ ID NO: 25), or human IgG1m(zav) allotype (SEQ ID NO: 26), or human IgG1m(f) allotype with a recombinant deletion of the C-terminal lysine (also referred to as IgG1-delK; SEQ ID NO: 35), or variants thereof; human IgG4 heavy chain constant region (hinge, CH2, C H3, SEQ ID NO: 8; CH1, hinge, CH2, CH3, SEQ ID NO: 46) or variants thereof; human IgG3 heavy chain constant region of human IgG3 allotype IGHG3*01 (SEQ ID NO: 40) or human IgG3 allotype IGHG3*04 (also called IgG3rch2; SEQ ID NO: 41), or variants thereof; mouse IgG2a heavy chain constant region (SEQ ID NO: 38) or variants thereof; or human kappa light chain (SEQ ID NO: 6), or human lambda LC (SEQ ID NO: 7), or mouse kappa LC (SEQ ID NO: 49) constant region were cloned into the pcDNA3.3 expression vector (ThermoFisher Scientific, US). The desired substitutions to generate the variants were introduced by gene synthesis and are described below. The CD20 antibody variants in this application have VH and VL sequences derived from the type I anti-human CD20 antibodies previously described (Engelberts et al., 2020). The HLA-DR antibody variants in this application have VH and VL sequences derived from the previously described HLA-DR antibodies IgG1-HLA-DR-4 (US Pat. No. 6,894,149) and IgG1-HLA-DR-1D09C3 (US Pat. No. 7,521,047).The CD3 antibodies and variants thereof in this application have VH and VL sequences derived from previously described CD3 antibodies (i.e., CD3-huCLB-T3 / 4: Labrijn et al., PNAS, 2013 Mar 26; 110(13): 5145-50, and Engelberts et al., 2020). The HER2 antibody variants in this application include VH and VL sequences derived from IgG1-HER2-1014-169 (WO2012143524). A human IgG1 antibody, i.e., an HIV1 gp120 specific antibody, containing the VH / VL of b12 was used as a negative control in some experiments (Barbas et al., J Mol Biol. 1993 Apr 5; 230(3): 812-2).

[0179] Non-activating antibody variants A wild-type human IgG1 heavy chain constant region (i.e., hinge, CH2 and CH3 regions) with a F405L mutation (SEQ ID NO: 9) or a K409R mutation (SEQ ID NO: 10) or a wild-type-like variant thereof is used in the control antibody as indicated. In some examples, a wild-type antibody variant with an IgG4 constant region (SEQ ID NO: 8) or an IgG4 variant thereof with a S228P mutation in SEQ ID NO: 8 is used as a control.

[0180] The non-activating Fc domain prevents the antibody from interacting with Fc receptors or C1q present on immune cells to activate the classical complement pathway. As described herein in the following examples, several non-activating antibody variants were produced and tested for their ability to silence Fc function. In some examples, a subset of non-activating antibody variants was tested for pharmacokinetic properties, immunogenicity or developmental potential.

[0181] Various non-activating antibody variants were generated using different VH and VL sequences as shown and described above. The following substitutions, where the amino acids are as defined by Eu numbering, were introduced into the IgG1m(f)HC region of SEQ ID NO: 1 in order to allow the generation of bispecific non-activating antibody variants, also in combination with either the F405L or K409R mutation: L234F-L235E-D265A (ia US Pat. No. 10,590,206) alone (SEQ ID NO: 3), or in combination with F405L (SEQ ID NO: 13) or K409R (SEQ ID NO: 14), L234F-L235E-G236R alone (SEQ ID NO: 2), or in combination with F405L (SEQ ID NO: 11) or K409R (SEQ ID NO: 12), L234A-L235A-P329G (Schlothauer et al., 2016, Protein Eng. Design and Selection) alone or in combination with either F405L or K409R, G236R-L328R (US2006 / 0235208A1, Moore et al., 2011, MAbs), E233P-L234V-L235A-delG236-S267K in combination with either F405L or K409R (Moore at al., 2019, Methods), N297G in combination with either F405L or K409R (Tao and Morrison, 1989, JI), and L234A-L235E-G237A-A330S-P331S in combination with either F405L or K409R (US8613926k). The following mutations were introduced into the IgG1m(f)HC region of SEQ ID NO: 35, which had a recombinant deletion of the HC C-terminal lysine (also called IgG1-delK): L234F-L235E-G236R (SEQ ID NO: 36) and L234F-L235E-D265A (SEQ ID NO: 37). The following mutations were introduced into the IgG1m(fa)HC region of SEQ ID NO: 23: L234F-L235E-G236R (SEQ ID NO: 27) and L234F-L235E-D265A (SEQ ID NO: 28). The following mutations were introduced into the IgG1m(za)HC region of SEQ ID NO: 24: L234F-L235E-G236R (SEQ ID NO: 29) and L234F-L235E-D265A (SEQ ID NO: 30).The following mutations were introduced into the IgG1m (zax) HC region of SEQ ID NO: 25: L234F-L235E-G236R (SEQ ID NO: 33) and L234F-L235E-D265A (SEQ ID NO: 34). The following mutations were introduced into the IgG1m (zav) HC region of SEQ ID NO: 26: L234F-L235E-G236R (SEQ ID NO: 31) and L234F-L235E-D265A (SEQ ID NO: 32). The following mutations were introduced into the IgG3 (IGHG3*01) HC region of SEQ ID NO: 40: L234F-L235E-G236R (SEQ ID NO: 42) and L234F-L235E-D265A (SEQ ID NO: 43). The following mutations were introduced into the HC region of IgG3 of SEQ ID NO: 41 (IGHG3*04; also called IgG3rch2): L234F-L235E-G236R (SEQ ID NO: 44) and L234F-L235E-D265A (SEQ ID NO: 45). The following mutations were introduced into the IgG4HC region of SEQ ID NO: 8, alone or in combination with the F405L-R409K mutation, allowing the generation of bispecific non-activating antibody variants: S228P-E233P-F234V-L235A-delG236 (WO2015 / 143079), or in combination with F405L-R409K, S228P-F234A-L235A (Allegre et al., Transplantation, 1994, Jun 15;57(11):1537-43 and Angal et al., Mol Immunol, 1993, Jan;30(1):105-8), or in combination with F405L-R409K, L235E-G236R (SEQ ID NO: 47), and L235E-D265A (SEQ ID NO: 48). The following mutations were introduced into the murine IgG2a HC region of SEQ ID NO:38: L234F-L235E-G236R (SEQ ID NO:39), L234A-L235A (Arduin et al., Mol Immunol, 2015 Feb;63(2):456-63), and L234A-L235A-P329G (Lo et al., J Biol Chem, 2017 Mar 3;292(9):3900-3908).

[0182] Transient expression Antibodies were expressed as human IgG1κ, IgG1λ, IgG3κ, IgG4κ, or mouse IgG2aκ. Plasmid DNA mixtures encoding both the heavy and light chains of the antibodies were transiently transfected into Expi293F™ cells (Thermo Fisher Scientific, Cat. No. A14527) using ExpiFectamine (Thermo Fisher Scientific, Cat. No. A14525) according to the manufacturer's instructions. Briefly, both DNA (1:1 HC / LC ratio) and ExpiFectamine are separately diluted to 20 μg / ml and 5.4% (v / v), respectively, in Opti-MEM I (Thermo Fisher Scientific, Cat. No. 51985034). Both solutions were incubated for 5 min, mixed, incubated for 10–20 min, and then diluted to 3 × 10 6 DNA in fresh Opti-MEM supplemented with pen / strep (final DNA concentration 1 μg / ml). 6 Add ExpiFectamine293F cells / ml. Approximately 16-24 hours after transfection, add ExpiFectamine293 Transfection Enhancer I (0.5(v / v)%) and II (5(v / v)%). Supernatants are typically harvested 5 days after transfection. Antibody concentrations in the supernatants were measured by absorbance at 280 nm. Antibodies were purified as follows:

[0183] Antibody purification and quality assessment Antibodies were purified by Protein A affinity chromatography. Culture supernatants were filtered through a 0.20 μM dead-end filter, loaded onto a 5 mL MabSelect SuRe column (GE Healthcare / Cytiva), washed with 0.02 M sodium citrate-NaOH (pH 5.0), and eluted with 0.02 M sodium citrate-NaOH (pH 3). The eluate was diluted in PBS (8.65 mM NaCl prepared for Genmab by HyClone / Cytiva) and diluted with 0.02 M sodium citrate-NaOH (pH 3). 2 HPO 4 Anhydrous, 1.9 mM NaH 2 PO 4Monohydrate, 140.3 mM NaCl, pH 7.4 buffer). If necessary, the protein was further purified by preparative size-exclusion chromatography (SEC) to remove aggregates. After buffer exchange or SEC, samples were sterile filtered through 0.2 μm dead-end filters. The quality of the purified protein was analyzed by mass spectrometry, capillary electrophoresis (non-reduced and reduced CE-SDS) and high performance size-exclusion chromatography (HP-SEC). Concentration was measured by absorbance at 280 nm. Purified antibodies were stored at 2-8 °C.

[0184] Generation of bispecific antibodies IgG1 bispecific antibodies were generated by controlled Fab-Arm exchange (cFAE) between a monospecific antibody carrying a F405L mutation in the CH3 domain and a monospecific antibody carrying a K409R mutation in the CH3 domain, in addition to the inactivating mutations listed above, essentially as previously reported (Labrijn et al., 2014, Nature Protocols Oct;9(10):2450-63). For the IgG4 bispecific variants, one monospecific antibody had a natural arginine (R) at position 409 and the other monospecific antibody had a F405L-R409K mutation in the antibody's CH3 domain. Briefly, both monospecific antibodies were mixed at equimolar concentrations and incubated with 75 mM 2-mercaptoethylamine-HCl (2-MEA) for 5 h at 31 °C. The 2-MEA was then removed by buffer exchange with PBS using Slide-A-Lyzer Dialysis Cassettes (10K MWCO; ThermoFisher Scientific) overnight at 4 °C. The dialysis buffer was exchanged twice. The generated bispecific antibodies were recovered from the cassettes and the concentration was measured by absorbance at 280 nm. The purified bispecific antibodies were stored at 2-8 °C and typically analyzed by CE-SDS and HP-SEC to assess the monomeric state of the generated bispecific antibodies, and by mass spectrometry using an Orbitrap Q-Exactive Plus mass spectrometer (ThermoFischer Scientific) to determine the efficiency of cFAE and bispecific antibody content.

[0185] Generation of F(ab')2 fragments F(ab')2 fragments targeting HLA-DR were generated using the FragIT kit (Genovis AB) essentially following the manufacturer's recommendations. Briefly, a spin column with FragIT, a resin with the FabRICATOR enzyme that digests IgG at a specific site below the hinge region generating a homogenous pool of F(ab')2 and Fc / 2 fragments coupled to agarose beads, is equilibrated with digestion buffer (Genovis AB). Subsequently, anti-human HLA-DR antibody variant IgG1-HLA-DR-1D09C3 with the E430G mutation is applied to the column and incubated for 15 min at room temperature (RT). After incubation, the sample is eluted and collected from the column. Purified F(ab')2 fragments were analyzed by capillary electrophoresis on sodium dodecyl sulfate-polyacrylamide gels (CE-SDS). Concentration was measured by absorbance at 280 nm. Purified antibodies were stored at 2-8 °C.

[0186] [Example 2] Target binding of anti-human CD20 antibody and its non-activating variant to Raji cells Anti-human CD20 with wild type Fc and variants with K409R, L234F-L235E-D265A-K409R, L234F-L235E-G236R-K409R, L234A-L235A-P329G-K409R, G236R-L328R-K409R, E233P-L234V-L235A-G236del-S267K-K409R, N297G-K409R, L234A-L235E-G237A-A330S-P331S-K409R Target binding of IgG1 and IgG4 antibodies with wild-type Fc and variants with S228P, S228P-F234A-L235A or S228P-E233P-F234V-L235A-G236del was assessed using Raji cells expressing CD20. Anti-HIV1 gp120 IgG1 wild-type antibody (IgG1-b12) was used as a non-binding control.

[0187] For this, FACS buffer (1x PBS, Cat. No. BE17-517Q, Lonza; 0.1% bovine serum albumin, BSA, Cat. No. 10735086001, Merck; 0.02% NaN 3 , Catalog No. 41920044-3, Bio-world) in 100 mL of Raji cells (3 × 10 4 Cells (cat. no. CCL-86, ATCC) were incubated in polystyrene round-bottom 96-well plates (cat. no. 650180, Greiner bio-on) at 4 °C with a concentration series of purified antibodies targeting CD20 (final concentrations of 0.0013 to 20 μg / mL; 5-fold dilutions) in a total volume of 50 μL for 30 min. Afterwards, cells were washed twice by the addition of 150 μL of FACS buffer, followed by pelleting the cells by centrifugation and removing the supernatant. Bound antibodies were eluted with 50 μL of goat F(ab') 2 Staining was performed by adding anti-human κ-PE (final concentration 2.5 μg / mL, Cat. No. 2062-09, Southern Biotech) for 30 min at 4°C. Cells were then washed twice by adding 150 μL of FACS buffer, followed by pelleting the cells by centrifugation and removing the supernatant. Binding of antibody variants to human CD20 was detected by measuring median fluorescence intensity by flow cytometry on an Intellicyt iQue screener (Sartorius). Data were analyzed using a nonlinear agonist dose-response model in GraphPad PRISM (version 8.4.1, GraphPad Software). Data are means ± SEM from four independent experiments.

[0188] Evaluation of target binding of anti-human CD20 IgG1 and IgG4 antibody variants revealed similar target binding for all anti-human CD20 IgG1 antibody variants with mutations in the heavy chain constant region (Figure 1A-B). All anti-human CD20 IgG4 antibody variants, including those with S228P-F234A-L235A or S228P-E233P-F234V-L235A-G236del inactivating mutations in the heavy chain constant region, showed somewhat reduced target binding compared to the IgG1 antibody variants.

[0189] In summary, the introduction of non-activating mutations in the heavy chains of anti-human CD20 IgG1 and IgG4 antibodies did not affect their ability to bind to their target CD20.

[0190] [Example 3] Complement-dependent cytotoxicity by anti-human CD20 antibodies and their non-activating variants For therapeutic monoclonal antibodies whose therapeutic effect is independent of interactions with effector molecules of the immune system (such as the complement protein C1q and Fcγ receptors) and whose adverse effects or dose-limiting toxicity are highly dependent on this interaction, a non-activating Fc domain may be considered to increase the therapeutic window. Engagement of the antibody with the C1q protein initiates the classical complement pathway, resulting in complement-dependent cytotoxicity (CDC). Here, the ability to induce CDC was evaluated for type I anti-human CD20 antibodies selected for their efficient and potent induction of CDC (Glennie et al., Mol Immunology, 2007, Sep;44(16):3823-37), as well as for variants of such antibodies with non-activating mutations in the constant heavy chain region.

[0191] Anti-human CD20 IgG1 antibodies as wild type or with the K409R mutation or inactivated variants thereof with the L234F-L235E-D265A-K409R, L234F-L235E-G236R-K409R, L234A-L235A-P329G-K409R, G236R-L328R-K409R, E233P-L234V-L235A-G236del-S267K-K409R, N297G-K409R, L234A-L235E-G237A-A330S-P331S-K409R mutations, as well as wild type, anti-human CD20 IgG1 antibodies with the S228P mutation. IgG4 antibodies and anti-human CD20 IgG4 antibodies carrying the S228P-F234A-L235A or S228P-E233P-F234V-L235A-G236del mutations were tested at a range of concentrations (0.0024-10 μg / mL final concentration; 4-fold dilutions) in an in vitro CDC assay on Raji cells using 20% ​​normal human serum (NHS, Cat. No. M0008, Sanquin) as a complement source. Raji cells (3 × 10 ) were cultured in RPMI-1640 medium (catalog no. BE12-115F, Lonza) containing 0.1% (w / v) bovine serum albumin (BSA, catalog no. 10735086001, Merck) and penicillin-streptomycin (Pen / Strep, final concentrations of 50 units / mL penicillin potassium and 50 μg / mL streptomycin sulfate, catalog no. DE17-603E, Lonza). 4Cells / well) were incubated in polystyrene round-bottom 96-well plates (catalog no. 650180, Greiner bio-one) for 15 min at room temperature (RT) with a range of concentrations of purified antibodies targeting CD20 in a total volume of 80 μL. Then, 20 μL of NHS (final concentration 20% (v / v)) was added and the cells were incubated for 45 min at 37 °C. After placing the plates on ice, the reaction was stopped by pelleting the cells by centrifugation and replacing the supernatant with 30 μL of 2 μg / mL propidium iodide (PI, catalog no. P4170, Sigma Aldrich) in PBS (catalog no. SH3A3830.03, GE Healthcare). The number of PI-positive cells was determined by flow cytometry on an Intellicyt iQue screener (Sartorius). The percentage of PI-positive cells, corresponding to the percentage of cell lysis, was calculated as (number of PI-positive cells / total number of cells) × 100%. Data were analyzed using a non-linear agonist dose-response model and the area under the dose-response curve (AUC) per experimental replicate was calculated using log-transformed concentrations in GraphPad PRISM (version 8.4.1, GraphPad Software) without antibody control as baseline, followed by normalization per experimental replicate to the AUC value measured for the non-binding control antibody IgG1-b12 (0%) and the AUC value measured for the wild-type IgG1 antibody variant (anti-human CD20 IgG1, 100%). Data are means ± SEM from three independent experiments.

[0192] Evaluation of the CDC capacity of anti-human CD20 IgG1 and IgG4 variants revealed that they could be separated into two groups according to their ability to abrogate CDC by introducing deactivating mutations in the heavy chain constant region. The ability to induce CDC was abrogated for anti-human CD20 IgG1 antibody variants harboring the K409R mutation as well as the L234F-L235E-G236R, L234A-L235A-P329G, G236R-L328R, E233P-L234V-L235A-G236del-S267K, N297G, and L234A-L235E-G237A-A330S-P331S inactivating mutations, as well as for IgG4 antibody variants harboring the S228P-F234A-L235A or S228P-E233P-F234V-L235A-G236del inactivating mutations. However, residual CDC was observed for the anti-human CD20 IgG1 variant carrying the mutations L234F-L235E-D265A-K409R compared to wild-type IgG1 (Figure 2).

[0193] In conclusion, IgG1 antibodies with the novel variants L234F-L235E-G236R and K409R mutations showed an almost complete lack of activation of the classical complement pathway. A similar effect was observed with antibodies with only the novel L234F-L235E-G236R mutations but without the K409R mutation. This represents a significant advance over IgG1 antibody variants with L234F-L235E-D265A and K409R mutations, in which residual CDC was observed, among others.

[0194] [Example 4] Evaluation of C1q binding to anti-human CD20 antibodies and their non-activating variants In Example 3, it was shown that the potency of inducing CDC was strongly reduced by anti-human CD20 IgG1 and IgG4 antibody variants with mutations that suppress Fc-mediated effector functions. Here, the binding of the complement protein C1q to anti-human CD20 IgG1 antibody variants with deactivating mutations in the constant heavy chain region was assessed using CD20-expressing Raji cells.

[0195] Anti-human CD20 antibodies and their non-activating variants were tested at various concentrations (0.014-10 μg / mL final concentration; 3-fold dilutions) using a C1q binding assay on Raji cells with 20% normal human serum (NHS, M0008, Sanquin) as a source of C1q. Raji cells (1 × 10 cells) were cultured in RPMI-1640 medium (catalog no. BE12-115F, Lonza) containing 0.1% (w / v) bovine serum albumin (BSA, catalog no. 10735086001, Merck) and penicillin-streptomycin (Pen / Strep, final concentration 50 units / mL penicillin potassium and 50 μg / mL streptomycin sulfate, catalog no. DE17-603E, Lonza). 5 Cells / well) were incubated in polystyrene round-bottom 96-well plates (catalog no. 650180, Greiner bio-one) for 15 min at 37 °C with a range of concentrations of purified antibodies targeting CD20 in a total volume of 80 μL. To prevent activation of CDC upon addition of NHS, cells were cooled by placing the plate on ice. Subsequently, 20 μL of NHS (final concentration 20% (v / v)) was added and cells were incubated for 45 min at 4 °C, after which the cells were pelleted by centrifugation and the supernatant removed. Next, 150 μL of FACS buffer (1x PBS, catalog no. BE17-517Q, Lonza; 0.1% bovine serum albumin, BSA; 0.02% NaN 3The cells were washed twice with 50 μL of polyclonal rabbit anti-human C1q complement-FITC (final concentration 20 μg / mL, Dako, catalog number 41920044-3, Bio-world), after which the cells were pelleted by centrifugation and the supernatant was removed. Bound C1q was stained by adding 50 μL of polyclonal rabbit anti-human C1q complement-FITC (final concentration 20 μg / mL, Dako, catalog number F0254, Agilent Technologies) for 30 min at 4 °C. The cells were then washed twice by adding 150 μL of FACS buffer, followed by pelleting the cells by centrifugation and removing the supernatant. C1q binding to the antibody variants was detected by measuring the median fluorescence intensity-FITC by flow cytometry on an Intellicyt iQue screener (Sartorius). Data were analyzed using a nonlinear agonist dose-response model in GraphPad PRISM (version 8.4.1, GraphPad Software). Data are the mean (± SD) from triplicate measurements of one experiment.

[0196] Assessment of C1q binding to anti-human CD20 IgG1 variants revealed that anti-human CD20 IgG1 with K409R mutation efficiently engages complement protein C1q upon binding to CD20 on target Raji cells, whereas anti-HIV1 gp120 antibody (IgG1-b12; CD20 non-binding control) does not. Introduction of either L234F-L235E-G236R or L234F-L235E-D265A inactivating mutations dramatically reduced C1q binding (Figure 3). C1q binding is more strongly reduced for anti-human CD20 IgG1 variants with L234F-L235E-G236R mutations compared to anti-human CD20 IgG1 variants with L234F-L235E-D265A mutations. This is consistent with the effect of these inactivating mutations on CDC activation, as shown in Example 3.

[0197] In conclusion, a non-activating anti-human CD20 IgG1 variant carrying the L234F-L235E-G236R inactivating mutations shows a strongly reduced C1q binding, consistent with the observed reduction in CDC.

[0198] [Example 5] Complement-dependent cytotoxicity by anti-human HLA-DR antibodies and their non-activating variants In Example 3, activation of the classical complement pathway by anti-human CD20 antibody variants with non-activating mutations in the constant heavy chain was evaluated using an in vitro CDC assay. The data in Example 3 revealed that anti-human CD20 IgG1 variants with the novel L234F-L235E-G236R and K409R mutations prevented activation of the classical complement pathway. In this example, we further evaluated the CDC capacity of the non-activating antibody variants using two antibodies targeting human HLA-DR, which is a potent inducer of CDC when used with an IgG1 backbone.

[0199] In vitro CDC assays were performed on Raji cells with 20% NHS as a complement source essentially as described in Example 3. Briefly, anti-human HLA-DR antibodies IgG1-HLA-DR-4 and IgG1-HLA-DR-1D09C3 with the K409R mutation or their variants with the inactivating mutations L234F-L235E-D265A or L234F-L235E-G236R combined with K409R in the constant heavy chain region, as well as HLA-DR targeting F(ab')2 fragments, were tested in a range of concentrations (0.014-10 μg / mL final concentration; 3-fold dilutions). The number of PI-positive cells as a measure of cell lysis was determined by flow cytometry on an Intellicyt iQue screener (Sartorius). Data were analyzed using a nonlinear agonist dose-response model, and the area under the dose-response curve (AUC) per experimental replicate was calculated using log-transformed concentrations in GraphPad PRISM (version 8.4.1, GraphPad Software) without antibody control as baseline, and then normalized per experimental replicate to the AUC value measured for the wild-type IgG1 antibody variant (100%). Data are based on five (wild-type and L234F-L235E-D265A-K409R variant) or two (L234F-L235E-G236R-K409R variant, or F(ab')2 fragment) independent experiments.

[0200] Evaluation of the CDC capacity of anti-human HLA-DR antibody variants revealed that antibody variants carrying the L234F-L235E-D265A mutation in addition to the K409R mutation retained the ability to induce CDC in Raji cells (Figure 4A, Figure 4B). However, introduction of the L234F-L235E-G236R inactivating mutation in the IgG1-HLA-DR-4-K409R and IgG1-HLA-DR-1D09C3-K409R antibody variants dramatically reduced their potency to induce CDC. Moreover, as observed for IgG1-HLA-DR-1D09C3, the potency to induce CDC by the L234F-L235E-G236R-K409R antibody variant is at the same level as the CDC induced by an HLA-DR-targeting F(ab')2 fragment that lacks the Fc region and therefore does not interact with the complement protein C1q (Figure 4B).

[0201] In summary, antibody variants carrying the L234F-L235E-G236R mutations efficiently reduced CDC to levels similar to those induced by F(ab')2 fragments that cannot engage the human complement system.

[0202] [Example 6] Binding to Fc.GAMMA. receptors by anti-human CD20 antibodies and their non-activating variants As such, for therapeutic monoclonal antibodies whose therapeutic effect is independent of interactions with effector molecules of the immune system (such as the complement protein C1q and Fcγ receptors) and whose adverse effects or dose-limiting toxicity are highly dependent on this interaction, a non-activating Fc domain may be considered to increase the therapeutic window. The data in Examples 3-5 showed that the introduction of non-activating mutations into the constant heavy chain region of the antibody reduces its ability to engage the complement protein C1q and reduces the induction of CDC. Here, we assessed binding of IgG1 or IgG4 anti-human CD20 antibodies and antibody variants thereof with non-activating mutations in the constant heavy chain region to the monomeric extracellular domain (ECD) of human FcγRIa (SEQ ID NO: 15) or the dimeric ECDs of human FcγRIIa allotype 131H (SEQ ID NO: 16), human FcγRIIa allotype 131R (SEQ ID NO: 17), human FcγRIIb (SEQ ID NO: 18), human FcγRIIIa allotype 158F (SEQ ID NO: 19) and human FcγRIIIa allotype 158V (SEQ ID NO: 20) in an ELISA assay.

[0203] For this purpose, anti-human CD20 antibodies IgG1 wild type, IgG1-K409R, L234F-L235E-D265A-K409R, L234F-L235E-G236R-K409R, L234A-L235A-P329G-K409R, G236R-L328R-K409R, E233P-L234V-L235A-G236del-S267K-K409R, N297G-K409R, L234A-L235 IgG4 wild type, IgG4-S228P, and its inactive variants with E-G237A-A330S-P331S-K409R mutations, as well as inactive variants with S228P-F234A-L235A or S228P-E233P-F234V-L235A-G236del mutations, were tested in a range of concentrations (0.0013 to 20 μg / mL in 5-fold dilution steps) for binding to human Fcγ receptors. To detect binding to monomeric and dimeric FcγR variants, 96-well Microlon ELISA plates (Greiner, Germany) were filled with 1 μg / mL goat F(ab') in PBS. 2 -Anti-human IgG-F(ab') 2Plates were coated with 100 μl / well of a dilution series of anti-human CD20 IgG1 and IgG4 antibody variants in PBST / 0.2% BSA with washing between incubations (PBS supplemented with 0.05% Tween 20 (catalog no. P1379, Sigma); PBST; +0.2% BSA; PBST / 0.2% BSA), at RT for 1 h with shaking, followed by incubation with 100 μl / well of monomeric or dimeric His-tagged C-terminally biotinylated FcγR ECD variants (1 μg / mL) in PBST / 0.2% BSA for 1 h with shaking at RT. Finally, after washing as above, the plates were incubated with 100 μL / well streptavidin-polyHRP (CLB, Cat. No. M2032, 1:10.000) in PBST / 0.2% BSA as antibody detector for 30 min at room temperature with shaking. Development with 1 mg / mL ABTS (Cat. Nos. 11112422001 and 11112597001, Roche) was performed for approximately 10 min (Ia), 15 min (IIa-131H, IIa-131R, IIIa-158V, IIIa-158F) or 30 min (IIb). The reaction was then stopped with 100 μL / well 2% oxalic acid (Riedel de Haen, Cat. No. 33506). Absorbance was measured at 405 nm using a microplate reader (BioTek, Winoosky, VT).

[0204] To measure the immobilization of antibody variants and evaluate whether the introduced non-activating mutations affect capture to the ELISA plate, 96-well Microron ELISA plates (catalog no. 655092, Greiner) were filled with 1 μg / mL goat F(ab') in PBS. 2 -Anti-human IgG-F(ab') 2(catalog no. 109-006-097, Jackson Laboratory) overnight at 4° C. The ELISA plates were then washed and blocked with 200 μL / well PBS supplemented with 0.2% BSA (PBS / 0.2% BSA) for 1 h at room temperature (RT). With washing between the different incubation steps (PBS supplemented with 0.05% Tween 20 (PBST) + 0.2% BSA; PBST / 0.2% BSA), the plates were successively incubated with a dilution series of anti-human CD20 IgG1 and IgG4 antibody variants (0.0013 to 20 μg / mL in 5-fold dilution steps) at 100 μl / well in PBST / 0.2% BSA for 1 h at room temperature with shaking, and with HRP-labeled goat anti-human IgG-Fcγ (catalog no. 109-035-098, Jackson Laboratory; 1:10,000) at 100 μl / well in PBST / 0.2% BSA with antibody detection for 30 min at room temperature with shaking. Color development was performed with 1 mg / mL ABTS (catalog numbers 11112422001 and 11112597001, Roche) for approximately 5 minutes, after which the reaction was stopped with 100 μL / well of 2% oxalic acid. The absorbance was measured at 405 nm using a microplate reader.

[0205] Data were analyzed using a nonlinear agonist dose-response model and the area under the dose-response curve (AUC) per experimental replicate was calculated using log-transformed concentrations in GraphPad PRISM (version 8.4.1, GraphPad Software) using the ELISA background signal (no antibody control) as baseline, followed by normalization per experimental replicate to the AUC value measured for the wild-type IgG1 antibody variant (anti-human CD20 IgG1, 100%). Data are based on three independent replicates.

[0206] All anti-human CD20 IgG1 and IgG4 antibody variants were immobilized on ELISA plates in a similar manner, and a slight reduction was observed for antibody variants with an IgG4 backbone (Figure 5). It was therefore concluded that the introduction of non-activating mutations in the heavy chain region of anti-human CD20 IgG1 and IgG4 antibody variants does not affect immobilization by anti-human IgG-F(ab')2 fragments. Evaluation of the binding of anti-human CD20 IgG1 variants to FcγRIa, FcγRIIa, FcγRIIb and FcγRIIIa by ELISA revealed that all anti-human CD20 IgG1 antibody variants with non-activating mutations in the constant heavy chain region do not interact with Fcγ receptors (Figures 6A-F). Furthermore, the non-activating IgG4 variants were also unable to interact with Fcγ receptors, except for the IgG4 antibody variant harboring the S228P-E233P-F234V-L235A-delG236 mutation, which showed residual binding to FcγRIIa-131R (Figure 6C).

[0207] In summary, the IgG1 antibody variant with the L234F-L235E-G236R inactivating mutation did not exhibit FcγR binding, similar to previously described inactivating Fc variants such as L234F-L235E-D265A and L234A-L235A-P329G.

[0208] [Example 7] Activation and signal transduction via Fcγ receptors by anti-human CD20 antibodies and their non-activating variants In Example 6, binding of anti-human CD20 IgG1 or IgG4 antibodies and their variants with non-activating mutations in the constant heavy chain to FcγRIa, FcγRIIa, FcγRIIb and FcγRIIIa was investigated. All non-activating antibody variants tested showed no binding to the tested FcγRs, except for the IgG4 antibody variant with S228P-E233P-F234V-L235A-delG236 mutation, which showed residual binding to FcγRIIa-131R. However, in the ELISA binding assay, only the effect on direct binding is evaluated. There is no effect on antigen binding of Fc receptors to effector cells, target-mediated antibody clustering and subsequent target-mediated clustering. Here, we investigated whether introduction of non-activating mutations in the heavy chain constant regions of anti-human CD20 IgG1 or IgG4 antibodies, as described in Example 6, affects FcγR activation and signaling in a Promega reporter assay using target-expressing Raji cells and Jurkat reporter cell lines expressing the indicated FcγRs.

[0209] Activation of FcγR-mediated signaling by the above-mentioned anti-human CD20 IgG1 and IgG4 antibody variants was quantified using a reporter bioassay (Promega, FcγRIa: Catalog No. CS1781C08; FcγRIIa allotype 131H: Catalog No. G9991; FcγRIIa allotype 131R: Catalog No. CS1781B08; FcγRIIb: Catalog No. CS1781E04; FcγRIIIa allotype 158F: Catalog No. G9790; FcγRIIIa allotype 158V: Catalog No. G7010) using CD20-expressing Raji cells as target cells. As effector cells, the reporter cell kit contains Jurkat human T cells engineered to stably express the indicated FcγR and nuclear factor of activated T cells (NFAT) response element driving the expression of firefly luciferase. The assay is performed according to the manufacturer's recommendations. Briefly, Raji cells (5,000 cells / well) were seeded into 384-wells of white OptiPlates (Perkin Elmer, catalog no. 6007290) in assay buffer (Promega, catalog no. G719A) supplemented with 12% low IgG serum (Promega, catalog no. G711A) and incubated at 37 °C / 5% CO in a total volume of 30 μL containing an antibody concentration series (FcγRIIa, FcγRIIb and FcγRIIIa, final concentrations of 0.001–15 μg / mL in 5-fold dilutions; FcγRIa, final concentrations of 0.00000006–15 μg / mL in 25-fold dilutions) and thawed Promega BioAssay Effector cells (30,000 cells / well). 2The plates were incubated at RT for 5 hours. The plates were then incubated at room temperature (RT) for 15 minutes, followed by the addition of 30 μL of Bio Glo Assay Luciferase Reagent. The plates were then incubated at room temperature for 5 minutes. Luciferase production was quantified by luminescence readout on an EnVision Multilabel Reader (Perkin Elmer). The background luminescence signal determined from the medium (no Raji cells, no antibody, no effector cells) sample was used to normalize the luminescence signal. Data were analyzed using a nonlinear agonist dose-response model, and the area under the dose-response curve (AUC) per experimental replicate was calculated using the log-transformed concentration in GraphPad PRISM (version 8.4.1, GraphPad Software) of the no antibody control (only Raji cells and effector cells) as the baseline. For each experiment, AUC values ​​were normalized to the reporter activity observed for cells incubated with non-binding control IgG1-b12 (0%) and wild-type anti-human CD20 IgG1 (100%). Data are based on three independent replicates.

[0210] Assessment of FcγRIa-, FcγRIIa-, FcγRIIb-, and FcγRIIIa-mediated activation using Promega reporter assays revealed that introduction of L234F-L235E-G236R in combination with the K409R mutation in the IgG1 heavy chain constant region prevented FcγR-mediated activation, as did the anti-human CD20 IgG1 nonactivating variants L234F-L235E-D265A-K409R, L234A-L235A-P329G-K409R, G236R-L328R-K409R, and E233P-L234V-L235A-G236del-S267K-K409R (Figure 7A-F). With the exception of partial FcγRIa-mediated activation by anti-human CD20 IgG1 antibody variants with N297G and K409R mutations (FIG. 7A), and partial FcγRIIa- and FcγRIIb-mediated activation by IgG4 antibody variants with S228P-F234A-L235A or S228P-E233P-F234V-L235A-G236del mutations, and partial FcγRIIa- and FcγRIIb-mediated activation by anti-human CD20 IgG1 antibody variants with L234A-L235E-G237A-A330S-P331S-K409R mutations (FIGS. 7B-D), lack of FcγR-mediated activation was also observed for other IgG1 or IgG4 antibody variants with non-activating mutations in the heavy chain constant region. As controls, antibody variants anti-human CD20 IgG1 and anti-human CD20 IgG1-K409R induced activation of all FcγRs tested, whereas antibody variants IgG4 and IgG4-S228P induced FcγRIIa-, FcγRIIb-, and FcγRIa-mediated activation, but were devoid of FcγRIIIa-mediated activation (Figures 7A-F).

[0211] For some variants partial FcγR mediated activation of one or more of the Fcγ receptors was observed, i.e. the ability to induce FcγR mediated activation by IgG1 variants carrying the N297G-K409R or L234A-L235E-G237A-A330S-P331S-K409R mutations and IgG4 antibody variants carrying the S228P-F234A-L235A or S228P-E233P-F234V-L235A-G236del mutations, advantageously an IgG1 antibody variant carrying in addition to the K409R mutation the novel L234F-L235E-G236R inactivating mutations in the heavy chain constant region, was efficiently abrogated.

[0212] [Example 8] Binding to neonatal Fc receptors by anti-human CD20 antibodies and their nonactivating variants. Neonatal Fc receptor (FcRn) is responsible for the long plasma half-life of IgG by protecting it from degradation. During antibody internalization, FcRn engages with the antibody Fc region in endosomes, where the interaction is stable in a weakly acidic environment (pH 6.0). Upon recycling to the plasma membrane, where the environment is neutral (pH 7.4), the interaction is disrupted and the antibody is released back into circulation. This affects the plasma half-life of IgG. Here, ELISA was performed to evaluate the binding of anti-human CD20 IgG1 and IgG4 antibodies and their variants containing non-activating mutations in the constant heavy chain region to human FcRn, as described in Examples 6 and 7.

[0213] Streptawell 96-well plates (Roche, Cat. No. 1734776001) were coated with 5 μg / mL (100 μL / well) of recombinantly produced biotinylated extracellular domain of human FcRn with a C-terminal His tag (FcRnECDHis; SEQ ID NO: 21) as a dimer containing β2 microglobulin (B2M; SEQ ID NO: 22) (FcRnECDHis-B2M-BIO), i.e., the extracellular domain of human FcRn, diluted in PBS supplemented with 0.05% Tween 20 (PBST) + 0.2% BSA for 2 h with shaking at room temperature (RT). Plates were then washed 3 times with PBST. Serially diluted antibody samples (final concentrations of 0.0013 to 20 μg / mL in 5-fold dilutions in PBST / 0.2% BSA, pH 6.0 or pH 7.4) were added and incubated for 1 h with shaking at room temperature. Plates were washed with PBST / 0.2% BSA, pH 6.0 or pH 7.4. Horseradish peroxidase (HRP)-conjugated polyclonal goat anti-human kappa light chain (1:5,000; Sigma, Cat. No. A-7164) diluted in PBST / 0.2% BSA, pH 6.0 or pH 7.4 was added and plates were incubated with shaking for 1 hour at room temperature. After washing with PBST / 0.2% BSA, pH 6.0 or pH 7.4, 100 μL of 2,2'-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid (ABTS; 1 mg / mL; Roche, Cat. Nos. 11112422001 and 11112597001) was added as substrate and the plate was incubated for 15 min at RT protected from light. The reaction was stopped with 100 μL of 2% oxalic acid (Riedel de Haen, Cat. No. 33506) and incubated for 10 min at room temperature. Absorbance was read at 405 nm using a microplate reader (BioTek, Winoosky, VT). Data were analyzed using a nonlinear agonist dose-response model and the ELISA background signal (no antibody control) was used as baseline and analyzed using GraphPad The area under the dose-response curve (AUC) per experimental replicate was calculated using log-transformed concentrations in PRISM (version 8.4.1, GraphPad Software). Data are based on three independent replicates.

[0214] FcRn binding ELISA assays showed that introduction of non-activating mutations in the constant heavy chain region of anti-human CD20 IgG1 or IgG4 antibodies did not inhibit FcRn binding at pH 6.0. Conversely, at pH 7.4, all tested IgG1 or IgG4 antibody variants, including anti-human CD20 antibody IgG1 and IgG4 wild type, showed no binding to human FcRn.

[0215] Taken together, these results indicate that the antibody IgG1 variant with the novel L234F-L235E-G236R inactivating mutations, like other variants tested, retained similar FcRn binding properties as the wild-type IgG1 antibody.

[0216] [Example 9] Induction of antibody-dependent cellular cytotoxicity by anti-human CD20 antibodies and their non-activating variants FcγR binding to non-activating anti-human CD20 antibodies and activation and signaling by non-activating anti-human CD20 antibodies were evaluated in Examples 6 and 7, respectively. Natural killer (NK) cell-mediated antibody-dependent cellular cytotoxicity (ADCC) is dependent on binding to FcγRIIIa. Here, anti-human CD20 IgG1 wild type, IgG1-K409R, L234F-L235E-D265A-K409R, L234F-L235E-G236R-K409R, L234A-L235A-P329G-K409R, G 236R-L328R-K409R, E233P-L234V-L235A-G236del-S267K-K409R, N297G-K409R, L234A-L235E-G237A-A330S-P The ability to induce ADCC by its inactivating variant with the 331S-K409R mutation as well as by IgG4 wild type, IgG4-S228P and inactivating variants with the S228P-F234A-L235A mutations, or the S228P-E233P-F234V-L235A-G236del mutation was assessed using the Perkin Elmer DELFIA® EuTDA TRF (time resolved fluorescence) cytotoxicity assay using CD20 expressing Raji cells and peripheral blood mononuclear cells (PBMCs) as a source of NK cells.

[0217] Buffy coats (Sanquin Blood Bank) were obtained from whole blood drawn from healthy volunteers, anticoagulated with citrate phosphate dextrose (20% final concentration (v / v)) to prevent coagulation activation, and diluted 3.6-fold in phosphate-buffered saline (PBS, Catalog No. SH3A3830.03, GE Healthcare). Peripheral blood mononuclear cells (PBMCs) were isolated from the PBS-diluted buffy coats by density gradient centrifugation using Leucosep™ tubes (Cat. No. 227290, Greiner Bio-One) containing Lymphocyte Separation Medium (Cat. No. 25-072-Cl, Corning) as described in the manufacturer's instructions with some modifications. Briefly, density gradient centrifugation was performed by centrifugation at 800×g for 20 min at room temperature (RT) with the centrifuge brake set at 3. The enriched cell fraction was then washed three times with 50 mL of PBS, followed by centrifugation at 300×g for 10 min. Isolated PBMCs were resuspended in culture medium (RPMI-1640 with 2 mM L-glutamine and 25 mM Hepes, Cat. No. BE12-115F, Lonza; supplemented with 10% donor bovine serum with iron, DBSI, Cat. No. 20371, Life Technologies). To determine the ADCC capacity of anti-CD20 wild-type antibodies and their non-activating variants, the DELFIA® EuTDA TRF (Time Resolved Fluorescence) Cytotoxicity Kit (Cat. No. AD0116, Perkin Elmer) was used according to the manufacturer's instructions. Two separate sets of experiments were performed.

[0218] For the first set of experiments, Raji cells were cultured at 1 × 10 6They were resuspended at a concentration of 10000 cells / mL and labeled with 0.16% (v / v) bis(acetoxymethyl) 2,2':6',2"-terpyridine-6,6"-dicarboxylate reagent solution (DELFIA BATDA Reagent, Cat. No. C136-100, Perkin Elmer) for 20 min at 37 °C in a water bath. The hydrophobic BATDA label can freely cross the cell membrane and is converted intracellularly to the hydrophilic TDA label (2,2':6',2"-terpyridine-6,6"-dicarboxylic acid), which no longer crosses the cell membrane. The labeled cells were then washed three times with culture medium to remove excess BATDA reagent. Subsequently, the cells were mixed with a concentration range of wild-type anti-CD20 antibody and its non-activated variant (final concentration of 0.01-10000 ng / mL; 10-fold dilutions) and incubated for 15 min at RT. Freshly isolated PBMCs were added to the mixture at an E:T ratio of 100:1 in culture medium in a total volume of 200 μL in a V-bottom 96-well plate (catalog no. 651101, Greiner bio-one). The plates were incubated at 4 °C for 1 h at 37 °C for 2 h at 37 °C for 1 h at 37 °C. 2 The plates were incubated for 2 hours at 37°C in a humidified (85%) air atmosphere at 10°C. Spontaneous BATDA release from the labeled cells, representing 0% ADCC, was determined in the absence of PBMCs and antibodies, and maximum BATDA release, representing 100% ADCC, was determined by incubating the labeled cells with DELFIA® lysis buffer (5% final concentration (v / v), Cat. No. 4005-0010, Perkin Elmer). After 2 hours, the plates were spun down at 500×g for 5 minutes and 20 μL of cell-free supernatant was transferred to a flat-bottom 96-well white opaque OptiPlate (Cat. No. 6005290, Perkin Elmer). Then, 200 μL of DELFIA Europium solution (Cat. No. C135-100, Perkin Elmer) was added to the transferred supernatant and the samples were incubated for 15 minutes at RT in the dark. The fluorescence of the EuTDA chelate formed from the released TDA label was measured in an EnVision multilabel plate reader (Perkin Elmer). The percentage specific release, as a measure of ADCC, was calculated using the following formula: % specific emission = ((fluorescent sample - fluorescence spontaneous emission) / (fluorescence maximum - fluorescence spontaneous emission)) x 100% Data were analyzed using a non-linear agonist dose-response model and the area under the dose-response curve (AUC) per experimental replicate was calculated using log-transformed concentrations in GraphPad PRISM (version 8.4.1, GraphPad Software) with the background strain (no antibody control) as baseline, followed by normalization per experimental replicate to the AUC value measured for the non-binding negative control IgG1-b12 (0%) and the AUC value measured for the wild-type IgG1 antibody variant (anti-human CD20 IgG1, 100%). Data are means (± SEM) obtained from four (wild-type and K409R variants) or two (L234F-L235E-D265A-K409R and L234F-L235E-G236R-K409R variants) independent donors.

[0219] A second set of experiments was performed and analyzed essentially as described above. However, instead of a concentration range, the ability to induce ADCC by anti-human CD20 IgG1 or IgG4 antibodies and variants thereof with non-activating mutations in the constant heavy chain region was evaluated at a final antibody concentration of 10ug / mL. Data were normalized to non-binding negative control IgG1-b12 (0%) and wild-type IgG1 antibody variant (anti-human CD20 IgG1, 100%) for each experimental replicate and visualized with GraphPad PRISM (version 8.4.1, GraphPad Software). Data are the mean (±SEM) obtained from six donors from two independent experiments.

[0220] Evaluation of NK cell-mediated ADCC using the Perkin Elmer DELFIA® EuTDA TRF cytotoxicity kit revealed that wild-type anti-human CD20 IgG1 or variants with the K409R mutation efficiently induced ADCC of CD20-expressing Raji cells (Figure 8). In contrast, consistent with the absence of FcγRIIIa binding and activation and signaling via it (Examples 6 and 7), anti-human CD20 wild-type IgG4 antibodies or variants with the S228P hinge stabilizing mutation did not induce NK cell-mediated ADCC (Figure 8B). Evaluation of the ADCC induction ability of non-activating antibody variants showed that the introduction of L234F-L235E-G236R and K409R mutations into the heavy chain constant region of anti-human CD20 IgG1 resulted in the induction of NK cell-mediated ADCC in anti-human CD20 IgG1 non-activating variant L234F-L235E-D265A-K409R (Figures 8A and 8B) and anti-human CD20 Similar to the IgG1 non-activating variants L234A-L235A-P329G-K409R, G236R-L328R-K409R, E233P-L234V-L235A-G236del-S267K-K409R, N297G-K409R, and L234A-L235E-G237A-A330S-P331S-K409R, the ADCC of CD20-expressing Raji cells was found to be reduced by approximately 69 to 98% (Figure 8B). Furthermore, similar to wild-type IgG4, the IgG4 non-activating antibody variants S228P-F234A-L235A or S228P-E233P-F234V-L235A-G236del were also unable to induce ADCC (Figure 8B).

[0221] In conclusion, the ability to induce ADCC was efficiently abolished by an antibody variant carrying novel L234F-L235E-G236R inactivating mutations in the heavy chain constant region.

[0222] [Example 10] T cell activation in PBMC cultures by anti-human CD3 antibodies and their non-activating variants Upregulation of CD69 on T cells was assessed as a measure of early T cell activation by anti-human CD3huCLB-T3 / 4 IgG1 and IgG4 antibodies and their variants with non-activating mutations in the constant heavy chain region by FACS analysis.

[0223] PBMCs were isolated from the buffy coat by density gradient separation as described in Example 9, washed with PBS, and resuspended in culture medium (RPMI-1640 with 2 mM L-glutamine and 25 mM Hepes, Cat. No. BE12-115F, Lonza; supplemented with 10% donor bovine serum with iron, DBSI, Cat. No. 20371, Life Technologies). Dose response series of anti-human CD3 IgG1-F405L and its non-activating variants with the L234F-L235E-D265A, L234F-L235E-G236R, L234A-L235A-P329G, G236R-L328R, E233P-L234V-L235A-G236del-S267K, or L234A-L235E-G237A-A330S-P331S mutations, anti-human CD3 with the hinge stabilizing mutation S228P in addition to F405L IgG4 or inactive variants carrying F405L and R409K plus S228P-E233P-F234V-L235A-G236del or S228P-F234A-L235A mutations were prepared in culture medium (final concentration of 0.001–1000 ng / ml, 10-fold dilutions) and incubated with PBMCs (1.5 × 10 5 The cells were added to wells of a 96-well round-bottom plate containing 100 μl of PBS (16–24 h of incubation) and 100 μl of PBS (1x PBS, Cat. No. BE17-517Q, Lonza; 0.1% bovine serum albumin, BSA; 0.02% NaN 3The cells were washed with FACS buffer (catalog no. 41920044-3, Bio-world) and stained with mouse anti-human CD28-PE (catalog no. 130-092-921; Miltenyi Biotec; T cell marker) and mouse anti-human CD69-APC antibodies (catalog no. 340560; BD Biosciences) for 30 min at 4 °C. Unbound antibodies were removed by washing twice with FACS buffer, and the cells were then resuspended in FACS buffer (150 μL / well) and the percentage of CD28-positive cells and CD69-positive cells in the PBMC mixture was measured with a Fortessa flow cytometer (BD). Data were analyzed using a CD28 + Cellular CD69 + The area under the dose-response curve (AUC) per PBMC donor for each experimental replicate was calculated using log-transformed concentrations in GraphPad PRISM (version 8.4.1, GraphPad Software) using background staining (no antibody control) as baseline, then normalized to the AUC value measured for the wild-type IgG1 antibody variant (IgG1-F405L, 100%) for each donor per experimental replicate. Data are means ± SEM from six donors from three independent replicates.

[0224] Assessment of CD69 upregulation on T cells as a measure of early T cell activation demonstrated that anti-human CD3 antibody IgG1, which has the L234F-L235E-G236R non-activating mutations in addition to the F405L mutation, significantly increased CD69 upregulation in comparison to the other variants tested, i.e. anti-human CD3 IgG1-F405L variants containing the non-activating mutations L234F-L235E-D265A, L234A-L235A-P329G, G236R-L328R, E233P-L234V-L235A-G236del-S267K, or L234A-L235E-G237A-A330S-P331S, as well as anti-human CD3 IgG1-F405L variants containing the non-activating mutations S228P-E233P-F234V-L235A-G236del. Figure 9A and 9B show that, like the IgG4-F405L-R409K variant, the anti-human CD3 IgG1 antibody variant with the N297G F405L mutation and the anti-human CD3 IgG4 antibody variant with the S228P-F234A-L235A F405L-R409K mutations prevented the upregulation of CD69 on T cells in PBMC cocultures. In contrast, the anti-human CD3 IgG1 antibody variant with the N297G F405L mutation and the anti-human CD3 IgG4 antibody variant with the S228P-F234A-L235A F405L-R409K mutations failed to prevent the upregulation of CD69 on T cells in PBMC cocultures.

[0225] In summary, T cell activation, as measured by CD69 upregulation in PBMC cocultures, can be prevented by introduction of novel L234F-L235E-G236R inactivating mutations in an anti-human CD3 IgG1-type antibody.

[0226] [Example 11] In vitro T cell-mediated cytotoxicity induced by non-activating antibody variants In Example 10, it was shown that introduction of the non-activating mutations L234F-L235E-G236R in the constant heavy chain region of an IgG1 variant of a CD3-targeting antibody can efficiently prevent T cell activation. Here, T cell-mediated cytotoxicity was evaluated for CD3xHER2, CD3xb12 and b12xHER2 bispecific wild-type IgG1 and IgG4 antibodies and their variants with non-activating mutations in the constant heavy chain region.

[0227] Bispecific molecules were generated by controlled Fab arm exchange (cFAE) as described in Example 1. T cell-mediated cytotoxicity with wild-type bispecific antibodies CD3xHER2 (anti-human CD3 (huCLB-T3 / 4) IgG1 or IgG4 and anti-human HER2 IgG1 or IgG4), CD3xb12 (anti-human CD3 (huCLB-T3 / 4) IgG1 or IgG4 and non-binding control antibody anti-HIV1 gp120 (b12) IgG1 or IgG4), or b12xHER2 (non-binding control antibody anti-HIV1 gp120 (b12) IgG1 or IgG4 and anti-human HER2 IgG1 or IgG4) and their variants with non-activating mutations in the constant heavy chain region was evaluated. PBMCs were isolated from buffy coats derived from healthy donors by density gradient separation as described in Example 9, washed with PBS, and resuspended in culture medium (RPMI-1640 with 2 mM L-glutamine and 25 mM Hepes; supplemented with 10% donor calf serum with iron (DBSI)). HER2-expressing SK-OV-3 cells (catalog no. HTB-77, ATCC) were cultured in McCoy's 5A medium (Lonza, catalog no. BE12-168F) supplemented with 10% (vol / vol) heat-inactivated DBSI and penicillin-streptomycin (Pen / Strep, final concentration 50 units / mL penicillin potassium and 50 μg / mL streptomycin sulfate (Lonza, catalog no. DE17-603E) at 5% (vol / vol) CO. 2 The cells were maintained at 37°C in a humidified incubator. SK-OV-3 cells were cultured to near confluence. The cells were trypsinized, resuspended in culture medium, and then passed through a cell strainer to obtain a single cell suspension. 2.5 × 10 4 Seed SK-OV-3 cells into each well of a 96-well culture plate and culture the cells at 37 °C and 5% CO. 2 The cells were incubated at 4 °C for 4 h to allow them to adhere to the plate. Then, 1 × 10 cells were added to each well of the 96-well plate containing the SK-OV-3 target cells. 5PBMCs were added to obtain an effector to target (E:T) ratio of 4:1. Subsequently, dose response series of bispecific CD3xHER2, CD3xb12 and b12xHER2 wild type and its non-activating variants as described above were prepared in culture medium (0.001-1000ng / mL final concentration, 10-fold dilutions) and added to wells of a 96-well culture plate containing SK-OV-3 cells and PBMCs. Incubation of SK-OV-3 target cells with 2μM staurosporine (catalog no. S6942-200, Sigma) was used as a reference for 100% tumor cell killing. A medium control (SK-OV-3 cells, no antibody, no PBMCs) was used as a reference for 0% tumor cell killing. The plates were incubated at 37°C, 5% CO 2 After 3 days, the plates were washed twice with PBS and 150 μL of culture medium containing 10% Alamar Blue (Invitrogen, Cat. No. DAL1100) was added to each well. The plates were incubated at 37° C., 5% CO 2 The cells were incubated at 4°C for 4 hours. Absorbance at 590 nm was measured (Envision, Perkin Elmer, Waltham, MA). Data were visualized in GraphPad PRISM (version 8.4.1, GraphPad Software) as dose response to the percentage of viable SK-OV-3 cells, calculated for each donor per experimental replicate. Data are means ± SEM from six donors from three independent replicates.

[0228] All bispecific CD3xHER2 antibody variants with non-activating mutations in the constant heavy chain region induced dose-dependent cytotoxicity of SK-OV-3 cells with an efficiency comparable to that of bispecific CD3xHER2 antibody variants with Fc regions with wild-type-like functions and therefore without non-activating mutations (Figure 10A). Wild-type-like bispecific CD3xb12 antibodies induced non-specific killing of SK-OV-3 cells, albeit to a lesser extent than wild-type-like bispecific CD3xHER2 antibody variants. In contrast, bispecific CD3xb12 antibody variants with L234F-L235E-G236R non-activating mutations did not show cytotoxicity of SK-OV-3 cells, similar to other non-activating bispecific CD3xb12 variants tested, except for CD3xb12 bispecific antibodies with N297G mutations, which still showed partial non-specific cytotoxicity of SK-OV-3 cells at the highest concentration tested (Figure 10B). This is consistent with the observed activation of T cells in PBMC cultures (upregulation of CD69) observed for this variant in Example 10. A slight non-specific cytotoxicity of SK-OV-3 cells was observed with the wild-type-like bispecific b12xHER2 antibody variant. No non-specific cytotoxicity was observed with the other bispecific b12xHER2 antibody variants tested, except for the residual non-specific cytotoxicity with the bispecific b12xHER2 variant carrying the E233P-L234V-L235A-G236del-S267K mutations (Figure 10C).

[0229] Overall, the introduction of novel L234F-L235E-G236R mutations in the constant heavy chain region of a bispecific antibody variant targeting a cancer antigen and T cells allows to efficiently circumvent non-specific cytotoxicity and retain the ability to induce specific T cell-mediated cytotoxicity.

[0230] [Example 12] Pharmacokinetic (PK) analysis of non-activating antibody variants The pharmacokinetic properties of anti-human CD3 (huCLB-T3 / 4) and anti-human CD20 IgG1 antibody variants with inactivating mutations in the constant heavy chain region were analyzed in mouse studies.

[0231] The mice in this study were housed at the Central Laboratory Animal Facility (Utrecht, the Netherlands). All mice were individually housed in ventilated cages and provided with food and water ad libitum. All experiments were conducted in accordance with the Dutch Animal Protection Act (WoD), translated from Directive (2010 / 63 / EU), and approved by the Dutch Central Committee on Animal Experiments and the local ethics committee. SCID mice (C.B-17 / IcrHan@Hsd-Prkdc<scid, Envigo) were injected intravenously with 500 μg of antibody (wild-type anti-human CD3 IgG1, its variants having the F405L mutation alone or in combination with the inactivating mutations L234F-L235E-D265A or L234F-L235E-G236R, wild-type anti-human CD20 IgG1, and its variants having the K409R mutation alone or in combination with the inactivating mutations L234F-L235E-D265A or L234F-L235E-G236R) using three mice per group. Blood samples (50 μL) were collected from the facial vein 10 minutes, 4 hours, 1 day, 2 days, 8 days, 14 days, and 21 days after antibody administration. Blood was collected into heparin-containing vials and subsequently centrifuged at 10,000 g for 5 minutes. Plasma was stored at -20 °C until determination of antibody concentration.

[0232] Specific human IgG concentrations were determined by total hIgG ELISA. Anti-human IgG (catalog number M9105, lot number 8000260395 Sanquin, The Netherlands) coated at a concentration of 2 μg / mL on 96-well Microlon ELISA plates (Greiner, Germany) was used as capture antibody. Plates were then blocked with PBS supplemented with 0.2% BSA, after which samples were added, serially diluted in ELISA buffer (PBS supplemented with 0.05% Tween 20 and 0.2% bovine serum albumin) and incubated for 1 h at room temperature (RT) on a plate shaker. Plates were then incubated with goat anti-human IgG immunoglobulin (catalog number 109-035-098, Jackson) and developed with 2,2'-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid; ABTS; Roche). The reaction was stopped by adding 2% oxalic acid (Riedel de Haen, Cat. No. 33506). A standard curve was run with each material injected. Absorbance was measured at 405 nm in a microplate reader (Biotek, Winooski, VT).

[0233] Serum IgG concentrations of anti-human CD20 IgG1 antibodies and their variants carrying the K409R, L234F-L235E-D265A-K409R, or L234F-L235E-G236R-K409R mutations (Figure 11A) and anti-human CD3 IgG1 antibodies and their variants carrying the F405L, L234F-L235E-D265A-F405L, or L234F-L235E-G236R-F405L mutations (Figure 11B) were comparable. Plasma IgG concentrations measured for all anti-human CD20 IgG1 and anti-human CD3 IgG1 antibody variants injected into mice were consistent with those predicted by the two-compartment model pharmacokinetics of wild-type human IgG1 in SCID mice. The calculated clearance values ​​for all anti-human CD20 IgG1 and anti-human CD3 IgG1 antibody variants were similar to their wild-type counterparts (FIG. 11C; 21 days post-injection). The variation between groups mainly reflects differences in distribution phase that may be caused by glycosylation differences between batches and may not be format related. t1 / 2 beta appeared to be comparable for all mutations.

[0234] In summary, introduction of inactivating mutations, including the novel L234F-L235E-G236R mutations, does not affect the pharmacokinetics of IgG1 antibody variants in mice.

[0235] [Example 13] Immunogenicity assessment of antibody variants harboring L234F-L235A-G236R inactivating mutations in the Fc region The clinical immunogenicity risk of the constant domain of an IgG1 antibody variant with the L234F-L235E-G236R mutations (IgG1-FER) was assessed using Abzena's iTope™ and TCED™ in silico technologies.

[0236] iTope™ software predicts favorable interactions between the amino acid side chains of all possible 9-mer peptides in a test protein sequence and the open-ended binding grooves of 34 human MHC class II alleles (Perry et al., 2008, Drugs RD 9(6), pp.385-96). The alleles selected represent the most common HLA-DR alleles found worldwide, with no weighting due to those most commonly found in any particular ethnic group. By comparing these predictions with the MHC class II binding sequences present in the wild-type human IgG1(m)f reference, it is possible to identify novel binding sequences. Peptides predicted to bind with moderate and high affinity to 50% or more of the MHC class II alleles are considered to correlate with the presence of T cell epitopes (Hill et al., 2003, Arthritis Res Ther 5(1), pp.R40-8). TCED™ is a database of known T cell epitopes identified by ex vivo immunogenicity studies using a variety of proteins, primarily antibodies (Bryson et al., 2010, BioDrugs 24(1), pp. 1-8). The database is interrogated by BLAST search to see if peptides with predicted promiscuous medium or high affinity are also present in the database.

[0237] Each 9-mer was scored based on its potential "fit" and interaction with MHC class II molecules. Peptide scores calculated by the software are between 0 and 1. Peptides that yielded high average binding scores (>0.55 with the iTope™ scoring function) were highlighted, and if more than 50% of the MHC class II binding peptides (i.e., 17 out of 34 alleles) had high binding affinity (score >0.6), such peptides were selected as candidates for CD4 +Promiscuous high affinity MHC class II binding peptides were defined as those considered at high risk of containing a T cell epitope. Promising moderate affinity MHC class II binding peptides bind to ≥ 50% of alleles with a binding score of > 0.55 (without a majority of > 0.6). Further analysis of the sequences was performed using TCED™. These criteria were modified in the case of a large aromatic amino acid (i.e., F, W, Y) present at the p1 anchor position where 20 of the 34 alleles had an open p1 pocket allowing binding of the large aromatic residue. If this occurs, the promiscuous peptide is defined as binding to ≥ 10 of the subset of 20 alleles. Sequences were used to interrogate TCED™ by BLAST search to identify high sequence homology between peptides (T cell epitopes) derived from unrelated proteins / antibodies that stimulated T cell responses in previous ex vivo studies.

[0238] iTope™ analysis did not identify any promiscuous medium or high affinity MHC class II binding sequences present in IgG1-FER and absent in wild type IgG1(m)f. Therefore, based on this analysis, there is no apparent increased risk of clinical immunogenicity for IgG1-FER based antibodies.

[0239] [Example 14] Glycoanalysis of non-activated antibody variants IgG molecules have a single conserved Asn(N)-linked glycosylation site in the CH2 region (at position N297 for IgG1). The core structure of this Asn-linked glycan is composed of N-acetylglucosamine (GlcNAc) and mannose residues. Further elongation can be performed with galactose, sialic acid, core fucosylation and bicleavable GlcNAc (Vidarsson et al., 2014, Front. Immunology Oct 20;5:520), resulting in IgG1 molecules heterogeneously glycosylated at N297. Glycans play an important role in protein conformation, stability and biological function (Costa et al., Crit Rev Biotechnol 34(4):281-99(2014)). Therefore, glycosylation changes in IgG1 molecules are undesirable, as efficacy and pharmacokinetic properties may be affected. Furthermore, glycan heterogeneity needs to be analyzed and controlled during production, and charged glycans can further increase the complexity of charge-based analytical assays such as imaging capillary isoelectric focusing (iCIEF) used for release testing or characterization.

[0240] In this example, N-linked glycan profiling was performed on wild-type anti-human CD20 IgG1 antibody and its non-activated variants with L234F-L235E-D265A or L234F-L235E-G236R mutations in addition to the K409R mutation. Additionally, glycan profiling was performed on anti-human CD3 (huCLB-T3 / 4) IgG1 antibody with L234F-L235E-F405L mutations or its variants further with D265A or G236R mutations.

[0241] Assessment of IgG1 N-linked glycosylation was performed by two different methods as indicated (Table 2). Anti-human CD20 IgG1 wild type, anti-human CD20 IgG1 with L234F-L235E-D265A-K409R mutations, and anti-human CD3 IgG1 variants with either L234F-L235E-F405L or L234F-L235E-D265A-F405L mutations were analyzed by 2-aminobenzamide (2-AB) labeling. N-linked glycan profiling was performed by normal-phase HPLC analysis of 2-AB labeled glycans using a Waters Alliance 2695 Separations Module (Waters). N-linked glycans were released from the antibodies by incubation with peptide-N-glycosidase F (catalog no. GKE-5006D, PROzyme). The antibodies were then ethanol precipitated (ice-cold) and removed. The supernatant containing the released glycans was dried under vacuum. The resulting glycans were solubilized and subsequently labeled with fluorophore 2-AB (from the LudgerTag 2-AB Glycan Labeling Kit, Cat. No. LT-KAB-A2, Ludger) label at the reducing end by reductive amination for HPLC analysis. HPLC profiles were then obtained using gradient elution in conjunction with fluorescence detection. The integral of the HPLC peak intensity was a measure of the percentage molar abundance of individual N-linked glycans (e.g., G0F, G1F, G2F, etc.) relative to the total population of oligosaccharides. Peaks were assigned based on an external glycan standard mixture (prepared by mixing distinct glycans from Ludger: NGA2 (=G0, #CN-NGA2-20U), NGA2F (=G0F, #CN-NGA2F-20U), NA2 (=G2, #CN-NA2-20U), NA2F (=G2F, #CN-NA2F-20U), MAN5 (#CN-MAN5-20U), MAN6 (#CN-MAN6-20U), MAN8 (#CN-MAN8-20U), MAN9 (#CN-MAN9-20U), A1 (#CN-A1-20U), A1F (#CN-A1F-20U), A2 (#CN-A2-20U), A2F (#CN-A2F-20U)).

[0242] Anti-human CD20 IgG1 variants with K409R plus L234F-L235E-G236R mutations and anti-human CD3 IgG1 variants with F405L plus L234F-L235E-G236R mutations were analyzed by liquid chromatography-mass spectrometry (LC-MS) analysis using an Orbitrap Q-Exactive Plus mass spectrometer (Thermo Fischer). Identification and relative quantification of N-linked glycosylation was performed on reduced glycosylated antibody heavy chains. Given the known primary amino acid sequence of the heavy chains, the mass identity of the attached N-linked glycans could be identified. Briefly, antibody samples were diluted to 200 μg / mL in PBS pH 7.4 (cat. no. 3623140, B. Braun) to a total volume of 50 μL. Next, 1 μL of 1 M dithiothreitol (DTT; Cat. No. D9163, Sigma) was added to 50 μL of sample and incubated at 37 °C for 1 h. Samples were then transferred to glass Qsert vials (Cat. No. 186001128c, Waters) and loaded onto the LC-MS. 1 μL was injected onto the LC column and antibodies were eluted from 23% (B) to 95% (B) using a mobile phase A (Milli-Q water with 0.1% formic acid, Cat. No. 56302-50ml-F, Fluka)-mobile phase B (0.1% formic acid in acetonitrile, Cat. No. 0001934101BS, Bio Solve) gradient at a flow rate of 0.2 mL / min in 2 min. The resulting raw m / z spectra were deconvoluted using Protein Deconvolution 4.0 (Thermo Fisher) software. As a result, the deconvoluted spectrum resulted in a reduction in glycosylated heavy chain mass. The resulting peak intensities (after spectral deconvolution) were a measure of the calculated percentage of molar abundance of individual N-linked glycans (e.g., G0F, G1F, G2F, etc.) relative to the total population of oligosaccharides. Figure 12 shows a schematic of the detected glycan species.

[0243] Knowing the relative abundance of the different oligosaccharides, coupled with knowledge of their structures (see Figure 12), allows the calculation of the total amount of galactosylated and charged glycans relative to the amount in the A2F structure. The amounts calculated here as percentages represent molar amounts, i.e. number of molecules and not mass. From Table 2, the wild-type IgG1 sequence is calculated to have a charged glycan abundance and galactosylation rate of 0-1% and 15-25%, respectively. The percentage of charged glycans and galactosylation of the non-activating L234F-L235E-D265A mutant was about 10% and about 60%.

[0244] Evaluation of glycosylation of anti-human CD20 IgG1 antibody variants (Table 2) revealed that introduction of the inactivating L234F-L235E-D265A mutation in combination with the K409R mutation increased galactosylation, denoted as G1 or G2 and its variants (G: galactose), and increased the presence of charged glycans, denoted as A1 or A2 and its variants (A: sialic acid), compared to the anti-human CD20 IgG1 wild-type antibody. A similar pattern was observed for the anti-human CD3 IgG1 antibody variants carrying the inactivating L234F-L235E-D265A mutation in addition to the F405L mutation. As the anti-human CD3 IgG1 antibody variants carrying the L234F-L235E-F405L mutation showed a glycosylation pattern similar to that of the anti-human CD20 IgG1 wild-type antibody, the increased galactosylation and the presence of charged glycans can be assigned to the presence of the D265A mutation. Consistent with this, introduction of the non-activating mutations L234F-L235E-G236R into either anti-human CD20 IgG1-K409R or anti-human CD3 IgG1-F405L did not result in increased galactosylation and increased presence of charged glycans.

[0245] In summary, introduction of the non-activating L234F-L235E-D265A mutations into the constant heavy chain region of an IgG1 antibody increases antibody glycosylation heterogeneity, increasing galactosylation and increasing the presence of charged glycans, whereas introduction of the non-activating L234F-L235E-G236R mutations into an IgG1 antibody allows retaining a glycan profile comparable to the wild-type constant region of IgG1. [Table 4] Table 2 shows the glycosylation analysis of anti-human CD20 IgG1 and anti-human CD3 IgG1 antibody variants with non-activating mutations in the constant heavy chain region. Anti-human CD20 IgG1 wild type (wt) or variants with L234F-L235E-D265A-K409R mutations, and anti-human CD3 IgG1 with L234F-L235E-F405L or L234F-L235E-D265A-F405L mutations were analyzed by 2-AB labeling. The glycan profiles of anti-human CD20 IgG1 and anti-human CD3 IgG1 antibody variants with L234F-L235E-G236R-K409R or L234F-L235E-G236R-F405L mutations, respectively, were evaluated by liquid chromatography-mass spectrometry (LC-MS). In contrast to the Orbitrap LC-MS method, the 2-AB labeling method is not able to assign G2 and mannose 6 and A2F and mannose 9 (not applicable; na) separately, so the sum of both is shown. G0F-GN, G1F-GN, and Man7 were not evaluated by the 2-AB method and are stated as not applicable (na). The variants tested were IgG1, IgG1-FER-K409R, IgG1-FEA-K409R, IgG1-FE-F405L, IgG1-FER-F405L, IgG1-FEA-F405L, FER:L234F-L235A-G236R, FE:L234F-L235E, and FEA:L234F-L235E-D265A. A schematic diagram of the detected glycan species is shown in Figure 12.

[0246] [Example 15] Efficiency of controlled Fab arm exchange with non-activating antibody variants Certain applications, such as T cell-mediated cytotoxicity of target cells as shown in Example 11, require the generation and use of bispecific antibody (bsAb) variants in which one F(ab) arm can engage target A and the other F(ab) arm engages target B.

[0247] Here, we evaluated the efficiency with which introducing a non-activating mutation in the constant heavy chain region of the parent monospecific antibodies, in addition to either the F405L or K409R mutation present in one of the monospecific antibodies, respectively, required for efficient heterodimer formation, would generate bsAb variants by controlled Fab arm exchange (cFAE) as described in Example 1. Briefly, the antibodies were mixed at equimolar concentrations and incubated with 75 mM 2-mercaptoethylamine-HCl (2-MEA; Cat. No. 30078, Sigma Aldrich) for 5 hours at 31° C. Afterwards, buffer exchange against PBS was achieved as described in Example 1 and concentrations were measured by absorbance at 280 nm using a NanoDrop ND-2000-EU spectrophotometer (Thermo Fisher). Mass spectrometry was performed to determine bispecific and residual homodimer content using an Orbitrap Q-Exactive Plus mass spectrometer (Thermo Fischer).

[0248] Evaluation of the efficiency of cFAE revealed that the introduction of the non-activating mutation L234F-L235E-G236R in addition to the F405L or K409R mutations, each present in one of the monospecific antibodies, resulted in an efficient cFAE, with at least 95% of the population being IgG1bsAb (BisG1), and the remaining population being one or both of the monospecific antibodies (shown in the figures as IgG1-A and IgG1-B; FIG. 13A). Furthermore, the efficiency of cFAE was evaluated when bsAb variants with an asymmetric distribution of non-activating mutations, i.e. L234F-L235E-G236R, present in one monospecific antibody variant in addition to the F405L or K409R mutations, and L234F-L235E-D265A, present in the other monospecific antibody variant in addition to the F405L or K409R mutations, were generated. Analysis revealed similar efficiency in generating bsAb variants (Figure 13B,C), as shown for variants harboring L234F-L235E-G236R inactivating mutations in both arms of the bsAb in addition to the F405L or K409R mutations, each present in one of the monospecific antibodies (Figure 13A).

[0249] In summary, antibody variants carrying the novel L234F-L235E-G236R inactivating mutations in the IgG1 constant heavy chain region retain the ability to efficiently form bsAb variants by controlled Fab arm exchange using the F405L and K409R mutations, each present in one of the monospecific antibodies, allowing heterodimerization and preventing the formation of monospecific homodimers.

[0250] [Example 16] Production levels of non-activating antibody variants In this example, production levels of antibody variants carrying either the F405L or K409R mutation as well as either the L234F-L235E-G236R or L234F-L235E-D265A inactivating mutations in the constant heavy chain region were assessed.

[0251] All antibody variants were produced in Expi293F cells as described in Example 1. To avoid differences in production titers due to the use of a particular clone for one non-activating antibody variant but not the other, and to allow direct comparison of production titers of antibody variants carrying either the L234F-L235E-D265A or L234F-L235E-G236R non-activating mutations in the constant heavy chain region, the same antibody clone is represented within each group, i.e., F405L or K409R antibody variant.

[0252] Analysis of production levels revealed that inactivating antibody variants with L234F-L235E-G236R mutations in addition to F405L mutations (black circles) were produced at similar levels to antibody variants with L234F-L235E-D265A inactivating mutations in addition to F405L mutations (white circles; Figure 14). Production levels of antibody variants with L234F-L235E-G236R inactivating mutations in addition to K409R mutations (black squares) were similar to antibody variants containing F405L mutations in addition to inactivating mutations, whereas a slight increase in average production levels was observed for antibody variants with L234F-L235E-D265A inactivating mutations in addition to K409R mutations (white squares; Figure 14).

[0253] In summary, no significant differences were observed in the production levels of non-activating antibody variants carrying either the F405L or K409R mutation plus the L234F-L235E-G236R mutation compared to variants carrying the F405L or K409R mutation plus the L234F-L235E-D265A non-activating mutation.

[0254] [Example 17] Assessment of stability and solubility of IgG1 non-activated antibody variants by PEG midpoint, DLS and DSF analysis The protein stability and solubility properties of anti-CD20 and anti-CD3 IgG1 antibody variants with non-activating mutations in the constant heavy chain region were evaluated using PEG-induced precipitation assays, differential scanning fluorimetry (DSF) and dynamic light scattering (DLS) assays.

[0255] Samples of anti-human CD20 and anti-human CD3 IgG1 (huCLB-T3 / 4) antibody variants were formulated in PBS pH 7.4 at a concentration of approximately 20 mg / mL (concentration range 18.7-21.6 mg / mL; filtration applied for anti-human CD3 IgG1 antibody variants). PBS was used as a non-optimal formulation to allow for better comparison of protein stability.

[0256] To assess the stability of conformational proteins, DSF was performed in an iQ5 Multicolor Real-Time PCR Detection System (Bio-Rad) capable of detecting changes in fluorescence intensity caused by the binding of the exogenous dye Sypro-Orange (5000-fold concentrated in DMSO, Cat. No. S5692, Sigma-Aldrich) to hydrophobic regions exposed by denatured IgG. Sypro-Orange was diluted 320-fold in PBS (Hyclone GE Healthcare, Cat. No. SH3A383.03) pH 7.4 or 30 mM sodium acetate pH 4 (Cat. No. 25022-1KG-R, Sigma-Aldrich) to a concentration of 75 mM. Thermal melting curves can be derived from measuring the increasing fluorescence during controlled stepwise thermal denaturation of the analyzed IgG. Thus, 5 μL samples (diluted in PBS; concentration range 1 mg / mL + / - 10%) of anti-human CD20 IgG1 antibody variants with either the K409R mutation, the L234F-L235E-D265A-K409R or the L234F-L235E-G236R-K409R mutations, or of anti-human CD3 IgG1 antibody variants with the F405L mutation, the L234F-L235E-D265A-F405L or the L234F-L235E-G236R-F405L mutations, mixed with 20 μL of 75 mM Sypro-Orange (in either PBS pH 7.4 or 30 mM sodium acetate pH 4) were prepared in duplicate in an iCycler iQ 96 well PCR plate. Fluorescence (excitation 485 nm, emission 575 nm) was recorded at increasing temperatures ranging from 25° C. to 95° C. with stepwise increments of 0.5° C. / increment and duration of 15 seconds plus the time required to record the fluorescence of all wells. Data was analyzed using Bio-Rad CFX Manager Software 3.0 and melting points were determined by the software from a graph of fluorescence versus temperature.

[0257] DLS analysis was performed to evaluate the tendency of the above mentioned antibody variants to aggregate in solution, as a measure of colloidal stability. 20 μL of the above mentioned antibody variants (concentration range 1 mg / mL + / - 10%) in PBS pH 7.4 were analyzed using Dynamics 7 software with a DynaPro Plate Reader II (Wyatt Technology). Samples were applied in triplicate to circular 384-well IQ-LV plates (Aurora Biotechnologies, Cat. No. 1011-00110), centrifuged for 3 min at 2,111 x g and covered with paraffin oil. Before measurements, the plates were centrifuged again for 3 min at 2,111 x g. Thermal scan measurements were performed at continuously increasing temperatures (1 °C increase / data point for each particular sample) throughout the experiment. The data were analyzed using a cumulant fitting procedure to determine the apparent radius. A cutoff value of 2.1 nm was used, excluding peaks with lower radii, which are often caused by software artifacts and are not consistently observed in each acquisition. A refractive index of 1.333 and a viscosity of 1.019 cP were used in PBS buffer at 25 °C (standard values ​​provided by the Dynamics software). Data were processed in Microsoft Excel to determine the onset of aggregation (T agg ) was determined by calculating the mean and standard deviation of the radius for all wells (n=10, first 10 measurements). agg The 99.99% confidence interval for each measurement was set individually, and the first consecutive value (from 25°C to 80°C) that fell outside the confidence interval was determined as T agg Five consecutive measurements were performed in triplicate per experiment.

[0258] A PEG-induced precipitation assay was performed to assess relative protein solubility. Two buffers were prepared. Buffer A: 50 mM phosphate buffer pH 7.0 (sodium phosphate monobasic; Fluka, catalog no. 17844 + sodium phosphate dihydrate, Fluka catalog no. 71633); Buffer B: 50 mM phosphate buffer pH 7.0 + 40% (w / v) PEG 8000 (Sigma-Aldrich, catalog no. P5413). Different amounts of buffer B were mixed with buffer A to generate a series of 11 different PEG concentrations ranging from 0% to 40% PEG. An aliquot (80 μL) of each PEG concentration buffer was added to a different well of a 96-well plate (UV Star® 96-well plate, half area, Greiner Bio-one, catalog no. 675801). Of each antibody sample (diluted to 1 mg / mL in PBS), 20 μL was added to the wells containing the series of PEG-containing buffers. The plate was covered with a seal and shaken at 750 rpm for 5 minutes. The plate was left at room temperature overnight followed by shaking for 5 minutes. The plate was then centrifuged at 4000 rpm for 20 minutes to remove the precipitated antibody. From each well, 80 μL was carefully transferred to a new UV Star 96-well plate without disturbing the centrifuged precipitate. The absorbance at 280 nm (A280, which correlates with the amount of solubilized protein) was measured on a Synergy™ 2 multimode microplate reader (Biotek Instruments, BioSPX), recalculated to a volume of 100 μL (path length correction) and the blank value (PEG without antibody) was subtracted. The corrected A280 values ​​were plotted against the PEG concentration in Graphpad Prism 8. The data was analyzed using nonlinear regression and the PEG midpoint (%) reflects the concentration of the test sample at which 50% of the antibody was precipitated (i.e., 50% loss in A280 compared to 0% PEG). The PEG midpoints were used as a measure of solubility, with higher PEG midpoints corresponding to better solubility.

[0259] The results of the above assays are summarized in Tables 3 and 4. DSF analysis revealed a melting temperature (T m ) was found, which has a T of 56.0°C at pH 4.0. mT of 1 and 62.0°C m 2. T recorded for anti-human CD20 IgG1-L234F-L235E-G236R-K409R at pH 7.4 m At pH 4.0, the variant with the L234F-L235E-G236R-K409R mutations had one T at 58.0 °C, which was comparable to that of the K409R-containing variant (67.8 °C vs. 68.0 °C, respectively). m In contrast, the anti-human CD20 IgG1 variant with the L234F-L235E-D265A-K409R mutations showed a T of 63.3°C at pH 7.4. m T of 1 and 68.5°C m 2 decrease, whereas at pH 4.0 there are three T m For the anti-human CD3 IgG1 antibody, variants with either the F405L mutation alone or the L234F-L235E-G236R-F405L mutations had the same T at pH 7.4 (68.0°C). m and had very similar T m 1 and T m At pH 7.4, the T of the IgG1-huCLB-T3 / 4-L234F-L235E-D265A-F405L antibody variant was m was decreased (63.0 °C) compared to the F405L and 234F-L235E-G236R-F405L-containing variants. Also, the first T recorded for the variant IgG1-huCLB-T3 / 4-L234F-L235E-D265A-F405L at pH 4.0 was m was decreased (48.5°C) compared to the other IgG1-huCLB-T3 / 4 variants, whereas the second T of 71.0°C m was observed for IgG1-huCLB-T3 / 4-L234F-L235E-D265A-F405L, which is consistent with other IgG1-huCLB-T3 / 4 variants.

[0260] The anti-human CD20 IgG1-K409R and anti-human CD20 IgG1-L234F-L235E-D265A-K409R antibody variants have the lowest aggregation temperatures (T agg 57.9°C and 58.5°C, respectively), followed by the L234F-L235E-G236R-K409R-containing variant (59.9°C). For the IgG1-huCLB-T3 / 4 variants, the lowest T was observed for the variant harboring the L234F-L235E-D265A-F405L mutation (58.7°C). agg was observed, followed by the F405L-containing variant and the L234F-L235E-G236R-F405L-containing variant (61.7°C and 62.0°C, respectively).

[0261] Anti-human CD20 IgG1 antibody variants with either the K409R, L234F-L235E-D265A-K409R or L234F-L235E-G236R-K409R mutations showed comparable relative solubility profiles as determined by PEG-induced precipitation assays. Overall, anti-human CD3 IgG1 variants showed lower relative solubility compared to anti-human CD20 IgG1 variants. Anti-human CD3 IgG1 variants with the L234F-L235E-D265A-F405L or L2345F-L23E-G236R-F405L mutations were comparable in terms of solubility. Both of these variants were comparatively slightly less soluble than the anti-human CD3 IgG1-F405L variant.

[0262] In summary, anti-human CD20 IgG1 antibody variants with non-activating mutations showed comparable profiles in terms of solubility and aggregation tendency compared to the K409R-containing control variant. However, variants with L234F-L235E-G236R-K409R mutations showed a comparable protein stability profile to the K409R-containing control variant, whereas the anti-human CD20 IgG1-L234F-L235E-D265A-K409R variant showed a lower protein stability profile. For anti-human CD3 IgG1 variants, variants with F405L or L234F-L235E-G236R-F405L mutations showed comparable aggregation tendency and protein stability profiles. The solubility of both non-activating variants was reduced compared to the F405L-containing control variant. Variants with the L234F-L235E-D265A-F405L mutations showed reduced protein stability and a slightly higher tendency to aggregate compared to variants with the F405L or L234F-L235E-G236R-F405L mutations. The solubility of the L234F-L235E-D265A-F405L-containing variant was comparable to that of the L234F-L235E-G236R-F405L-containing variant.

[0263] In conclusion, when formulated at high concentrations in PBS, anti-human CD20 and CD3 IgG1 variants with L234F-L235E-G236R mutations exhibited a more robust protein stability profile than the L234F-L235E-D265A-containing variants. Furthermore, the anti-human CD3 IgG1 variant with L234F-L235E-G236R mutations showed a slightly lower aggregation tendency than the L234F-L235E-D265A-containing variants.

[0264] Table 3 shows the protein conformational stability determined by DSF analysis for anti-human CD20 IgG1 and anti-CD3 IgG1 (huCLB-T3 / 4) antibody variants with non-activating mutations in the constant heavy chain region. DSF: Differential Scanning Fluorescence; m :Melting temperature.

[0265] Table 4 shows the protein solubility, as determined by the PEG midpoint assay, and the aggregation tendency, as determined by DLS analysis, for anti-human CD20 IgG1 and anti-human CD3 IgG1 antibody variants with non-activating mutations in the constant heavy chain region. agg : aggregation temperature; PEG: polyethylene glycol; DLS: dynamic light scattering. [Table 5] [Table 6] [Example 18] Effect on protein stability of IgG1 non-activating antibody variants after storage at different temperatures for 1 or 4 months The protein stability and solubility profiles of anti-CD20 and anti-CD3 IgG1 antibody variants with non-activating mutations in the constant heavy chain region were evaluated in Example 17. Here, the protein stability of such antibody variants was evaluated after storage at 2-8°C or 40°C for 1 or 4 months using different assays.

[0266] Samples of anti-human CD20 IgG1 and anti-human CD3 IgG1 (huCLB-T3 / 4) antibody variants were formulated in PBS pH 7.4 at a concentration of approximately 20 mg / mL (concentration range 18.7-21.6 mg / mL; filtration applied to anti-human CD3 IgG1 antibody variants). PBS was used as a non-optimal formulation to allow better comparison of protein stability. Anti-human CD20 IgG1 variants had either K409R mutations, L234F-L235E-D265A-K409R or L234F-L235E-G236R-K409R mutations, whereas anti-human CD3 IgG1 antibody variants had either F405L mutations, L234F-L235E-D265A-F405L or L234F-L235E-G236R-F405L mutations. Samples were incubated at 2–8° C for 1 month, at 40° C for 1 month, at 2–8° C for 4 months, and at 40° C for 4 months. All samples were then analyzed by high-performance size-exclusion chromatography (HP-SEC), capillary isoelectric focusing (cIEF), capillary electrophoresis-sodium dodecyl sulfate (CE-SDS), and dynamic light scattering (DLS).

[0267] HP-SEC analysis was performed using a Waters Alliance2795 Separation Module (Waters) equipped with a Waters 2487 Dual λ Absorbance Detector (Waters) using a TSK column (G3000SWxl; Tosoh Biosciences, catalog no. 6095006) and a TSK-gel SWxl guard column (Tosoh Biosciences, catalog no. 6095007). Samples (diluted to 5 mg / mL in PBS pH 7.4) were diluted with 0.1 M sodium sulfate (Na 2 SO 4 , Sigma-Aldrich, Cat. No. 31481) / 0.1 M Sodium Phosphate pH 6.8 (NaH 2 PO 4 , Sigma-Aldrich Catalog No. 17844 / Na 2 HPO 4 .2H 2The column was run at 1 mL / min using a 50 mL / min HPLC column (Sigma-Aldrich catalog no. 71633). Results were processed using Empower 3 software and expressed for each peak as a percentage of the total peak height.

[0268] cIEF analysis was performed using an ICE3 Analyzer (ProteinSimple). Each anti-human CD20 IgG1 variant was mixed with an assay mix containing a final concentration of 0.3 mg / mL of antibody, 0.35% methylcellulose (ProteinSimple, Cat. No. 101876); 2% Pharmalytes3-10 (GE Healthcare, Cat. No. 17-0456-01); 6% Pharmalytes8-10.5 (GE Healthcare, Cat. No. 17-0455-01); 0.5% pI marker 7.65 and 0.5% pI marker 10.10 (ProteinSimple, Cat. No. 102407 and Cat. No. 102232). Focusing was performed at 1500 V (prefocusing) for 1 min and 3000 V for 7 min. Anti-human CD3 IgG1 variants were mixed with an assay mix containing finally 0.3 mg / mL antibody; 3.2 M urea (Sigma-Aldrich, Cat. No. 33247-1kg); 0.35% methylcellulose; 2% Pharmalytes 3-10; 6% Pharmalytes 8-10.5; 0.5% pI marker 7.65 and 0.5% pI marker 10.10. Focusing was performed at 1500 V (prefocusing) for 2 min and 3000 V for 9 min. Whole capillary absorption images were taken by a charge-coupled device camera. After peak profile calibration, data were analyzed for pI and area (%) by Empower 3 software (Waters).

[0269] CE-SDS was performed using the LabChip® GXII Touch (Perkin Elmer, Catalog No. CLS138160) on the HT Protein Express LabChip (Perkin Elmer, Catalog No. 760499) using the HT Protein Express Reagent Kit (Perkin Elmer, Catalog No. CLS960008) with little modification. Samples were diluted to 1 mg / mL in PBS pH 7.4 and prepared with 2 μL of diluted sample + 7 μL of denaturing solution + 35 μL of MilliQ water. Samples were prepared in 96-well Bio-Rad HSP9601 plates (Cat. No. 4TI-0960). The analysis was performed under both non-reducing and reducing conditions (addition of DTT). Samples were denatured by incubation at 70° C. for 3 minutes. The chip was prepared according to the manufacturer's instructions and samples were run on the HT Antibody Assay 200 high sensitivity setting. Protein size (kDa) and purity (%) were analyzed using Labchip GXII software V5.3.2115.0.

[0270] Dynamic light scattering (DLS) was performed essentially as described in Example 17, except that DLS was performed at a constant temperature of 25° C. to determine the average particle size.

[0271] All results are summarized in Tables 5-7. As detected by HP-SEC, storage of anti-human CD20 IgG1 and IgG1-huCLB-T3 / 4 antibody variants for 4 months at 2-8°C did not significantly affect the percentage of multimers detected in the samples compared to 1 month exposure. The percentage of multimers present was highest in samples containing anti-human CD20 and CD3 IgG1 variants exposed to 40°C for 4 months and to a lesser extent in those exposed to 40°C for 1 month. Among the anti-human CD20 IgG1 samples at 40°C, samples of variants with L234F-L235E-D265A-K409R mutations contained a higher percentage of multimers than samples of variants with either K409R or L234F-L235E-G236R-K409R mutations. Furthermore, no differences in the rate of degradation were observed between samples containing anti-human CD20 IgG1 variants. In the case of anti-human CD3 IgG1 antibody variants, an enhanced proportion of multimers was detected in samples of the L234F-L235E-G236R-F405L-containing variant after 4 months at 40° C. and, to a lesser extent, 1 month at 40° C., although the proportion of multimers was comparable for samples of variants with the L234F-L235E-D265A-F405L or F405L mutations. Exposure of the anti-human CD3 IgG1-L234F-L235E-G236R-F405L variant to 40° C. for 1 month resulted in a relatively high proportion of protein degradation, which was progressively degraded after 4 months of exposure.

[0272] cIEF analysis was used to study the changes in the percentage of acidic, neutral and basic peaks of the antibody variants in response to different stress conditions. The change in the percentage of acidic protein present in the sample is used as a surrogate measure of deamidation. Of note, the neutral peak was split into two peaks for samples stored for 1 month at either temperature, which were summed as the percentage of neutral protein. An increase in the percentage of acidic protein was observed in all samples between 1 and 4 months of storage at 40 °C, whereas no such increase in the percentage of acidic protein between 1 and 4 months of storage was observed for samples stored at 2-8 °C. The percentage of acidic protein was comparable for all anti-human CD20 IgG1 antibody variants tested. Anti-human CD3 IgG1 variants carrying either the F405L mutation alone, the L234F-L235E-D265A-F405L or the L234F-L235E-G236R-F405L mutations also showed comparable percentages of acidic protein in all tested conditions, except for the F405L-containing variant, which reached a maximum of 79% acidic protein after 4 months of storage at 40°C, whereas the variants carrying the inactivating mutations reached 98% and 96%, respectively, in this condition.

[0273] The percentage of intact protein detected by CE-SDS served as a measure of protein integrity and degradation in response to the stress conditions tested. Anti-human CD20 IgG1 variants largely retained their intact structure after 1 or 4 months of storage of samples at 2-8 °C or after 1 month of storage at 40 °C. However, the percentage of intact IgG detected decreased after 4 months of exposure of samples to 40 °C for all anti-human CD20 IgG1 variants, with the strongest degradation observed for K409R-containing variants. A similar pattern was observed when studying the percentages of HC and LC, but to a lesser extent, except that variants with K409R or L234F-L235E-D265A-K409R mutations showed similar percentages of HC and LC. Anti-human CD3 IgG1 antibody variants showed a similar degradation pattern, with the L234F-L235E-G236R-F405L-containing variant showing the lowest percentage of intact IgG protein after exposure to 40° C. The percentages of HC and LC for the anti-human CD3 IgG1 variants showed a similar pattern overall, although the lowest percentages of HC and LC were detected for the variant with the L234F-L235E-D265A-F405L mutation.

[0274] DLS analysis was performed to determine the mean particle size (radius) after subjecting the antibody variant samples to the indicated stress conditions, which is a surrogate measure of the aggregation level. For all anti-human CD20 IgG1 variants, exposure of the samples to 40°C for 4 months substantially increased the mean particle radius compared to the other conditions tested. In all tested conditions, the highest mean particle radius was detected for the variants with the L234F-L235E-G236R-K409R variant compared to the K409R and L234F-L235E-D265A-K409R variants. Particles were already present in the starting material for samples containing the L234F-L235E-G236R-K409R-containing variants. For the anti-human CD3 IgG1 variants, a relatively large mean particle size was detected for the anti-human CD3 IgG1-F405L variant in all tested conditions. The cause of these larger radii is probably unrelated to the applied stress conditions, since the particles were already present in the starting material. No increase in the average grain size is observed after stress. Substantially lower average grain sizes were detected for variants with the inactivating mutations L234F-L235E-D265A-F405L or L234F-L235E-G236R-F405L.

[0275] In summary, incubation of anti-human CD20 and CD3 IgG1 antibody variants with non-activating mutations in the constant heavy chain region at 40°C for 1 to 4 months resulted in increased multimerization compared to variants stored at 2 to 8°C. More multimerization was observed with the anti-human CD20 IgG1 variant with L234F-L235E-D265A-K409R than with the variant with L234F-L235E-G236R-K409R mutations, whereas the anti-human CD3 IgG1 variant with L234F-L235E-G236R-F405L mutations showed more multimerization than the variant with L234F-L235E-D265A-F405L mutations. In general, no increase in the percentage of acidic protein, used as a surrogate measure of deamidated protein, was observed for any of the antibody variants tested when stored at 2-8 °C for 1 or 4 months. Samples stored at 40 °C for 1 month contained relatively more acidic protein compared to storage at 2-8 °C, but a further increase in the percentage of acidic protein was observed for all variants tested after 4 months of storage at 40 °C. Less intact protein was detected for all antibody variants after storage at 40 °C, which was most pronounced after 4 months of storage. The anti-human CD3 IgG1-L234F-L235E-G236R-F405L variant showed a slightly stronger decrease in the percentage of intact protein detected after 4 months of storage at 40 °C. Finally, more aggregation was observed for all anti-human CD20 IgG1 variants when stored at 40 °C for 4 months. Among the anti-human CD3 IgG1 variants, both variants with non-activating mutations contained substantially less aggregates than the F405L-containing variant.

[0276] Acceptable stability profiles were observed for both anti-human CD20 and CD3 IgG1 variants with either the L234F-L235E-D265A or L234F-L235E-G236R inactivating mutations. Storage at 40°C for 4 months reduced the stability profile of all antibody variants tested.

[0277] Table 5 shows the percentage of multimers, monomers and degraded proteins detected by HP-SEC analysis in samples containing anti-human CD20 and CD3 IgG1 (huCLB-T3 / 4) antibody variants with inactivating or single heterodimerization-promoting mutations in the constant heavy chain region stored for 1 month or 4 months at 2-8° C. or 40° C. HP-SEC: high performance size exclusion chromatography.

[0278] Table 6 shows the percentage of acidic, neutral and basic isoforms present in samples containing anti-human CD20 and CD3 IgG1 variants with inactivating or single heterodimerization-promoting mutations in the constant heavy chain region stored for 1 or 4 months at 2-8°C or 40°C as determined by cIEF analysis. The change in the percentage of acidic isoforms present in a sample is a surrogate for the level of sample deamidation. CIEF: capillary isoelectric focusing.

[0279] Table 7 shows the percentage of intact protein and the sum of HC and LC as determined by non-reduced and reduced CE-SDS analysis, and the mean radius of particles (nm) as determined by DLS analysis, detected in samples containing anti-human CD20 and CD3 IgG1 variants with inactivating or single heterodimerization-promoting mutations in the constant heavy chain region stored for 1 month or 4 months at 2-8 °C or 40 °C. CE-SDS: sodium dodecyl sulfate capillary electrophoresis; DLS: dynamic light scattering. [Table 7] [Table 8] [Table 9] [Example 19] Effect of freeze / thaw on the stability of IgG1 non-activating antibody variants The effect of freeze / thaw cycles on the stability of IgG1 antibody variants carrying deactivating mutations in the constant heavy chain region was assessed using the assay described in Example 18.

[0280] Samples of anti-human CD20 and CD3 IgG1 (huCLB-T3 / 4) antibody variants were formulated in PBS pH 7.4 at a concentration of approximately 20 mg / mL (concentration range 18.7-21.6 mg / mL; filtration applied to anti-human CD3 IgG1 antibody variants). PBS was used as a non-optimal formulation to allow better comparison of protein stability. Anti-human CD20 IgG1 variants had either K409R mutations, L234F-L235E-D265A-K409R or L234F-L235E-G236R-K409R mutations, whereas anti-human CD3 IgG1 antibody variants had either F405L mutations, L234F-L235E-D265A-F405L or L234F-L235E-G236R-F405L mutations. Two independent identical samples were frozen at <-65°C and thawed to room temperature (RT) for three cycles, and then protein stability was tested using HP-SEC, cIEF, CE-SDS and DLS as described in Example 18.

[0281] After three cycles of freezing and thawing, an increase in the percentage of multimers is observed in samples containing anti-human CD20 IgG1 variants as determined by HP-SEC, however the percentage of multimers present in the samples is low (ranging from 0.7% to 1.2% in the reference samples and from 1.8% to 2.8% in the freeze-thawed samples). For the anti-human CD3 IgG1 antibody variants, the percentage of detected multimers, ranging from 0.2% to 0.7%, was considered too low to allow reliable conclusions.

[0282] No substantial differences in the percentage of acidic protein were detected upon freezing and thawing of anti-human CD20 and CD3 IgG1 antibody variants as determined by cIEF analysis. Overall, variants carrying the L234F-L235E-D265A mutations in combination with either K409R or F405L showed the highest percentage of acidic protein even in the reference sample, likely due to sialylation of D265A.

[0283] Freezing and thawing the antibody variant samples also did not affect the percentage of intact protein for any of the IgG1 variants, which ranged from 99% to 100%. Similarly, the percentages of HC and LC were not affected by freezing and thawing the samples.

[0284] The mean particle size (radius), assessed by DLS analysis, was not affected by freezing and thawing of the samples. For the variants with the L234F-L235E-G236R-K409R mutations, relatively more aggregates of the anti-human CD20 IgG1 variants were detected, whereas for the anti-human CD3 IgG1 variants, relatively more aggregates were detected for the F405L-containing variants.

[0285] In summary, freezing and thawing samples containing anti-human CD20 and CD3 IgG1 antibody variants with the inactivating mutations L234F-L235E-D265A or L234F-L235E-G236R in combination with either K409R or F405L did not result in an associated increase in multimerization, deamidation, or proteolysis compared to reference samples. It was therefore concluded that antibody variants with either the L234F-L235E-G236R or L234F-L235E-D265A mutations are similarly capable of retaining protein stability upon repeated freeze / thaw cycles.

[0286] Table 8 shows the percentage of multimers and monomers detected by HP-SEC analysis in samples containing anti-human CD20 IgG1 and anti-human CD3 IgG1 antibody variants with non-activating mutations in the constant heavy chain region that were subjected to three freeze / thaw cycles. Values ​​shown are the average of two individual samples that were subjected to three freeze / thaw cycles. HP-SEC: High performance size exclusion chromatography.

[0287] Table 9 shows the percentage of acidic, neutral and basic isoforms present in samples containing anti-human CD20 and CD3 IgG1 antibody variants with inactivating mutations in the constant heavy chain region subjected to one or two freeze / thaw cycles, as determined by cIEF analysis. The change in the percentage of acidic proteins present in the sample is used as a surrogate measure of deamidation. Values ​​shown are the average of two individual samples subjected to three freeze / thaw cycles. CIEF: capillary isoelectric focusing.

[0288] Table 10 shows the percentage of intact protein and total HC+LC as determined by CE-SDS analysis, and the average radius (nm) as determined by DLS analysis detected in samples containing anti-human CD20 and CD3 IgG1 antibody variants with non-activating mutations in the constant heavy chain region that were subjected to three freeze / thaw cycles. Values ​​shown are the average of two individual samples that were subjected to three freeze / thaw cycles. CE-SDS: sodium dodecyl sulfate capillary electrophoresis; DLS: dynamic light scattering. [Table 10] [Table 11] [Table 12] [Example 20] Effect of low pH-induced stress on the stability of IgG1 non-activating antibody variants In Example 18, the impact of low or high temperature storage of IgG1 antibody variants carrying deactivating mutations in the constant heavy chain region was evaluated using a series of protein stability assays. The same assays were applied in Example 19 to evaluate the impact of freeze / thaw cycles on the stability of such IgG1 antibody variants. Here, since viral inactivation during antibody therapeutic development is often performed under low pH conditions, the impact of low pH induced stress is evaluated using the assays described in Example 18.

[0289] Samples of anti-human CD20 IgG1 and anti-human CD3 (huCLB-T3 / 4) IgG1 antibody variants were formulated in PBS pH 7.4 at a concentration of approximately 20 mg / mL (concentration range 18.7-21.6 mg / mL; filtration applied for anti-human CD3 antibody variants). Anti-human CD20 IgG1 variants had either K409R mutation, L234F-L235E-D265A-K409R or L234F-L235E-G236R-K409R mutations, whereas anti-human IgG1 antibody variants had either F405L mutation, L234F-L235E-D265A-F405L or L234F-L235E-G236R-F405L mutations. The indicated antibody variants were formulated in PBS (reference) and buffer exchanged into 0.02 M sodium citrate buffer (pH 3.0; Sigma-Aldrich, Cat. No. C1909-500G) for 1 h (at room temperature) or 24 h (at 2-8° C.), followed by another buffer exchange back into PBS. Protein stability was then tested using HP-SEC, cIEF, CE-SDS and DLS as described in Example 18.

[0290] As determined by HP-SEC analysis, the percentage of multimers present in anti-human CD20 IgG1 samples increased substantially at pH 3.0 for variants with L234F-L235E-D265A-K409R mutations, while both K409R and L234F-L235E-G236R-K409R-containing variants showed a lower increase in the percentage of multimers. The percentage of multimers observed in samples maintained under pH 3.0 conditions for 1 hour or 24 hours was similar. For anti-human CD3 IgG1 variants, both samples containing variants with L234F-L235E-D265A-F405L or L234F-L235E-G236R-F405L mutations showed an increased presence of multimers at pH 3.0 (1 hour and 24 hours). These percentages were higher than those detected in samples containing F405L-containing variants.

[0291] Data generated by cIEF analysis showed tailing on the basic side of the neutral peak for all anti-human CD20 IgG1 variants incubated for 1 or 24 hours at pH 3.0. Anti-human CD3 IgG1 variants carrying the inactivating mutations L234F-L235E-D265A-F405L or L234F-L235E-G236R-F405L showed an increased percentage of acidic protein at pH 3.0, which was not observed for the F405L-containing variants. Compared to the reference sample (pH 7.4), an increase in acidic protein was observed at pH 3.0 for both the L234F-L235E-D265A-F405L- and L234F-L235E-G236R-F405L-containing variants, although the L234F-L235E-D265A-F405L-containing variant already contained a higher percentage of acidic protein at pH 7.4 than the L234F-L235E-G236R-F405L-containing variant, likely due to sialylation of D265A. Of note, the increase in the percentage of acidic protein in samples containing anti-human CD3 IgG1 variants with non-activating mutations may reflect an increase in spikes observed in these samples at pH 3.0, which may be due to an increase in aggregates formed in the samples.

[0292] CE-SDS analysis did not reveal any differences in the percentage of intact IgG1 or total HC+LC between samples formulated in either PBS or pH 3.0 buffer. Furthermore, an increase in the mean particle size detected using DLS analysis was observed when anti-human CD20 IgG1-L234F-L235E-D265A-K409R was maintained at pH 3.0 for 1 or 24 hours. The larger mean particle size in the PBS reference sample containing the anti-human CD20 IgG1-L234F-L235E-G236R-K409R variant and in the sample containing the anti-human CD3 IgG1-F405L variant may be explained by the removal of aggregates during buffer exchange. Apart from these observations, no substantial differences in particle size were observed between samples formulated in PBS alone or after incubation in pH 3.0.

[0293] In summary, an increased presence of multimers was detected in samples containing anti-human CD20 or CD3 IgG1 antibody variants with either F405L or K409R mutations and / or non-activating mutations in the constant heavy chain region when the samples were incubated at pH 3.0 compared to samples formulated in PBS at physiological pH. Anti-human CD20 IgG1 variants with L234F-L235E-D265A-K409R mutations showed a greater increase in multimerization than variants with L234F-L235E-G236R-K409R mutations. A tailing of the neutral peak on the basic side was observed for all anti-human CD20 IgG1 variants tested at pH 3.0, while the results for anti-human CD3 IgG1 antibody variants with non-activating mutations were inconclusive due to spikes detected in samples kept at pH 3.0. Acidic conditions (pH 3.0) did not affect protein integrity, but a relative increase in mean particle size was detected for anti-human CD20 IgG1-L234F-L235E-D265A-K409R after incubation at pH 3.0.

[0294] Thus, when incubated at low pH, anti-human CD20 IgG1 antibody variants with L234F-L235E-G236R mutations showed less aggregation than variants with L234F-L235E-D265A mutations, indicating that L234F-L235E-G236R-containing antibody variants may be preferred over L234F-L235E-D265A-containing variants during viral inactivation procedures at low pH.

[0295] Table 11 shows the percentage of multimers and monomers detected by HP-SEC analysis for samples containing anti-human CD20 and CD3 IgG1 (huCLB-T3 / 4) antibody variants with non-activating mutations in the constant heavy chain region formulated in PBS or upon incubation in 0.02 M sodium citrate buffer (pH 3.0) for 1 hour or 24 hours.

[0296] Table 12 shows the percentage of acidic, neutral and basic isoforms present in samples containing anti-human CD20 IgG1 and anti-CD3 IgG1 (-huCLB-T3 / 4) antibody variants with non-activating mutations in the constant heavy chain region formulated in PBS or 0.02 M sodium citrate buffer (pH 3.0) for 1 or 24 hours, as determined by cIEF analysis.

[0297] Table 13 shows the percentage of intact IgG and total HC+LC as determined by CE-SDS analysis, and the mean particle radius (nm) as determined by DLS analysis detected in samples containing anti-human CD20 IgG1 and anti-CD3 IgG1-huCLB-T3 / 4 antibody variants with non-activating mutations in the constant heavy chain region formulated in PBS or 0.02 M sodium citrate buffer (pH 3.0) for 1 hour or 24 hours. [Table 13] [Table 14] [Table 15] [Example 21] Assessment of the stability of IgG1 non-activating antibody variants at low pH In Example 20, the effect of low pH on the stability of IgG1 antibody variants with deactivating mutations in the constant heavy chain region was evaluated using a series of protein stability assays, where the effect of low pH induced stress (pH 3.5) is assessed after 0.5, 1 and 4 hours at RT using the assays described in Example 18.

[0298] Samples of anti-human CD20 and human CD3 (huCLB-T3 / 4) antibody variants were formulated in PBS pH 7.4 at a concentration of approximately 5 mg / mL (concentration range 4.98-5.3 mg / mL). IgG1-CD20 variants harbored either the L234F-L235E-D265A-K409R or L234F-L235E-G236R-K409R mutations, whereas IgG1-CD3 antibody variants harbored either the L234F-L235E-D265A-F405L or L234F-L235E-G236R-F405L mutations. To assess the effect of exposure to low pH (3.5) conditions, samples containing the antibody variants were buffer exchanged into 0.02 M sodium citrate buffer (pH 3.5) for 0.5, 1 and 4 hours at room temperature, followed by another buffer exchange back into PBS. Protein stability was tested using HP-SEC, CE-SDS and DLS as described in Example 19.

[0299] As determined by HP-SEC analysis, no increase in the percentage of multimers was observed in samples containing anti-human CD20 IgG1 antibody variants with non-activating mutations at 0.5, 1 and 4 hours when the samples were incubated at pH 3.5. The percentage of multimers of anti-human CD3 IgG1 variants with L234F-L235E-G236R-F405L mutations increased slightly over time. The percentage of detected multimers strongly increased over time in samples containing IgG1-huCLB-T3 / 4-L234F-L235E-D265A-F405L variants at pH 3.5.

[0300] CE-SDS analysis did not reveal any change in the percentage of intact IgG or total HC+LC in any of the samples tested when the samples were incubated at either pH 7.4 (PBS) or pH 3.5. Also, no substantial difference in the mean particle size was detected between any of the samples maintained at either pH 7.4 or pH 3.5 as analyzed by DLS. The enhanced particle size measured for the reference sample of IgG1-CD20-L234F-L235E-G236R-K409R could be explained by the presence of aggregated particles in the applied reference batch, which were removed during the buffer exchange process and therefore not detected in the pH-stressed samples.

[0301] In summary, no increase in multimerization was observed for anti-human CD20 IgG1 antibody variants with non-activating mutations in the constant heavy chain region after exposure to pH 3.5 for 0.5, 1 or 4 hours. Anti-human CD3 IgG1 variants with L234F-L235E-G236R-F405L mutations showed a slight increase in multimerization at pH 3.5, but this increase was more pronounced for variants with L234F-L235E-D265A-F405L mutations. Thus, in agreement with the data shown in Example 20 and despite clone-dependent differences in assay results, it was confirmed that L234F-L235E-G236R-containing antibody variants may have advantages over L234F-L235E-D265A-containing variants for virus inactivation procedures at low pH.

[0302] Table 14 shows the percentage of multimers and monomers detected by HP-SEC analysis (degradation in all samples <0.2%) and the percentage of intact IgG and total HC+LC determined by CE-SDS analysis in samples containing anti-human CD20 and anti-human CD3 antibody variants with non-activating mutations in the constant heavy chain region dissolved in PBS or 0.02 M sodium citrate buffer (pH 3.5) for 0.5 h, 1 h or 4 h.

[0303] Table 14 also shows the mean particle radius (nm) determined by DLS analysis detected in samples containing IgG1-CD20 and IgG1-CD3 antibody variants with non-activating mutations in the constant heavy chain region dissolved in PBS or 0.02 M sodium citrate buffer (pH 3.5) for 0.5 h, 1 h or 4 h. [Table 16] [Example 22] Assessment of the stability at low pH of bispecific IgG1 antibody variants with deactivating mutations In Example 21, the effect of low pH induced stress (pH 3.5) on the stability of an IgG1 antibody variant with a non-activating mutation in the constant heavy chain region is evaluated after 0.5, 1 hour and 4 hours. Here, the effect of low pH induced stress (pH 3.5) on the stability of an anti-CD3 / CD20 bispecific antibody with a non-activating mutation is evaluated after 0.5, 1 hour and 4 hours using the assay described in Example 18.

[0304] Bispecific antibodies (bsAbs) were generated from anti-human CD3 and anti-human CD20 IgG1 antibodies with the non-activating mutations L234F-L235E-D265A or L234F-L235E-G236R in addition to either the K409R or F405L mutations that promote half-molecule heterodimerization with complementary half-molecules only under controlled reducing conditions, using the controlled Fab arm exchange procedure described in Example 1. This resulted in bsIgG1-CD3xCD20 antibodies with either L234F-L235E-D265A in both arms, or L234F-L235E-G236R in both arms (hereafter referred to as symmetric backbones), or bispecific antibodies with a combination of L234F-L235E-D265A mutations in one arm and L234F-L235E-G236R mutations in the other arm (hereafter referred to as asymmetric backbones). Samples of bsIgG1-CD3xCD20 antibody variants were formulated in PBS (pH 7.4) at a concentration of approximately 5 mg / mL (concentration range 3.209-5.304 mg / mL). To assess the effect of exposure to low pH (3.5) conditions, samples containing bsIgG1-CD3xCD20 antibodies were buffer exchanged into 0.02 M sodium citrate buffer (pH 3.5) for 0.5, 1 and 4 hours at room temperature, followed by another buffer exchange back into PBS. Protein stability was then tested using HP-SEC, CE-SDS and DLS as described in Example 18.

[0305] HP-SEC analysis showed that exposure of any of the bsIgG1-CD3xCD20 antibody variants with non-activating mutations to pH 3.5 at any of the time points tested had no significant effect on multimerization compared to the respective reference samples. Using CE-SDS, no effect of pH 3.5 exposure was observed for any of the bsAb variants tested, indicating that all bsAbs remained intact upon exposure to pH 3.5. Analysis of the average size by DLS showed that exposure to pH 3.5 of the symmetric bsAb with the L234F-L235E-D265A mutations as well as the asymmetric bsAb with the L234F-L235E-G236R mutations in the F405L-containing arm did not induce altered levels of aggregation. The small decrease in average size observed at pH 3.5 at either time point analyzed for the symmetric bsAb with the L234F-L235E-G236R mutation in the K409R-containing arm and the asymmetric bsAb with the L234F-L235E-G236R mutation may be due to removal of aggregates during the buffer exchange process.

[0306] In summary, IgG1bsAb variants with non-activating mutations in the constant heavy chain region retain their intact structure and are not increasingly susceptible to multimerization after exposure to low pH conditions for up to 4 hours at a concentration of about 5 mg / mL. These results indicate that bispecific antibody variants with either the L234F-L235E-G236R or L234F-L235E-D265A non-activating mutations are equally capable of retaining monomericity upon exposure to low pH conditions, a favorable property during therapeutic development in virus inactivation procedures performed at low pH.

[0307] Table 15 shows the percentage of multimers and monomers detected by HP-SEC analysis, as well as the percentage of intact IgG and total HC+LC determined by CE-SDS analysis detected in samples containing bispecific antibodies (bsAbs) generated from anti-human CD20 IgG1 and anti-human CD3 antibodies with inactivating mutations in the constant heavy chain region. BsAbs were generated with either L234F-L235E-D265A inactivating mutations in both arms, or L234F-L235E-G236R inactivating mutations in both arms, or a combination of L234F-L235E-D265A mutations in one arm and L234F-L235E-G236R mutations in the other arm. Antibody variants were dissolved in PBS or 0.02M sodium citrate buffer (pH 3.5) for 0.5 h, 1 h or 4 h before analysis.

[0308] Table 16 shows the mean radius (nm) and percentage of monomer mass determined by DLS analysis detected in samples containing bsIgG1-CD3xCD20 antibody variants with inactivating mutations in the constant heavy chain region. BsAbs were generated with either L234F-L235E-D265A inactivating mutations in both arms, or L234F-L235E-G236R inactivating mutations in both arms, or a combination of L234F-L235E-D265A mutations in one arm and L234F-L235E-G236R mutations in the other arm. Antibody variants were dissolved in PBS or 0.02 M sodium citrate buffer (pH 3.5) for 0.5 h, 1 h or 4 h before analysis. [Table 17] [Table 18] [Example 23] Assessment of the stability at low pH of anti-CD20 and anti-gp120 IgG1 antibody variants with deactivating mutations In Example 21, the effect of low pH induced stress (pH 3.5) on the stability of IgG1 antibody variants with non-activating mutations in the constant heavy chain region is evaluated after 0.5, 1 and 4 hours. Here, the effect of low pH induced stress (pH 3.5) on the stability of anti-human CD20 IgG1 and anti-gp120 (HIV1) IgG1-b12 antibodies with non-activating mutations in the constant heavy chain region in combination with the K409R mutation is evaluated after 0.5, 1, 2, 4 and 24 hours using the HP-SEC assay described in Example 18. The extent of multimerization under low pH conditions was analyzed for these antibody variants as a measure of protein instability, which is important in the context of viral inactivation procedures during antibody therapeutic development.

[0309] Samples of anti-human CD20 IgG1 and IgG1-b12 antibody variants were formulated in PBS at a concentration of approximately 0.5 mg / mL (concentration range 0.435-0.5 mg / mL). Anti-human CD20 IgG1 and IgG1-b12 variants carried either the L234F-L235E-D265A-K409R or L234F-L235E-G236R-K409R mutations. To assess the effect of exposure to low pH (3.5) conditions, samples containing antibody variants in PBS were acidified by dropwise addition of 2 M acetic acid (Fluka, Cat. No. 33209) to pH 3.5, followed by incubation for 0.5, 1, 2, 4 or 24 hours at room temperature using a plate shaker. After incubation, a sample from each sample tube was transferred to a tube containing 2 M Tris-HCl (pH 9.0; Sigma-Aldrich, Cat. No. T6066) to restore the pH to 7.4. Protein stability was then tested using HP-SEC as described in Example 18.

[0310] HP-SEC analysis revealed a rapid increase in multimerization in samples containing anti-human CD20 IgG1 variants with L234F-L235E-D265A-K409R mutations in response to incubation at pH 3.5, which topped out after 2 hours of incubation with 35.2% of multimers detected in the samples. In contrast, exposure of anti-human CD20 IgG1 variants with L234F-L235E-G236R-K409R mutations to pH 3.5 resulted in a stable and relatively slow increase in multimerization over time, with a maximum of 10.8% of multimers detected in samples analyzed after 24 hours of incubation.

[0311] A very similar pattern was observed for IgG1-b12 variants with non-activating mutations: variants with L234F-L235E-D265A-K409R mutations showed a rapid increase in multimerization, with up to 27.6% multimers detected after 2 h of incubation at pH 3.5, whereas a slower and lower increase in multimerization was observed for variants with L234F-L235E-G236R-K409R mutations at pH 3.5, with up to 18% multimers detected after 24 h of incubation at pH 3.5.

[0312] In summary, HP-SEC analysis revealed that the process of multimerization occurs more rapidly and extensively in antibody variants with L234F-L235E-D265A-K409R mutations than in variants with L234F-L235E-G236R-K409R mutations. Therefore, it was concluded that other antibody clone variants with L234F-L235E-G236R inactivating mutations also have a higher ability to retain monomericity at low pH conditions than L234F-L235E-D265A-containing variants. Retention of monomericity at low pH is a favorable feature during virus inactivation procedures performed at low pH.

[0313] Table 17 shows the percentage of multimers, monomers and degradation detected by HP-SEC analysis in samples containing anti-human CD20 IgG1 and anti-gp120 (HIV1) IgG1-b12 antibody variants with non-activating mutations in the constant heavy chain region dissolved in PBS or sodium acetate buffer (pH 3.5) for 0.5, 1, 2, 4 or 24 hours. [Table 19] Aspects of antibody therapeutic development in FER In Examples 15-23, various aspects related to antibody therapeutic developability were evaluated. First, no obvious differences were observed in the ability to form bispecific antibodies using antibody variants with either the L234F-L235E-D265A (FEA) or L234F-L235E-G236R (FER) mutations. Also, the production levels of antibody variants with either the FEA or FER mutations were similar. Highly enriched samples of antibody variants with FER mutations showed improved protein stability and reduced tendency to aggregate compared to antibody variants with FEA mutations. Exposure of antibody variants to 40°C for 4 months increased multimerization of all variants tested, the extent of which was antibody clone dependent. Furthermore, freeze-thaw cycles did not affect the protein stability of antibody variants with either the FEA or FER mutations. Importantly, antibody variants with FER mutations showed higher resistance to low pH-induced stress conditions compared to variants with FEA mutations.

[0314] The introduction of FEA mutations in the Fc region of IgG1 antibodies resulted in antibody variants that could be easily developed and are now widely accepted for clinical product development and clinical use. In view of the above, the protein stability profile and developability of antibody variants when an inactive format is desired could be further improved by introducing FER mutations instead. Thus, FER-containing variants may be preferred for the development of monospecific or multispecific antibodies for clinical use. In particular, antibody variants with FER mutations were shown to be less sensitive to exposure to low pH conditions, which is a beneficial property in procedures for viral inactivation that are often applied and required during the development of therapeutic antibodies for human use.

[0315] [Example 24] C1q binding and complement-dependent cytotoxicity by anti-human CD20 antibodies and their variants containing different sets of inactivating mutations in each heavy chain In Examples 3 and 5, it was shown that the introduction of L234F-L235E-G236R (FER) mutations into the heavy chain (HC) constant region of wild-type-like IgG1 anti-human CD20 and HLA-DR antibodies almost completely abolished the ability to induce CDC. This was consistent with the reduced C1q binding by antibody variants with FER mutations, as shown in Example 4. Here, the ability to bind complement factor C1q was evaluated for anti-human CD20 antibodies containing two HCs, either L234F-L235E-D265A (FEA) or FER mutations in both HCs, and for variants containing a FEA mutation in one HC and a FER mutation in the other HC. The ability of the antibody variants to induce CDC was also evaluated.

[0316] Wild-type anti-human CD20 antibody IgG1-CD20 and its variants with K409R or F405L mutations plus FEA or FER inactivating mutations in both HCs, or variants containing FEA mutations in one HC and FER mutations in the other HC (hereafter referred to as "asymmetric variants"), were tested in a C1q binding assay on Raji cells using 20% ​​normal human serum (NHS, M0008, Sanquin) as a source of C1q, according to the procedure described in Example 4. Asymmetric variants of anti-human CD20 antibodies were generated by controlled Fab arm exchange essentially as described in Example 1. C1q binding to the antibody variants was detected by measuring median fluorescence intensity-FITC by flow cytometry on an Intellicyt iQue screener (Sartorius). Data were analyzed using a non-linear agonist dose-response model and the area under the dose-response curve (AUC) per experimental replicate was calculated using log-transformed concentrations in GraphPad PRISM (version 8.4.1, GraphPad Software) without antibody control as baseline, followed by normalization per experimental replicate to the AUC value measured for the non-binding control antibody IgG1-b12 (0%) and the AUC value measured for the wild-type IgG1 antibody variant (IgG1-CD20, 100%). Data are means ± SEM from three independent experiments.

[0317] The same antibody variants tested in the C1q binding assay were tested in an in vitro CDC assay, with 5×10 antibodies per well essentially as described further in Example 3. 4Antibody variants were tested at a range of concentrations (0.014-10 μg / mL final concentration; 3-fold dilutions) using Raji cells. The number of PI-positive cells was determined by flow cytometry on an Intellicyt iQue screener (Sartorius). The percentage of PI-positive cells corresponding to the percentage of cell lysis was calculated as (number of PI-positive cells / total number of cells) × 100%. Data were analyzed using a nonlinear agonist dose-response model and the area under the dose-response curve (AUC) per experimental replicate was calculated using log-transformed concentrations in GraphPad PRISM (version 8.4.1, GraphPad Software) without antibody control as baseline, followed by normalization per experimental replicate to the AUC values ​​measured for the non-binding control antibody IgG1-b12 (0%) and the AUC values ​​measured for the wild-type IgG1 antibody variant (IgG1-CD20, 100%). Data are means ± SEM from three independent experiments.

[0318] Assessment of C1q binding to IgG1-CD20 variants revealed that wild-type IgG1-CD20 antibodies efficiently engaged the complement protein C1q upon binding to CD20 on target Raji cells. Introduction of either FEA or FER inactivating mutations, in combination with either F405L or K409R mutations, dramatically reduced C1q binding (Figure 16). C1q binding was more strongly reduced for IgG1-CD20 variants containing FER mutations than for IgG1-CD20 variants containing FEA mutations, regardless of whether both HCs contained either K409R or F405L mutations, or whether one HC contained a K409R mutation and the other contained a F405L mutation. A complete loss of C1q binding was observed for the asymmetric variant BisG1-CD20FEA-F405LxFER-K409R, whereas a strong reduction in C1q binding was observed for the other asymmetric variant BisG1-CD20FER-F405LxFEA-K409R (Figure 16).

[0319] Evaluation of the CDC-inducing ability of the same IgG1-CD20 variants revealed that the ability to eliminate CDC was most strongly suppressed by introducing FER mutations, regardless of whether both HCs contained either the K409R o...

Claims

1. 1. A protein comprising a first polypeptide and a second polypeptide, wherein said first and second polypeptides each comprise at least a hinge region, a CH2 region and a CH3 region, respectively, of a human IgG1 immunoglobulin heavy chain, and wherein at least one of said first and second polypeptides is modified to comprise substitutions of amino acids corresponding to amino acids at positions L234, L235 and G236, wherein the amino acid positions are as defined by Eu numbering.

2. 2. The protein of claim 1, wherein the amino acids at positions L234, L235 and G236 of at least one of the first and second polypeptides are substituted with F, E and R, respectively.

3. 3. A protein according to claim 1 or claim 2, wherein one of the first and second polypeptides comprises the substitutions of amino acids corresponding to the amino acids at positions L234, L235 and G236, and the other is modified and comprises substitutions of amino acids corresponding to the amino acids at positions L234, L235 and D265, preferably the substitutions being F, E and A, respectively.

4. 3. The protein of claim 1 or claim 2, wherein both the first and second polypeptides comprise the substitutions of amino acids corresponding to amino acids L234, L235 and G236.

5. The protein of any one of claims 1 to 4, wherein each of the first and second polypeptides comprises an immunoglobulin CH1 region.

6. The protein according to any one of claims 1 to 5, wherein the protein comprises a first and a second binding region.

7. 7. The protein of claim 6, wherein the first and second binding regions comprise a first immunoglobulin heavy chain variable region and a first immunoglobulin light chain variable region, respectively, and the second binding region comprises a second immunoglobulin heavy chain variable region and a second immunoglobulin light chain variable region.

8. The protein of claim 7 , wherein the immunoglobulin heavy and light chain variable regions are human or humanized immunoglobulin heavy and light chain variable regions.

9. 9. The protein of claim 7 or 8, wherein the first and second polypeptides are immunoglobulin heavy chains, and the first and second polypeptides comprise the first and second immunoglobulin heavy chain variable regions, respectively.

10. The protein of any one of claims 6 to 9, wherein the protein comprises a first immunoglobulin light chain constant region and a second immunoglobulin light chain constant region.

11. 11. The protein of claim 10, wherein the protein comprises a first and a second immunoglobulin light chain, the immunoglobulin light chain comprising the respective first and second immunoglobulin light chain variable regions and the respective first and second immunoglobulin constant light chain regions.

12. 12. The protein of claim 11 , wherein the first immunoglobulin light chain is connected to the first immunoglobulin heavy chain via a disulfide bridge and the second immunoglobulin light chain is connected to the second immunoglobulin heavy chain via a disulfide bridge, thereby forming the first binding region and the second binding region, respectively, and the first and second immunoglobulin heavy chains are connected via disulfide bridges.

13. The protein according to any one of claims 1 to 12, which is an antibody.

14. The protein according to any one of claims 4 to 12, which is a monospecific antibody.

15. The protein according to claims 4 to 14, wherein the first and second polypeptide chains have identical amino acid sequences.

16. 16. The protein of claim 15, wherein the first and second polypeptides comprise further amino acid substitutions, preferably at amino acids selected from the group consisting of T366, L368, K370, D399, F405, Y407 and K409, such as F405L or K409R.

17. The protein of any one of claims 14 to 16, wherein the monospecific antibody binds to an antigen selected from the group consisting of a cellular target, a cytokine, a toxin, a pathogen, a cancer antigen, and a plasma protein.

18. The protein according to any one of claims 1 to 13, which is a bispecific antibody.

19. 19. The bispecific antibody of claim 18, wherein the first and second polypeptides comprise further substitutions in their respective CH2 and CH3 regions such that the sequences of the respective CH2 and CH3 regions from the first and second polypeptides differ, said substitutions making it possible to obtain the polypeptide comprising the first and second polypeptides.

20. 20. The bispecific antibody of claim 19, wherein in the first polypeptide, at least one amino acid at a position corresponding to a position selected from the group consisting of T366, L368, K370, D399, F405, Y407 and K409 in a human IgG1 heavy chain is substituted, and in the second polypeptide, at least one amino acid at a position corresponding to a position selected from the group consisting of T366, L368, K370, D399, F405, Y407 and K409 in a human IgG1 heavy chain is substituted, and wherein the substitutions in the first and second polypeptides are not at the same position.

21. 20. The bispecific antibody of claim 19, wherein the amino acid at the position corresponding to F405 in said first polypeptide is L and the amino acid at the position corresponding to K409 in said second polypeptide is R, or vice versa.

22. 22. The bispecific antibody according to any one of claims 18 to 21, wherein said bispecific antibody has a modification consisting of substitution of amino acids at positions L234, L235 and G236 by F, E and R in both said first and second polypeptide, and substitution of the amino acid at position F405 by L in said first polypeptide and substitution of the amino acid at K409 by R in said second polypeptide, or vice versa.

23. 23. The bispecific antibody according to any one of claims 18 to 22, wherein the first and second polypeptides have a substitution consisting of a substitution according to any one of claims 1 to 4 and 19 to 22.

24. 24. The protein of any one of claims 1 to 17 or the bispecific antibody of any one of claims 18 to 23, wherein said first and second polypeptides comprise an amino acid sequence according to SEQ ID NO: 1, and said amino acid sequences comprised in said first and second polypeptides have amino acid substitutions as defined in claims 1 to 4 and 18 to 22.

25. 25. A protein or bispecific antibody as described in claim 24, wherein the amino acid sequence according to SEQ ID NO:1 does not contain a terminal lysine.

26. The bispecific antibody of any one of claims 18 to 25, wherein one of the binding regions binds to a cancer antigen.

27. 26. The bispecific antibody of any one of claims 18 to 25, wherein one of the binding regions binds to an effector cell, such as a T cell, a NK cell, a dendritic cell, a monocyte, a macrophage or a neutrophil.

28. 26. A bispecific antibody according to any one of claims 18 to 25, wherein one of the binding regions binds to an effector cell, such as a T cell, a NK cell, a dendritic cell, a monocyte, a macrophage or a neutrophil, and the other binding region binds to a cancer antigen.

29. 26. A nucleic acid encoding the first or second polypeptide according to any one of claims 4 to 17 and 24 to 25, wherein the first or second polypeptide comprises said substitutions of amino acids corresponding to amino acids L234, L235 and G236, preferably wherein said substitutions at positions L234, L235 and G236 are F, E and R, respectively.

30. 30. The nucleic acid of claim 29, wherein the first or second polypeptide is an immunoglobulin heavy chain.

31. A host cell comprising a nucleic acid according to claim 29 or 30.

32. A pharmaceutical composition comprising the protein according to any one of claims 1 to 17 or the bispecific antibody according to any one of claims 18 to 28, and a pharma- ceutically acceptable carrier.

33. A protein according to any one of claims 1 to 17, or a bispecific antibody according to any one of claims 18 to 28, or a pharmaceutical composition according to claim 32, for use in the treatment of a disease.

34. 34. The protein, bispecific antibody or pharmaceutical composition of claim 33 for uses comprising the treatment of cancer, an infectious disease or an autoimmune disease.

35. Use of a protein according to any one of claims 1 to 17, or a bispecific antibody according to any one of claims 18 to 28, or a pharmaceutical composition according to claim 32 for the manufacture of a medicament.

36. 36. The use according to claim 35 for the manufacture of a medicament for the treatment of cancer, an infectious disease, or an autoimmune disease.

37. 1. A method for preparing a bispecific antibody, comprising the steps of: a) a. an immunoglobulin heavy chain comprising at least the hinge, CH2 and CH3 regions, respectively, of a human IgG1 immunoglobulin heavy chain, comprising substitutions of amino acids at positions L234, L235 and G236 with F, E and R, respectively; b. an immunoglobulin light chain; providing a first antibody comprising: b) a. an immunoglobulin heavy chain comprising at least the hinge region, CH2 region, and CH3 region, respectively, of a human IgG1 immunoglobulin heavy chain, - substitutions at positions L234, L235 and D265, preferably F, E and A, respectively; Or, an immunoglobulin heavy chain comprising a substitution of the amino acids at positions L234, L235 and G236 by F, E and R, respectively; b. an immunoglobulin light chain; providing a second antibody comprising: c) wherein the sequences of the first and second CH3 regions of each of the first and second antibodies are different such that a heterodimeric interaction between the first and second CH3 regions is stronger than a homodimeric interaction of the first and second CH3 regions, respectively; d) incubating the first antibody with the second antibody under reducing conditions sufficient to allow cysteines in the hinge region to undergo disulfide bond isomerization; e) obtaining the bispecific antibody comprising the first immunoglobulin heavy chain and the first immunoglobulin light chain of the first antibody and the second immunoglobulin heavy chain and the second immunoglobulin light chain of the second antibody; A method for preparing a bispecific antibody comprising:

38. The method according to claim 37, wherein in step c) the sequence difference is according to any one of claims 19 to 25.

39. A method according to claim 37 or claim 38 for preparing a bispecific antibody as defined in any of claims 18 to 28.

Images