Preparation of compositions containing two or more antibodies

Abstract

To provide means and methods of producing at least two antibodies.SOLUTION: A method comprises: providing cells with nucleic acid that encodes antibodies; culturing the cells; collecting the antibodies from the culture; and separating produced antibodies from half antibodies by ion exchange chromatography (IEX). In some embodiments the antibodies exhibit IEX retention times that deviate by 10% or less from the average of the retention times of the individual antibodies under the IEX conditions used. The invention also relates to compositions of antibodies thus produced. In some aspects the invention relates to compositions comprising 2 to 10 recombinant antibodies, characterized in that the IEX retention times of at least two of the antibodies deviate by 10% or less from the average of the retention times of the individual antibodies under the IEX conditions. It also relates to compositions comprising 2 to 10 recombinant antibodies, characterized in that the pIs of at least two of the antibodies differ by 0.4 units or less from the average pI of the at least two antibodies.SELECTED DRAWING: None

Description

[Technical field]

[0001] The present invention relates to the field of antibodies, in particular to the field of therapeutic antibodies. The antibodies can be used for human therapy. More specifically, the present invention relates to the production and / or purification of multiple antibodies. A single host cell can produce multiple antibodies. Antibodies can also be produced by a mixture of host cells, each producing one of the antibodies. The present invention also relates to methods of producing compositions comprising such antibodies, and methods for purifying such antibodies. [Background technology]

[0002] Polyclonal antibodies are typically obtained from the blood of a subject. The advantage of polyclonal antibodies is that the pathogen is attacked through multiple targets and epitopes. The advantage of monoclonal or recombinant antibodies is their well-characterized specificity and function, which allows such antibodies to be used as precision drugs with well-defined spectra of action and toxicity. Summary of the Invention [Problem to be solved by the invention]

[0003] The specificity of monoclonal antibodies can also be a drawback, especially when multiple targets need to be addressed. It is possible to reduce this drawback by adding more antibodies to the drug, but considering that even a single therapeutic antibody can be expensive, it is easy to imagine that the cost of such a polyclonal combination would be very high.

[0004] The development of bispecific, trispecific and other multispecific antibodies has been successful in introducing some aspects of polyclonal antibodies into antibody therapeutics. Apart from increasing the number of targets, it has also been successful in introducing additional functions that were not previously achievable with classical monospecific mono- or polyclonal antibodies. The use of bispecific, trispecific and other multispecific antibodies in the same therapy can provide additional benefits.

[0005] The present invention provides an advancement in the art by describing a robust and economical method for the purification of multiple antibody therapeutics produced from a single host cell or alternatively a mixture of host cells. The present invention is particularly useful for the economical production and manufacturing of collections of two or more antibodies, preferably multispecific antibodies. [Means for solving the problem]

[0006] The present invention provides a method for producing at least two multispecific antibodies, comprising the steps of: providing a cell with nucleic acid encoding a multispecific antibody; Culturing the cells; and Recovering the multispecific antibody from the culture. separating the produced multispecific antibodies from half-antibodies, and optionally monospecific antibodies and / or other undesired antibody product related impurities, by ion exchange chromatography (IEX); The multispecific antibody is characterized in that it exhibits an IEX retention time that deviates from the average of the retention times of the individual multispecific antibodies by no more than 10% under the IEX conditions used. Preferably, the retention times of each half antibody and optionally monospecific antibodies and / or other undesired antibody product related impurities are outside the range spanned by the retention time of the multispecific antibody.

[0007] The present invention also provides a method for producing at least two multispecific antibodies, comprising the steps of: providing a cell with nucleic acid encoding a multispecific antibody; Culturing the cells; and Recovering the multispecific antibody from the culture. separating the produced multispecific antibodies from half-antibodies, and optionally monospecific antibodies and / or other undesired antibody product related impurities, by ion exchange chromatography (IEX); The method is characterized in that the multispecific antibodies exhibit IEX retention times that are essentially the same under the IEX conditions used, preferably with the retention times of each half antibody and optionally monospecific antibodies and / or other undesired antibody product related impurities falling outside the range spanned by the retention time of the antibody.

[0008] The present invention further provides a method for producing at least two antibodies, the antibodies including monospecific and / or multispecific antibodies, providing a cell with nucleic acid encoding an antibody; Culturing the cells; and recovering the antibody from the culture; and separating the produced antibodies from half antibodies by ion exchange chromatography (IEX), The method is characterized in that the antibodies exhibit IEX retention times that are essentially the same under the IEX conditions used, preferably with the retention times of each half antibody and optionally the monospecific antibody falling outside the range spanned by the retention time of the antibody.

[0009] The present invention also provides a composition comprising 2 to 10 recombinant antibodies obtainable by the methods described herein.

[0010] Also provided is a composition comprising 2 to 10 recombinant antibodies, such as multispecific antibodies, characterized in that the IEX retention times of at least two of the antibodies deviate by no more than 10% from the average of the retention times of the individual antibodies under the IEX conditions used.

[0011] Further provided is a composition comprising 2 to 10 recombinant antibodies, characterized in that at least two of the antibodies have essentially the same IEX retention time.

[0012] Also provided is a composition comprising 2-10 recombinant antibodies, characterized in that the isoelectric points (pI) of at least two of the antibodies preferably differ from the average pI of the at least two antibodies by no more than 0.4 units, 0.3, 0.2, and preferably no more than 0.1 units. The pI of each of the at least two antibodies preferably differs from each other by no more than 0.25 units.

[0013] As used herein, the term "antibody" refers to a protein molecule belonging to the immunoglobulin class of proteins that contains one or more domains that bind to an epitope on an antigen, such domains being derived from or sharing sequence homology with the variable region of the antibody. Antibodies are typically composed of a basic structural unit, each having two heavy chains and two light chains. Therapeutic antibodies are preferably as close as possible to the natural antibodies of the subject to be treated (e.g., human antibodies for human subjects). Antibodies with extended heavy and / or light chain variable regions are also included herein. The antibodies according to the present invention are not limited to any particular format or method of producing them.

[0014] A half antibody is a combination of a heavy and light chain that is not associated with other combinations of heavy and light chains and does not form an interface with other variable regions or variable region-like polypeptides. Other undesirable antibody product-related impurities can be free light chains not associated with heavy chains, free heavy chains not associated with light chains or another heavy chain, or incompletely assembled antibodies lacking light chains.

[0015] Suitable cells for antibody production are known in the art and include hybridoma cells, Chinese Hamster Ovary (CHO) cells, NS0 cells or PER-C6 cells, or various other cell lines known to the general skilled artisan. Various institutions and companies have developed cell lines for large-scale production of antibodies, for example for clinical use. Non-limiting examples of such cell lines are CHO cells, NS0 cells or PER.C6 cells or HEK293 cells, among others. In a particularly preferred embodiment, the cell is a human cell. Preferably, a cell transformed by an adenovirus E1 region or a functional equivalent thereof. In a preferred embodiment, the cell is a CHO cell or a variant thereof. Preferably, a variant utilizing the glutamine synthetase (GS) vector system for the expression of the antibody. In a preferred embodiment, the cell is a CHO cell.

[0016] The cells can be provided with nucleic acid encoding the antibody. The cells will express, assemble, and excrete the formed antibody into the cell culture supernatant. Introduction of a single heavy and light chain (or, more precisely, the nucleic acid encoding it) leads to the production of a monoclonal antibody with two heavy chains each associated with a light chain.

[0017] The method of the present invention can be carried out using a mixture of cells, including two or more types of cells, each producing a different antibody. The advantage of using such a mixture is that downstream processing of the recovered antibodies can be simplified and made more efficient. A further advantage of the method is that the antibody products are recovered and purified as one. Also, a mixture of two or more antibodies produced by the method of the present invention can reduce the number of tests required to obtain regulatory approval, compared to the number of tests required for each antibody separately.

[0018] In further embodiments, the cells are homogenous populations of cells consisting essentially of copies of a single cell with nucleic acids encoding each antibody. Co-expression of several heavy chains in a cell allows various combinations of heavy chains. Combination with additional light chains increases the number of combinations. Various methods have been developed in which the formation of specific combinations is favored over others. Heavy chain variants have been produced that specifically promote the formation of heavy chain heterodimers over homodimers, or conversely, that specifically promote the formation of heavy chain homodimers over heterodimers. Heavy chains with specific homo- or heterodimerization domains reduce the number of antibodies produced by such cells and / or increase the level of preferred antibodies over alternative combinations (e.g., higher production of heterodimers over homodimers).

[0019] The method of the invention is particularly suitable for the production of two or more multispecific antibodies, including bispecific antibodies. Various methods exist in the art for the production of bispecific antibodies. One approach is to utilize a common light chain that can form a functional variable domain with different heavy chains. One preferred method for producing bispecific antibodies involves the use of a transgenic animal that has a common chain in its genome, and immunization of such an animal with an antigen results in a repertoire of antibodies specific for that antigen based on the non-common chain, the repertoire consisting of different antibodies that contain the common chain and the rearranged cognate chain. One animal can be immunized with different antigens, or different animals can be immunized separately with each antigen. Nucleic acids encoding the non-common chains or their variable regions can be obtained from the animal(s), for example from B cells, spleen, or lymphoid tissue. These can be used to generate nucleic acids that can express each non-common chain, which can then be introduced into a production cell. The common chains can be introduced at the same or different times. The nucleic acid can be integrated into the nucleus, preferably the genome of a host cell, such that the host cell produces multispecific antibodies or multimers targeting multiple antigens, for which the transgenic animal has been immunized (see, e.g., WO2009 / 157771, incorporated herein by reference, for the purpose of generating variable regions that can be combined with a common chain to produce functional variable domains that are specific for different targets and / or different epitopes).

[0020] Cells producing a common light chain and two different heavy chains, each capable of forming a functional variable domain with a common light chain, produce bispecific antibodies, especially with two different heavy and light chain combinations. Similarly, cells producing a common light chain and three or more different heavy chains can form multispecific antibodies capable of targeting three or more antigens, or a combination of two or more multispecific antibodies capable of targeting three or more antigens. It is now possible to construct the standard format of an antibody (i.e., a constant portion and two variable domains) and add additional binding domains. In this way, multispecific antibodies can be made, such as having one or more single chain Fvs with additional binding specificities linked to the constant domain or one or more variable domains of the antibody. It is also possible to produce heavy chains with two or more variable regions. The additional heavy chain regions can advantageously be associated with different or common light chain variable regions. See US62 / 650467, incorporated herein by reference.

[0021] When a cell produces two or more multispecific antibodies, it may also produce some amount of half-antibodies and antibodies with the same heavy chain or homodimers. The latter amount can be reduced by including heterodimerization that promotes heavy chain modification. As mentioned above, there are various methods to induce heavy chain heterodimerization. The respective domains with modifications are collectively called heterodimerization domains. Heavy chains with heterodimerization domains that favor interaction are said to have compatible heterodimerization domains. This compatible heterodimerization domain is preferably a compatible immunoglobulin heavy chain CH3 heterodimerization domain. The art describes various methods by which such heavy chain CH3 heterodimerization can be achieved.

[0022] One preferred method for producing bispecific antibodies is disclosed in US Patent Nos. 9,248,181 and 9,358,286. Specifically, the preferred changes for producing essentially only bispecific full-length IgG molecules are the amino acid substitutions L351K and T366K (EU numbering) in the first CH3 domain ("KK-mutant" heavy chain) and the amino acid substitutions L351D and L368E in the second domain ("DE-mutant" heavy chain), or vice versa. As mentioned above, the DE-mutant and the KK-mutant preferentially pair to form heterodimers (so-called "DEKK" bispecific molecules). Homodimerization of DE-mutant heavy chains (DEDE homodimers) or KK-mutant heavy chains (KKKK homodimers) rarely occurs, such as due to the strong repulsion between charged residues at the CH3-CH3 interface between identical heavy chains.

[0023] In the present invention, the cells are preferably provided with a nucleic acid encoding a common light chain. To generate antibodies with different heavy chain variable regions but the same light chain variable region, various methods are available to the skilled artisan. WO2004 / 106375 describes a phage library using a common light chain variable region. Phage selection results in variable domains with the same light chain variable region but different heavy chain variable regions. Furthermore, non-human animals with a common chain and different cognate chains are described in WO2009 / 157771. Antibody selection in such animals results in variable domains with the same or similar common chain variable region but different cognate chain variable regions. WO2004 / 106375 and WO2009 / 157771 are incorporated herein by reference. The publications are specifically referred to for the generation of antibodies and nucleic acids encoding such antibodies with the same or similar common chain variable region and different cognate chain variable regions, preferably common light chain variable region and different heavy chain variable regions.

[0024] In a preferred embodiment, the cell produces two or more heavy chains and a common light chain. Each heavy and light chain may have one or more variable regions associated with each chain. In a preferred embodiment, the cell produces three or more heavy chains. One of the three heavy chains preferably contains a portion of a compatible heterodimerization domain. The two or more other heavy chains preferably contain other portions of a compatible heterodimerization domain. If the first heavy chain is symbolically represented by the letter "A" and the two other heavy chains are represented by the letters "B" and "C", the specific combination of heterodimerization domains leads mainly to the formation of combinations AB and AC. The combinations AA, BB, CC, and BC are disadvantaged by the inclusion of heterodimerization domains. Such a cell effectively produces only two bispecific antibodies AB and AC (see FIG. 1). In the present invention, the cell preferably produces two antibodies by producing at least three heavy chains. In a preferred embodiment, one of the heavy chains comprises a portion of a compatible heterodimerization domain and the two or more other heavy chains preferably comprise other portions of compatible heterodimerization domains. The heavy chain shared by two or more bi- or multispecific antibodies or multimers in the composition preferably has a CH3 DE portion of the heterodimerization domain. The other heavy chains in the bispecific or multispecific antibodies preferably have a CH3 KK portion.

[0025] At least two antibodies are preferably multispecific antibodies, preferably bispecific antibodies. In a preferred embodiment, at least two of the antibodies share the same heavy chain. The cells can produce several series of bispecific antibodies by including different heterodimerization domains in the heavy chains. Such implementation can lead to the production of bispecific antibodies with heavy chain combinations AB and CD more prominent than others. The combination AB can be advantageous by DE / KK heterodimerization domains as mentioned herein above, and CD by incorporation of a "knob in hole" heterodimerization domain, or other heterodimerization features known in the art, such as charge engineering. Shared heavy chains between different bi- or multispecific multimers or antibody combinations with shared heavy chains can be made by providing a portion of the heterodimerization domain in the shared heavy chain and a complementary portion of the heterodimerization domain in the various combined chains. For example, a CH3 DE portion in the shared heavy chain and a CH3 KK portion in the combined chains. One shared heavy chain and two combined heavy chains will in this setting result in the production by the cells of bi- or multispecific multimers or antibodies with heavy chain combinations AB and AC (or AxBC and AxDE for multispecific multimers). Other possible combinations are AB, AE, CD, and CF, or AB, AE, AG, and CD, etc.

[0026] Antibodies generally have a unique isoelectric point (typically in the pH range of 6-10) compared to other host cell proteins. Antibodies, such as multispecific antibodies, and multispecific multimers can be purified to relatively high purity via the methods disclosed herein. The methods may include a number of steps, such as culturing host cells and clarification of the harvest followed by protein capture. IEX chromatography, such as anion exchange chromatography, can be used to remove host cell DNA, and cation exchange chromatography (CIEX) can be used to remove, for example, host cell proteins, leached protein A, and potential aggregates. Additional steps, such as viral filtration or hydrophobic interaction chromatography, may be included.

[0027] Antibodies are typically excreted by the producing cells. Harvesting of such antibodies typically involves recovery of the cell supernatant followed by several purification steps to remove cell debris or aggregates whose presence is not desired. Clarification of the harvest can include filtration, centrifugation, or a combination of these of the culture supernatant of the antibody producing cells. Antibody protein capture is typically performed by affinity purification. Affinity purification can be performed in several ways. In many cases, it involves purification using columns with recombinant protein A, protein G, or protein L, bacterial proteins known for their high specific binding capacity for antibodies. Currently, various optimized variants that specifically bind antibodies are available. For example, recombinant protein A is available, recombinant protein A with nonessential domains removed, recombinant protein G with deleted albumin binding sites is also available, and modified protein L is also available. Bound antibodies can be recovered by elution of one or more of such columns. Anion and cation exchange chromatography can be used to further purify the preparation, for example, to purify bispecific or multispecific antibodies from half-antibodies and / or homodimeric antibodies and / or other undesired antibody product-related impurities. Hydrophobic interaction chromatography (HIC) is commonly used as an alternative polishing step in the antibody purification process. HIC offers orthogonal selectivity to ion exchange chromatography and can be an effective step for the clearance of aggregates and the reduction of host cell proteins. In the present invention, HIC can be used for analytical purposes after purification of two or more bispecific or multispecific antibodies or multimers has been achieved, to then quantify two or more species having similar, preferably essentially the same IEX retention time and / or similar, preferably essentially the same pI. Thus, HIC is used to quantify the relative amounts of purified molecules.

[0028] A hydrophobic interaction resin is selected as the stationary phase, and the pH and / or conductivity of the mobile phase are adjusted to achieve the required selectivity. Antibodies typically attract positive charges at low pH, affecting the polarity and hydrophobicity of the overall surface. pH conditions can be selected that allow for the separation of two or more antibodies in the preparation.

[0029] In some embodiments, the recovered antibodies are first separated from other proteins by affinity purification, preferably by protein A extraction. The affinity purified antibodies are then run on an anion exchange column under conditions that recover the antibodies in the flow-through fraction. The antibodies can then be run on one or more CIEX columns.

[0030] A preferred method for producing at least two antibodies is by production by a cell as a single composition comprising two or more antibodies.

[0031] The method for producing at least two antibodies preferably includes culturing host cells producing two or more antibodies, harvesting the cell supernatant, treating the harvested supernatant in a harvest clarification process, and filtering, preferably with a pore size cutoff that traps aggregates such as cells or cell debris and allows the antibody product to pass through. The antibodies are recovered from the cell culture by affinity chromatography using Protein A. The antibodies bound to the protein resin are eluted from the chromatography column after exposure to low pH, followed by neutralization of the eluate using a suitable buffer.

[0032] As used herein, the term "isoelectric point (pI)" refers to the pH at which the average net charge on the surface of a protein, i.e., the potential of the electrical double layer of the protein, is 0. In other words, the term refers to the point at which a group of proteins dissociates such that the number of cations and anions are equal and thus the net charge of the protein is zero.

[0033] As used herein, "pi" is calculated based on the primary amino acids by ExPASy, ProtParam tool, using default parameters as of the earliest filing date (priority date) of this application. ProtParam is a tool that allows the calculation of various physical and chemical parameters for a given protein stored in Swiss-Prot or TrEMBL or a user-entered protein sequence. The calculated parameters include the theoretical pI. Gasteiger E., Hoogland C., Gattiker A., ​​Duvaud S., Wilkins MR, Appel RD, Bairoch A., Protein Identification and Analysis Tools on the ExPASy Server, (In)John M. Walker(ed):The Proteomics Protocols Handbook, Humana Press(2005) pp.571-607.

[0034] The net charge on a protein's surface changes with pH in a manner determined by the protein's pI. At a pH equal to the protein's pI, the protein has no net charge. At a pH below the pI, the protein carries a net positive charge. When the buffer pH is above the protein's pI, it carries a net negative charge.

[0035] The pI of a protein may be determined by its primary amino acid sequence and therefore can be calculated, and then a buffer can be selected that ensures a known net charge of the protein of interest. Thus, a negatively charged cation exchange resin is selected when the protein of interest carries a net positive charge at the working pH.

[0036] Proteins with different pI values ​​have various degrees of charge at a given pH, thereby having different affinities for the positively charged surface groups on the particles of the anion exchange medium. Different proteins therefore bind to the resin with different strengths, facilitating the separation of the different proteins. Thus, by generating heterodimeric polypeptides with distinct pIs compared to homodimer and half-antibody contaminants and / or other antibody product-related impurities, the heterodimeric polypeptides can be easily purified using standard elution techniques, for example, by applying a pH gradient, or by applying a salt or conductivity gradient at a constant pH, or by applying a combination of pH and conductivity gradients. In an embodiment of the present invention, the constant parts and light chains in the antibodies have essentially the same sequence in different antibodies. Such antibodies typically differ essentially only in the amino acid sequence of the heavy chain variable region. For such antibodies, it is typically sufficient to calculate the pI of the heavy chain variable region shown herein. Then, the calculation is performed and a standard is set by using the pI of the heavy chain variable region instead of the pI of the (half) antibody. The average pI of the heavy chain variable region of the antibody indicates the retention time of the antibody on the CIEX column. The antibodies can have the same or different heterodimerization domains, if present. The antibodies preferably have the same heterodimerization domains. The antibodies may further differ somewhat in their exact amino acid sequence in the constant portions.

[0037] In the ion exchange chromatography (IEX) step(s), which is a process of separating ions and polar molecules based on their affinity to an ion exchanger. It works on a variety of charged molecules, including large proteins, small nucleotides, and amino acids. IEX is often used in protein purification. Water-soluble and charged proteins form ionic bonds with an insoluble stationary phase. The bound molecules can then be eluted and recovered using an eluent with a higher concentration of ions and / or a different pH. The concentration of salt or pH can be changed stepwise by gradually changing the mobile phase of the chromatographic experiment, or by gradually changing a combination of these. Two types of ion chromatography are anion exchange and cation exchange. Cation exchange chromatography or (CIEX) is preferred in the present invention. Antibody CIEX is preferably performed at physiological pH. The pH typically ranges from 5 to 9, preferably from 6 to 8.

[0038] CIEX is typically used when the molecule of interest is positively charged at the pH used for chromatography. The molecule is positively charged because the pH of the chromatography is less than the pI of the molecule. In this type of chromatography, the stationary phase is negatively charged and positively charged molecules are loaded and attracted to it. Anion exchange chromatography is where the stationary phase is positively charged and negatively charged molecules (meaning the pH of the chromatography is greater than the pI) are loaded and attracted to it.

[0039] Antibodies, such as multispecific antibodies, are typically bound to an IEX column in a binding step under conditions that promote binding of the antibodies, such as multispecific antibodies, to the substrate. The IEX column is typically then washed to remove unbound material. Elution from the column occurs in an elution step. The retention time of an antibody, such as a multispecific antibody, is typically calculated from the start of the elution step. The retention time is the amount of time the antibody spends on the column when the elution phase begins. If a sample contains several compounds, each compound in the sample will typically spend a different time on the column according to its chemical composition, i.e., each will have a different retention time. Retention times are usually quoted in seconds or minutes.

[0040] The retention times of antibodies, such as multispecific antibodies, are preferably essentially the same in the methods of the present invention. Different antibodies may have different retention times on the same column and conditions. In the present invention, it has been found that antibodies, such as multispecific antibodies, can be selected or designed to have IEX retention times that are close enough to allow co-purification of two or more antibodies, such as two or more multispecific antibodies, in a single IEX chromatography experiment. Retention times that deviate by 10% or less from the average of the retention times of the individual antibodies are typically close enough to allow co-purification of two or more antibodies, such as two or more multispecific antibodies, in a single IEX chromatography experiment.

[0041] A suitable CIEX HPLC method for antibody purification and / or analysis according to the present invention uses TSKgel SP-STAT (particle size 7 μm, 4.6 mM ID×10 cm L, Tosoh 21964) series ion exchange columns. The columns are packed with non-porous resin particles for high speed and high resolution analysis and isolation of biomolecules. The particles in TSKgel STAT columns contain an open access network of multi-layered ion exchange groups for loading capacity while the particle size makes these columns suitable for HPLC and FPLC systems.

[0042] The preferred method involves equilibration of a TSKgel SP-STAT (particle size 7 μm, 4.6 mM ID x 10 cm L, Tosoh 21964) using buffer A (sodium phosphate buffer, 25 mM, pH 6.0), followed by displacing the antibody from the column by increasing the salt concentration and running a gradient of buffer B (25 mM sodium phosphate, 1 mM NaCl, pH 6.0). The flow rate is set at 0.5 mL / min. The injected sample mass of test samples and controls is 10 μg, with injection volumes ranging from 10 to 100 μl. The chromatograms are analyzed for peak pattern, retention time, and peak area for the major peak observed based on the 220 nm results. This method can be scaled for larger amounts of antibody.

[0043] A typical graph of a CIEX chromatography experiment of an antibody preparation is shown in FIG. 2. The antibodies for this experiment were harvested from transfected cells as shown in the examples and purified from many other proteins in the medium using a protein A column. As shown in the examples, the antibodies were eluted by acid elution followed by neutralization and buffer exchange to PBS, pH 7.4. A sample of the antibody preparation was then loaded onto the CIEX column. After washing, associated proteins were eluted by applying a salt gradient. The CIEX conditions were the same as for the samples in FIG. 2A and FIG. 2B. Retention times are calculated relative to the peak top of the bispecific antibody. The retention times of two or more antibodies, such as a multispecific antibody, preferably deviate by no more than 10% from the average of the retention times of the two or more antibodies. Deviations of more than 10% typically result in inefficient separation of the antibodies from half antibodies and possibly of the multispecific antibody from homodimers and / or other antibody product-related impurities. In a preferred embodiment, the retention times of the two or more antibodies deviate by no more than 9% from the average of the retention times of the two antibodies. Preferably, the retention times of the two or more antibodies deviate by no more than 8%, 7%, 6%, or 5% from the average of the retention times of the two or more antibodies. In a preferred embodiment, the retention times of the two or more antibodies deviate by no more than 4% from the average of the retention times of the two or more antibodies. Preferably no more than 3%, preferably no more than 2%. A more similar retention time typically allows for increasingly more efficient separation of the multispecific antibodies from half antibodies and possibly homodimers and / or other antibody product related impurities, thus allowing for cleaner recovery of the two antibodies in the fractions of the IEX column.

[0044] In the means and methods of the invention, the average retention time of two or more antibodies, such as two or more multispecific antibodies, is calculated relative to the antibodies that are co-purified or recovered. Thus, in embodiments where the purified antibody is a multispecific antibody, the average retention time of two or more antibodies is calculated based on the multispecific antibody. The retention times of non-recovered antibodies, such as homodimeric antibodies, are not used in the average calculation.

[0045] Antibodies, such as multispecific antibodies, can be selected for co-purification in the methods of the invention by selecting antibodies, such as multispecific antibodies, that have essentially the same IEX retention time under the conditions used for IEX. Antibodies, such as multispecific antibodies, can also be tailored through appropriate modification of one or more variable regions to have essentially the same IEX retention time under the same or similar conditions used for IEX.

[0046] In one embodiment, the produced antibodies that are sought to be co-purified, such as bispecific and / or multispecific antibodies, have similar pI. The isoelectric points (pI) of at least two of said antibodies preferably differ from the average pI of said at least two antibodies by no more than 0.4 units, 0.3, 0.2, preferably 0.1 units. The pI of each of said at least two antibodies preferably differs from each other by no more than 0.25 units.

[0047] Small or no difference in the pI of the antibodies typically allows for good co-purification. Advantageously, the pI of each half antibody in an antibody is more different from the average. This difference facilitates good separation of the half antibodies from the "co" migrating full antibodies in the CIEX chromatography step.

[0048] In the embodiment where the antibodies sought to be co-purified are bispecific or multispecific antibodies, the pI of each variable domain of the antibodies sought to be co-purified differs from the average pI of the variable domains of the other antibody(ies) with which it is co-purified by more than 0.2, preferably 0.3, preferably 0.4, 0.5, 0.6, 0.7, 1.0, 1.2, 1.4, preferably 1.8 or 2.0 units. In this embodiment, the difference in pI of the variable domains in the antibodies is preferably at least 0.2 units greater than the difference between "x" and "y", preferably at least 0.3, 0.4, 0.5, preferably at least 0.6, 0.7, 0.8, 0.9, 1.0, 1.2, 1.5, 2.0 or 2.5 greater than the difference between "x" and "y", where "x" is the average pI of the two variable domains of the first said antibody and "y" is the average pI of the two variable domains of the second said antibody. The difference as referred to in the pI of variable domains within an antibody typically indicates better separation of antibody product related impurities from the antibodies that are sought to be co-purified and / or from monospecific antibodies.

[0049] Some compositions comprising two or more bispecific or multispecific antibodies have constant regions and light chains with similar amino acid sequences or essentially identical amino acid sequences. Such two or more bispecific and / or multispecific antibodies typically differ from each other essentially only in the amino acid sequences of the variable domains or essentially only in the heavy chain variable regions. In such cases, it is often not necessary to determine the pI of the whole antibody. Instead, the pI of the variable domains and / or the pI of the heavy chain variable regions can be determined. This provides a means to assess whether the antibodies can migrate close to each other in the CIEX chromatography step, in other words, whether the antibodies have sufficiently similar retention times to allow co-purification.

[0050] In one embodiment of the methods or compositions disclosed herein, two or more bispecific or multispecific antibodies have constant regions and light chains with identical amino acid sequences or have essentially identical amino acid sequences. Two or more bispecific or multispecific antibodies can be co-purified in a CIEX chromatography step, when the average pI of the variable domains in each antibody differs by no more than 0.7 units from the average pI of the variable domains of each antibody that is sought to be co-purified. In a preferred embodiment, the average pI "x" of the two variable domains of a first of said antibodies and the average pI "y" of the two variable domains of a second of said antibodies differs by no more than 0.6 units, preferably no more than 0.5 units from the average "x" and "y" of the first and second antibodies that are sought to be co-purified. "x" and "y" preferably differ by no more than 0.4, preferably 0.3, preferably 0.2, preferably 0.1 units from the average "x" and "y" of the first and second antibodies that are sought to be co-purified. Such bispecific and / or multispecific antibodies typically have retention times that are essentially the same. In this embodiment, the constant regions of the antibodies are essentially the same. The pI of such antibodies, in particular the average "x" and "y", refers to the overall pI of each antibody. In this embodiment, the pI of the variable domain of each antibody that is sought to be co-purified differs from the average pI of the variable domains in the antibody by more than 0.2, preferably 0.3, preferably 0.4, 0.5, 0.6, 0.7, 1.0, 1.2, 1.4, preferably 1.8 or 2.0 units. In this embodiment, the difference in pI of the variable domains in the antibody is preferably at least 0.2 units greater than the difference between "x" and "y", preferably at least 0.3, 0.4, 0.5, preferably at least 0.6, 0.7, 0.8, 0.9, 1.0, 1.2, 1.5, 2.0 or 2.5 units greater than the difference between "x" and "y". The differences as referred to in the pI of variable domains within an antibody typically indicate good resolution of half antibodies from antibodies that are sought to be co-purified and / or good resolution of monospecific antibodies.

[0051] In one embodiment of the methods or compositions disclosed herein, two or more bispecific or multispecific antibodies have constant regions and light chains with identical amino acid sequences or have essentially identical amino acid sequences. Two or more bispecific or multispecific antibodies can be co-purified in a CIEX chromatography step, when the average pI of the heavy chain variable regions in each antibody differs by no more than 0.7 units from the average pI of the heavy chain variable regions of each antibody that is sought to be co-purified. In a preferred embodiment, the average pI "m" of the two heavy chain variable regions of a first of said antibodies and the average pI "n" of the two heavy chain variable regions of a second of said antibodies differs by no more than 0.6 units, preferably no more than 0.5 units from the average of the first and second antibodies "m" and "n" that are sought to be co-purified. "m" and "n" are preferably 0.4, preferably 0.3, preferably 0.2, preferably 0.1 units from the average of the first and second antibodies "m" and "n" that are sought to be co-purified. Such bispecific and / or multispecific antibodies typically have retention times that are essentially the same. In this embodiment, the constant regions of the antibodies are essentially the same. The pI of such antibodies, in particular the average "m" and "n", refers to the overall pI of each antibody. In this embodiment, the pI of the heavy chain variable region of each antibody that is sought to be co-purified differs from the average pI of the heavy chain variable regions in the antibodies by more than 0.2, preferably by more than 0.3, preferably by more than 0.4, 0.5, 0.6, 0.7, 1.0, 1.2, 1.4, preferably by more than 1.8 or 2.0 units. In this embodiment, the difference in pI of the heavy chain variable regions in the antibodies is preferably at least 0.2 units greater than the difference between "m" and "n", preferably at least 0.3, 0.4, 0.5, preferably at least 0.6, 0.7, 0.8, 0.9, 1.0, 1.2, 1.5, 2.0 or 2.5 units greater than the difference between "m" and "n". The differences as referred to in the pI of the heavy chain variable regions within an antibody typically indicate good separation of half antibodies from antibodies that are sought to be co-purified and / or good separation of monospecific antibodies.

[0052] Antibodies, such as multispecific antibodies, can have or can be selected to have heavy and light chain combinations (half antibodies) or homodimers (e.g., monospecific antibodies) or other antibody product-related impurities that have a significantly different retention time than the retention time of the desired antibody, such as a full antibody or a multispecific antibody, under the IEX conditions used. In embodiments where the cells express a common light chain, the selection is typically on the heavy chain. The heavy chain can be engineered such that the half antibodies or homodimers have very different retention times. In preferred embodiments, the retention times of the half antibodies and / or homodimers differ by more than 10% from the average retention times of the respective antibodies or multispecific antibodies. In preferred embodiments, the average pI of the individual heavy and light chain combinations of antibodies that are not required to be co-purified differs by more than 0.5 units from the average pI of said heavy and light chains of at least two antibodies that are co-purified.

[0053] Furthermore, the present invention provides a composition comprising 2 to 10 recombinant antibodies obtainable by the methods described herein, characterized in that at least two of the antibodies have essentially the same IEX retention time.

[0054] Furthermore, the present invention provides a composition comprising 2 to 10 recombinant antibodies, characterized in that the pI of at least two of said antibodies differs from the average pI of said at least two antibodies by no more than 0.4 units, 0.3, 0.2, and preferably no more than 0.1 units, and preferably no more than 0.25 units.

[0055] The present invention further provides a composition comprising 2 to 10 recombinant antibodies, characterized in that the average pI "x" of the two variable domains of the first antibody and the average pI "y" of the two variable domains of the second antibody differ from the average of "x" and "y" of the first and second antibodies that are sought to be co-purified by 0.7, 0.6 and preferably 0.5 units or less. "x" and "y" preferably differ from the average of "x" and "y" of the first and second antibodies that are sought to be co-purified by 0.4, preferably 0.3, preferably 0.2, preferably 0.1 units or less. Antibodies such as such multispecific antibodies typically have retention times that are essentially the same. In this embodiment, the constant regions of the antibodies are essentially the same. The pI, particularly the average of the pIs of the different variable domains of the antibody, including the heavy and light chain variable regions, respectively, indicates the overall pI of the antibody.

[0056] The present invention further provides a composition comprising 2 to 10 recombinant antibodies, characterized in that the average pI "m" of the two heavy chain variable regions of the two variable domains of the first antibody and the average pI "n" of the two heavy chain variable regions of the two variable domains of the second antibody differ from the average of "m" and "n" of the first and second antibodies that are sought to be co-purified by 0.7, 0.6 and preferably 0.5 units or less. "m" and "n" are preferably 0.4 and preferably differ from the average of "m" and "n" of the first and second antibodies that are sought to be co-purified by 0.3, preferably 0.2, preferably 0.1 units or less. Antibodies such as such multispecific antibodies typically have retention times that are essentially the same. In this embodiment, the constant region and the light chain variable region of the antibody are essentially the same. The pI, particularly the average of the pIs of the different heavy chain variable regions of the antibody, indicates the overall pI of the antibody.

[0057] In a preferred embodiment, the IEX retention time and / or pI is essentially identical for all antibodies recovered in the composition. In a preferred embodiment, at least two of the antibodies are bispecific antibodies. Preferably, at least two of the antibodies share the same heavy chain.

[0058] In some embodiments, the common light chain variable region of one or both VH / VL joining regions is germline IgVk1-39. * In certain embodiments, the light chain variable region of one or both VH / VL joining regions comprises the kappa light chain V segment IgVκ1-39. * 01. IgVκ1-39 is the short form for immunoglobulin variable kappa 1-39 gene. This gene is also known as immunoglobulin kappa variable 1-39, IGKV139, IGKV1-39. The external Ids for this genetic element are HGNC:5740, Entrez Gene:28930, Ensembl:ENSG00000242371. The amino acid sequence of the V region is provided in SEQ ID NO:25. The V region can be combined with one of five J regions. The preferred J regions are jk1 and jk5, with the combined sequences designated IGKV1-39 / jk1 and IGKV1-39 / jk5. Alternative names include IgVκ1-39 * 01 / IGJκ1 * 01 or IgVκ1-39 * 01 / IGJκ5 * 01 (nomenclature according to the worldwide web of the IMGT database at imgt.org). In a specific embodiment, the light chain variable region of one or both of the VH / VL joining regions is the kappa light chain IgVk1-39 * 01 / IGJk1 * 01 or IgVk1-39 * 01 / IGJk1 * 05 (SEQ ID NO:26 and SEQ ID NO:27, respectively).

[0059] In some embodiments, one or both light chain variable regions of the VH / VL binding region of the bispecific antibody comprise an LCDR1 comprising the amino acid sequence QSISSY (SEQ ID NO: 22), an LCDR2 comprising the amino acid sequence AAS, and an LCDR3 comprising the amino acid sequence QQSYSTP (SEQ ID NO: 24) (i.e., the CDRs of IGKV1-39 according to IMGT). In some embodiments, one or both light chain variable regions of the VH / VL binding region of the bispecific antibody comprise an LCDR1 comprising the amino acid sequence QSISSY (SEQ ID NO: 22), an LCDR2 comprising the amino acid sequence AASLQS (SEQ ID NO: 23), and an LCDR3 comprising the amino acid sequence QQSYSTP (SEQ ID NO: 24).

[0060] In some embodiments, one or both of the VH / VL binding regions of the bispecific antibody comprise a light chain variable region comprising an amino acid sequence that is at least 90%, preferably at least 95%, more preferably at least 97%, more preferably at least 98%, more preferably at least 99% identical or 100% identical to the amino acid sequence set forth in SEQ ID NO: 26. In some embodiments, one or both of the VH / VL binding regions of the bispecific antibody comprise a light chain variable region comprising an amino acid sequence that is at least 90%, preferably at least 95%, more preferably at least 97%, more preferably at least 98%, more preferably at least 99% identical or 100% identical to the amino acid sequence set forth in SEQ ID NO: 27.

[0061] For example, in some embodiments, the variable light chain of one or both VH / VL binding regions of the bispecific antibody may have 0 to 10, preferably 0 to 5, amino acid insertions, deletions, substitutions, additions, or a combination thereof relative to SEQ ID NO: 26 or SEQ ID NO: 27. In some embodiments, the light chain variable region of one or both VH / VL binding regions of the bispecific antibody comprises 0 to 9, 0 to 8, 0 to 7, 0 to 6, 0 to 5, 0 to 4, preferably 0 to 3, preferably 0 to 2, preferably 0 to 1, preferably 0 amino acid insertions, deletions, substitutions, additions, or a combination thereof relative to the amino acid sequence shown.

[0062] In other embodiments, one or both light chain variable regions of the bispecific antibody VH / VL binding regions comprise the amino acid sequence of SEQ ID NO: 26 or SEQ ID NO: 27. In certain embodiments, both VH / VL binding regions of the bispecific antibody comprise identical VL regions. In one embodiment, both VLs of the bispecific antibody VH / VL binding regions comprise the amino acid sequence set forth in SEQ ID NO: 26. In one embodiment, both VLs of the bispecific antibody VH / VL binding regions comprise the amino acid sequence set forth in SEQ ID NO: 27.

[0063] Bispecific antibodies such as those disclosed in the methods herein can be provided in several formats. Many different formats of bispecific antibodies are known in the art. For example, bispecific antibody formats that are not conventional antibodies with two VH / VL combinations have at least a variable domain that includes a heavy chain variable region and a light chain variable region. This variable domain may be linked to a single chain Fv fragment, a monobody, a VH, and a Fab fragment that provides a second binding activity.

[0064] Bispecific antibodies, such as those disclosed in the methods provided herein, are generally of the human IgG subclass (e.g., IgG1, IgG2, IgG3, IgG4). In certain embodiments, the antibodies are of the human IgG1 subclass. Full-length IgG antibodies are preferred due to their advantageous half-life and reduced immunogenicity. Thus, in certain embodiments, the bispecific antibody is a full-length IgG molecule. In one embodiment, the bispecific antibody is a full-length IgG1 molecule.

[0065] In certain embodiments, the antibody comprises a fragment crystallizable (Fc). The Fc of the bispecific antibody is preferably composed of a human constant region. The constant region or Fc of the bispecific antibody may contain one or more, preferably no more than 10, preferably no more than 5, amino acids different from the constant region of a naturally occurring human antibody. For example, in certain embodiments, each Fab arm of the bispecific antibody may further comprise an Fc region that includes modifications that facilitate bispecific antibody formation, modifications that affect Fc-mediated effector functions, and / or other features described herein.

[0066] In one aspect, a pharmaceutical composition is provided that includes two or more antibodies as defined herein and a pharma- ceutically acceptable carrier. As used herein, the term "pharmaceutically acceptable" means approved by a government regulatory agency or listed in the United States Pharmacopeia or other generally recognized pharmacopoeias for use in animals, particularly humans, and includes any and all physiologically compatible solvents, salts, dispersion media, coatings, antibacterial and antifungal agents, isotonic and absorption delaying agents, and the like. The term "carrier" refers to a diluent, adjuvant, excipient, or vehicle with which the compound is administered. Such pharmaceutical carriers can be sterile liquids, such as water and oils, including those of petroleum, animal, vegetable, or synthetic origin, for example, peanut oil, soybean oil, mineral oil, sesame oil, glycerol polyethylene glycol ricinoleate, and the like. Water or saline solutions, as well as aqueous dextrose and glycerol solutions, can be used as carriers, particularly for injectable solutions. Liquid compositions for parenteral administration can be formulated for administration by injection or continuous infusion. Routes of administration by injection or infusion include intravesical, intratumoral, intravenous, intraperitoneal, intramuscular, intrathecal, and subcutaneous. Depending on the route of administration (e.g., intravenous, subcutaneous, intraarticular, etc.), the active compound can be coated with a material to protect it from the action of acids and other natural conditions that may inactivate the compound.

[0067] Pharmaceutical compositions suitable for administration to human patients are typically formulated for parenteral administration, for example, in a liquid carrier, or suitable for reconstitution into a liquid solution or suspension for intravenous administration. The compositions can be formulated in dosage unit form for ease of administration and uniformity of dosage.

[0068] Also included are solid form preparations which are intended to be converted, shortly before use, to liquid form preparations for either oral or parenteral administration. Such liquid forms include solutions, suspensions, and emulsions.

[0069] A "bispecific antibody" is an antibody as described herein in which one domain of the antibody binds to a first antigen while a second domain of the antibody binds to a second antigen, the first and second antigens being non-identical. The term "bispecific antibody" also encompasses antibodies in which one heavy chain variable region / light chain variable region (VH / VL) combination binds to a first epitope on an antigen and a second VH / VL combination binds to a second epitope. The term further includes antibodies in which the VH can specifically recognize a first antigen and the VL paired with the VH in the immunoglobulin variable region can specifically recognize a second antigen. The resulting VH / VL pair binds to either antigen 1 or antigen 2. Such so-called "two-in-one antibodies" are described, for example, in WO 2008 / 027236, WO 2010 / 108127, and Schaefer et al (Cancer Cell 20, 472-486, October 2011). The bispecific antibodies according to the invention are not limited to any particular bispecific format or method of producing them.

[0070] As used herein, when referring to nucleic acid or amino acid sequences, "percent (%) identity" is defined as the percentage of residues in a candidate sequence that are identical to the residues in a selected sequence after aligning the sequences for optimal comparison purposes. Percent sequence identity for comparing nucleic acid sequences is determined using the AlignX application of Vector NTI Program Advance 10.5.2 software using default settings, which use a modified ClustalW algorithm (Thompson, JD, Higgins, DG, and Gibson TJ (1994) Nuc. Acid Res. 22:4673-4680), a swgapdnarnt scoring matrix, a gap open penalty of 15, and a gap extension penalty of 6.66. Amino acid sequences were aligned using the AlignX application of Vector NTI Program Advance 11.5.2 software using default settings, which employs a modified ClustalW algorithm (Thompson, JD, Higgins, DG, and Gibson TJ 1994), a blosum62mt2 scoring matrix, a gap open penalty of 10, and a gap extension penalty of 0.1.

[0071] The term "common light chain" as used herein refers to the two light chains (or VL portions thereof) in a bispecific antibody. The two light chains (or VL portions thereof) may be identical or may have some amino acid sequence differences, although the binding specificity of the full-length antibody is not affected. The terms "common light chain", "common VL", "single light chain", "single VL", with or without the addition of the term "rearrangement", are all used interchangeably herein. "Common" refers to functional equivalents of light chains that are not identical in amino acid sequence. Many variants of the light chain exist, and mutations (deletions, substitutions, insertions and / or additions) exist that do not affect the formation of a functional binding region. The light chain of the present invention may also be a light chain with 0 to 10, preferably 0 to 5, amino acid insertions, deletions, substitutions, additions or combinations thereof, as specified herein above. It is within the scope of the definition of a common light chain used herein to prepare or find a light chain that is not identical but is still functionally equivalent, for example by introducing and testing conservative amino acid changes, changes in amino acids in regions that do not contribute or only partially contribute to binding specificity when paired with a heavy chain, etc. The term "full-length IgG" or "full-length antibody" according to the present invention is defined as including essentially a complete IgG, but not necessarily having all the functions of a complete IgG. For the avoidance of doubt, a full-length IgG comprises two heavy chains and two light chains. Each chain contains a constant (C) region and a variable (V) region, which can be broken down into domains designated CH1, CH2, CH3, VH and CL, VL. IgG antibodies bind to antigens via the variable region domains contained in the Fab portion, and after binding can interact with molecules and cells of the immune system via the constant domains, primarily the Fc portion. A full-length antibody according to the present invention encompasses IgG molecules in which there may be mutations that provide the desired properties. A full-length IgG should not have a deletion of a substantial portion of any of the regions. However, IgG molecules in which one or several amino acid residues have been deleted without essentially altering the binding properties of the resulting IgG molecule are included in the term "full length IgG".For example, such an IgG molecule may have a deletion of 1 to 10 amino acid residues, preferably in the non-CDR regions, where the deleted amino acids are not essential for the binding specificity of the IgG.

[0072] An antibody typically recognizes an epitope of an antigen, which may be present in other compounds as well, and an antibody according to the invention that "specifically recognizes" an antigen may also recognize other compounds as well, if such other compounds contain the same type of epitope. Thus, the term "specifically recognize" in relation to the interaction of an antigen and an antibody does not exclude the binding of the antibody to other compounds that contain the same type of epitope.

[0073] The term "epitope" or "antigenic determinant" refers to a site on an antigen to which an immunoglobulin or antibody specifically binds. Epitopes can be formed both from contiguous amino acids juxtaposed by tertiary folding of a protein (so-called linear and conformational epitopes) or from non-contiguous amino acids. Epitopes formed from contiguous linear amino acids are typically retained on exposure to denaturing solvents, while epitopes formed by tertiary folding typically lose conformation on treatment with denaturing solvents. Epitopes may typically include 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14 or 15 amino acids in a unique spatial conformation. Methods for determining the spatial conformation of epitopes are known to those of skill in the art and include techniques in the art, such as x-ray crystallography, HDX-MS and 2-dimensional nuclear magnetic resonance, pep scan, and alanine scan, depending on the nature of the epitope (see, e.g., Epitope Mapping Protocols in Methods in Molecular Biology, Vol. 66, GE Morris, Ed (1996)).

[0074] Although features are described herein as part of the same or separate embodiments for clarity and conciseness of description, it will be understood that the scope of the invention may include embodiments having all or a partial combination of the described features.

[0075] In order to make this specification easier to understand, certain terms are defined at the beginning. Additional definitions are described throughout the detailed description. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art, and conventional methods of immunology, protein chemistry, biochemistry, recombinant DNA technology and pharmacology are employed.

[0076] As used herein, the singular forms "a," "an," and "the" include plural referents unless the context clearly dictates otherwise. Use of the term "including," as well as other forms such as "include," "includes," and "included," is not limiting. [Brief description of the drawings]

[0077] [Figure 1] Schematic diagram of an embodiment in which the composition comprises two bispecific antibodies sharing a common arm. The diagram shows an antibody having a heavy chain (1) and a light chain (4). The four heavy chains have three different variable regions (5, 6, and 7). The heavy chains having a shared variable region (5) have a portion of a heterodimerization domain (3). The heavy chains having variable regions (6) and (7) have compatible portions of the heterodimerization domain (2). The preferred pairing of the heterodimerization regions (2) and (3) can lead to the formation of a bispecific antibody. [Figure 2A] Panel A: CIEX profile at 220 nm of bispecific antibody PB4516 production number 8 (p08). [Figure 2B] Panel B: CIEX profile at 220 nm of bispecific antibody PB6892 production number 4 (p04). [Figure 3A]FIG. 13 is a CIEX profile at 220 nm of bispecific antibody PB4516 production number 10 (p10). [Figure 3B] CIEX profile at 220 nm of bispecific antibody PB11244 production number 1 (p01). [Figure 3C] Two bispecific antibodies per box are identified in column PB (PBXXXX(X)). The heavy chain variable regions of PB11244 and PB4516 are shown in columns target 1 and target 2. The sequence of the light chain is the same for all antibodies, with the amino acid sequence of the common light chain IgKV1*39 / jk1 in SEQ ID NO: 26. The pI calculated with ExPASy, ProtParam tool, is shown for each heavy chain variable region in column pI. The pI difference between the two heavy chain variable regions is shown in the last column, demonstrating that the average VH pI difference between PB11244 and PB4516 is 0.08. [Figure 4] Figure 2 is a CIEX profile at 220 nm of an antibody preparation of individual colony cp12 of pool FST2. The CIEX profile shows sharp peaks for the co-eluting antibodies PB4516 and PB11244. The profile shows that the sample contains limited amounts of product-related impurities. It also shows good separation between the co-eluting bispecific antibody and the separately migrating product-related impurities. [Diagram 5] CIEX profile at 220 nm of antibody preparation of colony CP07. Colonies were picked from a collection of individual colonies of single colony FST2cp09. A second subcloning was performed to confirm that the FST2cp09-cp07 cell line is a clonal cell line. Bispecific antibody specific ELISA showed the presence of EGFR / HER2 bispecific antibody PB11244 at 743 μg / ml and EGFR / HER3 bispecific antibody PB4516 at 1134 μg / ml. [Figure 6] Retention time of antibody homodimer (PGXXXX) with two identical variable domains. The amino acid sequence of the heavy chain variable region has the sequence shown in MF of FIG. 8 and common light chain IgKV1*39 / jk1 in SEQ ID NO: 26. [Figure 7]Two bispecific antibodies per box are identified in the PB column (PBXXXX(X)). The heavy chain variable region (MFXXXX) of each bispecific antibody is shown in the Target 1 and Target 2 columns. The sequence of the light chain is the same for all antibodies, with the amino acid sequence of the common light chain IgKV1*39 / jk1 of SEQ ID NO: 26. The pI calculated with ExPASy, ProtParam tool, is shown for each heavy chain variable region in the pI column. The measured pI difference between the two heavy chain variable regions and the retention time are shown in the last two columns. The measured retention times and the calculated pI and average pI indicate that the bispecific antibody binding can be efficiently co-eluted in CIEX chromatography. The antibodies have an IgG1 constant region and a common light chain. The heavy chains with a shared heavy chain variable region (same MF) with CH3 DE heavy chain or KK heavy chain are shown. For CIEX chromatography performed with each bispecific, retention times are shown using the exemplary conditions in the Materials and Methods section described for CIEX chromatography. Many other CIEX chromatography conditions will result in suitable retention times between the listed bispecific antibody pairs with the pI values ​​provided. [Figure 8-1] The amino acid sequence of the heavy chain variable region (MFXXXX) of each antibody and the CDR of the light chain variable region, and the amino acid sequence of the common light chain variable region. [Figure 8-2] The amino acid sequence of the heavy chain variable region (MFXXXX) of each antibody and the CDR of the light chain variable region, and the amino acid sequence of the common light chain variable region. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0078] Working Example Example 1 Materials and Methods cell line HEK293 and CHO-K1 were maintained in Growth Medium.

[0079] Generation of bispecific antibodies Bispecific antibodies were generated using proprietary CH3 technology to ensure efficient heterodimerization and bispecific antibody formation, which uses charge-based point mutations in the CH3 region to allow efficient pairing of two different heavy chain molecules, as previously described (PCT / NL2013 / 050294, WO 2013 / 157954A1).

[0080] The VH genes were cloned into one of two different backbone IgG1 vectors. Depending on the binding partner, the VH was cloned into an IgG1 backbone containing a CH3 variant with the heterodimerization variant "DE" or into an IgG1 backbone containing the complementary CH3 heterodimerization variant "KK". In the case of bispecific or multispecific antibodies, two or more antibodies share a heavy chain. The shared chain preferably has the CH3 heterodimerization variant "DE" (also called DE heavy chain) and two or more unique heavy chains have the CH3 heterodimerization variant "KK" (also called KK-heavy chain).

[0081] HEK293 cells were transiently transfected with DNA-FUGENE mix and further cultured. Seven days after transfection, the supernatant was harvested and the medium was refreshed. 14 days after transfection, the supernatants were combined and filtered at 0.22 μM. The sterile supernatant was stored at 4°C. Suspension-adapted 293F cells were cultured in T125 flasks on a shaking plateau to a density of 3.0×10e6 cells / ml. Cells were seeded at a density of 0.3–0.5×10e6 viable cells / ml in each well of a 24-deep well plate. Cells were transiently transfected with individual sterile DNA:PEl-MIX and further cultured. Seven days after transfection, the supernatant was harvested and filtered at 0.22 μM. The sterile supernatant was stored at 4°C.

[0082] Generation of a stable cell line pool co-expressing two bispecific antibodies CHO cells were transfected with the three heavy chain constructs and the common light chain construct at a molar ratio of common light chain construct (cLC):EGFR heavy chain:HER2 heavy chain:HER3 heavy chain=2.5:2:1:1. Ten pools of stably transfected cells were obtained (AJ). ELISA analysis of anti-EGFR, anti-HER2, and anti-HER3 antibodies was performed on day 3 and day 6 supernatants of the ten pools. All three could be detected in all pools.

[0083] Generation of stable cell line clones co-expressing two bispecific antibodies Pools were plated on semi-solid medium and grown for 7–10 days. Single colonies were picked and seeded into 24-well culture plates. Colonies were replated before antibody was harvested from the culture supernatant.

[0084] Determination of antibody titers Anti-HER2 antibody titers of samples containing single bispecific antibodies were determined by ELISA against Erbb-2 Fc protein (R&D system). Anti-HER3 antibody titers of samples containing single bispecific antibodies were determined by ELISA against human Erbb-3-Fc protein (R&D system). Anti-EGFR antibody titers of samples containing single bispecific antibodies were determined by ELISA against human EGFR ECD-Fc protein (R&D system). Two-fold serial dilutions of antigen were used to coat the wells of an ELISA plate, starting at 5 μg / ml.

[0085] An ELISA assay for quantifying EFGRxHER2 and EGFRxHER3 bispecific antibodies in a composition containing a mixture of two bispecific antibodies was performed by coating an ELISA plate with EGFR-Fc (R&D system). After washing, the plate was incubated with the sample. After washing, the presence of bound bispecific antibodies with one EFGR arm and one HER2 arm was detected by incubating with labeled HER2-Fc. The presence of bound bispecific antibodies with one EFGR arm and one HER3 arm was detected by incubating with labeled HER3-Fc.

[0086] IgG purification Purification of IgG was performed using affinity chromatography. Purification was performed under sterile conditions using vacuum filtration. The pH of the medium was first adjusted to pH 8.0, and then the product was incubated with Protein A Sepharose CL-4B beads (50% v / v) (Pierce) for 2 hours at 25°C at 600 rpm on a shaking platform. The beads were then harvested by vacuum filtration. The beads were washed twice with PBS pH 7.4. IgG was eluted with 0.1 M citrate buffer at pH 3.0, and the IgG fraction was immediately neutralized with Tris pH 8.0. Buffer exchange was performed by centrifugation using Ultracel (Millipore). The sample was finally buffered in PBS pH 7.4.

[0087] Cation-Exchange Chromatography (CIEX) CIEX-HPLC chromatography was performed using TSKgel SP-STAT (particle size 7 μm, 4.6 mM ID×10 cm L, Tosoh 21964) series ion exchange columns. The columns are packed with non-porous resin particles for high speed and high resolution analysis and isolation of biomolecules. The particles in TSKgel STAT columns contain an open access network of multi-layered ion exchange groups for loading capacity while the particle size makes these columns suitable for HPLC and FPLC systems.

[0088] Buffer A (sodium phosphate buffer, 25 mM, pH 6.0) is used to equilibrate a TSKgel SP-STAT (particle size 7 μm, 4.6 mM ID x 10 cmL, Tosoh 21964), followed by transfer of the antibody from the column by increasing the salt concentration and running a buffer B (25 mM sodium phosphate, 1 mM NaCl, pH 6.0) gradient. The flow rate was set at 0.5 mL / min. The injected sample mass for all test samples and controls (in PBS) was 10 μg, and the injection volume was 10-100 μl. The chromatograms were analyzed for peak pattern, retention time, and peak area for the major peaks observed based on the 220 nm results.

[0089] result The CIEX profiles of bispecific antibodies PB4516p08 and PB6892p04 were compared (see FIG. 2). It was observed that the production of PB6892 contained a significant amount of impurities. Also, the retention time of the bispecific antibody fraction of PB6892 was significantly lower than that of the bispecific antibody fraction of PB4516. In case of simultaneous production by the same cells and co-purification using CIEX, the retention times of the two bispecific antibodies are preferably close to each other. For this reason, the variable region of the HER2 arm of PB6892 was replaced with a different variable region. A heavy chain with the variable region MF2032 was selected and used to produce the EGFR×HER2 bispecific antibody PB11244. The CIEX profiles of PB4516p10 and PB11244p01 are shown in FIG. 3. The retention times of the bispecific antibody fraction are 16.310 and 16.950, respectively. These retention times are sufficiently equal to allow co-purification using CIEX under the conditions shown. In addition, the figure shows that the retention times of the impurities are sufficiently different to allow efficient separation on analytical and preparative columns.The bispecific antibody preparations were produced in HEK293 cells.

[0090] For co-production, CHO-K1 cells were used. CHO cells were transfected with a construct containing three heavy chains with the variable regions of MF3755 (EGFR), M2032 (HER2) and MF3178 (HER3), respectively, and transfected into CHO-K1 cells together with the light chain variable region expression construct of SEQ ID NO: 26. Vector positive cells were selected and pooled. Ten separate pools of transfected CHO-K1 cells (identified as A-J) were generated.

[0091] Table 1 shows the amount of bispecific antibody PB4516 (EGFRxHER3) and PB11244 (EGFRxHER2) produced by each pool. The ratio and total amount of IgG produced are also shown. Pools F and J were selected for subcloning.

[0092] Table 2 shows the amount of bispecific antibodies PB4516 (EGFRxHER3) and PB11244 (EGFRxHER2) produced by each clone.

[0093] The CIEX profile was analyzed using antibodies produced by clone FST2cp12 (see FIG. 4). It is clear that the two bispecific antibodies efficiently co-elute in the same CIEX elution fraction.

[0094] Clone FST2cp09 was further subcloned to confirm the clonality of the cell line and to determine further CIEX profiles of the produced antibodies. Figure 5 shows the CIEX profiles. It is clear that the two bispecific antibodies efficiently co-elute in the same CIEX elution fraction. The relative contribution of the two bispecific antibodies in the co-elution is analyzed by ELISA and / or hydrophobic interaction column. Bispecific antibody specific ELISA showed the presence of EGFR / HER2 bispecific antibody PB11244 at 743 μg / ml and EGFR / HER3 bispecific antibody PB4516 at 1134 μg / ml.

[0095] Example 2 Generation of a stable cell line pool co-expressing two bispecific antibodies Cell lines expressing the two bispecific antibodies listed in Figure 7 are generated as follows: CHO cells are transfected with the three heavy chain constructs and the common light chain construct. The three heavy chains are identified by the heavy chain variable region (MFXXXX) shown in the box. The light chain is IgVk1 of SEQ ID NO: 26. * 39 / jk1. The two bispecific antibodies each have one heavy chain in common and one different heavy chain. For example, the first binding designated in FIG. 7 shares a common heavy chain that includes the same HER3 binding arm that includes a heavy chain variable region (MF3178) and a different second binding arm. PB4528 has an EGFR binding arm with a heavy chain variable region (MF4003), and PB4188 has a HER2 binding arm with a heavy chain variable region (MF3958). The shared heavy chain has a KK CH3 region of compatible DE / KK heterodimerization domains. The shared heavy chain arm may also have a DE CH3 region. For example, in FIG. 3-5, the two bispecific antibodies are co-purified PB11244 and PB4516. As shown in Figure 3c, PB11244 and PB4516 share the same EGFR binding arm with a heavy chain variable region (MF3755), PB11244 has a HER2 binding arm with a heavy chain variable region (MF2032), and PB4516 has a HER3 binding arm with a heavy chain variable region (MF3178). The shared heavy chain arm of this bispecific antibody pair has a DE CH3 region, while the different HER2 and HER3 binding arms have KK CH3 regions.

[0096] The molar ratio of common light chain construct (cLC):shared heavy chain construct:different heavy chain construct 1:different heavy chain construct 2 is 2.5:2:1:1. A pool of stably transfected cells is obtained. Antigen ELISA analysis is performed on the supernatant harvested from the pool. All three antigen binding species are detected in the pool. The CIEX retention times of the bispecific antibodies in each binding in Figure 7 are determined under similar CIEX conditions and are 8 thThe deviation from the average retention time is calculated by the formula 100x((AB) / (A+B)), where A is the retention time of the bispecific antibody with the longest retention time. For example, the deviation of the first bond is 100x((16.46-16.24) / (16.46+16.24))=0.67, i.e. 0.7%.

[0097] The antibodies in the collected supernatant are first separated from other proteins in the supernatant by protein A extraction, followed by acid elution and rapid neutralization. The collected antibodies are then buffer exchanged into PBS. The sample is then loaded onto a CIEX column and washed and eluted by increasing salt gradient. The absorbance of the eluate is measured at 220 nm and the retention time is calculated from the start of the salt gradient and the observation of the bispecific antibody peak. The bispecific antibodies are collected and each bispecific antibody in the collected eluate is verified by ELISA. The retention time of each bispecific antibody is shown in the last column. It is clear that many of the conjugates have retention times that efficiently co-elute on the CIEX column. It is also clear that CIEX chromatography provides good separation of the co-eluting bispecific antibodies and their respective homodimers (if present). Figure 6 shows the retention times of various antibodies with homodimers of heavy chains containing heavy chain variable regions present in the co-eluting bispecific antibodies. It is clear that the retention times of the homodimers are sufficiently different from the retention times of the respective bispecific antibodies. For example, in the first box of Figure 7, homodimers PG3178, PG3958 and PG4003 may be present in the product. The retention times of the respective homodimers are approximately 22, 12 and 13 (Figure 6, rows 1-3), and the retention times of the bispecific antibody containing the heavy chain variable region are approximately 19 and 19.5 (Figure 7, rows 1 and 2).

[0098] [Table 1] Table 1: Quantification of EGFRxHER2 and EGFRxHER3 bispecific antibody production in pools of cells. Culture supernatants of 10 pools (A-J) were evaluated. The ELISA assay is based on EGFR-Fc coating, binding of the produced antibodies, and detection of either labeled HER2-Fc or labeled HER3-Fc. Pools A-J were analyzed using two ELISA assays. Bispecific antibodies PB4516 and PB11244 are IgG1 heavy chain antibodies with compatible DE / KK heterodimerization domains. The heavy chains combine with a common light chain. The bispecific antibodies share the MF3755 heavy chain variable region on one heavy chain and have different heavy chain variable regions on the other IgG1 heavy chain, MF3178 for PB4516 and MF2032 for PB11244, respectively.

[0099] [Table 2] Table 2: Selected pools were used for single cell cloning. 18 colonies were picked from 3 pools. Two independent pools F (FST1 and FST2) and one pool J (JST1) were used for single cell cloning. The indicator "cp" followed by a number identifies the individual colonies of the pool. Picked colonies were expanded and used for antibody recovery. Individual colonies from the same pool produced different amounts and different ratios of each bispecific antibody.

Claims

1. 1. A method for producing at least two heterologous antibodies, comprising: The antibodies include at least one monospecific antibody and at least one multispecific antibody; providing a cell with nucleic acids encoding said monospecific and multispecific antibodies; Culturing the cells; and recovering the monospecific and multispecific antibodies from the culture; and and separating the produced monospecific and polyspecific antibodies from half antibodies by ion exchange chromatography (IEX), 1. A method for producing at least two heterologous antibodies, characterized in that each of said monospecific and multispecific antibodies exhibits an IEX retention time that deviates by no more than 10% from the average of the retention times of said individual monospecific and multispecific antibodies under the IEX conditions used.

2. 2. The method of claim 1, further comprising selecting at least two monospecific and multispecific antibodies that have essentially the same IEX retention time under the conditions used for IEX.

3. 3. The method of claim 1 or 2, wherein recovering the antibody from the culture comprises purifying the antibody from other proteins by antibody affinity purification.

4. 4. The method of claim 1, wherein recovering the antibody from the culture comprises purifying the antibody from other proteins by Protein A extraction.

5. 4. The method of claim 3, further comprising subjecting the affinity purified antibody to size exclusion chromatography, gel filtration chromatography and / or anion exchange chromatography.

6. 6. The method of claim 1, wherein following IEX, the recovered antibodies are quantitatively analyzed for relative expression levels by hydrophobic interaction chromatography (HIC).

7. The method of claim 6, wherein the specificity of the recovered antibodies is verified by ELISA.

8. 8. The method of claim 1, wherein the retention time of each half antibody is outside the range spanned by the retention time of the antibody.

9. The method of claim 8 , wherein the cell produces three heavy chains.

10. The method of claim 9, wherein the heavy chain comprises a domain for efficient heterodimerization of the heavy chain.

11. 11. The method of any one of claims 1 to 10, wherein the antibody has an isoelectric point (pi) that differs by no more than 0.4 units from the average pI of the at least two antibodies.

12. 12. The method of any one of claims 1 to 11, wherein the antibody is selected to have a heavy and light chain combination that has a retention time significantly different from the retention time of the whole antibody under the IEX conditions used.

13. 13. The method of claim 12, wherein the isoelectric point (pi) of the combined heavy and light chain differs by more than 0.4 units from the average pI of the at least two antibodies.

14. 14. The method of any one of claims 1 to 13, wherein the multispecific antibody has a heavy chain that comprises a CH3 domain that assists in heterodimerization of the heavy chain.

15. The method of any one of claims 1 to 14, wherein the antibody has an IgG heavy chain.

16. 16. The method of any one of claims 8 to 15, wherein the cells produce two or more heavy chains, each comprising a CH3 region, and one heavy chain comprises amino acid substitutions L351K and T366K (EU numbering) in the CH3 region and another heavy chain comprises amino acid substitutions L351D and L368E in the CH3 region.

17. 1. A method for producing at least two heterologous antibodies, comprising: The antibodies include at least one monospecific antibody and at least one multispecific antibody; providing a cell with nucleic acids encoding said monospecific and multispecific antibodies; Culturing the cells; and recovering the monospecific and multispecific antibodies from the culture; and and separating the produced monospecific and polyspecific antibodies from half antibodies by ion exchange chromatography (IEX), 1. A method for producing at least two heterologous antibodies, characterized in that each of said monospecific and multispecific antibodies exhibits an IEX retention time that deviates by no more than 10% from the average of the retention times of said individual monospecific and multispecific antibodies under the IEX conditions used, and wherein following said IEX, the recovered monospecific and multispecific antibodies are quantitatively analyzed for relative expression levels by Hydrophobic Interaction Chromatography (HIC) and the specificity of the recovered monospecific and multispecific antibodies is verified by ELISA.

18. 18. The method of any one of claims 1 to 17, wherein the IEX retention time is the amount of time each antibody spends on the column, calculated from the start of the elution step to the top of the elution peak of the individual antibody.

19. 19. The method of any one of claims 1 to 18, wherein the cell produces i) a common light chain, and ii) two or more IgG heavy chains, each comprising a CH3 region, wherein one heavy chain comprises amino acid substitutions L351K and T366K (EU numbering) in the CH3 region, and another heavy chain comprises amino acid substitutions L351D and L368E in the CH3 region.

20. 1. A composition comprising 2 to 10 heterogeneous recombinant antibodies, said antibodies comprising at least one monospecific antibody and at least one multispecific antibody, wherein at least one of the monospecific antibodies and at least one of the multispecific antibodies have an IEX retention time that deviates by no more than 10% from the average of the retention times of the individual monospecific and multispecific antibodies under said IEX conditions.

21. 21. The composition of claim 20, wherein the IEX retention time is the amount of time each antibody spends on the column calculated from the start of the elution step to the top of the elution peak of the individual antibody.

22. 1. A composition comprising 2 to 10 heterogeneous recombinant antibodies, said antibodies comprising at least one monospecific antibody and at least one multispecific antibody, characterized in that at least one of the monospecific antibodies and at least one of the multispecific antibodies has an isoelectric point (pI) that differs by no more than 0.4 units from the average pI of said at least two antibodies.

23. 23. A composition according to any one of claims 20 to 22, characterized in that the IEX retention time and / or the pI are essentially the same for all of the antibodies.

Images