Related Applications

[0001 j This application claims the benefit of U.S. Provisional Application No.

62/535,354, filed July 21 , 2017, the contents of which are incorporated herein by reference in its entirety.

Field of the Invention

[0002| The invention relates to the generation of multispecific antibody mixtures include a subset of antibodies isolated from a mixture of two or more monospecific antibodies and one or more bispecific antibodies, wherein all antibodies in the subset have the same common heavy chain. The invention also relates to methods of isolating, purifying, or otherwise producing such a subset of antibodies by using at least one affinity chromatography step. The invention also relates to methods of using such a subset o antibodies in a variety of therapeutic indications.

Background of the Invention

[0003] During the last two decades, monoclonal antibodies (mAbs) have become an important therapeutic modality, bringing significant benefits for patients in various indications. The success of mAbs is in part due to their high specificity for their target antigen and low intrinsic toxicity. These properties greatly limit off-target side effects when compared to other classes of drugs. The vast majority of approved therapeutic mAbs are unmodified antibodies of the IgG isotype.

[0004J However, the targeting of a single protein that is enabled by standard mAbs might not always be sufficient to achieve significant therapeutic effect (Fischer Expert Opin. Drug Discov. 2008 3(8):833-839).

[0005| An obvious option to increase efficacy, is to use two mAbs in combination. This strategy is clinically pursued, for instance, for antibodies targeting immune checkpoint molecules such as anti-CTLA4 and anti-PD- 1 antibodies (Larkin et al., N Engl J Med 2015; 373:23-34; Harris et al., Cancer Biol Med. 2016 13(2): 171 -93). However, significant cost and development hurdles are associated with the development of mAbs combinations. In particular, two separate manufacturing processes have to be put in place, leading to significant increase in costs (Rasmussen et al., Archives of Biochemistry and Biophysics 2012 526:39-145). These issues become even more important if targeting three or more proteins or antigens is considered.

[0006] Several approaches for targeting multiple proteins or antigens have been used to date. The use of two mAbs in combination has been pursued, but is often hindered by significant cost and development hurdles. Bispecific antibodies (BiAbs) represent a rapidly developing alternative to achieve muitispeeifie targeting and over 60 formats have been described to date (Spiess et al., Mol Immunol. 2015 67:95-106; Brinkmann and Kontermann mAbs 2017 9: 182- 212), and another approach to achieve targeting of two or even more proteins is the generation of antibody mixtures or recombinant polyclonal mixtures. However, only a few of these muitispeeifie formats have been approved for therapeutic use.

[0007] Accordingly, there exists a need for generating muitispeeifie antibody mixtures that are able to target multiple proteins, epitopes, and/or antigens.

Summary of the Invention

[0008] The disclosure provides antibody mixtures that include a subset of antibodies isolated from a mixture of two or more monospecific antibodies and one or more bispecific antibodies, wherein all antibodies in the subset have the same common heavy chain. These mixtures are produced when multiple light chains are co-expressed with a common heavy chain in a single cell. In some embodiments, the purified subsets include only antibodies that contain at least a kappa light chain. In some embodiments, the purified subsets include only antibodies that contain at least a lambda light chain. In some embodiments, the purified subsets include only antibodies that contain a kappa light chain and a lambda light chain.

[0009] The disclosure also provides methods of isolating, purifying, or otherwise producing a subset of antibodies isolated from a mixture of two or more monospecific antibodies and one or more bispecific antibodies, wherein all antibodies in the subset have the same common heavy chain, by using at least one affinity chromatography step. In some embodiments, the purification step is performed using kappa constant or variable domain specific affinity chromatography media. In some embodiments, the purification step is performed using lambda constant or variable domain specific affinity chromatography media. In some embodiments, the purification step is performed using a two-step affinity chromatography process. In some embodiments, the first purification step is performed using kappa constant or variable domain specific affinity chromatography media and the second purification step using lambda constant specific affinity chromatography media. In some embodiments, the first purification step is performed using Lambda constant specific affinity chromatography media and the second purification step using Kappa constant or variable domain specific affinity chromatography media.

[0010] The multispecific antibody mixtures and methods provided herein are useful in any of a variety of therapeutic, diagnostic, and/or prophylactic indications. For example, the multispecific antibody mixtures are useful in treating, preventing and/or delaying the progression of, or alleviating a symptom of cancer or other neoplastic condition by administering an antibody mixture to a subject in which such treatment or prevention is desired. In some embodiments, the multispecific antibody mixtures described herein are useful in treating hematological malignancies and/or solid tumors. For example, the multispecific antibody mixtures described herein are useful in treating CD47 ' tumors, mesothelin ⁺ tumors, and combinations thereof. By way of non-limiting example, the multispecific antibody mixtures described herein are useful in treating non-Hodgkin's lymphoma (NHL), acute lymphocytic leukemia (ALL), acute myeloid leukemia (AML), chronic lymphocytic leukemia (CLL), chronic myelogenous leukemia (CML), multiple myeloma (MM), breast cancer, ovarian cancer, head and neck cancer, bladder cancer, melanoma, mesothelioma, colorectal cancer, cholangiocarcinoma, pancreatic cancer, including pancreatic adenocarcinoma, lung cancer, including lung adenocarcinoma, leiomyoma, leiomyosarcoma, kidney cancer, glioma, glioblastoma, endometrial cancer, esophageal cancer, biliary gastric cancer, and prostate cancer. Solid tumors include, e.g., breast tumors, ovarian tumors, lung tumors, pancreatic tumors, prostate tumors, melanoma tumors, colorectal tumors, lung tumors, head and neck tumors, bladder tumors, esophageal tumors, liver tumors, and kidney tumors.

[0011 j In some embodiments, the multispecific antibody mixtures are useful in treating, preventing and/or delaying the progression of, or alleviating a symptom of an autoimmune disease and/or inflammatory disorder by administering an antibody mixture to a subject in which such treatment or prevention is desired. Autoimmune diseases include, for example, Acquired Immunodeficiency Syndrome (AIDS, which is a viral disease with an autoimmune component), alopecia areata, ankylosing spondylitis, antiphospholipid syndrome, autoimmune Addison's disease, autoimmune hemolytic anemia, autoimmune hepatitis, autoimmune inner ear disease (AIED), autoimmune lymphoprol i ferative syndrome (ALPS), autoimmune thrombocytopenic purpura (ATP), Behcet's disease, cardiomyopathy, celiac sprue-dermatitis hepetiformis; chronic fatigue immune dysfunction syndrome (CFIDS), chronic inflammatory demyelinating polyneuropathy (CIPD), cicatricial pemphigoid, cold agglutinin disease, crest syndrome, Crohn's disease, Degos' disease, dermatomyositis-juveniie, discoid lupus, essential mixed cryoglobulinemia, fibromyalgia-fibromyositis, Graves' disease, Guillain-Barre syndrome, Hashimoto's thyroiditis, idiopathic pulmonary fibrosis, idiopathic thrombocytopenia purpura (ITP), IgA nephropathy, insulin-dependent diabetes mellitus, juvenile chronic arthritis (Still's disease), juvenile rheumatoid arthritis, Meniere's disease, mixed connective tissue disease, multiple sclerosis, myasthenia gravis, pernacious anemia, polyarteritis nodosa, polychondritis, polyglandular syndromes, polymyalgia rheumaiica, polymyositis and dermatomyositis, primary agammaglobulinemia, primary biliary cirrhosis, psoriasis, psoriatic arthritis, Raynaud's phenomena, Reiter's syndrome, rheumatic fever, rheumatoid arthritis, sarcoidosis, scleroderma (progressive systemic sclerosis (PSS), also known as systemic sclerosis (SS)), Sjogren's syndrome, stiff-man syndrome, systemic lupus erythematosus, Takayasu arteritis, temporal arteritis/giant cell arteritis, ulcerative colitis, uveitis, vitiligo and Wegener's granulomatosis. 10012] Inflammatory disorders include, for example, chronic and acute inflammatory disorders. Examples of inflammatory disorders include Alzheimer's disease, asthma, chronic obstructive pulmonary disease, atopic allergy, allergy, atherosclerosis, bronchial asthma, eczema, glomerulonephritis, graft vs. host disease, hemolytic anemias, osteoarthritis, sepsis, stroke, transplantation of tissue and organs, vasculitis, diabetic retinopathy and ventilator induced lung injury.

[0013] In some embodiments, the multispecific antibody mixtures are useful in retargeting

T cells.

]0014] Pharmaceutical compositions according to the invention can include a multispecific antibody mixture of the invention and a carrier. These pharmaceutical compositions can be included in kits, such as, for example, therapeutic kits and/or diagnostic kits.

Brief Description of the Drawings

[0015] Figures l A, I B, and I C are a series of schematic representations of various approaches to generate recombinant antibody mixtures. Figure 1 A depicts methods of mixing monoclonal antibodies (mAbs) that were expressed and purified independently to produce a mixture of mAbs. Figure I B depicts methods of mixing cell lines, also referred to as a polyclonal cell bank, to produce a mixture of only mAbs in a single fermentation reactor. Figure I C depicts methods of co-expressing multiple antibody chains in a single cell line to produce a mixture of inAbs and bispecific antibodies (BiAbs).

[0016] Figures 2A, 2B, and 2C are a series of schematic representations of methods of co- expressing multiple light chains with a single, common heavy chain and the resultant antibody mixtures produced. Figure 2A depicts how the co-expression of two antibody light chains with a single heavy chain results in secretion of three different antibodies: two mAbs each incorporating the same light chain in each Fab and a single BiAb incorporating a different light chain in each Fab. Figure 2B depicts how the co-expression of three antibody light chains with a single heavy chain results in secretion of six different antibodies: three mAbs each incorporating the same light chain in each Fab and three BiAb incorporating a different light chain in each Fab. Figure 2C depicts how the co-expression of four antibody light chains with a single heavy chain results in secretion of ten different antibodies: four mAbs each incorporating the same light chain in each Fab and six BiAb incorporating a different light chain in each Fab.

[0017] Figure 3 is a schematic representation of methods of co-expressing two lambda light chains, one kappa light chain, and a single common heavy chain in a single cell, which results in the secretion of a mixture of two IgGXX mAbs, one IgGKK mAb, two ¾Ο λ BiAbs, and one IgGXX BiAb (Mixture 1 ). From Mixture 1 , different submixtures can be isolated using different affinity chromatography steps as indicated.

[0018] Figure 4 is a schematic representation of methods of co-expressing two lambda light chains, two kappa light chains, and a single common heavy chain in a single cell, which results in the secretion of two IgGXX mAbs, two IgGKK mAbs, four IgGicX BiAbs, one IgGXX BiAb and one IgGKK BiAb (Mixture 1). From Mixture 1 , different submixtures can be isolated using different affinity chromatography steps as indicated.

[0019] Figure 5 is a schematic representation of methods of co-expressing three lambda light chains, one kappa light chain, and a single common heavy chain in a single cell, which results in the secretion of a mixture of three IgGXX mAbs, one IgGKK mAb, three IgGKX BiAbs, three IgGXX BiAb and one IgGKK BiAb (Mixture 1). From Mixture 1 , different submixtures can be isolated using different affinity chromatography steps as indicated.

[0020] Figure 6 is a general schematic representation of methods of the disclosure.

[0021] Figure 7 A is a schematic representation of two bispecific antibodies targeting CD47 and two epitopes of a tumor associated antigen (TAA). [00221 Figure 7B is a schematic representation of a method of purifying a subset containing two BiAbs (CD47 x TAA epitope 1 and CD47 x TAA epitope 2) by two step affinity chromatography using kappa- specific affinity media followed by lambda-specific affinity media or vice versa.

[00231 Figure 8A is a schematic representation of a first bispecific antibody targeting

CD47 and a first tumor associated antigen (TAA 1 ) and a second bispecific antibody targeting CD47 and a second tumor associated antigen (TAA 2).

[0024] Figure 8B is a schematic representation of a method of purifying a subset containing two BiAbs (CD47 x TAA 1 and CD47 x TAA 2) by two step affinity chromatography using kappa- specific affinity media followed by lambda-specific affinity media or vice versa.

[0025] Figure 9A is a schematic representation of bispecific antibodies targeting multiple tumor associated antigens (TAA 1, TAA 2, et seq.) and/or multiple epitopes on a TAA.

[0026] Figure 9B is a schematic representation of a method of purifying a subset containing

4 different BiAbs (CD47 x TAA 1; CD47 x TAA 2; PD-L1 x TAA 1; PD-L1 x TAA 2) by two step affinity chromatography using kappa-specific affinity media followed by lambda-specific affinity media or vice versa.

[0027J Figure 1 OA is a schematic representation of bispecific antibodies for retargeting T cells to multiple tumor associated antigens (TAA 1 , TAA 2, et seq.) and/or multiple epitopes on a TAA.

[0028] Figure 10B is a schematic representation of a method of purifying a subset containing 2 different BiAbs (CD3 x TAA 1 ; CD3 x TAA 2) by two step affinity chromatography using kappa-specific affinity media followed by lambda-specific affinity media or vice versa.

[0029] Figure 1 1 A is a schematic representation of a mixture of a mAb targeting a first

TAA, combined with a BiAb targeting the first and a second TAA for which bivalent targeting should be avoided.

[0030] Figure 1 I B is a schematic representation of a method of purifying a subset containing a BiAbs (TAA 1 x TAA 2) and an anti-T AA 1 mAb by two step affinity chromatography using kappa-specific affinity media followed by lambda-specific affinity media or vice versa.

[0031 ] Figure 12A, 12B and 12C depict a series of graphs depicting the level of Antibody

Dependent Cellular Phagocytosis (ADCP), shown as phagocytosis index, observed in a pancreas adenocarcinoma HPAC cell line in the presence of increasing concentrations of various bispecific antibodies and/or monoclonal antibodies as well as combinations of bispecific antibodies.

[0032J Figure 13 A, 13B and 13C depict a series of graphs depicting the level of ADCP, shown as phagocytosis index, observed in the ovarian adenocarcinoma CaOV cell line in the presence of increasing concentrations of various bispecific antibodies and/or monoclonal antibodies as well as combinations of bispecific antibodies.

[0033] Figures 14A, 14B and 14C depict a series of graphs depicting the level of ADCP, shown as phagocytosis index, observed in the gastric carcinoma NCI-N87 cell line in the presence of increasing concentrations of various bispecific antibodies and/or monoclonal antibodies as well as combinations of bispecific antibodies.

[0034] Figures 15 is a graph depicting the level of ADCP, shown as phagocytosis index, observed in the gastric carcinoma NCI-N87 cell line in the presence of increasing concentrations o various bispecific antibodies as well as combinations using different ratios of bispecific antibodies.

[0035] Figure 16A is a picture of an isoelectric focusing gel on which a sample of the IgG mixture obtained via the co-expression of one heavy chain and three di ferent light chains and purified from the supernatant by Protein A chromatography (lane 1 ). The K2O25O30 submixture composed of two bispecific antibodies was loaded in lane 2.

[0036J Figure 16B is a chromatogram obtained after hydrophobic interaction chromatography of the K2O25O30 submixture. Elution time, and peak characteristics are indicated in the inserted table.

[0037] Figures 17 is a graph depicting the level of ADCP, shown as phagocytosis index, observed in the gastric carcinoma NCI-N87 cell line in the presence of increasing concentrations of the K2Q25O30 submixture and an equimolar combination of 2025 and K2O30.

[00381 Figure 18 A depicts the anti-tumour activity of 2 CD47xMSLN κλ bodies (K2038 and K2041 ) and their combination (K2038+K2041) in a MS LN -transacted HepG2 mouse model of liver cancer. Figure 18B, shows the areas under the curves and statistical analysis. Figure 18C is a graph showing the tumor growth inhibition observed in the different treatment groups.

Detailed Description

[0039] The multispecific antibody mixtures and methods provided herein overcome limitations seen with other antibody formats and/Or antibody mixtures. Currently, most of the approved monoclonal antibody formats are unmodified antibodies of the IgG l isotype. However, targeting a single antigen is not always sufficient to achieve the desired therapeutic effects (Fischer Expert Opin. Drug Discov. 2008 3(8):833-839). Two mAbs in combination have been used to increase efficiency. This strategy is clinically pursued, for instance, for antibodies targeting immune checkpoint molecules such as anti-CTLA4 and anti-PD- 1 antibodies (Larkin et al., N Engl J Med 2015; 373:23-34; Harris et al., Cancer Biol Med. 2016 13(2): 171 -93). However, significant cost and development hurdles are associated with the development of mAbs combinations. In particular, two separate manufacturing processes have to be put in place, leading to significant increase in costs (Rasmussen et al., Archives of Biochemistry and Biophysics 2012 526:39 145). These issues become even more important if targeting three or more proteins or antigens is considered.

[0040] Bispecific antibodies (BiAbs) have been developed as an alternative means to achieve multispecific targeting, and over 60 formats have been described to date (Spiess et al., Mol Immunol. 2015 67:95- 106; Brinkmann and Kontermann mAbs 2017 9:182-212). As two binding sites are incorporated into the same molecule, unique modes of actions that are not supported by mAbs or mAb combinations, are enabled by BiAbs (Fischer and Leger Pathobiology. 2007 74( l ):3- 14). Examples of such unique modes of action are retargeting of T-cell or NK cells to tumors cells, BiAb delivery to the central nervous system via increased transportation across the blood brain barrier, coagulation Factor VIII mimetic activity, selective targeting of receptors expressed on multiple cell types.

[0041 ] Yet another approach to achieve targeting of two or even more proteins is the generation of antibody mixtures or recombinant polyclonal mixtures. In contrast to a combination of two mAbs described above, in which each mAb is produced separately, the antibodies in the mixture are produced together as a mixture. Various approaches to generate mixtures have been described (see e.g., Raju and Strohl Expert Opin. Biol. Ther. 2013 1 ( 10): 1347- 1352; Wang et al., 2013 Current Opinion in Chemical Engineering 2013 2: 1-1 1). A general challenge for the production of a recombinant polyclonal mixture, is to achieve consistency between batches so that each component of the mixtures remains constant and thus the overall composition and biological activity of the mixture is consistent.

[0042] In one approach, different stable cell lines, each expressing a single monoclonal antibody, are mixed top produce a polyclonal cell bank that is used in a single bioreactor for production. In this case, the different antibodies are secreted by different cell lines into the medium and all antibodies are purified together to obtain the final recombinant antibody mixture (Rasmussen et al., Archives of Biochemistry and Biophysics 2012 526:39-145). In this case, the reproducible growth and productivity characteristics of the individual cell lines during fermentation must be extremely well controlled to ensure consistency between batches of the recombinant polyclonal mixture. Faster growth or increased productivity of one cell line has a direct impact on the mixture composition. Achieving such a level of control is not straightforward and represents significant hurdle of this approach. Nevertheless, very complex mixtures of up to 25 independent antibodies have been developed using polyclonal cells lines and evaluated in clinical trials (Hjelmstrom et al., Blood 2008 1 12: 1987).

[0043] Another strategy is to co-express multiple antibody heavy and/or light chains within a single cell. In this case, different heavy and light chains can pair and, thus, a mixture can be generated. The complexity of the resulting mixture depends on the number of different chain that are co-expressed. In this approach, ensuring that the different possible pairings reconstitute a functional antigen binding site is critical to avoid production of non-functional molecules. This pairing issue can be solved by the use of a common heavy or common light chain or by engineering of protein interfaces to preferentially form the desired pairing (See e.g., Fischer et al., Nat. Comms 201 56:61 13 doi: 10.1038/ncomms71 13). In contrast to the previous approach relying on mixing of independent cell lines, this approach leads to the generation of both mAbs and BiAbs. A significant advantage is that, once a cell line expressing the different antibody chains in a stable manner is identified, the fermentation and production is simplified as it is similar to a standard niAb process. However, i all pairing are allowed, the complexity of the mixture components (mAbs and BiAbs) can be quite significant and thus antibodies relying on a common heavy or light chain are pref erred.

[0044] Different approaches to generate recombinant antibody mixtures, their benefits and limitations are listed in Table 1 and illustrated in Figures 1 A- 1C.

Table 1. Recombinant Antibody Mixtures

Mixing of purified mAbs Composition can be High cost of goods monoclonal precisely controlled by (COGs) due to

antibodies mixing purified mAbs independent

manufacturing of each component before mixing

This issue increases with each additional antibody composing in the mixture

Mix of cells lines mAbs Complex mixtures can be Control of fermentation (polyclonal cell bank) achieved reaction and polyclonal

Reduced COGs cell line stability are

critical to ensure batch consistency

Co-expression of mAbs and Single cell line is used as Use of common chain to multiple antibody BiAbs in standard mAb process ensure productive chains in a single cell Presence of BiAb that pairings

line enable novel modes of

action

Reduced COGs

[00451 Several mixtures have been developed using some of the approaches described above and have reached clinical stage indicating that this therapeutic modality is viable f interest. Mixtures enable unique modes of actions that cannot be achieved with a single antibody.

[0046] For instance, targeting the receptors of the HER ErbB family has been actively explored with antibody mixtures. This family of receptor tyrosine kinases includes growth factor receptor EGFR/ErbB l , HER-2/ErbB2, HER-3/ErbB3 and HER-4/ErbB4. Several monoclonal antibodies targeting EGFR and HER2 are approved for clinical use. Studies have reported that mixtures of antibodies targeting multiple epitopes and thus able to simultaneously engage these targets lead to increased inhibition of cancer cell growth in in vitro and in vivo experiments. This has led to the development of several antibody mixtures directed against members of the HER/ErbB family. One example is Sym004, a mixture of two anti EGFR mAbs that showed significantly superior activity in preclinical models compared to the approved anti-EGFR mAb Cetuximab (Koefoed et al., mAbs 201 1 3:6, 584-595). MM- 151 is another mixture that is composed of three human mAbs targeting distinct, non-overlapping epitopes on EGFR (Arena et al., Science Translational Medicine 2016 8(324), 324ral 4). This mixture provides more effective blockade of the receptor and limits the emergence of resistance when compared to single mAb therapy. A more complex mixture, Sym013, composed of 6 mAbs targeting EGFR, HER2 and HER3 (two mAbs directed against non-overlapping epitope on each target), has also reached clinical development stage. The overall superior activity observed with mixtures against this receptor family can be explained by several factors such as a more complete shutdown of this partially redundant signaling pathways, increased internalization and degradation via antibody mediated receptor cross-linking, increase Fc-mediated cell killing.

[0047] Another therapeutic area in which antibody mixtures have demonstrated striking superiority is the treatment of infectious diseases (Oleksiewicz et at., Archives o Biochemistry and Biophysics 2012 526: 124 13 1 ; Pohl et al.. Infection and Immunity 2013 81 (6): 1880- 1888). For instance, multi-epitope targeting of soluble toxins such as Botulinum toxin A increases not only neutralization of the toxin but also its clearance from circulation. Overall, rnAb approaches have not proven very effective for the treatment of infections in contrast to other therapeutic areas. This might ref lect that a polyclonal antibody response similar to that of the natural immune system is required for effective protection against complex organisms such as pathogens.

[0048] Rozrolimupab is a striking example of how complex antibody mixtures can. This mixture of 25 mAbs against the Rhesus D antigen is produced using the polyclonal cell bank approach, which is exemplified in Figure I B, and has been developed for the treatment of Immune thrombocytopenic purpura (ITP). A benefit o this complex mixture is that it provides effective coverage of the numerous Rhesus D variants present in the population (Robak et al., Blood. 2012 120( 18): 3670-76).

[0049] These examples highlight the mechanistic advantages that mixtures can provide in comparison to single mAb therapy. These mechanisms include but are not limited to: increased internalization and degradation of receptors, fast and superior clearance of soluble targets, increased Fc-dependent effector functions, synergistic effect of targeting multiple epitopes on a single target or multiple targets and pathways, better coverage of target variants and prevention of escape mutations.

[0050] The present invention provides the mean of generating defined antibody mixture of either i) bispecific antibodies (BiAbs) only or ii) monoclonal antibodies (mAbs) and BiAbs. The method relies on co-expression of a single antibody heavy chain that is common to all Fv regions of the antibodies in the mixture and several light chains of either kappa or lambda families. This co-expression and random incorporation into IgG molecules leads to the secretion from a single cell of a mixture of monospecific mAbs or BiAb that all contain the same heavy chain. The complete mixture can then be purified using for instance an affinity reagent binding the Fc region of an IgG such Protein A, as previously described in US20140179547. The invention improves over previous methods as it allows to selectively purify different subsets of the secreted mixture. In particular, it allows for the simple and cost effective generation of mixtures o BiAbs. This opens the possibility of using mixtures of BiAbs for modes of action that are not enabled by a single mAb or mAb mixtures. The present invention combines the benefit o using a single production cell line and the possibility to control the composition of the final purified mixtures to maximize the desired biological activity.

[00511 Depending on the number of light chains that are co-expressed with a common heavy chain, different types of antibody mixtures can be generated. This process can be generalized as follows:

For

n = the number of different light chains expressed with a common heavy chain

Then

The number of different mAbs= n

The number of different BiAbs = (n ² -n)/2

The total number of different antibodies = n+(n ² -n)/2

[0052] Table 2 lists these numbers for the co-expression of 2 to 10 different light chains and 3 examples are further detailed below:

1 - If two antibody light chains are co-expressed with a single heavy chain three different antibodies are secreted: two mAbs each incorporating the same light chain in each Fab and a single BiAb incorporating a different light chain in each Fab (see Figure 2 A). However, the abundance of the three forms depends on the relative expression and assembly of the two light chains. Provided expression and assembly are equivalent the theoretical distribution is 25% of each mAb and 50% of the BiAb.

2- If three antibody light chains are co-expressed with a single heavy chain, 6 different antibodies are secreted: three mAbs each incorporating the same light chain in each Fab and three BiAb incorporating a different light chain in each Fab (see Figure 2B). Provided expression and assembly are equivalent the theoretical distribution is 1 1.1% for each mAb and 22.2% for each of the BiAbs. 3- If four antibody light chains are co-expressed with a single heavy chain, 10 different antibodies are secreted: four mAbs each incorporating the same light chain in each Fab and six BiAb incorporating a different light chain in each Fab (see Figure 2C). Provided expression and assembly are equivalent the theoretical distribution is 6.25% for each mAb and 12.5% for each of the BiAbs.

Table 2. Summary of Antibody Mixtures

[0053 J Depending on whether the different light chains that are co-expressed are of the kappa type or of the lambda type, the distribution of mAbs and BiAbs containing kappa or lambda light chains will vary, thus further diversifying the types of mixtures that can be generated by this approach. For instance, if two lambda light chains and one kappa light chain are co-expressed with a common heavy chain, two IgGXl mAbs, one IgGra mAb, two IgGic BiAbs, and one IgG X BiAb will be produced as shown in Figure 3. If four light chain are co-expressed, the molecule distribution will vary if two kappa and two lambda light chains are used (Figure 4) or if three lambdas and one kappa light chains are used (Figure 5).

[0054] The resulting number of antibody forms generated in a given multispecific mixture can be generalized as follows:

For

n = number of kappa light chains

m = number of lambda light chains

Then The number of IgGKK niAbs = n

The number of IgGH mAbs = m

The number of IgGiA BiAbs = n x m

The number of IgGKK BiAbs = (n ² -n)/2

The number of IgG BiAbs = (m ² -m)/2

[0055] Provided expression and assembly of the different light chains is identical, the resulting theoretical distribution of each form can be generalized as follows:

For

n = number of kappa light chains

m = number of lambda light chains

Then

The proportion of IgGKK mAbs = n/(n x m)

The proportion of IgG . mAbs = m/(n x m)

The proportion of IgGx BiAbs = (n x m)x2/(n x m)

The proportion of IgGKK BiAbs = (n ² -n)/(n x m)

The proportion of IgG BiAbs = (m ² -m)/(n x m)

10056] The present invention provides the mean of purifying a well-defined subset of either i) BiAbs only or ii) mAbs and BiAb from a wide variety of different mixtures that can be generated by the co-expression of varying numbers of kappa and or lambda light chains, as described above. The invention relies on multistep affinity chromatography using resins that bind specifically to the Fc portion and resins that bind specifically to the kappa constant region or lambda constant region. In this way, subsets of antibody forms composing these complex mixtures can be readily isolated and their composition tailored depending on which type of resins are used and the desired mode of action.

[0057] For instance, the co-expression of two lambda light chains and one kappa light chain with a common heavy chain generates a mixture that is composed of two IgGAA, mAbs, one IgGxx mAb, two ¾ϋ λ BiAbs, and one IgGAl BiAb (Figure 3). All these antibody forms contain an Fc portion and can be easily and effectively purified using, for instance, Protein A chromatography. From this mixture (Mixture 1, indicated in Figure 3), different defined subsets of antibodies can be readily isolated using affinity chromatography media binding specifically either to the kappa or lambda chain, thereby allowing for capture of all antibody forms containing at least one kappa or lambda chain, respectively (see Figure 3, submixtures 2 and 3, respectively). Submixture 2 will contain one IgGrac mAb and two IgGi . BiAbs. Submixture 3 will contain two IgGXk mAbs, two IgGi l BiAbs, and one IgGXk BiAb. A submixture containing two IgGi BiAbs (without mAbs) can be isolated via two consecutive affinity chromatography steps using media binding specifically to the kappa and lambda chain (Figure 3, submixture 4).

[0058J If four different light chains are co-expressed, the resulting mixtures are more complex and depend on the proportion of kappa and lambda light chains that are co-expressed. Two examples are described in Figures 4 and 5 in which two lambda light chains and two kappa light chains or three lambda light chains and one kappa light chain are co-expressed, respectively.

[0059] The co-expression of two lambda light chains and two kappa light chain with a common heavy chain generates a mixture (Mixture 1, Figure 4) that is composed of two IgGlX mAbs, two IgGi K mAbs, four IgGid. BiAbs, one IgGXl BiAb and one ¾GKK BiAb. All these antibody forms contain and Fc portion and can effectively be purified using for instance Protein A chromatography. As described above, different defined subsets of antibodies can be readily isolated from this initial Mixture 1 using affinity chromatography media binding specifically either to the kappa or lambda chain and thus allowing to capture all antibody forms containing at least one kappa or lambda chain, respectively (Figure 4, submixture 2 and 3, respectively). Submixture 2 will contain two IgGicK mAbs, four IgGid, BiAbs and one IgGio BiAb. Submixture 3 will contain two IgGAl mAbs, four IgGx BiAbs and one IgGXX BiAb. Finally, a mixture containing four IgGi BiAbs (without mAbs or BiAbs containing only lambda or only kappa chains) can be isolated via two consecutive affinity chromatography steps using media binding specifically to the kappa and lambda chain(Figure 4, submixture 4).

[0060] In the situation in which 3 lambda light chains and one kappa light chain are co- expressed with a common heavy chain, a mixture (Mixture 1 , Figure 5) that is composed of three IgGAl mAbs, one IgGioc mAbs, three ¾ΰκλ BiAbs, three IgGXX BiAb and one IgGra BiAb is generated. All these antibody forms contain and Fc portion and can effectively be purified using for instance Protein A chromatography. As described above, different defined subsets of antibodies can be readily isolated from this initial Mixture 1 (Figure 5) using affinity chromatography media binding specifically either to the kappa or lambda light chain constant domains such as the CaptureSelect Fab Kappa and CaptureSelect Fab Lambda affinity matrices (GE Healthcare) and thus allowing to capture all antibody forms containing at least one kappa or lambda chain, respectively (Figure 5, submixtures 2 and 3, respectively). Submixture 2 will contain one IgGi i mAb and three IgGi . BiAbs. Submixture 3 will contain three IgGXl mAbs, three IgGicX BiAbs and three IgGXk BiAbs. Finally, a submixture containing three IgGi BiAbs (without mAbs of BiAbs containing only lambda or only kappa chains) can be isolated via two consecutive affinity chromatography steps using media binding specifically to the kappa and lambda chain (Figure 5, submixture 4).

[0061] The two situations described above highlight that although in each case four light chains are co-expressed, the nature of the chain expressed (i.e., two kappa light chains and two lambda light chains or 3 lambda light chains and one kappa light chain) lead to the generation and purification of very different mixture subsets that can readily and effectively be isolated by applying the method of the invention. These situations represent only selected examples and do not limit the application of the invention to other situations.

[0062] It is obvious that the invention can generally be used to isolate subsets of mAbs and/or BiAbs from complex mixtures, based on their respective content of kappa and lambda light chains. In any situation, it is possible, using affinity chromatography media binding specifically to the constant or variable domain of either kappa or lambda antibody light chains, to purify three types of mixture subsets:

1. All antibody molecules containing at least one kappa light chain, including both mAbs and BiAbs

2. All antibody molecules containing at least one lambda light chain, including both mAbs and BiAbs

3. All BiAbs molecules containing one kappa light chain and one lambda light chain

[0063] The invention can also be further applied to purify antibody subsets of antibody mixtures containing hybrid light chains as described in patent US20140179547. These hybrid chains are composed of either:

1. A variable kappa domain fused to a constant lambda domain or

2. A variable lambda domain fused to a constant kappa domain [0064] Similarly to full length lambda or kappa light chains, these hybrid molecules can be selectively separated depending of their variable and constant domain composition by using chromatography media binding either to:

• kappa constant domains (such as CaptureSelect™ LC-kappa (Hu) affinity matrix (Life Technologies, Zug, Switzerland)

• some kappa variable domains (such as Protein L)

• lambda constant domains (such as CaptureSelect™ LC- lambda (Hu) affinity matrix (Life Technologies, Zug, Switzerland)

[0065! The general application of the invention is described in Figure 6.

[0066] The mu!tispecific antibody mixtures of the disclosure are useful in a variety of indications. Subsets of antibody mixtures can have multiple applications for the development of therapeutic modalities relying on modes of actions that are not enabled by mAbs, mAb mixtures or BiAbs.

10067] In particular the invention allows for the straightforward generation of fully human

BiAb mixtures using a single cell line, which represents a significant advantage. Indeed, as described above the combination of two individual mAbs leads to significant increases in costs. This limitation is obviously even more significant if two individual BiAb have to be combined, as generation BiAb is more complex and costs higher than for standard mAbs. Examples of potentially interesting mixtures of BiAb as well as how these mixtures can be generated using the invention, are described in more detail below.

(0068] Multi-epitope targeting of a tumor associated antigen (TAA) combined with

CD47 blockade: CD47 is a ubiquitously expressed receptor that acts as a checkpoint of the innate immune system, repressing phagocytosis via interaction with SIRPa (See e.g., Oldenborg, P. A., CD47: A Cell Surface Glycoprotein Which Regulates Multiple Functions of Hematopoietic Cells in Health and Disease, ISRN Hematol. 2013; 2013:614619; Soto-Pantoja DR, et a!., Therapeutic opportunities for targeting the ubiquitous cell surface receptor CD47 (2012), Expert Opin Ther Targets. 2013 Jan; 17( l):89- 103; Sick E, et al., CD47 Update: a multi faced actor in the tumor microenvironment of potential therapeutic interest, Br J Pharmacol. 2012 Dec; 167(7): 1415-30). Blockade of CD47 with a BiAb approach avoids toxicities observed with mAbs directed against CD47. A CD47xTAA BiAb allows for restricted inhibition of CD47 only on TAA expressing cells thus avoiding toxicities and poor pharmacokinetic properties. A mixture of two BiAb targeting two epitopes on a single receptor each combining an anti-TAA arm and an anti-CD47 arms would lead to superior anti-CD47 blockade, increased Fc coverage and ultimately better tumor cell killing (Figure 7A). Indeed, in such a situation, for each TAA molecule, two CD47 receptors can be blocked and two Fc are present at the target cell surf ace enhancing recruitment of effector cells via Fc gamma receptor interaction.

[006 j Co-expression of a common heavy chain along with:

» A first lambda light chain driving specificity against a first epitope on a TAA

• A second lambda light chain driving specificity against a second epitope on a TAA

• A kappa light chain driving specificity against CD47

[0070] From the resulting mixture a subset containing two BiAbs (CD47xTAA epitope 1 and CD47xTAA epitope 2) can be purified by two step affinity chromatography using kappa specific affinity media followed by lambda specific affinity media or vice versa (Figure 7B).

[0071] Multi-TAA targeting combined with CD47 blockade: The example above can be applied to two TAAs expressed by a target cancer cell. A mixture of two BiAb targeting two TAAs, each combining an anti-TAA arm and an anti-CD47 arm would also lead to superior anti- CD47 blockade, increased Fc coverage, further combined with effects linked to the blockade of the TAA and ultimately better tumor cell killing (Figure 8A).

[0072] Co-expression of a common heavy chain along with:

• A first lambda light chain driving specificity against a first TAA

• A second lambda light chain driving specificity against a second TAA

• A kappa light chain driving specificity against CD47

[0073] From the resulting mixture a subset containing two BiAbs (CD47xTAA 1 and

CD47xTAA 2) can be purified by two step affinity chromatography using kappa specific affinity media followed by lambda specific affinity media or vice versa (Figure 8B).

[0074] Multi-TAA or epitope targeting combined with blockade of several checkpoint molecules: The approach used in the examples above can be extended to two TAAs (or two epitopes on the same TAA) expressed by a target cancer cell combined with arms blocking two checkpoint receptors, for instance CD47 and PD-L 1. Such a mixture would lead to CD47 blockade, PD-Ll blockade, increased Fc coverage, and ultimately better tumor cell killing and potentially durable response of the immune system while avoiding toxicities linked to general monospecific blockade of immune checkpoint (Figure 9A).

[0075] Co-expression of a common heavy chain along with:

• A first lambda light chain driving specificity against a first TAA

• A second lambda light chain driving specificity against a second TAA

• A kappa light chain driving specificity against CD47

• A kappa light chain driving specificity against PD-L 1

[0076] From the resulting mixture a subset containing 4 different BiAbs (CD47xTAA 1 ;

CD47xTAA 2; PD-L l xTAA 1 ; PD-Ll xTAA 2) can be purified by two step affinity chromatography using kappa specific affinity media followed by lambda specific affinity media or vice versa (Figure 9B).

10077 J Miilti TAA retargeting of T cells: T cell retargeting is a clinically validated and widely pursued approach in oncology (Chames and Baty MAbs. 2009 l(6):539-47). A mixture of BiAb would allow to retarget T cells to two different TAAs (or epitopes on the same TAA) potentially improving treatment efficacy (Figure 1 OA).

[00781 Co-expression of a common heavy chain along with:

• A first lambda light chain driving specificity against a first TAA

• A second lambda light chain driving specificity against a second TAA

• A kappa light chain driving specificity against CD3

[0079] From the resulting mixture a subset containing 2 different BiAbs (CD3 TAA 1 ;

CD3xTAA 2) can be purified by two step affinity chromatography using kappa specific affinity media followed by lambda specific affinity media or vice versa (Figure 10B).

[0080] The examples above are illustrative and do not limit the possible application that can be pursued by applying the invention. Furthermore is obvious that any kappa chain provided in any example can be replaced by a lambda chain and vice versa.

[0081 ] Another important type application of the invention is the generation of m Ab and

BiAb mixtures that exclude one or several forms of mAbs. This feature is important as for some targets monospecific, bivalent engagement by a mAb is detrimental and can lead to toxicities and other unwanted effects. For instance it is know that anti-cMet antibodies lead unwanted agonistic activity that have led to the development of monovalent antibodies. Generating a mixture containing anti-cMet antibodies but avoiding the anti c-Met mAb is straightforward using the present invention. As described above targeting of CD47 with a mAb leads to significant toxicities in human. Anti-CD47 mAbs can be removed from mixtures of mAbs targeting other receptors and BiAb targeting CD47 in conjunction with another receptor. Similarly, monoclonal anti-CD3 would also need to be eliminated from anti mixture containing a CD3 binding component.

[0082] Combined monovalent and bivalent targeting of two tumor associated antigens: For instance, a mixture of a mAb targeting a first TAA, combined with a BiAb targeting the first and a second TAA for which bivalent targeting should be avoided. This strategy could be applied for TAAs such as cMet and EGFR (Figure 1 1 A).

[0083] Co-expression of a common heavy chain along with:

• A lambda light chain driving specificity against a first TAA 1

• A kappa light chain driving specificity against a second TAA 2

[0084] From the resulting mixture a subset containing a BiAbs (TAA 1 xTAA 2) and an anti-TAA l mAb can be purified by affinity chromatography using lambda specific affinity media (Figure 1 1 B).

[0085] Again other applications of subset of antibody mixtures can be rationalized and the examples above do not limit the scope of the invention.

[0086] The same principle applies for any antibody isotype as well as antibodies form non- human species provided affinity or other chromatography reagents are available to separate the different submixtures of antibodies.

[0087] Furthermore, the same principle of the invention can be applied to F(ab')2 formats in which a single VHCH1 is co-expressed with two or more VKCK or VXCX thus leading to the secretion of a mixture of monospecific and bispecific F(ab')2 molecules. These can then be separated into well-defined submixtures according to method of the invention using affinity chromatography media binding to portion of the kappa or lambda light chains. Similarly, hybrid V CX and VXCi can also be used when performing the method of the invention.

EXAMPLES

EXAMPLE 1 : Selection of antibody candidates for bispecific antibody generation [00881 Four antibodies targeting different epitopes on hMSLN and an anti-hCD 19 antibody, all containing a lambda light chain, as well as an anti-hCD47 antibody containing a kappa light chain were selected for the generation of antibody mixtures. These antibodies all contain the same heavy chain, which are described in PCT Publication No. WO 2014/087248 and in co-pending patent application US Patent Application No. 62/51 1 ,669, filed May 26, 2017 and entitled "Anti-CD47 x Anti-Mesothelin Antibodies and Methods of Use Thereof, are suitable for the generation of bispecifie antibodies based on the κλ body format as described in the Patent Application Publication No. US20140179547. The selected antibodies are listed in Table 3 and the sequences are shown below.

Table 3.

Antibody Specificity Light chain

025 Anti-hMSLN Lambda

O30 Anti-hMSLN Lambda

035 Anti-hMSLN Lambda

038 Anti-hMSLN Lambda

041 Anti-hMSLN Lambda

L7-2 Anti-hCD19 Lambda

K2 Anti-hCD47 Kappa

[0089] Each of the anti-hMSLN, anti-hCD 19, and anti-hCD47 antibodies in Table 3 includes a common heavy chain (SEQ ID NO: 2) encoded by the nucleic acid sequence shown in SEQ ID NO: 1 :

>COMMON-HC-NT ( SEQ I D NO : 1 )

GAGGTGCAGCTGTTGGAGTCTGGGGGAGGCTTGGTACAGCCTGGGGGGTCCCTGAGACTC TCCTGTGCAG CCTCTGGATTCACCTTTAGCAGCTATGCCATGAGCTGGGTCCGCCAGGCTCCAGGGAAGG GGCTGGAGTG GGTCTCAGCTATTAGTGGTAGTGGTGGTAGCACATACTACGCAGACTCCGTGAAGGGCCG GTTCACCATC TCCAGAGACAATTCCAAGAACACGCTGTATCTGCAAATGAACAGCCTGAGAGCCGAGGAC ACGGCCGTAT ATTACTGTGCGAAAAGTTATGGTGCTTTTGACTACTGGGGCCAGGGAACCCTGGTCACAG TCTCGAGCGC CTCCACCAAGGGCCCATCGGTCTTCCCCCTGGCACCCTCCTCCAAGAGCACCTCTGGGGG CACAGCGGCC CTGGGCTGCCTGGTCAAGGACTACTTCCCCGAACCGGTGACAGTCTCGTGGAACTCAGGA GCCCTGACCA GCGGCGTGCACACCTTCCCGGCTGTCCTACAGTCCTCAGGACTCTACTCCCTCAGCAGCG TGGTGACTGT GCCCTCCAGCAGCTTGGGCACCCAGACCTACATCTGCAACGTGAATCACAAGCCCAGCAA CACCAAGGTG GACAAGAGAGTTGAGCCCAAATCTTGTGACAAAACTCACACATGCCCACCGTGCCCAGCA CCTGAACTCC TGGGGGGACCGTCAGTCTTCCTCTTCCCCCCAAAACCCAAGGACACCCTCATGATCTCCC GGACCCCTGA GGTCACATGCGTGGTGGTGGACGTGAGCCACGAAGACCCTGAGGTCAAGTTCAACTGGTA CGTGGACGGC GTGGAGGTGCATAATGCCAAGACAAAGCCGCGGGAGGAGCAGTACAACAGCACGTACCGT GTGGTCAGCG TCCTCACCGTCCTGCACCAGGACTGGCTGAATGGCAAGGAGTACAAGTGCAAGGTCTCCA ACAAAGCCCT CCCAGCCCCCATCGAGAAAACCATCTCCAAAGCCAAAGGGCAGCCCCGAGAACCACAGGT GTATACCCTG CCCCCATCTCGGGAGGAGATGACCAAGAACCAGGTCAGCCTGACTTGCCTGGTCAAAGGC TTCTATCCCA GCGACATCGCCGTGGAGTGGGAGAGCAACGGGCAGCCGGAGAACAACTACAAGACCACGC CTCCCGTGCT GGACTCCGACGGCTCCTTCTTCCTCTATAGCAAGCTCACCGTGGACAAGTCCAGGTGGCA GCAGGGGAAC GTCTTCTCATGCTCCGTGATGCATGAGGCTCTGCACAACCACTACACGCAGAAGAGCCTC TCCCTGTCTC CGGGTTAA

>COMMON-HC-AA (SEQ ID NO: 2)

EVQLLESGGGLVQPGGSLRLSCAASGFTFSSYMSWVRQAPGKGLEWVSAISGSGGSTYYA DSVKGRFTI SRDNSKNTLYLQMNSLRAEDTAVYYCAKSYGAEDYWGQGTLVT¥SSASTKGPSVFPLAP SSKSTSGGTAA LGCLVKDYFPEPVTVSWNSGALTSGVH FPAVLQSSGLYSLSSVV VPSSSLGTQTYICNVNHKPSNTKV DKRVEPKSCDKTHTCPPCPAPELLGGPSVFLFPPKPKDTL ISRTPEVTCVVVDVSHEDPEVKFN YVDG VEVHNAKTKPREEQYNSTYRVVSVLTVLHQDWLNGKEYKCKVSNKALPAPIEKTISKAKG QPREPQVYTL PPSREEMTKNQVSLTCLVKGFYPSDIAVEWESNGQPENNYKTTPPVLDSDGSFFLYSKLT VDKSRWQQGN VFSCSVMHEALHNHYTQKSLSLSPG

[0090] Each of the anti-hMSLN, anti-hC 1 , and anti-hCD47 antibodies in Table 3 includes a common variable heavy domain (SEQ ID NO: 4) encoded by the nucleic acid sequence shown in SEQ ID NO: 3:

>COMMON-VH-NT (SEQ ID NO: 3)

GAGGTGCAGCTGTTGGAGTCTGGGGGAGGCTTGGTACAGCCTGGGGGGTCCCTGAGACTC TCCTGTGCAG CCTCTGGATTCACCTTTAGCAGCTATGCCATGAGCTGGGTCCGCCAGGCTCCAGGGAAGG GGCTGGAGTG GGTCTCAGCTATTAGTGGTAGTGGTGGTAGCACATACTACGCAGACTCCGTGAAGGGCCG GTTCACCATC TCCAGAGACAATTCCAAGAACACGCTGTATCTGCAAATGAACAGCCTGAGAGCCGAGGAC ACGGCCGTAT ATTACTGTGCGAAAAGTTATGGTGCTTTTGACTACTGGGGCCAGGGAACCCTGGTCACAG TCTCGAGC

>COM ON-VH~AA { SEQ ID NO: 4) EVQLLESGGGLVQPGGSLRLSCAASGFTFSSYAMSWVRQAPGKGLEWVSAISGSGGSTYY ADSVKGRFTI SRDNSKNTLYLQMNSLRAEDTAVYYCAKSYGAFDYWGQGTLVTVSS

[0091] The 025 antibody includes a common heavy chain (SEQ ID NO: 2) encoded by the nucleic acid sequence shown in SEQ ID NO: 1 and includes a lambda light chain (SEQ ID NO: 6) encoded by the nucleic acid sequence shown in SEQ ID NO: 5. The variable region of the lambda light chain is bolded in the amino acid sequence below.

>025-LC-NT (SEQ ID NO: 5)

CAGCCTGTGCTGACTCAGCCGGCTTCCCTCTCTGCATCTCCTGGAGCATCAGCCAGTCTC ACCTGCACCT TGCACAGTGGCATCTCTGTTAAGGATTACAGGATATACTGGTACCAGCAGAAGCCAGGGC GTCCTCCCCA GTATCTCCTGAGGTACAAGTCTAATTCAGATATGCAGCAGGGATCTGGAGTCCCCAGCCG CTTCTCTGGG

TCCAAAGATGCTTCGGCCAATGCAGGGATTTTACTCATCTCTGGGCTCCAGTCTGAG GATGAGGCTGACT ATTACTGTATGATTTGGCACCATGGCCATGGGACTAGTCTTGTGTTCGGCGGAGGGACCA AGCTGACCGT CCTAGGTCAGCCCAAGGCTGCCCCCTCGGTCACTCTGTTCCCGCCCTCCTCTGAGGAGCT TCAAGCCAAC AAGGCCACACTGGTGTGTCTCATAAGTGACTTCTACCCGGGAGCCGTGACAGTGGCTTGG AAAGCAGATA GCAGCCCCGTCAAGGCGGGAGTGGAGACCACCACACCCTCCAAACAAAGCAACAACAAGT ACGCGGCCAG CAGCTATCTGAGCCTGACGCCTGAGCAGTGGAAGTCCCACAGAAGCTACAGCTGCCAGGT CACGCATGAA GGGAGCACCGTGGAGAAGACAGTGGCCCCTACAGAATGTTCATAA

>025-LC-AA (SEQ ID NO: 6)

QPVLTQPASLSASPGASASLTCTLHSGISVKDTRIYWYQQKPGRPPQYLLRYKSNSDMQQ GSGVPSRFSG SKDASANAGILLISGLQSEDEADYYCMIWHHGHGTSLVFGGGTKL VIGQPKAAPSVTLFPPSSEELQAN KATLVCLISDFYPGAVTVA KADSSPVKAGVETTTPSKQSNNKYAASSYLSLTPEQ KSHRSYSCQVTHE GSTVEKTVAPTECS

[0092] The 025 antibody includes a common variable heavy domain (SEQ ID NO: 4) encoded by the nucleic acid sequence shown in SEQ ID NO: 3 and includes a lambda variable light domain (SEQ ID NO: 8) encoded by the nucleic acid sequence shown in SEQ ID NO: 7. >0 5-VH-NT (SEQ ID NO: 7)

CAGCCTGTGCTGACTCAGCCGGCTTCCCTCTCTGCATCTCCTGGAGCATCAGCCAGTCTC ACCTGCACCT TGCACAGTGGCATCTCTGTTAAGGATTACAGGATATACTGGTACCAGCAGAAGCCAGGGC GTCCTCCCCA GTATCTCCTGAGGTACAAGTCTAATTCAGATATGCAGCAGGGATCTGGAGTCCCCAGCCG CTTCTCTGGG

TCCAAAGATGCTTCGGCCAATGCAGGGATTTTACTCATCTCTGGGCTCCAGTCTGAG GATGAGGCTGACT ATTACIGTATGATTTGGCACCATGGCCATGGGACTAGTCTIGTGTTCGGCGGAGGGACCA AGCTGACCGT CCTA

>023-VK-AA {SEQ ID NO: 8)

QPVLTQPASLSASPGASASLTCTLHSGISVKDYRIYWYQQKPGRPPQYLLRYKSNSDMQQ GSGVPSRFSG SKDASANAGILLISGLQSEDEADYYCMIWHHGHGTSLVFGGGTKLTVL

(0093] The 030 antibody includes a common heavy chain (SEQ ID NO: 2) encoded by the nucleic acid sequence shown in SEQ ID NO: 1 and includes a lambda light chain (SEQ ID NO: 10) encoded by the nucleic acid sequence shown in SEQ ID NO: 9. The variable region of the lambda light chain is bolded in the amino acid sequence below.

>O30-LC-NT (SEQ ID NO: 9)

CAGTCTGTGCTGACTCAGCCACCCTCAGCGTCTGGGACCCCCGGGCAGAGGGTCACCATC TCTTGTTCTG GAAGCAGCTCCAACATCGCGCATGGGCCTGTAAACTGGTACCAGCAGCTCCCAGGAACGG CCCCCAAACT

CCTCATCTATGCTACTAATCATCGGCCCTCAGGGGTCCCTGACCGATTTTCTGGCTC CAAGTCTGGCACC ACAGCCTCCCTGACCATCAGTGGGCTCCAGTCTGAGGATGAGGCTGATTATTACTGTGCT GCATATGATC TTACGGGCTGGTTTGCGTATGCTGTGTTCGGCGGAGGGACCAAGCTGACCGTCCTAGGTC AGCCCAAGGC TGCCCCCTCGGTCACTCTGTTCCCGCCCTCCTCTGAGGAGCTTCAAGCCAACAAGGCCAC ACTGGTGTGT CTCATAAGTGACTTCTACCCGGGAGCCGTGACAGTGGCTTGGAAAGCAGATAGCAGCCCC GTCAAGGCGG GAGTGGAGACCACCACACCCTCCAAACAAAGCAACAACAAGTACGCGGCCAGCAGCTATC TGAGCCTGAC GCCTGAGCAGTGGAAGTCCCACAGAAGCTACAGCTGCCAGGTCACGCATGAAGGGAGCAC CGTGGAGAAG ACAGTGGCCCCTACAGAATGTTCATAA

>O50-LC-AA (SEQ ID NO: 10)

QSVLTQPPSASGTPGQRV ISCSGSSSNIAHGP¥NWYQQLPGTAPKLLITATNHRPSG¥PDRFSGSKSGT TASLTISGLQSEDEADYXCAAYDLTGWFAYAVFGGGTKLTVLGQPKAAPSVTLFPPSSEE LQANKATLVC LISDFYPGAVTVA KADSSPVKAGVETTTPSKQSNNKYAASSYLSLTPEQWKSHRSYSCQVTHEGSTVEK TVAPTECS

[0094] The 030 antibody includes a common variable heavy domain (SEQ ID NO: 4) encoded by the nucleic acid sequence shown in SEQ ID NO: 3 and includes a lambda variable light domain (SEQ ID NO: 12) encoded by the nucleic acid sequence shown in SEQ ID NO: 11. >O30-VL-NT (SEQ ID NO: 11)

CAGTCTGTGCTGACTCAGCCACCCTCAGCGTCTGGGACCCCCGGGCAGAGGGTCACCATC TCTTGTTCTG GAAGCAGCTCCAACATCGCGCATGGGCCTGTAAACTGGTACCAGCAGCTCCCAGGAACGG CCCCCAAACT CCTCATCTATGCTACTAATCATCGGCCCTCAGGGGTCCCTGACCGATTTTCTGGCTCCAA GTCTGGCACC ACAGCCTCCCTGACCATCAGTGGGCTCCAGTCTGAGGATGAGGCTGATTATTACTGTGCT GCATATGATC TTACGGGCTGGTTTGCGTATGCTGTGTTCGGCGGAGGGACCAAGCTGACCGTCCTA

>O30--VL-AA (SEQ ID NO: 12)

QSVLTQPPSASGTPGQRVTISCSGSSSNIAHGPVNWYQQLPGTAPKLLIYATNHRPSGVP DRFSGSKSGT TASLTISGLQSEDEADYYCAAYDLTGWFAYAVFGGGTKLTVL

[0095] The 035 antibody includes a common heavy chain (SEQ ID NO: 2) encoded by the nucleic acid sequence shown in SEQ ID NO: 1 and includes a lambda light chain (SEQ ID NO: 14) encoded by the nucleic acid sequence shown in SEQ ID NO: 13. The variable region of the lambda light chain is bolded in the amino acid sequence below.

>035-LC-NT (SEQ ID NO: 13)

CAGCCTGTGCTGACTCAGCCGGTTTCCCTCTCTGCATCTCCTGGAGCATCAGTCAGTCTC ACCTGCACCT

TGCGCAGTGACATCAGGGTTAGAGATTACAGGATATTCTGGTACCAGCAGAAGCCAG GGAGTCCTCCCCA GTATCTCCTGAGGTACAAAACCGACTCAGATAAGCAGCAGGGCTCTGGAGTCCCCAGCCG CTTCTCTGGG TCCAAAGATGCTTCGGCCAATGCAGGGATTTTACTCATCTCTGGGCTCCAGTCTGAGGAT GAGGCTGACT ATTACTGTATGATTTGGCACCGCACCACGGGCACTAGTCTTGTGTTCGGCGGAGGGACCA AGCTGACCGT CCTAGGTCAGCCCAAGGCTGCCCCCTCGGTCACTCTGTTCCCGCCCTCCTCTGAGGAGCT TCAAGCCAAC AAGGCCACACTGGTGTGTCTCATAAGTGACTTCTACCCGGGAGCCGTGACAGTGGCTTGG AAAGCAGATA GCAGCCCCGTCAAGGCGGGAGTGGAGACCACCACACCCTCCAAACAAAGCAACAACAAGT ACGCGGCCAG CAGCTATCTGAGCCTGACGCCTGAGCAGTGGAAGTCCCACAGAAGCTACAGCTGCCAGGT CACGCATGAA GGGAGCACCGTGGAGAAGACAGTGGCCCCTACAGAATGTTCATAA

X '5-LC-AA (SEQ ID NO: 14)

QPVLTQPVSLSASPGASVSLTCTLRSDIRVRDYRIFWYQQKPGSPPQYLLRYKTDSDKQQ GSGVPSRFSG SKDASANAGILLISGLQSEDEAD YC IWHRTTGTSLVFGGGTKLTVLGQP AAPSVTLFPPSSEELQAN

KATLVCLISDFYPGAVTVAWKADSSPVKAGVETTTPSKQSNNKYAASSYLSLTPEQW KSHRSYSCQVTHE GS EKTVAPTECS [0096] The 035 antibody includes a common variable heavy domain (SEQ ID NO: 4) encoded by the nucleic acid sequence shown in SEQ ID NO: 3 and includes a lambda variable light domain (SEQ ID NO: 16) encoded by the nucleic acid sequence shown in SEQ ID NO: 15.

>035-VL-NT (SEQ ID NO: 15)

CAGCCTGTGCTGACTCAGCCGGTTTCCCTCTCTGCATCTCCTGGAGCATCAGTCAGTCTC ACCTGCACCT TGCGCAGTGACATCAGGGTTAGAGATTACAGGATATTCTGGTACCAGCAGAAGCCAGGGA GTCCTCCCCA GTATCTCCTGAGGTACAAAACCGACTCAGATAAGCAGCAGGGCTCTGGAGTCCCCAGCCG CTTCTCTGGG TCCAAAGATGCTTCGGCCAATGCAGGGATTTTACTCATCTCTGGGCTCCAGTCTGAGGAT GAGGCTGACT ATTACTGTATGATTTGGCACCGCACCACGGGCACTAGTCTTGTGTTCGGCGGAGGGACCA AGCTGACCGT CCTA

>055-VI.,-AA (SEQ ID NO: 16)

QPVLTQPVSLSASPGASVSLTCTLRSDIRVRDYRIF YQQKPGSPPQYLLRYKTDSDKQQGSGVPSRFSG SKDASANAGILLISGLQSEDEADYYCMIWHRTTGTSLVFGGGTKLTVL

[0097] The 038 antibody includes a common heavy chain (SEQ ID NO: 2) encoded by the nucleic acid sequence shown in SEQ ID NO: 1 and includes a lambda light chain (SEQ ID NO: 29) encoded by the nucleic acid sequence shown in SEQ ID NO: 30

>03S-LC--NT (SEQ ID NO: 29)

CAGCCTGTGCTGACTCAGCCGGCTTCCCTCTCTGCATCTCCTGGGGCATCAGCCAGTCTC ACCTGCACCT TGCGCAGTGGCATCAACGTTAGAGATTACAGGATATTCTGGTACCAGCAGAAGCCAGGGA GTCCTCCCCA GTATCTCCTGAGGTACAAAAGCGCATCAGATAAGCAGCAGGGCTCTGGAGTCCCCAGCCG CTTCTCTGGG TCCAAAGATGCTTCGGCCAATGCAGGGATTTTACTCATCTCTGGGCTCCAGTCTGAGGAT GAGGCTGACT ATTACTGTATGATTTGGCACCACGATTCGGAGGGGCATGCTTTTGTGTTCGGCGGAGGGA CCAAGCTGAC CGTCCTAGGTCAGCCCAAGGCTGCCCCCTCGGTCACTCTGTTCCCGCCCTCCTCTGAGGA GCTTCAAGCC AACAAGGCCACACTGGTGTGTCTCATAAGTGACTTCTACCCGGGAGCCGTGACAGTGGCT TGGAAAGCAG ATAGCAGCCCCGTCAAGGCGGGAGTGGAGACCACCACACCCTCCAAACAAAGCAACAACA AGTACGCGGC CAGCAGCTATCTGAGCCTGACGCCTGAGCAGTGGAAGTCCCACAGAAGCTACAGCTGCCA GGTCACGCAT GAAGGGAGCACCGTGGAGAAGACAGTGGCCCCTACAGAATGTTCATAA

>038 -LC-AA (SEQ ID NO: 30) QPVLTQPASLSASPGASASLTCTLRSGIN¥RDTRIF YQQKPGSPPQYLLRYKSASDKQQGSGVPS FSG SKDASANAGILLISGLQSEDEADYYCMIIfHHDSEGHAFVFGGGTKL VLGQPKAAPSVTLFPPSSEELQA NKATLVCLISDFYPGAVTVA KADSSPVKAGVETTTPSKQSNNKYAASSYLSLTPEQ KSHRSYSCQVTH

EGSTVEKTVAPTECS

[0098| The 038 antibody includes a common variable heavy domain (SEQ ID NO: 4) encoded by the nucleic acid sequence shown in SEQ ID NO: 3 and includes a lambda variable light domain (SEQ ID NO: 31) encoded by the nucleic acid sequence shown in SEQ ID NO: 32.

>03S-VL"NT (SEQ ID NO: 31)

CAGCCTGTGCTGACTCAGCCGGCTTCCCTCTCTGCATCTCCTGGGGCATCAGCCAGTCTC ACCTGCACCTTGCGCAGTGGCATCAACGTTAGAGATTACAGGATATTCTGGTACCAGCAG AAGCCAGGGAGTCCTCCCCAGTATCTCCTGAGGTACAAAAGCGCATCAGATAAGCAGCAG GGCTCTGGAGTCCCCAGCCGCTTCTCTGGGTCCAAAGATGCTTCGGCCAATGCAGGGATT TTACTCATCTCTGGGCTCCAGTCTGAGGATGAGGCTGACTATTACTGTATGATTTGGCAC CACGATTCGGAGGGGCATGCTTTTGTGTTCGGCGGAGGGACCAAGCTGACCGTCCTA

>038-VL-AA (SEQ ID NO: 32)

QPVLTQPASLSASPGASASLTCTLRSGINVRDYRIFWYQQKPGSPPQYLLRYKSASDKQQ GSGVPSRFSG SKDASANAGILLISGLQSEDEADYYCMI HHDSEGHAFVFGGGTKLTVL

[0099] The 041 antibody includes a common heavy chain (SEQ ID NO: 2) encoded by the nucleic acid sequence shown in SEQ ID NO: 1 and includes a lambda light chain (SEQ ID NO: 18) encoded by the nucleic acid sequence shown in SEQ ID NO: 17. The variable region of the lambda light chain is bolded in the amino acid sequence below.

>04i-LC-NT (SEQ ID NO: 17)

TCCTATGTGCTGACTCAGCCACCCTCAGTGTCAGTGGCCCCAGGAAAGACGGCCAGGATT ACCTGTGGGG GAAACAAAATTGGACACCGCGCCGTGCACTGGTACCAGCAGAAGCCAGGCCAGGCCCCTG TGCTGGTCAT CTATTATACCTATGAGCGGCCCTCAGGGATTCCTGAGCGATTCTCTGGCTCCAACTCTGG GAACACGGCC ACCCTGACCATCAGCAGGGTCGAAGCCGGGGATGAGGCCGACTATTACTGTCAGGTGTGG GATTGGTACA GCGAGGGGGGGGTTGTGTTCGGCGGAGGGACCAAGCTGACCGTCCTAGGTCAGCCCAAGG CTGCCCCCTC GGTCACTCTGTTCCCGCCCTCCTCTGAGGAGCTTCAAGCCAACAAGGCCACACTGGTGTG TCTCATAAGT GACTTCTACCCGGGAGCCGTGACAGTGGCTTGGAAAGCAGATAGCAGCCCCGTCAAGGCG GGAGTGGAGA CCACCACACCCTCCAAACAAAGCAACAACAAGTACGCGGCCAGCAGCTATCTGAGCCTGA CGCCTGAGCA GTGGAAGTCCCACAGAAGCTACAGCTGCCAGGTCACGCATGAAGGGAGCACCGTGGAGAA GACAGTGGCC

CCTACAGAATGTTCATAA

>041 -LC-AA (SEQ ID NO: 18)

SYVLTQPPSVSVAPGKTARITCGGNKIGHRAVHWTQQKPGQAP¥LVIYYTYERPSGIPE RFS6SNSGNTA TLTISRVEAGDEADYYCQVWDWYSEGGV GGGTMiTVLGQPKAAPSVTLFPPSSEELQANKATLVCLIS DFYPGAVTVAWKADSSPVKAGVETTTPSKQSNNKYAASSYLSLTPEQ KSHRSYSCQVTHEGSTVEKTVA PTECS

[00100] The 041 antibody includes a common variable heavy domain (SEQ ID NO: 4) encoded by the nucleic acid sequence shown in SEQ ID NO: 3 and includes a lambda variable light domain (SEQ ID NO: 20) encoded by the nucleic acid sequence shown in SEQ ID NO: 19.

>041-VL-NT (SEQ ID NO: 19)

TCCTATGTGCTGACTCAGCCACCCTCAGTGTCAGTGGCCCCAGGAAAGACGGCCAGGATT ACCTGTGGGG GAAACAAAATTGGACACCGCGCCGTGCACTGGTACCAGCAGAAGCCAGGCCAGGCCCCTG TGCTGGTCAT CTATTATACCTATGAGCGGCCCTCAGGGATTCCTGAGCGATTCTCTGGCTCCAACTCTGG GAACACGGCC ACCCTGACCATCAGCAGGGTCGAAGCCGGGGATGAGGCCGACTATTACTGTCAGGTGTGG GATTGGTACA GCGAGGGGGGGGTTGTGTTCGGCGGAGGGACCAAGCTGACCGTCCTA

>042-VL-AA (SEQ ID NO: 20)

SYVLTQPPSVSVAPGKTARITCGGNKIGHRAVHWYQQKPGQAPVLVIYYTYERPSGIPER FSGSNSGNT TLTISRVEAGDEADYYCQVWDWYSEGGVVFGGGTKLTVL

(00101J The L7-2 antibody includes a common heavy chain (SEQ ID NO: 2) encoded by the nucleic acid sequence shown in SEQ ID NO: 1 and includes a lambda light chain (SEQ ID NO: 22) encoded by the nucleic acid sequence shown in SEQ ID NO: 21.

>L7-2-VL~NT (SEQ ID NO: 21)

AATTTTATGCTGACTCAGCCCCACTCTGTGTCGGAGTCTCCGGGGAAGACGGTAACCATC TCCTGCACCC GCAGCAGTGGCTCTATCGAAGATAAGTATGTGCAGTGGTACCAGCAGCGCCCGGGCAGTT CCCCCACCAT TGTGATCTATTATGATAACGAAAGACCCTCTGGGGTCCCTGATCGGTTCTCTGGCTCCAT CGACAGCTCC TCCAACTCTGCCTCCCTCACCATCTCTGGACTGAAGACTGAGGACGAGGCTGACTACTAC TGTCAGACCT ACGACCAGAGCCTGTATGGTTGGGTGTTCGGCGGAGGGACCAAGCTGACCGTCCTAGGTC AGCCCAAGGC TGCCCCCTCGGTCACTCTGTTCCCGCCCTCCTCTGAGGAGCTTCAAGCCAACAAGGCCAC ACTGGTGTGT

CTCATAAGTGACTTCTACCCGGGAGCCGTGACAGTGGCTTGGAAAGCAGATAGCAGC CCCGTCAAGGCGG GAGTGGAGACCACCACACCCTCCAAACAAAGCAACAACAAGTACGCGGCCAGCAGCTATC TGAGCCTGAC GCCTGAGCAGTGGAAGTCCCACAGAAGCTACAGCTGCCAGGTCACGCATGAAGGGAGCAC CGTGGAGAAG ACAGTGGCCCCT CAGAATGTTCATAA

>L7-2-LC-AA (SEQ ID NO: 22)

NFMLTQPHSVSESPGKTVTISCTRSSGSIEDKYVQWYQQ PGSSPTIVIYYD ERPSGVPDRFSGSIDSS

SNSASLTISGLKTEDEADYYCQTYDQSLYGWVFGGGTKLTVLGQPKAAPSVTLFPPS SEELQANKATLVC LISDFYPGAVTVAWKADSSPVKAG¥ETTTPSKQSNNKYAASSYLSLTPEQWKSHRSYSC QVTHEGSTVEK TVAPTECS

[00102| The C2 antibody includes a common variable heavy domain (SEQ ID NO: 1 14) encoded by the nucleic acid sequence shown in SEQ ID NO: 1 13 and includes a lambda variable light domain (SEQ ID NO: 24) encoded by the nucleic acid sequence shown in SEQ ID NO: 23.

>L7-2-VL-NT (SEQ ID NO: 23}

AATTTTATGCTGACTCAGCCCCACTCTGTGTCGGAGTCTCCGGGGAAGACGGTAACCATC TCCTGCACCC

GCAGCAGTGGCTCTATCGAAGATAAGTATGTGCAGTGGTACCAGCAGCGCCCGGGCA GTTCCCCCACCAT TGTGATCTATTATGATAACGAAAGACCCTCTGGGGTCCCTGATCGGTTCTCTGGCTCCAT CGACAGCTCC TCCAACTCTGCCTCCCTCACCATCTCTGGACTGAAGACTGAGGACGAGGCTGACTACTAC TGTCAGACCT ACGACCAGAGCCTGTATGGTTGGGTGTTCGGCGGAGGGACCAAGCTGACCGTCCTA

>L7-2-VL-AA (SEQ ID NO: 24)

NFMLTQPHSVSESPGKTVTISCTRSSGSIEDKYVQWYQQRPGSSPTIVIYYDNERPSGVP DRFSGSIDSS SNSASLTISGLKTEDEADYYCQTYDQSLYG VFGGGTKLTVL

[00103] The K2 antibody includes a common heavy chain (SEQ ID NO: 2) encoded by the nucleic acid sequence shown in SEQ ID NO: 1 and includes a kappa light chain (SEQ ID NO: 26) encoded by the nucleic acid sequence shown in SEQ ID NO: 25.

>K2~LC-NT (SEQ ID NO: 25)

GACATCCAGATGACCCAGTCTCCATCCTCCCTGTCTGCATCTGTAGGAGACAGAGTCACC ATCACTTGCC GGGCAAGTCAGAGCATTAGCAGCTATTTAAATTGGTATCAGCAGAAACCAGGGAAAGCCC CTAAGCTCCT GATCTATGCTGCATCCAGTTTGCAAAGTGGGGTCCCATCAAGGTTCAGTGGCAGTGGATC TGGGACAGAT

TTCACTCTCACCATCAGCAGTCTGCAACCTGAAGATTTTGCAACTTACTACTGTCAG CAGATGCACCCGC GCGCCCCGAAGACCTTCGGCCAAGGGACCAAGGTGGAAATCAAACGTACGGTGGCTGCAC CATCTGTCTT CATCTTCCCGCCATCTGATGAGCAGTTGAAATCTGGAACTGCCTCTGTTGTGTGCCTGCT GAATAACTTC TATCCCAGAGAGGCCAAAGTACAGTGGAAGGTGGATAACGCCCTCCAATCGGGTAACTCC CAGGAGAGTG TCACAGAGCAGGACAGCAAGGACAGCACCTACAGCCTCAGCAGCACCCTGACGCTGAGCA AAGCAGACTA CGAGAAACACAAAGTCTACGCCTGCGAAGTCACCCATCAGGGCCTGAGCTCGCCCGTCAC AAAGAGCTTC AACAGGGGAGAGTGTTAA

>K2-LC-AA (SEQ ID NO: 26)

DIQMQSPSSLSASVGDRV ITCRASQSISSYLNWYQQKPGKAPKLLIYAASSLQSGVPSRFSGSGSGTD FTLTISSLQPEDFATXYCQQMHPRAPKTFGQGTKVEIKRTVAAPSVFIFPPSDEQLKSGT ASVVCLLNNF YPREAKVQWKVDNALQSGNSQESVTEQDSKDSTYSLSSTLTLSKADYEKHKVYACEVTHQ GLSSPVTKSF NRGEC

[00104] The K2 antibody includes a common variable heavy domain (SEQ ID NO: 4) encoded by the nucleic acid sequence shown in SEQ ID NO: 3 and includes a kappa variable light domain (SEQ ID NO: 28) encoded by the nucleic acid sequence shown in SEQ ID NO: 27.

>K2-VL-NT (SEQ ID NO: 27)

GACATCCAGATGACCCAGTCTCCAICCTCCCTGTCTGCATCTGTAGGAGACAGAGICACC ATCACITGCC GGGCAAGTCAGAGCATTAGCAGCTATTTAAATTGGTATCAGCAGAAACCAGGGAAAGCCC CTAAGCTCCT GATCTATGCTGCATCCAGTTTGCAAAGTGGGGTCCCATCAAGGTTCAGTGGCAGTGGATC TGGGACAGAT TTCACTCTCACCATCAGCAGTCTGCAACCTGAAGATTTTGCAACTTACTACTGTCAGCAG ATGCACCCGC GCGCCCCGAAGACCTTCGGCCAAGGGACCAAGGTGGAAATCAAA

>K2 VL-AA (SEQ ID NO: 28)

DIQMTQSPSSLSASVGDRVTITCRASQSISSYLNWYQQKPGKAPKLLIYAASSLQSGVPS RFSGSGSGTD FTLTISSLQPEDFATYYCQQMHPRAPKTFGQGTKVEIK

EXAMPLE 2: Expression and purification of bispecific antibodies carrying a Lambda and a Kappa light chain.

[00105] The simultaneous expression of one heavy chain and two light chains in the same cell can lead to the assembly of three different antibodies. Simultaneous expression can be achieved in different ways such as the transfection of multiple vectors expressing one of the chains to be co-expressed or by using vectors that drive multiple gene expression. A vector ρΝονί Ηλ was previously generated to allow for the co-expression of one heavy chain, one Kappa light chain and one Lambda light chain as described in US 2012/0184716 and WO 2012/023053, each of which is hereby incorporated by reference in its entirety. The expression of the three genes is driven by human cytomegalovirus promoters (hCMV), and the vector also contains a glutamine synthetase gene (GS) that enables the selection and establishment of stable cell lines. The VL genes of the anti-hMSLN IgGl, the anti-hCDl9IgG or the anti-hCD47 IgCrK were cloned in the vector ρΝονίκΗλ, for transient expression in mammalian cells. Peak cells were amplified and split in T175 flasks at a concentration of 8 x 10 ⁶ cells per flask in 45 nxL culture media containing fetal bovine serum. 30 μg of plasmid DNA were trans fee ted into the cells using Lipofectamine 2000 transfection reagent according to manufacturer's instructions. Antibody concentration in the serum-containing supernatant of trans ected cells was measured at several time points during the production using the Bio-Layer Interferometry (BLI) technology. An OctetRED96 instrument and Protein A biosensors were used for quantitation (Pall, Basel, Switzerland). 200 of supernatant were used to determine IgG concentration; biosensors were pre-conditioned and regenerated using 10 mM glycine pH 1.7 and IgG calibrators diluted in conditioned PEAK cell medium were prepared for standard curve generation. Concentrations were determined using the dose response 5 L weighted Y standard curve equation and an initial slope binding rate equation. According to antibody concentration, supernatants were harvested 7 to 10 days after trans ection and clarified by centrifugation at 1300 g for 10 min. The purification process was composed of three affinity steps. First, the CaptureSelect™ IgG-CH l affinity matrix (Thermo Fisher Scientific, Waltham, MA) was washed with PBS and then added in the clarified supernatant. After incubation overnight at +4°C, supernatants were centrifuged at 1000 g for 10 min, flow through was stored and resin washed twice with PBS. Then, the resin was transferred on spin columns and a solution containing 50 mM glycine at pH 3.0 was used for elution. Several elution fractions were generated, pooled and desalted against 25mM Histidine/125mM NaCl pH6.() buffer using 50 kDaAmicon® Ultra Centrifugal filter units (Merck KGaA, Darmstadt, Germany). The final product, containing total human IgGs from the supernatant, was quantified using a Nanodrop spectrophotometer (NanoDrop Technologies, Wilmington, DE) and incubated for 30 min at RT and 20 rpm with the appropriate volume of CaptureSelect™ LC-kappa (Hu) affinity matrix (Thermo Fisher Scientific, Waltham, MA). Incubation, resin recovery, elution and desalting steps were performed as described previously. The last affinity purification step was performed using the CaptureSelect™ LC-lambda (Hu) affinity matrix (Thermo Fisher Scientific, Waltham, MA) applying the same process as for the two previous purifications. The final product was quantified using the Nanodrop. Purified bispecific antibodies were analyzed by electrophoresis in denaturing and reducing conditions. The Agilent 2100 Bioanalyzer was used with the Protein 80 kit as described by the manufacturer (Agilent Technologies, Santa Clara, CA, USA). 4 μΕ of purified samples were mixed with sample buffer supplemented with dithiothreitol (DTT; Sigma Aldrich, St. Louis, MO). Samples were heated at 95°C for 5 min and then loaded on the chip. An aliquot from the first purification step (containing the bispecific antibody and both monospecific mAbs) and an aliquot of the final product were loaded on an IsoElectric Focusing (IEF) gel to evaluate the purity of the final purified bispecific antibody (absence of mAb contamination). The aggregate level was determined by SEC- HPLC. Finally, binding o the bispecific antibodies on both targets was assessed using the OctetRED96. Briefly, biotinylated targets (hMSLN, hCD19, hCD47 and an irrelevant target) were loaded on a streptavidin biosensor. Then this biosensor was dipped into a solution containing the bispecific antibody and binding was monitored in real time. All samples were tested for endotoxin contamination using the Limu!us Amebocyte Lysate test (LAL; Charles River Laboratories, Wilmington, MA). The following bispecific antibodies were expressed an purified:

Table 4.

Bispecific antibody Specificities

K2025 Anti-hCD47; anti-hMSLN

K2O30 Anti-hCD47; anti-hMSLN

K2035 Anti-hCD47; anti-hMSLN

K2038 Anti-hCD47; anti-hMSLN

K2041 Anti-hCD47; anti-hMSLN

K2L7-2 Anti-hCD47; anti-hCD 19

EXAMPLE 3: Activity in phagocytosis assays of individual bispecific antibodies targeting hMSLN and hCD47 and combinations of bispecific antibodies

[00106] Bispecific antibodies targeting a protein or antigen expressed at the surface of a tumor cell and capable o co-engaging CD47 on the same tumor cell can mediate an increase in phagocytic activity of macrophage by preventing the inhibitory signal mediated by the interaction of CD47 with SIRPa. This principle has been described in in PCT Publication No. WO 2014/087248 and in co-pending patent application US Patent Application No. 62/51 1 ,669, filed May 26, 2017 and entitled "Anti-CD47 x Anti-Mesothelin Antibodies and Methods of Use Thereof. Here we tested the hypothesis of multi-epitope targeting of a tumor associated antigen (TAA), in this case hMSLN combined with CD47 blockade as illustrated in Figure 7. In order to evaluate the benefit of having a higher density of Fc and higher CD47 blockade per molecule of MSLN, we compared the phagocytic activity mediated by either individual hCD47/hMSLN bispecific antibodies or by pairwise combinations of bispecific antibodies. The hCD47/hCD 19 bispecific antibody K2L7-2 was used as a monovalent anti-hCD47 control as no hCD 19 is present in the assay. Three tumor cell lines expressing different levels of hCD47 and hMSLN were tested: NCI-N87, HP AC and Caov-3. The levels of cell surface expression of CD47 and Mesothelin for NCI-N87 cells were 43,000 and 27,000, respectively. The levels of cell surface expression of CD47 and Mesothelin for HP AC cells were 105,000 and 13,000, respectively. The levels of cell surface expression of CD47 and Mesothelin for Caov-3 cells were 220,000 and 38,000, respectively.

[001071 The assay was performed with human macrophages differentiated from peripheral blood monocytes and NCI-N87, HP AC or CaOV3 as target cells. Macrophages were co-incubated with Calcein AM-labeled target cells (effector: target ratio 1 : 1) for 2.5 hours at 37°C in the presence of increasing concentrations of bispecific or monovalent antibody. At the end o the incubation period, supernatants were replaced by complete culture medium. The plates were imaged with the CX5 Imaging platform and 1500 macrophages were acquired and analyzed per conditions. Phagocytosis was evidenced by double-positive events and the phagocytosis indexes were calculated by the software.

[00108J These dose response experiments indicate that the phagocytic activity mediated by the combination of K2O30 + K2041 was superior to the activity of either K2O30 or K2041 on all three cell types (Figure 12 A, 13A and 14A). Similarly, the phagocytic activity mediated by the combination of K20 5 + K2041 was superior to the activity of either K2035 or K2041 on all there cell types (Figure 12B, 13B and 14B). In addition, the phagocytic activity mediated by the combination of K2038 + K2041 was superior to the activity of either K2038 or K2041 on NCI- N87 and HPAC cells (Figure 12C and 14C). The combination of K2O30 +K2025 was superior to the activity of either K2O30 or K2025 on NCI-N87 cells (Figure 15). In all case the negative control IgG (referred to as NI-080 I , described in PCT Publication No. WO2008/ 106200) or the monovalent anti-hCD47 control only induced minimal or no phagocytic activity. In all these examples, the two bispecific antibodies were added in equal amounts (i.e., at a 50:50 ratio). In order to evaluate whether slight changes in this equimolar ratio would influence the activity, K2O30 and K2025 were mixed at 40:60 and 60:40 ratios and tested. All ratios gave similar results indicating slight differences in ratio between the two bispecific antibodies do not significantly impact the increase in phagocytic activity of the combination (Figure 15 A).

EXAMPLE 4: Expression of four antibody chains from a single vector

[00109] The vector ρΝονίκΗλ was modified to enable the expression of an additional light chain. The new vector, pNoviH3 L contains four promoters driving the expression of one heavy chain and three light chains. The K2 anti-CD47 Kappa light chain and the two anti-hMSLN Lambda light chains O30 and 025 were cloned into this single vector. All coding sequences and cloning junctions were verified by sequencing. This vector was used in transient transfection as described in Example 2 to verify its functionality, L e. , that bispecific antibodies could be produced. The vector was then linearized for electroporation in Chinese Hamster Ovary (CHO) cells.

EXAMPLE 5: Expression of antibody mixtures containing bispecific and monospecific antibodies.

[001 lOj In the studies presented herein, stable CHO lines were transfected and grown using a chemically defined, animal component- free (CDACF) manufacturing process. After transfection by electroporation and selection with MSX, a screening by FACS was performed. The highest producing pools were selected for production in fed batch conditions. Total IgG productivity was assessed for different pools by Octet technology. Affinity purification was performed as described in Example 2. The material after the Protein A chromatography step, thus containing all IgG forms, was analyzed for polypeptide content using an Agilent Bioanalyzer, by isoelectric focusing gel (IEF) and the aggregate level determined by SEC-HPLC. Finally, binding of the bispecific antibodies on both targets was assessed using the OctetRED96. Briefly, biotinylated targets (hMSLN, hCD47 and an irrelevant target) were loaded on a streptavidin biosensor. Then this biosensor was dipped into a solution containing the bispecific antibody and binding was monitored in real time. All samples were tested for endotoxin contamination using the Limulus Amebocyte Lysate test (LAL; Charles River Laboratories, Wilmington, MA).

EXAMPLE 6: Purification of a submixturc containing two bispecific antibodies

[00111] The mixture of all IgG forms (Bispecific and monospecific IgGs) that was expressed and purified in Example 5 was further subjected to two steps of chromatography using media that specifically interact with the Kappa or Lambda light chain constant domains such as the CaptureSelect Fab Kappa and CaptureSelect Fab Lambda affinity matrices (GE Healthcare). These two steps allowed to recover all the IgGs forms containing both a kappa light chain and a lambda light chain as indicated in Figure 6 and 7B. The total IgG was then applied to a column containing CaptureSelect Fab kappa affinity matrix (GE Healthcare) equilibrated with ten volumes of PBS. The column was then washed with 5- 10 column volumes of PBS. All the immunoglobulin molecules bearing a kappa light chain were eluted from the column by applying 5 column volumes 0.1 M Glycine pH 3.0 and fractions were collected. The fractions containing the antibody were pooled before buffer exchange on an Amicon buffer exchange column (Merck) equilibrated with PBS. The antibody was then applied on a second column containing CaptureSelect Fab lambda affinity matrix equilibrated with ten volumes of PBS. The column was then washed with 5- 10 column volumes of PBS. All the immunoglobulin molecules bearing only kappa light chain do bind to the column and were found in the flowthrough. Antibodies carrying a lambda light chain were eluted from the column by applying 5 column volumes 0.1 M Glycine pH 3.0 and fraction were collected. The fractions containing the bispecific antibodies were pooled before buffer exchange on an Amicon buffer exchange column (Merck) equilibrated with PBS. The flow through and elution fraction of each purification step was analyzed by SDS-PAGE and indicated that the antibodies bearing lambda light chains were found in the flow through of the CaptureSelect Fab kappa affinity matrix and, conversely, that antibodies bearing kappa light chains were found in the flow through of the CaptureSelect Fab lambda affinity matrix.

[00112] The mixture of all IgG forms after obtained after the Protein A step and the K2O25O30 submixture containing two bispecific antibody forms that was purified as described above were analyzed by IEF (Figure 16A). As expected 6 bands corresponding to the different mono- and bispecific antibody forms were visible for the mixture obtained after Protein A. After the three affinity purification steps, two bands corresponding to the bispecific forms containing a kappa and lambda light chain were visible. The purified K2O25O30 submixture was further analyzed by hydrophobic interaction chromatography (HIC) to determine the relative content of the two bispecific antibody forms. The integration of the two distinct peaks revealed that K2O25O30 contains 69% of K2025 and 3 1 % of K2O30 (Figure 16B).

EXAMPLE 7: Activity in phagocytosis assays of a submixture containing two bispecific antibodies targeting hMSL and hCD47 and a combination of the same two bispecific antibodies

[001131 The activity of the K2O25O30 submixture isolated in Example 6 was then compared to the equimolar combination of K2025 and K2O30 in an in vitro phagocytosis assay using NCI-N87 cells as described in Example 3. The activities of K2O25O30 and the equimolar combination of K2025 and K2O30 were identical (Figure 17), indicating that the submixture recapitulates the increase in activity of the combination compared to the activity of the individual bispecific antibodies as described in Example 3. EXAMPLE 8: in vivo Antitumor Activity of Bispcciflc Antibodies combinations

[001141 The anti-tumor activity of 2 CD47xMSLN κλ bodies (K2038 and K2041 ) and one combinations of CD47xMSLN κλ bodies (K2038+K2041) were evaluated in a MSLN-transfected HepG2 model of liver cancer. 3.10 ⁶ HepG2-MSLN cells were implanted subcutaneous ly in NOD/SCID mice. Tumor volumes were measured 2 to 3 times per week (using the following formula: (length x width ² )/2). Treatment was initiated when the tumor reached between 150 to 200mm ³ . Mice were randomized into 8 groups (6-7 mice per group). This experiment compared the effect of combinations of CD47xMSLN κλ-bodies to the effect of the CD47xMSLN κλ-bodies alone and the MSLN Mab Amatuximab. Antibody was injected i.v. once a week until the end of the experiment (d28). All the antibodies were administered at 60mg/kg per injection. The tumor volume measurement was used to calculate area under the curve (AUC) for each individual mouse. For statistical analyses, a one-way ANOVA was performed followed by multiple comparison test (Tukey's multiple comparison), using GraphPad Prism. p< 0.05 is considered to be statistically significant. Percentage of tumor growth inhibition (TGI), in comparison to isotype control group, was also determined based on tumor volumes, using the formula: %TGI = { 1- [(Tt T0)/(Vt-V0)]} X 100; with Tt = median tumor volume of treated at time t; TO = median tumor volume of treated at time 0; Vt = median tumor volume of control at time t and V0 = median tumor volume of control at time 0.

[00115J As shown in Figure 18, treatment with the 2 CD47xMSLN κλ bodies significantly reduced tumor growth as compared to hlgG 1 control, with a TGI of 21 % for K2041 and 46% for K2038. The combination demonstrated a superior effect on tumor growth, as compared to the single approach, with a TGI of 80% for K2041+K2038. Finally, both the single CD47xMSLN κλ bodies and their combination exhibited a superior anti-tumor effect as compared to Amatuximab, a neutralizing mAb recognizing MSLN, currently under clinical investigation.

Other Embodiments

100116] While the invention has been described in conjunction with the detailed description thereof, the foregoing description is intended to illustrate and not limit the scope of the invention, which is defined by the scope of the appended claims. Other aspects, advantages, and modifications are within the scope of the following claims.