Bi-specific antibodies

Bispecific antibodies can bind two different antigens. IgG type antibodies have two binding sites with different variable regions. An IgG variable region is made up of a variable light-chain sequence (VL) and a variable heavy chain sequence (VH). The light chains (LCs) of common LC antibodies are identical for both variable regions, leaving the heavy chain (HC) for generating different specificites. Thus, recombinant host cells for production of common LC bispecific antibodies carry genes for both hCs, with different specificites (A and B) along with one LC gene. A, B and the light cahins are expressed independently in those host cells, which then assemble them into three IgG types –AA, AB, and BB –for secretyion into a culture. By purefuly random assembly, the three types should be produced in a ratio of 1:2:1 (AA, AB, and BB). Genetic engineering methods can modify heavy cahin Fc regions such that A and B are assembled prereably, thus reducing the formation of AA and BB. (Bakker “purifying common light-chain bispecific antibodies: a twin-column, countercurrent chromatography platform process) BioProcess International 11(5): 36-44 (2013).

Bispecific antibodies offer opportunities for increasing specificity, broadening potency, and utilizing novel mechanisms of action that cannot be acheived with traditional mAb. Cross-linking two different receptors using a bispecific antibody to inhibit a signalling pathway has shown utility and in one example a cell surface tyrosine phosphatase was recruited into an EgE receptor complex to decrease activity of the phosphorylated IgE receptor. This approach was more effective than blocking the ligand binding site because inhibition of signalting by the bispecific antibody occured even in the presence of high cocnetrations of ligand. The use of bispecific antibodies has also shown application in recruting cytotoxic T cells where T cell activation was acheived in proximity to tumor cells by the bispecific antibody binding receptors simultanteously on the two different cell types. (Scheer, US 2013/0017200)

Paticular Types of Bi-specific Antibodies that are Purified

kappa-lamda antibodies:

–Using KappaSelect and LambdaSelect Resins:

An appealing bispecific format for therapeuti use is an unmodified human IgG. This format would share stability, pharmacokinetic and other sought-after drug-like properties of therapeutic mAbs, while enabling novel modes of action. An approach previously treid was to co-express the H and L chains of two different antibodies in a single cell. However, the random assembly of the fourt chains resulted in a complex mixture of ten molecules, substantially challenging development from yield, cost and purity perspectives. A more selective approach is to use antibodies that share a common chain such that concomitant expression of two heavy and a common light chain in the same cell results in a mixture containing only two mAbs and one BiAb. However, the downstream purificaiton of the BiAb from this mixture is challening and relies of differences between the physicochemical properites (for example, overall charge or hydrophobicity) of teh BiAb and the two mAbs. A completely different approach is to create in vitro display libraries with a common heavy chain that are used to select against two different antigens. This allows the idolation of candidates with different target specificties that share the same heavy chain but carry either a kappa or lambda light chains. Three different chains (one heavy and two light) are then co-epxressed in a single cell to generate a mixutre containing two mAbs speies (one kappa and one lambda) and a BiAb constaining a kappa and lambda light chain. A BiAb assembled in this manner can then be efficiently purified form the mAb species using highly selective affinity resins binding to either human kappa or lambda constant domains. Base on its structure, this fully human BiAb format is referred to a s a “kappa-lambda” body. (Fischer, “exploting light chains for the scalable generation and platform puriicaiotn of native human bispecific IgG” Nature Communications, 2015).

—-Protein A – Kappa Select –Lambda Select

Fischer (US2014/0179547; see also US Patent 16/042,889, published as 2019/0031714 and US 2012/0184716 (producing three or more monospecific antibodies and three or more bispecific antibodies). ) discloses a method of purifying a bispecific bivalent antibody that shares the same heavy chain (antibodies bearing both a kappa and a lambda light chain) by the use of affinity chromatography such as CaptureSelect Fab Kappa and CaptureSelect Fab Lambda that specifically interacts with the Kappa or Lambda light chain constant domains constant domains. The method uses a three step procedure: (1) Protein A to capture IgG (mono and bi), (2) Kappa select capture IgG containing a Kappa light chain(s) and (3) Lambda select to capture IgG containing a Lambda light chain.

Fischer “exploiting light chains for teh scalable generation and platform purificaiton of native human bispecific IgG” Nature communications, 2014) discloses assembly and purification of kappa/lambda-bodies using a three step affinity process. First, a non-distinguishing affinity capture step was perforemd using either a protein A or Capture Select CH1 risn that binds to the Fc or CH1 doamin of human IgG, respecitvely. The second affinity-capture purificaiton using KappaSelect binds to human kappa constant region and allows for recovery of the IgGk and the kappa/lambda body, while the IgGlambda is eliminated in the flow through. Third, the kappa/lambda body is purified by a LambdaFabSelect affinity rsin that binds selectively to human lambda constant region, eliminating the IgGkappa species.

—-Protein A- Lambda Select –Kappa Select:

Hultberg (US 2012/0156206) disclsoes purification of camelid-derived bispecific cMET antibodies using the three step purificiaotn method of ProtA sepharose to select for only properly assembled Mabs containing two variable heavy and light chainsand then Labda-select and Kappa-select baes to separate the parental Mabs from the bispecific Mabs.

—-Differential elution of bi-specific kappa/lambda antibody:

Elson (US13/655,955, published as US 2013/0317200) discloses purification of a kappa-lamda-antibody (termed a kappalamda-body) which consists of a common IgG1 heavy chain and two different light chains that drive specificity for two independent targets. One of the light chains contains a kappa constant region while the other contains a lamda constant region. Purification of the kappa-lamda antibody can be performed by sequential binding to KappaSelect and LambdaFabSelect affinity resins which have specificity and affinity for either the kappa or lamda constant region. The kappa-lamda body binds to either KappaSelect or LambdaFab Select resins with a weaker affinity than the monospecific kappa-Mab (for KappaSelect) or monospecific lamda-MAb (for LambdaFabSelect) due to the fact that it contains only one of each light chain rather than two for the monoclonal format. In one embodiment Elson discloses using a higher pH step elution to preferentially elute a bispecific kappa-lamda-body from a KappaSelect affinity resin over monospecific kappa-MAB which elutes at a lower pH, as the monospecific MAb presumably has a higher affinity to the resin owing to the presence of two kappa chains in the monospecific format as opposed to a single kappa chain in the kappa-lampda body. For example, after column loaiding at 10 mg/mL and wash step with 50mM Sodium Phosphate, 250 mM Sodium Chloride, pH 7.0 (5 column volumes), a pH step-elution (pH 3.0 follwed by pH 2.5 and pH 2.0) was performed using a 50mM glycine buffer adjusted to the relevant pH. The bispecific kappa-lamda body product preferentially elutes at the ihger pH step.

–Using Mixed Mode and HIC Resins

Elson (US13/655,955, published as US 2013/0317200; see also 16/032,873, published as US 2019/0040117) also disclose reducing free light chains from the kappa-lamda antibodies using Mep HyperCel mixed mode chromatography by applying a cell culture supernatant to the resin, eluting the mAb with an acetate buffered elution buffer at pH 5.0 (eluate) and removing free light cahins which are strongly bound to the resin at pH 2.1.

Fourque (US15/068916, published as US 2016/02646485, now US Patent No: 10,457,749; see also US Patent Application 16/601,121, published as US 2020/0181287) discloses a method for purifying a bispecific antibody having a first light chain with a kappa constant region and a second light chain with a labda constant region (kappa/lamba-body) from a mixture using a mixed mode resin such as TOYOPEARL MX-Trp 60M and HIC such as a TOYOPEARL Butyl 600M resin.

For different H and L antibody chains (“crossMab”):

–Cation Exchange:

Neumann (US 2014/0081000) discloses that recently a new bispecific bivalent antibody form called the “CrossMab” which is a bispecific antibody having four different antibody chains, i.e., two different light chains and two different heavy chains has been described (Proc. Natl. Acad. Sci, USA 108 (2011) 11187-11192). Neumann discloses that during the expression different product related impurities such as the 3/4 antibody, an antibody that is missing one light chain, displaying a MW of about 125 kDa compared to a MW of 150 kDa for a comlet four chain antibody is formed. Neumann discloses a method for purifying the full lenght four chain antibody with CEX by applying a solution comprising a non-ionic polymer and an additive to which the antibody has been adsorbed whereby the four chain full lenght antibody in monomeric form remains adsorbed to the CEX and then recoveirng the four chain full lenght antibody in monomeric form by applying a solution comprising a non-ionic polymer such as PEG, an additive such as zwitterions, and an elution compound (see also purificaiton of antibodies by CEX).

VH3 domain containing bispecific antibody

A subset of antibodies bind Protien A in the variable region of the heavy chain. In particular, IgGs that contain H chains derived from the VH3 gene family have been shown to belong to this fmaily. Becasue of this, on SpA-based resins, many bsAbs exhibit poor resolution of the two bidning species as well as retention of the non-bnding Fc*Fc* parental antibody. This is becasue VH binding reduces the avidity difference between the bispecific and the FcFc parental and the Fc*Fc* parental is also retained by VH binding. This pehnomenon can be resolved through the use of the alkali stabilized MabSelect SuRe ligand. This is because whereas all five domains of SpA (E, D, A, B, and C) binding antibodies via the Fc region, only domains D and E exhibit significant Fab binding. As the MabSelect SuRe ligand is a tetramer of the Z-domain, a protein-engineered verison of thenative SpA B domain, it would be expected to lack significant fab binding, and this has been demonstrated in multiple studies. (Tustian, Biotechnol. Prog., 2018, 34(3))

–Protein A chromatography:

Haywood (US 15/566,231, published as US 2018/0100007) discloses purification of a human VH3 domain containing antibdoy such as an Fab, Fab’, F(ab’)2, Fv and scFv as well as such types of antibodies that include more than one VH3 domain such as those in the format of diabodies, tetrabodies, minibodies, and domain antibodies. using Protein A chromatography where the VH3 domain containing antibody is recovered in monomeric form.

Types of Purification Techniques for Purification of Bi-specific Antibodies

Protein A/G or Combination:

–By modifying bi-specific antibody to create differential binding

Blein (US 14/431207, published as US 2015/0239991; see also US Patent Application 16/275821, published as US 20200010568) discloses a method for the purification of a hetro-dimeric immunoglobulin from a mixture where the immunoglobulin has a modification in a H1 and/or a CH2 and/or a CH3 region to reduce or eliminate binding to Protein G, applying the mixture to Protein G and eluting the hetero-dimeric immunoglobulin from the resin. In a second embodiment, the modification is in the VH3 region and the mixture is applied to Protein A and the hetero-dimeric immunoglobulin fragment is eluted from Protein A. In one embdoiment,t he amino acid substitution is 81E or 82aS. In a third embodiment, the modification is in a first heavy chain of the hetero-dimer that reduces or eliminates binding to a first affinity reagent and there is a modification in the second heavy chain that reduces or eliminates binding to a second affinity reagent, the mixture is applied to the first affinity reagent and hetero-dimers are eluted and this eluate is applied to a second column with a second affinity reagent and the hetero-dimers are eluted. When the first affinity reagetn is Protein A, the seond is Protein G or vice versa.

Gramer (US 2014/0303356) discloses purifcation of a heterodimeric protein from at least a first and second homodimeric proteins using Protein A/G. In one embodiment, either the first or second homodimeric protein may have been engineered/ modified so taht it does not bind Protein A/G. This facilitates separation of the homodimeric protein which does not bind to ProteinA form the hterodimeric protein.

Igawa (EP 2522724 A1) discloses methods for purifying multi specific antibodies by altering amino acid residues in the H chain constant and/or variable region so as to result in altered protein A binding which also exhibit plasma retention comparable or longer than that of human IgG1. In one embodiment the amino acid substtiution is from the the amino aicd residues of positions 250-255, 308-317 and 430-436 in the Fc domain or antibody have chain constant region.

Lindhofer (US 5,945,311) discloses a method for producing a bi-secific antibody by providing a quadroma fused from hybridomas of which one produces a first antibody which has an affinity to the binding domain of protein A where the antibodies are of the subclass IgG1, IgG2, IgG4 or rat antibodies of the subclass IgG2c and the other hybrodoma produces a second antibody which in comparision with the first antibody has a smaller or no affinity to the binding domain of the protein A the second antibodies being rat antibodies of the subclass IgG1, IgG2a, IgG2b, IgG3 or IgG3. The quadorma culture is applied to protein A, washed and the bi-specific antibodies are eluted in a pH range 5.6-6.01 which is at least 0.5 units above the pH at which the antibodies with greater affintiy to the binding domain of prtoein A are still bonded.

Zhang (WO 2017/034770) discloses a multispecific heterodimeric antibody that includes a first and second H chain wherein the CH3 of the first H chain includes at least two amino acid modifications positively charged amino acids and the CH3 domain of the second H chain includes at least 2 amino acid mutations negatively charged amino acids. The increased negative charges in the short H chain CH3 domain does not bind to protein A whereas the positivley charged long chain Fc is capable of binding to protein A. As a result, short chain homodimers will flow through the column as they no longer bind to protein, the homodimers haivng two long chain will bind strongly to protein wheras the ehterodimers will be to protien A with lower affinity due to the fact that only long cahin Fc provides bining to protein A. By eluting the column with an elution solution having a pH of 4 or above, the herodimer will be eluted out, but the long chain homidmers will not be eluted out due to their stronger binding to the column.

Yeung (WO2010/075548) discloses variant antibodies with one or more amino acid modification in the VH region that have altered binding to Staphylococcus aureus protein A and methods of using such antibodies.

—-CH3 of IgG3 in second H chain or His435Arg subsitution in CH3 region in second H chain:

Examples of the Fc region include human IgG type Fc that may be of the IgG1, IgG2, IgG3 and IgG4 isotypes. The “Fc region” refers to the hinge porition, CH2 domain and CH3 domain in an antibody. According to EU numbering by Kabat, a Human IgG class Fc region refers to the region from cysteine at position 226 to teh C temrinus or from porline at position 230 to teh C terminus. The human CH2 domain refers to positions 231-340 and the CH3 domain refers to positions 341-447. Tanaka (Chugai, Tokyo, US Patent application 16/061,454, published as US 2019/0330268)

Davis (US 8,586,713 and 2010/0331527) discloses antigen binding proteins such as bispecific antibodies that have IgG CH2 and CH3 regions with different affinities for Protein A that allows rapid isolation by differential binding of the IgG regions to Protein A. In one embodiment, differentially modified heterodimeric human IgG2 and unmodified human homodimeric IgG2 were first enriched by a bind and wash process through a protein A column and then a step gradient elution was performed. In one embodiment, the bispecific antibody includes a first and second polypeptide, the first polypeptide having from N to C terminal a first antigen binding region that binds a first antigen followed by a constant region that includes a first CH3 region of human IgG selected from IgG1, IgG2, IgG4 anda combination thereof and a second ABP that seelectively binds a second antigen followed by a constant region that includes a second CH3 region of a human IgG selected form IgG1, IgG2, IgG4 and a combination thereof wherein the second CH3 region comprises a modificaiton such as a 435R modification that reduces or elimiantes binding of the second domain to Protein A. Davis disloses that the inability of IgG3 to bind Protein A is determined by a single amino acid residue, Arg435 (EU numbeirng: Arg95 by IMGT), which corresponding position in the other IgG subclasses is occupied by a histidine residue. It is therefore possible, instead of IgG3, to use an IgG1 sequence in which His435 is mutated to Arg. This single point mtuation in IgG1 should be sufificent to create the different binding affinities. The point mutation could, in theory, be potentially immunogenic. To avoid this pitfall, a dipeptide mutation H435R/Y436F (EU numbering) can be used. The resulting sequence in the vicinity of the alteration is identical to that of IgG3 and would thus be expected to be immunoglogically invisible because there would be non non-native short peptides available for presentation to T cells. It has been reported that this double mutant still does not bind Protein A. Finally, the dipeptide mutation does not include any of the residues that form the Fc dimer interface, so it is unlikely to interfere with the formation of heterodimers.

Shitar, (US 2007/0148165) discloses that the binding activity to protein A is decreased when His 435 in the human Igg1 heavy chain constant region is replaced with Arg derived from IgG3.

Tustian (US 14/808,171, published as US2106/0024147) discloses methods of purifying a heterodimeric protein such as a bispecific antibody by using affinity capture such as Protein A and elution processes at a particular pH range. The mixture of multimeric proteins contains (i) a first homodimer comprising two copies of a first polypeptide, (ii) a heterodimer comprising the first polypeptide and a second polypeptide and optionally (iii) a second homodimer comprising two compies of the second polypeptide. The first and second polypepides have different affinities for the affinity matrix such that the first homodimer, the heterodimer and second homidimer can be separated on the basis of differential binding to the affinity matrix. Differential binding can be manipulated by changing the pH and/or ionic strenght. In addition, a chaotropic agent such as CaCl2 or MgCl2 can be used to enhance the elution of each dimer species. In one embodiment the bispecific antibody contains a single common light chain and two distinct heavy chains and one of the heavy chains contains a substituted Fc sequence that reduces or eliminates binding of the Fc* to Protein A such as an Fc* sequence that contains H435R/Y436F substitutions in the CH3 domains. As a result, three products are creates; two of which are homodimeric for the heavy chains and one of which is the desired heterodimeric bispecific produc. The Fc* sequence allows selective purification of the FcFc* bispecific product on the affinity column due to intermedite bindng affinity for Protein A compared to the high avidity FcFc heavy chain homodimer or the weakly binding Fc*Fc* homodimer.

——For increasing the Dynamic Binding Activity of bi-specific antibody:

Binding capacities include statis binding capacity (SBC) and dynamic binding capacity (DBC). SBC refers to the upper limit of the amount of polypeptides that a resin can adsorb, and DBC refers to the degree to which polypeptides can be collected wehn a polypeptide containing solution is flowing through the column. A resin having a large dynamic binding capacity allows efficient polypeptide adsorption even under high linear flow rate, and polypeptide purification can be accomplished in a short time. DBC is determined by depicting the behavior in which a continuosuly loaded protein is discharged from the column as a breakthrough curve in a chromatogram by UV monitoring using a purification device connected to a UV detector. DBC can be determined by loading a column with a resin and allowing a polypeptide containing sample solution to flowthorugh the column at a specified linear flow rate. Then, the absorbance of the eluate is measured, and DBC is determined by identifying the mass of the added polypeptide when breakthrough (BT) for a specific proportion (for example, 5%) of absorbance of the added sample solution is measured. Specifically, a calculation method when using 5% BT is (1) the load fraction (IgG concentration: P g/L) is allowed to flow through the LC apparatus without passing it thorugh the column and the value of OD280 nm for 100% leakage (= 100% BT) is confirmed. This value is denoted as “a”. (2) the value obtained by multiplying 0.05 to a is defined as the OD280nm at 5% BT. This value is denoted as “b5%”. (3) the load fraction is allowed to flow continously through a set amount of equilibrated resin (r L), and when the OD280 nm value reaches by 5%, teh volumen of the load fraction is read form the chromatogram. This valued is denoted as “C5% L”. (4) the value obtained by the equation (P x c5%)/r is calculate as DBC5% which is the DBC at 5% BT. DBC5%=(PxC5%)/r (unit: g/L resin). When determining DBC10% the calculation is possible by determining c10% in a similar manner (Tanaka, Chugai, Tokyo, US Patent application 16/061,454, published as US 2019/0330268)

Tanaka (Chugai, Tokyo, US Patent application 16/061,454, published as US 2019/0330268) discloses a method for increasing the dynamic binding capacity of a bi-specific antiody for a Protein A resin which includes preparing a first polypeptide chain that binds to the Protein A resin and a second polypeptide that does not bind to the resin or shows weaker binding. In one embodiment, the Fc region includes a CH3 of IgG1, IgG2 or IgG4 and the second polypeptide of the Fc region includes a CH3 of IgG3. In one embodiment, position 435 of the first polypetpide chain is modified to be His and position 435 of the second polypetpide chain is modified to be Arg. Examples of other modifications include modifying positions 435 and 436 of the first chain to be His(H) and Tyr and positions 435 and 436 of the second chain to be Arg (R) and Phe (F). The increase in the DBC of the Fc region containing polypeptide for a Protein A resin is at least 5, 15, 20 or even 25 g/L resin or mroe when taking 5% BT as the standard.

–Using low conductivity wash

Falkenstein (US 15/900, 449, published as US 20180186866) discloses that HCP can be reduced if the conductivity of the aqueous solution used in the wash step is low (below 0.5 mS/cm). Advantageously futher process steps can be obviated before loading hte eluate to the next chromatographi material if a low conductivity aqueous solution wash step is ued in the preceeding affintiy chromatography step.

–Protein A – Mixed Mode (AEX or CEX) – Mixed Mode (AEX or CEX):

Giese (US Patent Applicaiton 16/221,369, published as US 2019/0256556) discloses a method of purifying a bi-specific/heterodimeric polypeptice with a multi-step chromatography method which includes affinity chromatography such as Protein A followed by two different multimodal ion exchange chromatography steps., such as a multimodal AEX followed by a multimodal CEX or vice versa. In one embodiment, a bispecific antibody against target proteins X1 and Y1 (anti-X1/anti-Y1 bispecific antibody) was assembled as follows. Each half-antibody (aX1 (knob) and a Y1 (hole) was independently subject to an affinity chromatography step using prtoein A resin.The half antibody pools obtained from the protein A chromatography step were then combined in a 1:1 molar ration, ( the pH was adjusted to pH 82. L-gltathione buffer was added). The resulting assembled pool was cooled and pH adjusted to pH 5.5 and then subjected to a multimodal CEX using Capto MMC resin in a bind and elute mode. Eluate form the multimodal CEX was then subject to multimodal AEX using Capto Adhere resin in a flow-through mode which removed residual impuriteis like DNA, host cell protein and endotxoins as well as product variants including half-antibodies, homodimers and aggregates.

IgG-CH1 Affinity Matrix

Fischer (WO/2013/136186) discloses using two immunoglobulin CH1 heavy chain constant domain sequences that differ by at least one amino acid in a bispecific antigen-binding protein. The amino acid difference results in an improved ability to isolate the biapecific protein form homodimers beause the difference results in a differentail ability of teh CH1 domain to bind the CaptureSelect® IgG-CH1 (BAC BV) affinity reagent. According to the invention, the bispecific antibody includes a first and a second polypeptide, the first polypetpide comprising from N-temrinal to C-temrinal a first antigen binding region that selectively binds a first antigen, followed by a constant region that includes a first CH1 region of a human IgG and a second polypeptide comprising from N-temrinal to C-terminal, a second antigen-binding region that selectively binds a second antigen, followed by a constant region that includes a second CH1 reigon of a human IgG wherein the second CH1 region includes a modificaiton that reduces or eliminates binding of the second CH1 domain to the CaptureSelect IgG-CH1 affinity reagent.

Cation Exchange

Igawa (US 2009/0263392) discloses purificaiton of bipspecific antibodies that includes modifying the amino acids present on the variable framework regions of two types of polypeptides that constitute the bispecific antibody so that they may be separated as by CEX based on their differening isoelectric points of the H chains.

Lebanon (US 15/565,494, published as US 2018/0079797) discloses a strong cation exchange for the separation of bispecific antibodies using various linear pH gradeints witha starting buffer A: CAPS, CHES, TAPS, HEPPSO, MOPSO, MES, acetic acid, hormic acid and NaCL, pH 4 and a final buffer composed of the same agents at pH 11.

Scheer (US 2011/0287009) disclsoes purificaiton of a bispecific antibody by weak CEX using a carboxymethyl resin with a pH gradient elution from 4.5 to 9.2 The buffer A and B consisted of sodium citrate, MES, HEPES, imidizole, Tris, CAPS and NaCL where the A buffer is adjusted to pH 4.2 with HCl and the B buffer is adjusted to pH 9.2 using NaOH.

Hydrophobic Interaction Chromatography:

Manzke (J Immunological Methods, 1992, 208(1), pp. 65-73) discloses a method for large scale production and single step purificaiton of bispecific antibodies using HIC.

Hydroxapatite:

–Conditions:

—-Calcium ion (NaCL or KCL) Gradients:

Bertl (US 15/048308, published as US 2016/0376304) discloses a method of purfying a bispecific antibody using HA where the bispecific antibody is eluted with a buffer having chloride ions such as one having NaCL or KCL. By increasing concetnration of sodium chloride in the presence of phosphate, the bispecific antibody was separated from unwanted species.

—-Phosphate gradients:

Ford (J Chromatography B , 754 (2001) 427-435) disclsoes purificaotn of bispecific antibodies recognising carcinoembryonic antigen and doxorubicin using Protein A affinity chromatography and HPLC hydroxyapatite affinity chromatography. Elution was with a 60-360 mM phosphate buffer gradient.

Tarditi (J Chromatography 599 (1992) 13-20) disclsoes isloation of bispecific monoclonal antibodies (bi MAbs) using Protein A followed by a high performance hydroxyapatite (HPHT) column. Purification was performed on the HPHT at a low rate of 0.4 ml/min using 10 mM phosphate buffer (pH 6.8) as buffer A and 350 mM phosphate buffer (pH 6.8 as buffer B). Gradients were selected after evaluating the retention times of parental MAbs. The general gradient rage was 0-80% buffer B. The next step was a linear scale-up of selected gradients. The HPHT chromatography has unique selectivity for IgG idiotypes and separated IgGs according to their net charge. As the active bi MAb secreted by one hybrid hybridoma is formed of two IgG moieties, each corresponding to half of the parental IgG, this combination will have a total net charge intermedaite between those of the parental MAbs. Thus, the biMAb will be eluted form the HPHT column with an ionic strenght intermedaite between those of the parental MAbs.

Reduction – Oxidation –Sequential Affinity:

Cariring (“A novel redox method for rapid production of functional bi-specific antibodies for use in early pilot studies” 6(7), 2011) discloses a chemical reduction-oxidation (redox) method for the production of purified bsAbs in a fraction of the time taken by traditional hybrid hybridoma technology by using the mild reducing agent 2-mercaptoethansulfonic acid sodium salt (MESNA) followed by dialysis under oxidizing conditions in order to allow the antibodies to reform. During this reaciton a mixture of antibodies is formed, including parental antibodies and bi-specific antibody. Bi-specific antibodies are purified over two sequential affinity columns. For example. bsAbs from different species (rat/mouse hybrid bsAbs) were purified first over an anti-rat IgG and secondonly an anti-mouse IgG affinity column. Cariring demonstrated that a similar schedule could be used to make bsAbs using two different antibodies form the same species and subclass. In order to purify these bsAbs the parental antibodies were conjugated to biotin or dinitrophenol prior to reduction. Purificaiton of the bsAbs was carried out by sequential purificaiton on anti-biotin and anti-DNP affinity columns.

Twin-Column Multicolumn Countercurrent Solvent Gradient purificaiton (MCSGP):

NCSGP is a chromatographyic process that allows for product isoaltion with high yeild and purity. In many cases (e.g., bispecific antibody isolation), the difficulty is caused by product-related ipurities taht elute closely to the product of interest. Often in preparative batch chromatography, the overlap remains even after reasonable elution-doncition optimization. The twin-column NCSGP process can be beneficial in through internal recylcing of the impure-side fractions. Those frations containing overlapping product and impurities are washed from one column inot the other, where the product is readsorbed and conserved. In phase I, the column are interconnected, and the chromatographic fraction containing the overlapping reigon of weakly adsorbing impurites and product elutes form the upstream into the downstream column. In between those two columns, the stream is diluted inline to ensure readsorption of eluted product on the downstream column. In phase 2, the columns are disconnected and pure product elutes form the upstream column. The overlapping region of product and strongly adsorbing impurity remains in the upstream column. In parallel, feed solution, is loaded onto the downstream column. In phase 3, the columns are again interconnected, and the chromatographic fraction containing the overlapping region is washed form the upstream into the downstream column. Again, product adsorption in the downstream column is ensured by inline dilution of the stream exiting the upstream column. Ini phase 4, the columns are disconnected again, and the strongly adsorbing impuriteies are washed out of the upstream column. Taht column is tripped, cleaned and equilibrated. In parallel, weakly adsorbing impurites elute form teh downstream column as elution is initated. After this last phase, the columns switch position so that the (equilibrated) upstream column becomes the downstream column and vice versa. Phases 1-4 repeat with the columns in this reverse order. On finishing phase 4 and another column-position switch, the columns have returend to their original position, so one cycle of the twin-column process is complete. continues to run over multiple cycles. (Bakker “purifying common light-chain bispecific antibodies: a twin-column, countercurrent chromatography platform process) BioProcess International 11(5): 36-44 (2013).